cpython/Modules/_threadmodule.c

2652 lines
77 KiB
C
Raw Normal View History

/* Thread module */
/* Interface to Sjoerd's portable C thread library */
1996-12-16 20:05:22 -04:00
#include "Python.h"
#include "pycore_interp.h" // _PyInterpreterState.threads.count
#include "pycore_lock.h"
#include "pycore_moduleobject.h" // _PyModule_GetState()
#include "pycore_modsupport.h" // _PyArg_NoKeywords()
#include "pycore_pylifecycle.h"
#include "pycore_pystate.h" // _PyThreadState_SetCurrent()
#include "pycore_sysmodule.h" // _PySys_GetAttr()
#include "pycore_time.h" // _PyTime_FromSeconds()
#include "pycore_weakref.h" // _PyWeakref_GET_REF()
#include <stddef.h> // offsetof()
#ifdef HAVE_SIGNAL_H
# include <signal.h> // SIGINT
#endif
// ThreadError is just an alias to PyExc_RuntimeError
#define ThreadError PyExc_RuntimeError
// Forward declarations
static struct PyModuleDef thread_module;
// Module state
typedef struct {
PyTypeObject *excepthook_type;
PyTypeObject *lock_type;
PyTypeObject *local_type;
PyTypeObject *local_dummy_type;
PyTypeObject *thread_handle_type;
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
// Linked list of handles to all non-daemon threads created by the
// threading module. We wait for these to finish at shutdown.
struct llist_node shutdown_handles;
} thread_module_state;
static inline thread_module_state*
get_thread_state(PyObject *module)
{
void *state = _PyModule_GetState(module);
assert(state != NULL);
return (thread_module_state *)state;
}
// _ThreadHandle type
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
// Handles state transitions according to the following diagram:
//
// NOT_STARTED -> STARTING -> RUNNING -> DONE
// | ^
// | |
// +----- error --------+
typedef enum {
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
THREAD_HANDLE_NOT_STARTED = 1,
THREAD_HANDLE_STARTING = 2,
THREAD_HANDLE_RUNNING = 3,
THREAD_HANDLE_DONE = 4,
} ThreadHandleState;
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
// A handle to wait for thread completion.
//
// This may be used to wait for threads that were spawned by the threading
// module as well as for the "main" thread of the threading module. In the
// former case an OS thread, identified by the `os_handle` field, will be
// associated with the handle. The handle "owns" this thread and ensures that
// the thread is either joined or detached after the handle is destroyed.
//
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
// Joining the handle is idempotent; the underlying OS thread, if any, is
// joined or detached only once. Concurrent join operations are serialized
// until it is their turn to execute or an earlier operation completes
// successfully. Once a join has completed successfully all future joins
// complete immediately.
//
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
// This must be separately reference counted because it may be destroyed
// in `thread_run()` after the PyThreadState has been destroyed.
typedef struct {
struct llist_node node; // linked list node (see _pythread_runtime_state)
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
// linked list node (see thread_module_state)
struct llist_node shutdown_node;
// The `ident`, `os_handle`, `has_os_handle`, and `state` fields are
// protected by `mutex`.
PyThread_ident_t ident;
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
PyThread_handle_t os_handle;
int has_os_handle;
// Holds a value from the `ThreadHandleState` enum.
int state;
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
PyMutex mutex;
// Set immediately before `thread_run` returns to indicate that the OS
// thread is about to exit. This is used to avoid false positives when
// detecting self-join attempts. See the comment in `ThreadHandle_join()`
// for a more detailed explanation.
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
PyEvent thread_is_exiting;
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
// Serializes calls to `join` and `set_done`.
_PyOnceFlag once;
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
Py_ssize_t refcount;
} ThreadHandle;
static inline int
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
get_thread_handle_state(ThreadHandle *handle)
{
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
PyMutex_Lock(&handle->mutex);
int state = handle->state;
PyMutex_Unlock(&handle->mutex);
return state;
}
static inline void
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
set_thread_handle_state(ThreadHandle *handle, ThreadHandleState state)
{
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
PyMutex_Lock(&handle->mutex);
handle->state = state;
PyMutex_Unlock(&handle->mutex);
}
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
static PyThread_ident_t
ThreadHandle_ident(ThreadHandle *handle)
{
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
PyMutex_Lock(&handle->mutex);
PyThread_ident_t ident = handle->ident;
PyMutex_Unlock(&handle->mutex);
return ident;
}
static int
ThreadHandle_get_os_handle(ThreadHandle *handle, PyThread_handle_t *os_handle)
{
PyMutex_Lock(&handle->mutex);
int has_os_handle = handle->has_os_handle;
if (has_os_handle) {
*os_handle = handle->os_handle;
}
PyMutex_Unlock(&handle->mutex);
return has_os_handle;
}
static void
add_to_shutdown_handles(thread_module_state *state, ThreadHandle *handle)
{
HEAD_LOCK(&_PyRuntime);
llist_insert_tail(&state->shutdown_handles, &handle->shutdown_node);
HEAD_UNLOCK(&_PyRuntime);
}
static void
clear_shutdown_handles(thread_module_state *state)
{
HEAD_LOCK(&_PyRuntime);
struct llist_node *node;
llist_for_each_safe(node, &state->shutdown_handles) {
llist_remove(node);
}
HEAD_UNLOCK(&_PyRuntime);
}
static void
remove_from_shutdown_handles(ThreadHandle *handle)
{
HEAD_LOCK(&_PyRuntime);
if (handle->shutdown_node.next != NULL) {
llist_remove(&handle->shutdown_node);
}
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
HEAD_UNLOCK(&_PyRuntime);
}
static ThreadHandle *
ThreadHandle_new(void)
{
ThreadHandle *self =
(ThreadHandle *)PyMem_RawCalloc(1, sizeof(ThreadHandle));
if (self == NULL) {
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
PyErr_NoMemory();
return NULL;
}
self->ident = 0;
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
self->os_handle = 0;
self->has_os_handle = 0;
self->thread_is_exiting = (PyEvent){0};
self->mutex = (PyMutex){_Py_UNLOCKED};
self->once = (_PyOnceFlag){0};
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
self->state = THREAD_HANDLE_NOT_STARTED;
self->refcount = 1;
HEAD_LOCK(&_PyRuntime);
llist_insert_tail(&_PyRuntime.threads.handles, &self->node);
HEAD_UNLOCK(&_PyRuntime);
return self;
}
static void
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
ThreadHandle_incref(ThreadHandle *self)
{
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
_Py_atomic_add_ssize(&self->refcount, 1);
}
static int
detach_thread(ThreadHandle *self)
{
if (!self->has_os_handle) {
return 0;
}
// This is typically short so no need to release the GIL
if (PyThread_detach_thread(self->os_handle)) {
fprintf(stderr, "detach_thread: failed detaching thread\n");
return -1;
}
return 0;
}
// NB: This may be called after the PyThreadState in `thread_run` has been
// deleted; it cannot call anything that relies on a valid PyThreadState
// existing.
static void
ThreadHandle_decref(ThreadHandle *self)
{
if (_Py_atomic_add_ssize(&self->refcount, -1) > 1) {
return;
}
// Remove ourself from the global list of handles
HEAD_LOCK(&_PyRuntime);
if (self->node.next != NULL) {
llist_remove(&self->node);
}
HEAD_UNLOCK(&_PyRuntime);
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
assert(self->shutdown_node.next == NULL);
// It's safe to access state non-atomically:
// 1. This is the destructor; nothing else holds a reference.
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
// 2. The refcount going to zero is a "synchronizes-with" event; all
// changes from other threads are visible.
if (self->state == THREAD_HANDLE_RUNNING && !detach_thread(self)) {
self->state = THREAD_HANDLE_DONE;
}
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
PyMem_RawFree(self);
}
void
_PyThread_AfterFork(struct _pythread_runtime_state *state)
{
// gh-115035: We mark ThreadHandles as not joinable early in the child's
// after-fork handler. We do this before calling any Python code to ensure
// that it happens before any ThreadHandles are deallocated, such as by a
// GC cycle.
PyThread_ident_t current = PyThread_get_thread_ident_ex();
struct llist_node *node;
llist_for_each_safe(node, &state->handles) {
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
ThreadHandle *handle = llist_data(node, ThreadHandle, node);
if (handle->ident == current) {
continue;
}
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
// Mark all threads as done. Any attempts to join or detach the
// underlying OS thread (if any) could crash. We are the only thread;
// it's safe to set this non-atomically.
handle->state = THREAD_HANDLE_DONE;
handle->once = (_PyOnceFlag){_Py_ONCE_INITIALIZED};
handle->mutex = (PyMutex){_Py_UNLOCKED};
_PyEvent_Notify(&handle->thread_is_exiting);
llist_remove(node);
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
remove_from_shutdown_handles(handle);
}
}
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
// bootstate is used to "bootstrap" new threads. Any arguments needed by
// `thread_run()`, which can only take a single argument due to platform
// limitations, are contained in bootstate.
struct bootstate {
PyThreadState *tstate;
PyObject *func;
PyObject *args;
PyObject *kwargs;
ThreadHandle *handle;
PyEvent handle_ready;
};
static void
thread_bootstate_free(struct bootstate *boot, int decref)
{
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
if (decref) {
Py_DECREF(boot->func);
Py_DECREF(boot->args);
Py_XDECREF(boot->kwargs);
}
ThreadHandle_decref(boot->handle);
PyMem_RawFree(boot);
}
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
static void
thread_run(void *boot_raw)
{
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
struct bootstate *boot = (struct bootstate *) boot_raw;
PyThreadState *tstate = boot->tstate;
// Wait until the handle is marked as running
PyEvent_Wait(&boot->handle_ready);
// `handle` needs to be manipulated after bootstate has been freed
ThreadHandle *handle = boot->handle;
ThreadHandle_incref(handle);
// gh-108987: If _thread.start_new_thread() is called before or while
// Python is being finalized, thread_run() can called *after*.
// _PyRuntimeState_SetFinalizing() is called. At this point, all Python
// threads must exit, except of the thread calling Py_Finalize() whch holds
// the GIL and must not exit.
//
// At this stage, tstate can be a dangling pointer (point to freed memory),
// it's ok to call _PyThreadState_MustExit() with a dangling pointer.
if (_PyThreadState_MustExit(tstate)) {
// Don't call PyThreadState_Clear() nor _PyThreadState_DeleteCurrent().
// These functions are called on tstate indirectly by Py_Finalize()
// which calls _PyInterpreterState_Clear().
//
// Py_DECREF() cannot be called because the GIL is not held: leak
// references on purpose. Python is being finalized anyway.
thread_bootstate_free(boot, 0);
goto exit;
}
_PyThreadState_Bind(tstate);
PyEval_AcquireThread(tstate);
_Py_atomic_add_ssize(&tstate->interp->threads.count, 1);
PyObject *res = PyObject_Call(boot->func, boot->args, boot->kwargs);
if (res == NULL) {
if (PyErr_ExceptionMatches(PyExc_SystemExit))
/* SystemExit is ignored silently */
PyErr_Clear();
else {
PyErr_FormatUnraisable(
"Exception ignored in thread started by %R", boot->func);
}
}
else {
Py_DECREF(res);
}
thread_bootstate_free(boot, 1);
_Py_atomic_add_ssize(&tstate->interp->threads.count, -1);
PyThreadState_Clear(tstate);
_PyThreadState_DeleteCurrent(tstate);
exit:
// Don't need to wait for this thread anymore
remove_from_shutdown_handles(handle);
_PyEvent_Notify(&handle->thread_is_exiting);
ThreadHandle_decref(handle);
// bpo-44434: Don't call explicitly PyThread_exit_thread(). On Linux with
// the glibc, pthread_exit() can abort the whole process if dlopen() fails
// to open the libgcc_s.so library (ex: EMFILE error).
return;
}
static int
force_done(ThreadHandle *handle)
{
assert(get_thread_handle_state(handle) == THREAD_HANDLE_STARTING);
_PyEvent_Notify(&handle->thread_is_exiting);
set_thread_handle_state(handle, THREAD_HANDLE_DONE);
return 0;
}
static int
ThreadHandle_start(ThreadHandle *self, PyObject *func, PyObject *args,
PyObject *kwargs)
{
// Mark the handle as starting to prevent any other threads from doing so
PyMutex_Lock(&self->mutex);
if (self->state != THREAD_HANDLE_NOT_STARTED) {
PyMutex_Unlock(&self->mutex);
PyErr_SetString(ThreadError, "thread already started");
return -1;
}
self->state = THREAD_HANDLE_STARTING;
PyMutex_Unlock(&self->mutex);
// Do all the heavy lifting outside of the mutex. All other operations on
// the handle should fail since the handle is in the starting state.
// gh-109795: Use PyMem_RawMalloc() instead of PyMem_Malloc(),
// because it should be possible to call thread_bootstate_free()
// without holding the GIL.
struct bootstate *boot = PyMem_RawMalloc(sizeof(struct bootstate));
if (boot == NULL) {
PyErr_NoMemory();
goto start_failed;
}
PyInterpreterState *interp = _PyInterpreterState_GET();
boot->tstate = _PyThreadState_New(interp, _PyThreadState_WHENCE_THREADING);
if (boot->tstate == NULL) {
PyMem_RawFree(boot);
if (!PyErr_Occurred()) {
PyErr_NoMemory();
}
goto start_failed;
}
boot->func = Py_NewRef(func);
boot->args = Py_NewRef(args);
boot->kwargs = Py_XNewRef(kwargs);
boot->handle = self;
ThreadHandle_incref(self);
boot->handle_ready = (PyEvent){0};
PyThread_ident_t ident;
PyThread_handle_t os_handle;
if (PyThread_start_joinable_thread(thread_run, boot, &ident, &os_handle)) {
PyThreadState_Clear(boot->tstate);
thread_bootstate_free(boot, 1);
PyErr_SetString(ThreadError, "can't start new thread");
goto start_failed;
}
// Mark the handle running
PyMutex_Lock(&self->mutex);
assert(self->state == THREAD_HANDLE_STARTING);
self->ident = ident;
self->has_os_handle = 1;
self->os_handle = os_handle;
self->state = THREAD_HANDLE_RUNNING;
PyMutex_Unlock(&self->mutex);
// Unblock the thread
_PyEvent_Notify(&boot->handle_ready);
return 0;
start_failed:
_PyOnceFlag_CallOnce(&self->once, (_Py_once_fn_t *)force_done, self);
return -1;
}
static int
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
join_thread(ThreadHandle *handle)
{
assert(get_thread_handle_state(handle) == THREAD_HANDLE_RUNNING);
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
PyThread_handle_t os_handle;
if (ThreadHandle_get_os_handle(handle, &os_handle)) {
int err = 0;
Py_BEGIN_ALLOW_THREADS
err = PyThread_join_thread(os_handle);
Py_END_ALLOW_THREADS
if (err) {
PyErr_SetString(ThreadError, "Failed joining thread");
return -1;
}
}
set_thread_handle_state(handle, THREAD_HANDLE_DONE);
return 0;
}
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
static int
check_started(ThreadHandle *self)
{
ThreadHandleState state = get_thread_handle_state(self);
if (state < THREAD_HANDLE_RUNNING) {
PyErr_SetString(ThreadError, "thread not started");
return -1;
}
return 0;
}
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
static int
ThreadHandle_join(ThreadHandle *self, PyTime_t timeout_ns)
{
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
if (check_started(self) < 0) {
return -1;
}
// We want to perform this check outside of the `_PyOnceFlag` to prevent
// deadlock in the scenario where another thread joins us and we then
// attempt to join ourselves. However, it's not safe to check thread
// identity once the handle's os thread has finished. We may end up reusing
// the identity stored in the handle and erroneously think we are
// attempting to join ourselves.
//
// To work around this, we set `thread_is_exiting` immediately before
// `thread_run` returns. We can be sure that we are not attempting to join
// ourselves if the handle's thread is about to exit.
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
if (!_PyEvent_IsSet(&self->thread_is_exiting) &&
ThreadHandle_ident(self) == PyThread_get_thread_ident_ex()) {
// PyThread_join_thread() would deadlock or error out.
PyErr_SetString(ThreadError, "Cannot join current thread");
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
return -1;
}
// Wait until the deadline for the thread to exit.
PyTime_t deadline = timeout_ns != -1 ? _PyDeadline_Init(timeout_ns) : 0;
int detach = 1;
while (!PyEvent_WaitTimed(&self->thread_is_exiting, timeout_ns, detach)) {
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
if (deadline) {
// _PyDeadline_Get will return a negative value if the deadline has
// been exceeded.
timeout_ns = Py_MAX(_PyDeadline_Get(deadline), 0);
}
if (timeout_ns) {
// Interrupted
if (Py_MakePendingCalls() < 0) {
return -1;
}
}
else {
// Timed out
return 0;
}
}
if (_PyOnceFlag_CallOnce(&self->once, (_Py_once_fn_t *)join_thread,
self) == -1) {
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
return -1;
}
assert(get_thread_handle_state(self) == THREAD_HANDLE_DONE);
return 0;
}
static int
set_done(ThreadHandle *handle)
{
assert(get_thread_handle_state(handle) == THREAD_HANDLE_RUNNING);
if (detach_thread(handle) < 0) {
PyErr_SetString(ThreadError, "failed detaching handle");
return -1;
}
_PyEvent_Notify(&handle->thread_is_exiting);
set_thread_handle_state(handle, THREAD_HANDLE_DONE);
return 0;
}
static int
ThreadHandle_set_done(ThreadHandle *self)
{
if (check_started(self) < 0) {
return -1;
}
if (_PyOnceFlag_CallOnce(&self->once, (_Py_once_fn_t *)set_done, self) ==
-1) {
return -1;
}
assert(get_thread_handle_state(self) == THREAD_HANDLE_DONE);
return 0;
}
// A wrapper around a ThreadHandle.
typedef struct {
PyObject_HEAD
ThreadHandle *handle;
} PyThreadHandleObject;
static PyThreadHandleObject *
PyThreadHandleObject_new(PyTypeObject *type)
{
ThreadHandle *handle = ThreadHandle_new();
if (handle == NULL) {
return NULL;
}
PyThreadHandleObject *self =
(PyThreadHandleObject *)type->tp_alloc(type, 0);
if (self == NULL) {
ThreadHandle_decref(handle);
return NULL;
}
self->handle = handle;
return self;
}
static PyObject *
PyThreadHandleObject_tp_new(PyTypeObject *type, PyObject *args, PyObject *kwds)
{
return (PyObject *)PyThreadHandleObject_new(type);
}
static int
PyThreadHandleObject_traverse(PyThreadHandleObject *self, visitproc visit,
void *arg)
{
Py_VISIT(Py_TYPE(self));
return 0;
}
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
static void
PyThreadHandleObject_dealloc(PyThreadHandleObject *self)
{
PyObject_GC_UnTrack(self);
PyTypeObject *tp = Py_TYPE(self);
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
ThreadHandle_decref(self->handle);
tp->tp_free(self);
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
Py_DECREF(tp);
}
static PyObject *
PyThreadHandleObject_repr(PyThreadHandleObject *self)
{
PyThread_ident_t ident = ThreadHandle_ident(self->handle);
return PyUnicode_FromFormat("<%s object: ident=%" PY_FORMAT_THREAD_IDENT_T ">",
Py_TYPE(self)->tp_name, ident);
}
static PyObject *
PyThreadHandleObject_get_ident(PyThreadHandleObject *self,
PyObject *Py_UNUSED(ignored))
{
return PyLong_FromUnsignedLongLong(ThreadHandle_ident(self->handle));
}
static PyObject *
PyThreadHandleObject_join(PyThreadHandleObject *self, PyObject *args)
{
PyObject *timeout_obj = NULL;
if (!PyArg_ParseTuple(args, "|O:join", &timeout_obj)) {
return NULL;
}
PyTime_t timeout_ns = -1;
if (timeout_obj != NULL && timeout_obj != Py_None) {
if (_PyTime_FromSecondsObject(&timeout_ns, timeout_obj,
_PyTime_ROUND_TIMEOUT) < 0) {
return NULL;
}
}
if (ThreadHandle_join(self->handle, timeout_ns) < 0) {
return NULL;
}
Py_RETURN_NONE;
}
static PyObject *
PyThreadHandleObject_is_done(PyThreadHandleObject *self,
PyObject *Py_UNUSED(ignored))
{
if (_PyEvent_IsSet(&self->handle->thread_is_exiting)) {
Py_RETURN_TRUE;
}
else {
Py_RETURN_FALSE;
}
}
static PyObject *
PyThreadHandleObject_set_done(PyThreadHandleObject *self,
PyObject *Py_UNUSED(ignored))
{
if (ThreadHandle_set_done(self->handle) < 0) {
return NULL;
}
Py_RETURN_NONE;
}
static PyGetSetDef ThreadHandle_getsetlist[] = {
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
{"ident", (getter)PyThreadHandleObject_get_ident, NULL, NULL},
{0},
};
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
static PyMethodDef ThreadHandle_methods[] = {
{"join", (PyCFunction)PyThreadHandleObject_join, METH_VARARGS, NULL},
{"_set_done", (PyCFunction)PyThreadHandleObject_set_done, METH_NOARGS, NULL},
{"is_done", (PyCFunction)PyThreadHandleObject_is_done, METH_NOARGS, NULL},
{0, 0}
};
static PyType_Slot ThreadHandle_Type_slots[] = {
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
{Py_tp_dealloc, (destructor)PyThreadHandleObject_dealloc},
{Py_tp_repr, (reprfunc)PyThreadHandleObject_repr},
{Py_tp_getset, ThreadHandle_getsetlist},
{Py_tp_traverse, PyThreadHandleObject_traverse},
{Py_tp_methods, ThreadHandle_methods},
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
{Py_tp_new, PyThreadHandleObject_tp_new},
{0, 0}
};
static PyType_Spec ThreadHandle_Type_spec = {
"_thread._ThreadHandle",
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
sizeof(PyThreadHandleObject),
0,
Py_TPFLAGS_DEFAULT | Py_TPFLAGS_IMMUTABLETYPE | Py_TPFLAGS_HAVE_GC,
ThreadHandle_Type_slots,
};
/* Lock objects */
typedef struct {
PyObject_HEAD
PyThread_type_lock lock_lock;
PyObject *in_weakreflist;
char locked; /* for sanity checking */
} lockobject;
static int
lock_traverse(lockobject *self, visitproc visit, void *arg)
{
Py_VISIT(Py_TYPE(self));
return 0;
}
static void
2000-07-10 07:03:58 -03:00
lock_dealloc(lockobject *self)
{
PyObject_GC_UnTrack(self);
if (self->in_weakreflist != NULL) {
PyObject_ClearWeakRefs((PyObject *) self);
}
if (self->lock_lock != NULL) {
/* Unlock the lock so it's safe to free it */
if (self->locked)
PyThread_release_lock(self->lock_lock);
PyThread_free_lock(self->lock_lock);
}
PyTypeObject *tp = Py_TYPE(self);
tp->tp_free((PyObject*)self);
Py_DECREF(tp);
}
static inline PyLockStatus
acquire_timed(PyThread_type_lock lock, PyTime_t timeout)
{
return PyThread_acquire_lock_timed_with_retries(lock, timeout);
}
static int
lock_acquire_parse_args(PyObject *args, PyObject *kwds,
PyTime_t *timeout)
{
char *kwlist[] = {"blocking", "timeout", NULL};
int blocking = 1;
PyObject *timeout_obj = NULL;
if (!PyArg_ParseTupleAndKeywords(args, kwds, "|pO:acquire", kwlist,
&blocking, &timeout_obj))
return -1;
// XXX Use PyThread_ParseTimeoutArg().
const PyTime_t unset_timeout = _PyTime_FromSeconds(-1);
*timeout = unset_timeout;
if (timeout_obj
&& _PyTime_FromSecondsObject(timeout,
timeout_obj, _PyTime_ROUND_TIMEOUT) < 0)
return -1;
if (!blocking && *timeout != unset_timeout ) {
PyErr_SetString(PyExc_ValueError,
"can't specify a timeout for a non-blocking call");
return -1;
}
if (*timeout < 0 && *timeout != unset_timeout) {
PyErr_SetString(PyExc_ValueError,
"timeout value must be a non-negative number");
return -1;
}
if (!blocking)
*timeout = 0;
else if (*timeout != unset_timeout) {
PyTime_t microseconds;
microseconds = _PyTime_AsMicroseconds(*timeout, _PyTime_ROUND_TIMEOUT);
if (microseconds > PY_TIMEOUT_MAX) {
PyErr_SetString(PyExc_OverflowError,
"timeout value is too large");
return -1;
}
}
return 0;
}
static PyObject *
lock_PyThread_acquire_lock(lockobject *self, PyObject *args, PyObject *kwds)
{
PyTime_t timeout;
if (lock_acquire_parse_args(args, kwds, &timeout) < 0)
return NULL;
PyLockStatus r = acquire_timed(self->lock_lock, timeout);
if (r == PY_LOCK_INTR) {
return NULL;
}
if (r == PY_LOCK_ACQUIRED)
self->locked = 1;
return PyBool_FromLong(r == PY_LOCK_ACQUIRED);
}
2002-06-13 17:33:02 -03:00
PyDoc_STRVAR(acquire_doc,
"acquire($self, /, blocking=True, timeout=-1)\n\
--\n\
\n\
Lock the lock. Without argument, this blocks if the lock is already\n\
locked (even by the same thread), waiting for another thread to release\n\
the lock, and return True once the lock is acquired.\n\
With an argument, this will only block if the argument is true,\n\
and the return value reflects whether the lock is acquired.\n\
The blocking operation is interruptible.");
PyDoc_STRVAR(acquire_lock_doc,
"acquire_lock($self, /, blocking=True, timeout=-1)\n\
--\n\
\n\
An obsolete synonym of acquire().");
PyDoc_STRVAR(enter_doc,
"__enter__($self, /)\n\
--\n\
\n\
Lock the lock.");
1996-12-16 20:05:22 -04:00
static PyObject *
lock_PyThread_release_lock(lockobject *self, PyObject *Py_UNUSED(ignored))
{
/* Sanity check: the lock must be locked */
if (!self->locked) {
PyErr_SetString(ThreadError, "release unlocked lock");
return NULL;
}
self->locked = 0;
PyThread_release_lock(self->lock_lock);
Py_RETURN_NONE;
}
2002-06-13 17:33:02 -03:00
PyDoc_STRVAR(release_doc,
"release($self, /)\n\
--\n\
\n\
Release the lock, allowing another thread that is blocked waiting for\n\
the lock to acquire the lock. The lock must be in the locked state,\n\
2002-06-13 17:33:02 -03:00
but it needn't be locked by the same thread that unlocks it.");
PyDoc_STRVAR(release_lock_doc,
"release_lock($self, /)\n\
--\n\
\n\
An obsolete synonym of release().");
PyDoc_STRVAR(lock_exit_doc,
"__exit__($self, /, *exc_info)\n\
--\n\
\n\
Release the lock.");
1996-12-16 20:05:22 -04:00
static PyObject *
lock_locked_lock(lockobject *self, PyObject *Py_UNUSED(ignored))
{
return PyBool_FromLong((long)self->locked);
}
2002-06-13 17:33:02 -03:00
PyDoc_STRVAR(locked_doc,
"locked($self, /)\n\
--\n\
\n\
2002-06-13 17:33:02 -03:00
Return whether the lock is in the locked state.");
PyDoc_STRVAR(locked_lock_doc,
"locked_lock($self, /)\n\
--\n\
\n\
An obsolete synonym of locked().");
static PyObject *
lock_repr(lockobject *self)
{
return PyUnicode_FromFormat("<%s %s object at %p>",
self->locked ? "locked" : "unlocked", Py_TYPE(self)->tp_name, self);
}
#ifdef HAVE_FORK
static PyObject *
lock__at_fork_reinit(lockobject *self, PyObject *Py_UNUSED(args))
{
if (_PyThread_at_fork_reinit(&self->lock_lock) < 0) {
PyErr_SetString(ThreadError, "failed to reinitialize lock at fork");
return NULL;
}
self->locked = 0;
Py_RETURN_NONE;
}
#endif /* HAVE_FORK */
static lockobject *newlockobject(PyObject *module);
static PyObject *
lock_new(PyTypeObject *type, PyObject *args, PyObject *kwargs)
{
// convert to AC?
if (!_PyArg_NoKeywords("lock", kwargs)) {
goto error;
}
if (!_PyArg_CheckPositional("lock", PyTuple_GET_SIZE(args), 0, 0)) {
goto error;
}
PyObject *module = PyType_GetModuleByDef(type, &thread_module);
assert(module != NULL);
return (PyObject *)newlockobject(module);
error:
return NULL;
}
1996-12-16 20:05:22 -04:00
static PyMethodDef lock_methods[] = {
{"acquire_lock", _PyCFunction_CAST(lock_PyThread_acquire_lock),
METH_VARARGS | METH_KEYWORDS, acquire_lock_doc},
{"acquire", _PyCFunction_CAST(lock_PyThread_acquire_lock),
METH_VARARGS | METH_KEYWORDS, acquire_doc},
{"release_lock", (PyCFunction)lock_PyThread_release_lock,
METH_NOARGS, release_lock_doc},
{"release", (PyCFunction)lock_PyThread_release_lock,
METH_NOARGS, release_doc},
{"locked_lock", (PyCFunction)lock_locked_lock,
METH_NOARGS, locked_lock_doc},
{"locked", (PyCFunction)lock_locked_lock,
METH_NOARGS, locked_doc},
{"__enter__", _PyCFunction_CAST(lock_PyThread_acquire_lock),
METH_VARARGS | METH_KEYWORDS, enter_doc},
{"__exit__", (PyCFunction)lock_PyThread_release_lock,
METH_VARARGS, lock_exit_doc},
#ifdef HAVE_FORK
{"_at_fork_reinit", (PyCFunction)lock__at_fork_reinit,
METH_NOARGS, NULL},
#endif
{NULL, NULL} /* sentinel */
};
PyDoc_STRVAR(lock_doc,
"lock()\n\
--\n\
\n\
A lock object is a synchronization primitive. To create a lock,\n\
call threading.Lock(). Methods are:\n\
\n\
acquire() -- lock the lock, possibly blocking until it can be obtained\n\
release() -- unlock of the lock\n\
locked() -- test whether the lock is currently locked\n\
\n\
A lock is not owned by the thread that locked it; another thread may\n\
unlock it. A thread attempting to lock a lock that it has already locked\n\
will block until another thread unlocks it. Deadlocks may ensue.");
static PyMemberDef lock_type_members[] = {
{"__weaklistoffset__", Py_T_PYSSIZET, offsetof(lockobject, in_weakreflist), Py_READONLY},
{NULL},
};
static PyType_Slot lock_type_slots[] = {
{Py_tp_dealloc, (destructor)lock_dealloc},
{Py_tp_repr, (reprfunc)lock_repr},
{Py_tp_doc, (void *)lock_doc},
{Py_tp_methods, lock_methods},
{Py_tp_traverse, lock_traverse},
{Py_tp_members, lock_type_members},
{Py_tp_new, lock_new},
{0, 0}
};
static PyType_Spec lock_type_spec = {
.name = "_thread.lock",
.basicsize = sizeof(lockobject),
.flags = (Py_TPFLAGS_DEFAULT | Py_TPFLAGS_HAVE_GC |
Py_TPFLAGS_IMMUTABLETYPE),
.slots = lock_type_slots,
};
/* Recursive lock objects */
typedef struct {
PyObject_HEAD
PyThread_type_lock rlock_lock;
PyThread_ident_t rlock_owner;
unsigned long rlock_count;
PyObject *in_weakreflist;
} rlockobject;
static int
rlock_traverse(rlockobject *self, visitproc visit, void *arg)
{
Py_VISIT(Py_TYPE(self));
return 0;
}
static void
rlock_dealloc(rlockobject *self)
{
PyObject_GC_UnTrack(self);
if (self->in_weakreflist != NULL)
PyObject_ClearWeakRefs((PyObject *) self);
/* self->rlock_lock can be NULL if PyThread_allocate_lock() failed
in rlock_new() */
if (self->rlock_lock != NULL) {
/* Unlock the lock so it's safe to free it */
if (self->rlock_count > 0)
PyThread_release_lock(self->rlock_lock);
PyThread_free_lock(self->rlock_lock);
}
PyTypeObject *tp = Py_TYPE(self);
tp->tp_free(self);
Py_DECREF(tp);
}
static bool
rlock_is_owned_by(rlockobject *self, PyThread_ident_t tid)
{
PyThread_ident_t owner_tid =
_Py_atomic_load_ullong_relaxed(&self->rlock_owner);
return owner_tid == tid && self->rlock_count > 0;
}
static PyObject *
rlock_acquire(rlockobject *self, PyObject *args, PyObject *kwds)
{
PyTime_t timeout;
PyThread_ident_t tid;
PyLockStatus r = PY_LOCK_ACQUIRED;
if (lock_acquire_parse_args(args, kwds, &timeout) < 0)
return NULL;
tid = PyThread_get_thread_ident_ex();
if (rlock_is_owned_by(self, tid)) {
unsigned long count = self->rlock_count + 1;
if (count <= self->rlock_count) {
PyErr_SetString(PyExc_OverflowError,
"Internal lock count overflowed");
return NULL;
}
self->rlock_count = count;
Py_RETURN_TRUE;
}
r = acquire_timed(self->rlock_lock, timeout);
if (r == PY_LOCK_ACQUIRED) {
assert(self->rlock_count == 0);
_Py_atomic_store_ullong_relaxed(&self->rlock_owner, tid);
self->rlock_count = 1;
}
else if (r == PY_LOCK_INTR) {
return NULL;
}
return PyBool_FromLong(r == PY_LOCK_ACQUIRED);
}
PyDoc_STRVAR(rlock_acquire_doc,
"acquire($self, /, blocking=True, timeout=-1)\n\
--\n\
\n\
Lock the lock. `blocking` indicates whether we should wait\n\
for the lock to be available or not. If `blocking` is False\n\
and another thread holds the lock, the method will return False\n\
immediately. If `blocking` is True and another thread holds\n\
the lock, the method will wait for the lock to be released,\n\
take it and then return True.\n\
(note: the blocking operation is interruptible.)\n\
\n\
In all other cases, the method will return True immediately.\n\
Precisely, if the current thread already holds the lock, its\n\
internal counter is simply incremented. If nobody holds the lock,\n\
the lock is taken and its internal counter initialized to 1.");
PyDoc_STRVAR(rlock_enter_doc,
"__enter__($self, /)\n\
--\n\
\n\
Lock the lock.");
static PyObject *
rlock_release(rlockobject *self, PyObject *Py_UNUSED(ignored))
{
PyThread_ident_t tid = PyThread_get_thread_ident_ex();
if (!rlock_is_owned_by(self, tid)) {
PyErr_SetString(PyExc_RuntimeError,
"cannot release un-acquired lock");
return NULL;
}
if (--self->rlock_count == 0) {
_Py_atomic_store_ullong_relaxed(&self->rlock_owner, 0);
PyThread_release_lock(self->rlock_lock);
}
Py_RETURN_NONE;
}
PyDoc_STRVAR(rlock_release_doc,
"release($self, /)\n\
--\n\
\n\
Release the lock, allowing another thread that is blocked waiting for\n\
the lock to acquire the lock. The lock must be in the locked state,\n\
and must be locked by the same thread that unlocks it; otherwise a\n\
`RuntimeError` is raised.\n\
\n\
Do note that if the lock was acquire()d several times in a row by the\n\
current thread, release() needs to be called as many times for the lock\n\
to be available for other threads.");
PyDoc_STRVAR(rlock_exit_doc,
"__exit__($self, /, *exc_info)\n\
--\n\
\n\
Release the lock.");
static PyObject *
rlock_acquire_restore(rlockobject *self, PyObject *args)
{
PyThread_ident_t owner;
unsigned long count;
int r = 1;
if (!PyArg_ParseTuple(args, "(k" Py_PARSE_THREAD_IDENT_T "):_acquire_restore",
&count, &owner))
return NULL;
if (!PyThread_acquire_lock(self->rlock_lock, 0)) {
Py_BEGIN_ALLOW_THREADS
r = PyThread_acquire_lock(self->rlock_lock, 1);
Py_END_ALLOW_THREADS
}
if (!r) {
PyErr_SetString(ThreadError, "couldn't acquire lock");
return NULL;
}
assert(self->rlock_count == 0);
_Py_atomic_store_ullong_relaxed(&self->rlock_owner, owner);
self->rlock_count = count;
Py_RETURN_NONE;
}
PyDoc_STRVAR(rlock_acquire_restore_doc,
"_acquire_restore($self, state, /)\n\
--\n\
\n\
For internal use by `threading.Condition`.");
static PyObject *
rlock_release_save(rlockobject *self, PyObject *Py_UNUSED(ignored))
{
PyThread_ident_t owner;
unsigned long count;
if (self->rlock_count == 0) {
PyErr_SetString(PyExc_RuntimeError,
"cannot release un-acquired lock");
return NULL;
}
owner = self->rlock_owner;
count = self->rlock_count;
self->rlock_count = 0;
_Py_atomic_store_ullong_relaxed(&self->rlock_owner, 0);
PyThread_release_lock(self->rlock_lock);
return Py_BuildValue("k" Py_PARSE_THREAD_IDENT_T, count, owner);
}
PyDoc_STRVAR(rlock_release_save_doc,
"_release_save($self, /)\n\
--\n\
\n\
For internal use by `threading.Condition`.");
static PyObject *
rlock_recursion_count(rlockobject *self, PyObject *Py_UNUSED(ignored))
{
PyThread_ident_t tid = PyThread_get_thread_ident_ex();
PyThread_ident_t owner =
_Py_atomic_load_ullong_relaxed(&self->rlock_owner);
return PyLong_FromUnsignedLong(owner == tid ? self->rlock_count : 0UL);
}
PyDoc_STRVAR(rlock_recursion_count_doc,
"_recursion_count($self, /)\n\
--\n\
\n\
For internal use by reentrancy checks.");
static PyObject *
rlock_is_owned(rlockobject *self, PyObject *Py_UNUSED(ignored))
{
PyThread_ident_t tid = PyThread_get_thread_ident_ex();
if (rlock_is_owned_by(self, tid)) {
Py_RETURN_TRUE;
}
Py_RETURN_FALSE;
}
PyDoc_STRVAR(rlock_is_owned_doc,
"_is_owned($self, /)\n\
--\n\
\n\
For internal use by `threading.Condition`.");
static PyObject *
rlock_new(PyTypeObject *type, PyObject *args, PyObject *kwds)
{
rlockobject *self = (rlockobject *) type->tp_alloc(type, 0);
if (self == NULL) {
return NULL;
}
self->in_weakreflist = NULL;
self->rlock_owner = 0;
self->rlock_count = 0;
self->rlock_lock = PyThread_allocate_lock();
if (self->rlock_lock == NULL) {
Py_DECREF(self);
PyErr_SetString(ThreadError, "can't allocate lock");
return NULL;
}
return (PyObject *) self;
}
static PyObject *
rlock_repr(rlockobject *self)
{
PyThread_ident_t owner =
_Py_atomic_load_ullong_relaxed(&self->rlock_owner);
return PyUnicode_FromFormat(
"<%s %s object owner=%" PY_FORMAT_THREAD_IDENT_T " count=%lu at %p>",
self->rlock_count ? "locked" : "unlocked",
Py_TYPE(self)->tp_name, owner,
self->rlock_count, self);
}
#ifdef HAVE_FORK
static PyObject *
rlock__at_fork_reinit(rlockobject *self, PyObject *Py_UNUSED(args))
{
if (_PyThread_at_fork_reinit(&self->rlock_lock) < 0) {
PyErr_SetString(ThreadError, "failed to reinitialize lock at fork");
return NULL;
}
self->rlock_owner = 0;
self->rlock_count = 0;
Py_RETURN_NONE;
}
#endif /* HAVE_FORK */
static PyMethodDef rlock_methods[] = {
{"acquire", _PyCFunction_CAST(rlock_acquire),
METH_VARARGS | METH_KEYWORDS, rlock_acquire_doc},
{"release", (PyCFunction)rlock_release,
METH_NOARGS, rlock_release_doc},
{"_is_owned", (PyCFunction)rlock_is_owned,
METH_NOARGS, rlock_is_owned_doc},
{"_acquire_restore", (PyCFunction)rlock_acquire_restore,
METH_VARARGS, rlock_acquire_restore_doc},
{"_release_save", (PyCFunction)rlock_release_save,
METH_NOARGS, rlock_release_save_doc},
{"_recursion_count", (PyCFunction)rlock_recursion_count,
METH_NOARGS, rlock_recursion_count_doc},
{"__enter__", _PyCFunction_CAST(rlock_acquire),
METH_VARARGS | METH_KEYWORDS, rlock_enter_doc},
{"__exit__", (PyCFunction)rlock_release,
METH_VARARGS, rlock_exit_doc},
#ifdef HAVE_FORK
{"_at_fork_reinit", (PyCFunction)rlock__at_fork_reinit,
METH_NOARGS, NULL},
#endif
{NULL, NULL} /* sentinel */
};
static PyMemberDef rlock_type_members[] = {
{"__weaklistoffset__", Py_T_PYSSIZET, offsetof(rlockobject, in_weakreflist), Py_READONLY},
{NULL},
};
static PyType_Slot rlock_type_slots[] = {
{Py_tp_dealloc, (destructor)rlock_dealloc},
{Py_tp_repr, (reprfunc)rlock_repr},
{Py_tp_methods, rlock_methods},
{Py_tp_alloc, PyType_GenericAlloc},
{Py_tp_new, rlock_new},
{Py_tp_members, rlock_type_members},
{Py_tp_traverse, rlock_traverse},
{0, 0},
};
static PyType_Spec rlock_type_spec = {
.name = "_thread.RLock",
.basicsize = sizeof(rlockobject),
.flags = (Py_TPFLAGS_DEFAULT | Py_TPFLAGS_BASETYPE |
Py_TPFLAGS_HAVE_GC | Py_TPFLAGS_IMMUTABLETYPE),
.slots = rlock_type_slots,
};
static lockobject *
newlockobject(PyObject *module)
{
thread_module_state *state = get_thread_state(module);
PyTypeObject *type = state->lock_type;
lockobject *self = (lockobject *)type->tp_alloc(type, 0);
if (self == NULL) {
return NULL;
}
self->lock_lock = PyThread_allocate_lock();
self->locked = 0;
self->in_weakreflist = NULL;
if (self->lock_lock == NULL) {
Py_DECREF(self);
PyErr_SetString(ThreadError, "can't allocate lock");
return NULL;
}
return self;
}
/* Thread-local objects */
/* Quick overview:
We need to be able to reclaim reference cycles as soon as possible
(both when a thread is being terminated, or a thread-local object
becomes unreachable from user data). Constraints:
- it must not be possible for thread-state dicts to be involved in
reference cycles (otherwise the cyclic GC will refuse to consider
objects referenced from a reachable thread-state dict, even though
local_dealloc would clear them)
- the death of a thread-state dict must still imply destruction of the
corresponding local dicts in all thread-local objects.
Our implementation uses small "localdummy" objects in order to break
the reference chain. These trivial objects are hashable (using the
default scheme of identity hashing) and weakrefable.
Each thread-state holds two separate localdummy objects:
- `threading_local_key` is used as a key to retrieve the locals dictionary
for the thread in any `threading.local` object.
- `threading_local_sentinel` is used to signal when a thread is being
destroyed. Consequently, the associated thread-state must hold the only
reference.
Each `threading.local` object contains a dict mapping localdummy keys to
locals dicts and a set containing weak references to localdummy
sentinels. Each sentinel weak reference has a callback that removes itself
and the locals dict for the key from the `threading.local` object when
called.
Therefore:
- The thread-state only holds strong references to localdummy objects, which
cannot participate in cycles.
- Only outside objects (application- or library-level) hold strong
references to the thread-local objects.
- As soon as thread-state's sentinel dummy is destroyed the callbacks for
all weakrefs attached to the sentinel are called, and destroy the
corresponding local dicts from thread-local objects.
- As soon as a thread-local object is destroyed, its local dicts are
destroyed.
- The GC can do its work correctly when a thread-local object is dangling,
without any interference from the thread-state dicts.
This dual key arrangement is necessary to ensure that `threading.local`
values can be retrieved from finalizers. If we were to only keep a mapping
of localdummy weakrefs to locals dicts it's possible that the weakrefs would
be cleared before finalizers were called (GC currently clears weakrefs that
are garbage before invoking finalizers), causing lookups in finalizers to
fail.
*/
typedef struct {
PyObject_HEAD
PyObject *weakreflist; /* List of weak references to self */
} localdummyobject;
static void
localdummy_dealloc(localdummyobject *self)
{
if (self->weakreflist != NULL)
PyObject_ClearWeakRefs((PyObject *) self);
PyTypeObject *tp = Py_TYPE(self);
tp->tp_free((PyObject*)self);
Py_DECREF(tp);
}
static PyMemberDef local_dummy_type_members[] = {
{"__weaklistoffset__", Py_T_PYSSIZET, offsetof(localdummyobject, weakreflist), Py_READONLY},
{NULL},
};
static PyType_Slot local_dummy_type_slots[] = {
{Py_tp_dealloc, (destructor)localdummy_dealloc},
{Py_tp_doc, "Thread-local dummy"},
{Py_tp_members, local_dummy_type_members},
{0, 0}
};
static PyType_Spec local_dummy_type_spec = {
.name = "_thread._localdummy",
.basicsize = sizeof(localdummyobject),
.flags = (Py_TPFLAGS_DEFAULT | Py_TPFLAGS_DISALLOW_INSTANTIATION |
Py_TPFLAGS_IMMUTABLETYPE),
.slots = local_dummy_type_slots,
};
typedef struct {
PyObject_HEAD
PyObject *args;
PyObject *kw;
PyObject *weakreflist; /* List of weak references to self */
/* A {localdummy -> localdict} dict */
PyObject *localdicts;
/* A set of weakrefs to thread sentinels localdummies*/
PyObject *thread_watchdogs;
} localobject;
/* Forward declaration */
static int create_localsdict(localobject *self, thread_module_state *state,
PyObject **localsdict, PyObject **sentinel_wr);
static PyObject *clear_locals(PyObject *meth_self, PyObject *dummyweakref);
/* Create a weakref to the sentinel localdummy for the current thread */
static PyObject *
create_sentinel_wr(localobject *self)
{
static PyMethodDef wr_callback_def = {
"clear_locals", (PyCFunction) clear_locals, METH_O
};
PyThreadState *tstate = PyThreadState_Get();
/* We use a weak reference to self in the callback closure
in order to avoid spurious reference cycles */
PyObject *self_wr = PyWeakref_NewRef((PyObject *) self, NULL);
if (self_wr == NULL) {
return NULL;
}
PyObject *args = PyTuple_New(2);
if (args == NULL) {
Py_DECREF(self_wr);
return NULL;
}
PyTuple_SET_ITEM(args, 0, self_wr);
PyTuple_SET_ITEM(args, 1, Py_NewRef(tstate->threading_local_key));
PyObject *cb = PyCFunction_New(&wr_callback_def, args);
Py_DECREF(args);
if (cb == NULL) {
return NULL;
}
PyObject *wr = PyWeakref_NewRef(tstate->threading_local_sentinel, cb);
Py_DECREF(cb);
return wr;
}
static PyObject *
local_new(PyTypeObject *type, PyObject *args, PyObject *kw)
{
if (type->tp_init == PyBaseObject_Type.tp_init) {
int rc = 0;
if (args != NULL)
rc = PyObject_IsTrue(args);
if (rc == 0 && kw != NULL)
rc = PyObject_IsTrue(kw);
if (rc != 0) {
if (rc > 0) {
PyErr_SetString(PyExc_TypeError,
"Initialization arguments are not supported");
}
return NULL;
}
}
PyObject *module = PyType_GetModuleByDef(type, &thread_module);
assert(module != NULL);
thread_module_state *state = get_thread_state(module);
localobject *self = (localobject *)type->tp_alloc(type, 0);
if (self == NULL) {
return NULL;
}
self->args = Py_XNewRef(args);
self->kw = Py_XNewRef(kw);
self->localdicts = PyDict_New();
if (self->localdicts == NULL) {
goto err;
}
self->thread_watchdogs = PySet_New(NULL);
if (self->thread_watchdogs == NULL) {
goto err;
}
PyObject *localsdict = NULL;
PyObject *sentinel_wr = NULL;
if (create_localsdict(self, state, &localsdict, &sentinel_wr) < 0) {
goto err;
}
Py_DECREF(localsdict);
Py_DECREF(sentinel_wr);
return (PyObject *)self;
err:
Py_DECREF(self);
return NULL;
}
static int
local_traverse(localobject *self, visitproc visit, void *arg)
{
Py_VISIT(Py_TYPE(self));
Py_VISIT(self->args);
Py_VISIT(self->kw);
Py_VISIT(self->localdicts);
Py_VISIT(self->thread_watchdogs);
return 0;
}
static int
local_clear(localobject *self)
{
Py_CLEAR(self->args);
Py_CLEAR(self->kw);
Py_CLEAR(self->localdicts);
Py_CLEAR(self->thread_watchdogs);
return 0;
}
static void
local_dealloc(localobject *self)
{
/* Weakrefs must be invalidated right now, otherwise they can be used
from code called below, which is very dangerous since Py_REFCNT(self) == 0 */
if (self->weakreflist != NULL) {
PyObject_ClearWeakRefs((PyObject *) self);
}
PyObject_GC_UnTrack(self);
local_clear(self);
PyTypeObject *tp = Py_TYPE(self);
tp->tp_free((PyObject*)self);
Py_DECREF(tp);
}
/* Create the TLS key and sentinel if they don't exist */
static int
create_localdummies(thread_module_state *state)
{
PyThreadState *tstate = _PyThreadState_GET();
if (tstate->threading_local_key != NULL) {
return 0;
}
PyTypeObject *ld_type = state->local_dummy_type;
tstate->threading_local_key = ld_type->tp_alloc(ld_type, 0);
if (tstate->threading_local_key == NULL) {
return -1;
}
tstate->threading_local_sentinel = ld_type->tp_alloc(ld_type, 0);
if (tstate->threading_local_sentinel == NULL) {
Py_CLEAR(tstate->threading_local_key);
return -1;
}
return 0;
}
/* Insert a localsdict and sentinel weakref for the current thread, placing
strong references in localsdict and sentinel_wr, respectively.
*/
static int
create_localsdict(localobject *self, thread_module_state *state,
PyObject **localsdict, PyObject **sentinel_wr)
{
PyThreadState *tstate = _PyThreadState_GET();
PyObject *ldict = NULL;
PyObject *wr = NULL;
if (create_localdummies(state) < 0) {
goto err;
}
/* Create and insert the locals dict and sentinel weakref */
ldict = PyDict_New();
if (ldict == NULL) {
goto err;
}
if (PyDict_SetItem(self->localdicts, tstate->threading_local_key, ldict) <
0) {
goto err;
}
wr = create_sentinel_wr(self);
if (wr == NULL) {
PyObject *exc = PyErr_GetRaisedException();
if (PyDict_DelItem(self->localdicts, tstate->threading_local_key) <
0) {
PyErr_WriteUnraisable((PyObject *)self);
}
PyErr_SetRaisedException(exc);
goto err;
}
if (PySet_Add(self->thread_watchdogs, wr) < 0) {
PyObject *exc = PyErr_GetRaisedException();
if (PyDict_DelItem(self->localdicts, tstate->threading_local_key) <
0) {
PyErr_WriteUnraisable((PyObject *)self);
}
PyErr_SetRaisedException(exc);
goto err;
}
*localsdict = ldict;
*sentinel_wr = wr;
return 0;
err:
Py_XDECREF(ldict);
Py_XDECREF(wr);
return -1;
}
/* Return a strong reference to the locals dict for the current thread,
creating it if necessary.
*/
static PyObject *
_ldict(localobject *self, thread_module_state *state)
{
if (create_localdummies(state) < 0) {
return NULL;
}
/* Check if a localsdict already exists */
PyObject *ldict;
PyThreadState *tstate = _PyThreadState_GET();
if (PyDict_GetItemRef(self->localdicts, tstate->threading_local_key,
&ldict) < 0) {
return NULL;
}
if (ldict != NULL) {
return ldict;
}
/* threading.local hasn't been instantiated for this thread */
PyObject *wr;
if (create_localsdict(self, state, &ldict, &wr) < 0) {
return NULL;
}
/* run __init__ if we're a subtype of `threading.local` */
if (Py_TYPE(self)->tp_init != PyBaseObject_Type.tp_init &&
Py_TYPE(self)->tp_init((PyObject *)self, self->args, self->kw) < 0) {
/* we need to get rid of ldict from thread so
we create a new one the next time we do an attr
access */
PyObject *exc = PyErr_GetRaisedException();
if (PyDict_DelItem(self->localdicts, tstate->threading_local_key) <
0) {
PyErr_WriteUnraisable((PyObject *)self);
PyErr_Clear();
}
if (PySet_Discard(self->thread_watchdogs, wr) < 0) {
PyErr_WriteUnraisable((PyObject *)self);
}
PyErr_SetRaisedException(exc);
Py_DECREF(ldict);
Py_DECREF(wr);
return NULL;
}
Py_DECREF(wr);
return ldict;
}
static int
local_setattro(localobject *self, PyObject *name, PyObject *v)
{
PyObject *module = PyType_GetModuleByDef(Py_TYPE(self), &thread_module);
assert(module != NULL);
thread_module_state *state = get_thread_state(module);
PyObject *ldict = _ldict(self, state);
if (ldict == NULL) {
goto err;
}
bpo-46541: Replace core use of _Py_IDENTIFIER() with statically initialized global objects. (gh-30928) We're no longer using _Py_IDENTIFIER() (or _Py_static_string()) in any core CPython code. It is still used in a number of non-builtin stdlib modules. The replacement is: PyUnicodeObject (not pointer) fields under _PyRuntimeState, statically initialized as part of _PyRuntime. A new _Py_GET_GLOBAL_IDENTIFIER() macro facilitates lookup of the fields (along with _Py_GET_GLOBAL_STRING() for non-identifier strings). https://bugs.python.org/issue46541#msg411799 explains the rationale for this change. The core of the change is in: * (new) Include/internal/pycore_global_strings.h - the declarations for the global strings, along with the macros * Include/internal/pycore_runtime_init.h - added the static initializers for the global strings * Include/internal/pycore_global_objects.h - where the struct in pycore_global_strings.h is hooked into _PyRuntimeState * Tools/scripts/generate_global_objects.py - added generation of the global string declarations and static initializers I've also added a --check flag to generate_global_objects.py (along with make check-global-objects) to check for unused global strings. That check is added to the PR CI config. The remainder of this change updates the core code to use _Py_GET_GLOBAL_IDENTIFIER() instead of _Py_IDENTIFIER() and the related _Py*Id functions (likewise for _Py_GET_GLOBAL_STRING() instead of _Py_static_string()). This includes adding a few functions where there wasn't already an alternative to _Py*Id(), replacing the _Py_Identifier * parameter with PyObject *. The following are not changed (yet): * stop using _Py_IDENTIFIER() in the stdlib modules * (maybe) get rid of _Py_IDENTIFIER(), etc. entirely -- this may not be doable as at least one package on PyPI using this (private) API * (maybe) intern the strings during runtime init https://bugs.python.org/issue46541
2022-02-08 16:39:07 -04:00
int r = PyObject_RichCompareBool(name, &_Py_ID(__dict__), Py_EQ);
if (r == -1) {
goto err;
}
if (r == 1) {
PyErr_Format(PyExc_AttributeError,
"'%.100s' object attribute '%U' is read-only",
Py_TYPE(self)->tp_name, name);
goto err;
}
int st =
_PyObject_GenericSetAttrWithDict((PyObject *)self, name, v, ldict);
Py_DECREF(ldict);
return st;
err:
Py_XDECREF(ldict);
return -1;
}
static PyObject *local_getattro(localobject *, PyObject *);
static PyMemberDef local_type_members[] = {
{"__weaklistoffset__", Py_T_PYSSIZET, offsetof(localobject, weakreflist), Py_READONLY},
{NULL},
};
static PyType_Slot local_type_slots[] = {
{Py_tp_dealloc, (destructor)local_dealloc},
{Py_tp_getattro, (getattrofunc)local_getattro},
{Py_tp_setattro, (setattrofunc)local_setattro},
{Py_tp_doc, "_local()\n--\n\nThread-local data"},
{Py_tp_traverse, (traverseproc)local_traverse},
{Py_tp_clear, (inquiry)local_clear},
{Py_tp_new, local_new},
{Py_tp_members, local_type_members},
{0, 0}
};
static PyType_Spec local_type_spec = {
.name = "_thread._local",
.basicsize = sizeof(localobject),
.flags = (Py_TPFLAGS_DEFAULT | Py_TPFLAGS_BASETYPE | Py_TPFLAGS_HAVE_GC |
Py_TPFLAGS_IMMUTABLETYPE),
.slots = local_type_slots,
};
static PyObject *
local_getattro(localobject *self, PyObject *name)
{
PyObject *module = PyType_GetModuleByDef(Py_TYPE(self), &thread_module);
assert(module != NULL);
thread_module_state *state = get_thread_state(module);
PyObject *ldict = _ldict(self, state);
if (ldict == NULL)
return NULL;
bpo-46541: Replace core use of _Py_IDENTIFIER() with statically initialized global objects. (gh-30928) We're no longer using _Py_IDENTIFIER() (or _Py_static_string()) in any core CPython code. It is still used in a number of non-builtin stdlib modules. The replacement is: PyUnicodeObject (not pointer) fields under _PyRuntimeState, statically initialized as part of _PyRuntime. A new _Py_GET_GLOBAL_IDENTIFIER() macro facilitates lookup of the fields (along with _Py_GET_GLOBAL_STRING() for non-identifier strings). https://bugs.python.org/issue46541#msg411799 explains the rationale for this change. The core of the change is in: * (new) Include/internal/pycore_global_strings.h - the declarations for the global strings, along with the macros * Include/internal/pycore_runtime_init.h - added the static initializers for the global strings * Include/internal/pycore_global_objects.h - where the struct in pycore_global_strings.h is hooked into _PyRuntimeState * Tools/scripts/generate_global_objects.py - added generation of the global string declarations and static initializers I've also added a --check flag to generate_global_objects.py (along with make check-global-objects) to check for unused global strings. That check is added to the PR CI config. The remainder of this change updates the core code to use _Py_GET_GLOBAL_IDENTIFIER() instead of _Py_IDENTIFIER() and the related _Py*Id functions (likewise for _Py_GET_GLOBAL_STRING() instead of _Py_static_string()). This includes adding a few functions where there wasn't already an alternative to _Py*Id(), replacing the _Py_Identifier * parameter with PyObject *. The following are not changed (yet): * stop using _Py_IDENTIFIER() in the stdlib modules * (maybe) get rid of _Py_IDENTIFIER(), etc. entirely -- this may not be doable as at least one package on PyPI using this (private) API * (maybe) intern the strings during runtime init https://bugs.python.org/issue46541
2022-02-08 16:39:07 -04:00
int r = PyObject_RichCompareBool(name, &_Py_ID(__dict__), Py_EQ);
if (r == 1) {
return ldict;
}
if (r == -1) {
Py_DECREF(ldict);
return NULL;
}
if (!Py_IS_TYPE(self, state->local_type)) {
/* use generic lookup for subtypes */
PyObject *res =
_PyObject_GenericGetAttrWithDict((PyObject *)self, name, ldict, 0);
Py_DECREF(ldict);
return res;
}
/* Optimization: just look in dict ourselves */
PyObject *value;
if (PyDict_GetItemRef(ldict, name, &value) != 0) {
// found or error
Py_DECREF(ldict);
return value;
}
/* Fall back on generic to get __class__ and __dict__ */
PyObject *res =
_PyObject_GenericGetAttrWithDict((PyObject *)self, name, ldict, 0);
Py_DECREF(ldict);
return res;
}
/* Called when a dummy is destroyed, indicating that the owning thread is being
* cleared. */
static PyObject *
clear_locals(PyObject *locals_and_key, PyObject *dummyweakref)
{
PyObject *localweakref = PyTuple_GetItem(locals_and_key, 0);
localobject *self = (localobject *)_PyWeakref_GET_REF(localweakref);
if (self == NULL) {
Py_RETURN_NONE;
}
/* If the thread-local object is still alive and not being cleared,
remove the corresponding local dict */
if (self->localdicts != NULL) {
PyObject *key = PyTuple_GetItem(locals_and_key, 1);
if (PyDict_Pop(self->localdicts, key, NULL) < 0) {
PyErr_WriteUnraisable((PyObject*)self);
}
}
if (self->thread_watchdogs != NULL) {
if (PySet_Discard(self->thread_watchdogs, dummyweakref) < 0) {
PyErr_WriteUnraisable((PyObject *)self);
}
}
Py_DECREF(self);
Py_RETURN_NONE;
}
/* Module functions */
static PyObject *
thread_daemon_threads_allowed(PyObject *module, PyObject *Py_UNUSED(ignored))
{
PyInterpreterState *interp = _PyInterpreterState_GET();
if (interp->feature_flags & Py_RTFLAGS_DAEMON_THREADS) {
Py_RETURN_TRUE;
}
else {
Py_RETURN_FALSE;
}
}
PyDoc_STRVAR(daemon_threads_allowed_doc,
"daemon_threads_allowed($module, /)\n\
--\n\
\n\
Return True if daemon threads are allowed in the current interpreter,\n\
and False otherwise.\n");
static int
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
do_start_new_thread(thread_module_state *state, PyObject *func, PyObject *args,
PyObject *kwargs, ThreadHandle *handle, int daemon)
{
PyInterpreterState *interp = _PyInterpreterState_GET();
if (!_PyInterpreterState_HasFeature(interp, Py_RTFLAGS_THREADS)) {
PyErr_SetString(PyExc_RuntimeError,
"thread is not supported for isolated subinterpreters");
return -1;
}
if (_PyInterpreterState_GetFinalizing(interp) != NULL) {
PyErr_SetString(PyExc_PythonFinalizationError,
"can't create new thread at interpreter shutdown");
return -1;
}
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
if (!daemon) {
// Add the handle before starting the thread to avoid adding a handle
// to a thread that has already finished (i.e. if the thread finishes
// before the call to `ThreadHandle_start()` below returns).
add_to_shutdown_handles(state, handle);
}
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
if (ThreadHandle_start(handle, func, args, kwargs) < 0) {
if (!daemon) {
remove_from_shutdown_handles(handle);
}
return -1;
}
return 0;
}
1996-12-16 20:05:22 -04:00
static PyObject *
thread_PyThread_start_new_thread(PyObject *module, PyObject *fargs)
{
PyObject *func, *args, *kwargs = NULL;
thread_module_state *state = get_thread_state(module);
if (!PyArg_UnpackTuple(fargs, "start_new_thread", 2, 3,
&func, &args, &kwargs))
return NULL;
if (!PyCallable_Check(func)) {
PyErr_SetString(PyExc_TypeError,
"first arg must be callable");
return NULL;
}
if (!PyTuple_Check(args)) {
PyErr_SetString(PyExc_TypeError,
"2nd arg must be a tuple");
return NULL;
}
if (kwargs != NULL && !PyDict_Check(kwargs)) {
PyErr_SetString(PyExc_TypeError,
"optional 3rd arg must be a dictionary");
return NULL;
}
if (PySys_Audit("_thread.start_new_thread", "OOO",
func, args, kwargs ? kwargs : Py_None) < 0) {
return NULL;
}
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
ThreadHandle *handle = ThreadHandle_new();
if (handle == NULL) {
return NULL;
}
int st =
do_start_new_thread(state, func, args, kwargs, handle, /*daemon=*/1);
if (st < 0) {
ThreadHandle_decref(handle);
return NULL;
}
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
PyThread_ident_t ident = ThreadHandle_ident(handle);
ThreadHandle_decref(handle);
return PyLong_FromUnsignedLongLong(ident);
}
PyDoc_STRVAR(start_new_thread_doc,
"start_new_thread($module, function, args, kwargs={}, /)\n\
--\n\
\n\
Start a new thread and return its identifier.\n\
\n\
The thread will call the function with positional arguments from the\n\
tuple args and keyword arguments taken from the optional dictionary\n\
kwargs. The thread exits when the function returns; the return value\n\
is ignored. The thread will also exit when the function raises an\n\
unhandled exception; a stack trace will be printed unless the exception\n\
is SystemExit.");
PyDoc_STRVAR(start_new_doc,
"start_new($module, function, args, kwargs={}, /)\n\
--\n\
\n\
An obsolete synonym of start_new_thread().");
static PyObject *
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
thread_PyThread_start_joinable_thread(PyObject *module, PyObject *fargs,
PyObject *fkwargs)
{
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
static char *keywords[] = {"function", "handle", "daemon", NULL};
PyObject *func = NULL;
int daemon = 1;
thread_module_state *state = get_thread_state(module);
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
PyObject *hobj = NULL;
if (!PyArg_ParseTupleAndKeywords(fargs, fkwargs,
"O|Op:start_joinable_thread", keywords,
&func, &hobj, &daemon)) {
return NULL;
}
if (!PyCallable_Check(func)) {
PyErr_SetString(PyExc_TypeError,
"thread function must be callable");
return NULL;
}
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
if (hobj == NULL) {
hobj = Py_None;
}
else if (hobj != Py_None && !Py_IS_TYPE(hobj, state->thread_handle_type)) {
PyErr_SetString(PyExc_TypeError, "'handle' must be a _ThreadHandle");
return NULL;
}
if (PySys_Audit("_thread.start_joinable_thread", "OiO", func, daemon,
hobj) < 0) {
return NULL;
}
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
if (hobj == Py_None) {
hobj = (PyObject *)PyThreadHandleObject_new(state->thread_handle_type);
if (hobj == NULL) {
return NULL;
}
}
else {
Py_INCREF(hobj);
}
PyObject* args = PyTuple_New(0);
if (args == NULL) {
return NULL;
}
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
int st = do_start_new_thread(state, func, args,
/*kwargs=*/ NULL, ((PyThreadHandleObject*)hobj)->handle, daemon);
Py_DECREF(args);
if (st < 0) {
Py_DECREF(hobj);
return NULL;
}
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
return (PyObject *) hobj;
}
PyDoc_STRVAR(start_joinable_doc,
"start_joinable_thread($module, /, function, handle=None, daemon=True)\n\
--\n\
\n\
*For internal use only*: start a new thread.\n\
\n\
Like start_new_thread(), this starts a new thread calling the given function.\n\
Unlike start_new_thread(), this returns a handle object with methods to join\n\
or detach the given thread.\n\
This function is not for third-party code, please use the\n\
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
`threading` module instead. During finalization the runtime will not wait for\n\
the thread to exit if daemon is True. If handle is provided it must be a\n\
newly created thread._ThreadHandle instance.");
1996-12-16 20:05:22 -04:00
static PyObject *
thread_PyThread_exit_thread(PyObject *self, PyObject *Py_UNUSED(ignored))
{
PyErr_SetNone(PyExc_SystemExit);
return NULL;
}
2002-06-13 17:33:02 -03:00
PyDoc_STRVAR(exit_doc,
"exit($module, /)\n\
--\n\
\n\
This is synonymous to ``raise SystemExit''. It will cause the current\n\
2002-06-13 17:33:02 -03:00
thread to exit silently unless the exception is caught.");
PyDoc_STRVAR(exit_thread_doc,
"exit_thread($module, /)\n\
--\n\
\n\
An obsolete synonym of exit().");
2003-06-13 18:59:45 -03:00
static PyObject *
thread_PyThread_interrupt_main(PyObject *self, PyObject *args)
2003-06-13 18:59:45 -03:00
{
int signum = SIGINT;
if (!PyArg_ParseTuple(args, "|i:signum", &signum)) {
return NULL;
}
if (PyErr_SetInterruptEx(signum)) {
PyErr_SetString(PyExc_ValueError, "signal number out of range");
return NULL;
}
Py_RETURN_NONE;
2003-06-13 18:59:45 -03:00
}
PyDoc_STRVAR(interrupt_doc,
"interrupt_main($module, signum=signal.SIGINT, /)\n\
--\n\
\n\
Simulate the arrival of the given signal in the main thread,\n\
where the corresponding signal handler will be executed.\n\
If *signum* is omitted, SIGINT is assumed.\n\
A subthread can use this function to interrupt the main thread.\n\
2003-06-13 18:59:45 -03:00
\n\
Note: the default signal handler for SIGINT raises ``KeyboardInterrupt``."
2003-06-13 18:59:45 -03:00
);
1996-12-16 20:05:22 -04:00
static PyObject *
thread_PyThread_allocate_lock(PyObject *module, PyObject *Py_UNUSED(ignored))
{
return (PyObject *) newlockobject(module);
}
PyDoc_STRVAR(allocate_lock_doc,
"allocate_lock($module, /)\n\
--\n\
\n\
Create a new lock object. See help(type(threading.Lock())) for\n\
information about locks.");
PyDoc_STRVAR(allocate_doc,
"allocate($module, /)\n\
--\n\
\n\
An obsolete synonym of allocate_lock().");
1996-12-16 20:05:22 -04:00
static PyObject *
thread_get_ident(PyObject *self, PyObject *Py_UNUSED(ignored))
{
PyThread_ident_t ident = PyThread_get_thread_ident_ex();
if (ident == PYTHREAD_INVALID_THREAD_ID) {
PyErr_SetString(ThreadError, "no current thread ident");
return NULL;
}
return PyLong_FromUnsignedLongLong(ident);
}
2002-06-13 17:33:02 -03:00
PyDoc_STRVAR(get_ident_doc,
"get_ident($module, /)\n\
--\n\
\n\
Return a non-zero integer that uniquely identifies the current thread\n\
amongst other threads that exist simultaneously.\n\
This may be used to identify per-thread resources.\n\
Even though on some platforms threads identities may appear to be\n\
allocated consecutive numbers starting at 1, this behavior should not\n\
be relied upon, and the number should be seen purely as a magic cookie.\n\
2002-06-13 17:33:02 -03:00
A thread's identity may be reused for another thread after it exits.");
#ifdef PY_HAVE_THREAD_NATIVE_ID
static PyObject *
thread_get_native_id(PyObject *self, PyObject *Py_UNUSED(ignored))
{
unsigned long native_id = PyThread_get_thread_native_id();
return PyLong_FromUnsignedLong(native_id);
}
PyDoc_STRVAR(get_native_id_doc,
"get_native_id($module, /)\n\
--\n\
\n\
Return a non-negative integer identifying the thread as reported\n\
by the OS (kernel). This may be used to uniquely identify a\n\
particular thread within a system.");
#endif
static PyObject *
thread__count(PyObject *self, PyObject *Py_UNUSED(ignored))
{
PyInterpreterState *interp = _PyInterpreterState_GET();
return PyLong_FromSsize_t(_Py_atomic_load_ssize(&interp->threads.count));
}
PyDoc_STRVAR(_count_doc,
"_count($module, /)\n\
--\n\
\n\
Return the number of currently running Python threads, excluding\n\
the main thread. The returned number comprises all threads created\n\
through `start_new_thread()` as well as `threading.Thread`, and not\n\
yet finished.\n\
\n\
This function is meant for internal and specialized purposes only.\n\
In most applications `threading.enumerate()` should be used instead.");
Merged revisions 46753-51188 via svnmerge from svn+ssh://pythondev@svn.python.org/python/trunk ........ r46755 | brett.cannon | 2006-06-08 18:23:04 +0200 (Thu, 08 Jun 2006) | 4 lines Make binascii.hexlify() use s# for its arguments instead of t# to actually match its documentation stating it accepts any read-only buffer. ........ r46757 | brett.cannon | 2006-06-08 19:00:45 +0200 (Thu, 08 Jun 2006) | 8 lines Buffer objects would return the read or write buffer for a wrapped object when the char buffer was requested. Now it actually returns the char buffer if available or raises a TypeError if it isn't (as is raised for the other buffer types if they are not present but requested). Not a backport candidate since it does change semantics of the buffer object (although it could be argued this is enough of a bug to bother backporting). ........ r46760 | andrew.kuchling | 2006-06-09 03:10:17 +0200 (Fri, 09 Jun 2006) | 1 line Update functools section ........ r46762 | tim.peters | 2006-06-09 04:11:02 +0200 (Fri, 09 Jun 2006) | 6 lines Whitespace normalization. Since test_file is implicated in mysterious test failures when followed by test_optparse, if I had any brains I'd look at the checkin that last changed test_file ;-) ........ r46763 | tim.peters | 2006-06-09 05:09:42 +0200 (Fri, 09 Jun 2006) | 5 lines To boost morale :-), force test_optparse to run immediately after test_file until we can figure out how to fix it. (See python-dev; at the moment we don't even know which checkin caused the problem.) ........ r46764 | tim.peters | 2006-06-09 05:51:41 +0200 (Fri, 09 Jun 2006) | 6 lines AutoFileTests.tearDown(): Removed mysterious undocumented try/except. Remove TESTFN. Throughout: used open() instead of file(), and wrapped long lines. ........ r46765 | tim.peters | 2006-06-09 06:02:06 +0200 (Fri, 09 Jun 2006) | 8 lines testUnicodeOpen(): I have no idea why, but making this test clean up after itself appears to fix the test failures when test_optparse follows test_file. test_main(): Get rid of TESTFN no matter what. That's also enough to fix the mystery failures. Doesn't hurt to fix them twice :-) ........ r46766 | tim.peters | 2006-06-09 07:12:40 +0200 (Fri, 09 Jun 2006) | 6 lines Remove the temporary hack to force test_optparse to run immediately after test_file. At least 8 buildbot boxes passed since the underlying problem got fixed, and they all failed before the fix, so there's no point to this anymore. ........ r46767 | neal.norwitz | 2006-06-09 07:54:18 +0200 (Fri, 09 Jun 2006) | 1 line Fix grammar and reflow ........ r46769 | andrew.kuchling | 2006-06-09 12:22:35 +0200 (Fri, 09 Jun 2006) | 1 line Markup fix ........ r46773 | andrew.kuchling | 2006-06-09 15:15:57 +0200 (Fri, 09 Jun 2006) | 1 line [Bug #1472827] Make saxutils.XMLGenerator handle \r\n\t in attribute values by escaping them properly. 2.4 bugfix candidate. ........ r46778 | kristjan.jonsson | 2006-06-09 18:28:01 +0200 (Fri, 09 Jun 2006) | 2 lines Turn off warning about deprecated CRT functions on for VisualStudio .NET 2005. Make the definition #ARRAYSIZE conditional. VisualStudio .NET 2005 already has it defined using a better gimmick. ........ r46779 | phillip.eby | 2006-06-09 18:40:18 +0200 (Fri, 09 Jun 2006) | 2 lines Import wsgiref into the stdlib, as of the external version 0.1-r2181. ........ r46783 | andrew.kuchling | 2006-06-09 18:44:40 +0200 (Fri, 09 Jun 2006) | 1 line Add note about XMLGenerator bugfix ........ r46784 | andrew.kuchling | 2006-06-09 18:46:51 +0200 (Fri, 09 Jun 2006) | 1 line Add note about wsgiref ........ r46785 | brett.cannon | 2006-06-09 19:05:48 +0200 (Fri, 09 Jun 2006) | 2 lines Fix inconsistency in naming within an enum. ........ r46787 | tim.peters | 2006-06-09 19:47:00 +0200 (Fri, 09 Jun 2006) | 2 lines Whitespace normalization. ........ r46792 | georg.brandl | 2006-06-09 20:29:52 +0200 (Fri, 09 Jun 2006) | 3 lines Test file.__exit__. ........ r46794 | brett.cannon | 2006-06-09 20:40:46 +0200 (Fri, 09 Jun 2006) | 2 lines svn:ignore .pyc and .pyo files. ........ r46795 | georg.brandl | 2006-06-09 20:45:48 +0200 (Fri, 09 Jun 2006) | 3 lines RFE #1491485: str/unicode.endswith()/startswith() now accept a tuple as first argument. ........ r46798 | andrew.kuchling | 2006-06-09 21:03:16 +0200 (Fri, 09 Jun 2006) | 1 line Describe startswith()/endswiith() change; add reminder about wsgiref ........ r46799 | tim.peters | 2006-06-09 21:24:44 +0200 (Fri, 09 Jun 2006) | 11 lines Implementing a happy idea from Georg Brandl: make runtest() try to clean up files and directories the tests often leave behind by mistake. This is the first time in history I don't have a bogus "db_home" directory after running the tests ;-) Also worked on runtest's docstring, to say something about all the arguments, and to document the non-obvious return values. New functions runtest_inner() and cleanup_test_droppings() in support of the above. ........ r46800 | andrew.kuchling | 2006-06-09 21:43:25 +0200 (Fri, 09 Jun 2006) | 1 line Remove unused variable ........ r46801 | andrew.kuchling | 2006-06-09 21:56:05 +0200 (Fri, 09 Jun 2006) | 1 line Add some wsgiref text ........ r46803 | thomas.heller | 2006-06-09 21:59:11 +0200 (Fri, 09 Jun 2006) | 1 line set eol-style svn property ........ r46804 | thomas.heller | 2006-06-09 22:01:01 +0200 (Fri, 09 Jun 2006) | 1 line set eol-style svn property ........ r46805 | georg.brandl | 2006-06-09 22:43:48 +0200 (Fri, 09 Jun 2006) | 3 lines Make use of new str.startswith/endswith semantics. Occurences in email and compiler were ignored due to backwards compat requirements. ........ r46806 | brett.cannon | 2006-06-10 00:31:23 +0200 (Sat, 10 Jun 2006) | 4 lines An object with __call__ as an attribute, when called, will have that attribute checked for __call__ itself, and will continue to look until it finds an object without the attribute. This can lead to an infinite recursion. Closes bug #532646, again. Will be backported. ........ r46808 | brett.cannon | 2006-06-10 00:45:54 +0200 (Sat, 10 Jun 2006) | 2 lines Fix bug introduced in rev. 46806 by not having variable declaration at the top of a block. ........ r46812 | georg.brandl | 2006-06-10 08:40:50 +0200 (Sat, 10 Jun 2006) | 4 lines Apply perky's fix for #1503157: "/".join([u"", u""]) raising OverflowError. Also improve error message on overflow. ........ r46817 | martin.v.loewis | 2006-06-10 10:14:03 +0200 (Sat, 10 Jun 2006) | 2 lines Port cygwin kill_python changes from 2.4 branch. ........ r46818 | armin.rigo | 2006-06-10 12:57:40 +0200 (Sat, 10 Jun 2006) | 4 lines SF bug #1503294. PyThreadState_GET() complains if the tstate is NULL, but only in debug mode. ........ r46819 | martin.v.loewis | 2006-06-10 14:23:46 +0200 (Sat, 10 Jun 2006) | 4 lines Patch #1495999: Part two of Windows CE changes. - update header checks, using autoconf - provide dummies for getenv, environ, and GetVersion - adjust MSC_VER check in socketmodule.c ........ r46820 | skip.montanaro | 2006-06-10 16:09:11 +0200 (Sat, 10 Jun 2006) | 1 line document the class, not its initializer ........ r46821 | greg.ward | 2006-06-10 18:40:01 +0200 (Sat, 10 Jun 2006) | 4 lines Sync with Optik docs (rev 518): * restore "Extending optparse" section * document ALWAYS_TYPED_ACTIONS (SF #1449311) ........ r46824 | thomas.heller | 2006-06-10 21:51:46 +0200 (Sat, 10 Jun 2006) | 8 lines Upgrade to ctypes version 0.9.9.7. Summary of changes: - support for 'variable sized' data - support for anonymous structure/union fields - fix severe bug with certain arrays or structures containing more than 256 fields ........ r46825 | thomas.heller | 2006-06-10 21:55:36 +0200 (Sat, 10 Jun 2006) | 8 lines Upgrade to ctypes version 0.9.9.7. Summary of changes: - support for 'variable sized' data - support for anonymous structure/union fields - fix severe bug with certain arrays or structures containing more than 256 fields ........ r46826 | fred.drake | 2006-06-10 22:01:34 +0200 (Sat, 10 Jun 2006) | 4 lines SF patch #1303595: improve description of __builtins__, explaining how it varies between __main__ and other modules, and strongly suggest not touching it but using __builtin__ if absolutely necessary ........ r46827 | fred.drake | 2006-06-10 22:02:58 +0200 (Sat, 10 Jun 2006) | 1 line credit for SF patch #1303595 ........ r46831 | thomas.heller | 2006-06-10 22:29:34 +0200 (Sat, 10 Jun 2006) | 2 lines New docs for ctypes. ........ r46834 | thomas.heller | 2006-06-10 23:07:19 +0200 (Sat, 10 Jun 2006) | 1 line Fix a wrong printf format. ........ r46835 | thomas.heller | 2006-06-10 23:17:58 +0200 (Sat, 10 Jun 2006) | 1 line Fix the second occurrence of the problematic printf format. ........ r46837 | thomas.heller | 2006-06-10 23:56:03 +0200 (Sat, 10 Jun 2006) | 1 line Don't use C++ comment. ........ r46838 | thomas.heller | 2006-06-11 00:01:50 +0200 (Sun, 11 Jun 2006) | 1 line Handle failure of PyMem_Realloc. ........ r46839 | skip.montanaro | 2006-06-11 00:38:13 +0200 (Sun, 11 Jun 2006) | 2 lines Suppress warning on MacOSX about possible use before set of proc. ........ r46840 | tim.peters | 2006-06-11 00:51:45 +0200 (Sun, 11 Jun 2006) | 8 lines shuffle() doscstring: Removed warning about sequence length versus generator period. While this was a real weakness of the older WH generator for lists with just a few dozen elements, and so could potentially bite the naive ;-), the Twister should show excellent behavior up to at least 600 elements. Module docstring: reflowed some jarringly short lines. ........ r46844 | greg.ward | 2006-06-11 02:40:49 +0200 (Sun, 11 Jun 2006) | 4 lines Bug #1361643: fix textwrap.dedent() so it handles tabs appropriately, i.e. do *not* expand tabs, but treat them as whitespace that is not equivalent to spaces. Add a couple of test cases. Clarify docs. ........ r46850 | neal.norwitz | 2006-06-11 07:44:18 +0200 (Sun, 11 Jun 2006) | 5 lines Fix Coverity # 146. newDBSequenceObject would deref dbobj, so it can't be NULL. We know it's not NULL from the ParseTuple and DbObject_Check will verify it's not NULL. ........ r46851 | neal.norwitz | 2006-06-11 07:45:25 +0200 (Sun, 11 Jun 2006) | 4 lines Wrap some long lines Top/Bottom factor out some common expressions Add a XXX comment about widing offset. ........ r46852 | neal.norwitz | 2006-06-11 07:45:47 +0200 (Sun, 11 Jun 2006) | 1 line Add versionadded to doc ........ r46853 | neal.norwitz | 2006-06-11 07:47:14 +0200 (Sun, 11 Jun 2006) | 3 lines Update doc to make it agree with code. Bottom factor out some common code. ........ r46854 | neal.norwitz | 2006-06-11 07:48:14 +0200 (Sun, 11 Jun 2006) | 3 lines f_code can't be NULL based on Frame_New and other code that derefs it. So there doesn't seem to be much point to checking here. ........ r46855 | neal.norwitz | 2006-06-11 09:26:27 +0200 (Sun, 11 Jun 2006) | 1 line Fix errors found by pychecker ........ r46856 | neal.norwitz | 2006-06-11 09:26:50 +0200 (Sun, 11 Jun 2006) | 1 line warnings was imported at module scope, no need to import again ........ r46857 | neal.norwitz | 2006-06-11 09:27:56 +0200 (Sun, 11 Jun 2006) | 5 lines Fix errors found by pychecker. I think these changes are correct, but I'm not sure. Could someone who knows how this module works test it? It can at least start on the cmd line. ........ r46858 | neal.norwitz | 2006-06-11 10:35:14 +0200 (Sun, 11 Jun 2006) | 1 line Fix errors found by pychecker ........ r46859 | ronald.oussoren | 2006-06-11 16:33:36 +0200 (Sun, 11 Jun 2006) | 4 lines This patch improves the L&F of IDLE on OSX. The changes are conditionalized on being in an IDLE.app bundle on darwin. This does a slight reorganisation of the menus and adds support for file-open events. ........ r46860 | greg.ward | 2006-06-11 16:42:41 +0200 (Sun, 11 Jun 2006) | 1 line SF #1366250: optparse docs: fix inconsistency in variable name; minor tweaks. ........ r46861 | greg.ward | 2006-06-11 18:24:11 +0200 (Sun, 11 Jun 2006) | 3 lines Bug #1498146: fix optparse to handle Unicode strings in option help, description, and epilog. ........ r46862 | thomas.heller | 2006-06-11 19:04:22 +0200 (Sun, 11 Jun 2006) | 2 lines Release the GIL during COM method calls, to avoid deadlocks in Python coded COM objects. ........ r46863 | tim.peters | 2006-06-11 21:42:51 +0200 (Sun, 11 Jun 2006) | 2 lines Whitespace normalization. ........ r46864 | tim.peters | 2006-06-11 21:43:49 +0200 (Sun, 11 Jun 2006) | 2 lines Add missing svn:eol-style property to text files. ........ r46865 | ronald.oussoren | 2006-06-11 21:45:57 +0200 (Sun, 11 Jun 2006) | 2 lines Remove message about using make frameworkinstall, that's no longer necesssary ........ r46866 | ronald.oussoren | 2006-06-11 22:23:29 +0200 (Sun, 11 Jun 2006) | 2 lines Use configure to substitute the correct prefix instead of hardcoding ........ r46867 | ronald.oussoren | 2006-06-11 22:24:45 +0200 (Sun, 11 Jun 2006) | 4 lines - Change fixapplepython23.py to ensure that it will run with /usr/bin/python on intel macs. - Fix some minor problems in the installer for OSX ........ r46868 | neal.norwitz | 2006-06-11 22:25:56 +0200 (Sun, 11 Jun 2006) | 5 lines Try to fix several networking tests. The problem is that if hosts have a search path setup, some of these hosts resolve to the wrong address. By appending a period to the hostname, the hostname should only resolve to what we want it to resolve to. Hopefully this doesn't break different bots. ........ r46869 | neal.norwitz | 2006-06-11 22:42:02 +0200 (Sun, 11 Jun 2006) | 7 lines Try to fix another networking test. The problem is that if hosts have a search path setup, some of these hosts resolve to the wrong address. By appending a period to the hostname, the hostname should only resolve to what we want it to resolve to. Hopefully this doesn't break different bots. Also add more info to failure message to aid debugging test failure. ........ r46870 | neal.norwitz | 2006-06-11 22:46:46 +0200 (Sun, 11 Jun 2006) | 4 lines Fix test on PPC64 buildbot. It raised an IOError (really an URLError which derives from an IOError). That seems valid. Env Error includes both OSError and IOError, so this seems like a reasonable fix. ........ r46871 | tim.peters | 2006-06-11 22:52:59 +0200 (Sun, 11 Jun 2006) | 10 lines compare_generic_iter(): Fixed the failure of test_wsgiref's testFileWrapper when running with -O. test_simple_validation_error still fails under -O. That appears to be because wsgiref's validate.py uses `assert` statements all over the place to check arguments for sanity. That should all be changed (it's not a logical error in the software if a user passes bogus arguments, so this isn't a reasonable use for `assert` -- checking external preconditions should generally raise ValueError or TypeError instead, as appropriate). ........ r46872 | neal.norwitz | 2006-06-11 23:38:38 +0200 (Sun, 11 Jun 2006) | 1 line Get test to pass on S/390. Shout if you think this change is incorrect. ........ r46873 | neal.norwitz | 2006-06-12 04:05:55 +0200 (Mon, 12 Jun 2006) | 1 line Cleanup Py_ssize_t a little (get rid of second #ifdef) ........ r46874 | neal.norwitz | 2006-06-12 04:06:17 +0200 (Mon, 12 Jun 2006) | 1 line Fix some Py_ssize_t issues ........ r46875 | neal.norwitz | 2006-06-12 04:06:42 +0200 (Mon, 12 Jun 2006) | 1 line Fix some Py_ssize_t issues ........ r46876 | neal.norwitz | 2006-06-12 04:07:24 +0200 (Mon, 12 Jun 2006) | 2 lines Cleanup: Remove import of types to get StringTypes, we can just use basestring. ........ r46877 | neal.norwitz | 2006-06-12 04:07:57 +0200 (Mon, 12 Jun 2006) | 1 line Don't truncate if size_t is bigger than uint ........ r46878 | neal.norwitz | 2006-06-12 04:08:41 +0200 (Mon, 12 Jun 2006) | 1 line Don't leak the list object if there's an error allocating the item storage. Backport candidate ........ r46879 | neal.norwitz | 2006-06-12 04:09:03 +0200 (Mon, 12 Jun 2006) | 1 line Fix typo. Backport if anyone cares. :-) ........ r46880 | neal.norwitz | 2006-06-12 04:09:34 +0200 (Mon, 12 Jun 2006) | 1 line Fix indentation of case and a Py_ssize_t issue. ........ r46881 | neal.norwitz | 2006-06-12 04:11:18 +0200 (Mon, 12 Jun 2006) | 3 lines Get rid of f_restricted too. Doc the other 4 ints that were already removed at the NeedForSpeed sprint. ........ r46882 | neal.norwitz | 2006-06-12 04:13:21 +0200 (Mon, 12 Jun 2006) | 1 line Fix the socket tests so they can be run concurrently. Backport candidate ........ r46883 | neal.norwitz | 2006-06-12 04:16:10 +0200 (Mon, 12 Jun 2006) | 1 line i and j are initialized below when used. No need to do it twice ........ r46884 | neal.norwitz | 2006-06-12 05:05:03 +0200 (Mon, 12 Jun 2006) | 1 line Remove unused import ........ r46885 | neal.norwitz | 2006-06-12 05:05:40 +0200 (Mon, 12 Jun 2006) | 1 line Impl ssize_t ........ r46886 | neal.norwitz | 2006-06-12 05:33:09 +0200 (Mon, 12 Jun 2006) | 6 lines Patch #1503046, Conditional compilation of zlib.(de)compressobj.copy copy is only in newer versions of zlib. This should allow zlibmodule to work with older versions like the Tru64 buildbot. ........ r46887 | phillip.eby | 2006-06-12 06:04:32 +0200 (Mon, 12 Jun 2006) | 2 lines Sync w/external release 0.1.2. Please see PEP 360 before making changes to external packages. ........ r46888 | martin.v.loewis | 2006-06-12 06:26:31 +0200 (Mon, 12 Jun 2006) | 2 lines Get rid of function pointer cast. ........ r46889 | thomas.heller | 2006-06-12 08:05:57 +0200 (Mon, 12 Jun 2006) | 3 lines I don't know how that happend, but the entire file contents was duplicated. Thanks to Simon Percivall for the heads up. ........ r46890 | nick.coghlan | 2006-06-12 10:19:37 +0200 (Mon, 12 Jun 2006) | 1 line Fix site module docstring to match the code ........ r46891 | nick.coghlan | 2006-06-12 10:23:02 +0200 (Mon, 12 Jun 2006) | 1 line Fix site module docstring to match the code for Mac OSX, too ........ r46892 | nick.coghlan | 2006-06-12 10:27:13 +0200 (Mon, 12 Jun 2006) | 1 line The site module documentation also described the Windows behaviour incorrectly. ........ r46893 | nick.coghlan | 2006-06-12 12:17:11 +0200 (Mon, 12 Jun 2006) | 1 line Make the -m switch conform to the documentation of sys.path by behaving like the -c switch ........ r46894 | kristjan.jonsson | 2006-06-12 17:45:12 +0200 (Mon, 12 Jun 2006) | 2 lines Fix the CRT argument error handling for VisualStudio .NET 2005. Install a CRT error handler and disable the assertion for debug builds. This causes CRT to set errno to EINVAL. This update fixes crash cases in the test suite where the default CRT error handler would cause process exit. ........ r46899 | thomas.heller | 2006-06-12 22:56:48 +0200 (Mon, 12 Jun 2006) | 1 line Add pep-291 compatibility markers. ........ r46901 | ka-ping.yee | 2006-06-13 01:47:52 +0200 (Tue, 13 Jun 2006) | 5 lines Add the uuid module. This module has been tested so far on Windows XP (Python 2.4 and 2.5a2), Mac OS X (Python 2.3, 2.4, and 2.5a2), and Linux (Python 2.4 and 2.5a2). ........ r46902 | tim.peters | 2006-06-13 02:30:01 +0200 (Tue, 13 Jun 2006) | 2 lines Whitespace normalization. ........ r46903 | tim.peters | 2006-06-13 02:30:50 +0200 (Tue, 13 Jun 2006) | 2 lines Added missing svn:eol-style property to text files. ........ r46905 | tim.peters | 2006-06-13 05:30:07 +0200 (Tue, 13 Jun 2006) | 5 lines get_matching_blocks(): rewrote code & comments so they match; added more comments about why it's this way at all; and removed what looked like needless expense (sorting (i, j, k) triples directly should give exactly the same order as sorting (i, (i, j, k)) pairs). ........ r46906 | neal.norwitz | 2006-06-13 06:08:53 +0200 (Tue, 13 Jun 2006) | 1 line Don't fail if another process is listening on our port. ........ r46908 | neal.norwitz | 2006-06-13 10:28:19 +0200 (Tue, 13 Jun 2006) | 2 lines Initialize the type object so pychecker can't crash the interpreter. ........ r46909 | neal.norwitz | 2006-06-13 10:41:06 +0200 (Tue, 13 Jun 2006) | 1 line Verify the crash due to EncodingMap not initialized does not return ........ r46910 | thomas.heller | 2006-06-13 10:56:14 +0200 (Tue, 13 Jun 2006) | 3 lines Add some windows datatypes that were missing from this file, and add the aliases defined in windows header files for the structures. ........ r46911 | thomas.heller | 2006-06-13 11:40:14 +0200 (Tue, 13 Jun 2006) | 3 lines Add back WCHAR, UINT, DOUBLE, _LARGE_INTEGER, _ULARGE_INTEGER. VARIANT_BOOL is a special _ctypes data type, not c_short. ........ r46912 | ronald.oussoren | 2006-06-13 13:19:56 +0200 (Tue, 13 Jun 2006) | 4 lines Linecache contains support for PEP302 loaders, but fails to deal with loaders that return None to indicate that the module is valid but no source is available. This patch fixes that. ........ r46913 | andrew.kuchling | 2006-06-13 13:57:04 +0200 (Tue, 13 Jun 2006) | 1 line Mention uuid module ........ r46915 | walter.doerwald | 2006-06-13 14:02:12 +0200 (Tue, 13 Jun 2006) | 2 lines Fix passing errors to the encoder and decoder functions. ........ r46917 | walter.doerwald | 2006-06-13 14:04:43 +0200 (Tue, 13 Jun 2006) | 3 lines errors is an attribute in the incremental decoder not an argument. ........ r46919 | andrew.macintyre | 2006-06-13 17:04:24 +0200 (Tue, 13 Jun 2006) | 11 lines Patch #1454481: Make thread stack size runtime tunable. Heavily revised, comprising revisions: 46640 - original trunk revision (backed out in r46655) 46647 - markup fix (backed out in r46655) 46692:46918 merged from branch aimacintyre-sf1454481 branch tested on buildbots (Windows buildbots had problems not related to these changes). ........ r46920 | brett.cannon | 2006-06-13 18:06:55 +0200 (Tue, 13 Jun 2006) | 2 lines Remove unused variable. ........ r46921 | andrew.kuchling | 2006-06-13 18:41:41 +0200 (Tue, 13 Jun 2006) | 1 line Add ability to set stack size ........ r46923 | marc-andre.lemburg | 2006-06-13 19:04:26 +0200 (Tue, 13 Jun 2006) | 2 lines Update pybench to version 2.0. ........ r46924 | marc-andre.lemburg | 2006-06-13 19:07:14 +0200 (Tue, 13 Jun 2006) | 2 lines Revert wrong svn copy. ........ r46925 | andrew.macintyre | 2006-06-13 19:14:36 +0200 (Tue, 13 Jun 2006) | 2 lines fix exception usage ........ r46927 | tim.peters | 2006-06-13 20:37:07 +0200 (Tue, 13 Jun 2006) | 2 lines Whitespace normalization. ........ r46928 | marc-andre.lemburg | 2006-06-13 20:56:56 +0200 (Tue, 13 Jun 2006) | 9 lines Updated to pybench 2.0. See svn.python.org/external/pybench-2.0 for the original import of that version. Note that platform.py was not copied over from pybench-2.0 since it is already part of Python 2.5. ........ r46929 | andrew.macintyre | 2006-06-13 21:02:35 +0200 (Tue, 13 Jun 2006) | 5 lines Increase the small thread stack size to get the test to pass reliably on the one buildbot that insists on more than 32kB of thread stack. ........ r46930 | marc-andre.lemburg | 2006-06-13 21:20:07 +0200 (Tue, 13 Jun 2006) | 2 lines Whitespace normalization. ........ r46931 | thomas.heller | 2006-06-13 22:18:43 +0200 (Tue, 13 Jun 2006) | 2 lines More docs for ctypes. ........ r46932 | brett.cannon | 2006-06-13 23:34:24 +0200 (Tue, 13 Jun 2006) | 2 lines Ignore .pyc and .pyo files in Pybench. ........ r46933 | brett.cannon | 2006-06-13 23:46:41 +0200 (Tue, 13 Jun 2006) | 7 lines If a classic class defined a __coerce__() method that just returned its two arguments in reverse, the interpreter would infinitely recourse trying to get a coercion that worked. So put in a recursion check after a coercion is made and the next call to attempt to use the coerced values. Fixes bug #992017 and closes crashers/coerce.py . ........ r46936 | gerhard.haering | 2006-06-14 00:24:47 +0200 (Wed, 14 Jun 2006) | 3 lines Merged changes from external pysqlite 2.3.0 release. Documentation updates will follow in a few hours at the latest. Then we should be ready for beta1. ........ r46937 | brett.cannon | 2006-06-14 00:26:13 +0200 (Wed, 14 Jun 2006) | 2 lines Missed test for rev. 46933; infinite recursion from __coerce__() returning its arguments reversed. ........ r46938 | gerhard.haering | 2006-06-14 00:53:48 +0200 (Wed, 14 Jun 2006) | 2 lines Updated documentation for pysqlite 2.3.0 API. ........ r46939 | tim.peters | 2006-06-14 06:09:25 +0200 (Wed, 14 Jun 2006) | 10 lines SequenceMatcher.get_matching_blocks(): This now guarantees that adjacent triples in the result list describe non-adjacent matching blocks. That's _nice_ to have, and Guido said he wanted it. Not a bugfix candidate: Guido or not ;-), this changes visible endcase semantics (note that some tests had to change), and nothing about this was documented before. Since it was working as designed, and behavior was consistent with the docs, it wasn't "a bug". ........ r46940 | tim.peters | 2006-06-14 06:13:00 +0200 (Wed, 14 Jun 2006) | 2 lines Repaired typo in new comment. ........ r46941 | tim.peters | 2006-06-14 06:15:27 +0200 (Wed, 14 Jun 2006) | 2 lines Whitespace normalization. ........ r46942 | fred.drake | 2006-06-14 06:25:02 +0200 (Wed, 14 Jun 2006) | 3 lines - make some disabled tests run what they intend when enabled - remove some over-zealous triple-quoting ........ r46943 | fred.drake | 2006-06-14 07:04:47 +0200 (Wed, 14 Jun 2006) | 3 lines add tests for two cases that are handled correctly in the current code, but that SF patch 1504676 as written mis-handles ........ r46944 | fred.drake | 2006-06-14 07:15:51 +0200 (Wed, 14 Jun 2006) | 1 line explain an XXX in more detail ........ r46945 | martin.v.loewis | 2006-06-14 07:21:04 +0200 (Wed, 14 Jun 2006) | 1 line Patch #1455898: Incremental mode for "mbcs" codec. ........ r46946 | georg.brandl | 2006-06-14 08:08:31 +0200 (Wed, 14 Jun 2006) | 3 lines Bug #1339007: Shelf objects now don't raise an exception in their __del__ method when initialization failed. ........ r46948 | thomas.heller | 2006-06-14 08:18:15 +0200 (Wed, 14 Jun 2006) | 1 line Fix docstring. ........ r46949 | georg.brandl | 2006-06-14 08:29:07 +0200 (Wed, 14 Jun 2006) | 2 lines Bug #1501122: mention __gt__ &co in description of comparison order. ........ r46951 | thomas.heller | 2006-06-14 09:08:38 +0200 (Wed, 14 Jun 2006) | 1 line Write more docs. ........ r46952 | georg.brandl | 2006-06-14 10:31:39 +0200 (Wed, 14 Jun 2006) | 3 lines Bug #1153163: describe __add__ vs __radd__ behavior when adding objects of same type/of subclasses of the other. ........ r46954 | georg.brandl | 2006-06-14 10:42:11 +0200 (Wed, 14 Jun 2006) | 3 lines Bug #1202018: add some common mime.types locations. ........ r46955 | georg.brandl | 2006-06-14 10:50:03 +0200 (Wed, 14 Jun 2006) | 3 lines Bug #1117556: SimpleHTTPServer now tries to find and use the system's mime.types file for determining MIME types. ........ r46957 | thomas.heller | 2006-06-14 11:09:08 +0200 (Wed, 14 Jun 2006) | 1 line Document paramflags. ........ r46958 | thomas.heller | 2006-06-14 11:20:11 +0200 (Wed, 14 Jun 2006) | 1 line Add an __all__ list, since this module does 'from ctypes import *'. ........ r46959 | andrew.kuchling | 2006-06-14 15:59:15 +0200 (Wed, 14 Jun 2006) | 1 line Add item ........ r46961 | georg.brandl | 2006-06-14 18:46:43 +0200 (Wed, 14 Jun 2006) | 3 lines Bug #805015: doc error in PyUnicode_FromEncodedObject. ........ r46962 | gerhard.haering | 2006-06-15 00:28:37 +0200 (Thu, 15 Jun 2006) | 10 lines - Added version checks in C code to make sure we don't trigger bugs in older SQLite versions. - Added version checks in test suite so that we don't execute tests that we know will fail with older (buggy) SQLite versions. Now, all tests should run against all SQLite versions from 3.0.8 until 3.3.6 (latest one now). The sqlite3 module can be built against all these SQLite versions and the sqlite3 module does its best to not trigger bugs in SQLite, but using SQLite 3.3.3 or later is recommended. ........ r46963 | tim.peters | 2006-06-15 00:38:13 +0200 (Thu, 15 Jun 2006) | 2 lines Whitespace normalization. ........ r46964 | neal.norwitz | 2006-06-15 06:54:29 +0200 (Thu, 15 Jun 2006) | 9 lines Speculative checkin (requires approval of Gerhard Haering) This backs out the test changes in 46962 which prevented crashes by not running the tests via a version check. All the version checks added in that rev were removed from the tests. Code was added to the error handler in connection.c that seems to work with older versions of sqlite including 3.1.3. ........ r46965 | neal.norwitz | 2006-06-15 07:55:49 +0200 (Thu, 15 Jun 2006) | 1 line Try to narrow window of failure on slow/busy boxes (ppc64 buildbot) ........ r46966 | martin.v.loewis | 2006-06-15 08:45:05 +0200 (Thu, 15 Jun 2006) | 2 lines Make import/lookup of mbcs fail on non-Windows systems. ........ r46967 | ronald.oussoren | 2006-06-15 10:14:18 +0200 (Thu, 15 Jun 2006) | 2 lines Patch #1446489 (zipfile: support for ZIP64) ........ r46968 | neal.norwitz | 2006-06-15 10:16:44 +0200 (Thu, 15 Jun 2006) | 6 lines Re-revert this change. Install the version check and don't run the test until Gerhard has time to fully debug the issue. This affects versions before 3.2.1 (possibly only versions earlier than 3.1.3). Based on discussion on python-checkins. ........ r46969 | gregory.p.smith | 2006-06-15 10:52:32 +0200 (Thu, 15 Jun 2006) | 6 lines - bsddb: multithreaded DB access using the simple bsddb module interface now works reliably. It has been updated to use automatic BerkeleyDB deadlock detection and the bsddb.dbutils.DeadlockWrap wrapper to retry database calls that would previously deadlock. [SF python bug #775414] ........ r46970 | gregory.p.smith | 2006-06-15 11:23:52 +0200 (Thu, 15 Jun 2006) | 2 lines minor documentation cleanup. mention the bsddb.db interface explicitly by name. ........ r46971 | neal.norwitz | 2006-06-15 11:57:03 +0200 (Thu, 15 Jun 2006) | 5 lines Steal the trick from test_compiler to print out a slow msg. This will hopefully get the buildbots to pass. Not sure this test will be feasible or even work. But everything is red now, so it can't get much worse. ........ r46972 | neal.norwitz | 2006-06-15 12:24:49 +0200 (Thu, 15 Jun 2006) | 1 line Print some more info to get an idea of how much longer the test will last ........ r46981 | tim.peters | 2006-06-15 20:04:40 +0200 (Thu, 15 Jun 2006) | 6 lines Try to reduce the extreme peak memory and disk-space use of this test. It probably still requires more disk space than most buildbots have, and in any case is still so intrusive that if we don't find another way to test this I'm taking my buildbot offline permanently ;-) ........ r46982 | tim.peters | 2006-06-15 20:06:29 +0200 (Thu, 15 Jun 2006) | 2 lines Whitespace normalization. ........ r46983 | tim.peters | 2006-06-15 20:07:28 +0200 (Thu, 15 Jun 2006) | 2 lines Add missing svn:eol-style property to text files. ........ r46984 | tim.peters | 2006-06-15 20:38:19 +0200 (Thu, 15 Jun 2006) | 2 lines Oops -- I introduced an off-by-6436159488 error. ........ r46990 | neal.norwitz | 2006-06-16 06:30:34 +0200 (Fri, 16 Jun 2006) | 1 line Disable this test until we can determine what to do about it ........ r46991 | neal.norwitz | 2006-06-16 06:31:06 +0200 (Fri, 16 Jun 2006) | 1 line Param name is dir, not directory. Update docstring. Backport candidate ........ r46992 | neal.norwitz | 2006-06-16 06:31:28 +0200 (Fri, 16 Jun 2006) | 1 line Add missing period in comment. ........ r46993 | neal.norwitz | 2006-06-16 06:32:43 +0200 (Fri, 16 Jun 2006) | 1 line Fix whitespace, there are memory leaks in this module. ........ r46995 | fred.drake | 2006-06-17 01:45:06 +0200 (Sat, 17 Jun 2006) | 3 lines SF patch 1504676: Make sgmllib char and entity references pluggable (implementation/tests contributed by Sam Ruby) ........ r46996 | fred.drake | 2006-06-17 03:07:54 +0200 (Sat, 17 Jun 2006) | 1 line fix change that broke the htmllib tests ........ r46998 | martin.v.loewis | 2006-06-17 11:15:14 +0200 (Sat, 17 Jun 2006) | 3 lines Patch #763580: Add name and value arguments to Tkinter variable classes. ........ r46999 | martin.v.loewis | 2006-06-17 11:20:41 +0200 (Sat, 17 Jun 2006) | 2 lines Patch #1096231: Add default argument to wm_iconbitmap. ........ r47000 | martin.v.loewis | 2006-06-17 11:25:15 +0200 (Sat, 17 Jun 2006) | 2 lines Patch #1494750: Destroy master after deleting children. ........ r47003 | george.yoshida | 2006-06-17 18:31:52 +0200 (Sat, 17 Jun 2006) | 2 lines markup fix ........ r47005 | george.yoshida | 2006-06-17 18:39:13 +0200 (Sat, 17 Jun 2006) | 4 lines Update url. Old url returned status code:301 Moved permanently. ........ r47007 | martin.v.loewis | 2006-06-17 20:44:27 +0200 (Sat, 17 Jun 2006) | 2 lines Patch #812986: Update the canvas even if not tracing. ........ r47008 | martin.v.loewis | 2006-06-17 21:03:26 +0200 (Sat, 17 Jun 2006) | 2 lines Patch #815924: Restore ability to pass type= and icon= ........ r47009 | neal.norwitz | 2006-06-18 00:37:45 +0200 (Sun, 18 Jun 2006) | 1 line Fix typo in docstring ........ r47010 | neal.norwitz | 2006-06-18 00:38:15 +0200 (Sun, 18 Jun 2006) | 1 line Fix memory leak reported by valgrind while running test_subprocess ........ r47011 | fred.drake | 2006-06-18 04:57:35 +0200 (Sun, 18 Jun 2006) | 1 line remove unnecessary markup ........ r47013 | neal.norwitz | 2006-06-18 21:35:01 +0200 (Sun, 18 Jun 2006) | 7 lines Prevent spurious leaks when running regrtest.py -R. There may be more issues that crop up from time to time, but this change seems to have been pretty stable (no spurious warnings) for about a week. Other modules which use threads may require similar use of threading_setup/threading_cleanup from test_support. ........ r47014 | neal.norwitz | 2006-06-18 21:37:40 +0200 (Sun, 18 Jun 2006) | 9 lines The hppa ubuntu box sometimes hangs forever in these tests. My guess is that the wait is failing for some reason. Use WNOHANG, so we won't wait until the buildbot kills the test suite. I haven't been able to reproduce the failure, so I'm not sure if this will help or not. Hopefully, this change will cause the test to fail, rather than hang. That will be better since we will get the rest of the test results. It may also help us debug the real problem. ........ r47015 | neal.norwitz | 2006-06-18 22:10:24 +0200 (Sun, 18 Jun 2006) | 1 line Revert 47014 until it is more robust ........ r47016 | thomas.heller | 2006-06-18 23:27:04 +0200 (Sun, 18 Jun 2006) | 6 lines Fix typos. Fix doctest example. Mention in the tutorial that 'errcheck' is explained in the ref manual. Use better wording in some places. Remoce code examples that shouldn't be in the tutorial. Remove some XXX notices. ........ r47017 | georg.brandl | 2006-06-19 00:17:29 +0200 (Mon, 19 Jun 2006) | 3 lines Patch #1507676: improve exception messages in abstract.c, object.c and typeobject.c. ........ r47018 | neal.norwitz | 2006-06-19 07:40:44 +0200 (Mon, 19 Jun 2006) | 1 line Use Py_ssize_t ........ r47019 | georg.brandl | 2006-06-19 08:35:54 +0200 (Mon, 19 Jun 2006) | 3 lines Add news entry about error msg improvement. ........ r47020 | thomas.heller | 2006-06-19 09:07:49 +0200 (Mon, 19 Jun 2006) | 2 lines Try to repair the failing test on the OpenBSD buildbot. Trial and error... ........ r47021 | tim.peters | 2006-06-19 09:45:16 +0200 (Mon, 19 Jun 2006) | 2 lines Whitespace normalization. ........ r47022 | walter.doerwald | 2006-06-19 10:07:50 +0200 (Mon, 19 Jun 2006) | 4 lines Patch #1506645: add Python wrappers for the curses functions is_term_resized, resize_term and resizeterm. This uses three separate configure checks (one for each function). ........ r47023 | walter.doerwald | 2006-06-19 10:14:09 +0200 (Mon, 19 Jun 2006) | 2 lines Make check order match in configure and configure.in. ........ r47024 | tim.peters | 2006-06-19 10:14:28 +0200 (Mon, 19 Jun 2006) | 3 lines Repair KeyError when running test_threaded_import under -R, as reported by Neal on python-dev. ........ r47025 | thomas.heller | 2006-06-19 10:32:46 +0200 (Mon, 19 Jun 2006) | 3 lines Next try to fix the OpenBSD buildbot tests: Use ctypes.util.find_library to locate the C runtime library on platforms where is returns useful results. ........ r47026 | tim.peters | 2006-06-19 11:09:44 +0200 (Mon, 19 Jun 2006) | 13 lines TestHelp.make_parser(): This was making a permanent change to os.environ (setting envar COLUMNS), which at least caused test_float_default() to fail if the tests were run more than once. This repairs the test_optparse -R failures Neal reported on python-dev. It also explains some seemingly bizarre test_optparse failures we saw a couple weeks ago on the buildbots, when test_optparse failed due to test_file failing to clean up after itself, and then test_optparse failed in an entirely different way when regrtest's -w option ran test_optparse a second time. It's now obvious that make_parser() permanently changing os.environ was responsible for the second half of that. ........ r47027 | anthony.baxter | 2006-06-19 14:04:15 +0200 (Mon, 19 Jun 2006) | 2 lines Preparing for 2.5b1. ........ r47029 | fred.drake | 2006-06-19 19:31:16 +0200 (Mon, 19 Jun 2006) | 1 line remove non-working document formats from edist ........ r47030 | gerhard.haering | 2006-06-19 23:17:35 +0200 (Mon, 19 Jun 2006) | 5 lines Fixed a memory leak that was introduced with incorrect usage of the Python weak reference API in pysqlite 2.2.1. Bumbed pysqlite version number to upcoming pysqlite 2.3.1 release. ........ r47032 | ka-ping.yee | 2006-06-20 00:49:36 +0200 (Tue, 20 Jun 2006) | 2 lines Remove Python 2.3 compatibility comment. ........ r47033 | trent.mick | 2006-06-20 01:21:25 +0200 (Tue, 20 Jun 2006) | 2 lines Upgrade pyexpat to expat 2.0.0 (http://python.org/sf/1462338). ........ r47034 | trent.mick | 2006-06-20 01:57:41 +0200 (Tue, 20 Jun 2006) | 3 lines [ 1295808 ] expat symbols should be namespaced in pyexpat (http://python.org/sf/1295808) ........ r47039 | andrew.kuchling | 2006-06-20 13:52:16 +0200 (Tue, 20 Jun 2006) | 1 line Uncomment wsgiref section ........ r47040 | andrew.kuchling | 2006-06-20 14:15:09 +0200 (Tue, 20 Jun 2006) | 1 line Add four library items ........ r47041 | andrew.kuchling | 2006-06-20 14:19:54 +0200 (Tue, 20 Jun 2006) | 1 line Terminology and typography fixes ........ r47042 | andrew.kuchling | 2006-06-20 15:05:12 +0200 (Tue, 20 Jun 2006) | 1 line Add introductory paragraphs summarizing the release; minor edits ........ r47043 | andrew.kuchling | 2006-06-20 15:11:29 +0200 (Tue, 20 Jun 2006) | 1 line Minor edits and rearrangements; markup fix ........ r47044 | andrew.kuchling | 2006-06-20 15:20:30 +0200 (Tue, 20 Jun 2006) | 1 line [Bug #1504456] Mention xml -> xmlcore change ........ r47047 | brett.cannon | 2006-06-20 19:30:26 +0200 (Tue, 20 Jun 2006) | 2 lines Raise TestSkipped when the test socket connection is refused. ........ r47049 | brett.cannon | 2006-06-20 21:20:17 +0200 (Tue, 20 Jun 2006) | 2 lines Fix typo of exception name. ........ r47053 | brett.cannon | 2006-06-21 18:57:57 +0200 (Wed, 21 Jun 2006) | 5 lines At the C level, tuple arguments are passed in directly to the exception constructor, meaning it is treated as *args, not as a single argument. This means using the 'message' attribute won't work (until Py3K comes around), and so one must grab from 'arg' to get the error number. ........ r47054 | andrew.kuchling | 2006-06-21 19:10:18 +0200 (Wed, 21 Jun 2006) | 1 line Link to LibRef module documentation ........ r47055 | andrew.kuchling | 2006-06-21 19:17:10 +0200 (Wed, 21 Jun 2006) | 1 line Note some of Barry's work ........ r47056 | andrew.kuchling | 2006-06-21 19:17:28 +0200 (Wed, 21 Jun 2006) | 1 line Bump version ........ r47057 | georg.brandl | 2006-06-21 19:45:17 +0200 (Wed, 21 Jun 2006) | 3 lines fix [ 1509132 ] compiler module builds incorrect AST for TryExceptFinally ........ r47058 | georg.brandl | 2006-06-21 19:52:36 +0200 (Wed, 21 Jun 2006) | 3 lines Make test_fcntl aware of netbsd3. ........ r47059 | georg.brandl | 2006-06-21 19:53:17 +0200 (Wed, 21 Jun 2006) | 3 lines Patch #1509001: expected skips for netbsd3. ........ r47060 | gerhard.haering | 2006-06-21 22:55:04 +0200 (Wed, 21 Jun 2006) | 2 lines Removed call to enable_callback_tracebacks that slipped in by accident. ........ r47061 | armin.rigo | 2006-06-21 23:58:50 +0200 (Wed, 21 Jun 2006) | 13 lines Fix for an obscure bug introduced by revs 46806 and 46808, with a test. The problem of checking too eagerly for recursive calls is the following: if a RuntimeError is caused by recursion, and if code needs to normalize it immediately (as in the 2nd test), then PyErr_NormalizeException() needs a call to the RuntimeError class to instantiate it, and this hits the recursion limit again... causing PyErr_NormalizeException() to never finish. Moved this particular recursion check to slot_tp_call(), which is not involved in instantiating built-in exceptions. Backport candidate. ........ r47064 | neal.norwitz | 2006-06-22 08:30:50 +0200 (Thu, 22 Jun 2006) | 3 lines Copy the wsgiref package during make install. ........ r47065 | neal.norwitz | 2006-06-22 08:35:30 +0200 (Thu, 22 Jun 2006) | 1 line Reset the doc date to today for the automatic doc builds ........ r47067 | andrew.kuchling | 2006-06-22 15:10:23 +0200 (Thu, 22 Jun 2006) | 1 line Mention how to suppress warnings ........ r47069 | georg.brandl | 2006-06-22 16:46:17 +0200 (Thu, 22 Jun 2006) | 3 lines Set lineno correctly on list, tuple and dict literals. ........ r47070 | georg.brandl | 2006-06-22 16:46:46 +0200 (Thu, 22 Jun 2006) | 4 lines Test for correct compilation of try-except-finally stmt. Test for correct lineno on list, tuple, dict literals. ........ r47071 | fred.drake | 2006-06-22 17:50:08 +0200 (Thu, 22 Jun 2006) | 1 line fix markup nit ........ r47072 | brett.cannon | 2006-06-22 18:49:14 +0200 (Thu, 22 Jun 2006) | 6 lines 'warning's was improperly requiring that a command-line Warning category be both a subclass of Warning and a subclass of types.ClassType. The latter is no longer true thanks to new-style exceptions. Closes bug #1510580. Thanks to AMK for the test. ........ r47073 | ronald.oussoren | 2006-06-22 20:33:54 +0200 (Thu, 22 Jun 2006) | 3 lines MacOSX: Add a message to the first screen of the installer that tells users how to avoid updates to their shell profile. ........ r47074 | georg.brandl | 2006-06-22 21:02:18 +0200 (Thu, 22 Jun 2006) | 3 lines Fix my name ;) ........ r47075 | thomas.heller | 2006-06-22 21:07:36 +0200 (Thu, 22 Jun 2006) | 2 lines Small fixes, mostly in the markup. ........ r47076 | peter.astrand | 2006-06-22 22:06:46 +0200 (Thu, 22 Jun 2006) | 1 line Make it possible to run test_subprocess.py on Python 2.2, which lacks test_support.is_resource_enabled. ........ r47077 | peter.astrand | 2006-06-22 22:21:26 +0200 (Thu, 22 Jun 2006) | 1 line Applied patch #1506758: Prevent MemoryErrors with large MAXFD. ........ r47079 | neal.norwitz | 2006-06-23 05:32:44 +0200 (Fri, 23 Jun 2006) | 1 line Fix refleak ........ r47080 | fred.drake | 2006-06-23 08:03:45 +0200 (Fri, 23 Jun 2006) | 9 lines - SF bug #853506: IP6 address parsing in sgmllib ('[' and ']' were not accepted in unquoted attribute values) - cleaned up tests of character and entity reference decoding so the tests cover the documented relationships among handle_charref, handle_entityref, convert_charref, convert_codepoint, and convert_entityref, without bringing up Unicode issues that sgmllib cannot be involved in ........ r47085 | andrew.kuchling | 2006-06-23 21:23:40 +0200 (Fri, 23 Jun 2006) | 11 lines Fit Makefile for the Python doc environment better; this is a step toward including the howtos in the build process. * Put LaTeX output in ../paper-<whatever>/. * Put HTML output in ../html/ * Explain some of the Makefile variables * Remove some cruft dating to my environment (e.g. the 'web' target) This makefile isn't currently invoked by the documentation build process, so these changes won't destabilize anything. ........ r47086 | hyeshik.chang | 2006-06-23 23:16:18 +0200 (Fri, 23 Jun 2006) | 5 lines Bug #1511381: codec_getstreamcodec() in codec.c is corrected to omit a default "error" argument for NULL pointer. This allows the parser to take a codec from cjkcodecs again. (Reported by Taewook Kang and reviewed by Walter Doerwald) ........ r47091 | ronald.oussoren | 2006-06-25 22:44:16 +0200 (Sun, 25 Jun 2006) | 6 lines Workaround for bug #1512124 Without this patch IDLE will get unresponsive when you open the debugger window on OSX. This is both using the system Tcl/Tk on Tiger as the latest universal download from tk-components.sf.net. ........ r47092 | ronald.oussoren | 2006-06-25 23:14:19 +0200 (Sun, 25 Jun 2006) | 3 lines Drop the calldll demo's for macos, calldll isn't present anymore, no need to keep the demo's around. ........ r47093 | ronald.oussoren | 2006-06-25 23:15:58 +0200 (Sun, 25 Jun 2006) | 3 lines Use a path without a double slash to compile the .py files after installation (macosx, binary installer). This fixes bug #1508369 for python 2.5. ........ r47094 | ronald.oussoren | 2006-06-25 23:19:06 +0200 (Sun, 25 Jun 2006) | 3 lines Also install the .egg-info files in Lib. This will cause wsgiref.egg-info to be installed. ........ r47097 | andrew.kuchling | 2006-06-26 14:40:02 +0200 (Mon, 26 Jun 2006) | 1 line [Bug #1511998] Various comments from Nick Coghlan; thanks! ........ r47098 | andrew.kuchling | 2006-06-26 14:43:43 +0200 (Mon, 26 Jun 2006) | 1 line Describe workaround for PyRange_New()'s removal ........ r47099 | andrew.kuchling | 2006-06-26 15:08:24 +0200 (Mon, 26 Jun 2006) | 5 lines [Bug #1512163] Fix typo. This change will probably break tests on FreeBSD buildbots, but I'll check in a fix for that next. ........ r47100 | andrew.kuchling | 2006-06-26 15:12:16 +0200 (Mon, 26 Jun 2006) | 9 lines [Bug #1512163] Use one set of locking methods, lockf(); remove the flock() calls. On FreeBSD, the two methods lockf() and flock() end up using the same mechanism and the second one fails. A Linux man page claims that the two methods are orthogonal (so locks acquired one way don't interact with locks acquired the other way) but that clearly must be false. ........ r47101 | andrew.kuchling | 2006-06-26 15:23:10 +0200 (Mon, 26 Jun 2006) | 5 lines Add a test for a conflicting lock. On slow machines, maybe the time intervals (2 sec, 0.5 sec) will be too tight. I'll see how the buildbots like it. ........ r47103 | andrew.kuchling | 2006-06-26 16:33:24 +0200 (Mon, 26 Jun 2006) | 1 line Windows doesn't have os.fork(). I'll just disable this test for now ........ r47106 | andrew.kuchling | 2006-06-26 19:00:35 +0200 (Mon, 26 Jun 2006) | 9 lines Attempt to fix build failure on OS X and Debian alpha; the symptom is consistent with os.wait() returning immediately because some other subprocess had previously exited; the test suite then immediately tries to lock the mailbox and gets an error saying it's already locked. To fix this, do a waitpid() so the test suite only continues once the intended child process has exited. ........ r47113 | neal.norwitz | 2006-06-27 06:06:46 +0200 (Tue, 27 Jun 2006) | 1 line Ignore some more warnings in the dynamic linker on an older gentoo ........ r47114 | neal.norwitz | 2006-06-27 06:09:13 +0200 (Tue, 27 Jun 2006) | 6 lines Instead of doing a make test, run the regression tests out of the installed copy. This will hopefully catch problems where directories are added under Lib/ but not to Makefile.pre.in. This breaks out the 2 runs of the test suite with and without -O which is also nicer. ........ r47115 | neal.norwitz | 2006-06-27 06:12:58 +0200 (Tue, 27 Jun 2006) | 5 lines Fix SF bug #1513032, 'make install' failure on FreeBSD 5.3. No need to install lib-old, it's empty in 2.5. ........ r47116 | neal.norwitz | 2006-06-27 06:23:06 +0200 (Tue, 27 Jun 2006) | 1 line Test unimportant change to verify buildbot does not try to build ........ r47117 | neal.norwitz | 2006-06-27 06:26:30 +0200 (Tue, 27 Jun 2006) | 1 line Try again: test unimportant change to verify buildbot does not try to build ........ r47118 | neal.norwitz | 2006-06-27 06:28:56 +0200 (Tue, 27 Jun 2006) | 1 line Verify buildbot picks up these changes (really needs testing after last change to Makefile.pre.in) ........ r47121 | vinay.sajip | 2006-06-27 09:34:37 +0200 (Tue, 27 Jun 2006) | 1 line Removed buggy exception handling in doRollover of rotating file handlers. Exceptions now propagate to caller. ........ r47123 | ronald.oussoren | 2006-06-27 12:08:25 +0200 (Tue, 27 Jun 2006) | 3 lines MacOSX: fix rather dumb buglet that made it impossible to create extensions on OSX 10.3 when using a binary distribution build on 10.4. ........ r47125 | tim.peters | 2006-06-27 13:52:49 +0200 (Tue, 27 Jun 2006) | 2 lines Whitespace normalization. ........ r47128 | ronald.oussoren | 2006-06-27 14:53:52 +0200 (Tue, 27 Jun 2006) | 8 lines Use staticly build copies of zlib and bzip2 to build the OSX installer, that way the resulting binaries have a better change of running on 10.3. This patch also updates the search logic for sleepycat db3/4, without this patch you cannot use a sleepycat build with a non-standard prefix; with this you can (at least on OSX) if you add the prefix to CPPFLAGS/LDFLAGS at configure-time. This change is needed to build the binary installer for OSX. ........ r47131 | ronald.oussoren | 2006-06-27 17:45:32 +0200 (Tue, 27 Jun 2006) | 5 lines macosx: Install a libpython2.5.a inside the framework as a symlink to the actual dylib at the root of the framework, that way tools that expect a unix-like install (python-config, but more importantly external products like mod_python) work correctly. ........ r47137 | neal.norwitz | 2006-06-28 07:03:22 +0200 (Wed, 28 Jun 2006) | 4 lines According to the man pages on Gentoo Linux and Tru64, EACCES or EAGAIN can be returned if fcntl (lockf) fails. This fixes the test failure on Tru64 by checking for either error rather than just EAGAIN. ........ r47139 | neal.norwitz | 2006-06-28 08:28:31 +0200 (Wed, 28 Jun 2006) | 5 lines Fix bug #1512695: cPickle.loads could crash if it was interrupted with a KeyboardInterrupt since PyTuple_Pack was passed a NULL. Will backport. ........ r47142 | nick.coghlan | 2006-06-28 12:41:47 +0200 (Wed, 28 Jun 2006) | 1 line Make full module name available as __module_name__ even when __name__ is set to something else (like '__main__') ........ r47143 | armin.rigo | 2006-06-28 12:49:51 +0200 (Wed, 28 Jun 2006) | 2 lines A couple of crashers of the "won't fix" kind. ........ r47147 | andrew.kuchling | 2006-06-28 16:25:20 +0200 (Wed, 28 Jun 2006) | 1 line [Bug #1508766] Add docs for uuid module; docs written by George Yoshida, with minor rearrangements by me. ........ r47148 | andrew.kuchling | 2006-06-28 16:27:21 +0200 (Wed, 28 Jun 2006) | 1 line [Bug #1508766] Add docs for uuid module; this puts the module in the 'Internet Protocols' section. Arguably this module could also have gone in the chapters on strings or encodings, maybe even the crypto chapter. Fred, please move if you see fit. ........ r47151 | georg.brandl | 2006-06-28 22:23:25 +0200 (Wed, 28 Jun 2006) | 3 lines Fix end_fill(). ........ r47153 | trent.mick | 2006-06-28 22:30:41 +0200 (Wed, 28 Jun 2006) | 2 lines Mention the expat upgrade and pyexpat fix I put in 2.5b1. ........ r47154 | fred.drake | 2006-06-29 02:51:53 +0200 (Thu, 29 Jun 2006) | 6 lines SF bug #1504333: sgmlib should allow angle brackets in quoted values (modified patch by Sam Ruby; changed to use separate REs for start and end tags to reduce matching cost for end tags; extended tests; updated to avoid breaking previous changes to support IPv6 addresses in unquoted attribute values) ........ r47156 | fred.drake | 2006-06-29 04:57:48 +0200 (Thu, 29 Jun 2006) | 1 line document recent bugfixes in sgmllib ........ r47158 | neal.norwitz | 2006-06-29 06:10:08 +0200 (Thu, 29 Jun 2006) | 10 lines Add new utility function, reap_children(), to test_support. This should be called at the end of each test that spawns children (perhaps it should be called from regrtest instead?). This will hopefully prevent some of the unexplained failures in the buildbots (hppa and alpha) during tests that spawn children. The problems were not reproducible. There were many zombies that remained at the end of several tests. In the worst case, this shouldn't cause any more problems, though it may not help either. Time will tell. ........ r47159 | neal.norwitz | 2006-06-29 07:48:14 +0200 (Thu, 29 Jun 2006) | 5 lines This should fix the buildbot failure on s/390 which can't connect to gmail.org. It makes the error message consistent and always sends to stderr. It would be much better for all the networking tests to hit only python.org. ........ r47161 | thomas.heller | 2006-06-29 20:34:15 +0200 (Thu, 29 Jun 2006) | 3 lines Protect the thread api calls in the _ctypes extension module within #ifdef WITH_THREADS/#endif blocks. Found by Sam Rushing. ........ r47162 | martin.v.loewis | 2006-06-29 20:58:44 +0200 (Thu, 29 Jun 2006) | 2 lines Patch #1509163: MS Toolkit Compiler no longer available ........ r47163 | skip.montanaro | 2006-06-29 21:20:09 +0200 (Thu, 29 Jun 2006) | 1 line add string methods to index ........ r47164 | vinay.sajip | 2006-06-30 02:13:08 +0200 (Fri, 30 Jun 2006) | 1 line Fixed bug in fileConfig() which failed to clear logging._handlerList ........ r47166 | tim.peters | 2006-06-30 08:18:39 +0200 (Fri, 30 Jun 2006) | 2 lines Whitespace normalization. ........ r47170 | neal.norwitz | 2006-06-30 09:32:16 +0200 (Fri, 30 Jun 2006) | 1 line Silence compiler warning ........ r47171 | neal.norwitz | 2006-06-30 09:32:46 +0200 (Fri, 30 Jun 2006) | 1 line Another problem reported by Coverity. Backport candidate. ........ r47175 | thomas.heller | 2006-06-30 19:44:54 +0200 (Fri, 30 Jun 2006) | 2 lines Revert the use of PY_FORMAT_SIZE_T in PyErr_Format. ........ r47176 | tim.peters | 2006-06-30 20:34:51 +0200 (Fri, 30 Jun 2006) | 2 lines Remove now-unused fidding with PY_FORMAT_SIZE_T. ........ r47177 | georg.brandl | 2006-06-30 20:47:56 +0200 (Fri, 30 Jun 2006) | 3 lines Document decorator usage of property. ........ r47181 | fred.drake | 2006-06-30 21:29:25 +0200 (Fri, 30 Jun 2006) | 4 lines - consistency nit: always include "()" in \function and \method (*should* be done by the presentation, but that requires changes all over) - avoid spreading the __name meme ........ r47188 | vinay.sajip | 2006-07-01 12:45:20 +0200 (Sat, 01 Jul 2006) | 1 line Added entry for fileConfig() bugfix. ........ r47189 | vinay.sajip | 2006-07-01 12:47:20 +0200 (Sat, 01 Jul 2006) | 1 line Added duplicate call to fileConfig() to ensure that it cleans up after itself correctly. ........ r47190 | martin.v.loewis | 2006-07-01 17:33:37 +0200 (Sat, 01 Jul 2006) | 2 lines Release all forwarded functions in .close. Fixes #1513223. ........ r47191 | fred.drake | 2006-07-01 18:28:20 +0200 (Sat, 01 Jul 2006) | 7 lines SF bug #1296433 (Expat bug #1515266): Unchecked calls to character data handler would cause a segfault. This merges in Expat's lib/xmlparse.c revisions 1.154 and 1.155, which fix this and a closely related problem (the later does not affect Python). Moved the crasher test to the tests for xml.parsers.expat. ........ r47197 | gerhard.haering | 2006-07-02 19:48:30 +0200 (Sun, 02 Jul 2006) | 4 lines The sqlite3 module did cut off data from the SQLite database at the first null character before sending it to a custom converter. This has been fixed now. ........ r47198 | martin.v.loewis | 2006-07-02 20:44:00 +0200 (Sun, 02 Jul 2006) | 1 line Correct arithmetic in access on Win32. Fixes #1513646. ........ r47203 | thomas.heller | 2006-07-03 09:58:09 +0200 (Mon, 03 Jul 2006) | 1 line Cleanup: Remove commented out code. ........ r47204 | thomas.heller | 2006-07-03 09:59:50 +0200 (Mon, 03 Jul 2006) | 1 line Don't run the doctests with Python 2.3 because it doesn't have the ELLIPSIS flag. ........ r47205 | thomas.heller | 2006-07-03 10:04:05 +0200 (Mon, 03 Jul 2006) | 7 lines Fixes so that _ctypes can be compiled with the MingW compiler. It seems that the definition of '__attribute__(x)' was responsible for the compiler ignoring the '__fastcall' attribute on the ffi_closure_SYSV function in libffi_msvc/ffi.c, took me quite some time to figure this out. ........ r47206 | thomas.heller | 2006-07-03 10:08:14 +0200 (Mon, 03 Jul 2006) | 11 lines Add a new function uses_seh() to the _ctypes extension module. This will return True if Windows Structured Exception handling (SEH) is used when calling functions, False otherwise. Currently, only MSVC supports SEH. Fix the test so that it doesn't crash when run with MingW compiled _ctypes. Note that two tests are still failing when mingw is used, I suspect structure layout differences and function calling conventions between MSVC and MingW. ........ r47207 | tim.peters | 2006-07-03 10:23:19 +0200 (Mon, 03 Jul 2006) | 2 lines Whitespace normalization. ........ r47208 | martin.v.loewis | 2006-07-03 11:44:00 +0200 (Mon, 03 Jul 2006) | 3 lines Only setup canvas when it is first created. Fixes #1514703 ........ r47209 | martin.v.loewis | 2006-07-03 12:05:30 +0200 (Mon, 03 Jul 2006) | 3 lines Reimplement turtle.circle using a polyline, to allow correct filling of arcs. Also fixes #1514693. ........ r47210 | martin.v.loewis | 2006-07-03 12:19:49 +0200 (Mon, 03 Jul 2006) | 3 lines Bug #1514693: Update turtle's heading when switching between degrees and radians. ........ r47211 | martin.v.loewis | 2006-07-03 13:12:06 +0200 (Mon, 03 Jul 2006) | 2 lines Document functions added in 2.3 and 2.5. ........ r47212 | martin.v.loewis | 2006-07-03 14:19:50 +0200 (Mon, 03 Jul 2006) | 3 lines Bug #1417699: Reject locale-specific decimal point in float() and atof(). ........ r47213 | martin.v.loewis | 2006-07-03 14:28:58 +0200 (Mon, 03 Jul 2006) | 3 lines Bug #1267547: Put proper recursive setup.py call into the spec file generated by bdist_rpm. ........ r47215 | martin.v.loewis | 2006-07-03 15:01:35 +0200 (Mon, 03 Jul 2006) | 3 lines Patch #825417: Fix timeout processing in expect, read_until. Will backport to 2.4. ........ r47218 | martin.v.loewis | 2006-07-03 15:47:40 +0200 (Mon, 03 Jul 2006) | 2 lines Put method-wrappers into trashcan. Fixes #927248. ........ r47219 | andrew.kuchling | 2006-07-03 16:07:30 +0200 (Mon, 03 Jul 2006) | 1 line [Bug #1515932] Clarify description of slice assignment ........ r47220 | andrew.kuchling | 2006-07-03 16:16:09 +0200 (Mon, 03 Jul 2006) | 4 lines [Bug #1511911] Clarify description of optional arguments to sorted() by improving the xref to the section on lists, and by copying the explanations of the arguments (with a slight modification). ........ r47223 | kristjan.jonsson | 2006-07-03 16:59:05 +0200 (Mon, 03 Jul 2006) | 1 line Fix build problems with the platform SDK on windows. It is not sufficient to test for the C compiler version when determining if we have the secure CRT from microsoft. Must test with an undocumented macro, __STDC_SECURE_LIB__ too. ........ r47224 | ronald.oussoren | 2006-07-04 14:30:22 +0200 (Tue, 04 Jul 2006) | 7 lines Sync the darwin/x86 port libffi with the copy in PyObjC. This fixes a number of bugs in that port. The most annoying ones were due to some subtle differences between the document ABI and the actual implementation :-( (there are no python unittests that fail without this patch, but without it some of libffi's unittests fail). ........ r47234 | georg.brandl | 2006-07-05 10:21:00 +0200 (Wed, 05 Jul 2006) | 3 lines Remove remaining references to OverflowWarning. ........ r47236 | thomas.heller | 2006-07-05 11:13:56 +0200 (Wed, 05 Jul 2006) | 3 lines Fix the bitfield test when _ctypes is compiled with MingW. Structures containing bitfields may have different layout on MSVC and MingW . ........ r47237 | thomas.wouters | 2006-07-05 13:03:49 +0200 (Wed, 05 Jul 2006) | 15 lines Fix bug in passing tuples to string.Template. All other values (with working str() or repr()) would work, just not multi-value tuples. Probably not a backport candidate, since it changes the behaviour of passing a single-element tuple: >>> string.Template("$foo").substitute(dict(foo=(1,))) '(1,)' versus '1' ........ r47241 | georg.brandl | 2006-07-05 16:18:45 +0200 (Wed, 05 Jul 2006) | 2 lines Patch #1517490: fix glitches in filter() docs. ........ r47244 | georg.brandl | 2006-07-05 17:50:05 +0200 (Wed, 05 Jul 2006) | 2 lines no need to elaborate "string". ........ r47251 | neal.norwitz | 2006-07-06 06:28:59 +0200 (Thu, 06 Jul 2006) | 3 lines Fix refleaks reported by Shane Hathaway in SF patch #1515361. This change contains only the changes related to leaking the copy variable. ........ r47253 | fred.drake | 2006-07-06 07:13:22 +0200 (Thu, 06 Jul 2006) | 4 lines - back out Expat change; the final fix to Expat will be different - change the pyexpat wrapper to not be so sensitive to this detail of the Expat implementation (the ex-crasher test still passes) ........ r47257 | neal.norwitz | 2006-07-06 08:45:08 +0200 (Thu, 06 Jul 2006) | 1 line Add a NEWS entry for a recent pyexpat fix ........ r47258 | martin.v.loewis | 2006-07-06 08:55:58 +0200 (Thu, 06 Jul 2006) | 2 lines Add sqlite3.dll to the DLLs component, not to the TkDLLs component. Fixes #1517388. ........ r47259 | martin.v.loewis | 2006-07-06 09:05:21 +0200 (Thu, 06 Jul 2006) | 1 line Properly quote compileall and Lib paths in case TARGETDIR has a space. ........ r47260 | thomas.heller | 2006-07-06 09:50:18 +0200 (Thu, 06 Jul 2006) | 5 lines Revert the change done in svn revision 47206: Add a new function uses_seh() to the _ctypes extension module. This will return True if Windows Structured Exception handling (SEH) is used when calling functions, False otherwise. ........ r47261 | armin.rigo | 2006-07-06 09:58:18 +0200 (Thu, 06 Jul 2006) | 3 lines A couple of examples about how to attack the fact that _PyType_Lookup() returns a borrowed ref. Many of the calls are open to attack. ........ r47262 | thomas.heller | 2006-07-06 10:28:14 +0200 (Thu, 06 Jul 2006) | 2 lines The test that calls a function with invalid arguments and catches the resulting Windows access violation will not be run by default. ........ r47263 | thomas.heller | 2006-07-06 10:48:35 +0200 (Thu, 06 Jul 2006) | 5 lines Patch #1517790: It is now possible to use custom objects in the ctypes foreign function argtypes sequence as long as they provide a from_param method, no longer is it required that the object is a ctypes type. ........ r47264 | thomas.heller | 2006-07-06 10:58:40 +0200 (Thu, 06 Jul 2006) | 2 lines Document the Struture and Union constructors. ........ r47265 | thomas.heller | 2006-07-06 11:11:22 +0200 (Thu, 06 Jul 2006) | 2 lines Document the changes in svn revision 47263, from patch #1517790. ........ r47267 | ronald.oussoren | 2006-07-06 12:13:35 +0200 (Thu, 06 Jul 2006) | 7 lines This patch solves the problem Skip was seeing with zlib, this patch ensures that configure uses similar compiler flags as setup.py when doing the zlib test. Without this patch configure would use the first shared library on the linker path, with this patch it uses the first shared or static library on that path just like setup.py. ........ r47268 | thomas.wouters | 2006-07-06 12:48:28 +0200 (Thu, 06 Jul 2006) | 4 lines NEWS entry for r47267: fixing configure's zlib probing. ........ r47269 | fredrik.lundh | 2006-07-06 14:29:24 +0200 (Thu, 06 Jul 2006) | 3 lines added XMLParser alias for cElementTree compatibility ........ r47271 | nick.coghlan | 2006-07-06 14:53:04 +0200 (Thu, 06 Jul 2006) | 1 line Revert the __module_name__ changes made in rev 47142. We'll revisit this in Python 2.6 ........ r47272 | nick.coghlan | 2006-07-06 15:04:56 +0200 (Thu, 06 Jul 2006) | 1 line Update the tutorial section on relative imports ........ r47273 | nick.coghlan | 2006-07-06 15:35:27 +0200 (Thu, 06 Jul 2006) | 1 line Ignore ImportWarning by default ........ r47274 | nick.coghlan | 2006-07-06 15:41:34 +0200 (Thu, 06 Jul 2006) | 1 line Cover ImportWarning, PendingDeprecationWarning and simplefilter() in the warnings module docs ........ r47275 | nick.coghlan | 2006-07-06 15:47:18 +0200 (Thu, 06 Jul 2006) | 1 line Add NEWS entries for the ImportWarning change and documentation update ........ r47276 | andrew.kuchling | 2006-07-06 15:57:28 +0200 (Thu, 06 Jul 2006) | 1 line ImportWarning is now silent by default ........ r47277 | thomas.heller | 2006-07-06 17:06:05 +0200 (Thu, 06 Jul 2006) | 2 lines Document the correct return type of PyLong_AsUnsignedLongLongMask. ........ r47278 | hyeshik.chang | 2006-07-06 17:21:52 +0200 (Thu, 06 Jul 2006) | 2 lines Add a testcase for r47086 which fixed a bug in codec_getstreamcodec(). ........ r47279 | hyeshik.chang | 2006-07-06 17:39:24 +0200 (Thu, 06 Jul 2006) | 3 lines Test using all CJK encodings for the testcases which don't require specific encodings. ........ r47280 | martin.v.loewis | 2006-07-06 21:28:03 +0200 (Thu, 06 Jul 2006) | 2 lines Properly generate logical file ids. Fixes #1515998. Also correct typo in Control.mapping. ........ r47287 | neal.norwitz | 2006-07-07 08:03:15 +0200 (Fri, 07 Jul 2006) | 17 lines Restore rev 47014: The hppa ubuntu box sometimes hangs forever in these tests. My guess is that the wait is failing for some reason. Use WNOHANG, so we won't wait until the buildbot kills the test suite. I haven't been able to reproduce the failure, so I'm not sure if this will help or not. Hopefully, this change will cause the test to fail, rather than hang. That will be better since we will get the rest of the test results. It may also help us debug the real problem. *** The reason this originally failed was because there were many zombie children outstanding before rev 47158 cleaned them up. There are still hangs in test_subprocess that need to be addressed, but that will take more work. This should close some holes. ........ r47289 | georg.brandl | 2006-07-07 10:15:12 +0200 (Fri, 07 Jul 2006) | 3 lines Fix RFC number. ........ r50489 | neal.norwitz | 2006-07-08 07:31:37 +0200 (Sat, 08 Jul 2006) | 1 line Fix SF bug #1519018: 'as' is now validated properly in import statements ........ r50490 | georg.brandl | 2006-07-08 14:15:27 +0200 (Sat, 08 Jul 2006) | 3 lines Add an additional test for bug #1519018. ........ r50491 | tim.peters | 2006-07-08 21:55:05 +0200 (Sat, 08 Jul 2006) | 2 lines Whitespace normalization. ........ r50493 | neil.schemenauer | 2006-07-09 18:16:34 +0200 (Sun, 09 Jul 2006) | 2 lines Fix AST compiler bug #1501934: incorrect LOAD/STORE_GLOBAL generation. ........ r50495 | neil.schemenauer | 2006-07-09 23:19:29 +0200 (Sun, 09 Jul 2006) | 2 lines Fix SF bug 1441486: bad unary minus folding in compiler. ........ r50497 | neal.norwitz | 2006-07-10 00:14:42 +0200 (Mon, 10 Jul 2006) | 4 lines On 64 bit systems, int literals that use less than 64 bits are now ints rather than longs. This also fixes the test for eval(-sys.maxint - 1). ........ r50500 | neal.norwitz | 2006-07-10 02:04:44 +0200 (Mon, 10 Jul 2006) | 4 lines Bug #1512814, Fix incorrect lineno's when code at module scope started after line 256. ........ r50501 | neal.norwitz | 2006-07-10 02:05:34 +0200 (Mon, 10 Jul 2006) | 1 line Fix doco. Backport candidate. ........ r50503 | neal.norwitz | 2006-07-10 02:23:17 +0200 (Mon, 10 Jul 2006) | 5 lines Part of SF patch #1484695. This removes dead code. The chksum was already verified in .frombuf() on the lines above. If there was a problem an exception is raised, so there was no way this condition could have been true. ........ r50504 | neal.norwitz | 2006-07-10 03:18:57 +0200 (Mon, 10 Jul 2006) | 3 lines Patch #1516912: improve Modules support for OpenVMS. ........ r50506 | neal.norwitz | 2006-07-10 04:36:41 +0200 (Mon, 10 Jul 2006) | 7 lines Patch #1504046: Add documentation for xml.etree. /F wrote the text docs, Englebert Gruber massaged it to latex and I did some more massaging to try and improve the consistency and fix some name mismatches between the declaration and text. ........ r50509 | martin.v.loewis | 2006-07-10 09:23:48 +0200 (Mon, 10 Jul 2006) | 2 lines Introduce DISTUTILS_USE_SDK as a flag to determine whether the SDK environment should be used. Fixes #1508010. ........ r50510 | martin.v.loewis | 2006-07-10 09:26:41 +0200 (Mon, 10 Jul 2006) | 1 line Change error message to indicate that VS2003 is necessary to build extension modules, not the .NET SDK. ........ r50511 | martin.v.loewis | 2006-07-10 09:29:41 +0200 (Mon, 10 Jul 2006) | 1 line Add svn:ignore. ........ r50512 | anthony.baxter | 2006-07-10 09:41:04 +0200 (Mon, 10 Jul 2006) | 1 line preparing for 2.5b2 ........ r50513 | thomas.heller | 2006-07-10 11:10:28 +0200 (Mon, 10 Jul 2006) | 2 lines Fix bug #1518190: accept any integer or long value in the ctypes.c_void_p constructor. ........ r50514 | thomas.heller | 2006-07-10 11:31:06 +0200 (Mon, 10 Jul 2006) | 3 lines Fixed a segfault when ctypes.wintypes were imported on non-Windows machines. ........ r50516 | thomas.heller | 2006-07-10 13:11:10 +0200 (Mon, 10 Jul 2006) | 3 lines Assigning None to pointer type structure fields possible overwrote wrong fields. ........ r50517 | thomas.heller | 2006-07-10 13:17:37 +0200 (Mon, 10 Jul 2006) | 5 lines Moved the ctypes news entries from the 'Library' section into the 'Extension Modules' section where they belong, probably. This destroyes the original order of the news entries, don't know if that is important or not. ........ r50526 | phillip.eby | 2006-07-10 21:03:29 +0200 (Mon, 10 Jul 2006) | 2 lines Fix SF#1516184 and add a test to prevent regression. ........ r50528 | phillip.eby | 2006-07-10 21:18:35 +0200 (Mon, 10 Jul 2006) | 2 lines Fix SF#1457312: bad socket error handling in distutils "upload" command. ........ r50537 | peter.astrand | 2006-07-10 22:39:49 +0200 (Mon, 10 Jul 2006) | 1 line Make it possible to run test_subprocess.py with Python 2.2, which lacks test_support.reap_children(). ........ r50541 | tim.peters | 2006-07-10 23:08:24 +0200 (Mon, 10 Jul 2006) | 5 lines After approval from Anthony, merge the tim-current_frames branch into the trunk. This adds a new sys._current_frames() function, which returns a dict mapping thread id to topmost thread stack frame. ........ r50542 | tim.peters | 2006-07-10 23:11:49 +0200 (Mon, 10 Jul 2006) | 2 lines Whitespace normalization. ........ r50553 | martin.v.loewis | 2006-07-11 00:11:28 +0200 (Tue, 11 Jul 2006) | 4 lines Patch #1519566: Remove unused _tofill member. Make begin_fill idempotent. Update demo2 to demonstrate filling of concave shapes. ........ r50567 | anthony.baxter | 2006-07-11 04:04:09 +0200 (Tue, 11 Jul 2006) | 4 lines #1494314: Fix a regression with high-numbered sockets in 2.4.3. This means that select() on sockets > FD_SETSIZE (typically 1024) work again. The patch makes sockets use poll() internally where available. ........ r50568 | tim.peters | 2006-07-11 04:17:48 +0200 (Tue, 11 Jul 2006) | 2 lines Whitespace normalization. ........ r50575 | thomas.heller | 2006-07-11 18:42:05 +0200 (Tue, 11 Jul 2006) | 1 line Add missing Py_DECREF. ........ r50576 | thomas.heller | 2006-07-11 18:44:25 +0200 (Tue, 11 Jul 2006) | 1 line Add missing Py_DECREFs. ........ r50579 | andrew.kuchling | 2006-07-11 19:20:16 +0200 (Tue, 11 Jul 2006) | 1 line Bump version number; add sys._current_frames ........ r50582 | thomas.heller | 2006-07-11 20:28:35 +0200 (Tue, 11 Jul 2006) | 3 lines When a foreign function is retrived by calling __getitem__ on a ctypes library instance, do not set it as attribute. ........ r50583 | thomas.heller | 2006-07-11 20:40:50 +0200 (Tue, 11 Jul 2006) | 2 lines Change the ctypes version number to 1.0.0. ........ r50597 | neal.norwitz | 2006-07-12 07:26:17 +0200 (Wed, 12 Jul 2006) | 3 lines Bug #1520864: unpacking singleton tuples in for loop (for x, in) work again. ........ r50598 | neal.norwitz | 2006-07-12 07:26:35 +0200 (Wed, 12 Jul 2006) | 1 line Fix function name in error msg ........ r50599 | neal.norwitz | 2006-07-12 07:27:46 +0200 (Wed, 12 Jul 2006) | 4 lines Fix uninitialized memory read reported by Valgrind when running doctest. This could happen if size == 0. ........ r50600 | neal.norwitz | 2006-07-12 09:28:29 +0200 (Wed, 12 Jul 2006) | 1 line Actually change the MAGIC #. Create a new section for 2.5c1 and mention the impact of changing the MAGIC #. ........ r50601 | thomas.heller | 2006-07-12 10:43:47 +0200 (Wed, 12 Jul 2006) | 3 lines Fix #1467450: ctypes now uses RTLD_GLOBAL by default on OSX 10.3 to load shared libraries. ........ r50604 | thomas.heller | 2006-07-12 16:25:18 +0200 (Wed, 12 Jul 2006) | 3 lines Fix the wrong description of LibraryLoader.LoadLibrary, and document the DEFAULT_MODE constant. ........ r50607 | georg.brandl | 2006-07-12 17:31:17 +0200 (Wed, 12 Jul 2006) | 3 lines Accept long options "--help" and "--version". ........ r50617 | thomas.heller | 2006-07-13 11:53:47 +0200 (Thu, 13 Jul 2006) | 3 lines A misspelled preprocessor symbol caused ctypes to be always compiled without thread support. Replaced WITH_THREADS with WITH_THREAD. ........ r50619 | thomas.heller | 2006-07-13 19:01:14 +0200 (Thu, 13 Jul 2006) | 3 lines Fix #1521375. When running with root priviledges, 'gcc -o /dev/null' did overwrite /dev/null. Use a temporary file instead of /dev/null. ........ r50620 | thomas.heller | 2006-07-13 19:05:13 +0200 (Thu, 13 Jul 2006) | 2 lines Fix misleading words. ........ r50622 | andrew.kuchling | 2006-07-13 19:37:26 +0200 (Thu, 13 Jul 2006) | 1 line Typo fix ........ r50629 | georg.brandl | 2006-07-14 09:12:54 +0200 (Fri, 14 Jul 2006) | 3 lines Patch #1521874: grammar errors in doanddont.tex. ........ r50630 | neal.norwitz | 2006-07-14 09:20:04 +0200 (Fri, 14 Jul 2006) | 1 line Try to improve grammar further. ........ r50631 | martin.v.loewis | 2006-07-14 11:58:55 +0200 (Fri, 14 Jul 2006) | 1 line Extend build_ssl to Win64, using VSExtComp. ........ r50632 | martin.v.loewis | 2006-07-14 14:10:09 +0200 (Fri, 14 Jul 2006) | 1 line Add debug output to analyse buildbot failure. ........ r50633 | martin.v.loewis | 2006-07-14 14:31:05 +0200 (Fri, 14 Jul 2006) | 1 line Fix Debug build of _ssl. ........ r50636 | andrew.kuchling | 2006-07-14 15:32:38 +0200 (Fri, 14 Jul 2006) | 1 line Mention new options ........ r50638 | peter.astrand | 2006-07-14 16:04:45 +0200 (Fri, 14 Jul 2006) | 1 line Bug #1223937: CalledProcessError.errno -> CalledProcessError.returncode. ........ r50640 | thomas.heller | 2006-07-14 17:01:05 +0200 (Fri, 14 Jul 2006) | 4 lines Make the prototypes of our private PyUnicode_FromWideChar and PyUnicode_AsWideChar replacement functions compatible to the official functions by using Py_ssize_t instead of int. ........ r50643 | thomas.heller | 2006-07-14 19:51:14 +0200 (Fri, 14 Jul 2006) | 3 lines Patch #1521817: The index range checking on ctypes arrays containing exactly one element is enabled again. ........ r50647 | thomas.heller | 2006-07-14 20:22:50 +0200 (Fri, 14 Jul 2006) | 2 lines Updates for the ctypes documentation. ........ r50655 | fredrik.lundh | 2006-07-14 23:45:48 +0200 (Fri, 14 Jul 2006) | 3 lines typo ........ r50664 | george.yoshida | 2006-07-15 18:03:49 +0200 (Sat, 15 Jul 2006) | 2 lines Bug #15187702 : ext/win-cookbook.html has a broken link to distutils ........ r50667 | bob.ippolito | 2006-07-15 18:53:15 +0200 (Sat, 15 Jul 2006) | 1 line Patch #1220874: Update the binhex module for Mach-O. ........ r50671 | fred.drake | 2006-07-16 03:21:20 +0200 (Sun, 16 Jul 2006) | 1 line clean up some link markup ........ r50673 | neal.norwitz | 2006-07-16 03:50:38 +0200 (Sun, 16 Jul 2006) | 4 lines Bug #1512814, Fix incorrect lineno's when code within a function had more than 255 blank lines. Byte codes need to go first, line #s second. ........ r50674 | neal.norwitz | 2006-07-16 04:00:32 +0200 (Sun, 16 Jul 2006) | 5 lines a & b were dereffed above, so they are known to be valid pointers. z is known to be NULL, nothing to DECREF. Reported by Klockwork, #107. ........ r50675 | neal.norwitz | 2006-07-16 04:02:57 +0200 (Sun, 16 Jul 2006) | 5 lines self is dereffed (and passed as first arg), so it's known to be good. func is returned from PyArg_ParseTuple and also dereffed. Reported by Klocwork, #30 (self one at least). ........ r50676 | neal.norwitz | 2006-07-16 04:05:35 +0200 (Sun, 16 Jul 2006) | 4 lines proto was dereffed above and is known to be good. No need for X. Reported by Klocwork, #39. ........ r50677 | neal.norwitz | 2006-07-16 04:15:27 +0200 (Sun, 16 Jul 2006) | 5 lines Fix memory leaks in some conditions. Reported by Klocwork #152. ........ r50678 | neal.norwitz | 2006-07-16 04:17:36 +0200 (Sun, 16 Jul 2006) | 4 lines Fix memory leak under some conditions. Reported by Klocwork, #98. ........ r50679 | neal.norwitz | 2006-07-16 04:22:30 +0200 (Sun, 16 Jul 2006) | 8 lines Use sizeof(buffer) instead of duplicating the constants to ensure they won't be wrong. The real change is to pass (bufsz - 1) to PyOS_ascii_formatd and 1 to strncat. strncat copies n+1 bytes from src (not dest). Reported by Klocwork #58. ........ r50680 | neal.norwitz | 2006-07-16 04:32:03 +0200 (Sun, 16 Jul 2006) | 5 lines Handle a NULL name properly. Reported by Klocwork #67 ........ r50681 | neal.norwitz | 2006-07-16 04:35:47 +0200 (Sun, 16 Jul 2006) | 6 lines PyFunction_SetDefaults() is documented as taking None or a tuple. A NULL would crash the PyTuple_Check(). Now make NULL return a SystemError. Reported by Klocwork #73. ........ r50683 | neal.norwitz | 2006-07-17 02:55:45 +0200 (Mon, 17 Jul 2006) | 5 lines Stop INCREFing name, then checking if it's NULL. name (f_name) should never be NULL so assert it. Fix one place where we could have passed NULL. Reported by Klocwork #66. ........ r50684 | neal.norwitz | 2006-07-17 02:57:15 +0200 (Mon, 17 Jul 2006) | 5 lines otherset is known to be non-NULL based on checks before and DECREF after. DECREF otherset rather than XDECREF in error conditions too. Reported by Klockwork #154. ........ r50685 | neal.norwitz | 2006-07-17 02:59:04 +0200 (Mon, 17 Jul 2006) | 7 lines Reported by Klocwork #151. v2 can be NULL if exception2 is NULL. I don't think that condition can happen, but I'm not sure it can't either. Now the code will protect against either being NULL. ........ r50686 | neal.norwitz | 2006-07-17 03:00:16 +0200 (Mon, 17 Jul 2006) | 1 line Add NEWS entry for a bunch of fixes due to warnings produced by Klocworks static analysis tool. ........ r50687 | fred.drake | 2006-07-17 07:47:52 +0200 (Mon, 17 Jul 2006) | 3 lines document xmlcore (still minimal; needs mention in each of the xml.* modules) SF bug #1504456 (partial) ........ r50688 | georg.brandl | 2006-07-17 15:23:46 +0200 (Mon, 17 Jul 2006) | 3 lines Remove usage of sets module (patch #1500609). ........ r50689 | georg.brandl | 2006-07-17 15:26:33 +0200 (Mon, 17 Jul 2006) | 3 lines Add missing NEWS item (#1522771) ........ r50690 | andrew.kuchling | 2006-07-17 18:47:54 +0200 (Mon, 17 Jul 2006) | 1 line Attribute more features ........ r50692 | kurt.kaiser | 2006-07-17 23:59:27 +0200 (Mon, 17 Jul 2006) | 8 lines Patch 1479219 - Tal Einat 1. 'as' highlighted as builtin in comment string on import line 2. Comments such as "#False identity" which start with a keyword immediately after the '#' character aren't colored as comments. 3. u or U beginning unicode string not correctly highlighted Closes bug 1325071 ........ r50693 | barry.warsaw | 2006-07-18 01:07:51 +0200 (Tue, 18 Jul 2006) | 16 lines decode_rfc2231(): Be more robust against buggy RFC 2231 encodings. Specifically, instead of raising a ValueError when there is a single tick in the parameter, simply return that the entire string unquoted, with None for both the charset and the language. Also, if there are more than 2 ticks in the parameter, interpret the first three parts as the standard RFC 2231 parts, then the rest of the parts as the encoded string. Test cases added. Original fewer-than-3-parts fix by Tokio Kikuchi. Resolves SF bug # 1218081. I will back port the fix and tests to Python 2.4 (email 3.0) and Python 2.3 (email 2.5). Also, bump the version number to email 4.0.1, removing the 'alpha' moniker. ........ r50695 | kurt.kaiser | 2006-07-18 06:03:16 +0200 (Tue, 18 Jul 2006) | 2 lines Rebinding Tab key was inserting 'tab' instead of 'Tab'. Bug 1179168. ........ r50696 | brett.cannon | 2006-07-18 06:41:36 +0200 (Tue, 18 Jul 2006) | 6 lines Fix bug #1520914. Starting in 2.4, time.strftime() began to check the bounds of values in the time tuple passed in. Unfortunately people came to rely on undocumented behaviour of setting unneeded values to 0, regardless of if it was within the valid range. Now those values force the value internally to the minimum value when 0 is passed in. ........ r50697 | facundo.batista | 2006-07-18 14:16:13 +0200 (Tue, 18 Jul 2006) | 1 line Comments and docs cleanups, and some little fixes, provided by Santiágo Peresón ........ r50704 | martin.v.loewis | 2006-07-18 19:46:31 +0200 (Tue, 18 Jul 2006) | 2 lines Patch #1524429: Use repr instead of backticks again. ........ r50706 | tim.peters | 2006-07-18 23:55:15 +0200 (Tue, 18 Jul 2006) | 2 lines Whitespace normalization. ........ r50708 | tim.peters | 2006-07-19 02:03:19 +0200 (Wed, 19 Jul 2006) | 18 lines SF bug 1524317: configure --without-threads fails to build Moved the code for _PyThread_CurrentFrames() up, so it's no longer in a huge "#ifdef WITH_THREAD" block (I didn't realize it /was/ in one). Changed test_sys's test_current_frames() so it passes with or without thread supported compiled in. Note that test_sys fails when Python is compiled without threads, but for an unrelated reason (the old test_exit() fails with an indirect ImportError on the `thread` module). There are also other unrelated compilation failures without threads, in extension modules (like ctypes); at least the core compiles again. Do we really support --without-threads? If so, there are several problems remaining. ........ r50713 | thomas.heller | 2006-07-19 11:09:32 +0200 (Wed, 19 Jul 2006) | 4 lines Make sure the _ctypes extension can be compiled when WITH_THREAD is not defined on Windows, even if that configuration is probably not supported at all. ........ r50715 | martin.v.loewis | 2006-07-19 19:18:32 +0200 (Wed, 19 Jul 2006) | 4 lines Revert r50706 (Whitespace normalization) and r50697: Comments and docs cleanups, and some little fixes per recommendation from Raymond Hettinger. ........ r50719 | phillip.eby | 2006-07-20 17:54:16 +0200 (Thu, 20 Jul 2006) | 4 lines Fix SF#1516184 (again) and add a test to prevent regression. (There was a problem with empty filenames still causing recursion) ........ r50720 | georg.brandl | 2006-07-20 18:28:39 +0200 (Thu, 20 Jul 2006) | 3 lines Guard for _active being None in __del__ method. ........ r50721 | vinay.sajip | 2006-07-20 18:28:39 +0200 (Thu, 20 Jul 2006) | 1 line Updated documentation for TimedRotatingFileHandler relating to how rollover files are named. The previous documentation was wrongly the same as for RotatingFileHandler. ........ r50731 | fred.drake | 2006-07-20 22:11:57 +0200 (Thu, 20 Jul 2006) | 1 line markup fix ........ r50739 | kurt.kaiser | 2006-07-21 00:22:52 +0200 (Fri, 21 Jul 2006) | 7 lines Avoid occasional failure to detect closing paren properly. Patch 1407280 Tal Einat M ParenMatch.py M NEWS.txt M CREDITS.txt ........ r50740 | vinay.sajip | 2006-07-21 01:20:12 +0200 (Fri, 21 Jul 2006) | 1 line Addressed SF#1524081 by using a dictionary to map level names to syslog priority names, rather than a string.lower(). ........ r50741 | neal.norwitz | 2006-07-21 07:29:58 +0200 (Fri, 21 Jul 2006) | 1 line Add some asserts that we got good params passed ........ r50742 | neal.norwitz | 2006-07-21 07:31:02 +0200 (Fri, 21 Jul 2006) | 5 lines Move the initialization of some pointers earlier. The problem is that if we call Py_DECREF(frame) like we do if allocating locals fails, frame_dealloc() will try to use these bogus values and crash. ........ r50743 | neal.norwitz | 2006-07-21 07:32:28 +0200 (Fri, 21 Jul 2006) | 4 lines Handle allocation failures gracefully. Found with failmalloc. Many (all?) of these could be backported. ........ r50745 | neal.norwitz | 2006-07-21 09:59:02 +0200 (Fri, 21 Jul 2006) | 1 line Speel initialise write. Tanks Anthony. ........ r50746 | neal.norwitz | 2006-07-21 09:59:47 +0200 (Fri, 21 Jul 2006) | 2 lines Handle more memory allocation failures without crashing. ........ r50754 | barry.warsaw | 2006-07-21 16:51:07 +0200 (Fri, 21 Jul 2006) | 23 lines More RFC 2231 improvements for the email 4.0 package. As Mark Sapiro rightly points out there are really two types of continued headers defined in this RFC (i.e. "encoded" parameters with the form "name*0*=" and unencoded parameters with the form "name*0="), but we were were handling them both the same way and that isn't correct. This patch should be much more RFC compliant in that only encoded params are %-decoded and the charset/language information is only extract if there are any encoded params in the segments. If there are no encoded params then the RFC says that there will be no charset/language parts. Note however that this will change the return value for Message.get_param() in some cases. For example, whereas before if you had all unencoded param continuations you would have still gotten a 3-tuple back from this method (with charset and language == None), you will now get just a string. I don't believe this is a backward incompatible change though because the documentation for this method already indicates that either return value is possible and that you must do an isinstance(val, tuple) check to discriminate between the two. (Yeah that API kind of sucks but we can't change /that/ without breaking code.) Test cases, some documentation updates, and a NEWS item accompany this patch. ........ r50759 | georg.brandl | 2006-07-21 19:36:31 +0200 (Fri, 21 Jul 2006) | 3 lines Fix check for empty list (vs. None). ........ r50771 | brett.cannon | 2006-07-22 00:44:07 +0200 (Sat, 22 Jul 2006) | 2 lines Remove an XXX marker in a comment. ........ r50773 | neal.norwitz | 2006-07-22 18:20:49 +0200 (Sat, 22 Jul 2006) | 1 line Fix more memory allocation issues found with failmalloc. ........ r50774 | neal.norwitz | 2006-07-22 19:00:57 +0200 (Sat, 22 Jul 2006) | 1 line Don't fail if the directory already exists ........ r50775 | greg.ward | 2006-07-23 04:25:53 +0200 (Sun, 23 Jul 2006) | 6 lines Be a lot smarter about whether this test passes: instead of assuming that a 2.93 sec audio file will always take 3.1 sec (as it did on the hardware I had when I first wrote the test), expect that it will take 2.93 sec +/- 10%, and only fail if it's outside of that range. Compute the expected ........ r50776 | kurt.kaiser | 2006-07-23 06:19:49 +0200 (Sun, 23 Jul 2006) | 2 lines Tooltips failed on new-syle class __init__ args. Bug 1027566 Loren Guthrie ........ r50777 | neal.norwitz | 2006-07-23 09:50:36 +0200 (Sun, 23 Jul 2006) | 1 line Handle more mem alloc issues found with failmalloc ........ r50778 | neal.norwitz | 2006-07-23 09:51:58 +0200 (Sun, 23 Jul 2006) | 5 lines If the for loop isn't entered, entryblock will be NULL. If passed to stackdepth_walk it will be dereffed. Not sure if I found with failmalloc or Klockwork #55. ........ r50779 | neal.norwitz | 2006-07-23 09:53:14 +0200 (Sun, 23 Jul 2006) | 4 lines Move the initialization of size_a down below the check for a being NULL. Reported by Klocwork #106 ........ r50780 | neal.norwitz | 2006-07-23 09:55:55 +0200 (Sun, 23 Jul 2006) | 9 lines Check the allocation of b_objects and return if there was a failure. Also fix a few memory leaks in other failure scenarios. It seems that if b_objects == Py_None, we will have an extra ref to b_objects. Add XXX comment so hopefully someone documents why the else isn't necessary or adds it in. Reported by Klocwork #20 ........ r50781 | neal.norwitz | 2006-07-23 09:57:11 +0200 (Sun, 23 Jul 2006) | 2 lines Fix memory leaks spotted by Klocwork #37. ........ r50782 | neal.norwitz | 2006-07-23 09:59:00 +0200 (Sun, 23 Jul 2006) | 5 lines nextlink can be NULL if teedataobject_new fails, so use XINCREF. Ensure that dataobj is never NULL. Reported by Klocwork #102 ........ r50783 | neal.norwitz | 2006-07-23 10:01:43 +0200 (Sun, 23 Jul 2006) | 8 lines Ensure we don't write beyond errText. I think I got this right, but it definitely could use some review to ensure I'm not off by one and there's no possible overflow/wrap-around of bytes_left. Reported by Klocwork #1. Fix a problem if there is a failure allocating self->db. Found with failmalloc. ........ r50784 | ronald.oussoren | 2006-07-23 11:41:09 +0200 (Sun, 23 Jul 2006) | 3 lines Without this patch CMD-W won't close EditorWindows on MacOS X. This solves part of bug #1517990. ........ r50785 | ronald.oussoren | 2006-07-23 11:46:11 +0200 (Sun, 23 Jul 2006) | 5 lines Fix for bug #1517996: Class and Path browsers show Tk menu This patch replaces the menubar that is used by AquaTk for windows without a menubar of their own by one that is more appropriate for IDLE. ........ r50786 | andrew.macintyre | 2006-07-23 14:57:02 +0200 (Sun, 23 Jul 2006) | 2 lines Build updates for OS/2 EMX port ........ r50787 | andrew.macintyre | 2006-07-23 15:00:04 +0200 (Sun, 23 Jul 2006) | 3 lines bugfix: PyThread_start_new_thread() returns the thread ID, not a flag; will backport. ........ r50789 | andrew.macintyre | 2006-07-23 15:04:00 +0200 (Sun, 23 Jul 2006) | 2 lines Get mailbox module working on OS/2 EMX port. ........ r50791 | greg.ward | 2006-07-23 18:05:51 +0200 (Sun, 23 Jul 2006) | 1 line Resync optparse with Optik 1.5.3: minor tweaks for/to tests. ........ r50794 | martin.v.loewis | 2006-07-24 07:05:22 +0200 (Mon, 24 Jul 2006) | 2 lines Update list of unsupported systems. Fixes #1510853. ........ r50795 | martin.v.loewis | 2006-07-24 12:26:33 +0200 (Mon, 24 Jul 2006) | 1 line Patch #1448199: Release GIL around ConnectRegistry. ........ r50796 | martin.v.loewis | 2006-07-24 13:54:53 +0200 (Mon, 24 Jul 2006) | 3 lines Patch #1232023: Don't include empty path component from registry, so that the current directory does not get added to sys.path. Also fixes #1526785. ........ r50797 | martin.v.loewis | 2006-07-24 14:54:17 +0200 (Mon, 24 Jul 2006) | 3 lines Bug #1524310: Properly report errors from FindNextFile in os.listdir. Will backport to 2.4. ........ r50800 | georg.brandl | 2006-07-24 15:28:57 +0200 (Mon, 24 Jul 2006) | 7 lines Patch #1523356: fix determining include dirs in python-config. Also don't install "python-config" when doing altinstall, but always install "python-config2.x" and make a link to it like with the main executable. ........ r50802 | georg.brandl | 2006-07-24 15:46:47 +0200 (Mon, 24 Jul 2006) | 3 lines Patch #1527744: right order of includes in order to have HAVE_CONIO_H defined properly. ........ r50803 | georg.brandl | 2006-07-24 16:09:56 +0200 (Mon, 24 Jul 2006) | 3 lines Patch #1515343: Fix printing of deprecated string exceptions with a value in the traceback module. ........ r50804 | kurt.kaiser | 2006-07-24 19:13:23 +0200 (Mon, 24 Jul 2006) | 7 lines EditorWindow failed when used stand-alone if sys.ps1 not set. Bug 1010370 Dave Florek M EditorWindow.py M PyShell.py M NEWS.txt ........ r50805 | kurt.kaiser | 2006-07-24 20:05:51 +0200 (Mon, 24 Jul 2006) | 6 lines - EditorWindow.test() was failing. Bug 1417598 M EditorWindow.py M ScriptBinding.py M NEWS.txt ........ r50808 | georg.brandl | 2006-07-24 22:11:35 +0200 (Mon, 24 Jul 2006) | 3 lines Repair accidental NameError. ........ r50809 | tim.peters | 2006-07-24 23:02:15 +0200 (Mon, 24 Jul 2006) | 2 lines Whitespace normalization. ........ r50810 | greg.ward | 2006-07-25 04:11:12 +0200 (Tue, 25 Jul 2006) | 3 lines Don't use standard assert: want tests to fail even when run with -O. Delete cruft. ........ r50811 | tim.peters | 2006-07-25 06:07:22 +0200 (Tue, 25 Jul 2006) | 10 lines current_frames_with_threads(): There's actually no way to guess /which/ line the spawned thread is in at the time sys._current_frames() is called: we know it finished enter_g.set(), but can't know whether the instruction counter has advanced to the following leave_g.wait(). The latter is overwhelming most likely, but not guaranteed, and I see that the "x86 Ubuntu dapper (icc) trunk" buildbot found it on the other line once. Changed the test so it passes in either case. ........ r50815 | martin.v.loewis | 2006-07-25 11:53:12 +0200 (Tue, 25 Jul 2006) | 2 lines Bug #1525817: Don't truncate short lines in IDLE's tool tips. ........ r50816 | martin.v.loewis | 2006-07-25 12:05:47 +0200 (Tue, 25 Jul 2006) | 3 lines Bug #978833: Really close underlying socket in _socketobject.close. Will backport to 2.4. ........ r50817 | martin.v.loewis | 2006-07-25 12:11:14 +0200 (Tue, 25 Jul 2006) | 1 line Revert incomplete checkin. ........ r50819 | georg.brandl | 2006-07-25 12:22:34 +0200 (Tue, 25 Jul 2006) | 4 lines Patch #1525766: correctly pass onerror arg to recursive calls of pkg.walk_packages. Also improve the docstrings. ........ r50825 | brett.cannon | 2006-07-25 19:32:20 +0200 (Tue, 25 Jul 2006) | 2 lines Add comment for changes to test_ossaudiodev. ........ r50826 | brett.cannon | 2006-07-25 19:34:36 +0200 (Tue, 25 Jul 2006) | 3 lines Fix a bug in the messages for an assert failure where not enough arguments to a string were being converted in the format. ........ r50828 | armin.rigo | 2006-07-25 20:09:57 +0200 (Tue, 25 Jul 2006) | 2 lines Document why is and is not a good way to fix the gc_inspection crasher. ........ r50829 | armin.rigo | 2006-07-25 20:11:07 +0200 (Tue, 25 Jul 2006) | 5 lines Added another crasher, which hit me today (I was not intentionally writing such code, of course, but it took some gdb time to figure out what my bug was). ........ r50830 | armin.rigo | 2006-07-25 20:38:39 +0200 (Tue, 25 Jul 2006) | 3 lines Document the crashers that will not go away soon as "won't fix", and explain why. ........ r50831 | ronald.oussoren | 2006-07-25 21:13:35 +0200 (Tue, 25 Jul 2006) | 3 lines Install the compatibility symlink to libpython.a on OSX using 'ln -sf' instead of 'ln -s', this avoid problems when reinstalling python. ........ r50832 | ronald.oussoren | 2006-07-25 21:20:54 +0200 (Tue, 25 Jul 2006) | 7 lines Fix for bug #1525447 (renaming to MacOSmodule.c would also work, but not without causing problems for anyone that is on a case-insensitive filesystem). Setup.py tries to compile the MacOS extension from MacOSmodule.c, while the actual file is named macosmodule.c. This is no problem on the (default) case-insensitive filesystem, but doesn't work on case-sensitive filesystems. ........ r50833 | ronald.oussoren | 2006-07-25 22:28:55 +0200 (Tue, 25 Jul 2006) | 7 lines Fix bug #1517990: IDLE keybindings on OSX This adds a new key definition for OSX, which is slightly different from the classic mac definition. Also add NEWS item for a couple of bugfixes I added recently. ........ r50834 | tim.peters | 2006-07-26 00:30:24 +0200 (Wed, 26 Jul 2006) | 2 lines Whitespace normalization. ........ r50839 | neal.norwitz | 2006-07-26 06:00:18 +0200 (Wed, 26 Jul 2006) | 1 line Hmm, only python2.x is installed, not plain python. Did that change recently? ........ r50840 | barry.warsaw | 2006-07-26 07:54:46 +0200 (Wed, 26 Jul 2006) | 6 lines Forward port some fixes that were in email 2.5 but for some reason didn't make it into email 4.0. Specifically, in Message.get_content_charset(), handle RFC 2231 headers that contain an encoding not known to Python, or a character in the data that isn't in the charset encoding. Also forward port the appropriate unit tests. ........ r50841 | georg.brandl | 2006-07-26 09:23:32 +0200 (Wed, 26 Jul 2006) | 3 lines NEWS entry for #1525766. ........ r50842 | georg.brandl | 2006-07-26 09:40:17 +0200 (Wed, 26 Jul 2006) | 3 lines Bug #1459963: properly capitalize HTTP header names. ........ r50843 | georg.brandl | 2006-07-26 10:03:10 +0200 (Wed, 26 Jul 2006) | 6 lines Part of bug #1523610: fix miscalculation of buffer length. Also add a guard against NULL in converttuple and add a test case (that previously would have crashed). ........ r50844 | martin.v.loewis | 2006-07-26 14:12:56 +0200 (Wed, 26 Jul 2006) | 3 lines Bug #978833: Really close underlying socket in _socketobject.close. Fix httplib.HTTPConnection.getresponse to not close the socket if it is still needed for the response. ........ r50845 | andrew.kuchling | 2006-07-26 19:16:52 +0200 (Wed, 26 Jul 2006) | 1 line [Bug #1471938] Fix build problem on Solaris 8 by conditionalizing the use of mvwgetnstr(); it was conditionalized a few lines below. Fix from Paul Eggert. I also tried out the STRICT_SYSV_CURSES case and am therefore removing the 'untested' comment. ........ r50846 | andrew.kuchling | 2006-07-26 19:18:01 +0200 (Wed, 26 Jul 2006) | 1 line Correct error message ........ r50847 | andrew.kuchling | 2006-07-26 19:19:39 +0200 (Wed, 26 Jul 2006) | 1 line Minor grammar fix ........ r50848 | andrew.kuchling | 2006-07-26 19:22:21 +0200 (Wed, 26 Jul 2006) | 1 line Put news item in right section ........ r50850 | andrew.kuchling | 2006-07-26 20:03:12 +0200 (Wed, 26 Jul 2006) | 1 line Use sys.exc_info() ........ r50851 | andrew.kuchling | 2006-07-26 20:15:45 +0200 (Wed, 26 Jul 2006) | 1 line Use sys.exc_info() ........ r50852 | phillip.eby | 2006-07-26 21:48:27 +0200 (Wed, 26 Jul 2006) | 4 lines Allow the 'onerror' argument to walk_packages() to catch any Exception, not just ImportError. This allows documentation tools to better skip unimportable packages. ........ r50854 | tim.peters | 2006-07-27 01:23:15 +0200 (Thu, 27 Jul 2006) | 2 lines Whitespace normalization. ........ r50855 | tim.peters | 2006-07-27 03:14:53 +0200 (Thu, 27 Jul 2006) | 21 lines Bug #1521947: possible bug in mystrtol.c with recent gcc. In general, C doesn't define anything about what happens when an operation on a signed integral type overflows, and PyOS_strtol() did several formally undefined things of that nature on signed longs. Some version of gcc apparently tries to exploit that now, and PyOS_strtol() could fail to detect overflow then. Tried to repair all that, although it seems at least as likely to me that we'll get screwed by bad platform definitions for LONG_MIN and/or LONG_MAX now. For that reason, I don't recommend backporting this. Note that I have no box on which this makes a lick of difference -- can't really test it, except to note that it didn't break anything on my boxes. Silent change: PyOS_strtol() used to return the hard-coded 0x7fffffff in case of overflow. Now it returns LONG_MAX. They're the same only on 32-bit boxes (although C doesn't guarantee that either ...). ........ r50856 | neal.norwitz | 2006-07-27 05:51:58 +0200 (Thu, 27 Jul 2006) | 6 lines Don't kill a normal instance of python running on windows when checking to kill a cygwin instance. build\\python.exe was matching a normal windows instance. Prefix that with a \\ to ensure build is a directory and not PCbuild. As discussed on python-dev. ........ r50857 | neal.norwitz | 2006-07-27 05:55:39 +0200 (Thu, 27 Jul 2006) | 5 lines Closure can't be NULL at this point since we know it's a tuple. Reported by Klocwork # 74. ........ r50858 | neal.norwitz | 2006-07-27 06:04:50 +0200 (Thu, 27 Jul 2006) | 1 line No functional change. Add comment and assert to describe why there cannot be overflow which was reported by Klocwork. Discussed on python-dev ........ r50859 | martin.v.loewis | 2006-07-27 08:38:16 +0200 (Thu, 27 Jul 2006) | 3 lines Bump distutils version to 2.5, as several new features have been introduced since 2.4. ........ r50860 | andrew.kuchling | 2006-07-27 14:18:20 +0200 (Thu, 27 Jul 2006) | 1 line Reformat docstring; fix typo ........ r50861 | georg.brandl | 2006-07-27 17:05:36 +0200 (Thu, 27 Jul 2006) | 6 lines Add test_main() methods. These three tests were never run by regrtest.py. We really need a simpler testing framework. ........ r50862 | tim.peters | 2006-07-27 17:09:20 +0200 (Thu, 27 Jul 2006) | 2 lines News for patch #1529686. ........ r50863 | tim.peters | 2006-07-27 17:11:00 +0200 (Thu, 27 Jul 2006) | 2 lines Whitespace normalization. ........ r50864 | georg.brandl | 2006-07-27 17:38:33 +0200 (Thu, 27 Jul 2006) | 3 lines Amend news entry. ........ r50865 | georg.brandl | 2006-07-27 18:08:15 +0200 (Thu, 27 Jul 2006) | 3 lines Make uuid test suite pass on this box by requesting output with LC_ALL=C. ........ r50866 | andrew.kuchling | 2006-07-27 20:37:33 +0200 (Thu, 27 Jul 2006) | 1 line Add example ........ r50867 | thomas.heller | 2006-07-27 20:39:55 +0200 (Thu, 27 Jul 2006) | 9 lines Remove code that is no longer used (ctypes.com). Fix the DllGetClassObject and DllCanUnloadNow so that they forward the call to the comtypes.server.inprocserver module. The latter was never documented, never used by published code, and didn't work anyway, so I think it does not deserve a NEWS entry (but I might be wrong). ........ r50868 | andrew.kuchling | 2006-07-27 20:41:21 +0200 (Thu, 27 Jul 2006) | 1 line Typo fix ('publically' is rare, poss. non-standard) ........ r50869 | andrew.kuchling | 2006-07-27 20:42:41 +0200 (Thu, 27 Jul 2006) | 1 line Add missing word ........ r50870 | andrew.kuchling | 2006-07-27 20:44:10 +0200 (Thu, 27 Jul 2006) | 1 line Repair typos ........ r50872 | andrew.kuchling | 2006-07-27 20:53:33 +0200 (Thu, 27 Jul 2006) | 1 line Update URL; add example ........ r50873 | andrew.kuchling | 2006-07-27 21:07:29 +0200 (Thu, 27 Jul 2006) | 1 line Add punctuation mark; add some examples ........ r50874 | andrew.kuchling | 2006-07-27 21:11:07 +0200 (Thu, 27 Jul 2006) | 1 line Mention base64 module; rewrite last sentence to be more positive ........ r50875 | andrew.kuchling | 2006-07-27 21:12:49 +0200 (Thu, 27 Jul 2006) | 1 line If binhex is higher-level than binascii, it should come first in the chapter ........ r50876 | tim.peters | 2006-07-27 22:47:24 +0200 (Thu, 27 Jul 2006) | 28 lines check_node(): stop spraying mystery output to stderr. When a node number disagrees, keep track of all sources & the node numbers they reported, and stick all that in the error message. Changed all callers to supply a non-empty "source" argument; made the "source" argument non-optional. On my box, test_uuid still fails, but with the less confusing output: AssertionError: different sources disagree on node: from source 'getnode1', node was 00038a000015 from source 'getnode2', node was 00038a000015 from source 'ipconfig', node was 001111b2b7bf Only the last one appears to be correct; e.g., C:\Code\python\PCbuild>getmac Physical Address Transport Name =================== ========================================================== 00-11-11-B2-B7-BF \Device\Tcpip_{190FB163-5AFD-4483-86A1-2FE16AC61FF1} 62-A1-AC-6C-FD-BE \Device\Tcpip_{8F77DF5A-EA3D-4F1D-975E-D472CEE6438A} E2-1F-01-C6-5D-88 \Device\Tcpip_{CD18F76B-2EF3-409F-9B8A-6481EE70A1E4} I can't find anything on my box with MAC 00-03-8a-00-00-15, and am not clear on where that comes from. ........ r50878 | andrew.kuchling | 2006-07-28 00:40:05 +0200 (Fri, 28 Jul 2006) | 1 line Reword paragraph ........ r50879 | andrew.kuchling | 2006-07-28 00:49:38 +0200 (Fri, 28 Jul 2006) | 1 line Add example ........ r50880 | andrew.kuchling | 2006-07-28 00:49:54 +0200 (Fri, 28 Jul 2006) | 1 line Add example ........ r50881 | barry.warsaw | 2006-07-28 01:43:15 +0200 (Fri, 28 Jul 2006) | 27 lines Patch #1520294: Support for getset and member descriptors in types.py, inspect.py, and pydoc.py. Specifically, this allows for querying the type of an object against these built-in C types and more importantly, for getting their docstrings printed in the interactive interpreter's help() function. This patch includes a new built-in module called _types which provides definitions of getset and member descriptors for use by the types.py module. These types are exposed as types.GetSetDescriptorType and types.MemberDescriptorType. Query functions are provided as inspect.isgetsetdescriptor() and inspect.ismemberdescriptor(). The implementations of these are robust enough to work with Python implementations other than CPython, which may not have these fundamental types. The patch also includes documentation and test suite updates. I commit these changes now under these guiding principles: 1. Silence is assent. The release manager has not said "no", and of the few people that cared enough to respond to the thread, the worst vote was "0". 2. It's easier to ask for forgiveness than permission. 3. It's so dang easy to revert stuff in svn, that you could view this as a forcing function. :) Windows build patches will follow. ........ r50882 | tim.peters | 2006-07-28 01:44:37 +0200 (Fri, 28 Jul 2006) | 4 lines Bug #1529297: The rewrite of doctest for Python 2.4 unintentionally lost that tests are sorted by name before being run. ``DocTestFinder`` has been changed to sort the list of tests it returns. ........ r50883 | tim.peters | 2006-07-28 01:45:48 +0200 (Fri, 28 Jul 2006) | 2 lines Whitespace normalization. ........ r50884 | tim.peters | 2006-07-28 01:46:36 +0200 (Fri, 28 Jul 2006) | 2 lines Add missing svn:eol-style property to text files. ........ r50885 | barry.warsaw | 2006-07-28 01:50:40 +0200 (Fri, 28 Jul 2006) | 4 lines Enable the building of the _types module on Windows. Note that this has only been tested for VS 2003 since that's all I have. ........ r50887 | tim.peters | 2006-07-28 02:23:15 +0200 (Fri, 28 Jul 2006) | 7 lines defdict_reduce(): Plug leaks. We didn't notice these before because test_defaultdict didn't actually do anything before Georg fixed that earlier today. Neal's next refleak run then showed test_defaultdict leaking 9 references on each run. That's repaired by this checkin. ........ r50888 | tim.peters | 2006-07-28 02:30:00 +0200 (Fri, 28 Jul 2006) | 2 lines News about the repaired memory leak in defaultdict. ........ r50889 | gregory.p.smith | 2006-07-28 03:35:25 +0200 (Fri, 28 Jul 2006) | 7 lines - pybsddb Bug #1527939: bsddb module DBEnv dbremove and dbrename methods now allow their database parameter to be None as the sleepycat API allows. Also adds an appropriate test case for DBEnv.dbrename and dbremove. ........ r50895 | neal.norwitz | 2006-07-28 06:22:34 +0200 (Fri, 28 Jul 2006) | 1 line Ensure the actual number matches the expected count ........ r50896 | tim.peters | 2006-07-28 06:51:59 +0200 (Fri, 28 Jul 2006) | 6 lines Live with that "the hardware address" is an ill-defined concept, and that different ways of trying to find "the hardware address" may return different results. Certainly true on both of my Windows boxes, and in different ways (see whining on python-dev). ........ r50897 | neal.norwitz | 2006-07-28 09:21:27 +0200 (Fri, 28 Jul 2006) | 3 lines Try to find the MAC addr on various flavours of Unix. This seems hopeless. The reduces the test_uuid failures, but there's still another method failing. ........ r50898 | martin.v.loewis | 2006-07-28 09:45:49 +0200 (Fri, 28 Jul 2006) | 2 lines Add UUID for upcoming 2.5b3. ........ r50899 | matt.fleming | 2006-07-28 13:27:27 +0200 (Fri, 28 Jul 2006) | 3 lines Allow socketmodule to compile on NetBSD -current, whose bluetooth API differs from both Linux and FreeBSD. Accepted by Neal Norwitz. ........ r50900 | andrew.kuchling | 2006-07-28 14:07:12 +0200 (Fri, 28 Jul 2006) | 1 line [Patch #1529811] Correction to description of r|* mode ........ r50901 | andrew.kuchling | 2006-07-28 14:18:22 +0200 (Fri, 28 Jul 2006) | 1 line Typo fix ........ r50902 | andrew.kuchling | 2006-07-28 14:32:43 +0200 (Fri, 28 Jul 2006) | 1 line Add example ........ r50903 | andrew.kuchling | 2006-07-28 14:33:19 +0200 (Fri, 28 Jul 2006) | 1 line Add example ........ r50904 | andrew.kuchling | 2006-07-28 14:45:55 +0200 (Fri, 28 Jul 2006) | 1 line Don't overwrite built-in name; add some blank lines for readability ........ r50905 | andrew.kuchling | 2006-07-28 14:48:07 +0200 (Fri, 28 Jul 2006) | 1 line Add example. Should I propagate this example to all the other DBM-ish modules, too? ........ r50912 | georg.brandl | 2006-07-28 20:31:39 +0200 (Fri, 28 Jul 2006) | 3 lines Patch #1529686: also run test_email_codecs with regrtest.py. ........ r50913 | georg.brandl | 2006-07-28 20:36:01 +0200 (Fri, 28 Jul 2006) | 3 lines Fix spelling. ........ r50915 | thomas.heller | 2006-07-28 21:42:40 +0200 (Fri, 28 Jul 2006) | 3 lines Remove a useless XXX comment. Cosmetic changes to the code so that the #ifdef _UNICODE block doesn't mess emacs code formatting. ........ r50916 | phillip.eby | 2006-07-28 23:12:07 +0200 (Fri, 28 Jul 2006) | 5 lines Bug #1529871: The speed enhancement patch #921466 broke Python's compliance with PEP 302. This was fixed by adding an ``imp.NullImporter`` type that is used in ``sys.path_importer_cache`` to cache non-directory paths and avoid excessive filesystem operations during imports. ........ r50917 | phillip.eby | 2006-07-28 23:31:54 +0200 (Fri, 28 Jul 2006) | 2 lines Fix svn merge spew. ........ r50918 | thomas.heller | 2006-07-28 23:43:20 +0200 (Fri, 28 Jul 2006) | 4 lines Patch #1529514: More openbsd platforms for ctypes. Regenerated Modules/_ctypes/libffi/configure with autoconf 2.59. Approved by Neal. ........ r50922 | georg.brandl | 2006-07-29 10:51:21 +0200 (Sat, 29 Jul 2006) | 2 lines Bug #835255: The "closure" argument to new.function() is now documented. ........ r50924 | georg.brandl | 2006-07-29 11:33:26 +0200 (Sat, 29 Jul 2006) | 3 lines Bug #1441397: The compiler module now recognizes module and function docstrings correctly as it did in Python 2.4. ........ r50925 | georg.brandl | 2006-07-29 12:25:46 +0200 (Sat, 29 Jul 2006) | 4 lines Revert rev 42617, it was introduced to work around bug #1441397. test_compiler now passes again. ........ r50926 | fred.drake | 2006-07-29 15:22:49 +0200 (Sat, 29 Jul 2006) | 1 line update target version number ........ r50927 | andrew.kuchling | 2006-07-29 15:56:48 +0200 (Sat, 29 Jul 2006) | 1 line Add example ........ r50928 | andrew.kuchling | 2006-07-29 16:04:47 +0200 (Sat, 29 Jul 2006) | 1 line Update URL ........ r50930 | andrew.kuchling | 2006-07-29 16:08:15 +0200 (Sat, 29 Jul 2006) | 1 line Reword paragraph to match the order of the subsequent sections ........ r50931 | andrew.kuchling | 2006-07-29 16:21:15 +0200 (Sat, 29 Jul 2006) | 1 line [Bug #1529157] Mention raw_input() and input(); while I'm at it, reword the description a bit ........ r50932 | andrew.kuchling | 2006-07-29 16:42:48 +0200 (Sat, 29 Jul 2006) | 1 line [Bug #1519571] Document some missing functions: setup(), title(), done() ........ r50933 | andrew.kuchling | 2006-07-29 16:43:55 +0200 (Sat, 29 Jul 2006) | 1 line Fix docstring punctuation ........ r50934 | andrew.kuchling | 2006-07-29 17:10:32 +0200 (Sat, 29 Jul 2006) | 1 line [Bug #1414697] Change docstring of set/frozenset types to specify that the contents are unique. Raymond, please feel free to edit or revert. ........ r50935 | andrew.kuchling | 2006-07-29 17:35:21 +0200 (Sat, 29 Jul 2006) | 1 line [Bug #1530382] Document SSL.server(), .issuer() methods ........ r50936 | andrew.kuchling | 2006-07-29 17:42:46 +0200 (Sat, 29 Jul 2006) | 1 line Typo fix ........ r50937 | andrew.kuchling | 2006-07-29 17:43:13 +0200 (Sat, 29 Jul 2006) | 1 line Tweak wording ........ r50938 | matt.fleming | 2006-07-29 17:55:30 +0200 (Sat, 29 Jul 2006) | 2 lines Fix typo ........ r50939 | andrew.kuchling | 2006-07-29 17:57:08 +0200 (Sat, 29 Jul 2006) | 6 lines [Bug #1528258] Mention that the 'data' argument can be None. The constructor docs referred the reader to the add_data() method's docs, but they weren't very helpful. I've simply copied an earlier explanation of 'data' that's more useful. ........ r50940 | andrew.kuchling | 2006-07-29 18:08:40 +0200 (Sat, 29 Jul 2006) | 1 line Set bug/patch count. Take a bow, everyone! ........ r50941 | fred.drake | 2006-07-29 18:56:15 +0200 (Sat, 29 Jul 2006) | 18 lines expunge the xmlcore changes: 41667, 41668 - initial switch to xmlcore 47044 - mention of xmlcore in What's New 50687 - mention of xmlcore in the library reference re-apply xmlcore changes to xml: 41674 - line ending changes (re-applied manually), directory props 41677 - add cElementTree wrapper 41678 - PSF licensing for etree 41812 - whitespace normalization 42724 - fix svn:eol-style settings 43681, 43682 - remove Python version-compatibility cruft from minidom 46773 - fix encoding of \r\n\t in attr values in saxutils 47269 - added XMLParser alias for cElementTree compatibility additional tests were added in Lib/test/test_sax.py that failed with the xmlcore changes; these relate to SF bugs #1511497, #1513611 ........ r50942 | andrew.kuchling | 2006-07-29 20:14:07 +0200 (Sat, 29 Jul 2006) | 17 lines Reorganize the docs for 'file' and 'open()' after some discussion with Fred. We want to encourage users to write open() when opening a file, but open() was described with a single paragraph and 'file' had lots of explanation of the mode and bufsize arguments. I've shrunk the description of 'file' to cross-reference to the 'File objects' section, and to open() for an explanation of the arguments. open() now has all the paragraphs about the mode string. The bufsize argument was moved up so that it isn't buried at the end; now there's 1 paragraph on mode, 1 on bufsize, and then 3 more on mode. Various other edits and rearrangements were made in the process. It's probably best to read the final text and not to try to make sense of the diffs. ........ r50943 | fred.drake | 2006-07-29 20:19:19 +0200 (Sat, 29 Jul 2006) | 1 line restore test un-intentionally removed in the xmlcore purge (revision 50941) ........ r50944 | fred.drake | 2006-07-29 20:33:29 +0200 (Sat, 29 Jul 2006) | 3 lines make the reference to older versions of the documentation a link to the right page on python.org ........ r50945 | fred.drake | 2006-07-29 21:09:01 +0200 (Sat, 29 Jul 2006) | 1 line document the footnote usage pattern ........ r50947 | fred.drake | 2006-07-29 21:14:10 +0200 (Sat, 29 Jul 2006) | 1 line emphasize and oddball nuance of LaTeX comment syntax ........ r50948 | andrew.kuchling | 2006-07-29 21:24:04 +0200 (Sat, 29 Jul 2006) | 1 line [Patch #1490989 from Skip Montanaro] Mention debugging builds in the API documentation. I've changed Skip's patch to point to Misc/SpecialBuilds and fiddled with the markup a bit. ........ r50949 | neal.norwitz | 2006-07-29 21:29:35 +0200 (Sat, 29 Jul 2006) | 6 lines Disable these tests until they are reliable across platforms. These problems may mask more important, real problems. One or both methods are known to fail on: Solaris, OpenBSD, Debian, Ubuntu. They pass on Windows and some Linux boxes. ........ r50950 | andrew.kuchling | 2006-07-29 21:50:37 +0200 (Sat, 29 Jul 2006) | 1 line [Patch #1068277] Clarify that os.path.exists() can return False depending on permissions. Fred approved committing this patch in December 2004! ........ r50952 | fred.drake | 2006-07-29 22:04:42 +0200 (Sat, 29 Jul 2006) | 6 lines SF bug #1193966: Weakref types documentation misplaced The information about supporting weakrefs with types defined in C extensions is moved to the Extending & Embedding manual. Py_TPFLAGS_HAVE_WEAKREFS is no longer mentioned since it is part of Py_TPFLAGS_DEFAULT. ........ r50953 | skip.montanaro | 2006-07-29 22:06:05 +0200 (Sat, 29 Jul 2006) | 4 lines Add a comment to the csv reader documentation that explains why the treatment of newlines changed in 2.5. Pulled almost verbatim from a comment by Andrew McNamara in <http://python.org/sf/1465014>. ........ r50954 | neal.norwitz | 2006-07-29 22:20:52 +0200 (Sat, 29 Jul 2006) | 3 lines If the executable doesn't exist, there's no reason to try to start it. This prevents garbage about command not found being printed on Solaris. ........ r50955 | fred.drake | 2006-07-29 22:21:25 +0200 (Sat, 29 Jul 2006) | 1 line fix minor markup error that introduced extra punctuation ........ r50957 | neal.norwitz | 2006-07-29 22:37:08 +0200 (Sat, 29 Jul 2006) | 3 lines Disable test_getnode too, since this is also unreliable. ........ r50958 | andrew.kuchling | 2006-07-29 23:27:12 +0200 (Sat, 29 Jul 2006) | 1 line Follow TeX's conventions for hyphens ........ r50959 | andrew.kuchling | 2006-07-29 23:30:21 +0200 (Sat, 29 Jul 2006) | 1 line Fix case for 'Unix' ........ r50960 | fred.drake | 2006-07-30 01:34:57 +0200 (Sun, 30 Jul 2006) | 1 line markup cleanups ........ r50961 | andrew.kuchling | 2006-07-30 02:27:34 +0200 (Sun, 30 Jul 2006) | 1 line Minor typo fixes ........ r50962 | andrew.kuchling | 2006-07-30 02:37:56 +0200 (Sun, 30 Jul 2006) | 1 line [Bug #793553] Correct description of keyword arguments for SSL authentication ........ r50963 | tim.peters | 2006-07-30 02:58:15 +0200 (Sun, 30 Jul 2006) | 2 lines Whitespace normalization. ........ r50964 | fred.drake | 2006-07-30 05:03:43 +0200 (Sun, 30 Jul 2006) | 1 line lots of markup nits, most commonly Unix/unix --> \UNIX ........ r50965 | fred.drake | 2006-07-30 07:41:28 +0200 (Sun, 30 Jul 2006) | 1 line update information on wxPython, from Robin Dunn ........ r50966 | fred.drake | 2006-07-30 07:49:49 +0200 (Sun, 30 Jul 2006) | 4 lines remove possibly-outdated comment on what GUI toolkit is most commonly used; it is hard to know whether this is right, and it does not add valuable reference information at any rate ........ r50967 | fred.drake | 2006-07-30 07:55:39 +0200 (Sun, 30 Jul 2006) | 3 lines - remove yet another reference to how commonly Tkinter is (thought to be) used - fix an internal section reference ........ r50968 | neal.norwitz | 2006-07-30 08:53:31 +0200 (Sun, 30 Jul 2006) | 4 lines Patch #1531113: Fix augmented assignment with yield expressions. Also fix a SystemError when trying to assign to yield expressions. ........ r50969 | neal.norwitz | 2006-07-30 08:55:48 +0200 (Sun, 30 Jul 2006) | 5 lines Add PyErr_WarnEx() so C code can pass the stacklevel to warnings.warn(). This provides the proper warning for struct.pack(). PyErr_Warn() is now deprecated in favor of PyErr_WarnEx(). As mentioned by Tim Peters on python-dev. ........ r50970 | neal.norwitz | 2006-07-30 08:57:04 +0200 (Sun, 30 Jul 2006) | 3 lines Bug #1515471: string.replace() accepts character buffers again. Pass the char* and size around rather than PyObject's. ........ r50971 | neal.norwitz | 2006-07-30 08:59:13 +0200 (Sun, 30 Jul 2006) | 1 line Whitespace normalization ........ r50973 | georg.brandl | 2006-07-30 12:53:32 +0200 (Sun, 30 Jul 2006) | 3 lines Clarify that __op__ methods must return NotImplemented if they don't support the operation. ........ r50974 | georg.brandl | 2006-07-30 13:07:23 +0200 (Sun, 30 Jul 2006) | 3 lines Bug #1002398: The documentation for os.path.sameopenfile now correctly refers to file descriptors, not file objects. ........ r50977 | martin.v.loewis | 2006-07-30 15:00:31 +0200 (Sun, 30 Jul 2006) | 3 lines Don't copy directory stat times in shutil.copytree on Windows Fixes #1525866. ........ r50978 | martin.v.loewis | 2006-07-30 15:14:05 +0200 (Sun, 30 Jul 2006) | 3 lines Base __version__ on sys.version_info, as distutils is no longer maintained separatedly. ........ r50979 | martin.v.loewis | 2006-07-30 15:27:31 +0200 (Sun, 30 Jul 2006) | 3 lines Mention Cygwin in distutils error message about a missing VS 2003. Fixes #1257728. ........ r50982 | martin.v.loewis | 2006-07-30 16:09:47 +0200 (Sun, 30 Jul 2006) | 5 lines Drop usage of test -e in configure as it is not portable. Fixes #1439538 Will backport to 2.4 Also regenerate pyconfig.h.in. ........ r50984 | georg.brandl | 2006-07-30 18:20:10 +0200 (Sun, 30 Jul 2006) | 3 lines Fix makefile changes for python-config. ........ r50985 | george.yoshida | 2006-07-30 18:37:37 +0200 (Sun, 30 Jul 2006) | 2 lines Rename struct.pack_to to struct.pack_into as changed in revision 46642. ........ r50986 | george.yoshida | 2006-07-30 18:41:30 +0200 (Sun, 30 Jul 2006) | 2 lines Typo fix ........ r50987 | neal.norwitz | 2006-07-30 21:18:13 +0200 (Sun, 30 Jul 2006) | 1 line Add some asserts and update comments ........ r50988 | neal.norwitz | 2006-07-30 21:18:38 +0200 (Sun, 30 Jul 2006) | 1 line Verify that the signal handlers were really called ........ r50989 | neal.norwitz | 2006-07-30 21:20:42 +0200 (Sun, 30 Jul 2006) | 3 lines Try to prevent hangs on Tru64/Alpha buildbot. I'm not certain this will help and may need to be reverted if it causes problems. ........ r50990 | georg.brandl | 2006-07-30 22:18:51 +0200 (Sun, 30 Jul 2006) | 2 lines Bug #1531349: right <-> left glitch in __rop__ description. ........ r50992 | tim.peters | 2006-07-31 03:46:03 +0200 (Mon, 31 Jul 2006) | 2 lines Whitespace normalization. ........ r50993 | andrew.mcnamara | 2006-07-31 04:27:48 +0200 (Mon, 31 Jul 2006) | 2 lines Redo the comment about the 2.5 change in quoted-newline handling. ........ r50994 | tim.peters | 2006-07-31 04:40:23 +0200 (Mon, 31 Jul 2006) | 10 lines ZipFile.close(): Killed one of the struct.pack deprecation warnings on Win32. Also added an XXX about the line: pos3 = self.fp.tell() `pos3` is never referenced, and I have no idea what the code intended to do instead. ........ r50996 | tim.peters | 2006-07-31 04:53:03 +0200 (Mon, 31 Jul 2006) | 8 lines ZipFile.close(): Kill the other struct.pack deprecation warning on Windows. Afraid I can't detect a pattern to when the pack formats decide to use a signed or unsigned format code -- appears nearly arbitrary to my eyes. So I left all the pack formats alone and changed the special-case data values instead. ........ r50997 | skip.montanaro | 2006-07-31 05:09:45 +0200 (Mon, 31 Jul 2006) | 1 line minor tweaks ........ r50998 | skip.montanaro | 2006-07-31 05:11:11 +0200 (Mon, 31 Jul 2006) | 1 line minor tweaks ........ r50999 | andrew.kuchling | 2006-07-31 14:20:24 +0200 (Mon, 31 Jul 2006) | 1 line Add refcounts for PyErr_WarnEx ........ r51000 | andrew.kuchling | 2006-07-31 14:39:05 +0200 (Mon, 31 Jul 2006) | 9 lines Document PyErr_WarnEx. (Bad Neal! No biscuit!) Is the explanation of the 'stacklevel' parameter clear? Please feel free to edit it. I don't have LaTeX installed on this machine, so haven't verified that the markup is correct. Will check tonight, or maybe the automatic doc build will tell me. ........ r51001 | andrew.kuchling | 2006-07-31 14:52:26 +0200 (Mon, 31 Jul 2006) | 1 line Add PyErr_WarnEx() ........ r51002 | andrew.kuchling | 2006-07-31 15:18:27 +0200 (Mon, 31 Jul 2006) | 1 line Mention csv newline changes ........ r51003 | andrew.kuchling | 2006-07-31 17:22:58 +0200 (Mon, 31 Jul 2006) | 1 line Typo fix ........ r51004 | andrew.kuchling | 2006-07-31 17:23:43 +0200 (Mon, 31 Jul 2006) | 1 line Remove reference to notation ........ r51005 | georg.brandl | 2006-07-31 18:00:34 +0200 (Mon, 31 Jul 2006) | 3 lines Fix function name. ........ r51006 | andrew.kuchling | 2006-07-31 18:10:24 +0200 (Mon, 31 Jul 2006) | 1 line [Bug #1514540] Instead of putting the standard types in a section, put them in a chapter of their own. This means string methods will now show up in the ToC. (Should the types come before or after the functions+exceptions+constants chapter? I've put them after, for now.) ........ r51007 | andrew.kuchling | 2006-07-31 18:22:05 +0200 (Mon, 31 Jul 2006) | 1 line [Bug #848556] Remove \d* from second alternative to avoid exponential case when repeating match ........ r51008 | andrew.kuchling | 2006-07-31 18:27:57 +0200 (Mon, 31 Jul 2006) | 1 line Update list of files; fix a typo ........ r51013 | andrew.kuchling | 2006-08-01 18:24:30 +0200 (Tue, 01 Aug 2006) | 1 line typo fix ........ r51018 | thomas.heller | 2006-08-01 18:54:43 +0200 (Tue, 01 Aug 2006) | 2 lines Fix a potential segfault and various potentail refcount leaks in the cast() function. ........ r51020 | thomas.heller | 2006-08-01 19:46:10 +0200 (Tue, 01 Aug 2006) | 1 line Minimal useful docstring for CopyComPointer. ........ r51021 | andrew.kuchling | 2006-08-01 20:16:15 +0200 (Tue, 01 Aug 2006) | 8 lines [Patch #1520905] Attempt to suppress core file created by test_subprocess.py. Patch by Douglas Greiman. The test_run_abort() testcase produces a core file on Unix systems, even though the test is successful. This can be confusing or alarming to someone who runs 'make test' and then finds that the Python interpreter apparently crashed. ........ r51023 | georg.brandl | 2006-08-01 20:49:24 +0200 (Tue, 01 Aug 2006) | 3 lines os.urandom no longer masks unrelated exceptions like SystemExit or KeyboardInterrupt. ........ r51025 | thomas.heller | 2006-08-01 21:14:15 +0200 (Tue, 01 Aug 2006) | 2 lines Speed up PyType_stgdict and PyObject_stgdict. ........ r51027 | ronald.oussoren | 2006-08-01 22:30:31 +0200 (Tue, 01 Aug 2006) | 3 lines Make sure the postinstall action that optionally updates the user's profile on MacOS X actually works correctly in all cases. ........ r51028 | ronald.oussoren | 2006-08-01 23:00:57 +0200 (Tue, 01 Aug 2006) | 4 lines This fixes bug #1527397: PythonLauncher runs scripts with the wrong working directory. It also fixes a bug where PythonLauncher failed to launch scripts when the scriptname (or the path to the script) contains quotes. ........ r51031 | tim.peters | 2006-08-02 05:27:46 +0200 (Wed, 02 Aug 2006) | 2 lines Whitespace normalization. ........ r51032 | tim.peters | 2006-08-02 06:12:36 +0200 (Wed, 02 Aug 2006) | 19 lines Try to squash struct.pack warnings on the "amd64 gentoo trunk" buildbot (& possibly other 64-bit boxes) during test_gzip. The native zlib crc32 function returns an unsigned 32-bit integer, which the Python wrapper implicitly casts to C long. Therefore the same crc can "look negative" on a 32-bit box but "look positive" on a 64-bit box. This patch papers over that platform difference when writing the crc to file. It may be better to change the Python wrapper, either to make the result "look positive" on all platforms (which means it may have to return a Python long at times on a 32-bit box), or to keep the sign the same across boxes. But that would be a visible change in what users see, while the current hack changes no visible behavior (well, apart from stopping the struct deprecation warning). Note that the module-level write32() function is no longer used. ........ r51033 | neal.norwitz | 2006-08-02 06:27:11 +0200 (Wed, 02 Aug 2006) | 4 lines Prevent memory leak on error. Reported by Klocwork #36 ........ r51034 | tim.peters | 2006-08-02 07:20:08 +0200 (Wed, 02 Aug 2006) | 9 lines _Stream.close(): Try to kill struct.pack() warnings when writing the crc to file on the "PPC64 Debian trunk" buildbot when running test_tarfile. This is again a case where the native zlib crc is an unsigned 32-bit int, but the Python wrapper implicitly casts it to signed C long, so that "the sign bit looks different" on different platforms. ........ r51035 | ronald.oussoren | 2006-08-02 08:10:10 +0200 (Wed, 02 Aug 2006) | 2 lines Updated documentation for the script that builds the OSX installer. ........ r51036 | neal.norwitz | 2006-08-02 08:14:22 +0200 (Wed, 02 Aug 2006) | 2 lines _PyWeakref_GetWeakrefCount() now returns a Py_ssize_t instead of long. ........ r51037 | neal.norwitz | 2006-08-02 08:15:10 +0200 (Wed, 02 Aug 2006) | 1 line v is already checked for NULL, so just DECREF it ........ r51038 | neal.norwitz | 2006-08-02 08:19:19 +0200 (Wed, 02 Aug 2006) | 1 line Let us know when there was a problem and the child had to kill the parent ........ r51039 | neal.norwitz | 2006-08-02 08:46:21 +0200 (Wed, 02 Aug 2006) | 5 lines Patch #1519025 and bug #926423: If a KeyboardInterrupt occurs during a socket operation on a socket with a timeout, the exception will be caught correctly. Previously, the exception was not caught. ........ r51040 | neal.norwitz | 2006-08-02 09:09:32 +0200 (Wed, 02 Aug 2006) | 1 line Add some explanation about Klocwork and Coverity static analysis ........ r51041 | anthony.baxter | 2006-08-02 09:43:09 +0200 (Wed, 02 Aug 2006) | 1 line pre-release machinations ........ r51043 | thomas.heller | 2006-08-02 13:35:31 +0200 (Wed, 02 Aug 2006) | 4 lines A few nore words about what ctypes does. Document that using the wrong calling convention can also raise 'ValueError: Procedure called with the wrong number of arguments'. ........ r51045 | thomas.heller | 2006-08-02 14:00:13 +0200 (Wed, 02 Aug 2006) | 1 line Fix a mistake. ........ r51046 | martin.v.loewis | 2006-08-02 15:53:55 +0200 (Wed, 02 Aug 2006) | 3 lines Correction of patch #1455898: In the mbcs decoder, set final=False for stream decoder, but final=True for the decode function. ........ r51049 | tim.peters | 2006-08-02 20:19:35 +0200 (Wed, 02 Aug 2006) | 2 lines Add missing svn:eol-style property to text files. ........ r51079 | neal.norwitz | 2006-08-04 06:50:21 +0200 (Fri, 04 Aug 2006) | 3 lines Bug #1531405, format_exception no longer raises an exception if str(exception) raised an exception. ........ r51080 | neal.norwitz | 2006-08-04 06:58:47 +0200 (Fri, 04 Aug 2006) | 11 lines Bug #1191458: tracing over for loops now produces a line event on each iteration. I'm not positive this is the best way to handle this. I'm also not sure that there aren't other cases where the lnotab is generated incorrectly. It would be great if people that use pdb or tracing could test heavily. Also: * Remove dead/duplicated code that wasn't used/necessary because we already handled the docstring prior to entering the loop. * add some debugging code into the compiler (#if 0'd out). ........ r51081 | neal.norwitz | 2006-08-04 07:09:28 +0200 (Fri, 04 Aug 2006) | 4 lines Bug #1333982: string/number constants were inappropriately stored in the byte code and co_consts even if they were not used, ie immediately popped off the stack. ........ r51082 | neal.norwitz | 2006-08-04 07:12:19 +0200 (Fri, 04 Aug 2006) | 1 line There were really two issues ........ r51084 | fred.drake | 2006-08-04 07:17:21 +0200 (Fri, 04 Aug 2006) | 1 line SF patch #1534048 (bug #1531003): fix typo in error message ........ r51085 | gregory.p.smith | 2006-08-04 07:17:47 +0200 (Fri, 04 Aug 2006) | 3 lines fix typos ........ r51087 | georg.brandl | 2006-08-04 08:03:53 +0200 (Fri, 04 Aug 2006) | 3 lines Fix bug caused by first decrefing, then increfing. ........ r51109 | neil.schemenauer | 2006-08-04 18:20:30 +0200 (Fri, 04 Aug 2006) | 5 lines Fix the 'compiler' package to generate correct code for MAKE_CLOSURE. In the 2.5 development cycle, MAKE_CLOSURE as changed to take free variables as a tuple rather than as individual items on the stack. Closes patch #1534084. ........ r51110 | georg.brandl | 2006-08-04 20:03:37 +0200 (Fri, 04 Aug 2006) | 3 lines Change fix for segfaulting property(), add a NEWS entry and a test. ........ r51111 | georg.brandl | 2006-08-04 20:07:34 +0200 (Fri, 04 Aug 2006) | 3 lines Better fix for bug #1531405, not executing str(value) twice. ........ r51112 | thomas.heller | 2006-08-04 20:17:40 +0200 (Fri, 04 Aug 2006) | 1 line On Windows, make PyErr_Warn an exported function again. ........ r51113 | thomas.heller | 2006-08-04 20:57:34 +0200 (Fri, 04 Aug 2006) | 4 lines Fix #1530448 - fix ctypes build failure on solaris 10. The '-mimpure-text' linker flag is required when linking _ctypes.so. ........ r51114 | thomas.heller | 2006-08-04 21:49:31 +0200 (Fri, 04 Aug 2006) | 3 lines Fix #1534738: win32 debug version of _msi must be _msi_d.pyd, not _msi.pyd. Fix the name of the pdb file as well. ........ r51115 | andrew.kuchling | 2006-08-04 22:37:43 +0200 (Fri, 04 Aug 2006) | 1 line Typo fixes ........ r51116 | andrew.kuchling | 2006-08-04 23:10:03 +0200 (Fri, 04 Aug 2006) | 1 line Fix mangled sentence ........ r51118 | tim.peters | 2006-08-05 00:00:35 +0200 (Sat, 05 Aug 2006) | 2 lines Whitespace normalization. ........ r51119 | bob.ippolito | 2006-08-05 01:59:21 +0200 (Sat, 05 Aug 2006) | 5 lines Fix #1530559, struct.pack raises TypeError where it used to convert. Passing float arguments to struct.pack when integers are expected now triggers a DeprecationWarning. ........ r51123 | georg.brandl | 2006-08-05 08:10:54 +0200 (Sat, 05 Aug 2006) | 3 lines Patch #1534922: correct and enhance unittest docs. ........ r51126 | georg.brandl | 2006-08-06 09:06:33 +0200 (Sun, 06 Aug 2006) | 2 lines Bug #1535182: really test the xreadlines() method of bz2 objects. ........ r51128 | georg.brandl | 2006-08-06 09:26:21 +0200 (Sun, 06 Aug 2006) | 4 lines Bug #1535081: A leading underscore has been added to the names of the md5 and sha modules, so add it in Modules/Setup.dist too. ........ r51129 | georg.brandl | 2006-08-06 10:23:54 +0200 (Sun, 06 Aug 2006) | 3 lines Bug #1535165: fixed a segfault in input() and raw_input() when sys.stdin is closed. ........ r51131 | georg.brandl | 2006-08-06 11:17:16 +0200 (Sun, 06 Aug 2006) | 2 lines Don't produce output in test_builtin. ........ r51133 | andrew.macintyre | 2006-08-06 14:37:03 +0200 (Sun, 06 Aug 2006) | 4 lines test_threading now skips testing alternate thread stack sizes on platforms that don't support changing thread stack size. ........ r51134 | andrew.kuchling | 2006-08-07 00:07:04 +0200 (Mon, 07 Aug 2006) | 2 lines [Patch #1464056] Ensure that we use the panelw library when linking with ncursesw. Once I see how the buildbots react, I'll backport this to 2.4. ........ r51137 | georg.brandl | 2006-08-08 13:52:34 +0200 (Tue, 08 Aug 2006) | 3 lines webbrowser: Silence stderr output if no gconftool or gnome browser found ........ r51138 | georg.brandl | 2006-08-08 13:56:21 +0200 (Tue, 08 Aug 2006) | 7 lines Remove "non-mapping" and "non-sequence" from TypeErrors raised by PyMapping_Size and PySequence_Size. Because len() tries first sequence, then mapping size, it will always raise a "non-mapping object has no len" error which is confusing. ........ r51139 | thomas.heller | 2006-08-08 19:37:00 +0200 (Tue, 08 Aug 2006) | 3 lines memcmp() can return values other than -1, 0, and +1 but tp_compare must not. ........ r51140 | thomas.heller | 2006-08-08 19:39:20 +0200 (Tue, 08 Aug 2006) | 1 line Remove accidently committed, duplicated test. ........ r51147 | andrew.kuchling | 2006-08-08 20:50:14 +0200 (Tue, 08 Aug 2006) | 1 line Reword paragraph to clarify ........ r51148 | andrew.kuchling | 2006-08-08 20:56:08 +0200 (Tue, 08 Aug 2006) | 1 line Move obmalloc item into C API section ........ r51149 | andrew.kuchling | 2006-08-08 21:00:14 +0200 (Tue, 08 Aug 2006) | 1 line 'Other changes' section now has only one item; move the item elsewhere and remove the section ........ r51150 | andrew.kuchling | 2006-08-08 21:00:34 +0200 (Tue, 08 Aug 2006) | 1 line Bump version number ........ r51151 | georg.brandl | 2006-08-08 22:11:22 +0200 (Tue, 08 Aug 2006) | 2 lines Bug #1536828: typo: TypeType should have been StringType. ........ r51153 | georg.brandl | 2006-08-08 22:13:13 +0200 (Tue, 08 Aug 2006) | 2 lines Bug #1536660: separate two words. ........ r51155 | georg.brandl | 2006-08-08 22:48:10 +0200 (Tue, 08 Aug 2006) | 3 lines ``str`` is now the same object as ``types.StringType``. ........ r51156 | tim.peters | 2006-08-09 02:52:26 +0200 (Wed, 09 Aug 2006) | 2 lines Whitespace normalization. ........ r51158 | georg.brandl | 2006-08-09 09:03:22 +0200 (Wed, 09 Aug 2006) | 4 lines Introduce an upper bound on tuple nesting depth in C argument format strings; fixes rest of #1523610. ........ r51160 | martin.v.loewis | 2006-08-09 09:57:39 +0200 (Wed, 09 Aug 2006) | 4 lines __hash__ may now return long int; the final hash value is obtained by invoking hash on the long int. Fixes #1536021. ........ r51168 | andrew.kuchling | 2006-08-09 15:03:41 +0200 (Wed, 09 Aug 2006) | 1 line [Bug #1536021] Mention __hash__ change ........ r51169 | andrew.kuchling | 2006-08-09 15:57:05 +0200 (Wed, 09 Aug 2006) | 1 line [Patch #1534027] Add notes on locale module changes ........ r51170 | andrew.kuchling | 2006-08-09 16:05:35 +0200 (Wed, 09 Aug 2006) | 1 line Add missing 'self' parameters ........ r51171 | andrew.kuchling | 2006-08-09 16:06:19 +0200 (Wed, 09 Aug 2006) | 1 line Reindent code ........ r51172 | armin.rigo | 2006-08-09 16:55:26 +0200 (Wed, 09 Aug 2006) | 2 lines Fix and test for an infinite C recursion. ........ r51173 | ronald.oussoren | 2006-08-09 16:56:33 +0200 (Wed, 09 Aug 2006) | 2 lines It's unlikely that future versions will require _POSIX_C_SOURCE ........ r51178 | armin.rigo | 2006-08-09 17:37:26 +0200 (Wed, 09 Aug 2006) | 2 lines Concatenation on a long string breaks (SF #1526585). ........ r51180 | kurt.kaiser | 2006-08-09 18:46:15 +0200 (Wed, 09 Aug 2006) | 8 lines 1. When used w/o subprocess, all exceptions were preceeded by an error message claiming they were IDLE internal errors (since 1.2a1). 2. Add Ronald Oussoren to CREDITS M NEWS.txt M PyShell.py M CREDITS.txt ........ r51181 | kurt.kaiser | 2006-08-09 19:47:15 +0200 (Wed, 09 Aug 2006) | 4 lines As a slight enhancement to the previous checkin, improve the internal error reporting by moving message to IDLE console. ........ r51182 | andrew.kuchling | 2006-08-09 20:23:14 +0200 (Wed, 09 Aug 2006) | 1 line Typo fix ........ r51183 | kurt.kaiser | 2006-08-09 22:34:46 +0200 (Wed, 09 Aug 2006) | 2 lines ToggleTab dialog was setting indent to 8 even if cancelled (since 1.2a1). ........ r51184 | martin.v.loewis | 2006-08-10 01:42:18 +0200 (Thu, 10 Aug 2006) | 2 lines Add some commentary on -mimpure-text. ........ r51185 | tim.peters | 2006-08-10 02:58:49 +0200 (Thu, 10 Aug 2006) | 2 lines Add missing svn:eol-style property to text files. ........ r51186 | kurt.kaiser | 2006-08-10 03:41:17 +0200 (Thu, 10 Aug 2006) | 2 lines Changing tokenize (39046) to detect dedent broke tabnanny check (since 1.2a1) ........ r51187 | tim.peters | 2006-08-10 05:01:26 +0200 (Thu, 10 Aug 2006) | 13 lines test_copytree_simple(): This was leaving behind two new temp directories each time it ran, at least on Windows. Several changes: explicitly closed all files; wrapped long lines; stopped suppressing errors when removing a file or directory fails (removing /shouldn't/ fail!); and changed what appeared to be incorrect usage of os.removedirs() (that doesn't remove empty directories at and /under/ the given path, instead it must be given an empty leaf directory and then deletes empty directories moving /up/ the path -- could be that the conceptually simpler shutil.rmtree() was really actually intended here). ........
2006-08-11 11:57:12 -03:00
static PyObject *
thread_stack_size(PyObject *self, PyObject *args)
{
size_t old_size;
Py_ssize_t new_size = 0;
int rc;
if (!PyArg_ParseTuple(args, "|n:stack_size", &new_size))
return NULL;
if (new_size < 0) {
PyErr_SetString(PyExc_ValueError,
"size must be 0 or a positive value");
return NULL;
}
old_size = PyThread_get_stacksize();
rc = PyThread_set_stacksize((size_t) new_size);
if (rc == -1) {
PyErr_Format(PyExc_ValueError,
"size not valid: %zd bytes",
new_size);
return NULL;
}
if (rc == -2) {
PyErr_SetString(ThreadError,
"setting stack size not supported");
return NULL;
}
return PyLong_FromSsize_t((Py_ssize_t) old_size);
Merged revisions 46753-51188 via svnmerge from svn+ssh://pythondev@svn.python.org/python/trunk ........ r46755 | brett.cannon | 2006-06-08 18:23:04 +0200 (Thu, 08 Jun 2006) | 4 lines Make binascii.hexlify() use s# for its arguments instead of t# to actually match its documentation stating it accepts any read-only buffer. ........ r46757 | brett.cannon | 2006-06-08 19:00:45 +0200 (Thu, 08 Jun 2006) | 8 lines Buffer objects would return the read or write buffer for a wrapped object when the char buffer was requested. Now it actually returns the char buffer if available or raises a TypeError if it isn't (as is raised for the other buffer types if they are not present but requested). Not a backport candidate since it does change semantics of the buffer object (although it could be argued this is enough of a bug to bother backporting). ........ r46760 | andrew.kuchling | 2006-06-09 03:10:17 +0200 (Fri, 09 Jun 2006) | 1 line Update functools section ........ r46762 | tim.peters | 2006-06-09 04:11:02 +0200 (Fri, 09 Jun 2006) | 6 lines Whitespace normalization. Since test_file is implicated in mysterious test failures when followed by test_optparse, if I had any brains I'd look at the checkin that last changed test_file ;-) ........ r46763 | tim.peters | 2006-06-09 05:09:42 +0200 (Fri, 09 Jun 2006) | 5 lines To boost morale :-), force test_optparse to run immediately after test_file until we can figure out how to fix it. (See python-dev; at the moment we don't even know which checkin caused the problem.) ........ r46764 | tim.peters | 2006-06-09 05:51:41 +0200 (Fri, 09 Jun 2006) | 6 lines AutoFileTests.tearDown(): Removed mysterious undocumented try/except. Remove TESTFN. Throughout: used open() instead of file(), and wrapped long lines. ........ r46765 | tim.peters | 2006-06-09 06:02:06 +0200 (Fri, 09 Jun 2006) | 8 lines testUnicodeOpen(): I have no idea why, but making this test clean up after itself appears to fix the test failures when test_optparse follows test_file. test_main(): Get rid of TESTFN no matter what. That's also enough to fix the mystery failures. Doesn't hurt to fix them twice :-) ........ r46766 | tim.peters | 2006-06-09 07:12:40 +0200 (Fri, 09 Jun 2006) | 6 lines Remove the temporary hack to force test_optparse to run immediately after test_file. At least 8 buildbot boxes passed since the underlying problem got fixed, and they all failed before the fix, so there's no point to this anymore. ........ r46767 | neal.norwitz | 2006-06-09 07:54:18 +0200 (Fri, 09 Jun 2006) | 1 line Fix grammar and reflow ........ r46769 | andrew.kuchling | 2006-06-09 12:22:35 +0200 (Fri, 09 Jun 2006) | 1 line Markup fix ........ r46773 | andrew.kuchling | 2006-06-09 15:15:57 +0200 (Fri, 09 Jun 2006) | 1 line [Bug #1472827] Make saxutils.XMLGenerator handle \r\n\t in attribute values by escaping them properly. 2.4 bugfix candidate. ........ r46778 | kristjan.jonsson | 2006-06-09 18:28:01 +0200 (Fri, 09 Jun 2006) | 2 lines Turn off warning about deprecated CRT functions on for VisualStudio .NET 2005. Make the definition #ARRAYSIZE conditional. VisualStudio .NET 2005 already has it defined using a better gimmick. ........ r46779 | phillip.eby | 2006-06-09 18:40:18 +0200 (Fri, 09 Jun 2006) | 2 lines Import wsgiref into the stdlib, as of the external version 0.1-r2181. ........ r46783 | andrew.kuchling | 2006-06-09 18:44:40 +0200 (Fri, 09 Jun 2006) | 1 line Add note about XMLGenerator bugfix ........ r46784 | andrew.kuchling | 2006-06-09 18:46:51 +0200 (Fri, 09 Jun 2006) | 1 line Add note about wsgiref ........ r46785 | brett.cannon | 2006-06-09 19:05:48 +0200 (Fri, 09 Jun 2006) | 2 lines Fix inconsistency in naming within an enum. ........ r46787 | tim.peters | 2006-06-09 19:47:00 +0200 (Fri, 09 Jun 2006) | 2 lines Whitespace normalization. ........ r46792 | georg.brandl | 2006-06-09 20:29:52 +0200 (Fri, 09 Jun 2006) | 3 lines Test file.__exit__. ........ r46794 | brett.cannon | 2006-06-09 20:40:46 +0200 (Fri, 09 Jun 2006) | 2 lines svn:ignore .pyc and .pyo files. ........ r46795 | georg.brandl | 2006-06-09 20:45:48 +0200 (Fri, 09 Jun 2006) | 3 lines RFE #1491485: str/unicode.endswith()/startswith() now accept a tuple as first argument. ........ r46798 | andrew.kuchling | 2006-06-09 21:03:16 +0200 (Fri, 09 Jun 2006) | 1 line Describe startswith()/endswiith() change; add reminder about wsgiref ........ r46799 | tim.peters | 2006-06-09 21:24:44 +0200 (Fri, 09 Jun 2006) | 11 lines Implementing a happy idea from Georg Brandl: make runtest() try to clean up files and directories the tests often leave behind by mistake. This is the first time in history I don't have a bogus "db_home" directory after running the tests ;-) Also worked on runtest's docstring, to say something about all the arguments, and to document the non-obvious return values. New functions runtest_inner() and cleanup_test_droppings() in support of the above. ........ r46800 | andrew.kuchling | 2006-06-09 21:43:25 +0200 (Fri, 09 Jun 2006) | 1 line Remove unused variable ........ r46801 | andrew.kuchling | 2006-06-09 21:56:05 +0200 (Fri, 09 Jun 2006) | 1 line Add some wsgiref text ........ r46803 | thomas.heller | 2006-06-09 21:59:11 +0200 (Fri, 09 Jun 2006) | 1 line set eol-style svn property ........ r46804 | thomas.heller | 2006-06-09 22:01:01 +0200 (Fri, 09 Jun 2006) | 1 line set eol-style svn property ........ r46805 | georg.brandl | 2006-06-09 22:43:48 +0200 (Fri, 09 Jun 2006) | 3 lines Make use of new str.startswith/endswith semantics. Occurences in email and compiler were ignored due to backwards compat requirements. ........ r46806 | brett.cannon | 2006-06-10 00:31:23 +0200 (Sat, 10 Jun 2006) | 4 lines An object with __call__ as an attribute, when called, will have that attribute checked for __call__ itself, and will continue to look until it finds an object without the attribute. This can lead to an infinite recursion. Closes bug #532646, again. Will be backported. ........ r46808 | brett.cannon | 2006-06-10 00:45:54 +0200 (Sat, 10 Jun 2006) | 2 lines Fix bug introduced in rev. 46806 by not having variable declaration at the top of a block. ........ r46812 | georg.brandl | 2006-06-10 08:40:50 +0200 (Sat, 10 Jun 2006) | 4 lines Apply perky's fix for #1503157: "/".join([u"", u""]) raising OverflowError. Also improve error message on overflow. ........ r46817 | martin.v.loewis | 2006-06-10 10:14:03 +0200 (Sat, 10 Jun 2006) | 2 lines Port cygwin kill_python changes from 2.4 branch. ........ r46818 | armin.rigo | 2006-06-10 12:57:40 +0200 (Sat, 10 Jun 2006) | 4 lines SF bug #1503294. PyThreadState_GET() complains if the tstate is NULL, but only in debug mode. ........ r46819 | martin.v.loewis | 2006-06-10 14:23:46 +0200 (Sat, 10 Jun 2006) | 4 lines Patch #1495999: Part two of Windows CE changes. - update header checks, using autoconf - provide dummies for getenv, environ, and GetVersion - adjust MSC_VER check in socketmodule.c ........ r46820 | skip.montanaro | 2006-06-10 16:09:11 +0200 (Sat, 10 Jun 2006) | 1 line document the class, not its initializer ........ r46821 | greg.ward | 2006-06-10 18:40:01 +0200 (Sat, 10 Jun 2006) | 4 lines Sync with Optik docs (rev 518): * restore "Extending optparse" section * document ALWAYS_TYPED_ACTIONS (SF #1449311) ........ r46824 | thomas.heller | 2006-06-10 21:51:46 +0200 (Sat, 10 Jun 2006) | 8 lines Upgrade to ctypes version 0.9.9.7. Summary of changes: - support for 'variable sized' data - support for anonymous structure/union fields - fix severe bug with certain arrays or structures containing more than 256 fields ........ r46825 | thomas.heller | 2006-06-10 21:55:36 +0200 (Sat, 10 Jun 2006) | 8 lines Upgrade to ctypes version 0.9.9.7. Summary of changes: - support for 'variable sized' data - support for anonymous structure/union fields - fix severe bug with certain arrays or structures containing more than 256 fields ........ r46826 | fred.drake | 2006-06-10 22:01:34 +0200 (Sat, 10 Jun 2006) | 4 lines SF patch #1303595: improve description of __builtins__, explaining how it varies between __main__ and other modules, and strongly suggest not touching it but using __builtin__ if absolutely necessary ........ r46827 | fred.drake | 2006-06-10 22:02:58 +0200 (Sat, 10 Jun 2006) | 1 line credit for SF patch #1303595 ........ r46831 | thomas.heller | 2006-06-10 22:29:34 +0200 (Sat, 10 Jun 2006) | 2 lines New docs for ctypes. ........ r46834 | thomas.heller | 2006-06-10 23:07:19 +0200 (Sat, 10 Jun 2006) | 1 line Fix a wrong printf format. ........ r46835 | thomas.heller | 2006-06-10 23:17:58 +0200 (Sat, 10 Jun 2006) | 1 line Fix the second occurrence of the problematic printf format. ........ r46837 | thomas.heller | 2006-06-10 23:56:03 +0200 (Sat, 10 Jun 2006) | 1 line Don't use C++ comment. ........ r46838 | thomas.heller | 2006-06-11 00:01:50 +0200 (Sun, 11 Jun 2006) | 1 line Handle failure of PyMem_Realloc. ........ r46839 | skip.montanaro | 2006-06-11 00:38:13 +0200 (Sun, 11 Jun 2006) | 2 lines Suppress warning on MacOSX about possible use before set of proc. ........ r46840 | tim.peters | 2006-06-11 00:51:45 +0200 (Sun, 11 Jun 2006) | 8 lines shuffle() doscstring: Removed warning about sequence length versus generator period. While this was a real weakness of the older WH generator for lists with just a few dozen elements, and so could potentially bite the naive ;-), the Twister should show excellent behavior up to at least 600 elements. Module docstring: reflowed some jarringly short lines. ........ r46844 | greg.ward | 2006-06-11 02:40:49 +0200 (Sun, 11 Jun 2006) | 4 lines Bug #1361643: fix textwrap.dedent() so it handles tabs appropriately, i.e. do *not* expand tabs, but treat them as whitespace that is not equivalent to spaces. Add a couple of test cases. Clarify docs. ........ r46850 | neal.norwitz | 2006-06-11 07:44:18 +0200 (Sun, 11 Jun 2006) | 5 lines Fix Coverity # 146. newDBSequenceObject would deref dbobj, so it can't be NULL. We know it's not NULL from the ParseTuple and DbObject_Check will verify it's not NULL. ........ r46851 | neal.norwitz | 2006-06-11 07:45:25 +0200 (Sun, 11 Jun 2006) | 4 lines Wrap some long lines Top/Bottom factor out some common expressions Add a XXX comment about widing offset. ........ r46852 | neal.norwitz | 2006-06-11 07:45:47 +0200 (Sun, 11 Jun 2006) | 1 line Add versionadded to doc ........ r46853 | neal.norwitz | 2006-06-11 07:47:14 +0200 (Sun, 11 Jun 2006) | 3 lines Update doc to make it agree with code. Bottom factor out some common code. ........ r46854 | neal.norwitz | 2006-06-11 07:48:14 +0200 (Sun, 11 Jun 2006) | 3 lines f_code can't be NULL based on Frame_New and other code that derefs it. So there doesn't seem to be much point to checking here. ........ r46855 | neal.norwitz | 2006-06-11 09:26:27 +0200 (Sun, 11 Jun 2006) | 1 line Fix errors found by pychecker ........ r46856 | neal.norwitz | 2006-06-11 09:26:50 +0200 (Sun, 11 Jun 2006) | 1 line warnings was imported at module scope, no need to import again ........ r46857 | neal.norwitz | 2006-06-11 09:27:56 +0200 (Sun, 11 Jun 2006) | 5 lines Fix errors found by pychecker. I think these changes are correct, but I'm not sure. Could someone who knows how this module works test it? It can at least start on the cmd line. ........ r46858 | neal.norwitz | 2006-06-11 10:35:14 +0200 (Sun, 11 Jun 2006) | 1 line Fix errors found by pychecker ........ r46859 | ronald.oussoren | 2006-06-11 16:33:36 +0200 (Sun, 11 Jun 2006) | 4 lines This patch improves the L&F of IDLE on OSX. The changes are conditionalized on being in an IDLE.app bundle on darwin. This does a slight reorganisation of the menus and adds support for file-open events. ........ r46860 | greg.ward | 2006-06-11 16:42:41 +0200 (Sun, 11 Jun 2006) | 1 line SF #1366250: optparse docs: fix inconsistency in variable name; minor tweaks. ........ r46861 | greg.ward | 2006-06-11 18:24:11 +0200 (Sun, 11 Jun 2006) | 3 lines Bug #1498146: fix optparse to handle Unicode strings in option help, description, and epilog. ........ r46862 | thomas.heller | 2006-06-11 19:04:22 +0200 (Sun, 11 Jun 2006) | 2 lines Release the GIL during COM method calls, to avoid deadlocks in Python coded COM objects. ........ r46863 | tim.peters | 2006-06-11 21:42:51 +0200 (Sun, 11 Jun 2006) | 2 lines Whitespace normalization. ........ r46864 | tim.peters | 2006-06-11 21:43:49 +0200 (Sun, 11 Jun 2006) | 2 lines Add missing svn:eol-style property to text files. ........ r46865 | ronald.oussoren | 2006-06-11 21:45:57 +0200 (Sun, 11 Jun 2006) | 2 lines Remove message about using make frameworkinstall, that's no longer necesssary ........ r46866 | ronald.oussoren | 2006-06-11 22:23:29 +0200 (Sun, 11 Jun 2006) | 2 lines Use configure to substitute the correct prefix instead of hardcoding ........ r46867 | ronald.oussoren | 2006-06-11 22:24:45 +0200 (Sun, 11 Jun 2006) | 4 lines - Change fixapplepython23.py to ensure that it will run with /usr/bin/python on intel macs. - Fix some minor problems in the installer for OSX ........ r46868 | neal.norwitz | 2006-06-11 22:25:56 +0200 (Sun, 11 Jun 2006) | 5 lines Try to fix several networking tests. The problem is that if hosts have a search path setup, some of these hosts resolve to the wrong address. By appending a period to the hostname, the hostname should only resolve to what we want it to resolve to. Hopefully this doesn't break different bots. ........ r46869 | neal.norwitz | 2006-06-11 22:42:02 +0200 (Sun, 11 Jun 2006) | 7 lines Try to fix another networking test. The problem is that if hosts have a search path setup, some of these hosts resolve to the wrong address. By appending a period to the hostname, the hostname should only resolve to what we want it to resolve to. Hopefully this doesn't break different bots. Also add more info to failure message to aid debugging test failure. ........ r46870 | neal.norwitz | 2006-06-11 22:46:46 +0200 (Sun, 11 Jun 2006) | 4 lines Fix test on PPC64 buildbot. It raised an IOError (really an URLError which derives from an IOError). That seems valid. Env Error includes both OSError and IOError, so this seems like a reasonable fix. ........ r46871 | tim.peters | 2006-06-11 22:52:59 +0200 (Sun, 11 Jun 2006) | 10 lines compare_generic_iter(): Fixed the failure of test_wsgiref's testFileWrapper when running with -O. test_simple_validation_error still fails under -O. That appears to be because wsgiref's validate.py uses `assert` statements all over the place to check arguments for sanity. That should all be changed (it's not a logical error in the software if a user passes bogus arguments, so this isn't a reasonable use for `assert` -- checking external preconditions should generally raise ValueError or TypeError instead, as appropriate). ........ r46872 | neal.norwitz | 2006-06-11 23:38:38 +0200 (Sun, 11 Jun 2006) | 1 line Get test to pass on S/390. Shout if you think this change is incorrect. ........ r46873 | neal.norwitz | 2006-06-12 04:05:55 +0200 (Mon, 12 Jun 2006) | 1 line Cleanup Py_ssize_t a little (get rid of second #ifdef) ........ r46874 | neal.norwitz | 2006-06-12 04:06:17 +0200 (Mon, 12 Jun 2006) | 1 line Fix some Py_ssize_t issues ........ r46875 | neal.norwitz | 2006-06-12 04:06:42 +0200 (Mon, 12 Jun 2006) | 1 line Fix some Py_ssize_t issues ........ r46876 | neal.norwitz | 2006-06-12 04:07:24 +0200 (Mon, 12 Jun 2006) | 2 lines Cleanup: Remove import of types to get StringTypes, we can just use basestring. ........ r46877 | neal.norwitz | 2006-06-12 04:07:57 +0200 (Mon, 12 Jun 2006) | 1 line Don't truncate if size_t is bigger than uint ........ r46878 | neal.norwitz | 2006-06-12 04:08:41 +0200 (Mon, 12 Jun 2006) | 1 line Don't leak the list object if there's an error allocating the item storage. Backport candidate ........ r46879 | neal.norwitz | 2006-06-12 04:09:03 +0200 (Mon, 12 Jun 2006) | 1 line Fix typo. Backport if anyone cares. :-) ........ r46880 | neal.norwitz | 2006-06-12 04:09:34 +0200 (Mon, 12 Jun 2006) | 1 line Fix indentation of case and a Py_ssize_t issue. ........ r46881 | neal.norwitz | 2006-06-12 04:11:18 +0200 (Mon, 12 Jun 2006) | 3 lines Get rid of f_restricted too. Doc the other 4 ints that were already removed at the NeedForSpeed sprint. ........ r46882 | neal.norwitz | 2006-06-12 04:13:21 +0200 (Mon, 12 Jun 2006) | 1 line Fix the socket tests so they can be run concurrently. Backport candidate ........ r46883 | neal.norwitz | 2006-06-12 04:16:10 +0200 (Mon, 12 Jun 2006) | 1 line i and j are initialized below when used. No need to do it twice ........ r46884 | neal.norwitz | 2006-06-12 05:05:03 +0200 (Mon, 12 Jun 2006) | 1 line Remove unused import ........ r46885 | neal.norwitz | 2006-06-12 05:05:40 +0200 (Mon, 12 Jun 2006) | 1 line Impl ssize_t ........ r46886 | neal.norwitz | 2006-06-12 05:33:09 +0200 (Mon, 12 Jun 2006) | 6 lines Patch #1503046, Conditional compilation of zlib.(de)compressobj.copy copy is only in newer versions of zlib. This should allow zlibmodule to work with older versions like the Tru64 buildbot. ........ r46887 | phillip.eby | 2006-06-12 06:04:32 +0200 (Mon, 12 Jun 2006) | 2 lines Sync w/external release 0.1.2. Please see PEP 360 before making changes to external packages. ........ r46888 | martin.v.loewis | 2006-06-12 06:26:31 +0200 (Mon, 12 Jun 2006) | 2 lines Get rid of function pointer cast. ........ r46889 | thomas.heller | 2006-06-12 08:05:57 +0200 (Mon, 12 Jun 2006) | 3 lines I don't know how that happend, but the entire file contents was duplicated. Thanks to Simon Percivall for the heads up. ........ r46890 | nick.coghlan | 2006-06-12 10:19:37 +0200 (Mon, 12 Jun 2006) | 1 line Fix site module docstring to match the code ........ r46891 | nick.coghlan | 2006-06-12 10:23:02 +0200 (Mon, 12 Jun 2006) | 1 line Fix site module docstring to match the code for Mac OSX, too ........ r46892 | nick.coghlan | 2006-06-12 10:27:13 +0200 (Mon, 12 Jun 2006) | 1 line The site module documentation also described the Windows behaviour incorrectly. ........ r46893 | nick.coghlan | 2006-06-12 12:17:11 +0200 (Mon, 12 Jun 2006) | 1 line Make the -m switch conform to the documentation of sys.path by behaving like the -c switch ........ r46894 | kristjan.jonsson | 2006-06-12 17:45:12 +0200 (Mon, 12 Jun 2006) | 2 lines Fix the CRT argument error handling for VisualStudio .NET 2005. Install a CRT error handler and disable the assertion for debug builds. This causes CRT to set errno to EINVAL. This update fixes crash cases in the test suite where the default CRT error handler would cause process exit. ........ r46899 | thomas.heller | 2006-06-12 22:56:48 +0200 (Mon, 12 Jun 2006) | 1 line Add pep-291 compatibility markers. ........ r46901 | ka-ping.yee | 2006-06-13 01:47:52 +0200 (Tue, 13 Jun 2006) | 5 lines Add the uuid module. This module has been tested so far on Windows XP (Python 2.4 and 2.5a2), Mac OS X (Python 2.3, 2.4, and 2.5a2), and Linux (Python 2.4 and 2.5a2). ........ r46902 | tim.peters | 2006-06-13 02:30:01 +0200 (Tue, 13 Jun 2006) | 2 lines Whitespace normalization. ........ r46903 | tim.peters | 2006-06-13 02:30:50 +0200 (Tue, 13 Jun 2006) | 2 lines Added missing svn:eol-style property to text files. ........ r46905 | tim.peters | 2006-06-13 05:30:07 +0200 (Tue, 13 Jun 2006) | 5 lines get_matching_blocks(): rewrote code & comments so they match; added more comments about why it's this way at all; and removed what looked like needless expense (sorting (i, j, k) triples directly should give exactly the same order as sorting (i, (i, j, k)) pairs). ........ r46906 | neal.norwitz | 2006-06-13 06:08:53 +0200 (Tue, 13 Jun 2006) | 1 line Don't fail if another process is listening on our port. ........ r46908 | neal.norwitz | 2006-06-13 10:28:19 +0200 (Tue, 13 Jun 2006) | 2 lines Initialize the type object so pychecker can't crash the interpreter. ........ r46909 | neal.norwitz | 2006-06-13 10:41:06 +0200 (Tue, 13 Jun 2006) | 1 line Verify the crash due to EncodingMap not initialized does not return ........ r46910 | thomas.heller | 2006-06-13 10:56:14 +0200 (Tue, 13 Jun 2006) | 3 lines Add some windows datatypes that were missing from this file, and add the aliases defined in windows header files for the structures. ........ r46911 | thomas.heller | 2006-06-13 11:40:14 +0200 (Tue, 13 Jun 2006) | 3 lines Add back WCHAR, UINT, DOUBLE, _LARGE_INTEGER, _ULARGE_INTEGER. VARIANT_BOOL is a special _ctypes data type, not c_short. ........ r46912 | ronald.oussoren | 2006-06-13 13:19:56 +0200 (Tue, 13 Jun 2006) | 4 lines Linecache contains support for PEP302 loaders, but fails to deal with loaders that return None to indicate that the module is valid but no source is available. This patch fixes that. ........ r46913 | andrew.kuchling | 2006-06-13 13:57:04 +0200 (Tue, 13 Jun 2006) | 1 line Mention uuid module ........ r46915 | walter.doerwald | 2006-06-13 14:02:12 +0200 (Tue, 13 Jun 2006) | 2 lines Fix passing errors to the encoder and decoder functions. ........ r46917 | walter.doerwald | 2006-06-13 14:04:43 +0200 (Tue, 13 Jun 2006) | 3 lines errors is an attribute in the incremental decoder not an argument. ........ r46919 | andrew.macintyre | 2006-06-13 17:04:24 +0200 (Tue, 13 Jun 2006) | 11 lines Patch #1454481: Make thread stack size runtime tunable. Heavily revised, comprising revisions: 46640 - original trunk revision (backed out in r46655) 46647 - markup fix (backed out in r46655) 46692:46918 merged from branch aimacintyre-sf1454481 branch tested on buildbots (Windows buildbots had problems not related to these changes). ........ r46920 | brett.cannon | 2006-06-13 18:06:55 +0200 (Tue, 13 Jun 2006) | 2 lines Remove unused variable. ........ r46921 | andrew.kuchling | 2006-06-13 18:41:41 +0200 (Tue, 13 Jun 2006) | 1 line Add ability to set stack size ........ r46923 | marc-andre.lemburg | 2006-06-13 19:04:26 +0200 (Tue, 13 Jun 2006) | 2 lines Update pybench to version 2.0. ........ r46924 | marc-andre.lemburg | 2006-06-13 19:07:14 +0200 (Tue, 13 Jun 2006) | 2 lines Revert wrong svn copy. ........ r46925 | andrew.macintyre | 2006-06-13 19:14:36 +0200 (Tue, 13 Jun 2006) | 2 lines fix exception usage ........ r46927 | tim.peters | 2006-06-13 20:37:07 +0200 (Tue, 13 Jun 2006) | 2 lines Whitespace normalization. ........ r46928 | marc-andre.lemburg | 2006-06-13 20:56:56 +0200 (Tue, 13 Jun 2006) | 9 lines Updated to pybench 2.0. See svn.python.org/external/pybench-2.0 for the original import of that version. Note that platform.py was not copied over from pybench-2.0 since it is already part of Python 2.5. ........ r46929 | andrew.macintyre | 2006-06-13 21:02:35 +0200 (Tue, 13 Jun 2006) | 5 lines Increase the small thread stack size to get the test to pass reliably on the one buildbot that insists on more than 32kB of thread stack. ........ r46930 | marc-andre.lemburg | 2006-06-13 21:20:07 +0200 (Tue, 13 Jun 2006) | 2 lines Whitespace normalization. ........ r46931 | thomas.heller | 2006-06-13 22:18:43 +0200 (Tue, 13 Jun 2006) | 2 lines More docs for ctypes. ........ r46932 | brett.cannon | 2006-06-13 23:34:24 +0200 (Tue, 13 Jun 2006) | 2 lines Ignore .pyc and .pyo files in Pybench. ........ r46933 | brett.cannon | 2006-06-13 23:46:41 +0200 (Tue, 13 Jun 2006) | 7 lines If a classic class defined a __coerce__() method that just returned its two arguments in reverse, the interpreter would infinitely recourse trying to get a coercion that worked. So put in a recursion check after a coercion is made and the next call to attempt to use the coerced values. Fixes bug #992017 and closes crashers/coerce.py . ........ r46936 | gerhard.haering | 2006-06-14 00:24:47 +0200 (Wed, 14 Jun 2006) | 3 lines Merged changes from external pysqlite 2.3.0 release. Documentation updates will follow in a few hours at the latest. Then we should be ready for beta1. ........ r46937 | brett.cannon | 2006-06-14 00:26:13 +0200 (Wed, 14 Jun 2006) | 2 lines Missed test for rev. 46933; infinite recursion from __coerce__() returning its arguments reversed. ........ r46938 | gerhard.haering | 2006-06-14 00:53:48 +0200 (Wed, 14 Jun 2006) | 2 lines Updated documentation for pysqlite 2.3.0 API. ........ r46939 | tim.peters | 2006-06-14 06:09:25 +0200 (Wed, 14 Jun 2006) | 10 lines SequenceMatcher.get_matching_blocks(): This now guarantees that adjacent triples in the result list describe non-adjacent matching blocks. That's _nice_ to have, and Guido said he wanted it. Not a bugfix candidate: Guido or not ;-), this changes visible endcase semantics (note that some tests had to change), and nothing about this was documented before. Since it was working as designed, and behavior was consistent with the docs, it wasn't "a bug". ........ r46940 | tim.peters | 2006-06-14 06:13:00 +0200 (Wed, 14 Jun 2006) | 2 lines Repaired typo in new comment. ........ r46941 | tim.peters | 2006-06-14 06:15:27 +0200 (Wed, 14 Jun 2006) | 2 lines Whitespace normalization. ........ r46942 | fred.drake | 2006-06-14 06:25:02 +0200 (Wed, 14 Jun 2006) | 3 lines - make some disabled tests run what they intend when enabled - remove some over-zealous triple-quoting ........ r46943 | fred.drake | 2006-06-14 07:04:47 +0200 (Wed, 14 Jun 2006) | 3 lines add tests for two cases that are handled correctly in the current code, but that SF patch 1504676 as written mis-handles ........ r46944 | fred.drake | 2006-06-14 07:15:51 +0200 (Wed, 14 Jun 2006) | 1 line explain an XXX in more detail ........ r46945 | martin.v.loewis | 2006-06-14 07:21:04 +0200 (Wed, 14 Jun 2006) | 1 line Patch #1455898: Incremental mode for "mbcs" codec. ........ r46946 | georg.brandl | 2006-06-14 08:08:31 +0200 (Wed, 14 Jun 2006) | 3 lines Bug #1339007: Shelf objects now don't raise an exception in their __del__ method when initialization failed. ........ r46948 | thomas.heller | 2006-06-14 08:18:15 +0200 (Wed, 14 Jun 2006) | 1 line Fix docstring. ........ r46949 | georg.brandl | 2006-06-14 08:29:07 +0200 (Wed, 14 Jun 2006) | 2 lines Bug #1501122: mention __gt__ &co in description of comparison order. ........ r46951 | thomas.heller | 2006-06-14 09:08:38 +0200 (Wed, 14 Jun 2006) | 1 line Write more docs. ........ r46952 | georg.brandl | 2006-06-14 10:31:39 +0200 (Wed, 14 Jun 2006) | 3 lines Bug #1153163: describe __add__ vs __radd__ behavior when adding objects of same type/of subclasses of the other. ........ r46954 | georg.brandl | 2006-06-14 10:42:11 +0200 (Wed, 14 Jun 2006) | 3 lines Bug #1202018: add some common mime.types locations. ........ r46955 | georg.brandl | 2006-06-14 10:50:03 +0200 (Wed, 14 Jun 2006) | 3 lines Bug #1117556: SimpleHTTPServer now tries to find and use the system's mime.types file for determining MIME types. ........ r46957 | thomas.heller | 2006-06-14 11:09:08 +0200 (Wed, 14 Jun 2006) | 1 line Document paramflags. ........ r46958 | thomas.heller | 2006-06-14 11:20:11 +0200 (Wed, 14 Jun 2006) | 1 line Add an __all__ list, since this module does 'from ctypes import *'. ........ r46959 | andrew.kuchling | 2006-06-14 15:59:15 +0200 (Wed, 14 Jun 2006) | 1 line Add item ........ r46961 | georg.brandl | 2006-06-14 18:46:43 +0200 (Wed, 14 Jun 2006) | 3 lines Bug #805015: doc error in PyUnicode_FromEncodedObject. ........ r46962 | gerhard.haering | 2006-06-15 00:28:37 +0200 (Thu, 15 Jun 2006) | 10 lines - Added version checks in C code to make sure we don't trigger bugs in older SQLite versions. - Added version checks in test suite so that we don't execute tests that we know will fail with older (buggy) SQLite versions. Now, all tests should run against all SQLite versions from 3.0.8 until 3.3.6 (latest one now). The sqlite3 module can be built against all these SQLite versions and the sqlite3 module does its best to not trigger bugs in SQLite, but using SQLite 3.3.3 or later is recommended. ........ r46963 | tim.peters | 2006-06-15 00:38:13 +0200 (Thu, 15 Jun 2006) | 2 lines Whitespace normalization. ........ r46964 | neal.norwitz | 2006-06-15 06:54:29 +0200 (Thu, 15 Jun 2006) | 9 lines Speculative checkin (requires approval of Gerhard Haering) This backs out the test changes in 46962 which prevented crashes by not running the tests via a version check. All the version checks added in that rev were removed from the tests. Code was added to the error handler in connection.c that seems to work with older versions of sqlite including 3.1.3. ........ r46965 | neal.norwitz | 2006-06-15 07:55:49 +0200 (Thu, 15 Jun 2006) | 1 line Try to narrow window of failure on slow/busy boxes (ppc64 buildbot) ........ r46966 | martin.v.loewis | 2006-06-15 08:45:05 +0200 (Thu, 15 Jun 2006) | 2 lines Make import/lookup of mbcs fail on non-Windows systems. ........ r46967 | ronald.oussoren | 2006-06-15 10:14:18 +0200 (Thu, 15 Jun 2006) | 2 lines Patch #1446489 (zipfile: support for ZIP64) ........ r46968 | neal.norwitz | 2006-06-15 10:16:44 +0200 (Thu, 15 Jun 2006) | 6 lines Re-revert this change. Install the version check and don't run the test until Gerhard has time to fully debug the issue. This affects versions before 3.2.1 (possibly only versions earlier than 3.1.3). Based on discussion on python-checkins. ........ r46969 | gregory.p.smith | 2006-06-15 10:52:32 +0200 (Thu, 15 Jun 2006) | 6 lines - bsddb: multithreaded DB access using the simple bsddb module interface now works reliably. It has been updated to use automatic BerkeleyDB deadlock detection and the bsddb.dbutils.DeadlockWrap wrapper to retry database calls that would previously deadlock. [SF python bug #775414] ........ r46970 | gregory.p.smith | 2006-06-15 11:23:52 +0200 (Thu, 15 Jun 2006) | 2 lines minor documentation cleanup. mention the bsddb.db interface explicitly by name. ........ r46971 | neal.norwitz | 2006-06-15 11:57:03 +0200 (Thu, 15 Jun 2006) | 5 lines Steal the trick from test_compiler to print out a slow msg. This will hopefully get the buildbots to pass. Not sure this test will be feasible or even work. But everything is red now, so it can't get much worse. ........ r46972 | neal.norwitz | 2006-06-15 12:24:49 +0200 (Thu, 15 Jun 2006) | 1 line Print some more info to get an idea of how much longer the test will last ........ r46981 | tim.peters | 2006-06-15 20:04:40 +0200 (Thu, 15 Jun 2006) | 6 lines Try to reduce the extreme peak memory and disk-space use of this test. It probably still requires more disk space than most buildbots have, and in any case is still so intrusive that if we don't find another way to test this I'm taking my buildbot offline permanently ;-) ........ r46982 | tim.peters | 2006-06-15 20:06:29 +0200 (Thu, 15 Jun 2006) | 2 lines Whitespace normalization. ........ r46983 | tim.peters | 2006-06-15 20:07:28 +0200 (Thu, 15 Jun 2006) | 2 lines Add missing svn:eol-style property to text files. ........ r46984 | tim.peters | 2006-06-15 20:38:19 +0200 (Thu, 15 Jun 2006) | 2 lines Oops -- I introduced an off-by-6436159488 error. ........ r46990 | neal.norwitz | 2006-06-16 06:30:34 +0200 (Fri, 16 Jun 2006) | 1 line Disable this test until we can determine what to do about it ........ r46991 | neal.norwitz | 2006-06-16 06:31:06 +0200 (Fri, 16 Jun 2006) | 1 line Param name is dir, not directory. Update docstring. Backport candidate ........ r46992 | neal.norwitz | 2006-06-16 06:31:28 +0200 (Fri, 16 Jun 2006) | 1 line Add missing period in comment. ........ r46993 | neal.norwitz | 2006-06-16 06:32:43 +0200 (Fri, 16 Jun 2006) | 1 line Fix whitespace, there are memory leaks in this module. ........ r46995 | fred.drake | 2006-06-17 01:45:06 +0200 (Sat, 17 Jun 2006) | 3 lines SF patch 1504676: Make sgmllib char and entity references pluggable (implementation/tests contributed by Sam Ruby) ........ r46996 | fred.drake | 2006-06-17 03:07:54 +0200 (Sat, 17 Jun 2006) | 1 line fix change that broke the htmllib tests ........ r46998 | martin.v.loewis | 2006-06-17 11:15:14 +0200 (Sat, 17 Jun 2006) | 3 lines Patch #763580: Add name and value arguments to Tkinter variable classes. ........ r46999 | martin.v.loewis | 2006-06-17 11:20:41 +0200 (Sat, 17 Jun 2006) | 2 lines Patch #1096231: Add default argument to wm_iconbitmap. ........ r47000 | martin.v.loewis | 2006-06-17 11:25:15 +0200 (Sat, 17 Jun 2006) | 2 lines Patch #1494750: Destroy master after deleting children. ........ r47003 | george.yoshida | 2006-06-17 18:31:52 +0200 (Sat, 17 Jun 2006) | 2 lines markup fix ........ r47005 | george.yoshida | 2006-06-17 18:39:13 +0200 (Sat, 17 Jun 2006) | 4 lines Update url. Old url returned status code:301 Moved permanently. ........ r47007 | martin.v.loewis | 2006-06-17 20:44:27 +0200 (Sat, 17 Jun 2006) | 2 lines Patch #812986: Update the canvas even if not tracing. ........ r47008 | martin.v.loewis | 2006-06-17 21:03:26 +0200 (Sat, 17 Jun 2006) | 2 lines Patch #815924: Restore ability to pass type= and icon= ........ r47009 | neal.norwitz | 2006-06-18 00:37:45 +0200 (Sun, 18 Jun 2006) | 1 line Fix typo in docstring ........ r47010 | neal.norwitz | 2006-06-18 00:38:15 +0200 (Sun, 18 Jun 2006) | 1 line Fix memory leak reported by valgrind while running test_subprocess ........ r47011 | fred.drake | 2006-06-18 04:57:35 +0200 (Sun, 18 Jun 2006) | 1 line remove unnecessary markup ........ r47013 | neal.norwitz | 2006-06-18 21:35:01 +0200 (Sun, 18 Jun 2006) | 7 lines Prevent spurious leaks when running regrtest.py -R. There may be more issues that crop up from time to time, but this change seems to have been pretty stable (no spurious warnings) for about a week. Other modules which use threads may require similar use of threading_setup/threading_cleanup from test_support. ........ r47014 | neal.norwitz | 2006-06-18 21:37:40 +0200 (Sun, 18 Jun 2006) | 9 lines The hppa ubuntu box sometimes hangs forever in these tests. My guess is that the wait is failing for some reason. Use WNOHANG, so we won't wait until the buildbot kills the test suite. I haven't been able to reproduce the failure, so I'm not sure if this will help or not. Hopefully, this change will cause the test to fail, rather than hang. That will be better since we will get the rest of the test results. It may also help us debug the real problem. ........ r47015 | neal.norwitz | 2006-06-18 22:10:24 +0200 (Sun, 18 Jun 2006) | 1 line Revert 47014 until it is more robust ........ r47016 | thomas.heller | 2006-06-18 23:27:04 +0200 (Sun, 18 Jun 2006) | 6 lines Fix typos. Fix doctest example. Mention in the tutorial that 'errcheck' is explained in the ref manual. Use better wording in some places. Remoce code examples that shouldn't be in the tutorial. Remove some XXX notices. ........ r47017 | georg.brandl | 2006-06-19 00:17:29 +0200 (Mon, 19 Jun 2006) | 3 lines Patch #1507676: improve exception messages in abstract.c, object.c and typeobject.c. ........ r47018 | neal.norwitz | 2006-06-19 07:40:44 +0200 (Mon, 19 Jun 2006) | 1 line Use Py_ssize_t ........ r47019 | georg.brandl | 2006-06-19 08:35:54 +0200 (Mon, 19 Jun 2006) | 3 lines Add news entry about error msg improvement. ........ r47020 | thomas.heller | 2006-06-19 09:07:49 +0200 (Mon, 19 Jun 2006) | 2 lines Try to repair the failing test on the OpenBSD buildbot. Trial and error... ........ r47021 | tim.peters | 2006-06-19 09:45:16 +0200 (Mon, 19 Jun 2006) | 2 lines Whitespace normalization. ........ r47022 | walter.doerwald | 2006-06-19 10:07:50 +0200 (Mon, 19 Jun 2006) | 4 lines Patch #1506645: add Python wrappers for the curses functions is_term_resized, resize_term and resizeterm. This uses three separate configure checks (one for each function). ........ r47023 | walter.doerwald | 2006-06-19 10:14:09 +0200 (Mon, 19 Jun 2006) | 2 lines Make check order match in configure and configure.in. ........ r47024 | tim.peters | 2006-06-19 10:14:28 +0200 (Mon, 19 Jun 2006) | 3 lines Repair KeyError when running test_threaded_import under -R, as reported by Neal on python-dev. ........ r47025 | thomas.heller | 2006-06-19 10:32:46 +0200 (Mon, 19 Jun 2006) | 3 lines Next try to fix the OpenBSD buildbot tests: Use ctypes.util.find_library to locate the C runtime library on platforms where is returns useful results. ........ r47026 | tim.peters | 2006-06-19 11:09:44 +0200 (Mon, 19 Jun 2006) | 13 lines TestHelp.make_parser(): This was making a permanent change to os.environ (setting envar COLUMNS), which at least caused test_float_default() to fail if the tests were run more than once. This repairs the test_optparse -R failures Neal reported on python-dev. It also explains some seemingly bizarre test_optparse failures we saw a couple weeks ago on the buildbots, when test_optparse failed due to test_file failing to clean up after itself, and then test_optparse failed in an entirely different way when regrtest's -w option ran test_optparse a second time. It's now obvious that make_parser() permanently changing os.environ was responsible for the second half of that. ........ r47027 | anthony.baxter | 2006-06-19 14:04:15 +0200 (Mon, 19 Jun 2006) | 2 lines Preparing for 2.5b1. ........ r47029 | fred.drake | 2006-06-19 19:31:16 +0200 (Mon, 19 Jun 2006) | 1 line remove non-working document formats from edist ........ r47030 | gerhard.haering | 2006-06-19 23:17:35 +0200 (Mon, 19 Jun 2006) | 5 lines Fixed a memory leak that was introduced with incorrect usage of the Python weak reference API in pysqlite 2.2.1. Bumbed pysqlite version number to upcoming pysqlite 2.3.1 release. ........ r47032 | ka-ping.yee | 2006-06-20 00:49:36 +0200 (Tue, 20 Jun 2006) | 2 lines Remove Python 2.3 compatibility comment. ........ r47033 | trent.mick | 2006-06-20 01:21:25 +0200 (Tue, 20 Jun 2006) | 2 lines Upgrade pyexpat to expat 2.0.0 (http://python.org/sf/1462338). ........ r47034 | trent.mick | 2006-06-20 01:57:41 +0200 (Tue, 20 Jun 2006) | 3 lines [ 1295808 ] expat symbols should be namespaced in pyexpat (http://python.org/sf/1295808) ........ r47039 | andrew.kuchling | 2006-06-20 13:52:16 +0200 (Tue, 20 Jun 2006) | 1 line Uncomment wsgiref section ........ r47040 | andrew.kuchling | 2006-06-20 14:15:09 +0200 (Tue, 20 Jun 2006) | 1 line Add four library items ........ r47041 | andrew.kuchling | 2006-06-20 14:19:54 +0200 (Tue, 20 Jun 2006) | 1 line Terminology and typography fixes ........ r47042 | andrew.kuchling | 2006-06-20 15:05:12 +0200 (Tue, 20 Jun 2006) | 1 line Add introductory paragraphs summarizing the release; minor edits ........ r47043 | andrew.kuchling | 2006-06-20 15:11:29 +0200 (Tue, 20 Jun 2006) | 1 line Minor edits and rearrangements; markup fix ........ r47044 | andrew.kuchling | 2006-06-20 15:20:30 +0200 (Tue, 20 Jun 2006) | 1 line [Bug #1504456] Mention xml -> xmlcore change ........ r47047 | brett.cannon | 2006-06-20 19:30:26 +0200 (Tue, 20 Jun 2006) | 2 lines Raise TestSkipped when the test socket connection is refused. ........ r47049 | brett.cannon | 2006-06-20 21:20:17 +0200 (Tue, 20 Jun 2006) | 2 lines Fix typo of exception name. ........ r47053 | brett.cannon | 2006-06-21 18:57:57 +0200 (Wed, 21 Jun 2006) | 5 lines At the C level, tuple arguments are passed in directly to the exception constructor, meaning it is treated as *args, not as a single argument. This means using the 'message' attribute won't work (until Py3K comes around), and so one must grab from 'arg' to get the error number. ........ r47054 | andrew.kuchling | 2006-06-21 19:10:18 +0200 (Wed, 21 Jun 2006) | 1 line Link to LibRef module documentation ........ r47055 | andrew.kuchling | 2006-06-21 19:17:10 +0200 (Wed, 21 Jun 2006) | 1 line Note some of Barry's work ........ r47056 | andrew.kuchling | 2006-06-21 19:17:28 +0200 (Wed, 21 Jun 2006) | 1 line Bump version ........ r47057 | georg.brandl | 2006-06-21 19:45:17 +0200 (Wed, 21 Jun 2006) | 3 lines fix [ 1509132 ] compiler module builds incorrect AST for TryExceptFinally ........ r47058 | georg.brandl | 2006-06-21 19:52:36 +0200 (Wed, 21 Jun 2006) | 3 lines Make test_fcntl aware of netbsd3. ........ r47059 | georg.brandl | 2006-06-21 19:53:17 +0200 (Wed, 21 Jun 2006) | 3 lines Patch #1509001: expected skips for netbsd3. ........ r47060 | gerhard.haering | 2006-06-21 22:55:04 +0200 (Wed, 21 Jun 2006) | 2 lines Removed call to enable_callback_tracebacks that slipped in by accident. ........ r47061 | armin.rigo | 2006-06-21 23:58:50 +0200 (Wed, 21 Jun 2006) | 13 lines Fix for an obscure bug introduced by revs 46806 and 46808, with a test. The problem of checking too eagerly for recursive calls is the following: if a RuntimeError is caused by recursion, and if code needs to normalize it immediately (as in the 2nd test), then PyErr_NormalizeException() needs a call to the RuntimeError class to instantiate it, and this hits the recursion limit again... causing PyErr_NormalizeException() to never finish. Moved this particular recursion check to slot_tp_call(), which is not involved in instantiating built-in exceptions. Backport candidate. ........ r47064 | neal.norwitz | 2006-06-22 08:30:50 +0200 (Thu, 22 Jun 2006) | 3 lines Copy the wsgiref package during make install. ........ r47065 | neal.norwitz | 2006-06-22 08:35:30 +0200 (Thu, 22 Jun 2006) | 1 line Reset the doc date to today for the automatic doc builds ........ r47067 | andrew.kuchling | 2006-06-22 15:10:23 +0200 (Thu, 22 Jun 2006) | 1 line Mention how to suppress warnings ........ r47069 | georg.brandl | 2006-06-22 16:46:17 +0200 (Thu, 22 Jun 2006) | 3 lines Set lineno correctly on list, tuple and dict literals. ........ r47070 | georg.brandl | 2006-06-22 16:46:46 +0200 (Thu, 22 Jun 2006) | 4 lines Test for correct compilation of try-except-finally stmt. Test for correct lineno on list, tuple, dict literals. ........ r47071 | fred.drake | 2006-06-22 17:50:08 +0200 (Thu, 22 Jun 2006) | 1 line fix markup nit ........ r47072 | brett.cannon | 2006-06-22 18:49:14 +0200 (Thu, 22 Jun 2006) | 6 lines 'warning's was improperly requiring that a command-line Warning category be both a subclass of Warning and a subclass of types.ClassType. The latter is no longer true thanks to new-style exceptions. Closes bug #1510580. Thanks to AMK for the test. ........ r47073 | ronald.oussoren | 2006-06-22 20:33:54 +0200 (Thu, 22 Jun 2006) | 3 lines MacOSX: Add a message to the first screen of the installer that tells users how to avoid updates to their shell profile. ........ r47074 | georg.brandl | 2006-06-22 21:02:18 +0200 (Thu, 22 Jun 2006) | 3 lines Fix my name ;) ........ r47075 | thomas.heller | 2006-06-22 21:07:36 +0200 (Thu, 22 Jun 2006) | 2 lines Small fixes, mostly in the markup. ........ r47076 | peter.astrand | 2006-06-22 22:06:46 +0200 (Thu, 22 Jun 2006) | 1 line Make it possible to run test_subprocess.py on Python 2.2, which lacks test_support.is_resource_enabled. ........ r47077 | peter.astrand | 2006-06-22 22:21:26 +0200 (Thu, 22 Jun 2006) | 1 line Applied patch #1506758: Prevent MemoryErrors with large MAXFD. ........ r47079 | neal.norwitz | 2006-06-23 05:32:44 +0200 (Fri, 23 Jun 2006) | 1 line Fix refleak ........ r47080 | fred.drake | 2006-06-23 08:03:45 +0200 (Fri, 23 Jun 2006) | 9 lines - SF bug #853506: IP6 address parsing in sgmllib ('[' and ']' were not accepted in unquoted attribute values) - cleaned up tests of character and entity reference decoding so the tests cover the documented relationships among handle_charref, handle_entityref, convert_charref, convert_codepoint, and convert_entityref, without bringing up Unicode issues that sgmllib cannot be involved in ........ r47085 | andrew.kuchling | 2006-06-23 21:23:40 +0200 (Fri, 23 Jun 2006) | 11 lines Fit Makefile for the Python doc environment better; this is a step toward including the howtos in the build process. * Put LaTeX output in ../paper-<whatever>/. * Put HTML output in ../html/ * Explain some of the Makefile variables * Remove some cruft dating to my environment (e.g. the 'web' target) This makefile isn't currently invoked by the documentation build process, so these changes won't destabilize anything. ........ r47086 | hyeshik.chang | 2006-06-23 23:16:18 +0200 (Fri, 23 Jun 2006) | 5 lines Bug #1511381: codec_getstreamcodec() in codec.c is corrected to omit a default "error" argument for NULL pointer. This allows the parser to take a codec from cjkcodecs again. (Reported by Taewook Kang and reviewed by Walter Doerwald) ........ r47091 | ronald.oussoren | 2006-06-25 22:44:16 +0200 (Sun, 25 Jun 2006) | 6 lines Workaround for bug #1512124 Without this patch IDLE will get unresponsive when you open the debugger window on OSX. This is both using the system Tcl/Tk on Tiger as the latest universal download from tk-components.sf.net. ........ r47092 | ronald.oussoren | 2006-06-25 23:14:19 +0200 (Sun, 25 Jun 2006) | 3 lines Drop the calldll demo's for macos, calldll isn't present anymore, no need to keep the demo's around. ........ r47093 | ronald.oussoren | 2006-06-25 23:15:58 +0200 (Sun, 25 Jun 2006) | 3 lines Use a path without a double slash to compile the .py files after installation (macosx, binary installer). This fixes bug #1508369 for python 2.5. ........ r47094 | ronald.oussoren | 2006-06-25 23:19:06 +0200 (Sun, 25 Jun 2006) | 3 lines Also install the .egg-info files in Lib. This will cause wsgiref.egg-info to be installed. ........ r47097 | andrew.kuchling | 2006-06-26 14:40:02 +0200 (Mon, 26 Jun 2006) | 1 line [Bug #1511998] Various comments from Nick Coghlan; thanks! ........ r47098 | andrew.kuchling | 2006-06-26 14:43:43 +0200 (Mon, 26 Jun 2006) | 1 line Describe workaround for PyRange_New()'s removal ........ r47099 | andrew.kuchling | 2006-06-26 15:08:24 +0200 (Mon, 26 Jun 2006) | 5 lines [Bug #1512163] Fix typo. This change will probably break tests on FreeBSD buildbots, but I'll check in a fix for that next. ........ r47100 | andrew.kuchling | 2006-06-26 15:12:16 +0200 (Mon, 26 Jun 2006) | 9 lines [Bug #1512163] Use one set of locking methods, lockf(); remove the flock() calls. On FreeBSD, the two methods lockf() and flock() end up using the same mechanism and the second one fails. A Linux man page claims that the two methods are orthogonal (so locks acquired one way don't interact with locks acquired the other way) but that clearly must be false. ........ r47101 | andrew.kuchling | 2006-06-26 15:23:10 +0200 (Mon, 26 Jun 2006) | 5 lines Add a test for a conflicting lock. On slow machines, maybe the time intervals (2 sec, 0.5 sec) will be too tight. I'll see how the buildbots like it. ........ r47103 | andrew.kuchling | 2006-06-26 16:33:24 +0200 (Mon, 26 Jun 2006) | 1 line Windows doesn't have os.fork(). I'll just disable this test for now ........ r47106 | andrew.kuchling | 2006-06-26 19:00:35 +0200 (Mon, 26 Jun 2006) | 9 lines Attempt to fix build failure on OS X and Debian alpha; the symptom is consistent with os.wait() returning immediately because some other subprocess had previously exited; the test suite then immediately tries to lock the mailbox and gets an error saying it's already locked. To fix this, do a waitpid() so the test suite only continues once the intended child process has exited. ........ r47113 | neal.norwitz | 2006-06-27 06:06:46 +0200 (Tue, 27 Jun 2006) | 1 line Ignore some more warnings in the dynamic linker on an older gentoo ........ r47114 | neal.norwitz | 2006-06-27 06:09:13 +0200 (Tue, 27 Jun 2006) | 6 lines Instead of doing a make test, run the regression tests out of the installed copy. This will hopefully catch problems where directories are added under Lib/ but not to Makefile.pre.in. This breaks out the 2 runs of the test suite with and without -O which is also nicer. ........ r47115 | neal.norwitz | 2006-06-27 06:12:58 +0200 (Tue, 27 Jun 2006) | 5 lines Fix SF bug #1513032, 'make install' failure on FreeBSD 5.3. No need to install lib-old, it's empty in 2.5. ........ r47116 | neal.norwitz | 2006-06-27 06:23:06 +0200 (Tue, 27 Jun 2006) | 1 line Test unimportant change to verify buildbot does not try to build ........ r47117 | neal.norwitz | 2006-06-27 06:26:30 +0200 (Tue, 27 Jun 2006) | 1 line Try again: test unimportant change to verify buildbot does not try to build ........ r47118 | neal.norwitz | 2006-06-27 06:28:56 +0200 (Tue, 27 Jun 2006) | 1 line Verify buildbot picks up these changes (really needs testing after last change to Makefile.pre.in) ........ r47121 | vinay.sajip | 2006-06-27 09:34:37 +0200 (Tue, 27 Jun 2006) | 1 line Removed buggy exception handling in doRollover of rotating file handlers. Exceptions now propagate to caller. ........ r47123 | ronald.oussoren | 2006-06-27 12:08:25 +0200 (Tue, 27 Jun 2006) | 3 lines MacOSX: fix rather dumb buglet that made it impossible to create extensions on OSX 10.3 when using a binary distribution build on 10.4. ........ r47125 | tim.peters | 2006-06-27 13:52:49 +0200 (Tue, 27 Jun 2006) | 2 lines Whitespace normalization. ........ r47128 | ronald.oussoren | 2006-06-27 14:53:52 +0200 (Tue, 27 Jun 2006) | 8 lines Use staticly build copies of zlib and bzip2 to build the OSX installer, that way the resulting binaries have a better change of running on 10.3. This patch also updates the search logic for sleepycat db3/4, without this patch you cannot use a sleepycat build with a non-standard prefix; with this you can (at least on OSX) if you add the prefix to CPPFLAGS/LDFLAGS at configure-time. This change is needed to build the binary installer for OSX. ........ r47131 | ronald.oussoren | 2006-06-27 17:45:32 +0200 (Tue, 27 Jun 2006) | 5 lines macosx: Install a libpython2.5.a inside the framework as a symlink to the actual dylib at the root of the framework, that way tools that expect a unix-like install (python-config, but more importantly external products like mod_python) work correctly. ........ r47137 | neal.norwitz | 2006-06-28 07:03:22 +0200 (Wed, 28 Jun 2006) | 4 lines According to the man pages on Gentoo Linux and Tru64, EACCES or EAGAIN can be returned if fcntl (lockf) fails. This fixes the test failure on Tru64 by checking for either error rather than just EAGAIN. ........ r47139 | neal.norwitz | 2006-06-28 08:28:31 +0200 (Wed, 28 Jun 2006) | 5 lines Fix bug #1512695: cPickle.loads could crash if it was interrupted with a KeyboardInterrupt since PyTuple_Pack was passed a NULL. Will backport. ........ r47142 | nick.coghlan | 2006-06-28 12:41:47 +0200 (Wed, 28 Jun 2006) | 1 line Make full module name available as __module_name__ even when __name__ is set to something else (like '__main__') ........ r47143 | armin.rigo | 2006-06-28 12:49:51 +0200 (Wed, 28 Jun 2006) | 2 lines A couple of crashers of the "won't fix" kind. ........ r47147 | andrew.kuchling | 2006-06-28 16:25:20 +0200 (Wed, 28 Jun 2006) | 1 line [Bug #1508766] Add docs for uuid module; docs written by George Yoshida, with minor rearrangements by me. ........ r47148 | andrew.kuchling | 2006-06-28 16:27:21 +0200 (Wed, 28 Jun 2006) | 1 line [Bug #1508766] Add docs for uuid module; this puts the module in the 'Internet Protocols' section. Arguably this module could also have gone in the chapters on strings or encodings, maybe even the crypto chapter. Fred, please move if you see fit. ........ r47151 | georg.brandl | 2006-06-28 22:23:25 +0200 (Wed, 28 Jun 2006) | 3 lines Fix end_fill(). ........ r47153 | trent.mick | 2006-06-28 22:30:41 +0200 (Wed, 28 Jun 2006) | 2 lines Mention the expat upgrade and pyexpat fix I put in 2.5b1. ........ r47154 | fred.drake | 2006-06-29 02:51:53 +0200 (Thu, 29 Jun 2006) | 6 lines SF bug #1504333: sgmlib should allow angle brackets in quoted values (modified patch by Sam Ruby; changed to use separate REs for start and end tags to reduce matching cost for end tags; extended tests; updated to avoid breaking previous changes to support IPv6 addresses in unquoted attribute values) ........ r47156 | fred.drake | 2006-06-29 04:57:48 +0200 (Thu, 29 Jun 2006) | 1 line document recent bugfixes in sgmllib ........ r47158 | neal.norwitz | 2006-06-29 06:10:08 +0200 (Thu, 29 Jun 2006) | 10 lines Add new utility function, reap_children(), to test_support. This should be called at the end of each test that spawns children (perhaps it should be called from regrtest instead?). This will hopefully prevent some of the unexplained failures in the buildbots (hppa and alpha) during tests that spawn children. The problems were not reproducible. There were many zombies that remained at the end of several tests. In the worst case, this shouldn't cause any more problems, though it may not help either. Time will tell. ........ r47159 | neal.norwitz | 2006-06-29 07:48:14 +0200 (Thu, 29 Jun 2006) | 5 lines This should fix the buildbot failure on s/390 which can't connect to gmail.org. It makes the error message consistent and always sends to stderr. It would be much better for all the networking tests to hit only python.org. ........ r47161 | thomas.heller | 2006-06-29 20:34:15 +0200 (Thu, 29 Jun 2006) | 3 lines Protect the thread api calls in the _ctypes extension module within #ifdef WITH_THREADS/#endif blocks. Found by Sam Rushing. ........ r47162 | martin.v.loewis | 2006-06-29 20:58:44 +0200 (Thu, 29 Jun 2006) | 2 lines Patch #1509163: MS Toolkit Compiler no longer available ........ r47163 | skip.montanaro | 2006-06-29 21:20:09 +0200 (Thu, 29 Jun 2006) | 1 line add string methods to index ........ r47164 | vinay.sajip | 2006-06-30 02:13:08 +0200 (Fri, 30 Jun 2006) | 1 line Fixed bug in fileConfig() which failed to clear logging._handlerList ........ r47166 | tim.peters | 2006-06-30 08:18:39 +0200 (Fri, 30 Jun 2006) | 2 lines Whitespace normalization. ........ r47170 | neal.norwitz | 2006-06-30 09:32:16 +0200 (Fri, 30 Jun 2006) | 1 line Silence compiler warning ........ r47171 | neal.norwitz | 2006-06-30 09:32:46 +0200 (Fri, 30 Jun 2006) | 1 line Another problem reported by Coverity. Backport candidate. ........ r47175 | thomas.heller | 2006-06-30 19:44:54 +0200 (Fri, 30 Jun 2006) | 2 lines Revert the use of PY_FORMAT_SIZE_T in PyErr_Format. ........ r47176 | tim.peters | 2006-06-30 20:34:51 +0200 (Fri, 30 Jun 2006) | 2 lines Remove now-unused fidding with PY_FORMAT_SIZE_T. ........ r47177 | georg.brandl | 2006-06-30 20:47:56 +0200 (Fri, 30 Jun 2006) | 3 lines Document decorator usage of property. ........ r47181 | fred.drake | 2006-06-30 21:29:25 +0200 (Fri, 30 Jun 2006) | 4 lines - consistency nit: always include "()" in \function and \method (*should* be done by the presentation, but that requires changes all over) - avoid spreading the __name meme ........ r47188 | vinay.sajip | 2006-07-01 12:45:20 +0200 (Sat, 01 Jul 2006) | 1 line Added entry for fileConfig() bugfix. ........ r47189 | vinay.sajip | 2006-07-01 12:47:20 +0200 (Sat, 01 Jul 2006) | 1 line Added duplicate call to fileConfig() to ensure that it cleans up after itself correctly. ........ r47190 | martin.v.loewis | 2006-07-01 17:33:37 +0200 (Sat, 01 Jul 2006) | 2 lines Release all forwarded functions in .close. Fixes #1513223. ........ r47191 | fred.drake | 2006-07-01 18:28:20 +0200 (Sat, 01 Jul 2006) | 7 lines SF bug #1296433 (Expat bug #1515266): Unchecked calls to character data handler would cause a segfault. This merges in Expat's lib/xmlparse.c revisions 1.154 and 1.155, which fix this and a closely related problem (the later does not affect Python). Moved the crasher test to the tests for xml.parsers.expat. ........ r47197 | gerhard.haering | 2006-07-02 19:48:30 +0200 (Sun, 02 Jul 2006) | 4 lines The sqlite3 module did cut off data from the SQLite database at the first null character before sending it to a custom converter. This has been fixed now. ........ r47198 | martin.v.loewis | 2006-07-02 20:44:00 +0200 (Sun, 02 Jul 2006) | 1 line Correct arithmetic in access on Win32. Fixes #1513646. ........ r47203 | thomas.heller | 2006-07-03 09:58:09 +0200 (Mon, 03 Jul 2006) | 1 line Cleanup: Remove commented out code. ........ r47204 | thomas.heller | 2006-07-03 09:59:50 +0200 (Mon, 03 Jul 2006) | 1 line Don't run the doctests with Python 2.3 because it doesn't have the ELLIPSIS flag. ........ r47205 | thomas.heller | 2006-07-03 10:04:05 +0200 (Mon, 03 Jul 2006) | 7 lines Fixes so that _ctypes can be compiled with the MingW compiler. It seems that the definition of '__attribute__(x)' was responsible for the compiler ignoring the '__fastcall' attribute on the ffi_closure_SYSV function in libffi_msvc/ffi.c, took me quite some time to figure this out. ........ r47206 | thomas.heller | 2006-07-03 10:08:14 +0200 (Mon, 03 Jul 2006) | 11 lines Add a new function uses_seh() to the _ctypes extension module. This will return True if Windows Structured Exception handling (SEH) is used when calling functions, False otherwise. Currently, only MSVC supports SEH. Fix the test so that it doesn't crash when run with MingW compiled _ctypes. Note that two tests are still failing when mingw is used, I suspect structure layout differences and function calling conventions between MSVC and MingW. ........ r47207 | tim.peters | 2006-07-03 10:23:19 +0200 (Mon, 03 Jul 2006) | 2 lines Whitespace normalization. ........ r47208 | martin.v.loewis | 2006-07-03 11:44:00 +0200 (Mon, 03 Jul 2006) | 3 lines Only setup canvas when it is first created. Fixes #1514703 ........ r47209 | martin.v.loewis | 2006-07-03 12:05:30 +0200 (Mon, 03 Jul 2006) | 3 lines Reimplement turtle.circle using a polyline, to allow correct filling of arcs. Also fixes #1514693. ........ r47210 | martin.v.loewis | 2006-07-03 12:19:49 +0200 (Mon, 03 Jul 2006) | 3 lines Bug #1514693: Update turtle's heading when switching between degrees and radians. ........ r47211 | martin.v.loewis | 2006-07-03 13:12:06 +0200 (Mon, 03 Jul 2006) | 2 lines Document functions added in 2.3 and 2.5. ........ r47212 | martin.v.loewis | 2006-07-03 14:19:50 +0200 (Mon, 03 Jul 2006) | 3 lines Bug #1417699: Reject locale-specific decimal point in float() and atof(). ........ r47213 | martin.v.loewis | 2006-07-03 14:28:58 +0200 (Mon, 03 Jul 2006) | 3 lines Bug #1267547: Put proper recursive setup.py call into the spec file generated by bdist_rpm. ........ r47215 | martin.v.loewis | 2006-07-03 15:01:35 +0200 (Mon, 03 Jul 2006) | 3 lines Patch #825417: Fix timeout processing in expect, read_until. Will backport to 2.4. ........ r47218 | martin.v.loewis | 2006-07-03 15:47:40 +0200 (Mon, 03 Jul 2006) | 2 lines Put method-wrappers into trashcan. Fixes #927248. ........ r47219 | andrew.kuchling | 2006-07-03 16:07:30 +0200 (Mon, 03 Jul 2006) | 1 line [Bug #1515932] Clarify description of slice assignment ........ r47220 | andrew.kuchling | 2006-07-03 16:16:09 +0200 (Mon, 03 Jul 2006) | 4 lines [Bug #1511911] Clarify description of optional arguments to sorted() by improving the xref to the section on lists, and by copying the explanations of the arguments (with a slight modification). ........ r47223 | kristjan.jonsson | 2006-07-03 16:59:05 +0200 (Mon, 03 Jul 2006) | 1 line Fix build problems with the platform SDK on windows. It is not sufficient to test for the C compiler version when determining if we have the secure CRT from microsoft. Must test with an undocumented macro, __STDC_SECURE_LIB__ too. ........ r47224 | ronald.oussoren | 2006-07-04 14:30:22 +0200 (Tue, 04 Jul 2006) | 7 lines Sync the darwin/x86 port libffi with the copy in PyObjC. This fixes a number of bugs in that port. The most annoying ones were due to some subtle differences between the document ABI and the actual implementation :-( (there are no python unittests that fail without this patch, but without it some of libffi's unittests fail). ........ r47234 | georg.brandl | 2006-07-05 10:21:00 +0200 (Wed, 05 Jul 2006) | 3 lines Remove remaining references to OverflowWarning. ........ r47236 | thomas.heller | 2006-07-05 11:13:56 +0200 (Wed, 05 Jul 2006) | 3 lines Fix the bitfield test when _ctypes is compiled with MingW. Structures containing bitfields may have different layout on MSVC and MingW . ........ r47237 | thomas.wouters | 2006-07-05 13:03:49 +0200 (Wed, 05 Jul 2006) | 15 lines Fix bug in passing tuples to string.Template. All other values (with working str() or repr()) would work, just not multi-value tuples. Probably not a backport candidate, since it changes the behaviour of passing a single-element tuple: >>> string.Template("$foo").substitute(dict(foo=(1,))) '(1,)' versus '1' ........ r47241 | georg.brandl | 2006-07-05 16:18:45 +0200 (Wed, 05 Jul 2006) | 2 lines Patch #1517490: fix glitches in filter() docs. ........ r47244 | georg.brandl | 2006-07-05 17:50:05 +0200 (Wed, 05 Jul 2006) | 2 lines no need to elaborate "string". ........ r47251 | neal.norwitz | 2006-07-06 06:28:59 +0200 (Thu, 06 Jul 2006) | 3 lines Fix refleaks reported by Shane Hathaway in SF patch #1515361. This change contains only the changes related to leaking the copy variable. ........ r47253 | fred.drake | 2006-07-06 07:13:22 +0200 (Thu, 06 Jul 2006) | 4 lines - back out Expat change; the final fix to Expat will be different - change the pyexpat wrapper to not be so sensitive to this detail of the Expat implementation (the ex-crasher test still passes) ........ r47257 | neal.norwitz | 2006-07-06 08:45:08 +0200 (Thu, 06 Jul 2006) | 1 line Add a NEWS entry for a recent pyexpat fix ........ r47258 | martin.v.loewis | 2006-07-06 08:55:58 +0200 (Thu, 06 Jul 2006) | 2 lines Add sqlite3.dll to the DLLs component, not to the TkDLLs component. Fixes #1517388. ........ r47259 | martin.v.loewis | 2006-07-06 09:05:21 +0200 (Thu, 06 Jul 2006) | 1 line Properly quote compileall and Lib paths in case TARGETDIR has a space. ........ r47260 | thomas.heller | 2006-07-06 09:50:18 +0200 (Thu, 06 Jul 2006) | 5 lines Revert the change done in svn revision 47206: Add a new function uses_seh() to the _ctypes extension module. This will return True if Windows Structured Exception handling (SEH) is used when calling functions, False otherwise. ........ r47261 | armin.rigo | 2006-07-06 09:58:18 +0200 (Thu, 06 Jul 2006) | 3 lines A couple of examples about how to attack the fact that _PyType_Lookup() returns a borrowed ref. Many of the calls are open to attack. ........ r47262 | thomas.heller | 2006-07-06 10:28:14 +0200 (Thu, 06 Jul 2006) | 2 lines The test that calls a function with invalid arguments and catches the resulting Windows access violation will not be run by default. ........ r47263 | thomas.heller | 2006-07-06 10:48:35 +0200 (Thu, 06 Jul 2006) | 5 lines Patch #1517790: It is now possible to use custom objects in the ctypes foreign function argtypes sequence as long as they provide a from_param method, no longer is it required that the object is a ctypes type. ........ r47264 | thomas.heller | 2006-07-06 10:58:40 +0200 (Thu, 06 Jul 2006) | 2 lines Document the Struture and Union constructors. ........ r47265 | thomas.heller | 2006-07-06 11:11:22 +0200 (Thu, 06 Jul 2006) | 2 lines Document the changes in svn revision 47263, from patch #1517790. ........ r47267 | ronald.oussoren | 2006-07-06 12:13:35 +0200 (Thu, 06 Jul 2006) | 7 lines This patch solves the problem Skip was seeing with zlib, this patch ensures that configure uses similar compiler flags as setup.py when doing the zlib test. Without this patch configure would use the first shared library on the linker path, with this patch it uses the first shared or static library on that path just like setup.py. ........ r47268 | thomas.wouters | 2006-07-06 12:48:28 +0200 (Thu, 06 Jul 2006) | 4 lines NEWS entry for r47267: fixing configure's zlib probing. ........ r47269 | fredrik.lundh | 2006-07-06 14:29:24 +0200 (Thu, 06 Jul 2006) | 3 lines added XMLParser alias for cElementTree compatibility ........ r47271 | nick.coghlan | 2006-07-06 14:53:04 +0200 (Thu, 06 Jul 2006) | 1 line Revert the __module_name__ changes made in rev 47142. We'll revisit this in Python 2.6 ........ r47272 | nick.coghlan | 2006-07-06 15:04:56 +0200 (Thu, 06 Jul 2006) | 1 line Update the tutorial section on relative imports ........ r47273 | nick.coghlan | 2006-07-06 15:35:27 +0200 (Thu, 06 Jul 2006) | 1 line Ignore ImportWarning by default ........ r47274 | nick.coghlan | 2006-07-06 15:41:34 +0200 (Thu, 06 Jul 2006) | 1 line Cover ImportWarning, PendingDeprecationWarning and simplefilter() in the warnings module docs ........ r47275 | nick.coghlan | 2006-07-06 15:47:18 +0200 (Thu, 06 Jul 2006) | 1 line Add NEWS entries for the ImportWarning change and documentation update ........ r47276 | andrew.kuchling | 2006-07-06 15:57:28 +0200 (Thu, 06 Jul 2006) | 1 line ImportWarning is now silent by default ........ r47277 | thomas.heller | 2006-07-06 17:06:05 +0200 (Thu, 06 Jul 2006) | 2 lines Document the correct return type of PyLong_AsUnsignedLongLongMask. ........ r47278 | hyeshik.chang | 2006-07-06 17:21:52 +0200 (Thu, 06 Jul 2006) | 2 lines Add a testcase for r47086 which fixed a bug in codec_getstreamcodec(). ........ r47279 | hyeshik.chang | 2006-07-06 17:39:24 +0200 (Thu, 06 Jul 2006) | 3 lines Test using all CJK encodings for the testcases which don't require specific encodings. ........ r47280 | martin.v.loewis | 2006-07-06 21:28:03 +0200 (Thu, 06 Jul 2006) | 2 lines Properly generate logical file ids. Fixes #1515998. Also correct typo in Control.mapping. ........ r47287 | neal.norwitz | 2006-07-07 08:03:15 +0200 (Fri, 07 Jul 2006) | 17 lines Restore rev 47014: The hppa ubuntu box sometimes hangs forever in these tests. My guess is that the wait is failing for some reason. Use WNOHANG, so we won't wait until the buildbot kills the test suite. I haven't been able to reproduce the failure, so I'm not sure if this will help or not. Hopefully, this change will cause the test to fail, rather than hang. That will be better since we will get the rest of the test results. It may also help us debug the real problem. *** The reason this originally failed was because there were many zombie children outstanding before rev 47158 cleaned them up. There are still hangs in test_subprocess that need to be addressed, but that will take more work. This should close some holes. ........ r47289 | georg.brandl | 2006-07-07 10:15:12 +0200 (Fri, 07 Jul 2006) | 3 lines Fix RFC number. ........ r50489 | neal.norwitz | 2006-07-08 07:31:37 +0200 (Sat, 08 Jul 2006) | 1 line Fix SF bug #1519018: 'as' is now validated properly in import statements ........ r50490 | georg.brandl | 2006-07-08 14:15:27 +0200 (Sat, 08 Jul 2006) | 3 lines Add an additional test for bug #1519018. ........ r50491 | tim.peters | 2006-07-08 21:55:05 +0200 (Sat, 08 Jul 2006) | 2 lines Whitespace normalization. ........ r50493 | neil.schemenauer | 2006-07-09 18:16:34 +0200 (Sun, 09 Jul 2006) | 2 lines Fix AST compiler bug #1501934: incorrect LOAD/STORE_GLOBAL generation. ........ r50495 | neil.schemenauer | 2006-07-09 23:19:29 +0200 (Sun, 09 Jul 2006) | 2 lines Fix SF bug 1441486: bad unary minus folding in compiler. ........ r50497 | neal.norwitz | 2006-07-10 00:14:42 +0200 (Mon, 10 Jul 2006) | 4 lines On 64 bit systems, int literals that use less than 64 bits are now ints rather than longs. This also fixes the test for eval(-sys.maxint - 1). ........ r50500 | neal.norwitz | 2006-07-10 02:04:44 +0200 (Mon, 10 Jul 2006) | 4 lines Bug #1512814, Fix incorrect lineno's when code at module scope started after line 256. ........ r50501 | neal.norwitz | 2006-07-10 02:05:34 +0200 (Mon, 10 Jul 2006) | 1 line Fix doco. Backport candidate. ........ r50503 | neal.norwitz | 2006-07-10 02:23:17 +0200 (Mon, 10 Jul 2006) | 5 lines Part of SF patch #1484695. This removes dead code. The chksum was already verified in .frombuf() on the lines above. If there was a problem an exception is raised, so there was no way this condition could have been true. ........ r50504 | neal.norwitz | 2006-07-10 03:18:57 +0200 (Mon, 10 Jul 2006) | 3 lines Patch #1516912: improve Modules support for OpenVMS. ........ r50506 | neal.norwitz | 2006-07-10 04:36:41 +0200 (Mon, 10 Jul 2006) | 7 lines Patch #1504046: Add documentation for xml.etree. /F wrote the text docs, Englebert Gruber massaged it to latex and I did some more massaging to try and improve the consistency and fix some name mismatches between the declaration and text. ........ r50509 | martin.v.loewis | 2006-07-10 09:23:48 +0200 (Mon, 10 Jul 2006) | 2 lines Introduce DISTUTILS_USE_SDK as a flag to determine whether the SDK environment should be used. Fixes #1508010. ........ r50510 | martin.v.loewis | 2006-07-10 09:26:41 +0200 (Mon, 10 Jul 2006) | 1 line Change error message to indicate that VS2003 is necessary to build extension modules, not the .NET SDK. ........ r50511 | martin.v.loewis | 2006-07-10 09:29:41 +0200 (Mon, 10 Jul 2006) | 1 line Add svn:ignore. ........ r50512 | anthony.baxter | 2006-07-10 09:41:04 +0200 (Mon, 10 Jul 2006) | 1 line preparing for 2.5b2 ........ r50513 | thomas.heller | 2006-07-10 11:10:28 +0200 (Mon, 10 Jul 2006) | 2 lines Fix bug #1518190: accept any integer or long value in the ctypes.c_void_p constructor. ........ r50514 | thomas.heller | 2006-07-10 11:31:06 +0200 (Mon, 10 Jul 2006) | 3 lines Fixed a segfault when ctypes.wintypes were imported on non-Windows machines. ........ r50516 | thomas.heller | 2006-07-10 13:11:10 +0200 (Mon, 10 Jul 2006) | 3 lines Assigning None to pointer type structure fields possible overwrote wrong fields. ........ r50517 | thomas.heller | 2006-07-10 13:17:37 +0200 (Mon, 10 Jul 2006) | 5 lines Moved the ctypes news entries from the 'Library' section into the 'Extension Modules' section where they belong, probably. This destroyes the original order of the news entries, don't know if that is important or not. ........ r50526 | phillip.eby | 2006-07-10 21:03:29 +0200 (Mon, 10 Jul 2006) | 2 lines Fix SF#1516184 and add a test to prevent regression. ........ r50528 | phillip.eby | 2006-07-10 21:18:35 +0200 (Mon, 10 Jul 2006) | 2 lines Fix SF#1457312: bad socket error handling in distutils "upload" command. ........ r50537 | peter.astrand | 2006-07-10 22:39:49 +0200 (Mon, 10 Jul 2006) | 1 line Make it possible to run test_subprocess.py with Python 2.2, which lacks test_support.reap_children(). ........ r50541 | tim.peters | 2006-07-10 23:08:24 +0200 (Mon, 10 Jul 2006) | 5 lines After approval from Anthony, merge the tim-current_frames branch into the trunk. This adds a new sys._current_frames() function, which returns a dict mapping thread id to topmost thread stack frame. ........ r50542 | tim.peters | 2006-07-10 23:11:49 +0200 (Mon, 10 Jul 2006) | 2 lines Whitespace normalization. ........ r50553 | martin.v.loewis | 2006-07-11 00:11:28 +0200 (Tue, 11 Jul 2006) | 4 lines Patch #1519566: Remove unused _tofill member. Make begin_fill idempotent. Update demo2 to demonstrate filling of concave shapes. ........ r50567 | anthony.baxter | 2006-07-11 04:04:09 +0200 (Tue, 11 Jul 2006) | 4 lines #1494314: Fix a regression with high-numbered sockets in 2.4.3. This means that select() on sockets > FD_SETSIZE (typically 1024) work again. The patch makes sockets use poll() internally where available. ........ r50568 | tim.peters | 2006-07-11 04:17:48 +0200 (Tue, 11 Jul 2006) | 2 lines Whitespace normalization. ........ r50575 | thomas.heller | 2006-07-11 18:42:05 +0200 (Tue, 11 Jul 2006) | 1 line Add missing Py_DECREF. ........ r50576 | thomas.heller | 2006-07-11 18:44:25 +0200 (Tue, 11 Jul 2006) | 1 line Add missing Py_DECREFs. ........ r50579 | andrew.kuchling | 2006-07-11 19:20:16 +0200 (Tue, 11 Jul 2006) | 1 line Bump version number; add sys._current_frames ........ r50582 | thomas.heller | 2006-07-11 20:28:35 +0200 (Tue, 11 Jul 2006) | 3 lines When a foreign function is retrived by calling __getitem__ on a ctypes library instance, do not set it as attribute. ........ r50583 | thomas.heller | 2006-07-11 20:40:50 +0200 (Tue, 11 Jul 2006) | 2 lines Change the ctypes version number to 1.0.0. ........ r50597 | neal.norwitz | 2006-07-12 07:26:17 +0200 (Wed, 12 Jul 2006) | 3 lines Bug #1520864: unpacking singleton tuples in for loop (for x, in) work again. ........ r50598 | neal.norwitz | 2006-07-12 07:26:35 +0200 (Wed, 12 Jul 2006) | 1 line Fix function name in error msg ........ r50599 | neal.norwitz | 2006-07-12 07:27:46 +0200 (Wed, 12 Jul 2006) | 4 lines Fix uninitialized memory read reported by Valgrind when running doctest. This could happen if size == 0. ........ r50600 | neal.norwitz | 2006-07-12 09:28:29 +0200 (Wed, 12 Jul 2006) | 1 line Actually change the MAGIC #. Create a new section for 2.5c1 and mention the impact of changing the MAGIC #. ........ r50601 | thomas.heller | 2006-07-12 10:43:47 +0200 (Wed, 12 Jul 2006) | 3 lines Fix #1467450: ctypes now uses RTLD_GLOBAL by default on OSX 10.3 to load shared libraries. ........ r50604 | thomas.heller | 2006-07-12 16:25:18 +0200 (Wed, 12 Jul 2006) | 3 lines Fix the wrong description of LibraryLoader.LoadLibrary, and document the DEFAULT_MODE constant. ........ r50607 | georg.brandl | 2006-07-12 17:31:17 +0200 (Wed, 12 Jul 2006) | 3 lines Accept long options "--help" and "--version". ........ r50617 | thomas.heller | 2006-07-13 11:53:47 +0200 (Thu, 13 Jul 2006) | 3 lines A misspelled preprocessor symbol caused ctypes to be always compiled without thread support. Replaced WITH_THREADS with WITH_THREAD. ........ r50619 | thomas.heller | 2006-07-13 19:01:14 +0200 (Thu, 13 Jul 2006) | 3 lines Fix #1521375. When running with root priviledges, 'gcc -o /dev/null' did overwrite /dev/null. Use a temporary file instead of /dev/null. ........ r50620 | thomas.heller | 2006-07-13 19:05:13 +0200 (Thu, 13 Jul 2006) | 2 lines Fix misleading words. ........ r50622 | andrew.kuchling | 2006-07-13 19:37:26 +0200 (Thu, 13 Jul 2006) | 1 line Typo fix ........ r50629 | georg.brandl | 2006-07-14 09:12:54 +0200 (Fri, 14 Jul 2006) | 3 lines Patch #1521874: grammar errors in doanddont.tex. ........ r50630 | neal.norwitz | 2006-07-14 09:20:04 +0200 (Fri, 14 Jul 2006) | 1 line Try to improve grammar further. ........ r50631 | martin.v.loewis | 2006-07-14 11:58:55 +0200 (Fri, 14 Jul 2006) | 1 line Extend build_ssl to Win64, using VSExtComp. ........ r50632 | martin.v.loewis | 2006-07-14 14:10:09 +0200 (Fri, 14 Jul 2006) | 1 line Add debug output to analyse buildbot failure. ........ r50633 | martin.v.loewis | 2006-07-14 14:31:05 +0200 (Fri, 14 Jul 2006) | 1 line Fix Debug build of _ssl. ........ r50636 | andrew.kuchling | 2006-07-14 15:32:38 +0200 (Fri, 14 Jul 2006) | 1 line Mention new options ........ r50638 | peter.astrand | 2006-07-14 16:04:45 +0200 (Fri, 14 Jul 2006) | 1 line Bug #1223937: CalledProcessError.errno -> CalledProcessError.returncode. ........ r50640 | thomas.heller | 2006-07-14 17:01:05 +0200 (Fri, 14 Jul 2006) | 4 lines Make the prototypes of our private PyUnicode_FromWideChar and PyUnicode_AsWideChar replacement functions compatible to the official functions by using Py_ssize_t instead of int. ........ r50643 | thomas.heller | 2006-07-14 19:51:14 +0200 (Fri, 14 Jul 2006) | 3 lines Patch #1521817: The index range checking on ctypes arrays containing exactly one element is enabled again. ........ r50647 | thomas.heller | 2006-07-14 20:22:50 +0200 (Fri, 14 Jul 2006) | 2 lines Updates for the ctypes documentation. ........ r50655 | fredrik.lundh | 2006-07-14 23:45:48 +0200 (Fri, 14 Jul 2006) | 3 lines typo ........ r50664 | george.yoshida | 2006-07-15 18:03:49 +0200 (Sat, 15 Jul 2006) | 2 lines Bug #15187702 : ext/win-cookbook.html has a broken link to distutils ........ r50667 | bob.ippolito | 2006-07-15 18:53:15 +0200 (Sat, 15 Jul 2006) | 1 line Patch #1220874: Update the binhex module for Mach-O. ........ r50671 | fred.drake | 2006-07-16 03:21:20 +0200 (Sun, 16 Jul 2006) | 1 line clean up some link markup ........ r50673 | neal.norwitz | 2006-07-16 03:50:38 +0200 (Sun, 16 Jul 2006) | 4 lines Bug #1512814, Fix incorrect lineno's when code within a function had more than 255 blank lines. Byte codes need to go first, line #s second. ........ r50674 | neal.norwitz | 2006-07-16 04:00:32 +0200 (Sun, 16 Jul 2006) | 5 lines a & b were dereffed above, so they are known to be valid pointers. z is known to be NULL, nothing to DECREF. Reported by Klockwork, #107. ........ r50675 | neal.norwitz | 2006-07-16 04:02:57 +0200 (Sun, 16 Jul 2006) | 5 lines self is dereffed (and passed as first arg), so it's known to be good. func is returned from PyArg_ParseTuple and also dereffed. Reported by Klocwork, #30 (self one at least). ........ r50676 | neal.norwitz | 2006-07-16 04:05:35 +0200 (Sun, 16 Jul 2006) | 4 lines proto was dereffed above and is known to be good. No need for X. Reported by Klocwork, #39. ........ r50677 | neal.norwitz | 2006-07-16 04:15:27 +0200 (Sun, 16 Jul 2006) | 5 lines Fix memory leaks in some conditions. Reported by Klocwork #152. ........ r50678 | neal.norwitz | 2006-07-16 04:17:36 +0200 (Sun, 16 Jul 2006) | 4 lines Fix memory leak under some conditions. Reported by Klocwork, #98. ........ r50679 | neal.norwitz | 2006-07-16 04:22:30 +0200 (Sun, 16 Jul 2006) | 8 lines Use sizeof(buffer) instead of duplicating the constants to ensure they won't be wrong. The real change is to pass (bufsz - 1) to PyOS_ascii_formatd and 1 to strncat. strncat copies n+1 bytes from src (not dest). Reported by Klocwork #58. ........ r50680 | neal.norwitz | 2006-07-16 04:32:03 +0200 (Sun, 16 Jul 2006) | 5 lines Handle a NULL name properly. Reported by Klocwork #67 ........ r50681 | neal.norwitz | 2006-07-16 04:35:47 +0200 (Sun, 16 Jul 2006) | 6 lines PyFunction_SetDefaults() is documented as taking None or a tuple. A NULL would crash the PyTuple_Check(). Now make NULL return a SystemError. Reported by Klocwork #73. ........ r50683 | neal.norwitz | 2006-07-17 02:55:45 +0200 (Mon, 17 Jul 2006) | 5 lines Stop INCREFing name, then checking if it's NULL. name (f_name) should never be NULL so assert it. Fix one place where we could have passed NULL. Reported by Klocwork #66. ........ r50684 | neal.norwitz | 2006-07-17 02:57:15 +0200 (Mon, 17 Jul 2006) | 5 lines otherset is known to be non-NULL based on checks before and DECREF after. DECREF otherset rather than XDECREF in error conditions too. Reported by Klockwork #154. ........ r50685 | neal.norwitz | 2006-07-17 02:59:04 +0200 (Mon, 17 Jul 2006) | 7 lines Reported by Klocwork #151. v2 can be NULL if exception2 is NULL. I don't think that condition can happen, but I'm not sure it can't either. Now the code will protect against either being NULL. ........ r50686 | neal.norwitz | 2006-07-17 03:00:16 +0200 (Mon, 17 Jul 2006) | 1 line Add NEWS entry for a bunch of fixes due to warnings produced by Klocworks static analysis tool. ........ r50687 | fred.drake | 2006-07-17 07:47:52 +0200 (Mon, 17 Jul 2006) | 3 lines document xmlcore (still minimal; needs mention in each of the xml.* modules) SF bug #1504456 (partial) ........ r50688 | georg.brandl | 2006-07-17 15:23:46 +0200 (Mon, 17 Jul 2006) | 3 lines Remove usage of sets module (patch #1500609). ........ r50689 | georg.brandl | 2006-07-17 15:26:33 +0200 (Mon, 17 Jul 2006) | 3 lines Add missing NEWS item (#1522771) ........ r50690 | andrew.kuchling | 2006-07-17 18:47:54 +0200 (Mon, 17 Jul 2006) | 1 line Attribute more features ........ r50692 | kurt.kaiser | 2006-07-17 23:59:27 +0200 (Mon, 17 Jul 2006) | 8 lines Patch 1479219 - Tal Einat 1. 'as' highlighted as builtin in comment string on import line 2. Comments such as "#False identity" which start with a keyword immediately after the '#' character aren't colored as comments. 3. u or U beginning unicode string not correctly highlighted Closes bug 1325071 ........ r50693 | barry.warsaw | 2006-07-18 01:07:51 +0200 (Tue, 18 Jul 2006) | 16 lines decode_rfc2231(): Be more robust against buggy RFC 2231 encodings. Specifically, instead of raising a ValueError when there is a single tick in the parameter, simply return that the entire string unquoted, with None for both the charset and the language. Also, if there are more than 2 ticks in the parameter, interpret the first three parts as the standard RFC 2231 parts, then the rest of the parts as the encoded string. Test cases added. Original fewer-than-3-parts fix by Tokio Kikuchi. Resolves SF bug # 1218081. I will back port the fix and tests to Python 2.4 (email 3.0) and Python 2.3 (email 2.5). Also, bump the version number to email 4.0.1, removing the 'alpha' moniker. ........ r50695 | kurt.kaiser | 2006-07-18 06:03:16 +0200 (Tue, 18 Jul 2006) | 2 lines Rebinding Tab key was inserting 'tab' instead of 'Tab'. Bug 1179168. ........ r50696 | brett.cannon | 2006-07-18 06:41:36 +0200 (Tue, 18 Jul 2006) | 6 lines Fix bug #1520914. Starting in 2.4, time.strftime() began to check the bounds of values in the time tuple passed in. Unfortunately people came to rely on undocumented behaviour of setting unneeded values to 0, regardless of if it was within the valid range. Now those values force the value internally to the minimum value when 0 is passed in. ........ r50697 | facundo.batista | 2006-07-18 14:16:13 +0200 (Tue, 18 Jul 2006) | 1 line Comments and docs cleanups, and some little fixes, provided by Santiágo Peresón ........ r50704 | martin.v.loewis | 2006-07-18 19:46:31 +0200 (Tue, 18 Jul 2006) | 2 lines Patch #1524429: Use repr instead of backticks again. ........ r50706 | tim.peters | 2006-07-18 23:55:15 +0200 (Tue, 18 Jul 2006) | 2 lines Whitespace normalization. ........ r50708 | tim.peters | 2006-07-19 02:03:19 +0200 (Wed, 19 Jul 2006) | 18 lines SF bug 1524317: configure --without-threads fails to build Moved the code for _PyThread_CurrentFrames() up, so it's no longer in a huge "#ifdef WITH_THREAD" block (I didn't realize it /was/ in one). Changed test_sys's test_current_frames() so it passes with or without thread supported compiled in. Note that test_sys fails when Python is compiled without threads, but for an unrelated reason (the old test_exit() fails with an indirect ImportError on the `thread` module). There are also other unrelated compilation failures without threads, in extension modules (like ctypes); at least the core compiles again. Do we really support --without-threads? If so, there are several problems remaining. ........ r50713 | thomas.heller | 2006-07-19 11:09:32 +0200 (Wed, 19 Jul 2006) | 4 lines Make sure the _ctypes extension can be compiled when WITH_THREAD is not defined on Windows, even if that configuration is probably not supported at all. ........ r50715 | martin.v.loewis | 2006-07-19 19:18:32 +0200 (Wed, 19 Jul 2006) | 4 lines Revert r50706 (Whitespace normalization) and r50697: Comments and docs cleanups, and some little fixes per recommendation from Raymond Hettinger. ........ r50719 | phillip.eby | 2006-07-20 17:54:16 +0200 (Thu, 20 Jul 2006) | 4 lines Fix SF#1516184 (again) and add a test to prevent regression. (There was a problem with empty filenames still causing recursion) ........ r50720 | georg.brandl | 2006-07-20 18:28:39 +0200 (Thu, 20 Jul 2006) | 3 lines Guard for _active being None in __del__ method. ........ r50721 | vinay.sajip | 2006-07-20 18:28:39 +0200 (Thu, 20 Jul 2006) | 1 line Updated documentation for TimedRotatingFileHandler relating to how rollover files are named. The previous documentation was wrongly the same as for RotatingFileHandler. ........ r50731 | fred.drake | 2006-07-20 22:11:57 +0200 (Thu, 20 Jul 2006) | 1 line markup fix ........ r50739 | kurt.kaiser | 2006-07-21 00:22:52 +0200 (Fri, 21 Jul 2006) | 7 lines Avoid occasional failure to detect closing paren properly. Patch 1407280 Tal Einat M ParenMatch.py M NEWS.txt M CREDITS.txt ........ r50740 | vinay.sajip | 2006-07-21 01:20:12 +0200 (Fri, 21 Jul 2006) | 1 line Addressed SF#1524081 by using a dictionary to map level names to syslog priority names, rather than a string.lower(). ........ r50741 | neal.norwitz | 2006-07-21 07:29:58 +0200 (Fri, 21 Jul 2006) | 1 line Add some asserts that we got good params passed ........ r50742 | neal.norwitz | 2006-07-21 07:31:02 +0200 (Fri, 21 Jul 2006) | 5 lines Move the initialization of some pointers earlier. The problem is that if we call Py_DECREF(frame) like we do if allocating locals fails, frame_dealloc() will try to use these bogus values and crash. ........ r50743 | neal.norwitz | 2006-07-21 07:32:28 +0200 (Fri, 21 Jul 2006) | 4 lines Handle allocation failures gracefully. Found with failmalloc. Many (all?) of these could be backported. ........ r50745 | neal.norwitz | 2006-07-21 09:59:02 +0200 (Fri, 21 Jul 2006) | 1 line Speel initialise write. Tanks Anthony. ........ r50746 | neal.norwitz | 2006-07-21 09:59:47 +0200 (Fri, 21 Jul 2006) | 2 lines Handle more memory allocation failures without crashing. ........ r50754 | barry.warsaw | 2006-07-21 16:51:07 +0200 (Fri, 21 Jul 2006) | 23 lines More RFC 2231 improvements for the email 4.0 package. As Mark Sapiro rightly points out there are really two types of continued headers defined in this RFC (i.e. "encoded" parameters with the form "name*0*=" and unencoded parameters with the form "name*0="), but we were were handling them both the same way and that isn't correct. This patch should be much more RFC compliant in that only encoded params are %-decoded and the charset/language information is only extract if there are any encoded params in the segments. If there are no encoded params then the RFC says that there will be no charset/language parts. Note however that this will change the return value for Message.get_param() in some cases. For example, whereas before if you had all unencoded param continuations you would have still gotten a 3-tuple back from this method (with charset and language == None), you will now get just a string. I don't believe this is a backward incompatible change though because the documentation for this method already indicates that either return value is possible and that you must do an isinstance(val, tuple) check to discriminate between the two. (Yeah that API kind of sucks but we can't change /that/ without breaking code.) Test cases, some documentation updates, and a NEWS item accompany this patch. ........ r50759 | georg.brandl | 2006-07-21 19:36:31 +0200 (Fri, 21 Jul 2006) | 3 lines Fix check for empty list (vs. None). ........ r50771 | brett.cannon | 2006-07-22 00:44:07 +0200 (Sat, 22 Jul 2006) | 2 lines Remove an XXX marker in a comment. ........ r50773 | neal.norwitz | 2006-07-22 18:20:49 +0200 (Sat, 22 Jul 2006) | 1 line Fix more memory allocation issues found with failmalloc. ........ r50774 | neal.norwitz | 2006-07-22 19:00:57 +0200 (Sat, 22 Jul 2006) | 1 line Don't fail if the directory already exists ........ r50775 | greg.ward | 2006-07-23 04:25:53 +0200 (Sun, 23 Jul 2006) | 6 lines Be a lot smarter about whether this test passes: instead of assuming that a 2.93 sec audio file will always take 3.1 sec (as it did on the hardware I had when I first wrote the test), expect that it will take 2.93 sec +/- 10%, and only fail if it's outside of that range. Compute the expected ........ r50776 | kurt.kaiser | 2006-07-23 06:19:49 +0200 (Sun, 23 Jul 2006) | 2 lines Tooltips failed on new-syle class __init__ args. Bug 1027566 Loren Guthrie ........ r50777 | neal.norwitz | 2006-07-23 09:50:36 +0200 (Sun, 23 Jul 2006) | 1 line Handle more mem alloc issues found with failmalloc ........ r50778 | neal.norwitz | 2006-07-23 09:51:58 +0200 (Sun, 23 Jul 2006) | 5 lines If the for loop isn't entered, entryblock will be NULL. If passed to stackdepth_walk it will be dereffed. Not sure if I found with failmalloc or Klockwork #55. ........ r50779 | neal.norwitz | 2006-07-23 09:53:14 +0200 (Sun, 23 Jul 2006) | 4 lines Move the initialization of size_a down below the check for a being NULL. Reported by Klocwork #106 ........ r50780 | neal.norwitz | 2006-07-23 09:55:55 +0200 (Sun, 23 Jul 2006) | 9 lines Check the allocation of b_objects and return if there was a failure. Also fix a few memory leaks in other failure scenarios. It seems that if b_objects == Py_None, we will have an extra ref to b_objects. Add XXX comment so hopefully someone documents why the else isn't necessary or adds it in. Reported by Klocwork #20 ........ r50781 | neal.norwitz | 2006-07-23 09:57:11 +0200 (Sun, 23 Jul 2006) | 2 lines Fix memory leaks spotted by Klocwork #37. ........ r50782 | neal.norwitz | 2006-07-23 09:59:00 +0200 (Sun, 23 Jul 2006) | 5 lines nextlink can be NULL if teedataobject_new fails, so use XINCREF. Ensure that dataobj is never NULL. Reported by Klocwork #102 ........ r50783 | neal.norwitz | 2006-07-23 10:01:43 +0200 (Sun, 23 Jul 2006) | 8 lines Ensure we don't write beyond errText. I think I got this right, but it definitely could use some review to ensure I'm not off by one and there's no possible overflow/wrap-around of bytes_left. Reported by Klocwork #1. Fix a problem if there is a failure allocating self->db. Found with failmalloc. ........ r50784 | ronald.oussoren | 2006-07-23 11:41:09 +0200 (Sun, 23 Jul 2006) | 3 lines Without this patch CMD-W won't close EditorWindows on MacOS X. This solves part of bug #1517990. ........ r50785 | ronald.oussoren | 2006-07-23 11:46:11 +0200 (Sun, 23 Jul 2006) | 5 lines Fix for bug #1517996: Class and Path browsers show Tk menu This patch replaces the menubar that is used by AquaTk for windows without a menubar of their own by one that is more appropriate for IDLE. ........ r50786 | andrew.macintyre | 2006-07-23 14:57:02 +0200 (Sun, 23 Jul 2006) | 2 lines Build updates for OS/2 EMX port ........ r50787 | andrew.macintyre | 2006-07-23 15:00:04 +0200 (Sun, 23 Jul 2006) | 3 lines bugfix: PyThread_start_new_thread() returns the thread ID, not a flag; will backport. ........ r50789 | andrew.macintyre | 2006-07-23 15:04:00 +0200 (Sun, 23 Jul 2006) | 2 lines Get mailbox module working on OS/2 EMX port. ........ r50791 | greg.ward | 2006-07-23 18:05:51 +0200 (Sun, 23 Jul 2006) | 1 line Resync optparse with Optik 1.5.3: minor tweaks for/to tests. ........ r50794 | martin.v.loewis | 2006-07-24 07:05:22 +0200 (Mon, 24 Jul 2006) | 2 lines Update list of unsupported systems. Fixes #1510853. ........ r50795 | martin.v.loewis | 2006-07-24 12:26:33 +0200 (Mon, 24 Jul 2006) | 1 line Patch #1448199: Release GIL around ConnectRegistry. ........ r50796 | martin.v.loewis | 2006-07-24 13:54:53 +0200 (Mon, 24 Jul 2006) | 3 lines Patch #1232023: Don't include empty path component from registry, so that the current directory does not get added to sys.path. Also fixes #1526785. ........ r50797 | martin.v.loewis | 2006-07-24 14:54:17 +0200 (Mon, 24 Jul 2006) | 3 lines Bug #1524310: Properly report errors from FindNextFile in os.listdir. Will backport to 2.4. ........ r50800 | georg.brandl | 2006-07-24 15:28:57 +0200 (Mon, 24 Jul 2006) | 7 lines Patch #1523356: fix determining include dirs in python-config. Also don't install "python-config" when doing altinstall, but always install "python-config2.x" and make a link to it like with the main executable. ........ r50802 | georg.brandl | 2006-07-24 15:46:47 +0200 (Mon, 24 Jul 2006) | 3 lines Patch #1527744: right order of includes in order to have HAVE_CONIO_H defined properly. ........ r50803 | georg.brandl | 2006-07-24 16:09:56 +0200 (Mon, 24 Jul 2006) | 3 lines Patch #1515343: Fix printing of deprecated string exceptions with a value in the traceback module. ........ r50804 | kurt.kaiser | 2006-07-24 19:13:23 +0200 (Mon, 24 Jul 2006) | 7 lines EditorWindow failed when used stand-alone if sys.ps1 not set. Bug 1010370 Dave Florek M EditorWindow.py M PyShell.py M NEWS.txt ........ r50805 | kurt.kaiser | 2006-07-24 20:05:51 +0200 (Mon, 24 Jul 2006) | 6 lines - EditorWindow.test() was failing. Bug 1417598 M EditorWindow.py M ScriptBinding.py M NEWS.txt ........ r50808 | georg.brandl | 2006-07-24 22:11:35 +0200 (Mon, 24 Jul 2006) | 3 lines Repair accidental NameError. ........ r50809 | tim.peters | 2006-07-24 23:02:15 +0200 (Mon, 24 Jul 2006) | 2 lines Whitespace normalization. ........ r50810 | greg.ward | 2006-07-25 04:11:12 +0200 (Tue, 25 Jul 2006) | 3 lines Don't use standard assert: want tests to fail even when run with -O. Delete cruft. ........ r50811 | tim.peters | 2006-07-25 06:07:22 +0200 (Tue, 25 Jul 2006) | 10 lines current_frames_with_threads(): There's actually no way to guess /which/ line the spawned thread is in at the time sys._current_frames() is called: we know it finished enter_g.set(), but can't know whether the instruction counter has advanced to the following leave_g.wait(). The latter is overwhelming most likely, but not guaranteed, and I see that the "x86 Ubuntu dapper (icc) trunk" buildbot found it on the other line once. Changed the test so it passes in either case. ........ r50815 | martin.v.loewis | 2006-07-25 11:53:12 +0200 (Tue, 25 Jul 2006) | 2 lines Bug #1525817: Don't truncate short lines in IDLE's tool tips. ........ r50816 | martin.v.loewis | 2006-07-25 12:05:47 +0200 (Tue, 25 Jul 2006) | 3 lines Bug #978833: Really close underlying socket in _socketobject.close. Will backport to 2.4. ........ r50817 | martin.v.loewis | 2006-07-25 12:11:14 +0200 (Tue, 25 Jul 2006) | 1 line Revert incomplete checkin. ........ r50819 | georg.brandl | 2006-07-25 12:22:34 +0200 (Tue, 25 Jul 2006) | 4 lines Patch #1525766: correctly pass onerror arg to recursive calls of pkg.walk_packages. Also improve the docstrings. ........ r50825 | brett.cannon | 2006-07-25 19:32:20 +0200 (Tue, 25 Jul 2006) | 2 lines Add comment for changes to test_ossaudiodev. ........ r50826 | brett.cannon | 2006-07-25 19:34:36 +0200 (Tue, 25 Jul 2006) | 3 lines Fix a bug in the messages for an assert failure where not enough arguments to a string were being converted in the format. ........ r50828 | armin.rigo | 2006-07-25 20:09:57 +0200 (Tue, 25 Jul 2006) | 2 lines Document why is and is not a good way to fix the gc_inspection crasher. ........ r50829 | armin.rigo | 2006-07-25 20:11:07 +0200 (Tue, 25 Jul 2006) | 5 lines Added another crasher, which hit me today (I was not intentionally writing such code, of course, but it took some gdb time to figure out what my bug was). ........ r50830 | armin.rigo | 2006-07-25 20:38:39 +0200 (Tue, 25 Jul 2006) | 3 lines Document the crashers that will not go away soon as "won't fix", and explain why. ........ r50831 | ronald.oussoren | 2006-07-25 21:13:35 +0200 (Tue, 25 Jul 2006) | 3 lines Install the compatibility symlink to libpython.a on OSX using 'ln -sf' instead of 'ln -s', this avoid problems when reinstalling python. ........ r50832 | ronald.oussoren | 2006-07-25 21:20:54 +0200 (Tue, 25 Jul 2006) | 7 lines Fix for bug #1525447 (renaming to MacOSmodule.c would also work, but not without causing problems for anyone that is on a case-insensitive filesystem). Setup.py tries to compile the MacOS extension from MacOSmodule.c, while the actual file is named macosmodule.c. This is no problem on the (default) case-insensitive filesystem, but doesn't work on case-sensitive filesystems. ........ r50833 | ronald.oussoren | 2006-07-25 22:28:55 +0200 (Tue, 25 Jul 2006) | 7 lines Fix bug #1517990: IDLE keybindings on OSX This adds a new key definition for OSX, which is slightly different from the classic mac definition. Also add NEWS item for a couple of bugfixes I added recently. ........ r50834 | tim.peters | 2006-07-26 00:30:24 +0200 (Wed, 26 Jul 2006) | 2 lines Whitespace normalization. ........ r50839 | neal.norwitz | 2006-07-26 06:00:18 +0200 (Wed, 26 Jul 2006) | 1 line Hmm, only python2.x is installed, not plain python. Did that change recently? ........ r50840 | barry.warsaw | 2006-07-26 07:54:46 +0200 (Wed, 26 Jul 2006) | 6 lines Forward port some fixes that were in email 2.5 but for some reason didn't make it into email 4.0. Specifically, in Message.get_content_charset(), handle RFC 2231 headers that contain an encoding not known to Python, or a character in the data that isn't in the charset encoding. Also forward port the appropriate unit tests. ........ r50841 | georg.brandl | 2006-07-26 09:23:32 +0200 (Wed, 26 Jul 2006) | 3 lines NEWS entry for #1525766. ........ r50842 | georg.brandl | 2006-07-26 09:40:17 +0200 (Wed, 26 Jul 2006) | 3 lines Bug #1459963: properly capitalize HTTP header names. ........ r50843 | georg.brandl | 2006-07-26 10:03:10 +0200 (Wed, 26 Jul 2006) | 6 lines Part of bug #1523610: fix miscalculation of buffer length. Also add a guard against NULL in converttuple and add a test case (that previously would have crashed). ........ r50844 | martin.v.loewis | 2006-07-26 14:12:56 +0200 (Wed, 26 Jul 2006) | 3 lines Bug #978833: Really close underlying socket in _socketobject.close. Fix httplib.HTTPConnection.getresponse to not close the socket if it is still needed for the response. ........ r50845 | andrew.kuchling | 2006-07-26 19:16:52 +0200 (Wed, 26 Jul 2006) | 1 line [Bug #1471938] Fix build problem on Solaris 8 by conditionalizing the use of mvwgetnstr(); it was conditionalized a few lines below. Fix from Paul Eggert. I also tried out the STRICT_SYSV_CURSES case and am therefore removing the 'untested' comment. ........ r50846 | andrew.kuchling | 2006-07-26 19:18:01 +0200 (Wed, 26 Jul 2006) | 1 line Correct error message ........ r50847 | andrew.kuchling | 2006-07-26 19:19:39 +0200 (Wed, 26 Jul 2006) | 1 line Minor grammar fix ........ r50848 | andrew.kuchling | 2006-07-26 19:22:21 +0200 (Wed, 26 Jul 2006) | 1 line Put news item in right section ........ r50850 | andrew.kuchling | 2006-07-26 20:03:12 +0200 (Wed, 26 Jul 2006) | 1 line Use sys.exc_info() ........ r50851 | andrew.kuchling | 2006-07-26 20:15:45 +0200 (Wed, 26 Jul 2006) | 1 line Use sys.exc_info() ........ r50852 | phillip.eby | 2006-07-26 21:48:27 +0200 (Wed, 26 Jul 2006) | 4 lines Allow the 'onerror' argument to walk_packages() to catch any Exception, not just ImportError. This allows documentation tools to better skip unimportable packages. ........ r50854 | tim.peters | 2006-07-27 01:23:15 +0200 (Thu, 27 Jul 2006) | 2 lines Whitespace normalization. ........ r50855 | tim.peters | 2006-07-27 03:14:53 +0200 (Thu, 27 Jul 2006) | 21 lines Bug #1521947: possible bug in mystrtol.c with recent gcc. In general, C doesn't define anything about what happens when an operation on a signed integral type overflows, and PyOS_strtol() did several formally undefined things of that nature on signed longs. Some version of gcc apparently tries to exploit that now, and PyOS_strtol() could fail to detect overflow then. Tried to repair all that, although it seems at least as likely to me that we'll get screwed by bad platform definitions for LONG_MIN and/or LONG_MAX now. For that reason, I don't recommend backporting this. Note that I have no box on which this makes a lick of difference -- can't really test it, except to note that it didn't break anything on my boxes. Silent change: PyOS_strtol() used to return the hard-coded 0x7fffffff in case of overflow. Now it returns LONG_MAX. They're the same only on 32-bit boxes (although C doesn't guarantee that either ...). ........ r50856 | neal.norwitz | 2006-07-27 05:51:58 +0200 (Thu, 27 Jul 2006) | 6 lines Don't kill a normal instance of python running on windows when checking to kill a cygwin instance. build\\python.exe was matching a normal windows instance. Prefix that with a \\ to ensure build is a directory and not PCbuild. As discussed on python-dev. ........ r50857 | neal.norwitz | 2006-07-27 05:55:39 +0200 (Thu, 27 Jul 2006) | 5 lines Closure can't be NULL at this point since we know it's a tuple. Reported by Klocwork # 74. ........ r50858 | neal.norwitz | 2006-07-27 06:04:50 +0200 (Thu, 27 Jul 2006) | 1 line No functional change. Add comment and assert to describe why there cannot be overflow which was reported by Klocwork. Discussed on python-dev ........ r50859 | martin.v.loewis | 2006-07-27 08:38:16 +0200 (Thu, 27 Jul 2006) | 3 lines Bump distutils version to 2.5, as several new features have been introduced since 2.4. ........ r50860 | andrew.kuchling | 2006-07-27 14:18:20 +0200 (Thu, 27 Jul 2006) | 1 line Reformat docstring; fix typo ........ r50861 | georg.brandl | 2006-07-27 17:05:36 +0200 (Thu, 27 Jul 2006) | 6 lines Add test_main() methods. These three tests were never run by regrtest.py. We really need a simpler testing framework. ........ r50862 | tim.peters | 2006-07-27 17:09:20 +0200 (Thu, 27 Jul 2006) | 2 lines News for patch #1529686. ........ r50863 | tim.peters | 2006-07-27 17:11:00 +0200 (Thu, 27 Jul 2006) | 2 lines Whitespace normalization. ........ r50864 | georg.brandl | 2006-07-27 17:38:33 +0200 (Thu, 27 Jul 2006) | 3 lines Amend news entry. ........ r50865 | georg.brandl | 2006-07-27 18:08:15 +0200 (Thu, 27 Jul 2006) | 3 lines Make uuid test suite pass on this box by requesting output with LC_ALL=C. ........ r50866 | andrew.kuchling | 2006-07-27 20:37:33 +0200 (Thu, 27 Jul 2006) | 1 line Add example ........ r50867 | thomas.heller | 2006-07-27 20:39:55 +0200 (Thu, 27 Jul 2006) | 9 lines Remove code that is no longer used (ctypes.com). Fix the DllGetClassObject and DllCanUnloadNow so that they forward the call to the comtypes.server.inprocserver module. The latter was never documented, never used by published code, and didn't work anyway, so I think it does not deserve a NEWS entry (but I might be wrong). ........ r50868 | andrew.kuchling | 2006-07-27 20:41:21 +0200 (Thu, 27 Jul 2006) | 1 line Typo fix ('publically' is rare, poss. non-standard) ........ r50869 | andrew.kuchling | 2006-07-27 20:42:41 +0200 (Thu, 27 Jul 2006) | 1 line Add missing word ........ r50870 | andrew.kuchling | 2006-07-27 20:44:10 +0200 (Thu, 27 Jul 2006) | 1 line Repair typos ........ r50872 | andrew.kuchling | 2006-07-27 20:53:33 +0200 (Thu, 27 Jul 2006) | 1 line Update URL; add example ........ r50873 | andrew.kuchling | 2006-07-27 21:07:29 +0200 (Thu, 27 Jul 2006) | 1 line Add punctuation mark; add some examples ........ r50874 | andrew.kuchling | 2006-07-27 21:11:07 +0200 (Thu, 27 Jul 2006) | 1 line Mention base64 module; rewrite last sentence to be more positive ........ r50875 | andrew.kuchling | 2006-07-27 21:12:49 +0200 (Thu, 27 Jul 2006) | 1 line If binhex is higher-level than binascii, it should come first in the chapter ........ r50876 | tim.peters | 2006-07-27 22:47:24 +0200 (Thu, 27 Jul 2006) | 28 lines check_node(): stop spraying mystery output to stderr. When a node number disagrees, keep track of all sources & the node numbers they reported, and stick all that in the error message. Changed all callers to supply a non-empty "source" argument; made the "source" argument non-optional. On my box, test_uuid still fails, but with the less confusing output: AssertionError: different sources disagree on node: from source 'getnode1', node was 00038a000015 from source 'getnode2', node was 00038a000015 from source 'ipconfig', node was 001111b2b7bf Only the last one appears to be correct; e.g., C:\Code\python\PCbuild>getmac Physical Address Transport Name =================== ========================================================== 00-11-11-B2-B7-BF \Device\Tcpip_{190FB163-5AFD-4483-86A1-2FE16AC61FF1} 62-A1-AC-6C-FD-BE \Device\Tcpip_{8F77DF5A-EA3D-4F1D-975E-D472CEE6438A} E2-1F-01-C6-5D-88 \Device\Tcpip_{CD18F76B-2EF3-409F-9B8A-6481EE70A1E4} I can't find anything on my box with MAC 00-03-8a-00-00-15, and am not clear on where that comes from. ........ r50878 | andrew.kuchling | 2006-07-28 00:40:05 +0200 (Fri, 28 Jul 2006) | 1 line Reword paragraph ........ r50879 | andrew.kuchling | 2006-07-28 00:49:38 +0200 (Fri, 28 Jul 2006) | 1 line Add example ........ r50880 | andrew.kuchling | 2006-07-28 00:49:54 +0200 (Fri, 28 Jul 2006) | 1 line Add example ........ r50881 | barry.warsaw | 2006-07-28 01:43:15 +0200 (Fri, 28 Jul 2006) | 27 lines Patch #1520294: Support for getset and member descriptors in types.py, inspect.py, and pydoc.py. Specifically, this allows for querying the type of an object against these built-in C types and more importantly, for getting their docstrings printed in the interactive interpreter's help() function. This patch includes a new built-in module called _types which provides definitions of getset and member descriptors for use by the types.py module. These types are exposed as types.GetSetDescriptorType and types.MemberDescriptorType. Query functions are provided as inspect.isgetsetdescriptor() and inspect.ismemberdescriptor(). The implementations of these are robust enough to work with Python implementations other than CPython, which may not have these fundamental types. The patch also includes documentation and test suite updates. I commit these changes now under these guiding principles: 1. Silence is assent. The release manager has not said "no", and of the few people that cared enough to respond to the thread, the worst vote was "0". 2. It's easier to ask for forgiveness than permission. 3. It's so dang easy to revert stuff in svn, that you could view this as a forcing function. :) Windows build patches will follow. ........ r50882 | tim.peters | 2006-07-28 01:44:37 +0200 (Fri, 28 Jul 2006) | 4 lines Bug #1529297: The rewrite of doctest for Python 2.4 unintentionally lost that tests are sorted by name before being run. ``DocTestFinder`` has been changed to sort the list of tests it returns. ........ r50883 | tim.peters | 2006-07-28 01:45:48 +0200 (Fri, 28 Jul 2006) | 2 lines Whitespace normalization. ........ r50884 | tim.peters | 2006-07-28 01:46:36 +0200 (Fri, 28 Jul 2006) | 2 lines Add missing svn:eol-style property to text files. ........ r50885 | barry.warsaw | 2006-07-28 01:50:40 +0200 (Fri, 28 Jul 2006) | 4 lines Enable the building of the _types module on Windows. Note that this has only been tested for VS 2003 since that's all I have. ........ r50887 | tim.peters | 2006-07-28 02:23:15 +0200 (Fri, 28 Jul 2006) | 7 lines defdict_reduce(): Plug leaks. We didn't notice these before because test_defaultdict didn't actually do anything before Georg fixed that earlier today. Neal's next refleak run then showed test_defaultdict leaking 9 references on each run. That's repaired by this checkin. ........ r50888 | tim.peters | 2006-07-28 02:30:00 +0200 (Fri, 28 Jul 2006) | 2 lines News about the repaired memory leak in defaultdict. ........ r50889 | gregory.p.smith | 2006-07-28 03:35:25 +0200 (Fri, 28 Jul 2006) | 7 lines - pybsddb Bug #1527939: bsddb module DBEnv dbremove and dbrename methods now allow their database parameter to be None as the sleepycat API allows. Also adds an appropriate test case for DBEnv.dbrename and dbremove. ........ r50895 | neal.norwitz | 2006-07-28 06:22:34 +0200 (Fri, 28 Jul 2006) | 1 line Ensure the actual number matches the expected count ........ r50896 | tim.peters | 2006-07-28 06:51:59 +0200 (Fri, 28 Jul 2006) | 6 lines Live with that "the hardware address" is an ill-defined concept, and that different ways of trying to find "the hardware address" may return different results. Certainly true on both of my Windows boxes, and in different ways (see whining on python-dev). ........ r50897 | neal.norwitz | 2006-07-28 09:21:27 +0200 (Fri, 28 Jul 2006) | 3 lines Try to find the MAC addr on various flavours of Unix. This seems hopeless. The reduces the test_uuid failures, but there's still another method failing. ........ r50898 | martin.v.loewis | 2006-07-28 09:45:49 +0200 (Fri, 28 Jul 2006) | 2 lines Add UUID for upcoming 2.5b3. ........ r50899 | matt.fleming | 2006-07-28 13:27:27 +0200 (Fri, 28 Jul 2006) | 3 lines Allow socketmodule to compile on NetBSD -current, whose bluetooth API differs from both Linux and FreeBSD. Accepted by Neal Norwitz. ........ r50900 | andrew.kuchling | 2006-07-28 14:07:12 +0200 (Fri, 28 Jul 2006) | 1 line [Patch #1529811] Correction to description of r|* mode ........ r50901 | andrew.kuchling | 2006-07-28 14:18:22 +0200 (Fri, 28 Jul 2006) | 1 line Typo fix ........ r50902 | andrew.kuchling | 2006-07-28 14:32:43 +0200 (Fri, 28 Jul 2006) | 1 line Add example ........ r50903 | andrew.kuchling | 2006-07-28 14:33:19 +0200 (Fri, 28 Jul 2006) | 1 line Add example ........ r50904 | andrew.kuchling | 2006-07-28 14:45:55 +0200 (Fri, 28 Jul 2006) | 1 line Don't overwrite built-in name; add some blank lines for readability ........ r50905 | andrew.kuchling | 2006-07-28 14:48:07 +0200 (Fri, 28 Jul 2006) | 1 line Add example. Should I propagate this example to all the other DBM-ish modules, too? ........ r50912 | georg.brandl | 2006-07-28 20:31:39 +0200 (Fri, 28 Jul 2006) | 3 lines Patch #1529686: also run test_email_codecs with regrtest.py. ........ r50913 | georg.brandl | 2006-07-28 20:36:01 +0200 (Fri, 28 Jul 2006) | 3 lines Fix spelling. ........ r50915 | thomas.heller | 2006-07-28 21:42:40 +0200 (Fri, 28 Jul 2006) | 3 lines Remove a useless XXX comment. Cosmetic changes to the code so that the #ifdef _UNICODE block doesn't mess emacs code formatting. ........ r50916 | phillip.eby | 2006-07-28 23:12:07 +0200 (Fri, 28 Jul 2006) | 5 lines Bug #1529871: The speed enhancement patch #921466 broke Python's compliance with PEP 302. This was fixed by adding an ``imp.NullImporter`` type that is used in ``sys.path_importer_cache`` to cache non-directory paths and avoid excessive filesystem operations during imports. ........ r50917 | phillip.eby | 2006-07-28 23:31:54 +0200 (Fri, 28 Jul 2006) | 2 lines Fix svn merge spew. ........ r50918 | thomas.heller | 2006-07-28 23:43:20 +0200 (Fri, 28 Jul 2006) | 4 lines Patch #1529514: More openbsd platforms for ctypes. Regenerated Modules/_ctypes/libffi/configure with autoconf 2.59. Approved by Neal. ........ r50922 | georg.brandl | 2006-07-29 10:51:21 +0200 (Sat, 29 Jul 2006) | 2 lines Bug #835255: The "closure" argument to new.function() is now documented. ........ r50924 | georg.brandl | 2006-07-29 11:33:26 +0200 (Sat, 29 Jul 2006) | 3 lines Bug #1441397: The compiler module now recognizes module and function docstrings correctly as it did in Python 2.4. ........ r50925 | georg.brandl | 2006-07-29 12:25:46 +0200 (Sat, 29 Jul 2006) | 4 lines Revert rev 42617, it was introduced to work around bug #1441397. test_compiler now passes again. ........ r50926 | fred.drake | 2006-07-29 15:22:49 +0200 (Sat, 29 Jul 2006) | 1 line update target version number ........ r50927 | andrew.kuchling | 2006-07-29 15:56:48 +0200 (Sat, 29 Jul 2006) | 1 line Add example ........ r50928 | andrew.kuchling | 2006-07-29 16:04:47 +0200 (Sat, 29 Jul 2006) | 1 line Update URL ........ r50930 | andrew.kuchling | 2006-07-29 16:08:15 +0200 (Sat, 29 Jul 2006) | 1 line Reword paragraph to match the order of the subsequent sections ........ r50931 | andrew.kuchling | 2006-07-29 16:21:15 +0200 (Sat, 29 Jul 2006) | 1 line [Bug #1529157] Mention raw_input() and input(); while I'm at it, reword the description a bit ........ r50932 | andrew.kuchling | 2006-07-29 16:42:48 +0200 (Sat, 29 Jul 2006) | 1 line [Bug #1519571] Document some missing functions: setup(), title(), done() ........ r50933 | andrew.kuchling | 2006-07-29 16:43:55 +0200 (Sat, 29 Jul 2006) | 1 line Fix docstring punctuation ........ r50934 | andrew.kuchling | 2006-07-29 17:10:32 +0200 (Sat, 29 Jul 2006) | 1 line [Bug #1414697] Change docstring of set/frozenset types to specify that the contents are unique. Raymond, please feel free to edit or revert. ........ r50935 | andrew.kuchling | 2006-07-29 17:35:21 +0200 (Sat, 29 Jul 2006) | 1 line [Bug #1530382] Document SSL.server(), .issuer() methods ........ r50936 | andrew.kuchling | 2006-07-29 17:42:46 +0200 (Sat, 29 Jul 2006) | 1 line Typo fix ........ r50937 | andrew.kuchling | 2006-07-29 17:43:13 +0200 (Sat, 29 Jul 2006) | 1 line Tweak wording ........ r50938 | matt.fleming | 2006-07-29 17:55:30 +0200 (Sat, 29 Jul 2006) | 2 lines Fix typo ........ r50939 | andrew.kuchling | 2006-07-29 17:57:08 +0200 (Sat, 29 Jul 2006) | 6 lines [Bug #1528258] Mention that the 'data' argument can be None. The constructor docs referred the reader to the add_data() method's docs, but they weren't very helpful. I've simply copied an earlier explanation of 'data' that's more useful. ........ r50940 | andrew.kuchling | 2006-07-29 18:08:40 +0200 (Sat, 29 Jul 2006) | 1 line Set bug/patch count. Take a bow, everyone! ........ r50941 | fred.drake | 2006-07-29 18:56:15 +0200 (Sat, 29 Jul 2006) | 18 lines expunge the xmlcore changes: 41667, 41668 - initial switch to xmlcore 47044 - mention of xmlcore in What's New 50687 - mention of xmlcore in the library reference re-apply xmlcore changes to xml: 41674 - line ending changes (re-applied manually), directory props 41677 - add cElementTree wrapper 41678 - PSF licensing for etree 41812 - whitespace normalization 42724 - fix svn:eol-style settings 43681, 43682 - remove Python version-compatibility cruft from minidom 46773 - fix encoding of \r\n\t in attr values in saxutils 47269 - added XMLParser alias for cElementTree compatibility additional tests were added in Lib/test/test_sax.py that failed with the xmlcore changes; these relate to SF bugs #1511497, #1513611 ........ r50942 | andrew.kuchling | 2006-07-29 20:14:07 +0200 (Sat, 29 Jul 2006) | 17 lines Reorganize the docs for 'file' and 'open()' after some discussion with Fred. We want to encourage users to write open() when opening a file, but open() was described with a single paragraph and 'file' had lots of explanation of the mode and bufsize arguments. I've shrunk the description of 'file' to cross-reference to the 'File objects' section, and to open() for an explanation of the arguments. open() now has all the paragraphs about the mode string. The bufsize argument was moved up so that it isn't buried at the end; now there's 1 paragraph on mode, 1 on bufsize, and then 3 more on mode. Various other edits and rearrangements were made in the process. It's probably best to read the final text and not to try to make sense of the diffs. ........ r50943 | fred.drake | 2006-07-29 20:19:19 +0200 (Sat, 29 Jul 2006) | 1 line restore test un-intentionally removed in the xmlcore purge (revision 50941) ........ r50944 | fred.drake | 2006-07-29 20:33:29 +0200 (Sat, 29 Jul 2006) | 3 lines make the reference to older versions of the documentation a link to the right page on python.org ........ r50945 | fred.drake | 2006-07-29 21:09:01 +0200 (Sat, 29 Jul 2006) | 1 line document the footnote usage pattern ........ r50947 | fred.drake | 2006-07-29 21:14:10 +0200 (Sat, 29 Jul 2006) | 1 line emphasize and oddball nuance of LaTeX comment syntax ........ r50948 | andrew.kuchling | 2006-07-29 21:24:04 +0200 (Sat, 29 Jul 2006) | 1 line [Patch #1490989 from Skip Montanaro] Mention debugging builds in the API documentation. I've changed Skip's patch to point to Misc/SpecialBuilds and fiddled with the markup a bit. ........ r50949 | neal.norwitz | 2006-07-29 21:29:35 +0200 (Sat, 29 Jul 2006) | 6 lines Disable these tests until they are reliable across platforms. These problems may mask more important, real problems. One or both methods are known to fail on: Solaris, OpenBSD, Debian, Ubuntu. They pass on Windows and some Linux boxes. ........ r50950 | andrew.kuchling | 2006-07-29 21:50:37 +0200 (Sat, 29 Jul 2006) | 1 line [Patch #1068277] Clarify that os.path.exists() can return False depending on permissions. Fred approved committing this patch in December 2004! ........ r50952 | fred.drake | 2006-07-29 22:04:42 +0200 (Sat, 29 Jul 2006) | 6 lines SF bug #1193966: Weakref types documentation misplaced The information about supporting weakrefs with types defined in C extensions is moved to the Extending & Embedding manual. Py_TPFLAGS_HAVE_WEAKREFS is no longer mentioned since it is part of Py_TPFLAGS_DEFAULT. ........ r50953 | skip.montanaro | 2006-07-29 22:06:05 +0200 (Sat, 29 Jul 2006) | 4 lines Add a comment to the csv reader documentation that explains why the treatment of newlines changed in 2.5. Pulled almost verbatim from a comment by Andrew McNamara in <http://python.org/sf/1465014>. ........ r50954 | neal.norwitz | 2006-07-29 22:20:52 +0200 (Sat, 29 Jul 2006) | 3 lines If the executable doesn't exist, there's no reason to try to start it. This prevents garbage about command not found being printed on Solaris. ........ r50955 | fred.drake | 2006-07-29 22:21:25 +0200 (Sat, 29 Jul 2006) | 1 line fix minor markup error that introduced extra punctuation ........ r50957 | neal.norwitz | 2006-07-29 22:37:08 +0200 (Sat, 29 Jul 2006) | 3 lines Disable test_getnode too, since this is also unreliable. ........ r50958 | andrew.kuchling | 2006-07-29 23:27:12 +0200 (Sat, 29 Jul 2006) | 1 line Follow TeX's conventions for hyphens ........ r50959 | andrew.kuchling | 2006-07-29 23:30:21 +0200 (Sat, 29 Jul 2006) | 1 line Fix case for 'Unix' ........ r50960 | fred.drake | 2006-07-30 01:34:57 +0200 (Sun, 30 Jul 2006) | 1 line markup cleanups ........ r50961 | andrew.kuchling | 2006-07-30 02:27:34 +0200 (Sun, 30 Jul 2006) | 1 line Minor typo fixes ........ r50962 | andrew.kuchling | 2006-07-30 02:37:56 +0200 (Sun, 30 Jul 2006) | 1 line [Bug #793553] Correct description of keyword arguments for SSL authentication ........ r50963 | tim.peters | 2006-07-30 02:58:15 +0200 (Sun, 30 Jul 2006) | 2 lines Whitespace normalization. ........ r50964 | fred.drake | 2006-07-30 05:03:43 +0200 (Sun, 30 Jul 2006) | 1 line lots of markup nits, most commonly Unix/unix --> \UNIX ........ r50965 | fred.drake | 2006-07-30 07:41:28 +0200 (Sun, 30 Jul 2006) | 1 line update information on wxPython, from Robin Dunn ........ r50966 | fred.drake | 2006-07-30 07:49:49 +0200 (Sun, 30 Jul 2006) | 4 lines remove possibly-outdated comment on what GUI toolkit is most commonly used; it is hard to know whether this is right, and it does not add valuable reference information at any rate ........ r50967 | fred.drake | 2006-07-30 07:55:39 +0200 (Sun, 30 Jul 2006) | 3 lines - remove yet another reference to how commonly Tkinter is (thought to be) used - fix an internal section reference ........ r50968 | neal.norwitz | 2006-07-30 08:53:31 +0200 (Sun, 30 Jul 2006) | 4 lines Patch #1531113: Fix augmented assignment with yield expressions. Also fix a SystemError when trying to assign to yield expressions. ........ r50969 | neal.norwitz | 2006-07-30 08:55:48 +0200 (Sun, 30 Jul 2006) | 5 lines Add PyErr_WarnEx() so C code can pass the stacklevel to warnings.warn(). This provides the proper warning for struct.pack(). PyErr_Warn() is now deprecated in favor of PyErr_WarnEx(). As mentioned by Tim Peters on python-dev. ........ r50970 | neal.norwitz | 2006-07-30 08:57:04 +0200 (Sun, 30 Jul 2006) | 3 lines Bug #1515471: string.replace() accepts character buffers again. Pass the char* and size around rather than PyObject's. ........ r50971 | neal.norwitz | 2006-07-30 08:59:13 +0200 (Sun, 30 Jul 2006) | 1 line Whitespace normalization ........ r50973 | georg.brandl | 2006-07-30 12:53:32 +0200 (Sun, 30 Jul 2006) | 3 lines Clarify that __op__ methods must return NotImplemented if they don't support the operation. ........ r50974 | georg.brandl | 2006-07-30 13:07:23 +0200 (Sun, 30 Jul 2006) | 3 lines Bug #1002398: The documentation for os.path.sameopenfile now correctly refers to file descriptors, not file objects. ........ r50977 | martin.v.loewis | 2006-07-30 15:00:31 +0200 (Sun, 30 Jul 2006) | 3 lines Don't copy directory stat times in shutil.copytree on Windows Fixes #1525866. ........ r50978 | martin.v.loewis | 2006-07-30 15:14:05 +0200 (Sun, 30 Jul 2006) | 3 lines Base __version__ on sys.version_info, as distutils is no longer maintained separatedly. ........ r50979 | martin.v.loewis | 2006-07-30 15:27:31 +0200 (Sun, 30 Jul 2006) | 3 lines Mention Cygwin in distutils error message about a missing VS 2003. Fixes #1257728. ........ r50982 | martin.v.loewis | 2006-07-30 16:09:47 +0200 (Sun, 30 Jul 2006) | 5 lines Drop usage of test -e in configure as it is not portable. Fixes #1439538 Will backport to 2.4 Also regenerate pyconfig.h.in. ........ r50984 | georg.brandl | 2006-07-30 18:20:10 +0200 (Sun, 30 Jul 2006) | 3 lines Fix makefile changes for python-config. ........ r50985 | george.yoshida | 2006-07-30 18:37:37 +0200 (Sun, 30 Jul 2006) | 2 lines Rename struct.pack_to to struct.pack_into as changed in revision 46642. ........ r50986 | george.yoshida | 2006-07-30 18:41:30 +0200 (Sun, 30 Jul 2006) | 2 lines Typo fix ........ r50987 | neal.norwitz | 2006-07-30 21:18:13 +0200 (Sun, 30 Jul 2006) | 1 line Add some asserts and update comments ........ r50988 | neal.norwitz | 2006-07-30 21:18:38 +0200 (Sun, 30 Jul 2006) | 1 line Verify that the signal handlers were really called ........ r50989 | neal.norwitz | 2006-07-30 21:20:42 +0200 (Sun, 30 Jul 2006) | 3 lines Try to prevent hangs on Tru64/Alpha buildbot. I'm not certain this will help and may need to be reverted if it causes problems. ........ r50990 | georg.brandl | 2006-07-30 22:18:51 +0200 (Sun, 30 Jul 2006) | 2 lines Bug #1531349: right <-> left glitch in __rop__ description. ........ r50992 | tim.peters | 2006-07-31 03:46:03 +0200 (Mon, 31 Jul 2006) | 2 lines Whitespace normalization. ........ r50993 | andrew.mcnamara | 2006-07-31 04:27:48 +0200 (Mon, 31 Jul 2006) | 2 lines Redo the comment about the 2.5 change in quoted-newline handling. ........ r50994 | tim.peters | 2006-07-31 04:40:23 +0200 (Mon, 31 Jul 2006) | 10 lines ZipFile.close(): Killed one of the struct.pack deprecation warnings on Win32. Also added an XXX about the line: pos3 = self.fp.tell() `pos3` is never referenced, and I have no idea what the code intended to do instead. ........ r50996 | tim.peters | 2006-07-31 04:53:03 +0200 (Mon, 31 Jul 2006) | 8 lines ZipFile.close(): Kill the other struct.pack deprecation warning on Windows. Afraid I can't detect a pattern to when the pack formats decide to use a signed or unsigned format code -- appears nearly arbitrary to my eyes. So I left all the pack formats alone and changed the special-case data values instead. ........ r50997 | skip.montanaro | 2006-07-31 05:09:45 +0200 (Mon, 31 Jul 2006) | 1 line minor tweaks ........ r50998 | skip.montanaro | 2006-07-31 05:11:11 +0200 (Mon, 31 Jul 2006) | 1 line minor tweaks ........ r50999 | andrew.kuchling | 2006-07-31 14:20:24 +0200 (Mon, 31 Jul 2006) | 1 line Add refcounts for PyErr_WarnEx ........ r51000 | andrew.kuchling | 2006-07-31 14:39:05 +0200 (Mon, 31 Jul 2006) | 9 lines Document PyErr_WarnEx. (Bad Neal! No biscuit!) Is the explanation of the 'stacklevel' parameter clear? Please feel free to edit it. I don't have LaTeX installed on this machine, so haven't verified that the markup is correct. Will check tonight, or maybe the automatic doc build will tell me. ........ r51001 | andrew.kuchling | 2006-07-31 14:52:26 +0200 (Mon, 31 Jul 2006) | 1 line Add PyErr_WarnEx() ........ r51002 | andrew.kuchling | 2006-07-31 15:18:27 +0200 (Mon, 31 Jul 2006) | 1 line Mention csv newline changes ........ r51003 | andrew.kuchling | 2006-07-31 17:22:58 +0200 (Mon, 31 Jul 2006) | 1 line Typo fix ........ r51004 | andrew.kuchling | 2006-07-31 17:23:43 +0200 (Mon, 31 Jul 2006) | 1 line Remove reference to notation ........ r51005 | georg.brandl | 2006-07-31 18:00:34 +0200 (Mon, 31 Jul 2006) | 3 lines Fix function name. ........ r51006 | andrew.kuchling | 2006-07-31 18:10:24 +0200 (Mon, 31 Jul 2006) | 1 line [Bug #1514540] Instead of putting the standard types in a section, put them in a chapter of their own. This means string methods will now show up in the ToC. (Should the types come before or after the functions+exceptions+constants chapter? I've put them after, for now.) ........ r51007 | andrew.kuchling | 2006-07-31 18:22:05 +0200 (Mon, 31 Jul 2006) | 1 line [Bug #848556] Remove \d* from second alternative to avoid exponential case when repeating match ........ r51008 | andrew.kuchling | 2006-07-31 18:27:57 +0200 (Mon, 31 Jul 2006) | 1 line Update list of files; fix a typo ........ r51013 | andrew.kuchling | 2006-08-01 18:24:30 +0200 (Tue, 01 Aug 2006) | 1 line typo fix ........ r51018 | thomas.heller | 2006-08-01 18:54:43 +0200 (Tue, 01 Aug 2006) | 2 lines Fix a potential segfault and various potentail refcount leaks in the cast() function. ........ r51020 | thomas.heller | 2006-08-01 19:46:10 +0200 (Tue, 01 Aug 2006) | 1 line Minimal useful docstring for CopyComPointer. ........ r51021 | andrew.kuchling | 2006-08-01 20:16:15 +0200 (Tue, 01 Aug 2006) | 8 lines [Patch #1520905] Attempt to suppress core file created by test_subprocess.py. Patch by Douglas Greiman. The test_run_abort() testcase produces a core file on Unix systems, even though the test is successful. This can be confusing or alarming to someone who runs 'make test' and then finds that the Python interpreter apparently crashed. ........ r51023 | georg.brandl | 2006-08-01 20:49:24 +0200 (Tue, 01 Aug 2006) | 3 lines os.urandom no longer masks unrelated exceptions like SystemExit or KeyboardInterrupt. ........ r51025 | thomas.heller | 2006-08-01 21:14:15 +0200 (Tue, 01 Aug 2006) | 2 lines Speed up PyType_stgdict and PyObject_stgdict. ........ r51027 | ronald.oussoren | 2006-08-01 22:30:31 +0200 (Tue, 01 Aug 2006) | 3 lines Make sure the postinstall action that optionally updates the user's profile on MacOS X actually works correctly in all cases. ........ r51028 | ronald.oussoren | 2006-08-01 23:00:57 +0200 (Tue, 01 Aug 2006) | 4 lines This fixes bug #1527397: PythonLauncher runs scripts with the wrong working directory. It also fixes a bug where PythonLauncher failed to launch scripts when the scriptname (or the path to the script) contains quotes. ........ r51031 | tim.peters | 2006-08-02 05:27:46 +0200 (Wed, 02 Aug 2006) | 2 lines Whitespace normalization. ........ r51032 | tim.peters | 2006-08-02 06:12:36 +0200 (Wed, 02 Aug 2006) | 19 lines Try to squash struct.pack warnings on the "amd64 gentoo trunk" buildbot (& possibly other 64-bit boxes) during test_gzip. The native zlib crc32 function returns an unsigned 32-bit integer, which the Python wrapper implicitly casts to C long. Therefore the same crc can "look negative" on a 32-bit box but "look positive" on a 64-bit box. This patch papers over that platform difference when writing the crc to file. It may be better to change the Python wrapper, either to make the result "look positive" on all platforms (which means it may have to return a Python long at times on a 32-bit box), or to keep the sign the same across boxes. But that would be a visible change in what users see, while the current hack changes no visible behavior (well, apart from stopping the struct deprecation warning). Note that the module-level write32() function is no longer used. ........ r51033 | neal.norwitz | 2006-08-02 06:27:11 +0200 (Wed, 02 Aug 2006) | 4 lines Prevent memory leak on error. Reported by Klocwork #36 ........ r51034 | tim.peters | 2006-08-02 07:20:08 +0200 (Wed, 02 Aug 2006) | 9 lines _Stream.close(): Try to kill struct.pack() warnings when writing the crc to file on the "PPC64 Debian trunk" buildbot when running test_tarfile. This is again a case where the native zlib crc is an unsigned 32-bit int, but the Python wrapper implicitly casts it to signed C long, so that "the sign bit looks different" on different platforms. ........ r51035 | ronald.oussoren | 2006-08-02 08:10:10 +0200 (Wed, 02 Aug 2006) | 2 lines Updated documentation for the script that builds the OSX installer. ........ r51036 | neal.norwitz | 2006-08-02 08:14:22 +0200 (Wed, 02 Aug 2006) | 2 lines _PyWeakref_GetWeakrefCount() now returns a Py_ssize_t instead of long. ........ r51037 | neal.norwitz | 2006-08-02 08:15:10 +0200 (Wed, 02 Aug 2006) | 1 line v is already checked for NULL, so just DECREF it ........ r51038 | neal.norwitz | 2006-08-02 08:19:19 +0200 (Wed, 02 Aug 2006) | 1 line Let us know when there was a problem and the child had to kill the parent ........ r51039 | neal.norwitz | 2006-08-02 08:46:21 +0200 (Wed, 02 Aug 2006) | 5 lines Patch #1519025 and bug #926423: If a KeyboardInterrupt occurs during a socket operation on a socket with a timeout, the exception will be caught correctly. Previously, the exception was not caught. ........ r51040 | neal.norwitz | 2006-08-02 09:09:32 +0200 (Wed, 02 Aug 2006) | 1 line Add some explanation about Klocwork and Coverity static analysis ........ r51041 | anthony.baxter | 2006-08-02 09:43:09 +0200 (Wed, 02 Aug 2006) | 1 line pre-release machinations ........ r51043 | thomas.heller | 2006-08-02 13:35:31 +0200 (Wed, 02 Aug 2006) | 4 lines A few nore words about what ctypes does. Document that using the wrong calling convention can also raise 'ValueError: Procedure called with the wrong number of arguments'. ........ r51045 | thomas.heller | 2006-08-02 14:00:13 +0200 (Wed, 02 Aug 2006) | 1 line Fix a mistake. ........ r51046 | martin.v.loewis | 2006-08-02 15:53:55 +0200 (Wed, 02 Aug 2006) | 3 lines Correction of patch #1455898: In the mbcs decoder, set final=False for stream decoder, but final=True for the decode function. ........ r51049 | tim.peters | 2006-08-02 20:19:35 +0200 (Wed, 02 Aug 2006) | 2 lines Add missing svn:eol-style property to text files. ........ r51079 | neal.norwitz | 2006-08-04 06:50:21 +0200 (Fri, 04 Aug 2006) | 3 lines Bug #1531405, format_exception no longer raises an exception if str(exception) raised an exception. ........ r51080 | neal.norwitz | 2006-08-04 06:58:47 +0200 (Fri, 04 Aug 2006) | 11 lines Bug #1191458: tracing over for loops now produces a line event on each iteration. I'm not positive this is the best way to handle this. I'm also not sure that there aren't other cases where the lnotab is generated incorrectly. It would be great if people that use pdb or tracing could test heavily. Also: * Remove dead/duplicated code that wasn't used/necessary because we already handled the docstring prior to entering the loop. * add some debugging code into the compiler (#if 0'd out). ........ r51081 | neal.norwitz | 2006-08-04 07:09:28 +0200 (Fri, 04 Aug 2006) | 4 lines Bug #1333982: string/number constants were inappropriately stored in the byte code and co_consts even if they were not used, ie immediately popped off the stack. ........ r51082 | neal.norwitz | 2006-08-04 07:12:19 +0200 (Fri, 04 Aug 2006) | 1 line There were really two issues ........ r51084 | fred.drake | 2006-08-04 07:17:21 +0200 (Fri, 04 Aug 2006) | 1 line SF patch #1534048 (bug #1531003): fix typo in error message ........ r51085 | gregory.p.smith | 2006-08-04 07:17:47 +0200 (Fri, 04 Aug 2006) | 3 lines fix typos ........ r51087 | georg.brandl | 2006-08-04 08:03:53 +0200 (Fri, 04 Aug 2006) | 3 lines Fix bug caused by first decrefing, then increfing. ........ r51109 | neil.schemenauer | 2006-08-04 18:20:30 +0200 (Fri, 04 Aug 2006) | 5 lines Fix the 'compiler' package to generate correct code for MAKE_CLOSURE. In the 2.5 development cycle, MAKE_CLOSURE as changed to take free variables as a tuple rather than as individual items on the stack. Closes patch #1534084. ........ r51110 | georg.brandl | 2006-08-04 20:03:37 +0200 (Fri, 04 Aug 2006) | 3 lines Change fix for segfaulting property(), add a NEWS entry and a test. ........ r51111 | georg.brandl | 2006-08-04 20:07:34 +0200 (Fri, 04 Aug 2006) | 3 lines Better fix for bug #1531405, not executing str(value) twice. ........ r51112 | thomas.heller | 2006-08-04 20:17:40 +0200 (Fri, 04 Aug 2006) | 1 line On Windows, make PyErr_Warn an exported function again. ........ r51113 | thomas.heller | 2006-08-04 20:57:34 +0200 (Fri, 04 Aug 2006) | 4 lines Fix #1530448 - fix ctypes build failure on solaris 10. The '-mimpure-text' linker flag is required when linking _ctypes.so. ........ r51114 | thomas.heller | 2006-08-04 21:49:31 +0200 (Fri, 04 Aug 2006) | 3 lines Fix #1534738: win32 debug version of _msi must be _msi_d.pyd, not _msi.pyd. Fix the name of the pdb file as well. ........ r51115 | andrew.kuchling | 2006-08-04 22:37:43 +0200 (Fri, 04 Aug 2006) | 1 line Typo fixes ........ r51116 | andrew.kuchling | 2006-08-04 23:10:03 +0200 (Fri, 04 Aug 2006) | 1 line Fix mangled sentence ........ r51118 | tim.peters | 2006-08-05 00:00:35 +0200 (Sat, 05 Aug 2006) | 2 lines Whitespace normalization. ........ r51119 | bob.ippolito | 2006-08-05 01:59:21 +0200 (Sat, 05 Aug 2006) | 5 lines Fix #1530559, struct.pack raises TypeError where it used to convert. Passing float arguments to struct.pack when integers are expected now triggers a DeprecationWarning. ........ r51123 | georg.brandl | 2006-08-05 08:10:54 +0200 (Sat, 05 Aug 2006) | 3 lines Patch #1534922: correct and enhance unittest docs. ........ r51126 | georg.brandl | 2006-08-06 09:06:33 +0200 (Sun, 06 Aug 2006) | 2 lines Bug #1535182: really test the xreadlines() method of bz2 objects. ........ r51128 | georg.brandl | 2006-08-06 09:26:21 +0200 (Sun, 06 Aug 2006) | 4 lines Bug #1535081: A leading underscore has been added to the names of the md5 and sha modules, so add it in Modules/Setup.dist too. ........ r51129 | georg.brandl | 2006-08-06 10:23:54 +0200 (Sun, 06 Aug 2006) | 3 lines Bug #1535165: fixed a segfault in input() and raw_input() when sys.stdin is closed. ........ r51131 | georg.brandl | 2006-08-06 11:17:16 +0200 (Sun, 06 Aug 2006) | 2 lines Don't produce output in test_builtin. ........ r51133 | andrew.macintyre | 2006-08-06 14:37:03 +0200 (Sun, 06 Aug 2006) | 4 lines test_threading now skips testing alternate thread stack sizes on platforms that don't support changing thread stack size. ........ r51134 | andrew.kuchling | 2006-08-07 00:07:04 +0200 (Mon, 07 Aug 2006) | 2 lines [Patch #1464056] Ensure that we use the panelw library when linking with ncursesw. Once I see how the buildbots react, I'll backport this to 2.4. ........ r51137 | georg.brandl | 2006-08-08 13:52:34 +0200 (Tue, 08 Aug 2006) | 3 lines webbrowser: Silence stderr output if no gconftool or gnome browser found ........ r51138 | georg.brandl | 2006-08-08 13:56:21 +0200 (Tue, 08 Aug 2006) | 7 lines Remove "non-mapping" and "non-sequence" from TypeErrors raised by PyMapping_Size and PySequence_Size. Because len() tries first sequence, then mapping size, it will always raise a "non-mapping object has no len" error which is confusing. ........ r51139 | thomas.heller | 2006-08-08 19:37:00 +0200 (Tue, 08 Aug 2006) | 3 lines memcmp() can return values other than -1, 0, and +1 but tp_compare must not. ........ r51140 | thomas.heller | 2006-08-08 19:39:20 +0200 (Tue, 08 Aug 2006) | 1 line Remove accidently committed, duplicated test. ........ r51147 | andrew.kuchling | 2006-08-08 20:50:14 +0200 (Tue, 08 Aug 2006) | 1 line Reword paragraph to clarify ........ r51148 | andrew.kuchling | 2006-08-08 20:56:08 +0200 (Tue, 08 Aug 2006) | 1 line Move obmalloc item into C API section ........ r51149 | andrew.kuchling | 2006-08-08 21:00:14 +0200 (Tue, 08 Aug 2006) | 1 line 'Other changes' section now has only one item; move the item elsewhere and remove the section ........ r51150 | andrew.kuchling | 2006-08-08 21:00:34 +0200 (Tue, 08 Aug 2006) | 1 line Bump version number ........ r51151 | georg.brandl | 2006-08-08 22:11:22 +0200 (Tue, 08 Aug 2006) | 2 lines Bug #1536828: typo: TypeType should have been StringType. ........ r51153 | georg.brandl | 2006-08-08 22:13:13 +0200 (Tue, 08 Aug 2006) | 2 lines Bug #1536660: separate two words. ........ r51155 | georg.brandl | 2006-08-08 22:48:10 +0200 (Tue, 08 Aug 2006) | 3 lines ``str`` is now the same object as ``types.StringType``. ........ r51156 | tim.peters | 2006-08-09 02:52:26 +0200 (Wed, 09 Aug 2006) | 2 lines Whitespace normalization. ........ r51158 | georg.brandl | 2006-08-09 09:03:22 +0200 (Wed, 09 Aug 2006) | 4 lines Introduce an upper bound on tuple nesting depth in C argument format strings; fixes rest of #1523610. ........ r51160 | martin.v.loewis | 2006-08-09 09:57:39 +0200 (Wed, 09 Aug 2006) | 4 lines __hash__ may now return long int; the final hash value is obtained by invoking hash on the long int. Fixes #1536021. ........ r51168 | andrew.kuchling | 2006-08-09 15:03:41 +0200 (Wed, 09 Aug 2006) | 1 line [Bug #1536021] Mention __hash__ change ........ r51169 | andrew.kuchling | 2006-08-09 15:57:05 +0200 (Wed, 09 Aug 2006) | 1 line [Patch #1534027] Add notes on locale module changes ........ r51170 | andrew.kuchling | 2006-08-09 16:05:35 +0200 (Wed, 09 Aug 2006) | 1 line Add missing 'self' parameters ........ r51171 | andrew.kuchling | 2006-08-09 16:06:19 +0200 (Wed, 09 Aug 2006) | 1 line Reindent code ........ r51172 | armin.rigo | 2006-08-09 16:55:26 +0200 (Wed, 09 Aug 2006) | 2 lines Fix and test for an infinite C recursion. ........ r51173 | ronald.oussoren | 2006-08-09 16:56:33 +0200 (Wed, 09 Aug 2006) | 2 lines It's unlikely that future versions will require _POSIX_C_SOURCE ........ r51178 | armin.rigo | 2006-08-09 17:37:26 +0200 (Wed, 09 Aug 2006) | 2 lines Concatenation on a long string breaks (SF #1526585). ........ r51180 | kurt.kaiser | 2006-08-09 18:46:15 +0200 (Wed, 09 Aug 2006) | 8 lines 1. When used w/o subprocess, all exceptions were preceeded by an error message claiming they were IDLE internal errors (since 1.2a1). 2. Add Ronald Oussoren to CREDITS M NEWS.txt M PyShell.py M CREDITS.txt ........ r51181 | kurt.kaiser | 2006-08-09 19:47:15 +0200 (Wed, 09 Aug 2006) | 4 lines As a slight enhancement to the previous checkin, improve the internal error reporting by moving message to IDLE console. ........ r51182 | andrew.kuchling | 2006-08-09 20:23:14 +0200 (Wed, 09 Aug 2006) | 1 line Typo fix ........ r51183 | kurt.kaiser | 2006-08-09 22:34:46 +0200 (Wed, 09 Aug 2006) | 2 lines ToggleTab dialog was setting indent to 8 even if cancelled (since 1.2a1). ........ r51184 | martin.v.loewis | 2006-08-10 01:42:18 +0200 (Thu, 10 Aug 2006) | 2 lines Add some commentary on -mimpure-text. ........ r51185 | tim.peters | 2006-08-10 02:58:49 +0200 (Thu, 10 Aug 2006) | 2 lines Add missing svn:eol-style property to text files. ........ r51186 | kurt.kaiser | 2006-08-10 03:41:17 +0200 (Thu, 10 Aug 2006) | 2 lines Changing tokenize (39046) to detect dedent broke tabnanny check (since 1.2a1) ........ r51187 | tim.peters | 2006-08-10 05:01:26 +0200 (Thu, 10 Aug 2006) | 13 lines test_copytree_simple(): This was leaving behind two new temp directories each time it ran, at least on Windows. Several changes: explicitly closed all files; wrapped long lines; stopped suppressing errors when removing a file or directory fails (removing /shouldn't/ fail!); and changed what appeared to be incorrect usage of os.removedirs() (that doesn't remove empty directories at and /under/ the given path, instead it must be given an empty leaf directory and then deletes empty directories moving /up/ the path -- could be that the conceptually simpler shutil.rmtree() was really actually intended here). ........
2006-08-11 11:57:12 -03:00
}
PyDoc_STRVAR(stack_size_doc,
"stack_size($module, size=0, /)\n\
--\n\
Merged revisions 46753-51188 via svnmerge from svn+ssh://pythondev@svn.python.org/python/trunk ........ r46755 | brett.cannon | 2006-06-08 18:23:04 +0200 (Thu, 08 Jun 2006) | 4 lines Make binascii.hexlify() use s# for its arguments instead of t# to actually match its documentation stating it accepts any read-only buffer. ........ r46757 | brett.cannon | 2006-06-08 19:00:45 +0200 (Thu, 08 Jun 2006) | 8 lines Buffer objects would return the read or write buffer for a wrapped object when the char buffer was requested. Now it actually returns the char buffer if available or raises a TypeError if it isn't (as is raised for the other buffer types if they are not present but requested). Not a backport candidate since it does change semantics of the buffer object (although it could be argued this is enough of a bug to bother backporting). ........ r46760 | andrew.kuchling | 2006-06-09 03:10:17 +0200 (Fri, 09 Jun 2006) | 1 line Update functools section ........ r46762 | tim.peters | 2006-06-09 04:11:02 +0200 (Fri, 09 Jun 2006) | 6 lines Whitespace normalization. Since test_file is implicated in mysterious test failures when followed by test_optparse, if I had any brains I'd look at the checkin that last changed test_file ;-) ........ r46763 | tim.peters | 2006-06-09 05:09:42 +0200 (Fri, 09 Jun 2006) | 5 lines To boost morale :-), force test_optparse to run immediately after test_file until we can figure out how to fix it. (See python-dev; at the moment we don't even know which checkin caused the problem.) ........ r46764 | tim.peters | 2006-06-09 05:51:41 +0200 (Fri, 09 Jun 2006) | 6 lines AutoFileTests.tearDown(): Removed mysterious undocumented try/except. Remove TESTFN. Throughout: used open() instead of file(), and wrapped long lines. ........ r46765 | tim.peters | 2006-06-09 06:02:06 +0200 (Fri, 09 Jun 2006) | 8 lines testUnicodeOpen(): I have no idea why, but making this test clean up after itself appears to fix the test failures when test_optparse follows test_file. test_main(): Get rid of TESTFN no matter what. That's also enough to fix the mystery failures. Doesn't hurt to fix them twice :-) ........ r46766 | tim.peters | 2006-06-09 07:12:40 +0200 (Fri, 09 Jun 2006) | 6 lines Remove the temporary hack to force test_optparse to run immediately after test_file. At least 8 buildbot boxes passed since the underlying problem got fixed, and they all failed before the fix, so there's no point to this anymore. ........ r46767 | neal.norwitz | 2006-06-09 07:54:18 +0200 (Fri, 09 Jun 2006) | 1 line Fix grammar and reflow ........ r46769 | andrew.kuchling | 2006-06-09 12:22:35 +0200 (Fri, 09 Jun 2006) | 1 line Markup fix ........ r46773 | andrew.kuchling | 2006-06-09 15:15:57 +0200 (Fri, 09 Jun 2006) | 1 line [Bug #1472827] Make saxutils.XMLGenerator handle \r\n\t in attribute values by escaping them properly. 2.4 bugfix candidate. ........ r46778 | kristjan.jonsson | 2006-06-09 18:28:01 +0200 (Fri, 09 Jun 2006) | 2 lines Turn off warning about deprecated CRT functions on for VisualStudio .NET 2005. Make the definition #ARRAYSIZE conditional. VisualStudio .NET 2005 already has it defined using a better gimmick. ........ r46779 | phillip.eby | 2006-06-09 18:40:18 +0200 (Fri, 09 Jun 2006) | 2 lines Import wsgiref into the stdlib, as of the external version 0.1-r2181. ........ r46783 | andrew.kuchling | 2006-06-09 18:44:40 +0200 (Fri, 09 Jun 2006) | 1 line Add note about XMLGenerator bugfix ........ r46784 | andrew.kuchling | 2006-06-09 18:46:51 +0200 (Fri, 09 Jun 2006) | 1 line Add note about wsgiref ........ r46785 | brett.cannon | 2006-06-09 19:05:48 +0200 (Fri, 09 Jun 2006) | 2 lines Fix inconsistency in naming within an enum. ........ r46787 | tim.peters | 2006-06-09 19:47:00 +0200 (Fri, 09 Jun 2006) | 2 lines Whitespace normalization. ........ r46792 | georg.brandl | 2006-06-09 20:29:52 +0200 (Fri, 09 Jun 2006) | 3 lines Test file.__exit__. ........ r46794 | brett.cannon | 2006-06-09 20:40:46 +0200 (Fri, 09 Jun 2006) | 2 lines svn:ignore .pyc and .pyo files. ........ r46795 | georg.brandl | 2006-06-09 20:45:48 +0200 (Fri, 09 Jun 2006) | 3 lines RFE #1491485: str/unicode.endswith()/startswith() now accept a tuple as first argument. ........ r46798 | andrew.kuchling | 2006-06-09 21:03:16 +0200 (Fri, 09 Jun 2006) | 1 line Describe startswith()/endswiith() change; add reminder about wsgiref ........ r46799 | tim.peters | 2006-06-09 21:24:44 +0200 (Fri, 09 Jun 2006) | 11 lines Implementing a happy idea from Georg Brandl: make runtest() try to clean up files and directories the tests often leave behind by mistake. This is the first time in history I don't have a bogus "db_home" directory after running the tests ;-) Also worked on runtest's docstring, to say something about all the arguments, and to document the non-obvious return values. New functions runtest_inner() and cleanup_test_droppings() in support of the above. ........ r46800 | andrew.kuchling | 2006-06-09 21:43:25 +0200 (Fri, 09 Jun 2006) | 1 line Remove unused variable ........ r46801 | andrew.kuchling | 2006-06-09 21:56:05 +0200 (Fri, 09 Jun 2006) | 1 line Add some wsgiref text ........ r46803 | thomas.heller | 2006-06-09 21:59:11 +0200 (Fri, 09 Jun 2006) | 1 line set eol-style svn property ........ r46804 | thomas.heller | 2006-06-09 22:01:01 +0200 (Fri, 09 Jun 2006) | 1 line set eol-style svn property ........ r46805 | georg.brandl | 2006-06-09 22:43:48 +0200 (Fri, 09 Jun 2006) | 3 lines Make use of new str.startswith/endswith semantics. Occurences in email and compiler were ignored due to backwards compat requirements. ........ r46806 | brett.cannon | 2006-06-10 00:31:23 +0200 (Sat, 10 Jun 2006) | 4 lines An object with __call__ as an attribute, when called, will have that attribute checked for __call__ itself, and will continue to look until it finds an object without the attribute. This can lead to an infinite recursion. Closes bug #532646, again. Will be backported. ........ r46808 | brett.cannon | 2006-06-10 00:45:54 +0200 (Sat, 10 Jun 2006) | 2 lines Fix bug introduced in rev. 46806 by not having variable declaration at the top of a block. ........ r46812 | georg.brandl | 2006-06-10 08:40:50 +0200 (Sat, 10 Jun 2006) | 4 lines Apply perky's fix for #1503157: "/".join([u"", u""]) raising OverflowError. Also improve error message on overflow. ........ r46817 | martin.v.loewis | 2006-06-10 10:14:03 +0200 (Sat, 10 Jun 2006) | 2 lines Port cygwin kill_python changes from 2.4 branch. ........ r46818 | armin.rigo | 2006-06-10 12:57:40 +0200 (Sat, 10 Jun 2006) | 4 lines SF bug #1503294. PyThreadState_GET() complains if the tstate is NULL, but only in debug mode. ........ r46819 | martin.v.loewis | 2006-06-10 14:23:46 +0200 (Sat, 10 Jun 2006) | 4 lines Patch #1495999: Part two of Windows CE changes. - update header checks, using autoconf - provide dummies for getenv, environ, and GetVersion - adjust MSC_VER check in socketmodule.c ........ r46820 | skip.montanaro | 2006-06-10 16:09:11 +0200 (Sat, 10 Jun 2006) | 1 line document the class, not its initializer ........ r46821 | greg.ward | 2006-06-10 18:40:01 +0200 (Sat, 10 Jun 2006) | 4 lines Sync with Optik docs (rev 518): * restore "Extending optparse" section * document ALWAYS_TYPED_ACTIONS (SF #1449311) ........ r46824 | thomas.heller | 2006-06-10 21:51:46 +0200 (Sat, 10 Jun 2006) | 8 lines Upgrade to ctypes version 0.9.9.7. Summary of changes: - support for 'variable sized' data - support for anonymous structure/union fields - fix severe bug with certain arrays or structures containing more than 256 fields ........ r46825 | thomas.heller | 2006-06-10 21:55:36 +0200 (Sat, 10 Jun 2006) | 8 lines Upgrade to ctypes version 0.9.9.7. Summary of changes: - support for 'variable sized' data - support for anonymous structure/union fields - fix severe bug with certain arrays or structures containing more than 256 fields ........ r46826 | fred.drake | 2006-06-10 22:01:34 +0200 (Sat, 10 Jun 2006) | 4 lines SF patch #1303595: improve description of __builtins__, explaining how it varies between __main__ and other modules, and strongly suggest not touching it but using __builtin__ if absolutely necessary ........ r46827 | fred.drake | 2006-06-10 22:02:58 +0200 (Sat, 10 Jun 2006) | 1 line credit for SF patch #1303595 ........ r46831 | thomas.heller | 2006-06-10 22:29:34 +0200 (Sat, 10 Jun 2006) | 2 lines New docs for ctypes. ........ r46834 | thomas.heller | 2006-06-10 23:07:19 +0200 (Sat, 10 Jun 2006) | 1 line Fix a wrong printf format. ........ r46835 | thomas.heller | 2006-06-10 23:17:58 +0200 (Sat, 10 Jun 2006) | 1 line Fix the second occurrence of the problematic printf format. ........ r46837 | thomas.heller | 2006-06-10 23:56:03 +0200 (Sat, 10 Jun 2006) | 1 line Don't use C++ comment. ........ r46838 | thomas.heller | 2006-06-11 00:01:50 +0200 (Sun, 11 Jun 2006) | 1 line Handle failure of PyMem_Realloc. ........ r46839 | skip.montanaro | 2006-06-11 00:38:13 +0200 (Sun, 11 Jun 2006) | 2 lines Suppress warning on MacOSX about possible use before set of proc. ........ r46840 | tim.peters | 2006-06-11 00:51:45 +0200 (Sun, 11 Jun 2006) | 8 lines shuffle() doscstring: Removed warning about sequence length versus generator period. While this was a real weakness of the older WH generator for lists with just a few dozen elements, and so could potentially bite the naive ;-), the Twister should show excellent behavior up to at least 600 elements. Module docstring: reflowed some jarringly short lines. ........ r46844 | greg.ward | 2006-06-11 02:40:49 +0200 (Sun, 11 Jun 2006) | 4 lines Bug #1361643: fix textwrap.dedent() so it handles tabs appropriately, i.e. do *not* expand tabs, but treat them as whitespace that is not equivalent to spaces. Add a couple of test cases. Clarify docs. ........ r46850 | neal.norwitz | 2006-06-11 07:44:18 +0200 (Sun, 11 Jun 2006) | 5 lines Fix Coverity # 146. newDBSequenceObject would deref dbobj, so it can't be NULL. We know it's not NULL from the ParseTuple and DbObject_Check will verify it's not NULL. ........ r46851 | neal.norwitz | 2006-06-11 07:45:25 +0200 (Sun, 11 Jun 2006) | 4 lines Wrap some long lines Top/Bottom factor out some common expressions Add a XXX comment about widing offset. ........ r46852 | neal.norwitz | 2006-06-11 07:45:47 +0200 (Sun, 11 Jun 2006) | 1 line Add versionadded to doc ........ r46853 | neal.norwitz | 2006-06-11 07:47:14 +0200 (Sun, 11 Jun 2006) | 3 lines Update doc to make it agree with code. Bottom factor out some common code. ........ r46854 | neal.norwitz | 2006-06-11 07:48:14 +0200 (Sun, 11 Jun 2006) | 3 lines f_code can't be NULL based on Frame_New and other code that derefs it. So there doesn't seem to be much point to checking here. ........ r46855 | neal.norwitz | 2006-06-11 09:26:27 +0200 (Sun, 11 Jun 2006) | 1 line Fix errors found by pychecker ........ r46856 | neal.norwitz | 2006-06-11 09:26:50 +0200 (Sun, 11 Jun 2006) | 1 line warnings was imported at module scope, no need to import again ........ r46857 | neal.norwitz | 2006-06-11 09:27:56 +0200 (Sun, 11 Jun 2006) | 5 lines Fix errors found by pychecker. I think these changes are correct, but I'm not sure. Could someone who knows how this module works test it? It can at least start on the cmd line. ........ r46858 | neal.norwitz | 2006-06-11 10:35:14 +0200 (Sun, 11 Jun 2006) | 1 line Fix errors found by pychecker ........ r46859 | ronald.oussoren | 2006-06-11 16:33:36 +0200 (Sun, 11 Jun 2006) | 4 lines This patch improves the L&F of IDLE on OSX. The changes are conditionalized on being in an IDLE.app bundle on darwin. This does a slight reorganisation of the menus and adds support for file-open events. ........ r46860 | greg.ward | 2006-06-11 16:42:41 +0200 (Sun, 11 Jun 2006) | 1 line SF #1366250: optparse docs: fix inconsistency in variable name; minor tweaks. ........ r46861 | greg.ward | 2006-06-11 18:24:11 +0200 (Sun, 11 Jun 2006) | 3 lines Bug #1498146: fix optparse to handle Unicode strings in option help, description, and epilog. ........ r46862 | thomas.heller | 2006-06-11 19:04:22 +0200 (Sun, 11 Jun 2006) | 2 lines Release the GIL during COM method calls, to avoid deadlocks in Python coded COM objects. ........ r46863 | tim.peters | 2006-06-11 21:42:51 +0200 (Sun, 11 Jun 2006) | 2 lines Whitespace normalization. ........ r46864 | tim.peters | 2006-06-11 21:43:49 +0200 (Sun, 11 Jun 2006) | 2 lines Add missing svn:eol-style property to text files. ........ r46865 | ronald.oussoren | 2006-06-11 21:45:57 +0200 (Sun, 11 Jun 2006) | 2 lines Remove message about using make frameworkinstall, that's no longer necesssary ........ r46866 | ronald.oussoren | 2006-06-11 22:23:29 +0200 (Sun, 11 Jun 2006) | 2 lines Use configure to substitute the correct prefix instead of hardcoding ........ r46867 | ronald.oussoren | 2006-06-11 22:24:45 +0200 (Sun, 11 Jun 2006) | 4 lines - Change fixapplepython23.py to ensure that it will run with /usr/bin/python on intel macs. - Fix some minor problems in the installer for OSX ........ r46868 | neal.norwitz | 2006-06-11 22:25:56 +0200 (Sun, 11 Jun 2006) | 5 lines Try to fix several networking tests. The problem is that if hosts have a search path setup, some of these hosts resolve to the wrong address. By appending a period to the hostname, the hostname should only resolve to what we want it to resolve to. Hopefully this doesn't break different bots. ........ r46869 | neal.norwitz | 2006-06-11 22:42:02 +0200 (Sun, 11 Jun 2006) | 7 lines Try to fix another networking test. The problem is that if hosts have a search path setup, some of these hosts resolve to the wrong address. By appending a period to the hostname, the hostname should only resolve to what we want it to resolve to. Hopefully this doesn't break different bots. Also add more info to failure message to aid debugging test failure. ........ r46870 | neal.norwitz | 2006-06-11 22:46:46 +0200 (Sun, 11 Jun 2006) | 4 lines Fix test on PPC64 buildbot. It raised an IOError (really an URLError which derives from an IOError). That seems valid. Env Error includes both OSError and IOError, so this seems like a reasonable fix. ........ r46871 | tim.peters | 2006-06-11 22:52:59 +0200 (Sun, 11 Jun 2006) | 10 lines compare_generic_iter(): Fixed the failure of test_wsgiref's testFileWrapper when running with -O. test_simple_validation_error still fails under -O. That appears to be because wsgiref's validate.py uses `assert` statements all over the place to check arguments for sanity. That should all be changed (it's not a logical error in the software if a user passes bogus arguments, so this isn't a reasonable use for `assert` -- checking external preconditions should generally raise ValueError or TypeError instead, as appropriate). ........ r46872 | neal.norwitz | 2006-06-11 23:38:38 +0200 (Sun, 11 Jun 2006) | 1 line Get test to pass on S/390. Shout if you think this change is incorrect. ........ r46873 | neal.norwitz | 2006-06-12 04:05:55 +0200 (Mon, 12 Jun 2006) | 1 line Cleanup Py_ssize_t a little (get rid of second #ifdef) ........ r46874 | neal.norwitz | 2006-06-12 04:06:17 +0200 (Mon, 12 Jun 2006) | 1 line Fix some Py_ssize_t issues ........ r46875 | neal.norwitz | 2006-06-12 04:06:42 +0200 (Mon, 12 Jun 2006) | 1 line Fix some Py_ssize_t issues ........ r46876 | neal.norwitz | 2006-06-12 04:07:24 +0200 (Mon, 12 Jun 2006) | 2 lines Cleanup: Remove import of types to get StringTypes, we can just use basestring. ........ r46877 | neal.norwitz | 2006-06-12 04:07:57 +0200 (Mon, 12 Jun 2006) | 1 line Don't truncate if size_t is bigger than uint ........ r46878 | neal.norwitz | 2006-06-12 04:08:41 +0200 (Mon, 12 Jun 2006) | 1 line Don't leak the list object if there's an error allocating the item storage. Backport candidate ........ r46879 | neal.norwitz | 2006-06-12 04:09:03 +0200 (Mon, 12 Jun 2006) | 1 line Fix typo. Backport if anyone cares. :-) ........ r46880 | neal.norwitz | 2006-06-12 04:09:34 +0200 (Mon, 12 Jun 2006) | 1 line Fix indentation of case and a Py_ssize_t issue. ........ r46881 | neal.norwitz | 2006-06-12 04:11:18 +0200 (Mon, 12 Jun 2006) | 3 lines Get rid of f_restricted too. Doc the other 4 ints that were already removed at the NeedForSpeed sprint. ........ r46882 | neal.norwitz | 2006-06-12 04:13:21 +0200 (Mon, 12 Jun 2006) | 1 line Fix the socket tests so they can be run concurrently. Backport candidate ........ r46883 | neal.norwitz | 2006-06-12 04:16:10 +0200 (Mon, 12 Jun 2006) | 1 line i and j are initialized below when used. No need to do it twice ........ r46884 | neal.norwitz | 2006-06-12 05:05:03 +0200 (Mon, 12 Jun 2006) | 1 line Remove unused import ........ r46885 | neal.norwitz | 2006-06-12 05:05:40 +0200 (Mon, 12 Jun 2006) | 1 line Impl ssize_t ........ r46886 | neal.norwitz | 2006-06-12 05:33:09 +0200 (Mon, 12 Jun 2006) | 6 lines Patch #1503046, Conditional compilation of zlib.(de)compressobj.copy copy is only in newer versions of zlib. This should allow zlibmodule to work with older versions like the Tru64 buildbot. ........ r46887 | phillip.eby | 2006-06-12 06:04:32 +0200 (Mon, 12 Jun 2006) | 2 lines Sync w/external release 0.1.2. Please see PEP 360 before making changes to external packages. ........ r46888 | martin.v.loewis | 2006-06-12 06:26:31 +0200 (Mon, 12 Jun 2006) | 2 lines Get rid of function pointer cast. ........ r46889 | thomas.heller | 2006-06-12 08:05:57 +0200 (Mon, 12 Jun 2006) | 3 lines I don't know how that happend, but the entire file contents was duplicated. Thanks to Simon Percivall for the heads up. ........ r46890 | nick.coghlan | 2006-06-12 10:19:37 +0200 (Mon, 12 Jun 2006) | 1 line Fix site module docstring to match the code ........ r46891 | nick.coghlan | 2006-06-12 10:23:02 +0200 (Mon, 12 Jun 2006) | 1 line Fix site module docstring to match the code for Mac OSX, too ........ r46892 | nick.coghlan | 2006-06-12 10:27:13 +0200 (Mon, 12 Jun 2006) | 1 line The site module documentation also described the Windows behaviour incorrectly. ........ r46893 | nick.coghlan | 2006-06-12 12:17:11 +0200 (Mon, 12 Jun 2006) | 1 line Make the -m switch conform to the documentation of sys.path by behaving like the -c switch ........ r46894 | kristjan.jonsson | 2006-06-12 17:45:12 +0200 (Mon, 12 Jun 2006) | 2 lines Fix the CRT argument error handling for VisualStudio .NET 2005. Install a CRT error handler and disable the assertion for debug builds. This causes CRT to set errno to EINVAL. This update fixes crash cases in the test suite where the default CRT error handler would cause process exit. ........ r46899 | thomas.heller | 2006-06-12 22:56:48 +0200 (Mon, 12 Jun 2006) | 1 line Add pep-291 compatibility markers. ........ r46901 | ka-ping.yee | 2006-06-13 01:47:52 +0200 (Tue, 13 Jun 2006) | 5 lines Add the uuid module. This module has been tested so far on Windows XP (Python 2.4 and 2.5a2), Mac OS X (Python 2.3, 2.4, and 2.5a2), and Linux (Python 2.4 and 2.5a2). ........ r46902 | tim.peters | 2006-06-13 02:30:01 +0200 (Tue, 13 Jun 2006) | 2 lines Whitespace normalization. ........ r46903 | tim.peters | 2006-06-13 02:30:50 +0200 (Tue, 13 Jun 2006) | 2 lines Added missing svn:eol-style property to text files. ........ r46905 | tim.peters | 2006-06-13 05:30:07 +0200 (Tue, 13 Jun 2006) | 5 lines get_matching_blocks(): rewrote code & comments so they match; added more comments about why it's this way at all; and removed what looked like needless expense (sorting (i, j, k) triples directly should give exactly the same order as sorting (i, (i, j, k)) pairs). ........ r46906 | neal.norwitz | 2006-06-13 06:08:53 +0200 (Tue, 13 Jun 2006) | 1 line Don't fail if another process is listening on our port. ........ r46908 | neal.norwitz | 2006-06-13 10:28:19 +0200 (Tue, 13 Jun 2006) | 2 lines Initialize the type object so pychecker can't crash the interpreter. ........ r46909 | neal.norwitz | 2006-06-13 10:41:06 +0200 (Tue, 13 Jun 2006) | 1 line Verify the crash due to EncodingMap not initialized does not return ........ r46910 | thomas.heller | 2006-06-13 10:56:14 +0200 (Tue, 13 Jun 2006) | 3 lines Add some windows datatypes that were missing from this file, and add the aliases defined in windows header files for the structures. ........ r46911 | thomas.heller | 2006-06-13 11:40:14 +0200 (Tue, 13 Jun 2006) | 3 lines Add back WCHAR, UINT, DOUBLE, _LARGE_INTEGER, _ULARGE_INTEGER. VARIANT_BOOL is a special _ctypes data type, not c_short. ........ r46912 | ronald.oussoren | 2006-06-13 13:19:56 +0200 (Tue, 13 Jun 2006) | 4 lines Linecache contains support for PEP302 loaders, but fails to deal with loaders that return None to indicate that the module is valid but no source is available. This patch fixes that. ........ r46913 | andrew.kuchling | 2006-06-13 13:57:04 +0200 (Tue, 13 Jun 2006) | 1 line Mention uuid module ........ r46915 | walter.doerwald | 2006-06-13 14:02:12 +0200 (Tue, 13 Jun 2006) | 2 lines Fix passing errors to the encoder and decoder functions. ........ r46917 | walter.doerwald | 2006-06-13 14:04:43 +0200 (Tue, 13 Jun 2006) | 3 lines errors is an attribute in the incremental decoder not an argument. ........ r46919 | andrew.macintyre | 2006-06-13 17:04:24 +0200 (Tue, 13 Jun 2006) | 11 lines Patch #1454481: Make thread stack size runtime tunable. Heavily revised, comprising revisions: 46640 - original trunk revision (backed out in r46655) 46647 - markup fix (backed out in r46655) 46692:46918 merged from branch aimacintyre-sf1454481 branch tested on buildbots (Windows buildbots had problems not related to these changes). ........ r46920 | brett.cannon | 2006-06-13 18:06:55 +0200 (Tue, 13 Jun 2006) | 2 lines Remove unused variable. ........ r46921 | andrew.kuchling | 2006-06-13 18:41:41 +0200 (Tue, 13 Jun 2006) | 1 line Add ability to set stack size ........ r46923 | marc-andre.lemburg | 2006-06-13 19:04:26 +0200 (Tue, 13 Jun 2006) | 2 lines Update pybench to version 2.0. ........ r46924 | marc-andre.lemburg | 2006-06-13 19:07:14 +0200 (Tue, 13 Jun 2006) | 2 lines Revert wrong svn copy. ........ r46925 | andrew.macintyre | 2006-06-13 19:14:36 +0200 (Tue, 13 Jun 2006) | 2 lines fix exception usage ........ r46927 | tim.peters | 2006-06-13 20:37:07 +0200 (Tue, 13 Jun 2006) | 2 lines Whitespace normalization. ........ r46928 | marc-andre.lemburg | 2006-06-13 20:56:56 +0200 (Tue, 13 Jun 2006) | 9 lines Updated to pybench 2.0. See svn.python.org/external/pybench-2.0 for the original import of that version. Note that platform.py was not copied over from pybench-2.0 since it is already part of Python 2.5. ........ r46929 | andrew.macintyre | 2006-06-13 21:02:35 +0200 (Tue, 13 Jun 2006) | 5 lines Increase the small thread stack size to get the test to pass reliably on the one buildbot that insists on more than 32kB of thread stack. ........ r46930 | marc-andre.lemburg | 2006-06-13 21:20:07 +0200 (Tue, 13 Jun 2006) | 2 lines Whitespace normalization. ........ r46931 | thomas.heller | 2006-06-13 22:18:43 +0200 (Tue, 13 Jun 2006) | 2 lines More docs for ctypes. ........ r46932 | brett.cannon | 2006-06-13 23:34:24 +0200 (Tue, 13 Jun 2006) | 2 lines Ignore .pyc and .pyo files in Pybench. ........ r46933 | brett.cannon | 2006-06-13 23:46:41 +0200 (Tue, 13 Jun 2006) | 7 lines If a classic class defined a __coerce__() method that just returned its two arguments in reverse, the interpreter would infinitely recourse trying to get a coercion that worked. So put in a recursion check after a coercion is made and the next call to attempt to use the coerced values. Fixes bug #992017 and closes crashers/coerce.py . ........ r46936 | gerhard.haering | 2006-06-14 00:24:47 +0200 (Wed, 14 Jun 2006) | 3 lines Merged changes from external pysqlite 2.3.0 release. Documentation updates will follow in a few hours at the latest. Then we should be ready for beta1. ........ r46937 | brett.cannon | 2006-06-14 00:26:13 +0200 (Wed, 14 Jun 2006) | 2 lines Missed test for rev. 46933; infinite recursion from __coerce__() returning its arguments reversed. ........ r46938 | gerhard.haering | 2006-06-14 00:53:48 +0200 (Wed, 14 Jun 2006) | 2 lines Updated documentation for pysqlite 2.3.0 API. ........ r46939 | tim.peters | 2006-06-14 06:09:25 +0200 (Wed, 14 Jun 2006) | 10 lines SequenceMatcher.get_matching_blocks(): This now guarantees that adjacent triples in the result list describe non-adjacent matching blocks. That's _nice_ to have, and Guido said he wanted it. Not a bugfix candidate: Guido or not ;-), this changes visible endcase semantics (note that some tests had to change), and nothing about this was documented before. Since it was working as designed, and behavior was consistent with the docs, it wasn't "a bug". ........ r46940 | tim.peters | 2006-06-14 06:13:00 +0200 (Wed, 14 Jun 2006) | 2 lines Repaired typo in new comment. ........ r46941 | tim.peters | 2006-06-14 06:15:27 +0200 (Wed, 14 Jun 2006) | 2 lines Whitespace normalization. ........ r46942 | fred.drake | 2006-06-14 06:25:02 +0200 (Wed, 14 Jun 2006) | 3 lines - make some disabled tests run what they intend when enabled - remove some over-zealous triple-quoting ........ r46943 | fred.drake | 2006-06-14 07:04:47 +0200 (Wed, 14 Jun 2006) | 3 lines add tests for two cases that are handled correctly in the current code, but that SF patch 1504676 as written mis-handles ........ r46944 | fred.drake | 2006-06-14 07:15:51 +0200 (Wed, 14 Jun 2006) | 1 line explain an XXX in more detail ........ r46945 | martin.v.loewis | 2006-06-14 07:21:04 +0200 (Wed, 14 Jun 2006) | 1 line Patch #1455898: Incremental mode for "mbcs" codec. ........ r46946 | georg.brandl | 2006-06-14 08:08:31 +0200 (Wed, 14 Jun 2006) | 3 lines Bug #1339007: Shelf objects now don't raise an exception in their __del__ method when initialization failed. ........ r46948 | thomas.heller | 2006-06-14 08:18:15 +0200 (Wed, 14 Jun 2006) | 1 line Fix docstring. ........ r46949 | georg.brandl | 2006-06-14 08:29:07 +0200 (Wed, 14 Jun 2006) | 2 lines Bug #1501122: mention __gt__ &co in description of comparison order. ........ r46951 | thomas.heller | 2006-06-14 09:08:38 +0200 (Wed, 14 Jun 2006) | 1 line Write more docs. ........ r46952 | georg.brandl | 2006-06-14 10:31:39 +0200 (Wed, 14 Jun 2006) | 3 lines Bug #1153163: describe __add__ vs __radd__ behavior when adding objects of same type/of subclasses of the other. ........ r46954 | georg.brandl | 2006-06-14 10:42:11 +0200 (Wed, 14 Jun 2006) | 3 lines Bug #1202018: add some common mime.types locations. ........ r46955 | georg.brandl | 2006-06-14 10:50:03 +0200 (Wed, 14 Jun 2006) | 3 lines Bug #1117556: SimpleHTTPServer now tries to find and use the system's mime.types file for determining MIME types. ........ r46957 | thomas.heller | 2006-06-14 11:09:08 +0200 (Wed, 14 Jun 2006) | 1 line Document paramflags. ........ r46958 | thomas.heller | 2006-06-14 11:20:11 +0200 (Wed, 14 Jun 2006) | 1 line Add an __all__ list, since this module does 'from ctypes import *'. ........ r46959 | andrew.kuchling | 2006-06-14 15:59:15 +0200 (Wed, 14 Jun 2006) | 1 line Add item ........ r46961 | georg.brandl | 2006-06-14 18:46:43 +0200 (Wed, 14 Jun 2006) | 3 lines Bug #805015: doc error in PyUnicode_FromEncodedObject. ........ r46962 | gerhard.haering | 2006-06-15 00:28:37 +0200 (Thu, 15 Jun 2006) | 10 lines - Added version checks in C code to make sure we don't trigger bugs in older SQLite versions. - Added version checks in test suite so that we don't execute tests that we know will fail with older (buggy) SQLite versions. Now, all tests should run against all SQLite versions from 3.0.8 until 3.3.6 (latest one now). The sqlite3 module can be built against all these SQLite versions and the sqlite3 module does its best to not trigger bugs in SQLite, but using SQLite 3.3.3 or later is recommended. ........ r46963 | tim.peters | 2006-06-15 00:38:13 +0200 (Thu, 15 Jun 2006) | 2 lines Whitespace normalization. ........ r46964 | neal.norwitz | 2006-06-15 06:54:29 +0200 (Thu, 15 Jun 2006) | 9 lines Speculative checkin (requires approval of Gerhard Haering) This backs out the test changes in 46962 which prevented crashes by not running the tests via a version check. All the version checks added in that rev were removed from the tests. Code was added to the error handler in connection.c that seems to work with older versions of sqlite including 3.1.3. ........ r46965 | neal.norwitz | 2006-06-15 07:55:49 +0200 (Thu, 15 Jun 2006) | 1 line Try to narrow window of failure on slow/busy boxes (ppc64 buildbot) ........ r46966 | martin.v.loewis | 2006-06-15 08:45:05 +0200 (Thu, 15 Jun 2006) | 2 lines Make import/lookup of mbcs fail on non-Windows systems. ........ r46967 | ronald.oussoren | 2006-06-15 10:14:18 +0200 (Thu, 15 Jun 2006) | 2 lines Patch #1446489 (zipfile: support for ZIP64) ........ r46968 | neal.norwitz | 2006-06-15 10:16:44 +0200 (Thu, 15 Jun 2006) | 6 lines Re-revert this change. Install the version check and don't run the test until Gerhard has time to fully debug the issue. This affects versions before 3.2.1 (possibly only versions earlier than 3.1.3). Based on discussion on python-checkins. ........ r46969 | gregory.p.smith | 2006-06-15 10:52:32 +0200 (Thu, 15 Jun 2006) | 6 lines - bsddb: multithreaded DB access using the simple bsddb module interface now works reliably. It has been updated to use automatic BerkeleyDB deadlock detection and the bsddb.dbutils.DeadlockWrap wrapper to retry database calls that would previously deadlock. [SF python bug #775414] ........ r46970 | gregory.p.smith | 2006-06-15 11:23:52 +0200 (Thu, 15 Jun 2006) | 2 lines minor documentation cleanup. mention the bsddb.db interface explicitly by name. ........ r46971 | neal.norwitz | 2006-06-15 11:57:03 +0200 (Thu, 15 Jun 2006) | 5 lines Steal the trick from test_compiler to print out a slow msg. This will hopefully get the buildbots to pass. Not sure this test will be feasible or even work. But everything is red now, so it can't get much worse. ........ r46972 | neal.norwitz | 2006-06-15 12:24:49 +0200 (Thu, 15 Jun 2006) | 1 line Print some more info to get an idea of how much longer the test will last ........ r46981 | tim.peters | 2006-06-15 20:04:40 +0200 (Thu, 15 Jun 2006) | 6 lines Try to reduce the extreme peak memory and disk-space use of this test. It probably still requires more disk space than most buildbots have, and in any case is still so intrusive that if we don't find another way to test this I'm taking my buildbot offline permanently ;-) ........ r46982 | tim.peters | 2006-06-15 20:06:29 +0200 (Thu, 15 Jun 2006) | 2 lines Whitespace normalization. ........ r46983 | tim.peters | 2006-06-15 20:07:28 +0200 (Thu, 15 Jun 2006) | 2 lines Add missing svn:eol-style property to text files. ........ r46984 | tim.peters | 2006-06-15 20:38:19 +0200 (Thu, 15 Jun 2006) | 2 lines Oops -- I introduced an off-by-6436159488 error. ........ r46990 | neal.norwitz | 2006-06-16 06:30:34 +0200 (Fri, 16 Jun 2006) | 1 line Disable this test until we can determine what to do about it ........ r46991 | neal.norwitz | 2006-06-16 06:31:06 +0200 (Fri, 16 Jun 2006) | 1 line Param name is dir, not directory. Update docstring. Backport candidate ........ r46992 | neal.norwitz | 2006-06-16 06:31:28 +0200 (Fri, 16 Jun 2006) | 1 line Add missing period in comment. ........ r46993 | neal.norwitz | 2006-06-16 06:32:43 +0200 (Fri, 16 Jun 2006) | 1 line Fix whitespace, there are memory leaks in this module. ........ r46995 | fred.drake | 2006-06-17 01:45:06 +0200 (Sat, 17 Jun 2006) | 3 lines SF patch 1504676: Make sgmllib char and entity references pluggable (implementation/tests contributed by Sam Ruby) ........ r46996 | fred.drake | 2006-06-17 03:07:54 +0200 (Sat, 17 Jun 2006) | 1 line fix change that broke the htmllib tests ........ r46998 | martin.v.loewis | 2006-06-17 11:15:14 +0200 (Sat, 17 Jun 2006) | 3 lines Patch #763580: Add name and value arguments to Tkinter variable classes. ........ r46999 | martin.v.loewis | 2006-06-17 11:20:41 +0200 (Sat, 17 Jun 2006) | 2 lines Patch #1096231: Add default argument to wm_iconbitmap. ........ r47000 | martin.v.loewis | 2006-06-17 11:25:15 +0200 (Sat, 17 Jun 2006) | 2 lines Patch #1494750: Destroy master after deleting children. ........ r47003 | george.yoshida | 2006-06-17 18:31:52 +0200 (Sat, 17 Jun 2006) | 2 lines markup fix ........ r47005 | george.yoshida | 2006-06-17 18:39:13 +0200 (Sat, 17 Jun 2006) | 4 lines Update url. Old url returned status code:301 Moved permanently. ........ r47007 | martin.v.loewis | 2006-06-17 20:44:27 +0200 (Sat, 17 Jun 2006) | 2 lines Patch #812986: Update the canvas even if not tracing. ........ r47008 | martin.v.loewis | 2006-06-17 21:03:26 +0200 (Sat, 17 Jun 2006) | 2 lines Patch #815924: Restore ability to pass type= and icon= ........ r47009 | neal.norwitz | 2006-06-18 00:37:45 +0200 (Sun, 18 Jun 2006) | 1 line Fix typo in docstring ........ r47010 | neal.norwitz | 2006-06-18 00:38:15 +0200 (Sun, 18 Jun 2006) | 1 line Fix memory leak reported by valgrind while running test_subprocess ........ r47011 | fred.drake | 2006-06-18 04:57:35 +0200 (Sun, 18 Jun 2006) | 1 line remove unnecessary markup ........ r47013 | neal.norwitz | 2006-06-18 21:35:01 +0200 (Sun, 18 Jun 2006) | 7 lines Prevent spurious leaks when running regrtest.py -R. There may be more issues that crop up from time to time, but this change seems to have been pretty stable (no spurious warnings) for about a week. Other modules which use threads may require similar use of threading_setup/threading_cleanup from test_support. ........ r47014 | neal.norwitz | 2006-06-18 21:37:40 +0200 (Sun, 18 Jun 2006) | 9 lines The hppa ubuntu box sometimes hangs forever in these tests. My guess is that the wait is failing for some reason. Use WNOHANG, so we won't wait until the buildbot kills the test suite. I haven't been able to reproduce the failure, so I'm not sure if this will help or not. Hopefully, this change will cause the test to fail, rather than hang. That will be better since we will get the rest of the test results. It may also help us debug the real problem. ........ r47015 | neal.norwitz | 2006-06-18 22:10:24 +0200 (Sun, 18 Jun 2006) | 1 line Revert 47014 until it is more robust ........ r47016 | thomas.heller | 2006-06-18 23:27:04 +0200 (Sun, 18 Jun 2006) | 6 lines Fix typos. Fix doctest example. Mention in the tutorial that 'errcheck' is explained in the ref manual. Use better wording in some places. Remoce code examples that shouldn't be in the tutorial. Remove some XXX notices. ........ r47017 | georg.brandl | 2006-06-19 00:17:29 +0200 (Mon, 19 Jun 2006) | 3 lines Patch #1507676: improve exception messages in abstract.c, object.c and typeobject.c. ........ r47018 | neal.norwitz | 2006-06-19 07:40:44 +0200 (Mon, 19 Jun 2006) | 1 line Use Py_ssize_t ........ r47019 | georg.brandl | 2006-06-19 08:35:54 +0200 (Mon, 19 Jun 2006) | 3 lines Add news entry about error msg improvement. ........ r47020 | thomas.heller | 2006-06-19 09:07:49 +0200 (Mon, 19 Jun 2006) | 2 lines Try to repair the failing test on the OpenBSD buildbot. Trial and error... ........ r47021 | tim.peters | 2006-06-19 09:45:16 +0200 (Mon, 19 Jun 2006) | 2 lines Whitespace normalization. ........ r47022 | walter.doerwald | 2006-06-19 10:07:50 +0200 (Mon, 19 Jun 2006) | 4 lines Patch #1506645: add Python wrappers for the curses functions is_term_resized, resize_term and resizeterm. This uses three separate configure checks (one for each function). ........ r47023 | walter.doerwald | 2006-06-19 10:14:09 +0200 (Mon, 19 Jun 2006) | 2 lines Make check order match in configure and configure.in. ........ r47024 | tim.peters | 2006-06-19 10:14:28 +0200 (Mon, 19 Jun 2006) | 3 lines Repair KeyError when running test_threaded_import under -R, as reported by Neal on python-dev. ........ r47025 | thomas.heller | 2006-06-19 10:32:46 +0200 (Mon, 19 Jun 2006) | 3 lines Next try to fix the OpenBSD buildbot tests: Use ctypes.util.find_library to locate the C runtime library on platforms where is returns useful results. ........ r47026 | tim.peters | 2006-06-19 11:09:44 +0200 (Mon, 19 Jun 2006) | 13 lines TestHelp.make_parser(): This was making a permanent change to os.environ (setting envar COLUMNS), which at least caused test_float_default() to fail if the tests were run more than once. This repairs the test_optparse -R failures Neal reported on python-dev. It also explains some seemingly bizarre test_optparse failures we saw a couple weeks ago on the buildbots, when test_optparse failed due to test_file failing to clean up after itself, and then test_optparse failed in an entirely different way when regrtest's -w option ran test_optparse a second time. It's now obvious that make_parser() permanently changing os.environ was responsible for the second half of that. ........ r47027 | anthony.baxter | 2006-06-19 14:04:15 +0200 (Mon, 19 Jun 2006) | 2 lines Preparing for 2.5b1. ........ r47029 | fred.drake | 2006-06-19 19:31:16 +0200 (Mon, 19 Jun 2006) | 1 line remove non-working document formats from edist ........ r47030 | gerhard.haering | 2006-06-19 23:17:35 +0200 (Mon, 19 Jun 2006) | 5 lines Fixed a memory leak that was introduced with incorrect usage of the Python weak reference API in pysqlite 2.2.1. Bumbed pysqlite version number to upcoming pysqlite 2.3.1 release. ........ r47032 | ka-ping.yee | 2006-06-20 00:49:36 +0200 (Tue, 20 Jun 2006) | 2 lines Remove Python 2.3 compatibility comment. ........ r47033 | trent.mick | 2006-06-20 01:21:25 +0200 (Tue, 20 Jun 2006) | 2 lines Upgrade pyexpat to expat 2.0.0 (http://python.org/sf/1462338). ........ r47034 | trent.mick | 2006-06-20 01:57:41 +0200 (Tue, 20 Jun 2006) | 3 lines [ 1295808 ] expat symbols should be namespaced in pyexpat (http://python.org/sf/1295808) ........ r47039 | andrew.kuchling | 2006-06-20 13:52:16 +0200 (Tue, 20 Jun 2006) | 1 line Uncomment wsgiref section ........ r47040 | andrew.kuchling | 2006-06-20 14:15:09 +0200 (Tue, 20 Jun 2006) | 1 line Add four library items ........ r47041 | andrew.kuchling | 2006-06-20 14:19:54 +0200 (Tue, 20 Jun 2006) | 1 line Terminology and typography fixes ........ r47042 | andrew.kuchling | 2006-06-20 15:05:12 +0200 (Tue, 20 Jun 2006) | 1 line Add introductory paragraphs summarizing the release; minor edits ........ r47043 | andrew.kuchling | 2006-06-20 15:11:29 +0200 (Tue, 20 Jun 2006) | 1 line Minor edits and rearrangements; markup fix ........ r47044 | andrew.kuchling | 2006-06-20 15:20:30 +0200 (Tue, 20 Jun 2006) | 1 line [Bug #1504456] Mention xml -> xmlcore change ........ r47047 | brett.cannon | 2006-06-20 19:30:26 +0200 (Tue, 20 Jun 2006) | 2 lines Raise TestSkipped when the test socket connection is refused. ........ r47049 | brett.cannon | 2006-06-20 21:20:17 +0200 (Tue, 20 Jun 2006) | 2 lines Fix typo of exception name. ........ r47053 | brett.cannon | 2006-06-21 18:57:57 +0200 (Wed, 21 Jun 2006) | 5 lines At the C level, tuple arguments are passed in directly to the exception constructor, meaning it is treated as *args, not as a single argument. This means using the 'message' attribute won't work (until Py3K comes around), and so one must grab from 'arg' to get the error number. ........ r47054 | andrew.kuchling | 2006-06-21 19:10:18 +0200 (Wed, 21 Jun 2006) | 1 line Link to LibRef module documentation ........ r47055 | andrew.kuchling | 2006-06-21 19:17:10 +0200 (Wed, 21 Jun 2006) | 1 line Note some of Barry's work ........ r47056 | andrew.kuchling | 2006-06-21 19:17:28 +0200 (Wed, 21 Jun 2006) | 1 line Bump version ........ r47057 | georg.brandl | 2006-06-21 19:45:17 +0200 (Wed, 21 Jun 2006) | 3 lines fix [ 1509132 ] compiler module builds incorrect AST for TryExceptFinally ........ r47058 | georg.brandl | 2006-06-21 19:52:36 +0200 (Wed, 21 Jun 2006) | 3 lines Make test_fcntl aware of netbsd3. ........ r47059 | georg.brandl | 2006-06-21 19:53:17 +0200 (Wed, 21 Jun 2006) | 3 lines Patch #1509001: expected skips for netbsd3. ........ r47060 | gerhard.haering | 2006-06-21 22:55:04 +0200 (Wed, 21 Jun 2006) | 2 lines Removed call to enable_callback_tracebacks that slipped in by accident. ........ r47061 | armin.rigo | 2006-06-21 23:58:50 +0200 (Wed, 21 Jun 2006) | 13 lines Fix for an obscure bug introduced by revs 46806 and 46808, with a test. The problem of checking too eagerly for recursive calls is the following: if a RuntimeError is caused by recursion, and if code needs to normalize it immediately (as in the 2nd test), then PyErr_NormalizeException() needs a call to the RuntimeError class to instantiate it, and this hits the recursion limit again... causing PyErr_NormalizeException() to never finish. Moved this particular recursion check to slot_tp_call(), which is not involved in instantiating built-in exceptions. Backport candidate. ........ r47064 | neal.norwitz | 2006-06-22 08:30:50 +0200 (Thu, 22 Jun 2006) | 3 lines Copy the wsgiref package during make install. ........ r47065 | neal.norwitz | 2006-06-22 08:35:30 +0200 (Thu, 22 Jun 2006) | 1 line Reset the doc date to today for the automatic doc builds ........ r47067 | andrew.kuchling | 2006-06-22 15:10:23 +0200 (Thu, 22 Jun 2006) | 1 line Mention how to suppress warnings ........ r47069 | georg.brandl | 2006-06-22 16:46:17 +0200 (Thu, 22 Jun 2006) | 3 lines Set lineno correctly on list, tuple and dict literals. ........ r47070 | georg.brandl | 2006-06-22 16:46:46 +0200 (Thu, 22 Jun 2006) | 4 lines Test for correct compilation of try-except-finally stmt. Test for correct lineno on list, tuple, dict literals. ........ r47071 | fred.drake | 2006-06-22 17:50:08 +0200 (Thu, 22 Jun 2006) | 1 line fix markup nit ........ r47072 | brett.cannon | 2006-06-22 18:49:14 +0200 (Thu, 22 Jun 2006) | 6 lines 'warning's was improperly requiring that a command-line Warning category be both a subclass of Warning and a subclass of types.ClassType. The latter is no longer true thanks to new-style exceptions. Closes bug #1510580. Thanks to AMK for the test. ........ r47073 | ronald.oussoren | 2006-06-22 20:33:54 +0200 (Thu, 22 Jun 2006) | 3 lines MacOSX: Add a message to the first screen of the installer that tells users how to avoid updates to their shell profile. ........ r47074 | georg.brandl | 2006-06-22 21:02:18 +0200 (Thu, 22 Jun 2006) | 3 lines Fix my name ;) ........ r47075 | thomas.heller | 2006-06-22 21:07:36 +0200 (Thu, 22 Jun 2006) | 2 lines Small fixes, mostly in the markup. ........ r47076 | peter.astrand | 2006-06-22 22:06:46 +0200 (Thu, 22 Jun 2006) | 1 line Make it possible to run test_subprocess.py on Python 2.2, which lacks test_support.is_resource_enabled. ........ r47077 | peter.astrand | 2006-06-22 22:21:26 +0200 (Thu, 22 Jun 2006) | 1 line Applied patch #1506758: Prevent MemoryErrors with large MAXFD. ........ r47079 | neal.norwitz | 2006-06-23 05:32:44 +0200 (Fri, 23 Jun 2006) | 1 line Fix refleak ........ r47080 | fred.drake | 2006-06-23 08:03:45 +0200 (Fri, 23 Jun 2006) | 9 lines - SF bug #853506: IP6 address parsing in sgmllib ('[' and ']' were not accepted in unquoted attribute values) - cleaned up tests of character and entity reference decoding so the tests cover the documented relationships among handle_charref, handle_entityref, convert_charref, convert_codepoint, and convert_entityref, without bringing up Unicode issues that sgmllib cannot be involved in ........ r47085 | andrew.kuchling | 2006-06-23 21:23:40 +0200 (Fri, 23 Jun 2006) | 11 lines Fit Makefile for the Python doc environment better; this is a step toward including the howtos in the build process. * Put LaTeX output in ../paper-<whatever>/. * Put HTML output in ../html/ * Explain some of the Makefile variables * Remove some cruft dating to my environment (e.g. the 'web' target) This makefile isn't currently invoked by the documentation build process, so these changes won't destabilize anything. ........ r47086 | hyeshik.chang | 2006-06-23 23:16:18 +0200 (Fri, 23 Jun 2006) | 5 lines Bug #1511381: codec_getstreamcodec() in codec.c is corrected to omit a default "error" argument for NULL pointer. This allows the parser to take a codec from cjkcodecs again. (Reported by Taewook Kang and reviewed by Walter Doerwald) ........ r47091 | ronald.oussoren | 2006-06-25 22:44:16 +0200 (Sun, 25 Jun 2006) | 6 lines Workaround for bug #1512124 Without this patch IDLE will get unresponsive when you open the debugger window on OSX. This is both using the system Tcl/Tk on Tiger as the latest universal download from tk-components.sf.net. ........ r47092 | ronald.oussoren | 2006-06-25 23:14:19 +0200 (Sun, 25 Jun 2006) | 3 lines Drop the calldll demo's for macos, calldll isn't present anymore, no need to keep the demo's around. ........ r47093 | ronald.oussoren | 2006-06-25 23:15:58 +0200 (Sun, 25 Jun 2006) | 3 lines Use a path without a double slash to compile the .py files after installation (macosx, binary installer). This fixes bug #1508369 for python 2.5. ........ r47094 | ronald.oussoren | 2006-06-25 23:19:06 +0200 (Sun, 25 Jun 2006) | 3 lines Also install the .egg-info files in Lib. This will cause wsgiref.egg-info to be installed. ........ r47097 | andrew.kuchling | 2006-06-26 14:40:02 +0200 (Mon, 26 Jun 2006) | 1 line [Bug #1511998] Various comments from Nick Coghlan; thanks! ........ r47098 | andrew.kuchling | 2006-06-26 14:43:43 +0200 (Mon, 26 Jun 2006) | 1 line Describe workaround for PyRange_New()'s removal ........ r47099 | andrew.kuchling | 2006-06-26 15:08:24 +0200 (Mon, 26 Jun 2006) | 5 lines [Bug #1512163] Fix typo. This change will probably break tests on FreeBSD buildbots, but I'll check in a fix for that next. ........ r47100 | andrew.kuchling | 2006-06-26 15:12:16 +0200 (Mon, 26 Jun 2006) | 9 lines [Bug #1512163] Use one set of locking methods, lockf(); remove the flock() calls. On FreeBSD, the two methods lockf() and flock() end up using the same mechanism and the second one fails. A Linux man page claims that the two methods are orthogonal (so locks acquired one way don't interact with locks acquired the other way) but that clearly must be false. ........ r47101 | andrew.kuchling | 2006-06-26 15:23:10 +0200 (Mon, 26 Jun 2006) | 5 lines Add a test for a conflicting lock. On slow machines, maybe the time intervals (2 sec, 0.5 sec) will be too tight. I'll see how the buildbots like it. ........ r47103 | andrew.kuchling | 2006-06-26 16:33:24 +0200 (Mon, 26 Jun 2006) | 1 line Windows doesn't have os.fork(). I'll just disable this test for now ........ r47106 | andrew.kuchling | 2006-06-26 19:00:35 +0200 (Mon, 26 Jun 2006) | 9 lines Attempt to fix build failure on OS X and Debian alpha; the symptom is consistent with os.wait() returning immediately because some other subprocess had previously exited; the test suite then immediately tries to lock the mailbox and gets an error saying it's already locked. To fix this, do a waitpid() so the test suite only continues once the intended child process has exited. ........ r47113 | neal.norwitz | 2006-06-27 06:06:46 +0200 (Tue, 27 Jun 2006) | 1 line Ignore some more warnings in the dynamic linker on an older gentoo ........ r47114 | neal.norwitz | 2006-06-27 06:09:13 +0200 (Tue, 27 Jun 2006) | 6 lines Instead of doing a make test, run the regression tests out of the installed copy. This will hopefully catch problems where directories are added under Lib/ but not to Makefile.pre.in. This breaks out the 2 runs of the test suite with and without -O which is also nicer. ........ r47115 | neal.norwitz | 2006-06-27 06:12:58 +0200 (Tue, 27 Jun 2006) | 5 lines Fix SF bug #1513032, 'make install' failure on FreeBSD 5.3. No need to install lib-old, it's empty in 2.5. ........ r47116 | neal.norwitz | 2006-06-27 06:23:06 +0200 (Tue, 27 Jun 2006) | 1 line Test unimportant change to verify buildbot does not try to build ........ r47117 | neal.norwitz | 2006-06-27 06:26:30 +0200 (Tue, 27 Jun 2006) | 1 line Try again: test unimportant change to verify buildbot does not try to build ........ r47118 | neal.norwitz | 2006-06-27 06:28:56 +0200 (Tue, 27 Jun 2006) | 1 line Verify buildbot picks up these changes (really needs testing after last change to Makefile.pre.in) ........ r47121 | vinay.sajip | 2006-06-27 09:34:37 +0200 (Tue, 27 Jun 2006) | 1 line Removed buggy exception handling in doRollover of rotating file handlers. Exceptions now propagate to caller. ........ r47123 | ronald.oussoren | 2006-06-27 12:08:25 +0200 (Tue, 27 Jun 2006) | 3 lines MacOSX: fix rather dumb buglet that made it impossible to create extensions on OSX 10.3 when using a binary distribution build on 10.4. ........ r47125 | tim.peters | 2006-06-27 13:52:49 +0200 (Tue, 27 Jun 2006) | 2 lines Whitespace normalization. ........ r47128 | ronald.oussoren | 2006-06-27 14:53:52 +0200 (Tue, 27 Jun 2006) | 8 lines Use staticly build copies of zlib and bzip2 to build the OSX installer, that way the resulting binaries have a better change of running on 10.3. This patch also updates the search logic for sleepycat db3/4, without this patch you cannot use a sleepycat build with a non-standard prefix; with this you can (at least on OSX) if you add the prefix to CPPFLAGS/LDFLAGS at configure-time. This change is needed to build the binary installer for OSX. ........ r47131 | ronald.oussoren | 2006-06-27 17:45:32 +0200 (Tue, 27 Jun 2006) | 5 lines macosx: Install a libpython2.5.a inside the framework as a symlink to the actual dylib at the root of the framework, that way tools that expect a unix-like install (python-config, but more importantly external products like mod_python) work correctly. ........ r47137 | neal.norwitz | 2006-06-28 07:03:22 +0200 (Wed, 28 Jun 2006) | 4 lines According to the man pages on Gentoo Linux and Tru64, EACCES or EAGAIN can be returned if fcntl (lockf) fails. This fixes the test failure on Tru64 by checking for either error rather than just EAGAIN. ........ r47139 | neal.norwitz | 2006-06-28 08:28:31 +0200 (Wed, 28 Jun 2006) | 5 lines Fix bug #1512695: cPickle.loads could crash if it was interrupted with a KeyboardInterrupt since PyTuple_Pack was passed a NULL. Will backport. ........ r47142 | nick.coghlan | 2006-06-28 12:41:47 +0200 (Wed, 28 Jun 2006) | 1 line Make full module name available as __module_name__ even when __name__ is set to something else (like '__main__') ........ r47143 | armin.rigo | 2006-06-28 12:49:51 +0200 (Wed, 28 Jun 2006) | 2 lines A couple of crashers of the "won't fix" kind. ........ r47147 | andrew.kuchling | 2006-06-28 16:25:20 +0200 (Wed, 28 Jun 2006) | 1 line [Bug #1508766] Add docs for uuid module; docs written by George Yoshida, with minor rearrangements by me. ........ r47148 | andrew.kuchling | 2006-06-28 16:27:21 +0200 (Wed, 28 Jun 2006) | 1 line [Bug #1508766] Add docs for uuid module; this puts the module in the 'Internet Protocols' section. Arguably this module could also have gone in the chapters on strings or encodings, maybe even the crypto chapter. Fred, please move if you see fit. ........ r47151 | georg.brandl | 2006-06-28 22:23:25 +0200 (Wed, 28 Jun 2006) | 3 lines Fix end_fill(). ........ r47153 | trent.mick | 2006-06-28 22:30:41 +0200 (Wed, 28 Jun 2006) | 2 lines Mention the expat upgrade and pyexpat fix I put in 2.5b1. ........ r47154 | fred.drake | 2006-06-29 02:51:53 +0200 (Thu, 29 Jun 2006) | 6 lines SF bug #1504333: sgmlib should allow angle brackets in quoted values (modified patch by Sam Ruby; changed to use separate REs for start and end tags to reduce matching cost for end tags; extended tests; updated to avoid breaking previous changes to support IPv6 addresses in unquoted attribute values) ........ r47156 | fred.drake | 2006-06-29 04:57:48 +0200 (Thu, 29 Jun 2006) | 1 line document recent bugfixes in sgmllib ........ r47158 | neal.norwitz | 2006-06-29 06:10:08 +0200 (Thu, 29 Jun 2006) | 10 lines Add new utility function, reap_children(), to test_support. This should be called at the end of each test that spawns children (perhaps it should be called from regrtest instead?). This will hopefully prevent some of the unexplained failures in the buildbots (hppa and alpha) during tests that spawn children. The problems were not reproducible. There were many zombies that remained at the end of several tests. In the worst case, this shouldn't cause any more problems, though it may not help either. Time will tell. ........ r47159 | neal.norwitz | 2006-06-29 07:48:14 +0200 (Thu, 29 Jun 2006) | 5 lines This should fix the buildbot failure on s/390 which can't connect to gmail.org. It makes the error message consistent and always sends to stderr. It would be much better for all the networking tests to hit only python.org. ........ r47161 | thomas.heller | 2006-06-29 20:34:15 +0200 (Thu, 29 Jun 2006) | 3 lines Protect the thread api calls in the _ctypes extension module within #ifdef WITH_THREADS/#endif blocks. Found by Sam Rushing. ........ r47162 | martin.v.loewis | 2006-06-29 20:58:44 +0200 (Thu, 29 Jun 2006) | 2 lines Patch #1509163: MS Toolkit Compiler no longer available ........ r47163 | skip.montanaro | 2006-06-29 21:20:09 +0200 (Thu, 29 Jun 2006) | 1 line add string methods to index ........ r47164 | vinay.sajip | 2006-06-30 02:13:08 +0200 (Fri, 30 Jun 2006) | 1 line Fixed bug in fileConfig() which failed to clear logging._handlerList ........ r47166 | tim.peters | 2006-06-30 08:18:39 +0200 (Fri, 30 Jun 2006) | 2 lines Whitespace normalization. ........ r47170 | neal.norwitz | 2006-06-30 09:32:16 +0200 (Fri, 30 Jun 2006) | 1 line Silence compiler warning ........ r47171 | neal.norwitz | 2006-06-30 09:32:46 +0200 (Fri, 30 Jun 2006) | 1 line Another problem reported by Coverity. Backport candidate. ........ r47175 | thomas.heller | 2006-06-30 19:44:54 +0200 (Fri, 30 Jun 2006) | 2 lines Revert the use of PY_FORMAT_SIZE_T in PyErr_Format. ........ r47176 | tim.peters | 2006-06-30 20:34:51 +0200 (Fri, 30 Jun 2006) | 2 lines Remove now-unused fidding with PY_FORMAT_SIZE_T. ........ r47177 | georg.brandl | 2006-06-30 20:47:56 +0200 (Fri, 30 Jun 2006) | 3 lines Document decorator usage of property. ........ r47181 | fred.drake | 2006-06-30 21:29:25 +0200 (Fri, 30 Jun 2006) | 4 lines - consistency nit: always include "()" in \function and \method (*should* be done by the presentation, but that requires changes all over) - avoid spreading the __name meme ........ r47188 | vinay.sajip | 2006-07-01 12:45:20 +0200 (Sat, 01 Jul 2006) | 1 line Added entry for fileConfig() bugfix. ........ r47189 | vinay.sajip | 2006-07-01 12:47:20 +0200 (Sat, 01 Jul 2006) | 1 line Added duplicate call to fileConfig() to ensure that it cleans up after itself correctly. ........ r47190 | martin.v.loewis | 2006-07-01 17:33:37 +0200 (Sat, 01 Jul 2006) | 2 lines Release all forwarded functions in .close. Fixes #1513223. ........ r47191 | fred.drake | 2006-07-01 18:28:20 +0200 (Sat, 01 Jul 2006) | 7 lines SF bug #1296433 (Expat bug #1515266): Unchecked calls to character data handler would cause a segfault. This merges in Expat's lib/xmlparse.c revisions 1.154 and 1.155, which fix this and a closely related problem (the later does not affect Python). Moved the crasher test to the tests for xml.parsers.expat. ........ r47197 | gerhard.haering | 2006-07-02 19:48:30 +0200 (Sun, 02 Jul 2006) | 4 lines The sqlite3 module did cut off data from the SQLite database at the first null character before sending it to a custom converter. This has been fixed now. ........ r47198 | martin.v.loewis | 2006-07-02 20:44:00 +0200 (Sun, 02 Jul 2006) | 1 line Correct arithmetic in access on Win32. Fixes #1513646. ........ r47203 | thomas.heller | 2006-07-03 09:58:09 +0200 (Mon, 03 Jul 2006) | 1 line Cleanup: Remove commented out code. ........ r47204 | thomas.heller | 2006-07-03 09:59:50 +0200 (Mon, 03 Jul 2006) | 1 line Don't run the doctests with Python 2.3 because it doesn't have the ELLIPSIS flag. ........ r47205 | thomas.heller | 2006-07-03 10:04:05 +0200 (Mon, 03 Jul 2006) | 7 lines Fixes so that _ctypes can be compiled with the MingW compiler. It seems that the definition of '__attribute__(x)' was responsible for the compiler ignoring the '__fastcall' attribute on the ffi_closure_SYSV function in libffi_msvc/ffi.c, took me quite some time to figure this out. ........ r47206 | thomas.heller | 2006-07-03 10:08:14 +0200 (Mon, 03 Jul 2006) | 11 lines Add a new function uses_seh() to the _ctypes extension module. This will return True if Windows Structured Exception handling (SEH) is used when calling functions, False otherwise. Currently, only MSVC supports SEH. Fix the test so that it doesn't crash when run with MingW compiled _ctypes. Note that two tests are still failing when mingw is used, I suspect structure layout differences and function calling conventions between MSVC and MingW. ........ r47207 | tim.peters | 2006-07-03 10:23:19 +0200 (Mon, 03 Jul 2006) | 2 lines Whitespace normalization. ........ r47208 | martin.v.loewis | 2006-07-03 11:44:00 +0200 (Mon, 03 Jul 2006) | 3 lines Only setup canvas when it is first created. Fixes #1514703 ........ r47209 | martin.v.loewis | 2006-07-03 12:05:30 +0200 (Mon, 03 Jul 2006) | 3 lines Reimplement turtle.circle using a polyline, to allow correct filling of arcs. Also fixes #1514693. ........ r47210 | martin.v.loewis | 2006-07-03 12:19:49 +0200 (Mon, 03 Jul 2006) | 3 lines Bug #1514693: Update turtle's heading when switching between degrees and radians. ........ r47211 | martin.v.loewis | 2006-07-03 13:12:06 +0200 (Mon, 03 Jul 2006) | 2 lines Document functions added in 2.3 and 2.5. ........ r47212 | martin.v.loewis | 2006-07-03 14:19:50 +0200 (Mon, 03 Jul 2006) | 3 lines Bug #1417699: Reject locale-specific decimal point in float() and atof(). ........ r47213 | martin.v.loewis | 2006-07-03 14:28:58 +0200 (Mon, 03 Jul 2006) | 3 lines Bug #1267547: Put proper recursive setup.py call into the spec file generated by bdist_rpm. ........ r47215 | martin.v.loewis | 2006-07-03 15:01:35 +0200 (Mon, 03 Jul 2006) | 3 lines Patch #825417: Fix timeout processing in expect, read_until. Will backport to 2.4. ........ r47218 | martin.v.loewis | 2006-07-03 15:47:40 +0200 (Mon, 03 Jul 2006) | 2 lines Put method-wrappers into trashcan. Fixes #927248. ........ r47219 | andrew.kuchling | 2006-07-03 16:07:30 +0200 (Mon, 03 Jul 2006) | 1 line [Bug #1515932] Clarify description of slice assignment ........ r47220 | andrew.kuchling | 2006-07-03 16:16:09 +0200 (Mon, 03 Jul 2006) | 4 lines [Bug #1511911] Clarify description of optional arguments to sorted() by improving the xref to the section on lists, and by copying the explanations of the arguments (with a slight modification). ........ r47223 | kristjan.jonsson | 2006-07-03 16:59:05 +0200 (Mon, 03 Jul 2006) | 1 line Fix build problems with the platform SDK on windows. It is not sufficient to test for the C compiler version when determining if we have the secure CRT from microsoft. Must test with an undocumented macro, __STDC_SECURE_LIB__ too. ........ r47224 | ronald.oussoren | 2006-07-04 14:30:22 +0200 (Tue, 04 Jul 2006) | 7 lines Sync the darwin/x86 port libffi with the copy in PyObjC. This fixes a number of bugs in that port. The most annoying ones were due to some subtle differences between the document ABI and the actual implementation :-( (there are no python unittests that fail without this patch, but without it some of libffi's unittests fail). ........ r47234 | georg.brandl | 2006-07-05 10:21:00 +0200 (Wed, 05 Jul 2006) | 3 lines Remove remaining references to OverflowWarning. ........ r47236 | thomas.heller | 2006-07-05 11:13:56 +0200 (Wed, 05 Jul 2006) | 3 lines Fix the bitfield test when _ctypes is compiled with MingW. Structures containing bitfields may have different layout on MSVC and MingW . ........ r47237 | thomas.wouters | 2006-07-05 13:03:49 +0200 (Wed, 05 Jul 2006) | 15 lines Fix bug in passing tuples to string.Template. All other values (with working str() or repr()) would work, just not multi-value tuples. Probably not a backport candidate, since it changes the behaviour of passing a single-element tuple: >>> string.Template("$foo").substitute(dict(foo=(1,))) '(1,)' versus '1' ........ r47241 | georg.brandl | 2006-07-05 16:18:45 +0200 (Wed, 05 Jul 2006) | 2 lines Patch #1517490: fix glitches in filter() docs. ........ r47244 | georg.brandl | 2006-07-05 17:50:05 +0200 (Wed, 05 Jul 2006) | 2 lines no need to elaborate "string". ........ r47251 | neal.norwitz | 2006-07-06 06:28:59 +0200 (Thu, 06 Jul 2006) | 3 lines Fix refleaks reported by Shane Hathaway in SF patch #1515361. This change contains only the changes related to leaking the copy variable. ........ r47253 | fred.drake | 2006-07-06 07:13:22 +0200 (Thu, 06 Jul 2006) | 4 lines - back out Expat change; the final fix to Expat will be different - change the pyexpat wrapper to not be so sensitive to this detail of the Expat implementation (the ex-crasher test still passes) ........ r47257 | neal.norwitz | 2006-07-06 08:45:08 +0200 (Thu, 06 Jul 2006) | 1 line Add a NEWS entry for a recent pyexpat fix ........ r47258 | martin.v.loewis | 2006-07-06 08:55:58 +0200 (Thu, 06 Jul 2006) | 2 lines Add sqlite3.dll to the DLLs component, not to the TkDLLs component. Fixes #1517388. ........ r47259 | martin.v.loewis | 2006-07-06 09:05:21 +0200 (Thu, 06 Jul 2006) | 1 line Properly quote compileall and Lib paths in case TARGETDIR has a space. ........ r47260 | thomas.heller | 2006-07-06 09:50:18 +0200 (Thu, 06 Jul 2006) | 5 lines Revert the change done in svn revision 47206: Add a new function uses_seh() to the _ctypes extension module. This will return True if Windows Structured Exception handling (SEH) is used when calling functions, False otherwise. ........ r47261 | armin.rigo | 2006-07-06 09:58:18 +0200 (Thu, 06 Jul 2006) | 3 lines A couple of examples about how to attack the fact that _PyType_Lookup() returns a borrowed ref. Many of the calls are open to attack. ........ r47262 | thomas.heller | 2006-07-06 10:28:14 +0200 (Thu, 06 Jul 2006) | 2 lines The test that calls a function with invalid arguments and catches the resulting Windows access violation will not be run by default. ........ r47263 | thomas.heller | 2006-07-06 10:48:35 +0200 (Thu, 06 Jul 2006) | 5 lines Patch #1517790: It is now possible to use custom objects in the ctypes foreign function argtypes sequence as long as they provide a from_param method, no longer is it required that the object is a ctypes type. ........ r47264 | thomas.heller | 2006-07-06 10:58:40 +0200 (Thu, 06 Jul 2006) | 2 lines Document the Struture and Union constructors. ........ r47265 | thomas.heller | 2006-07-06 11:11:22 +0200 (Thu, 06 Jul 2006) | 2 lines Document the changes in svn revision 47263, from patch #1517790. ........ r47267 | ronald.oussoren | 2006-07-06 12:13:35 +0200 (Thu, 06 Jul 2006) | 7 lines This patch solves the problem Skip was seeing with zlib, this patch ensures that configure uses similar compiler flags as setup.py when doing the zlib test. Without this patch configure would use the first shared library on the linker path, with this patch it uses the first shared or static library on that path just like setup.py. ........ r47268 | thomas.wouters | 2006-07-06 12:48:28 +0200 (Thu, 06 Jul 2006) | 4 lines NEWS entry for r47267: fixing configure's zlib probing. ........ r47269 | fredrik.lundh | 2006-07-06 14:29:24 +0200 (Thu, 06 Jul 2006) | 3 lines added XMLParser alias for cElementTree compatibility ........ r47271 | nick.coghlan | 2006-07-06 14:53:04 +0200 (Thu, 06 Jul 2006) | 1 line Revert the __module_name__ changes made in rev 47142. We'll revisit this in Python 2.6 ........ r47272 | nick.coghlan | 2006-07-06 15:04:56 +0200 (Thu, 06 Jul 2006) | 1 line Update the tutorial section on relative imports ........ r47273 | nick.coghlan | 2006-07-06 15:35:27 +0200 (Thu, 06 Jul 2006) | 1 line Ignore ImportWarning by default ........ r47274 | nick.coghlan | 2006-07-06 15:41:34 +0200 (Thu, 06 Jul 2006) | 1 line Cover ImportWarning, PendingDeprecationWarning and simplefilter() in the warnings module docs ........ r47275 | nick.coghlan | 2006-07-06 15:47:18 +0200 (Thu, 06 Jul 2006) | 1 line Add NEWS entries for the ImportWarning change and documentation update ........ r47276 | andrew.kuchling | 2006-07-06 15:57:28 +0200 (Thu, 06 Jul 2006) | 1 line ImportWarning is now silent by default ........ r47277 | thomas.heller | 2006-07-06 17:06:05 +0200 (Thu, 06 Jul 2006) | 2 lines Document the correct return type of PyLong_AsUnsignedLongLongMask. ........ r47278 | hyeshik.chang | 2006-07-06 17:21:52 +0200 (Thu, 06 Jul 2006) | 2 lines Add a testcase for r47086 which fixed a bug in codec_getstreamcodec(). ........ r47279 | hyeshik.chang | 2006-07-06 17:39:24 +0200 (Thu, 06 Jul 2006) | 3 lines Test using all CJK encodings for the testcases which don't require specific encodings. ........ r47280 | martin.v.loewis | 2006-07-06 21:28:03 +0200 (Thu, 06 Jul 2006) | 2 lines Properly generate logical file ids. Fixes #1515998. Also correct typo in Control.mapping. ........ r47287 | neal.norwitz | 2006-07-07 08:03:15 +0200 (Fri, 07 Jul 2006) | 17 lines Restore rev 47014: The hppa ubuntu box sometimes hangs forever in these tests. My guess is that the wait is failing for some reason. Use WNOHANG, so we won't wait until the buildbot kills the test suite. I haven't been able to reproduce the failure, so I'm not sure if this will help or not. Hopefully, this change will cause the test to fail, rather than hang. That will be better since we will get the rest of the test results. It may also help us debug the real problem. *** The reason this originally failed was because there were many zombie children outstanding before rev 47158 cleaned them up. There are still hangs in test_subprocess that need to be addressed, but that will take more work. This should close some holes. ........ r47289 | georg.brandl | 2006-07-07 10:15:12 +0200 (Fri, 07 Jul 2006) | 3 lines Fix RFC number. ........ r50489 | neal.norwitz | 2006-07-08 07:31:37 +0200 (Sat, 08 Jul 2006) | 1 line Fix SF bug #1519018: 'as' is now validated properly in import statements ........ r50490 | georg.brandl | 2006-07-08 14:15:27 +0200 (Sat, 08 Jul 2006) | 3 lines Add an additional test for bug #1519018. ........ r50491 | tim.peters | 2006-07-08 21:55:05 +0200 (Sat, 08 Jul 2006) | 2 lines Whitespace normalization. ........ r50493 | neil.schemenauer | 2006-07-09 18:16:34 +0200 (Sun, 09 Jul 2006) | 2 lines Fix AST compiler bug #1501934: incorrect LOAD/STORE_GLOBAL generation. ........ r50495 | neil.schemenauer | 2006-07-09 23:19:29 +0200 (Sun, 09 Jul 2006) | 2 lines Fix SF bug 1441486: bad unary minus folding in compiler. ........ r50497 | neal.norwitz | 2006-07-10 00:14:42 +0200 (Mon, 10 Jul 2006) | 4 lines On 64 bit systems, int literals that use less than 64 bits are now ints rather than longs. This also fixes the test for eval(-sys.maxint - 1). ........ r50500 | neal.norwitz | 2006-07-10 02:04:44 +0200 (Mon, 10 Jul 2006) | 4 lines Bug #1512814, Fix incorrect lineno's when code at module scope started after line 256. ........ r50501 | neal.norwitz | 2006-07-10 02:05:34 +0200 (Mon, 10 Jul 2006) | 1 line Fix doco. Backport candidate. ........ r50503 | neal.norwitz | 2006-07-10 02:23:17 +0200 (Mon, 10 Jul 2006) | 5 lines Part of SF patch #1484695. This removes dead code. The chksum was already verified in .frombuf() on the lines above. If there was a problem an exception is raised, so there was no way this condition could have been true. ........ r50504 | neal.norwitz | 2006-07-10 03:18:57 +0200 (Mon, 10 Jul 2006) | 3 lines Patch #1516912: improve Modules support for OpenVMS. ........ r50506 | neal.norwitz | 2006-07-10 04:36:41 +0200 (Mon, 10 Jul 2006) | 7 lines Patch #1504046: Add documentation for xml.etree. /F wrote the text docs, Englebert Gruber massaged it to latex and I did some more massaging to try and improve the consistency and fix some name mismatches between the declaration and text. ........ r50509 | martin.v.loewis | 2006-07-10 09:23:48 +0200 (Mon, 10 Jul 2006) | 2 lines Introduce DISTUTILS_USE_SDK as a flag to determine whether the SDK environment should be used. Fixes #1508010. ........ r50510 | martin.v.loewis | 2006-07-10 09:26:41 +0200 (Mon, 10 Jul 2006) | 1 line Change error message to indicate that VS2003 is necessary to build extension modules, not the .NET SDK. ........ r50511 | martin.v.loewis | 2006-07-10 09:29:41 +0200 (Mon, 10 Jul 2006) | 1 line Add svn:ignore. ........ r50512 | anthony.baxter | 2006-07-10 09:41:04 +0200 (Mon, 10 Jul 2006) | 1 line preparing for 2.5b2 ........ r50513 | thomas.heller | 2006-07-10 11:10:28 +0200 (Mon, 10 Jul 2006) | 2 lines Fix bug #1518190: accept any integer or long value in the ctypes.c_void_p constructor. ........ r50514 | thomas.heller | 2006-07-10 11:31:06 +0200 (Mon, 10 Jul 2006) | 3 lines Fixed a segfault when ctypes.wintypes were imported on non-Windows machines. ........ r50516 | thomas.heller | 2006-07-10 13:11:10 +0200 (Mon, 10 Jul 2006) | 3 lines Assigning None to pointer type structure fields possible overwrote wrong fields. ........ r50517 | thomas.heller | 2006-07-10 13:17:37 +0200 (Mon, 10 Jul 2006) | 5 lines Moved the ctypes news entries from the 'Library' section into the 'Extension Modules' section where they belong, probably. This destroyes the original order of the news entries, don't know if that is important or not. ........ r50526 | phillip.eby | 2006-07-10 21:03:29 +0200 (Mon, 10 Jul 2006) | 2 lines Fix SF#1516184 and add a test to prevent regression. ........ r50528 | phillip.eby | 2006-07-10 21:18:35 +0200 (Mon, 10 Jul 2006) | 2 lines Fix SF#1457312: bad socket error handling in distutils "upload" command. ........ r50537 | peter.astrand | 2006-07-10 22:39:49 +0200 (Mon, 10 Jul 2006) | 1 line Make it possible to run test_subprocess.py with Python 2.2, which lacks test_support.reap_children(). ........ r50541 | tim.peters | 2006-07-10 23:08:24 +0200 (Mon, 10 Jul 2006) | 5 lines After approval from Anthony, merge the tim-current_frames branch into the trunk. This adds a new sys._current_frames() function, which returns a dict mapping thread id to topmost thread stack frame. ........ r50542 | tim.peters | 2006-07-10 23:11:49 +0200 (Mon, 10 Jul 2006) | 2 lines Whitespace normalization. ........ r50553 | martin.v.loewis | 2006-07-11 00:11:28 +0200 (Tue, 11 Jul 2006) | 4 lines Patch #1519566: Remove unused _tofill member. Make begin_fill idempotent. Update demo2 to demonstrate filling of concave shapes. ........ r50567 | anthony.baxter | 2006-07-11 04:04:09 +0200 (Tue, 11 Jul 2006) | 4 lines #1494314: Fix a regression with high-numbered sockets in 2.4.3. This means that select() on sockets > FD_SETSIZE (typically 1024) work again. The patch makes sockets use poll() internally where available. ........ r50568 | tim.peters | 2006-07-11 04:17:48 +0200 (Tue, 11 Jul 2006) | 2 lines Whitespace normalization. ........ r50575 | thomas.heller | 2006-07-11 18:42:05 +0200 (Tue, 11 Jul 2006) | 1 line Add missing Py_DECREF. ........ r50576 | thomas.heller | 2006-07-11 18:44:25 +0200 (Tue, 11 Jul 2006) | 1 line Add missing Py_DECREFs. ........ r50579 | andrew.kuchling | 2006-07-11 19:20:16 +0200 (Tue, 11 Jul 2006) | 1 line Bump version number; add sys._current_frames ........ r50582 | thomas.heller | 2006-07-11 20:28:35 +0200 (Tue, 11 Jul 2006) | 3 lines When a foreign function is retrived by calling __getitem__ on a ctypes library instance, do not set it as attribute. ........ r50583 | thomas.heller | 2006-07-11 20:40:50 +0200 (Tue, 11 Jul 2006) | 2 lines Change the ctypes version number to 1.0.0. ........ r50597 | neal.norwitz | 2006-07-12 07:26:17 +0200 (Wed, 12 Jul 2006) | 3 lines Bug #1520864: unpacking singleton tuples in for loop (for x, in) work again. ........ r50598 | neal.norwitz | 2006-07-12 07:26:35 +0200 (Wed, 12 Jul 2006) | 1 line Fix function name in error msg ........ r50599 | neal.norwitz | 2006-07-12 07:27:46 +0200 (Wed, 12 Jul 2006) | 4 lines Fix uninitialized memory read reported by Valgrind when running doctest. This could happen if size == 0. ........ r50600 | neal.norwitz | 2006-07-12 09:28:29 +0200 (Wed, 12 Jul 2006) | 1 line Actually change the MAGIC #. Create a new section for 2.5c1 and mention the impact of changing the MAGIC #. ........ r50601 | thomas.heller | 2006-07-12 10:43:47 +0200 (Wed, 12 Jul 2006) | 3 lines Fix #1467450: ctypes now uses RTLD_GLOBAL by default on OSX 10.3 to load shared libraries. ........ r50604 | thomas.heller | 2006-07-12 16:25:18 +0200 (Wed, 12 Jul 2006) | 3 lines Fix the wrong description of LibraryLoader.LoadLibrary, and document the DEFAULT_MODE constant. ........ r50607 | georg.brandl | 2006-07-12 17:31:17 +0200 (Wed, 12 Jul 2006) | 3 lines Accept long options "--help" and "--version". ........ r50617 | thomas.heller | 2006-07-13 11:53:47 +0200 (Thu, 13 Jul 2006) | 3 lines A misspelled preprocessor symbol caused ctypes to be always compiled without thread support. Replaced WITH_THREADS with WITH_THREAD. ........ r50619 | thomas.heller | 2006-07-13 19:01:14 +0200 (Thu, 13 Jul 2006) | 3 lines Fix #1521375. When running with root priviledges, 'gcc -o /dev/null' did overwrite /dev/null. Use a temporary file instead of /dev/null. ........ r50620 | thomas.heller | 2006-07-13 19:05:13 +0200 (Thu, 13 Jul 2006) | 2 lines Fix misleading words. ........ r50622 | andrew.kuchling | 2006-07-13 19:37:26 +0200 (Thu, 13 Jul 2006) | 1 line Typo fix ........ r50629 | georg.brandl | 2006-07-14 09:12:54 +0200 (Fri, 14 Jul 2006) | 3 lines Patch #1521874: grammar errors in doanddont.tex. ........ r50630 | neal.norwitz | 2006-07-14 09:20:04 +0200 (Fri, 14 Jul 2006) | 1 line Try to improve grammar further. ........ r50631 | martin.v.loewis | 2006-07-14 11:58:55 +0200 (Fri, 14 Jul 2006) | 1 line Extend build_ssl to Win64, using VSExtComp. ........ r50632 | martin.v.loewis | 2006-07-14 14:10:09 +0200 (Fri, 14 Jul 2006) | 1 line Add debug output to analyse buildbot failure. ........ r50633 | martin.v.loewis | 2006-07-14 14:31:05 +0200 (Fri, 14 Jul 2006) | 1 line Fix Debug build of _ssl. ........ r50636 | andrew.kuchling | 2006-07-14 15:32:38 +0200 (Fri, 14 Jul 2006) | 1 line Mention new options ........ r50638 | peter.astrand | 2006-07-14 16:04:45 +0200 (Fri, 14 Jul 2006) | 1 line Bug #1223937: CalledProcessError.errno -> CalledProcessError.returncode. ........ r50640 | thomas.heller | 2006-07-14 17:01:05 +0200 (Fri, 14 Jul 2006) | 4 lines Make the prototypes of our private PyUnicode_FromWideChar and PyUnicode_AsWideChar replacement functions compatible to the official functions by using Py_ssize_t instead of int. ........ r50643 | thomas.heller | 2006-07-14 19:51:14 +0200 (Fri, 14 Jul 2006) | 3 lines Patch #1521817: The index range checking on ctypes arrays containing exactly one element is enabled again. ........ r50647 | thomas.heller | 2006-07-14 20:22:50 +0200 (Fri, 14 Jul 2006) | 2 lines Updates for the ctypes documentation. ........ r50655 | fredrik.lundh | 2006-07-14 23:45:48 +0200 (Fri, 14 Jul 2006) | 3 lines typo ........ r50664 | george.yoshida | 2006-07-15 18:03:49 +0200 (Sat, 15 Jul 2006) | 2 lines Bug #15187702 : ext/win-cookbook.html has a broken link to distutils ........ r50667 | bob.ippolito | 2006-07-15 18:53:15 +0200 (Sat, 15 Jul 2006) | 1 line Patch #1220874: Update the binhex module for Mach-O. ........ r50671 | fred.drake | 2006-07-16 03:21:20 +0200 (Sun, 16 Jul 2006) | 1 line clean up some link markup ........ r50673 | neal.norwitz | 2006-07-16 03:50:38 +0200 (Sun, 16 Jul 2006) | 4 lines Bug #1512814, Fix incorrect lineno's when code within a function had more than 255 blank lines. Byte codes need to go first, line #s second. ........ r50674 | neal.norwitz | 2006-07-16 04:00:32 +0200 (Sun, 16 Jul 2006) | 5 lines a & b were dereffed above, so they are known to be valid pointers. z is known to be NULL, nothing to DECREF. Reported by Klockwork, #107. ........ r50675 | neal.norwitz | 2006-07-16 04:02:57 +0200 (Sun, 16 Jul 2006) | 5 lines self is dereffed (and passed as first arg), so it's known to be good. func is returned from PyArg_ParseTuple and also dereffed. Reported by Klocwork, #30 (self one at least). ........ r50676 | neal.norwitz | 2006-07-16 04:05:35 +0200 (Sun, 16 Jul 2006) | 4 lines proto was dereffed above and is known to be good. No need for X. Reported by Klocwork, #39. ........ r50677 | neal.norwitz | 2006-07-16 04:15:27 +0200 (Sun, 16 Jul 2006) | 5 lines Fix memory leaks in some conditions. Reported by Klocwork #152. ........ r50678 | neal.norwitz | 2006-07-16 04:17:36 +0200 (Sun, 16 Jul 2006) | 4 lines Fix memory leak under some conditions. Reported by Klocwork, #98. ........ r50679 | neal.norwitz | 2006-07-16 04:22:30 +0200 (Sun, 16 Jul 2006) | 8 lines Use sizeof(buffer) instead of duplicating the constants to ensure they won't be wrong. The real change is to pass (bufsz - 1) to PyOS_ascii_formatd and 1 to strncat. strncat copies n+1 bytes from src (not dest). Reported by Klocwork #58. ........ r50680 | neal.norwitz | 2006-07-16 04:32:03 +0200 (Sun, 16 Jul 2006) | 5 lines Handle a NULL name properly. Reported by Klocwork #67 ........ r50681 | neal.norwitz | 2006-07-16 04:35:47 +0200 (Sun, 16 Jul 2006) | 6 lines PyFunction_SetDefaults() is documented as taking None or a tuple. A NULL would crash the PyTuple_Check(). Now make NULL return a SystemError. Reported by Klocwork #73. ........ r50683 | neal.norwitz | 2006-07-17 02:55:45 +0200 (Mon, 17 Jul 2006) | 5 lines Stop INCREFing name, then checking if it's NULL. name (f_name) should never be NULL so assert it. Fix one place where we could have passed NULL. Reported by Klocwork #66. ........ r50684 | neal.norwitz | 2006-07-17 02:57:15 +0200 (Mon, 17 Jul 2006) | 5 lines otherset is known to be non-NULL based on checks before and DECREF after. DECREF otherset rather than XDECREF in error conditions too. Reported by Klockwork #154. ........ r50685 | neal.norwitz | 2006-07-17 02:59:04 +0200 (Mon, 17 Jul 2006) | 7 lines Reported by Klocwork #151. v2 can be NULL if exception2 is NULL. I don't think that condition can happen, but I'm not sure it can't either. Now the code will protect against either being NULL. ........ r50686 | neal.norwitz | 2006-07-17 03:00:16 +0200 (Mon, 17 Jul 2006) | 1 line Add NEWS entry for a bunch of fixes due to warnings produced by Klocworks static analysis tool. ........ r50687 | fred.drake | 2006-07-17 07:47:52 +0200 (Mon, 17 Jul 2006) | 3 lines document xmlcore (still minimal; needs mention in each of the xml.* modules) SF bug #1504456 (partial) ........ r50688 | georg.brandl | 2006-07-17 15:23:46 +0200 (Mon, 17 Jul 2006) | 3 lines Remove usage of sets module (patch #1500609). ........ r50689 | georg.brandl | 2006-07-17 15:26:33 +0200 (Mon, 17 Jul 2006) | 3 lines Add missing NEWS item (#1522771) ........ r50690 | andrew.kuchling | 2006-07-17 18:47:54 +0200 (Mon, 17 Jul 2006) | 1 line Attribute more features ........ r50692 | kurt.kaiser | 2006-07-17 23:59:27 +0200 (Mon, 17 Jul 2006) | 8 lines Patch 1479219 - Tal Einat 1. 'as' highlighted as builtin in comment string on import line 2. Comments such as "#False identity" which start with a keyword immediately after the '#' character aren't colored as comments. 3. u or U beginning unicode string not correctly highlighted Closes bug 1325071 ........ r50693 | barry.warsaw | 2006-07-18 01:07:51 +0200 (Tue, 18 Jul 2006) | 16 lines decode_rfc2231(): Be more robust against buggy RFC 2231 encodings. Specifically, instead of raising a ValueError when there is a single tick in the parameter, simply return that the entire string unquoted, with None for both the charset and the language. Also, if there are more than 2 ticks in the parameter, interpret the first three parts as the standard RFC 2231 parts, then the rest of the parts as the encoded string. Test cases added. Original fewer-than-3-parts fix by Tokio Kikuchi. Resolves SF bug # 1218081. I will back port the fix and tests to Python 2.4 (email 3.0) and Python 2.3 (email 2.5). Also, bump the version number to email 4.0.1, removing the 'alpha' moniker. ........ r50695 | kurt.kaiser | 2006-07-18 06:03:16 +0200 (Tue, 18 Jul 2006) | 2 lines Rebinding Tab key was inserting 'tab' instead of 'Tab'. Bug 1179168. ........ r50696 | brett.cannon | 2006-07-18 06:41:36 +0200 (Tue, 18 Jul 2006) | 6 lines Fix bug #1520914. Starting in 2.4, time.strftime() began to check the bounds of values in the time tuple passed in. Unfortunately people came to rely on undocumented behaviour of setting unneeded values to 0, regardless of if it was within the valid range. Now those values force the value internally to the minimum value when 0 is passed in. ........ r50697 | facundo.batista | 2006-07-18 14:16:13 +0200 (Tue, 18 Jul 2006) | 1 line Comments and docs cleanups, and some little fixes, provided by Santiágo Peresón ........ r50704 | martin.v.loewis | 2006-07-18 19:46:31 +0200 (Tue, 18 Jul 2006) | 2 lines Patch #1524429: Use repr instead of backticks again. ........ r50706 | tim.peters | 2006-07-18 23:55:15 +0200 (Tue, 18 Jul 2006) | 2 lines Whitespace normalization. ........ r50708 | tim.peters | 2006-07-19 02:03:19 +0200 (Wed, 19 Jul 2006) | 18 lines SF bug 1524317: configure --without-threads fails to build Moved the code for _PyThread_CurrentFrames() up, so it's no longer in a huge "#ifdef WITH_THREAD" block (I didn't realize it /was/ in one). Changed test_sys's test_current_frames() so it passes with or without thread supported compiled in. Note that test_sys fails when Python is compiled without threads, but for an unrelated reason (the old test_exit() fails with an indirect ImportError on the `thread` module). There are also other unrelated compilation failures without threads, in extension modules (like ctypes); at least the core compiles again. Do we really support --without-threads? If so, there are several problems remaining. ........ r50713 | thomas.heller | 2006-07-19 11:09:32 +0200 (Wed, 19 Jul 2006) | 4 lines Make sure the _ctypes extension can be compiled when WITH_THREAD is not defined on Windows, even if that configuration is probably not supported at all. ........ r50715 | martin.v.loewis | 2006-07-19 19:18:32 +0200 (Wed, 19 Jul 2006) | 4 lines Revert r50706 (Whitespace normalization) and r50697: Comments and docs cleanups, and some little fixes per recommendation from Raymond Hettinger. ........ r50719 | phillip.eby | 2006-07-20 17:54:16 +0200 (Thu, 20 Jul 2006) | 4 lines Fix SF#1516184 (again) and add a test to prevent regression. (There was a problem with empty filenames still causing recursion) ........ r50720 | georg.brandl | 2006-07-20 18:28:39 +0200 (Thu, 20 Jul 2006) | 3 lines Guard for _active being None in __del__ method. ........ r50721 | vinay.sajip | 2006-07-20 18:28:39 +0200 (Thu, 20 Jul 2006) | 1 line Updated documentation for TimedRotatingFileHandler relating to how rollover files are named. The previous documentation was wrongly the same as for RotatingFileHandler. ........ r50731 | fred.drake | 2006-07-20 22:11:57 +0200 (Thu, 20 Jul 2006) | 1 line markup fix ........ r50739 | kurt.kaiser | 2006-07-21 00:22:52 +0200 (Fri, 21 Jul 2006) | 7 lines Avoid occasional failure to detect closing paren properly. Patch 1407280 Tal Einat M ParenMatch.py M NEWS.txt M CREDITS.txt ........ r50740 | vinay.sajip | 2006-07-21 01:20:12 +0200 (Fri, 21 Jul 2006) | 1 line Addressed SF#1524081 by using a dictionary to map level names to syslog priority names, rather than a string.lower(). ........ r50741 | neal.norwitz | 2006-07-21 07:29:58 +0200 (Fri, 21 Jul 2006) | 1 line Add some asserts that we got good params passed ........ r50742 | neal.norwitz | 2006-07-21 07:31:02 +0200 (Fri, 21 Jul 2006) | 5 lines Move the initialization of some pointers earlier. The problem is that if we call Py_DECREF(frame) like we do if allocating locals fails, frame_dealloc() will try to use these bogus values and crash. ........ r50743 | neal.norwitz | 2006-07-21 07:32:28 +0200 (Fri, 21 Jul 2006) | 4 lines Handle allocation failures gracefully. Found with failmalloc. Many (all?) of these could be backported. ........ r50745 | neal.norwitz | 2006-07-21 09:59:02 +0200 (Fri, 21 Jul 2006) | 1 line Speel initialise write. Tanks Anthony. ........ r50746 | neal.norwitz | 2006-07-21 09:59:47 +0200 (Fri, 21 Jul 2006) | 2 lines Handle more memory allocation failures without crashing. ........ r50754 | barry.warsaw | 2006-07-21 16:51:07 +0200 (Fri, 21 Jul 2006) | 23 lines More RFC 2231 improvements for the email 4.0 package. As Mark Sapiro rightly points out there are really two types of continued headers defined in this RFC (i.e. "encoded" parameters with the form "name*0*=" and unencoded parameters with the form "name*0="), but we were were handling them both the same way and that isn't correct. This patch should be much more RFC compliant in that only encoded params are %-decoded and the charset/language information is only extract if there are any encoded params in the segments. If there are no encoded params then the RFC says that there will be no charset/language parts. Note however that this will change the return value for Message.get_param() in some cases. For example, whereas before if you had all unencoded param continuations you would have still gotten a 3-tuple back from this method (with charset and language == None), you will now get just a string. I don't believe this is a backward incompatible change though because the documentation for this method already indicates that either return value is possible and that you must do an isinstance(val, tuple) check to discriminate between the two. (Yeah that API kind of sucks but we can't change /that/ without breaking code.) Test cases, some documentation updates, and a NEWS item accompany this patch. ........ r50759 | georg.brandl | 2006-07-21 19:36:31 +0200 (Fri, 21 Jul 2006) | 3 lines Fix check for empty list (vs. None). ........ r50771 | brett.cannon | 2006-07-22 00:44:07 +0200 (Sat, 22 Jul 2006) | 2 lines Remove an XXX marker in a comment. ........ r50773 | neal.norwitz | 2006-07-22 18:20:49 +0200 (Sat, 22 Jul 2006) | 1 line Fix more memory allocation issues found with failmalloc. ........ r50774 | neal.norwitz | 2006-07-22 19:00:57 +0200 (Sat, 22 Jul 2006) | 1 line Don't fail if the directory already exists ........ r50775 | greg.ward | 2006-07-23 04:25:53 +0200 (Sun, 23 Jul 2006) | 6 lines Be a lot smarter about whether this test passes: instead of assuming that a 2.93 sec audio file will always take 3.1 sec (as it did on the hardware I had when I first wrote the test), expect that it will take 2.93 sec +/- 10%, and only fail if it's outside of that range. Compute the expected ........ r50776 | kurt.kaiser | 2006-07-23 06:19:49 +0200 (Sun, 23 Jul 2006) | 2 lines Tooltips failed on new-syle class __init__ args. Bug 1027566 Loren Guthrie ........ r50777 | neal.norwitz | 2006-07-23 09:50:36 +0200 (Sun, 23 Jul 2006) | 1 line Handle more mem alloc issues found with failmalloc ........ r50778 | neal.norwitz | 2006-07-23 09:51:58 +0200 (Sun, 23 Jul 2006) | 5 lines If the for loop isn't entered, entryblock will be NULL. If passed to stackdepth_walk it will be dereffed. Not sure if I found with failmalloc or Klockwork #55. ........ r50779 | neal.norwitz | 2006-07-23 09:53:14 +0200 (Sun, 23 Jul 2006) | 4 lines Move the initialization of size_a down below the check for a being NULL. Reported by Klocwork #106 ........ r50780 | neal.norwitz | 2006-07-23 09:55:55 +0200 (Sun, 23 Jul 2006) | 9 lines Check the allocation of b_objects and return if there was a failure. Also fix a few memory leaks in other failure scenarios. It seems that if b_objects == Py_None, we will have an extra ref to b_objects. Add XXX comment so hopefully someone documents why the else isn't necessary or adds it in. Reported by Klocwork #20 ........ r50781 | neal.norwitz | 2006-07-23 09:57:11 +0200 (Sun, 23 Jul 2006) | 2 lines Fix memory leaks spotted by Klocwork #37. ........ r50782 | neal.norwitz | 2006-07-23 09:59:00 +0200 (Sun, 23 Jul 2006) | 5 lines nextlink can be NULL if teedataobject_new fails, so use XINCREF. Ensure that dataobj is never NULL. Reported by Klocwork #102 ........ r50783 | neal.norwitz | 2006-07-23 10:01:43 +0200 (Sun, 23 Jul 2006) | 8 lines Ensure we don't write beyond errText. I think I got this right, but it definitely could use some review to ensure I'm not off by one and there's no possible overflow/wrap-around of bytes_left. Reported by Klocwork #1. Fix a problem if there is a failure allocating self->db. Found with failmalloc. ........ r50784 | ronald.oussoren | 2006-07-23 11:41:09 +0200 (Sun, 23 Jul 2006) | 3 lines Without this patch CMD-W won't close EditorWindows on MacOS X. This solves part of bug #1517990. ........ r50785 | ronald.oussoren | 2006-07-23 11:46:11 +0200 (Sun, 23 Jul 2006) | 5 lines Fix for bug #1517996: Class and Path browsers show Tk menu This patch replaces the menubar that is used by AquaTk for windows without a menubar of their own by one that is more appropriate for IDLE. ........ r50786 | andrew.macintyre | 2006-07-23 14:57:02 +0200 (Sun, 23 Jul 2006) | 2 lines Build updates for OS/2 EMX port ........ r50787 | andrew.macintyre | 2006-07-23 15:00:04 +0200 (Sun, 23 Jul 2006) | 3 lines bugfix: PyThread_start_new_thread() returns the thread ID, not a flag; will backport. ........ r50789 | andrew.macintyre | 2006-07-23 15:04:00 +0200 (Sun, 23 Jul 2006) | 2 lines Get mailbox module working on OS/2 EMX port. ........ r50791 | greg.ward | 2006-07-23 18:05:51 +0200 (Sun, 23 Jul 2006) | 1 line Resync optparse with Optik 1.5.3: minor tweaks for/to tests. ........ r50794 | martin.v.loewis | 2006-07-24 07:05:22 +0200 (Mon, 24 Jul 2006) | 2 lines Update list of unsupported systems. Fixes #1510853. ........ r50795 | martin.v.loewis | 2006-07-24 12:26:33 +0200 (Mon, 24 Jul 2006) | 1 line Patch #1448199: Release GIL around ConnectRegistry. ........ r50796 | martin.v.loewis | 2006-07-24 13:54:53 +0200 (Mon, 24 Jul 2006) | 3 lines Patch #1232023: Don't include empty path component from registry, so that the current directory does not get added to sys.path. Also fixes #1526785. ........ r50797 | martin.v.loewis | 2006-07-24 14:54:17 +0200 (Mon, 24 Jul 2006) | 3 lines Bug #1524310: Properly report errors from FindNextFile in os.listdir. Will backport to 2.4. ........ r50800 | georg.brandl | 2006-07-24 15:28:57 +0200 (Mon, 24 Jul 2006) | 7 lines Patch #1523356: fix determining include dirs in python-config. Also don't install "python-config" when doing altinstall, but always install "python-config2.x" and make a link to it like with the main executable. ........ r50802 | georg.brandl | 2006-07-24 15:46:47 +0200 (Mon, 24 Jul 2006) | 3 lines Patch #1527744: right order of includes in order to have HAVE_CONIO_H defined properly. ........ r50803 | georg.brandl | 2006-07-24 16:09:56 +0200 (Mon, 24 Jul 2006) | 3 lines Patch #1515343: Fix printing of deprecated string exceptions with a value in the traceback module. ........ r50804 | kurt.kaiser | 2006-07-24 19:13:23 +0200 (Mon, 24 Jul 2006) | 7 lines EditorWindow failed when used stand-alone if sys.ps1 not set. Bug 1010370 Dave Florek M EditorWindow.py M PyShell.py M NEWS.txt ........ r50805 | kurt.kaiser | 2006-07-24 20:05:51 +0200 (Mon, 24 Jul 2006) | 6 lines - EditorWindow.test() was failing. Bug 1417598 M EditorWindow.py M ScriptBinding.py M NEWS.txt ........ r50808 | georg.brandl | 2006-07-24 22:11:35 +0200 (Mon, 24 Jul 2006) | 3 lines Repair accidental NameError. ........ r50809 | tim.peters | 2006-07-24 23:02:15 +0200 (Mon, 24 Jul 2006) | 2 lines Whitespace normalization. ........ r50810 | greg.ward | 2006-07-25 04:11:12 +0200 (Tue, 25 Jul 2006) | 3 lines Don't use standard assert: want tests to fail even when run with -O. Delete cruft. ........ r50811 | tim.peters | 2006-07-25 06:07:22 +0200 (Tue, 25 Jul 2006) | 10 lines current_frames_with_threads(): There's actually no way to guess /which/ line the spawned thread is in at the time sys._current_frames() is called: we know it finished enter_g.set(), but can't know whether the instruction counter has advanced to the following leave_g.wait(). The latter is overwhelming most likely, but not guaranteed, and I see that the "x86 Ubuntu dapper (icc) trunk" buildbot found it on the other line once. Changed the test so it passes in either case. ........ r50815 | martin.v.loewis | 2006-07-25 11:53:12 +0200 (Tue, 25 Jul 2006) | 2 lines Bug #1525817: Don't truncate short lines in IDLE's tool tips. ........ r50816 | martin.v.loewis | 2006-07-25 12:05:47 +0200 (Tue, 25 Jul 2006) | 3 lines Bug #978833: Really close underlying socket in _socketobject.close. Will backport to 2.4. ........ r50817 | martin.v.loewis | 2006-07-25 12:11:14 +0200 (Tue, 25 Jul 2006) | 1 line Revert incomplete checkin. ........ r50819 | georg.brandl | 2006-07-25 12:22:34 +0200 (Tue, 25 Jul 2006) | 4 lines Patch #1525766: correctly pass onerror arg to recursive calls of pkg.walk_packages. Also improve the docstrings. ........ r50825 | brett.cannon | 2006-07-25 19:32:20 +0200 (Tue, 25 Jul 2006) | 2 lines Add comment for changes to test_ossaudiodev. ........ r50826 | brett.cannon | 2006-07-25 19:34:36 +0200 (Tue, 25 Jul 2006) | 3 lines Fix a bug in the messages for an assert failure where not enough arguments to a string were being converted in the format. ........ r50828 | armin.rigo | 2006-07-25 20:09:57 +0200 (Tue, 25 Jul 2006) | 2 lines Document why is and is not a good way to fix the gc_inspection crasher. ........ r50829 | armin.rigo | 2006-07-25 20:11:07 +0200 (Tue, 25 Jul 2006) | 5 lines Added another crasher, which hit me today (I was not intentionally writing such code, of course, but it took some gdb time to figure out what my bug was). ........ r50830 | armin.rigo | 2006-07-25 20:38:39 +0200 (Tue, 25 Jul 2006) | 3 lines Document the crashers that will not go away soon as "won't fix", and explain why. ........ r50831 | ronald.oussoren | 2006-07-25 21:13:35 +0200 (Tue, 25 Jul 2006) | 3 lines Install the compatibility symlink to libpython.a on OSX using 'ln -sf' instead of 'ln -s', this avoid problems when reinstalling python. ........ r50832 | ronald.oussoren | 2006-07-25 21:20:54 +0200 (Tue, 25 Jul 2006) | 7 lines Fix for bug #1525447 (renaming to MacOSmodule.c would also work, but not without causing problems for anyone that is on a case-insensitive filesystem). Setup.py tries to compile the MacOS extension from MacOSmodule.c, while the actual file is named macosmodule.c. This is no problem on the (default) case-insensitive filesystem, but doesn't work on case-sensitive filesystems. ........ r50833 | ronald.oussoren | 2006-07-25 22:28:55 +0200 (Tue, 25 Jul 2006) | 7 lines Fix bug #1517990: IDLE keybindings on OSX This adds a new key definition for OSX, which is slightly different from the classic mac definition. Also add NEWS item for a couple of bugfixes I added recently. ........ r50834 | tim.peters | 2006-07-26 00:30:24 +0200 (Wed, 26 Jul 2006) | 2 lines Whitespace normalization. ........ r50839 | neal.norwitz | 2006-07-26 06:00:18 +0200 (Wed, 26 Jul 2006) | 1 line Hmm, only python2.x is installed, not plain python. Did that change recently? ........ r50840 | barry.warsaw | 2006-07-26 07:54:46 +0200 (Wed, 26 Jul 2006) | 6 lines Forward port some fixes that were in email 2.5 but for some reason didn't make it into email 4.0. Specifically, in Message.get_content_charset(), handle RFC 2231 headers that contain an encoding not known to Python, or a character in the data that isn't in the charset encoding. Also forward port the appropriate unit tests. ........ r50841 | georg.brandl | 2006-07-26 09:23:32 +0200 (Wed, 26 Jul 2006) | 3 lines NEWS entry for #1525766. ........ r50842 | georg.brandl | 2006-07-26 09:40:17 +0200 (Wed, 26 Jul 2006) | 3 lines Bug #1459963: properly capitalize HTTP header names. ........ r50843 | georg.brandl | 2006-07-26 10:03:10 +0200 (Wed, 26 Jul 2006) | 6 lines Part of bug #1523610: fix miscalculation of buffer length. Also add a guard against NULL in converttuple and add a test case (that previously would have crashed). ........ r50844 | martin.v.loewis | 2006-07-26 14:12:56 +0200 (Wed, 26 Jul 2006) | 3 lines Bug #978833: Really close underlying socket in _socketobject.close. Fix httplib.HTTPConnection.getresponse to not close the socket if it is still needed for the response. ........ r50845 | andrew.kuchling | 2006-07-26 19:16:52 +0200 (Wed, 26 Jul 2006) | 1 line [Bug #1471938] Fix build problem on Solaris 8 by conditionalizing the use of mvwgetnstr(); it was conditionalized a few lines below. Fix from Paul Eggert. I also tried out the STRICT_SYSV_CURSES case and am therefore removing the 'untested' comment. ........ r50846 | andrew.kuchling | 2006-07-26 19:18:01 +0200 (Wed, 26 Jul 2006) | 1 line Correct error message ........ r50847 | andrew.kuchling | 2006-07-26 19:19:39 +0200 (Wed, 26 Jul 2006) | 1 line Minor grammar fix ........ r50848 | andrew.kuchling | 2006-07-26 19:22:21 +0200 (Wed, 26 Jul 2006) | 1 line Put news item in right section ........ r50850 | andrew.kuchling | 2006-07-26 20:03:12 +0200 (Wed, 26 Jul 2006) | 1 line Use sys.exc_info() ........ r50851 | andrew.kuchling | 2006-07-26 20:15:45 +0200 (Wed, 26 Jul 2006) | 1 line Use sys.exc_info() ........ r50852 | phillip.eby | 2006-07-26 21:48:27 +0200 (Wed, 26 Jul 2006) | 4 lines Allow the 'onerror' argument to walk_packages() to catch any Exception, not just ImportError. This allows documentation tools to better skip unimportable packages. ........ r50854 | tim.peters | 2006-07-27 01:23:15 +0200 (Thu, 27 Jul 2006) | 2 lines Whitespace normalization. ........ r50855 | tim.peters | 2006-07-27 03:14:53 +0200 (Thu, 27 Jul 2006) | 21 lines Bug #1521947: possible bug in mystrtol.c with recent gcc. In general, C doesn't define anything about what happens when an operation on a signed integral type overflows, and PyOS_strtol() did several formally undefined things of that nature on signed longs. Some version of gcc apparently tries to exploit that now, and PyOS_strtol() could fail to detect overflow then. Tried to repair all that, although it seems at least as likely to me that we'll get screwed by bad platform definitions for LONG_MIN and/or LONG_MAX now. For that reason, I don't recommend backporting this. Note that I have no box on which this makes a lick of difference -- can't really test it, except to note that it didn't break anything on my boxes. Silent change: PyOS_strtol() used to return the hard-coded 0x7fffffff in case of overflow. Now it returns LONG_MAX. They're the same only on 32-bit boxes (although C doesn't guarantee that either ...). ........ r50856 | neal.norwitz | 2006-07-27 05:51:58 +0200 (Thu, 27 Jul 2006) | 6 lines Don't kill a normal instance of python running on windows when checking to kill a cygwin instance. build\\python.exe was matching a normal windows instance. Prefix that with a \\ to ensure build is a directory and not PCbuild. As discussed on python-dev. ........ r50857 | neal.norwitz | 2006-07-27 05:55:39 +0200 (Thu, 27 Jul 2006) | 5 lines Closure can't be NULL at this point since we know it's a tuple. Reported by Klocwork # 74. ........ r50858 | neal.norwitz | 2006-07-27 06:04:50 +0200 (Thu, 27 Jul 2006) | 1 line No functional change. Add comment and assert to describe why there cannot be overflow which was reported by Klocwork. Discussed on python-dev ........ r50859 | martin.v.loewis | 2006-07-27 08:38:16 +0200 (Thu, 27 Jul 2006) | 3 lines Bump distutils version to 2.5, as several new features have been introduced since 2.4. ........ r50860 | andrew.kuchling | 2006-07-27 14:18:20 +0200 (Thu, 27 Jul 2006) | 1 line Reformat docstring; fix typo ........ r50861 | georg.brandl | 2006-07-27 17:05:36 +0200 (Thu, 27 Jul 2006) | 6 lines Add test_main() methods. These three tests were never run by regrtest.py. We really need a simpler testing framework. ........ r50862 | tim.peters | 2006-07-27 17:09:20 +0200 (Thu, 27 Jul 2006) | 2 lines News for patch #1529686. ........ r50863 | tim.peters | 2006-07-27 17:11:00 +0200 (Thu, 27 Jul 2006) | 2 lines Whitespace normalization. ........ r50864 | georg.brandl | 2006-07-27 17:38:33 +0200 (Thu, 27 Jul 2006) | 3 lines Amend news entry. ........ r50865 | georg.brandl | 2006-07-27 18:08:15 +0200 (Thu, 27 Jul 2006) | 3 lines Make uuid test suite pass on this box by requesting output with LC_ALL=C. ........ r50866 | andrew.kuchling | 2006-07-27 20:37:33 +0200 (Thu, 27 Jul 2006) | 1 line Add example ........ r50867 | thomas.heller | 2006-07-27 20:39:55 +0200 (Thu, 27 Jul 2006) | 9 lines Remove code that is no longer used (ctypes.com). Fix the DllGetClassObject and DllCanUnloadNow so that they forward the call to the comtypes.server.inprocserver module. The latter was never documented, never used by published code, and didn't work anyway, so I think it does not deserve a NEWS entry (but I might be wrong). ........ r50868 | andrew.kuchling | 2006-07-27 20:41:21 +0200 (Thu, 27 Jul 2006) | 1 line Typo fix ('publically' is rare, poss. non-standard) ........ r50869 | andrew.kuchling | 2006-07-27 20:42:41 +0200 (Thu, 27 Jul 2006) | 1 line Add missing word ........ r50870 | andrew.kuchling | 2006-07-27 20:44:10 +0200 (Thu, 27 Jul 2006) | 1 line Repair typos ........ r50872 | andrew.kuchling | 2006-07-27 20:53:33 +0200 (Thu, 27 Jul 2006) | 1 line Update URL; add example ........ r50873 | andrew.kuchling | 2006-07-27 21:07:29 +0200 (Thu, 27 Jul 2006) | 1 line Add punctuation mark; add some examples ........ r50874 | andrew.kuchling | 2006-07-27 21:11:07 +0200 (Thu, 27 Jul 2006) | 1 line Mention base64 module; rewrite last sentence to be more positive ........ r50875 | andrew.kuchling | 2006-07-27 21:12:49 +0200 (Thu, 27 Jul 2006) | 1 line If binhex is higher-level than binascii, it should come first in the chapter ........ r50876 | tim.peters | 2006-07-27 22:47:24 +0200 (Thu, 27 Jul 2006) | 28 lines check_node(): stop spraying mystery output to stderr. When a node number disagrees, keep track of all sources & the node numbers they reported, and stick all that in the error message. Changed all callers to supply a non-empty "source" argument; made the "source" argument non-optional. On my box, test_uuid still fails, but with the less confusing output: AssertionError: different sources disagree on node: from source 'getnode1', node was 00038a000015 from source 'getnode2', node was 00038a000015 from source 'ipconfig', node was 001111b2b7bf Only the last one appears to be correct; e.g., C:\Code\python\PCbuild>getmac Physical Address Transport Name =================== ========================================================== 00-11-11-B2-B7-BF \Device\Tcpip_{190FB163-5AFD-4483-86A1-2FE16AC61FF1} 62-A1-AC-6C-FD-BE \Device\Tcpip_{8F77DF5A-EA3D-4F1D-975E-D472CEE6438A} E2-1F-01-C6-5D-88 \Device\Tcpip_{CD18F76B-2EF3-409F-9B8A-6481EE70A1E4} I can't find anything on my box with MAC 00-03-8a-00-00-15, and am not clear on where that comes from. ........ r50878 | andrew.kuchling | 2006-07-28 00:40:05 +0200 (Fri, 28 Jul 2006) | 1 line Reword paragraph ........ r50879 | andrew.kuchling | 2006-07-28 00:49:38 +0200 (Fri, 28 Jul 2006) | 1 line Add example ........ r50880 | andrew.kuchling | 2006-07-28 00:49:54 +0200 (Fri, 28 Jul 2006) | 1 line Add example ........ r50881 | barry.warsaw | 2006-07-28 01:43:15 +0200 (Fri, 28 Jul 2006) | 27 lines Patch #1520294: Support for getset and member descriptors in types.py, inspect.py, and pydoc.py. Specifically, this allows for querying the type of an object against these built-in C types and more importantly, for getting their docstrings printed in the interactive interpreter's help() function. This patch includes a new built-in module called _types which provides definitions of getset and member descriptors for use by the types.py module. These types are exposed as types.GetSetDescriptorType and types.MemberDescriptorType. Query functions are provided as inspect.isgetsetdescriptor() and inspect.ismemberdescriptor(). The implementations of these are robust enough to work with Python implementations other than CPython, which may not have these fundamental types. The patch also includes documentation and test suite updates. I commit these changes now under these guiding principles: 1. Silence is assent. The release manager has not said "no", and of the few people that cared enough to respond to the thread, the worst vote was "0". 2. It's easier to ask for forgiveness than permission. 3. It's so dang easy to revert stuff in svn, that you could view this as a forcing function. :) Windows build patches will follow. ........ r50882 | tim.peters | 2006-07-28 01:44:37 +0200 (Fri, 28 Jul 2006) | 4 lines Bug #1529297: The rewrite of doctest for Python 2.4 unintentionally lost that tests are sorted by name before being run. ``DocTestFinder`` has been changed to sort the list of tests it returns. ........ r50883 | tim.peters | 2006-07-28 01:45:48 +0200 (Fri, 28 Jul 2006) | 2 lines Whitespace normalization. ........ r50884 | tim.peters | 2006-07-28 01:46:36 +0200 (Fri, 28 Jul 2006) | 2 lines Add missing svn:eol-style property to text files. ........ r50885 | barry.warsaw | 2006-07-28 01:50:40 +0200 (Fri, 28 Jul 2006) | 4 lines Enable the building of the _types module on Windows. Note that this has only been tested for VS 2003 since that's all I have. ........ r50887 | tim.peters | 2006-07-28 02:23:15 +0200 (Fri, 28 Jul 2006) | 7 lines defdict_reduce(): Plug leaks. We didn't notice these before because test_defaultdict didn't actually do anything before Georg fixed that earlier today. Neal's next refleak run then showed test_defaultdict leaking 9 references on each run. That's repaired by this checkin. ........ r50888 | tim.peters | 2006-07-28 02:30:00 +0200 (Fri, 28 Jul 2006) | 2 lines News about the repaired memory leak in defaultdict. ........ r50889 | gregory.p.smith | 2006-07-28 03:35:25 +0200 (Fri, 28 Jul 2006) | 7 lines - pybsddb Bug #1527939: bsddb module DBEnv dbremove and dbrename methods now allow their database parameter to be None as the sleepycat API allows. Also adds an appropriate test case for DBEnv.dbrename and dbremove. ........ r50895 | neal.norwitz | 2006-07-28 06:22:34 +0200 (Fri, 28 Jul 2006) | 1 line Ensure the actual number matches the expected count ........ r50896 | tim.peters | 2006-07-28 06:51:59 +0200 (Fri, 28 Jul 2006) | 6 lines Live with that "the hardware address" is an ill-defined concept, and that different ways of trying to find "the hardware address" may return different results. Certainly true on both of my Windows boxes, and in different ways (see whining on python-dev). ........ r50897 | neal.norwitz | 2006-07-28 09:21:27 +0200 (Fri, 28 Jul 2006) | 3 lines Try to find the MAC addr on various flavours of Unix. This seems hopeless. The reduces the test_uuid failures, but there's still another method failing. ........ r50898 | martin.v.loewis | 2006-07-28 09:45:49 +0200 (Fri, 28 Jul 2006) | 2 lines Add UUID for upcoming 2.5b3. ........ r50899 | matt.fleming | 2006-07-28 13:27:27 +0200 (Fri, 28 Jul 2006) | 3 lines Allow socketmodule to compile on NetBSD -current, whose bluetooth API differs from both Linux and FreeBSD. Accepted by Neal Norwitz. ........ r50900 | andrew.kuchling | 2006-07-28 14:07:12 +0200 (Fri, 28 Jul 2006) | 1 line [Patch #1529811] Correction to description of r|* mode ........ r50901 | andrew.kuchling | 2006-07-28 14:18:22 +0200 (Fri, 28 Jul 2006) | 1 line Typo fix ........ r50902 | andrew.kuchling | 2006-07-28 14:32:43 +0200 (Fri, 28 Jul 2006) | 1 line Add example ........ r50903 | andrew.kuchling | 2006-07-28 14:33:19 +0200 (Fri, 28 Jul 2006) | 1 line Add example ........ r50904 | andrew.kuchling | 2006-07-28 14:45:55 +0200 (Fri, 28 Jul 2006) | 1 line Don't overwrite built-in name; add some blank lines for readability ........ r50905 | andrew.kuchling | 2006-07-28 14:48:07 +0200 (Fri, 28 Jul 2006) | 1 line Add example. Should I propagate this example to all the other DBM-ish modules, too? ........ r50912 | georg.brandl | 2006-07-28 20:31:39 +0200 (Fri, 28 Jul 2006) | 3 lines Patch #1529686: also run test_email_codecs with regrtest.py. ........ r50913 | georg.brandl | 2006-07-28 20:36:01 +0200 (Fri, 28 Jul 2006) | 3 lines Fix spelling. ........ r50915 | thomas.heller | 2006-07-28 21:42:40 +0200 (Fri, 28 Jul 2006) | 3 lines Remove a useless XXX comment. Cosmetic changes to the code so that the #ifdef _UNICODE block doesn't mess emacs code formatting. ........ r50916 | phillip.eby | 2006-07-28 23:12:07 +0200 (Fri, 28 Jul 2006) | 5 lines Bug #1529871: The speed enhancement patch #921466 broke Python's compliance with PEP 302. This was fixed by adding an ``imp.NullImporter`` type that is used in ``sys.path_importer_cache`` to cache non-directory paths and avoid excessive filesystem operations during imports. ........ r50917 | phillip.eby | 2006-07-28 23:31:54 +0200 (Fri, 28 Jul 2006) | 2 lines Fix svn merge spew. ........ r50918 | thomas.heller | 2006-07-28 23:43:20 +0200 (Fri, 28 Jul 2006) | 4 lines Patch #1529514: More openbsd platforms for ctypes. Regenerated Modules/_ctypes/libffi/configure with autoconf 2.59. Approved by Neal. ........ r50922 | georg.brandl | 2006-07-29 10:51:21 +0200 (Sat, 29 Jul 2006) | 2 lines Bug #835255: The "closure" argument to new.function() is now documented. ........ r50924 | georg.brandl | 2006-07-29 11:33:26 +0200 (Sat, 29 Jul 2006) | 3 lines Bug #1441397: The compiler module now recognizes module and function docstrings correctly as it did in Python 2.4. ........ r50925 | georg.brandl | 2006-07-29 12:25:46 +0200 (Sat, 29 Jul 2006) | 4 lines Revert rev 42617, it was introduced to work around bug #1441397. test_compiler now passes again. ........ r50926 | fred.drake | 2006-07-29 15:22:49 +0200 (Sat, 29 Jul 2006) | 1 line update target version number ........ r50927 | andrew.kuchling | 2006-07-29 15:56:48 +0200 (Sat, 29 Jul 2006) | 1 line Add example ........ r50928 | andrew.kuchling | 2006-07-29 16:04:47 +0200 (Sat, 29 Jul 2006) | 1 line Update URL ........ r50930 | andrew.kuchling | 2006-07-29 16:08:15 +0200 (Sat, 29 Jul 2006) | 1 line Reword paragraph to match the order of the subsequent sections ........ r50931 | andrew.kuchling | 2006-07-29 16:21:15 +0200 (Sat, 29 Jul 2006) | 1 line [Bug #1529157] Mention raw_input() and input(); while I'm at it, reword the description a bit ........ r50932 | andrew.kuchling | 2006-07-29 16:42:48 +0200 (Sat, 29 Jul 2006) | 1 line [Bug #1519571] Document some missing functions: setup(), title(), done() ........ r50933 | andrew.kuchling | 2006-07-29 16:43:55 +0200 (Sat, 29 Jul 2006) | 1 line Fix docstring punctuation ........ r50934 | andrew.kuchling | 2006-07-29 17:10:32 +0200 (Sat, 29 Jul 2006) | 1 line [Bug #1414697] Change docstring of set/frozenset types to specify that the contents are unique. Raymond, please feel free to edit or revert. ........ r50935 | andrew.kuchling | 2006-07-29 17:35:21 +0200 (Sat, 29 Jul 2006) | 1 line [Bug #1530382] Document SSL.server(), .issuer() methods ........ r50936 | andrew.kuchling | 2006-07-29 17:42:46 +0200 (Sat, 29 Jul 2006) | 1 line Typo fix ........ r50937 | andrew.kuchling | 2006-07-29 17:43:13 +0200 (Sat, 29 Jul 2006) | 1 line Tweak wording ........ r50938 | matt.fleming | 2006-07-29 17:55:30 +0200 (Sat, 29 Jul 2006) | 2 lines Fix typo ........ r50939 | andrew.kuchling | 2006-07-29 17:57:08 +0200 (Sat, 29 Jul 2006) | 6 lines [Bug #1528258] Mention that the 'data' argument can be None. The constructor docs referred the reader to the add_data() method's docs, but they weren't very helpful. I've simply copied an earlier explanation of 'data' that's more useful. ........ r50940 | andrew.kuchling | 2006-07-29 18:08:40 +0200 (Sat, 29 Jul 2006) | 1 line Set bug/patch count. Take a bow, everyone! ........ r50941 | fred.drake | 2006-07-29 18:56:15 +0200 (Sat, 29 Jul 2006) | 18 lines expunge the xmlcore changes: 41667, 41668 - initial switch to xmlcore 47044 - mention of xmlcore in What's New 50687 - mention of xmlcore in the library reference re-apply xmlcore changes to xml: 41674 - line ending changes (re-applied manually), directory props 41677 - add cElementTree wrapper 41678 - PSF licensing for etree 41812 - whitespace normalization 42724 - fix svn:eol-style settings 43681, 43682 - remove Python version-compatibility cruft from minidom 46773 - fix encoding of \r\n\t in attr values in saxutils 47269 - added XMLParser alias for cElementTree compatibility additional tests were added in Lib/test/test_sax.py that failed with the xmlcore changes; these relate to SF bugs #1511497, #1513611 ........ r50942 | andrew.kuchling | 2006-07-29 20:14:07 +0200 (Sat, 29 Jul 2006) | 17 lines Reorganize the docs for 'file' and 'open()' after some discussion with Fred. We want to encourage users to write open() when opening a file, but open() was described with a single paragraph and 'file' had lots of explanation of the mode and bufsize arguments. I've shrunk the description of 'file' to cross-reference to the 'File objects' section, and to open() for an explanation of the arguments. open() now has all the paragraphs about the mode string. The bufsize argument was moved up so that it isn't buried at the end; now there's 1 paragraph on mode, 1 on bufsize, and then 3 more on mode. Various other edits and rearrangements were made in the process. It's probably best to read the final text and not to try to make sense of the diffs. ........ r50943 | fred.drake | 2006-07-29 20:19:19 +0200 (Sat, 29 Jul 2006) | 1 line restore test un-intentionally removed in the xmlcore purge (revision 50941) ........ r50944 | fred.drake | 2006-07-29 20:33:29 +0200 (Sat, 29 Jul 2006) | 3 lines make the reference to older versions of the documentation a link to the right page on python.org ........ r50945 | fred.drake | 2006-07-29 21:09:01 +0200 (Sat, 29 Jul 2006) | 1 line document the footnote usage pattern ........ r50947 | fred.drake | 2006-07-29 21:14:10 +0200 (Sat, 29 Jul 2006) | 1 line emphasize and oddball nuance of LaTeX comment syntax ........ r50948 | andrew.kuchling | 2006-07-29 21:24:04 +0200 (Sat, 29 Jul 2006) | 1 line [Patch #1490989 from Skip Montanaro] Mention debugging builds in the API documentation. I've changed Skip's patch to point to Misc/SpecialBuilds and fiddled with the markup a bit. ........ r50949 | neal.norwitz | 2006-07-29 21:29:35 +0200 (Sat, 29 Jul 2006) | 6 lines Disable these tests until they are reliable across platforms. These problems may mask more important, real problems. One or both methods are known to fail on: Solaris, OpenBSD, Debian, Ubuntu. They pass on Windows and some Linux boxes. ........ r50950 | andrew.kuchling | 2006-07-29 21:50:37 +0200 (Sat, 29 Jul 2006) | 1 line [Patch #1068277] Clarify that os.path.exists() can return False depending on permissions. Fred approved committing this patch in December 2004! ........ r50952 | fred.drake | 2006-07-29 22:04:42 +0200 (Sat, 29 Jul 2006) | 6 lines SF bug #1193966: Weakref types documentation misplaced The information about supporting weakrefs with types defined in C extensions is moved to the Extending & Embedding manual. Py_TPFLAGS_HAVE_WEAKREFS is no longer mentioned since it is part of Py_TPFLAGS_DEFAULT. ........ r50953 | skip.montanaro | 2006-07-29 22:06:05 +0200 (Sat, 29 Jul 2006) | 4 lines Add a comment to the csv reader documentation that explains why the treatment of newlines changed in 2.5. Pulled almost verbatim from a comment by Andrew McNamara in <http://python.org/sf/1465014>. ........ r50954 | neal.norwitz | 2006-07-29 22:20:52 +0200 (Sat, 29 Jul 2006) | 3 lines If the executable doesn't exist, there's no reason to try to start it. This prevents garbage about command not found being printed on Solaris. ........ r50955 | fred.drake | 2006-07-29 22:21:25 +0200 (Sat, 29 Jul 2006) | 1 line fix minor markup error that introduced extra punctuation ........ r50957 | neal.norwitz | 2006-07-29 22:37:08 +0200 (Sat, 29 Jul 2006) | 3 lines Disable test_getnode too, since this is also unreliable. ........ r50958 | andrew.kuchling | 2006-07-29 23:27:12 +0200 (Sat, 29 Jul 2006) | 1 line Follow TeX's conventions for hyphens ........ r50959 | andrew.kuchling | 2006-07-29 23:30:21 +0200 (Sat, 29 Jul 2006) | 1 line Fix case for 'Unix' ........ r50960 | fred.drake | 2006-07-30 01:34:57 +0200 (Sun, 30 Jul 2006) | 1 line markup cleanups ........ r50961 | andrew.kuchling | 2006-07-30 02:27:34 +0200 (Sun, 30 Jul 2006) | 1 line Minor typo fixes ........ r50962 | andrew.kuchling | 2006-07-30 02:37:56 +0200 (Sun, 30 Jul 2006) | 1 line [Bug #793553] Correct description of keyword arguments for SSL authentication ........ r50963 | tim.peters | 2006-07-30 02:58:15 +0200 (Sun, 30 Jul 2006) | 2 lines Whitespace normalization. ........ r50964 | fred.drake | 2006-07-30 05:03:43 +0200 (Sun, 30 Jul 2006) | 1 line lots of markup nits, most commonly Unix/unix --> \UNIX ........ r50965 | fred.drake | 2006-07-30 07:41:28 +0200 (Sun, 30 Jul 2006) | 1 line update information on wxPython, from Robin Dunn ........ r50966 | fred.drake | 2006-07-30 07:49:49 +0200 (Sun, 30 Jul 2006) | 4 lines remove possibly-outdated comment on what GUI toolkit is most commonly used; it is hard to know whether this is right, and it does not add valuable reference information at any rate ........ r50967 | fred.drake | 2006-07-30 07:55:39 +0200 (Sun, 30 Jul 2006) | 3 lines - remove yet another reference to how commonly Tkinter is (thought to be) used - fix an internal section reference ........ r50968 | neal.norwitz | 2006-07-30 08:53:31 +0200 (Sun, 30 Jul 2006) | 4 lines Patch #1531113: Fix augmented assignment with yield expressions. Also fix a SystemError when trying to assign to yield expressions. ........ r50969 | neal.norwitz | 2006-07-30 08:55:48 +0200 (Sun, 30 Jul 2006) | 5 lines Add PyErr_WarnEx() so C code can pass the stacklevel to warnings.warn(). This provides the proper warning for struct.pack(). PyErr_Warn() is now deprecated in favor of PyErr_WarnEx(). As mentioned by Tim Peters on python-dev. ........ r50970 | neal.norwitz | 2006-07-30 08:57:04 +0200 (Sun, 30 Jul 2006) | 3 lines Bug #1515471: string.replace() accepts character buffers again. Pass the char* and size around rather than PyObject's. ........ r50971 | neal.norwitz | 2006-07-30 08:59:13 +0200 (Sun, 30 Jul 2006) | 1 line Whitespace normalization ........ r50973 | georg.brandl | 2006-07-30 12:53:32 +0200 (Sun, 30 Jul 2006) | 3 lines Clarify that __op__ methods must return NotImplemented if they don't support the operation. ........ r50974 | georg.brandl | 2006-07-30 13:07:23 +0200 (Sun, 30 Jul 2006) | 3 lines Bug #1002398: The documentation for os.path.sameopenfile now correctly refers to file descriptors, not file objects. ........ r50977 | martin.v.loewis | 2006-07-30 15:00:31 +0200 (Sun, 30 Jul 2006) | 3 lines Don't copy directory stat times in shutil.copytree on Windows Fixes #1525866. ........ r50978 | martin.v.loewis | 2006-07-30 15:14:05 +0200 (Sun, 30 Jul 2006) | 3 lines Base __version__ on sys.version_info, as distutils is no longer maintained separatedly. ........ r50979 | martin.v.loewis | 2006-07-30 15:27:31 +0200 (Sun, 30 Jul 2006) | 3 lines Mention Cygwin in distutils error message about a missing VS 2003. Fixes #1257728. ........ r50982 | martin.v.loewis | 2006-07-30 16:09:47 +0200 (Sun, 30 Jul 2006) | 5 lines Drop usage of test -e in configure as it is not portable. Fixes #1439538 Will backport to 2.4 Also regenerate pyconfig.h.in. ........ r50984 | georg.brandl | 2006-07-30 18:20:10 +0200 (Sun, 30 Jul 2006) | 3 lines Fix makefile changes for python-config. ........ r50985 | george.yoshida | 2006-07-30 18:37:37 +0200 (Sun, 30 Jul 2006) | 2 lines Rename struct.pack_to to struct.pack_into as changed in revision 46642. ........ r50986 | george.yoshida | 2006-07-30 18:41:30 +0200 (Sun, 30 Jul 2006) | 2 lines Typo fix ........ r50987 | neal.norwitz | 2006-07-30 21:18:13 +0200 (Sun, 30 Jul 2006) | 1 line Add some asserts and update comments ........ r50988 | neal.norwitz | 2006-07-30 21:18:38 +0200 (Sun, 30 Jul 2006) | 1 line Verify that the signal handlers were really called ........ r50989 | neal.norwitz | 2006-07-30 21:20:42 +0200 (Sun, 30 Jul 2006) | 3 lines Try to prevent hangs on Tru64/Alpha buildbot. I'm not certain this will help and may need to be reverted if it causes problems. ........ r50990 | georg.brandl | 2006-07-30 22:18:51 +0200 (Sun, 30 Jul 2006) | 2 lines Bug #1531349: right <-> left glitch in __rop__ description. ........ r50992 | tim.peters | 2006-07-31 03:46:03 +0200 (Mon, 31 Jul 2006) | 2 lines Whitespace normalization. ........ r50993 | andrew.mcnamara | 2006-07-31 04:27:48 +0200 (Mon, 31 Jul 2006) | 2 lines Redo the comment about the 2.5 change in quoted-newline handling. ........ r50994 | tim.peters | 2006-07-31 04:40:23 +0200 (Mon, 31 Jul 2006) | 10 lines ZipFile.close(): Killed one of the struct.pack deprecation warnings on Win32. Also added an XXX about the line: pos3 = self.fp.tell() `pos3` is never referenced, and I have no idea what the code intended to do instead. ........ r50996 | tim.peters | 2006-07-31 04:53:03 +0200 (Mon, 31 Jul 2006) | 8 lines ZipFile.close(): Kill the other struct.pack deprecation warning on Windows. Afraid I can't detect a pattern to when the pack formats decide to use a signed or unsigned format code -- appears nearly arbitrary to my eyes. So I left all the pack formats alone and changed the special-case data values instead. ........ r50997 | skip.montanaro | 2006-07-31 05:09:45 +0200 (Mon, 31 Jul 2006) | 1 line minor tweaks ........ r50998 | skip.montanaro | 2006-07-31 05:11:11 +0200 (Mon, 31 Jul 2006) | 1 line minor tweaks ........ r50999 | andrew.kuchling | 2006-07-31 14:20:24 +0200 (Mon, 31 Jul 2006) | 1 line Add refcounts for PyErr_WarnEx ........ r51000 | andrew.kuchling | 2006-07-31 14:39:05 +0200 (Mon, 31 Jul 2006) | 9 lines Document PyErr_WarnEx. (Bad Neal! No biscuit!) Is the explanation of the 'stacklevel' parameter clear? Please feel free to edit it. I don't have LaTeX installed on this machine, so haven't verified that the markup is correct. Will check tonight, or maybe the automatic doc build will tell me. ........ r51001 | andrew.kuchling | 2006-07-31 14:52:26 +0200 (Mon, 31 Jul 2006) | 1 line Add PyErr_WarnEx() ........ r51002 | andrew.kuchling | 2006-07-31 15:18:27 +0200 (Mon, 31 Jul 2006) | 1 line Mention csv newline changes ........ r51003 | andrew.kuchling | 2006-07-31 17:22:58 +0200 (Mon, 31 Jul 2006) | 1 line Typo fix ........ r51004 | andrew.kuchling | 2006-07-31 17:23:43 +0200 (Mon, 31 Jul 2006) | 1 line Remove reference to notation ........ r51005 | georg.brandl | 2006-07-31 18:00:34 +0200 (Mon, 31 Jul 2006) | 3 lines Fix function name. ........ r51006 | andrew.kuchling | 2006-07-31 18:10:24 +0200 (Mon, 31 Jul 2006) | 1 line [Bug #1514540] Instead of putting the standard types in a section, put them in a chapter of their own. This means string methods will now show up in the ToC. (Should the types come before or after the functions+exceptions+constants chapter? I've put them after, for now.) ........ r51007 | andrew.kuchling | 2006-07-31 18:22:05 +0200 (Mon, 31 Jul 2006) | 1 line [Bug #848556] Remove \d* from second alternative to avoid exponential case when repeating match ........ r51008 | andrew.kuchling | 2006-07-31 18:27:57 +0200 (Mon, 31 Jul 2006) | 1 line Update list of files; fix a typo ........ r51013 | andrew.kuchling | 2006-08-01 18:24:30 +0200 (Tue, 01 Aug 2006) | 1 line typo fix ........ r51018 | thomas.heller | 2006-08-01 18:54:43 +0200 (Tue, 01 Aug 2006) | 2 lines Fix a potential segfault and various potentail refcount leaks in the cast() function. ........ r51020 | thomas.heller | 2006-08-01 19:46:10 +0200 (Tue, 01 Aug 2006) | 1 line Minimal useful docstring for CopyComPointer. ........ r51021 | andrew.kuchling | 2006-08-01 20:16:15 +0200 (Tue, 01 Aug 2006) | 8 lines [Patch #1520905] Attempt to suppress core file created by test_subprocess.py. Patch by Douglas Greiman. The test_run_abort() testcase produces a core file on Unix systems, even though the test is successful. This can be confusing or alarming to someone who runs 'make test' and then finds that the Python interpreter apparently crashed. ........ r51023 | georg.brandl | 2006-08-01 20:49:24 +0200 (Tue, 01 Aug 2006) | 3 lines os.urandom no longer masks unrelated exceptions like SystemExit or KeyboardInterrupt. ........ r51025 | thomas.heller | 2006-08-01 21:14:15 +0200 (Tue, 01 Aug 2006) | 2 lines Speed up PyType_stgdict and PyObject_stgdict. ........ r51027 | ronald.oussoren | 2006-08-01 22:30:31 +0200 (Tue, 01 Aug 2006) | 3 lines Make sure the postinstall action that optionally updates the user's profile on MacOS X actually works correctly in all cases. ........ r51028 | ronald.oussoren | 2006-08-01 23:00:57 +0200 (Tue, 01 Aug 2006) | 4 lines This fixes bug #1527397: PythonLauncher runs scripts with the wrong working directory. It also fixes a bug where PythonLauncher failed to launch scripts when the scriptname (or the path to the script) contains quotes. ........ r51031 | tim.peters | 2006-08-02 05:27:46 +0200 (Wed, 02 Aug 2006) | 2 lines Whitespace normalization. ........ r51032 | tim.peters | 2006-08-02 06:12:36 +0200 (Wed, 02 Aug 2006) | 19 lines Try to squash struct.pack warnings on the "amd64 gentoo trunk" buildbot (& possibly other 64-bit boxes) during test_gzip. The native zlib crc32 function returns an unsigned 32-bit integer, which the Python wrapper implicitly casts to C long. Therefore the same crc can "look negative" on a 32-bit box but "look positive" on a 64-bit box. This patch papers over that platform difference when writing the crc to file. It may be better to change the Python wrapper, either to make the result "look positive" on all platforms (which means it may have to return a Python long at times on a 32-bit box), or to keep the sign the same across boxes. But that would be a visible change in what users see, while the current hack changes no visible behavior (well, apart from stopping the struct deprecation warning). Note that the module-level write32() function is no longer used. ........ r51033 | neal.norwitz | 2006-08-02 06:27:11 +0200 (Wed, 02 Aug 2006) | 4 lines Prevent memory leak on error. Reported by Klocwork #36 ........ r51034 | tim.peters | 2006-08-02 07:20:08 +0200 (Wed, 02 Aug 2006) | 9 lines _Stream.close(): Try to kill struct.pack() warnings when writing the crc to file on the "PPC64 Debian trunk" buildbot when running test_tarfile. This is again a case where the native zlib crc is an unsigned 32-bit int, but the Python wrapper implicitly casts it to signed C long, so that "the sign bit looks different" on different platforms. ........ r51035 | ronald.oussoren | 2006-08-02 08:10:10 +0200 (Wed, 02 Aug 2006) | 2 lines Updated documentation for the script that builds the OSX installer. ........ r51036 | neal.norwitz | 2006-08-02 08:14:22 +0200 (Wed, 02 Aug 2006) | 2 lines _PyWeakref_GetWeakrefCount() now returns a Py_ssize_t instead of long. ........ r51037 | neal.norwitz | 2006-08-02 08:15:10 +0200 (Wed, 02 Aug 2006) | 1 line v is already checked for NULL, so just DECREF it ........ r51038 | neal.norwitz | 2006-08-02 08:19:19 +0200 (Wed, 02 Aug 2006) | 1 line Let us know when there was a problem and the child had to kill the parent ........ r51039 | neal.norwitz | 2006-08-02 08:46:21 +0200 (Wed, 02 Aug 2006) | 5 lines Patch #1519025 and bug #926423: If a KeyboardInterrupt occurs during a socket operation on a socket with a timeout, the exception will be caught correctly. Previously, the exception was not caught. ........ r51040 | neal.norwitz | 2006-08-02 09:09:32 +0200 (Wed, 02 Aug 2006) | 1 line Add some explanation about Klocwork and Coverity static analysis ........ r51041 | anthony.baxter | 2006-08-02 09:43:09 +0200 (Wed, 02 Aug 2006) | 1 line pre-release machinations ........ r51043 | thomas.heller | 2006-08-02 13:35:31 +0200 (Wed, 02 Aug 2006) | 4 lines A few nore words about what ctypes does. Document that using the wrong calling convention can also raise 'ValueError: Procedure called with the wrong number of arguments'. ........ r51045 | thomas.heller | 2006-08-02 14:00:13 +0200 (Wed, 02 Aug 2006) | 1 line Fix a mistake. ........ r51046 | martin.v.loewis | 2006-08-02 15:53:55 +0200 (Wed, 02 Aug 2006) | 3 lines Correction of patch #1455898: In the mbcs decoder, set final=False for stream decoder, but final=True for the decode function. ........ r51049 | tim.peters | 2006-08-02 20:19:35 +0200 (Wed, 02 Aug 2006) | 2 lines Add missing svn:eol-style property to text files. ........ r51079 | neal.norwitz | 2006-08-04 06:50:21 +0200 (Fri, 04 Aug 2006) | 3 lines Bug #1531405, format_exception no longer raises an exception if str(exception) raised an exception. ........ r51080 | neal.norwitz | 2006-08-04 06:58:47 +0200 (Fri, 04 Aug 2006) | 11 lines Bug #1191458: tracing over for loops now produces a line event on each iteration. I'm not positive this is the best way to handle this. I'm also not sure that there aren't other cases where the lnotab is generated incorrectly. It would be great if people that use pdb or tracing could test heavily. Also: * Remove dead/duplicated code that wasn't used/necessary because we already handled the docstring prior to entering the loop. * add some debugging code into the compiler (#if 0'd out). ........ r51081 | neal.norwitz | 2006-08-04 07:09:28 +0200 (Fri, 04 Aug 2006) | 4 lines Bug #1333982: string/number constants were inappropriately stored in the byte code and co_consts even if they were not used, ie immediately popped off the stack. ........ r51082 | neal.norwitz | 2006-08-04 07:12:19 +0200 (Fri, 04 Aug 2006) | 1 line There were really two issues ........ r51084 | fred.drake | 2006-08-04 07:17:21 +0200 (Fri, 04 Aug 2006) | 1 line SF patch #1534048 (bug #1531003): fix typo in error message ........ r51085 | gregory.p.smith | 2006-08-04 07:17:47 +0200 (Fri, 04 Aug 2006) | 3 lines fix typos ........ r51087 | georg.brandl | 2006-08-04 08:03:53 +0200 (Fri, 04 Aug 2006) | 3 lines Fix bug caused by first decrefing, then increfing. ........ r51109 | neil.schemenauer | 2006-08-04 18:20:30 +0200 (Fri, 04 Aug 2006) | 5 lines Fix the 'compiler' package to generate correct code for MAKE_CLOSURE. In the 2.5 development cycle, MAKE_CLOSURE as changed to take free variables as a tuple rather than as individual items on the stack. Closes patch #1534084. ........ r51110 | georg.brandl | 2006-08-04 20:03:37 +0200 (Fri, 04 Aug 2006) | 3 lines Change fix for segfaulting property(), add a NEWS entry and a test. ........ r51111 | georg.brandl | 2006-08-04 20:07:34 +0200 (Fri, 04 Aug 2006) | 3 lines Better fix for bug #1531405, not executing str(value) twice. ........ r51112 | thomas.heller | 2006-08-04 20:17:40 +0200 (Fri, 04 Aug 2006) | 1 line On Windows, make PyErr_Warn an exported function again. ........ r51113 | thomas.heller | 2006-08-04 20:57:34 +0200 (Fri, 04 Aug 2006) | 4 lines Fix #1530448 - fix ctypes build failure on solaris 10. The '-mimpure-text' linker flag is required when linking _ctypes.so. ........ r51114 | thomas.heller | 2006-08-04 21:49:31 +0200 (Fri, 04 Aug 2006) | 3 lines Fix #1534738: win32 debug version of _msi must be _msi_d.pyd, not _msi.pyd. Fix the name of the pdb file as well. ........ r51115 | andrew.kuchling | 2006-08-04 22:37:43 +0200 (Fri, 04 Aug 2006) | 1 line Typo fixes ........ r51116 | andrew.kuchling | 2006-08-04 23:10:03 +0200 (Fri, 04 Aug 2006) | 1 line Fix mangled sentence ........ r51118 | tim.peters | 2006-08-05 00:00:35 +0200 (Sat, 05 Aug 2006) | 2 lines Whitespace normalization. ........ r51119 | bob.ippolito | 2006-08-05 01:59:21 +0200 (Sat, 05 Aug 2006) | 5 lines Fix #1530559, struct.pack raises TypeError where it used to convert. Passing float arguments to struct.pack when integers are expected now triggers a DeprecationWarning. ........ r51123 | georg.brandl | 2006-08-05 08:10:54 +0200 (Sat, 05 Aug 2006) | 3 lines Patch #1534922: correct and enhance unittest docs. ........ r51126 | georg.brandl | 2006-08-06 09:06:33 +0200 (Sun, 06 Aug 2006) | 2 lines Bug #1535182: really test the xreadlines() method of bz2 objects. ........ r51128 | georg.brandl | 2006-08-06 09:26:21 +0200 (Sun, 06 Aug 2006) | 4 lines Bug #1535081: A leading underscore has been added to the names of the md5 and sha modules, so add it in Modules/Setup.dist too. ........ r51129 | georg.brandl | 2006-08-06 10:23:54 +0200 (Sun, 06 Aug 2006) | 3 lines Bug #1535165: fixed a segfault in input() and raw_input() when sys.stdin is closed. ........ r51131 | georg.brandl | 2006-08-06 11:17:16 +0200 (Sun, 06 Aug 2006) | 2 lines Don't produce output in test_builtin. ........ r51133 | andrew.macintyre | 2006-08-06 14:37:03 +0200 (Sun, 06 Aug 2006) | 4 lines test_threading now skips testing alternate thread stack sizes on platforms that don't support changing thread stack size. ........ r51134 | andrew.kuchling | 2006-08-07 00:07:04 +0200 (Mon, 07 Aug 2006) | 2 lines [Patch #1464056] Ensure that we use the panelw library when linking with ncursesw. Once I see how the buildbots react, I'll backport this to 2.4. ........ r51137 | georg.brandl | 2006-08-08 13:52:34 +0200 (Tue, 08 Aug 2006) | 3 lines webbrowser: Silence stderr output if no gconftool or gnome browser found ........ r51138 | georg.brandl | 2006-08-08 13:56:21 +0200 (Tue, 08 Aug 2006) | 7 lines Remove "non-mapping" and "non-sequence" from TypeErrors raised by PyMapping_Size and PySequence_Size. Because len() tries first sequence, then mapping size, it will always raise a "non-mapping object has no len" error which is confusing. ........ r51139 | thomas.heller | 2006-08-08 19:37:00 +0200 (Tue, 08 Aug 2006) | 3 lines memcmp() can return values other than -1, 0, and +1 but tp_compare must not. ........ r51140 | thomas.heller | 2006-08-08 19:39:20 +0200 (Tue, 08 Aug 2006) | 1 line Remove accidently committed, duplicated test. ........ r51147 | andrew.kuchling | 2006-08-08 20:50:14 +0200 (Tue, 08 Aug 2006) | 1 line Reword paragraph to clarify ........ r51148 | andrew.kuchling | 2006-08-08 20:56:08 +0200 (Tue, 08 Aug 2006) | 1 line Move obmalloc item into C API section ........ r51149 | andrew.kuchling | 2006-08-08 21:00:14 +0200 (Tue, 08 Aug 2006) | 1 line 'Other changes' section now has only one item; move the item elsewhere and remove the section ........ r51150 | andrew.kuchling | 2006-08-08 21:00:34 +0200 (Tue, 08 Aug 2006) | 1 line Bump version number ........ r51151 | georg.brandl | 2006-08-08 22:11:22 +0200 (Tue, 08 Aug 2006) | 2 lines Bug #1536828: typo: TypeType should have been StringType. ........ r51153 | georg.brandl | 2006-08-08 22:13:13 +0200 (Tue, 08 Aug 2006) | 2 lines Bug #1536660: separate two words. ........ r51155 | georg.brandl | 2006-08-08 22:48:10 +0200 (Tue, 08 Aug 2006) | 3 lines ``str`` is now the same object as ``types.StringType``. ........ r51156 | tim.peters | 2006-08-09 02:52:26 +0200 (Wed, 09 Aug 2006) | 2 lines Whitespace normalization. ........ r51158 | georg.brandl | 2006-08-09 09:03:22 +0200 (Wed, 09 Aug 2006) | 4 lines Introduce an upper bound on tuple nesting depth in C argument format strings; fixes rest of #1523610. ........ r51160 | martin.v.loewis | 2006-08-09 09:57:39 +0200 (Wed, 09 Aug 2006) | 4 lines __hash__ may now return long int; the final hash value is obtained by invoking hash on the long int. Fixes #1536021. ........ r51168 | andrew.kuchling | 2006-08-09 15:03:41 +0200 (Wed, 09 Aug 2006) | 1 line [Bug #1536021] Mention __hash__ change ........ r51169 | andrew.kuchling | 2006-08-09 15:57:05 +0200 (Wed, 09 Aug 2006) | 1 line [Patch #1534027] Add notes on locale module changes ........ r51170 | andrew.kuchling | 2006-08-09 16:05:35 +0200 (Wed, 09 Aug 2006) | 1 line Add missing 'self' parameters ........ r51171 | andrew.kuchling | 2006-08-09 16:06:19 +0200 (Wed, 09 Aug 2006) | 1 line Reindent code ........ r51172 | armin.rigo | 2006-08-09 16:55:26 +0200 (Wed, 09 Aug 2006) | 2 lines Fix and test for an infinite C recursion. ........ r51173 | ronald.oussoren | 2006-08-09 16:56:33 +0200 (Wed, 09 Aug 2006) | 2 lines It's unlikely that future versions will require _POSIX_C_SOURCE ........ r51178 | armin.rigo | 2006-08-09 17:37:26 +0200 (Wed, 09 Aug 2006) | 2 lines Concatenation on a long string breaks (SF #1526585). ........ r51180 | kurt.kaiser | 2006-08-09 18:46:15 +0200 (Wed, 09 Aug 2006) | 8 lines 1. When used w/o subprocess, all exceptions were preceeded by an error message claiming they were IDLE internal errors (since 1.2a1). 2. Add Ronald Oussoren to CREDITS M NEWS.txt M PyShell.py M CREDITS.txt ........ r51181 | kurt.kaiser | 2006-08-09 19:47:15 +0200 (Wed, 09 Aug 2006) | 4 lines As a slight enhancement to the previous checkin, improve the internal error reporting by moving message to IDLE console. ........ r51182 | andrew.kuchling | 2006-08-09 20:23:14 +0200 (Wed, 09 Aug 2006) | 1 line Typo fix ........ r51183 | kurt.kaiser | 2006-08-09 22:34:46 +0200 (Wed, 09 Aug 2006) | 2 lines ToggleTab dialog was setting indent to 8 even if cancelled (since 1.2a1). ........ r51184 | martin.v.loewis | 2006-08-10 01:42:18 +0200 (Thu, 10 Aug 2006) | 2 lines Add some commentary on -mimpure-text. ........ r51185 | tim.peters | 2006-08-10 02:58:49 +0200 (Thu, 10 Aug 2006) | 2 lines Add missing svn:eol-style property to text files. ........ r51186 | kurt.kaiser | 2006-08-10 03:41:17 +0200 (Thu, 10 Aug 2006) | 2 lines Changing tokenize (39046) to detect dedent broke tabnanny check (since 1.2a1) ........ r51187 | tim.peters | 2006-08-10 05:01:26 +0200 (Thu, 10 Aug 2006) | 13 lines test_copytree_simple(): This was leaving behind two new temp directories each time it ran, at least on Windows. Several changes: explicitly closed all files; wrapped long lines; stopped suppressing errors when removing a file or directory fails (removing /shouldn't/ fail!); and changed what appeared to be incorrect usage of os.removedirs() (that doesn't remove empty directories at and /under/ the given path, instead it must be given an empty leaf directory and then deletes empty directories moving /up/ the path -- could be that the conceptually simpler shutil.rmtree() was really actually intended here). ........
2006-08-11 11:57:12 -03:00
\n\
Return the thread stack size used when creating new threads. The\n\
optional size argument specifies the stack size (in bytes) to be used\n\
for subsequently created threads, and must be 0 (use platform or\n\
configured default) or a positive integer value of at least 32,768 (32k).\n\
If changing the thread stack size is unsupported, a ThreadError\n\
exception is raised. If the specified size is invalid, a ValueError\n\
exception is raised, and the stack size is unmodified. 32k bytes\n\
currently the minimum supported stack size value to guarantee\n\
sufficient stack space for the interpreter itself.\n\
\n\
Note that some platforms may have particular restrictions on values for\n\
the stack size, such as requiring a minimum stack size larger than 32 KiB or\n\
Merged revisions 46753-51188 via svnmerge from svn+ssh://pythondev@svn.python.org/python/trunk ........ r46755 | brett.cannon | 2006-06-08 18:23:04 +0200 (Thu, 08 Jun 2006) | 4 lines Make binascii.hexlify() use s# for its arguments instead of t# to actually match its documentation stating it accepts any read-only buffer. ........ r46757 | brett.cannon | 2006-06-08 19:00:45 +0200 (Thu, 08 Jun 2006) | 8 lines Buffer objects would return the read or write buffer for a wrapped object when the char buffer was requested. Now it actually returns the char buffer if available or raises a TypeError if it isn't (as is raised for the other buffer types if they are not present but requested). Not a backport candidate since it does change semantics of the buffer object (although it could be argued this is enough of a bug to bother backporting). ........ r46760 | andrew.kuchling | 2006-06-09 03:10:17 +0200 (Fri, 09 Jun 2006) | 1 line Update functools section ........ r46762 | tim.peters | 2006-06-09 04:11:02 +0200 (Fri, 09 Jun 2006) | 6 lines Whitespace normalization. Since test_file is implicated in mysterious test failures when followed by test_optparse, if I had any brains I'd look at the checkin that last changed test_file ;-) ........ r46763 | tim.peters | 2006-06-09 05:09:42 +0200 (Fri, 09 Jun 2006) | 5 lines To boost morale :-), force test_optparse to run immediately after test_file until we can figure out how to fix it. (See python-dev; at the moment we don't even know which checkin caused the problem.) ........ r46764 | tim.peters | 2006-06-09 05:51:41 +0200 (Fri, 09 Jun 2006) | 6 lines AutoFileTests.tearDown(): Removed mysterious undocumented try/except. Remove TESTFN. Throughout: used open() instead of file(), and wrapped long lines. ........ r46765 | tim.peters | 2006-06-09 06:02:06 +0200 (Fri, 09 Jun 2006) | 8 lines testUnicodeOpen(): I have no idea why, but making this test clean up after itself appears to fix the test failures when test_optparse follows test_file. test_main(): Get rid of TESTFN no matter what. That's also enough to fix the mystery failures. Doesn't hurt to fix them twice :-) ........ r46766 | tim.peters | 2006-06-09 07:12:40 +0200 (Fri, 09 Jun 2006) | 6 lines Remove the temporary hack to force test_optparse to run immediately after test_file. At least 8 buildbot boxes passed since the underlying problem got fixed, and they all failed before the fix, so there's no point to this anymore. ........ r46767 | neal.norwitz | 2006-06-09 07:54:18 +0200 (Fri, 09 Jun 2006) | 1 line Fix grammar and reflow ........ r46769 | andrew.kuchling | 2006-06-09 12:22:35 +0200 (Fri, 09 Jun 2006) | 1 line Markup fix ........ r46773 | andrew.kuchling | 2006-06-09 15:15:57 +0200 (Fri, 09 Jun 2006) | 1 line [Bug #1472827] Make saxutils.XMLGenerator handle \r\n\t in attribute values by escaping them properly. 2.4 bugfix candidate. ........ r46778 | kristjan.jonsson | 2006-06-09 18:28:01 +0200 (Fri, 09 Jun 2006) | 2 lines Turn off warning about deprecated CRT functions on for VisualStudio .NET 2005. Make the definition #ARRAYSIZE conditional. VisualStudio .NET 2005 already has it defined using a better gimmick. ........ r46779 | phillip.eby | 2006-06-09 18:40:18 +0200 (Fri, 09 Jun 2006) | 2 lines Import wsgiref into the stdlib, as of the external version 0.1-r2181. ........ r46783 | andrew.kuchling | 2006-06-09 18:44:40 +0200 (Fri, 09 Jun 2006) | 1 line Add note about XMLGenerator bugfix ........ r46784 | andrew.kuchling | 2006-06-09 18:46:51 +0200 (Fri, 09 Jun 2006) | 1 line Add note about wsgiref ........ r46785 | brett.cannon | 2006-06-09 19:05:48 +0200 (Fri, 09 Jun 2006) | 2 lines Fix inconsistency in naming within an enum. ........ r46787 | tim.peters | 2006-06-09 19:47:00 +0200 (Fri, 09 Jun 2006) | 2 lines Whitespace normalization. ........ r46792 | georg.brandl | 2006-06-09 20:29:52 +0200 (Fri, 09 Jun 2006) | 3 lines Test file.__exit__. ........ r46794 | brett.cannon | 2006-06-09 20:40:46 +0200 (Fri, 09 Jun 2006) | 2 lines svn:ignore .pyc and .pyo files. ........ r46795 | georg.brandl | 2006-06-09 20:45:48 +0200 (Fri, 09 Jun 2006) | 3 lines RFE #1491485: str/unicode.endswith()/startswith() now accept a tuple as first argument. ........ r46798 | andrew.kuchling | 2006-06-09 21:03:16 +0200 (Fri, 09 Jun 2006) | 1 line Describe startswith()/endswiith() change; add reminder about wsgiref ........ r46799 | tim.peters | 2006-06-09 21:24:44 +0200 (Fri, 09 Jun 2006) | 11 lines Implementing a happy idea from Georg Brandl: make runtest() try to clean up files and directories the tests often leave behind by mistake. This is the first time in history I don't have a bogus "db_home" directory after running the tests ;-) Also worked on runtest's docstring, to say something about all the arguments, and to document the non-obvious return values. New functions runtest_inner() and cleanup_test_droppings() in support of the above. ........ r46800 | andrew.kuchling | 2006-06-09 21:43:25 +0200 (Fri, 09 Jun 2006) | 1 line Remove unused variable ........ r46801 | andrew.kuchling | 2006-06-09 21:56:05 +0200 (Fri, 09 Jun 2006) | 1 line Add some wsgiref text ........ r46803 | thomas.heller | 2006-06-09 21:59:11 +0200 (Fri, 09 Jun 2006) | 1 line set eol-style svn property ........ r46804 | thomas.heller | 2006-06-09 22:01:01 +0200 (Fri, 09 Jun 2006) | 1 line set eol-style svn property ........ r46805 | georg.brandl | 2006-06-09 22:43:48 +0200 (Fri, 09 Jun 2006) | 3 lines Make use of new str.startswith/endswith semantics. Occurences in email and compiler were ignored due to backwards compat requirements. ........ r46806 | brett.cannon | 2006-06-10 00:31:23 +0200 (Sat, 10 Jun 2006) | 4 lines An object with __call__ as an attribute, when called, will have that attribute checked for __call__ itself, and will continue to look until it finds an object without the attribute. This can lead to an infinite recursion. Closes bug #532646, again. Will be backported. ........ r46808 | brett.cannon | 2006-06-10 00:45:54 +0200 (Sat, 10 Jun 2006) | 2 lines Fix bug introduced in rev. 46806 by not having variable declaration at the top of a block. ........ r46812 | georg.brandl | 2006-06-10 08:40:50 +0200 (Sat, 10 Jun 2006) | 4 lines Apply perky's fix for #1503157: "/".join([u"", u""]) raising OverflowError. Also improve error message on overflow. ........ r46817 | martin.v.loewis | 2006-06-10 10:14:03 +0200 (Sat, 10 Jun 2006) | 2 lines Port cygwin kill_python changes from 2.4 branch. ........ r46818 | armin.rigo | 2006-06-10 12:57:40 +0200 (Sat, 10 Jun 2006) | 4 lines SF bug #1503294. PyThreadState_GET() complains if the tstate is NULL, but only in debug mode. ........ r46819 | martin.v.loewis | 2006-06-10 14:23:46 +0200 (Sat, 10 Jun 2006) | 4 lines Patch #1495999: Part two of Windows CE changes. - update header checks, using autoconf - provide dummies for getenv, environ, and GetVersion - adjust MSC_VER check in socketmodule.c ........ r46820 | skip.montanaro | 2006-06-10 16:09:11 +0200 (Sat, 10 Jun 2006) | 1 line document the class, not its initializer ........ r46821 | greg.ward | 2006-06-10 18:40:01 +0200 (Sat, 10 Jun 2006) | 4 lines Sync with Optik docs (rev 518): * restore "Extending optparse" section * document ALWAYS_TYPED_ACTIONS (SF #1449311) ........ r46824 | thomas.heller | 2006-06-10 21:51:46 +0200 (Sat, 10 Jun 2006) | 8 lines Upgrade to ctypes version 0.9.9.7. Summary of changes: - support for 'variable sized' data - support for anonymous structure/union fields - fix severe bug with certain arrays or structures containing more than 256 fields ........ r46825 | thomas.heller | 2006-06-10 21:55:36 +0200 (Sat, 10 Jun 2006) | 8 lines Upgrade to ctypes version 0.9.9.7. Summary of changes: - support for 'variable sized' data - support for anonymous structure/union fields - fix severe bug with certain arrays or structures containing more than 256 fields ........ r46826 | fred.drake | 2006-06-10 22:01:34 +0200 (Sat, 10 Jun 2006) | 4 lines SF patch #1303595: improve description of __builtins__, explaining how it varies between __main__ and other modules, and strongly suggest not touching it but using __builtin__ if absolutely necessary ........ r46827 | fred.drake | 2006-06-10 22:02:58 +0200 (Sat, 10 Jun 2006) | 1 line credit for SF patch #1303595 ........ r46831 | thomas.heller | 2006-06-10 22:29:34 +0200 (Sat, 10 Jun 2006) | 2 lines New docs for ctypes. ........ r46834 | thomas.heller | 2006-06-10 23:07:19 +0200 (Sat, 10 Jun 2006) | 1 line Fix a wrong printf format. ........ r46835 | thomas.heller | 2006-06-10 23:17:58 +0200 (Sat, 10 Jun 2006) | 1 line Fix the second occurrence of the problematic printf format. ........ r46837 | thomas.heller | 2006-06-10 23:56:03 +0200 (Sat, 10 Jun 2006) | 1 line Don't use C++ comment. ........ r46838 | thomas.heller | 2006-06-11 00:01:50 +0200 (Sun, 11 Jun 2006) | 1 line Handle failure of PyMem_Realloc. ........ r46839 | skip.montanaro | 2006-06-11 00:38:13 +0200 (Sun, 11 Jun 2006) | 2 lines Suppress warning on MacOSX about possible use before set of proc. ........ r46840 | tim.peters | 2006-06-11 00:51:45 +0200 (Sun, 11 Jun 2006) | 8 lines shuffle() doscstring: Removed warning about sequence length versus generator period. While this was a real weakness of the older WH generator for lists with just a few dozen elements, and so could potentially bite the naive ;-), the Twister should show excellent behavior up to at least 600 elements. Module docstring: reflowed some jarringly short lines. ........ r46844 | greg.ward | 2006-06-11 02:40:49 +0200 (Sun, 11 Jun 2006) | 4 lines Bug #1361643: fix textwrap.dedent() so it handles tabs appropriately, i.e. do *not* expand tabs, but treat them as whitespace that is not equivalent to spaces. Add a couple of test cases. Clarify docs. ........ r46850 | neal.norwitz | 2006-06-11 07:44:18 +0200 (Sun, 11 Jun 2006) | 5 lines Fix Coverity # 146. newDBSequenceObject would deref dbobj, so it can't be NULL. We know it's not NULL from the ParseTuple and DbObject_Check will verify it's not NULL. ........ r46851 | neal.norwitz | 2006-06-11 07:45:25 +0200 (Sun, 11 Jun 2006) | 4 lines Wrap some long lines Top/Bottom factor out some common expressions Add a XXX comment about widing offset. ........ r46852 | neal.norwitz | 2006-06-11 07:45:47 +0200 (Sun, 11 Jun 2006) | 1 line Add versionadded to doc ........ r46853 | neal.norwitz | 2006-06-11 07:47:14 +0200 (Sun, 11 Jun 2006) | 3 lines Update doc to make it agree with code. Bottom factor out some common code. ........ r46854 | neal.norwitz | 2006-06-11 07:48:14 +0200 (Sun, 11 Jun 2006) | 3 lines f_code can't be NULL based on Frame_New and other code that derefs it. So there doesn't seem to be much point to checking here. ........ r46855 | neal.norwitz | 2006-06-11 09:26:27 +0200 (Sun, 11 Jun 2006) | 1 line Fix errors found by pychecker ........ r46856 | neal.norwitz | 2006-06-11 09:26:50 +0200 (Sun, 11 Jun 2006) | 1 line warnings was imported at module scope, no need to import again ........ r46857 | neal.norwitz | 2006-06-11 09:27:56 +0200 (Sun, 11 Jun 2006) | 5 lines Fix errors found by pychecker. I think these changes are correct, but I'm not sure. Could someone who knows how this module works test it? It can at least start on the cmd line. ........ r46858 | neal.norwitz | 2006-06-11 10:35:14 +0200 (Sun, 11 Jun 2006) | 1 line Fix errors found by pychecker ........ r46859 | ronald.oussoren | 2006-06-11 16:33:36 +0200 (Sun, 11 Jun 2006) | 4 lines This patch improves the L&F of IDLE on OSX. The changes are conditionalized on being in an IDLE.app bundle on darwin. This does a slight reorganisation of the menus and adds support for file-open events. ........ r46860 | greg.ward | 2006-06-11 16:42:41 +0200 (Sun, 11 Jun 2006) | 1 line SF #1366250: optparse docs: fix inconsistency in variable name; minor tweaks. ........ r46861 | greg.ward | 2006-06-11 18:24:11 +0200 (Sun, 11 Jun 2006) | 3 lines Bug #1498146: fix optparse to handle Unicode strings in option help, description, and epilog. ........ r46862 | thomas.heller | 2006-06-11 19:04:22 +0200 (Sun, 11 Jun 2006) | 2 lines Release the GIL during COM method calls, to avoid deadlocks in Python coded COM objects. ........ r46863 | tim.peters | 2006-06-11 21:42:51 +0200 (Sun, 11 Jun 2006) | 2 lines Whitespace normalization. ........ r46864 | tim.peters | 2006-06-11 21:43:49 +0200 (Sun, 11 Jun 2006) | 2 lines Add missing svn:eol-style property to text files. ........ r46865 | ronald.oussoren | 2006-06-11 21:45:57 +0200 (Sun, 11 Jun 2006) | 2 lines Remove message about using make frameworkinstall, that's no longer necesssary ........ r46866 | ronald.oussoren | 2006-06-11 22:23:29 +0200 (Sun, 11 Jun 2006) | 2 lines Use configure to substitute the correct prefix instead of hardcoding ........ r46867 | ronald.oussoren | 2006-06-11 22:24:45 +0200 (Sun, 11 Jun 2006) | 4 lines - Change fixapplepython23.py to ensure that it will run with /usr/bin/python on intel macs. - Fix some minor problems in the installer for OSX ........ r46868 | neal.norwitz | 2006-06-11 22:25:56 +0200 (Sun, 11 Jun 2006) | 5 lines Try to fix several networking tests. The problem is that if hosts have a search path setup, some of these hosts resolve to the wrong address. By appending a period to the hostname, the hostname should only resolve to what we want it to resolve to. Hopefully this doesn't break different bots. ........ r46869 | neal.norwitz | 2006-06-11 22:42:02 +0200 (Sun, 11 Jun 2006) | 7 lines Try to fix another networking test. The problem is that if hosts have a search path setup, some of these hosts resolve to the wrong address. By appending a period to the hostname, the hostname should only resolve to what we want it to resolve to. Hopefully this doesn't break different bots. Also add more info to failure message to aid debugging test failure. ........ r46870 | neal.norwitz | 2006-06-11 22:46:46 +0200 (Sun, 11 Jun 2006) | 4 lines Fix test on PPC64 buildbot. It raised an IOError (really an URLError which derives from an IOError). That seems valid. Env Error includes both OSError and IOError, so this seems like a reasonable fix. ........ r46871 | tim.peters | 2006-06-11 22:52:59 +0200 (Sun, 11 Jun 2006) | 10 lines compare_generic_iter(): Fixed the failure of test_wsgiref's testFileWrapper when running with -O. test_simple_validation_error still fails under -O. That appears to be because wsgiref's validate.py uses `assert` statements all over the place to check arguments for sanity. That should all be changed (it's not a logical error in the software if a user passes bogus arguments, so this isn't a reasonable use for `assert` -- checking external preconditions should generally raise ValueError or TypeError instead, as appropriate). ........ r46872 | neal.norwitz | 2006-06-11 23:38:38 +0200 (Sun, 11 Jun 2006) | 1 line Get test to pass on S/390. Shout if you think this change is incorrect. ........ r46873 | neal.norwitz | 2006-06-12 04:05:55 +0200 (Mon, 12 Jun 2006) | 1 line Cleanup Py_ssize_t a little (get rid of second #ifdef) ........ r46874 | neal.norwitz | 2006-06-12 04:06:17 +0200 (Mon, 12 Jun 2006) | 1 line Fix some Py_ssize_t issues ........ r46875 | neal.norwitz | 2006-06-12 04:06:42 +0200 (Mon, 12 Jun 2006) | 1 line Fix some Py_ssize_t issues ........ r46876 | neal.norwitz | 2006-06-12 04:07:24 +0200 (Mon, 12 Jun 2006) | 2 lines Cleanup: Remove import of types to get StringTypes, we can just use basestring. ........ r46877 | neal.norwitz | 2006-06-12 04:07:57 +0200 (Mon, 12 Jun 2006) | 1 line Don't truncate if size_t is bigger than uint ........ r46878 | neal.norwitz | 2006-06-12 04:08:41 +0200 (Mon, 12 Jun 2006) | 1 line Don't leak the list object if there's an error allocating the item storage. Backport candidate ........ r46879 | neal.norwitz | 2006-06-12 04:09:03 +0200 (Mon, 12 Jun 2006) | 1 line Fix typo. Backport if anyone cares. :-) ........ r46880 | neal.norwitz | 2006-06-12 04:09:34 +0200 (Mon, 12 Jun 2006) | 1 line Fix indentation of case and a Py_ssize_t issue. ........ r46881 | neal.norwitz | 2006-06-12 04:11:18 +0200 (Mon, 12 Jun 2006) | 3 lines Get rid of f_restricted too. Doc the other 4 ints that were already removed at the NeedForSpeed sprint. ........ r46882 | neal.norwitz | 2006-06-12 04:13:21 +0200 (Mon, 12 Jun 2006) | 1 line Fix the socket tests so they can be run concurrently. Backport candidate ........ r46883 | neal.norwitz | 2006-06-12 04:16:10 +0200 (Mon, 12 Jun 2006) | 1 line i and j are initialized below when used. No need to do it twice ........ r46884 | neal.norwitz | 2006-06-12 05:05:03 +0200 (Mon, 12 Jun 2006) | 1 line Remove unused import ........ r46885 | neal.norwitz | 2006-06-12 05:05:40 +0200 (Mon, 12 Jun 2006) | 1 line Impl ssize_t ........ r46886 | neal.norwitz | 2006-06-12 05:33:09 +0200 (Mon, 12 Jun 2006) | 6 lines Patch #1503046, Conditional compilation of zlib.(de)compressobj.copy copy is only in newer versions of zlib. This should allow zlibmodule to work with older versions like the Tru64 buildbot. ........ r46887 | phillip.eby | 2006-06-12 06:04:32 +0200 (Mon, 12 Jun 2006) | 2 lines Sync w/external release 0.1.2. Please see PEP 360 before making changes to external packages. ........ r46888 | martin.v.loewis | 2006-06-12 06:26:31 +0200 (Mon, 12 Jun 2006) | 2 lines Get rid of function pointer cast. ........ r46889 | thomas.heller | 2006-06-12 08:05:57 +0200 (Mon, 12 Jun 2006) | 3 lines I don't know how that happend, but the entire file contents was duplicated. Thanks to Simon Percivall for the heads up. ........ r46890 | nick.coghlan | 2006-06-12 10:19:37 +0200 (Mon, 12 Jun 2006) | 1 line Fix site module docstring to match the code ........ r46891 | nick.coghlan | 2006-06-12 10:23:02 +0200 (Mon, 12 Jun 2006) | 1 line Fix site module docstring to match the code for Mac OSX, too ........ r46892 | nick.coghlan | 2006-06-12 10:27:13 +0200 (Mon, 12 Jun 2006) | 1 line The site module documentation also described the Windows behaviour incorrectly. ........ r46893 | nick.coghlan | 2006-06-12 12:17:11 +0200 (Mon, 12 Jun 2006) | 1 line Make the -m switch conform to the documentation of sys.path by behaving like the -c switch ........ r46894 | kristjan.jonsson | 2006-06-12 17:45:12 +0200 (Mon, 12 Jun 2006) | 2 lines Fix the CRT argument error handling for VisualStudio .NET 2005. Install a CRT error handler and disable the assertion for debug builds. This causes CRT to set errno to EINVAL. This update fixes crash cases in the test suite where the default CRT error handler would cause process exit. ........ r46899 | thomas.heller | 2006-06-12 22:56:48 +0200 (Mon, 12 Jun 2006) | 1 line Add pep-291 compatibility markers. ........ r46901 | ka-ping.yee | 2006-06-13 01:47:52 +0200 (Tue, 13 Jun 2006) | 5 lines Add the uuid module. This module has been tested so far on Windows XP (Python 2.4 and 2.5a2), Mac OS X (Python 2.3, 2.4, and 2.5a2), and Linux (Python 2.4 and 2.5a2). ........ r46902 | tim.peters | 2006-06-13 02:30:01 +0200 (Tue, 13 Jun 2006) | 2 lines Whitespace normalization. ........ r46903 | tim.peters | 2006-06-13 02:30:50 +0200 (Tue, 13 Jun 2006) | 2 lines Added missing svn:eol-style property to text files. ........ r46905 | tim.peters | 2006-06-13 05:30:07 +0200 (Tue, 13 Jun 2006) | 5 lines get_matching_blocks(): rewrote code & comments so they match; added more comments about why it's this way at all; and removed what looked like needless expense (sorting (i, j, k) triples directly should give exactly the same order as sorting (i, (i, j, k)) pairs). ........ r46906 | neal.norwitz | 2006-06-13 06:08:53 +0200 (Tue, 13 Jun 2006) | 1 line Don't fail if another process is listening on our port. ........ r46908 | neal.norwitz | 2006-06-13 10:28:19 +0200 (Tue, 13 Jun 2006) | 2 lines Initialize the type object so pychecker can't crash the interpreter. ........ r46909 | neal.norwitz | 2006-06-13 10:41:06 +0200 (Tue, 13 Jun 2006) | 1 line Verify the crash due to EncodingMap not initialized does not return ........ r46910 | thomas.heller | 2006-06-13 10:56:14 +0200 (Tue, 13 Jun 2006) | 3 lines Add some windows datatypes that were missing from this file, and add the aliases defined in windows header files for the structures. ........ r46911 | thomas.heller | 2006-06-13 11:40:14 +0200 (Tue, 13 Jun 2006) | 3 lines Add back WCHAR, UINT, DOUBLE, _LARGE_INTEGER, _ULARGE_INTEGER. VARIANT_BOOL is a special _ctypes data type, not c_short. ........ r46912 | ronald.oussoren | 2006-06-13 13:19:56 +0200 (Tue, 13 Jun 2006) | 4 lines Linecache contains support for PEP302 loaders, but fails to deal with loaders that return None to indicate that the module is valid but no source is available. This patch fixes that. ........ r46913 | andrew.kuchling | 2006-06-13 13:57:04 +0200 (Tue, 13 Jun 2006) | 1 line Mention uuid module ........ r46915 | walter.doerwald | 2006-06-13 14:02:12 +0200 (Tue, 13 Jun 2006) | 2 lines Fix passing errors to the encoder and decoder functions. ........ r46917 | walter.doerwald | 2006-06-13 14:04:43 +0200 (Tue, 13 Jun 2006) | 3 lines errors is an attribute in the incremental decoder not an argument. ........ r46919 | andrew.macintyre | 2006-06-13 17:04:24 +0200 (Tue, 13 Jun 2006) | 11 lines Patch #1454481: Make thread stack size runtime tunable. Heavily revised, comprising revisions: 46640 - original trunk revision (backed out in r46655) 46647 - markup fix (backed out in r46655) 46692:46918 merged from branch aimacintyre-sf1454481 branch tested on buildbots (Windows buildbots had problems not related to these changes). ........ r46920 | brett.cannon | 2006-06-13 18:06:55 +0200 (Tue, 13 Jun 2006) | 2 lines Remove unused variable. ........ r46921 | andrew.kuchling | 2006-06-13 18:41:41 +0200 (Tue, 13 Jun 2006) | 1 line Add ability to set stack size ........ r46923 | marc-andre.lemburg | 2006-06-13 19:04:26 +0200 (Tue, 13 Jun 2006) | 2 lines Update pybench to version 2.0. ........ r46924 | marc-andre.lemburg | 2006-06-13 19:07:14 +0200 (Tue, 13 Jun 2006) | 2 lines Revert wrong svn copy. ........ r46925 | andrew.macintyre | 2006-06-13 19:14:36 +0200 (Tue, 13 Jun 2006) | 2 lines fix exception usage ........ r46927 | tim.peters | 2006-06-13 20:37:07 +0200 (Tue, 13 Jun 2006) | 2 lines Whitespace normalization. ........ r46928 | marc-andre.lemburg | 2006-06-13 20:56:56 +0200 (Tue, 13 Jun 2006) | 9 lines Updated to pybench 2.0. See svn.python.org/external/pybench-2.0 for the original import of that version. Note that platform.py was not copied over from pybench-2.0 since it is already part of Python 2.5. ........ r46929 | andrew.macintyre | 2006-06-13 21:02:35 +0200 (Tue, 13 Jun 2006) | 5 lines Increase the small thread stack size to get the test to pass reliably on the one buildbot that insists on more than 32kB of thread stack. ........ r46930 | marc-andre.lemburg | 2006-06-13 21:20:07 +0200 (Tue, 13 Jun 2006) | 2 lines Whitespace normalization. ........ r46931 | thomas.heller | 2006-06-13 22:18:43 +0200 (Tue, 13 Jun 2006) | 2 lines More docs for ctypes. ........ r46932 | brett.cannon | 2006-06-13 23:34:24 +0200 (Tue, 13 Jun 2006) | 2 lines Ignore .pyc and .pyo files in Pybench. ........ r46933 | brett.cannon | 2006-06-13 23:46:41 +0200 (Tue, 13 Jun 2006) | 7 lines If a classic class defined a __coerce__() method that just returned its two arguments in reverse, the interpreter would infinitely recourse trying to get a coercion that worked. So put in a recursion check after a coercion is made and the next call to attempt to use the coerced values. Fixes bug #992017 and closes crashers/coerce.py . ........ r46936 | gerhard.haering | 2006-06-14 00:24:47 +0200 (Wed, 14 Jun 2006) | 3 lines Merged changes from external pysqlite 2.3.0 release. Documentation updates will follow in a few hours at the latest. Then we should be ready for beta1. ........ r46937 | brett.cannon | 2006-06-14 00:26:13 +0200 (Wed, 14 Jun 2006) | 2 lines Missed test for rev. 46933; infinite recursion from __coerce__() returning its arguments reversed. ........ r46938 | gerhard.haering | 2006-06-14 00:53:48 +0200 (Wed, 14 Jun 2006) | 2 lines Updated documentation for pysqlite 2.3.0 API. ........ r46939 | tim.peters | 2006-06-14 06:09:25 +0200 (Wed, 14 Jun 2006) | 10 lines SequenceMatcher.get_matching_blocks(): This now guarantees that adjacent triples in the result list describe non-adjacent matching blocks. That's _nice_ to have, and Guido said he wanted it. Not a bugfix candidate: Guido or not ;-), this changes visible endcase semantics (note that some tests had to change), and nothing about this was documented before. Since it was working as designed, and behavior was consistent with the docs, it wasn't "a bug". ........ r46940 | tim.peters | 2006-06-14 06:13:00 +0200 (Wed, 14 Jun 2006) | 2 lines Repaired typo in new comment. ........ r46941 | tim.peters | 2006-06-14 06:15:27 +0200 (Wed, 14 Jun 2006) | 2 lines Whitespace normalization. ........ r46942 | fred.drake | 2006-06-14 06:25:02 +0200 (Wed, 14 Jun 2006) | 3 lines - make some disabled tests run what they intend when enabled - remove some over-zealous triple-quoting ........ r46943 | fred.drake | 2006-06-14 07:04:47 +0200 (Wed, 14 Jun 2006) | 3 lines add tests for two cases that are handled correctly in the current code, but that SF patch 1504676 as written mis-handles ........ r46944 | fred.drake | 2006-06-14 07:15:51 +0200 (Wed, 14 Jun 2006) | 1 line explain an XXX in more detail ........ r46945 | martin.v.loewis | 2006-06-14 07:21:04 +0200 (Wed, 14 Jun 2006) | 1 line Patch #1455898: Incremental mode for "mbcs" codec. ........ r46946 | georg.brandl | 2006-06-14 08:08:31 +0200 (Wed, 14 Jun 2006) | 3 lines Bug #1339007: Shelf objects now don't raise an exception in their __del__ method when initialization failed. ........ r46948 | thomas.heller | 2006-06-14 08:18:15 +0200 (Wed, 14 Jun 2006) | 1 line Fix docstring. ........ r46949 | georg.brandl | 2006-06-14 08:29:07 +0200 (Wed, 14 Jun 2006) | 2 lines Bug #1501122: mention __gt__ &co in description of comparison order. ........ r46951 | thomas.heller | 2006-06-14 09:08:38 +0200 (Wed, 14 Jun 2006) | 1 line Write more docs. ........ r46952 | georg.brandl | 2006-06-14 10:31:39 +0200 (Wed, 14 Jun 2006) | 3 lines Bug #1153163: describe __add__ vs __radd__ behavior when adding objects of same type/of subclasses of the other. ........ r46954 | georg.brandl | 2006-06-14 10:42:11 +0200 (Wed, 14 Jun 2006) | 3 lines Bug #1202018: add some common mime.types locations. ........ r46955 | georg.brandl | 2006-06-14 10:50:03 +0200 (Wed, 14 Jun 2006) | 3 lines Bug #1117556: SimpleHTTPServer now tries to find and use the system's mime.types file for determining MIME types. ........ r46957 | thomas.heller | 2006-06-14 11:09:08 +0200 (Wed, 14 Jun 2006) | 1 line Document paramflags. ........ r46958 | thomas.heller | 2006-06-14 11:20:11 +0200 (Wed, 14 Jun 2006) | 1 line Add an __all__ list, since this module does 'from ctypes import *'. ........ r46959 | andrew.kuchling | 2006-06-14 15:59:15 +0200 (Wed, 14 Jun 2006) | 1 line Add item ........ r46961 | georg.brandl | 2006-06-14 18:46:43 +0200 (Wed, 14 Jun 2006) | 3 lines Bug #805015: doc error in PyUnicode_FromEncodedObject. ........ r46962 | gerhard.haering | 2006-06-15 00:28:37 +0200 (Thu, 15 Jun 2006) | 10 lines - Added version checks in C code to make sure we don't trigger bugs in older SQLite versions. - Added version checks in test suite so that we don't execute tests that we know will fail with older (buggy) SQLite versions. Now, all tests should run against all SQLite versions from 3.0.8 until 3.3.6 (latest one now). The sqlite3 module can be built against all these SQLite versions and the sqlite3 module does its best to not trigger bugs in SQLite, but using SQLite 3.3.3 or later is recommended. ........ r46963 | tim.peters | 2006-06-15 00:38:13 +0200 (Thu, 15 Jun 2006) | 2 lines Whitespace normalization. ........ r46964 | neal.norwitz | 2006-06-15 06:54:29 +0200 (Thu, 15 Jun 2006) | 9 lines Speculative checkin (requires approval of Gerhard Haering) This backs out the test changes in 46962 which prevented crashes by not running the tests via a version check. All the version checks added in that rev were removed from the tests. Code was added to the error handler in connection.c that seems to work with older versions of sqlite including 3.1.3. ........ r46965 | neal.norwitz | 2006-06-15 07:55:49 +0200 (Thu, 15 Jun 2006) | 1 line Try to narrow window of failure on slow/busy boxes (ppc64 buildbot) ........ r46966 | martin.v.loewis | 2006-06-15 08:45:05 +0200 (Thu, 15 Jun 2006) | 2 lines Make import/lookup of mbcs fail on non-Windows systems. ........ r46967 | ronald.oussoren | 2006-06-15 10:14:18 +0200 (Thu, 15 Jun 2006) | 2 lines Patch #1446489 (zipfile: support for ZIP64) ........ r46968 | neal.norwitz | 2006-06-15 10:16:44 +0200 (Thu, 15 Jun 2006) | 6 lines Re-revert this change. Install the version check and don't run the test until Gerhard has time to fully debug the issue. This affects versions before 3.2.1 (possibly only versions earlier than 3.1.3). Based on discussion on python-checkins. ........ r46969 | gregory.p.smith | 2006-06-15 10:52:32 +0200 (Thu, 15 Jun 2006) | 6 lines - bsddb: multithreaded DB access using the simple bsddb module interface now works reliably. It has been updated to use automatic BerkeleyDB deadlock detection and the bsddb.dbutils.DeadlockWrap wrapper to retry database calls that would previously deadlock. [SF python bug #775414] ........ r46970 | gregory.p.smith | 2006-06-15 11:23:52 +0200 (Thu, 15 Jun 2006) | 2 lines minor documentation cleanup. mention the bsddb.db interface explicitly by name. ........ r46971 | neal.norwitz | 2006-06-15 11:57:03 +0200 (Thu, 15 Jun 2006) | 5 lines Steal the trick from test_compiler to print out a slow msg. This will hopefully get the buildbots to pass. Not sure this test will be feasible or even work. But everything is red now, so it can't get much worse. ........ r46972 | neal.norwitz | 2006-06-15 12:24:49 +0200 (Thu, 15 Jun 2006) | 1 line Print some more info to get an idea of how much longer the test will last ........ r46981 | tim.peters | 2006-06-15 20:04:40 +0200 (Thu, 15 Jun 2006) | 6 lines Try to reduce the extreme peak memory and disk-space use of this test. It probably still requires more disk space than most buildbots have, and in any case is still so intrusive that if we don't find another way to test this I'm taking my buildbot offline permanently ;-) ........ r46982 | tim.peters | 2006-06-15 20:06:29 +0200 (Thu, 15 Jun 2006) | 2 lines Whitespace normalization. ........ r46983 | tim.peters | 2006-06-15 20:07:28 +0200 (Thu, 15 Jun 2006) | 2 lines Add missing svn:eol-style property to text files. ........ r46984 | tim.peters | 2006-06-15 20:38:19 +0200 (Thu, 15 Jun 2006) | 2 lines Oops -- I introduced an off-by-6436159488 error. ........ r46990 | neal.norwitz | 2006-06-16 06:30:34 +0200 (Fri, 16 Jun 2006) | 1 line Disable this test until we can determine what to do about it ........ r46991 | neal.norwitz | 2006-06-16 06:31:06 +0200 (Fri, 16 Jun 2006) | 1 line Param name is dir, not directory. Update docstring. Backport candidate ........ r46992 | neal.norwitz | 2006-06-16 06:31:28 +0200 (Fri, 16 Jun 2006) | 1 line Add missing period in comment. ........ r46993 | neal.norwitz | 2006-06-16 06:32:43 +0200 (Fri, 16 Jun 2006) | 1 line Fix whitespace, there are memory leaks in this module. ........ r46995 | fred.drake | 2006-06-17 01:45:06 +0200 (Sat, 17 Jun 2006) | 3 lines SF patch 1504676: Make sgmllib char and entity references pluggable (implementation/tests contributed by Sam Ruby) ........ r46996 | fred.drake | 2006-06-17 03:07:54 +0200 (Sat, 17 Jun 2006) | 1 line fix change that broke the htmllib tests ........ r46998 | martin.v.loewis | 2006-06-17 11:15:14 +0200 (Sat, 17 Jun 2006) | 3 lines Patch #763580: Add name and value arguments to Tkinter variable classes. ........ r46999 | martin.v.loewis | 2006-06-17 11:20:41 +0200 (Sat, 17 Jun 2006) | 2 lines Patch #1096231: Add default argument to wm_iconbitmap. ........ r47000 | martin.v.loewis | 2006-06-17 11:25:15 +0200 (Sat, 17 Jun 2006) | 2 lines Patch #1494750: Destroy master after deleting children. ........ r47003 | george.yoshida | 2006-06-17 18:31:52 +0200 (Sat, 17 Jun 2006) | 2 lines markup fix ........ r47005 | george.yoshida | 2006-06-17 18:39:13 +0200 (Sat, 17 Jun 2006) | 4 lines Update url. Old url returned status code:301 Moved permanently. ........ r47007 | martin.v.loewis | 2006-06-17 20:44:27 +0200 (Sat, 17 Jun 2006) | 2 lines Patch #812986: Update the canvas even if not tracing. ........ r47008 | martin.v.loewis | 2006-06-17 21:03:26 +0200 (Sat, 17 Jun 2006) | 2 lines Patch #815924: Restore ability to pass type= and icon= ........ r47009 | neal.norwitz | 2006-06-18 00:37:45 +0200 (Sun, 18 Jun 2006) | 1 line Fix typo in docstring ........ r47010 | neal.norwitz | 2006-06-18 00:38:15 +0200 (Sun, 18 Jun 2006) | 1 line Fix memory leak reported by valgrind while running test_subprocess ........ r47011 | fred.drake | 2006-06-18 04:57:35 +0200 (Sun, 18 Jun 2006) | 1 line remove unnecessary markup ........ r47013 | neal.norwitz | 2006-06-18 21:35:01 +0200 (Sun, 18 Jun 2006) | 7 lines Prevent spurious leaks when running regrtest.py -R. There may be more issues that crop up from time to time, but this change seems to have been pretty stable (no spurious warnings) for about a week. Other modules which use threads may require similar use of threading_setup/threading_cleanup from test_support. ........ r47014 | neal.norwitz | 2006-06-18 21:37:40 +0200 (Sun, 18 Jun 2006) | 9 lines The hppa ubuntu box sometimes hangs forever in these tests. My guess is that the wait is failing for some reason. Use WNOHANG, so we won't wait until the buildbot kills the test suite. I haven't been able to reproduce the failure, so I'm not sure if this will help or not. Hopefully, this change will cause the test to fail, rather than hang. That will be better since we will get the rest of the test results. It may also help us debug the real problem. ........ r47015 | neal.norwitz | 2006-06-18 22:10:24 +0200 (Sun, 18 Jun 2006) | 1 line Revert 47014 until it is more robust ........ r47016 | thomas.heller | 2006-06-18 23:27:04 +0200 (Sun, 18 Jun 2006) | 6 lines Fix typos. Fix doctest example. Mention in the tutorial that 'errcheck' is explained in the ref manual. Use better wording in some places. Remoce code examples that shouldn't be in the tutorial. Remove some XXX notices. ........ r47017 | georg.brandl | 2006-06-19 00:17:29 +0200 (Mon, 19 Jun 2006) | 3 lines Patch #1507676: improve exception messages in abstract.c, object.c and typeobject.c. ........ r47018 | neal.norwitz | 2006-06-19 07:40:44 +0200 (Mon, 19 Jun 2006) | 1 line Use Py_ssize_t ........ r47019 | georg.brandl | 2006-06-19 08:35:54 +0200 (Mon, 19 Jun 2006) | 3 lines Add news entry about error msg improvement. ........ r47020 | thomas.heller | 2006-06-19 09:07:49 +0200 (Mon, 19 Jun 2006) | 2 lines Try to repair the failing test on the OpenBSD buildbot. Trial and error... ........ r47021 | tim.peters | 2006-06-19 09:45:16 +0200 (Mon, 19 Jun 2006) | 2 lines Whitespace normalization. ........ r47022 | walter.doerwald | 2006-06-19 10:07:50 +0200 (Mon, 19 Jun 2006) | 4 lines Patch #1506645: add Python wrappers for the curses functions is_term_resized, resize_term and resizeterm. This uses three separate configure checks (one for each function). ........ r47023 | walter.doerwald | 2006-06-19 10:14:09 +0200 (Mon, 19 Jun 2006) | 2 lines Make check order match in configure and configure.in. ........ r47024 | tim.peters | 2006-06-19 10:14:28 +0200 (Mon, 19 Jun 2006) | 3 lines Repair KeyError when running test_threaded_import under -R, as reported by Neal on python-dev. ........ r47025 | thomas.heller | 2006-06-19 10:32:46 +0200 (Mon, 19 Jun 2006) | 3 lines Next try to fix the OpenBSD buildbot tests: Use ctypes.util.find_library to locate the C runtime library on platforms where is returns useful results. ........ r47026 | tim.peters | 2006-06-19 11:09:44 +0200 (Mon, 19 Jun 2006) | 13 lines TestHelp.make_parser(): This was making a permanent change to os.environ (setting envar COLUMNS), which at least caused test_float_default() to fail if the tests were run more than once. This repairs the test_optparse -R failures Neal reported on python-dev. It also explains some seemingly bizarre test_optparse failures we saw a couple weeks ago on the buildbots, when test_optparse failed due to test_file failing to clean up after itself, and then test_optparse failed in an entirely different way when regrtest's -w option ran test_optparse a second time. It's now obvious that make_parser() permanently changing os.environ was responsible for the second half of that. ........ r47027 | anthony.baxter | 2006-06-19 14:04:15 +0200 (Mon, 19 Jun 2006) | 2 lines Preparing for 2.5b1. ........ r47029 | fred.drake | 2006-06-19 19:31:16 +0200 (Mon, 19 Jun 2006) | 1 line remove non-working document formats from edist ........ r47030 | gerhard.haering | 2006-06-19 23:17:35 +0200 (Mon, 19 Jun 2006) | 5 lines Fixed a memory leak that was introduced with incorrect usage of the Python weak reference API in pysqlite 2.2.1. Bumbed pysqlite version number to upcoming pysqlite 2.3.1 release. ........ r47032 | ka-ping.yee | 2006-06-20 00:49:36 +0200 (Tue, 20 Jun 2006) | 2 lines Remove Python 2.3 compatibility comment. ........ r47033 | trent.mick | 2006-06-20 01:21:25 +0200 (Tue, 20 Jun 2006) | 2 lines Upgrade pyexpat to expat 2.0.0 (http://python.org/sf/1462338). ........ r47034 | trent.mick | 2006-06-20 01:57:41 +0200 (Tue, 20 Jun 2006) | 3 lines [ 1295808 ] expat symbols should be namespaced in pyexpat (http://python.org/sf/1295808) ........ r47039 | andrew.kuchling | 2006-06-20 13:52:16 +0200 (Tue, 20 Jun 2006) | 1 line Uncomment wsgiref section ........ r47040 | andrew.kuchling | 2006-06-20 14:15:09 +0200 (Tue, 20 Jun 2006) | 1 line Add four library items ........ r47041 | andrew.kuchling | 2006-06-20 14:19:54 +0200 (Tue, 20 Jun 2006) | 1 line Terminology and typography fixes ........ r47042 | andrew.kuchling | 2006-06-20 15:05:12 +0200 (Tue, 20 Jun 2006) | 1 line Add introductory paragraphs summarizing the release; minor edits ........ r47043 | andrew.kuchling | 2006-06-20 15:11:29 +0200 (Tue, 20 Jun 2006) | 1 line Minor edits and rearrangements; markup fix ........ r47044 | andrew.kuchling | 2006-06-20 15:20:30 +0200 (Tue, 20 Jun 2006) | 1 line [Bug #1504456] Mention xml -> xmlcore change ........ r47047 | brett.cannon | 2006-06-20 19:30:26 +0200 (Tue, 20 Jun 2006) | 2 lines Raise TestSkipped when the test socket connection is refused. ........ r47049 | brett.cannon | 2006-06-20 21:20:17 +0200 (Tue, 20 Jun 2006) | 2 lines Fix typo of exception name. ........ r47053 | brett.cannon | 2006-06-21 18:57:57 +0200 (Wed, 21 Jun 2006) | 5 lines At the C level, tuple arguments are passed in directly to the exception constructor, meaning it is treated as *args, not as a single argument. This means using the 'message' attribute won't work (until Py3K comes around), and so one must grab from 'arg' to get the error number. ........ r47054 | andrew.kuchling | 2006-06-21 19:10:18 +0200 (Wed, 21 Jun 2006) | 1 line Link to LibRef module documentation ........ r47055 | andrew.kuchling | 2006-06-21 19:17:10 +0200 (Wed, 21 Jun 2006) | 1 line Note some of Barry's work ........ r47056 | andrew.kuchling | 2006-06-21 19:17:28 +0200 (Wed, 21 Jun 2006) | 1 line Bump version ........ r47057 | georg.brandl | 2006-06-21 19:45:17 +0200 (Wed, 21 Jun 2006) | 3 lines fix [ 1509132 ] compiler module builds incorrect AST for TryExceptFinally ........ r47058 | georg.brandl | 2006-06-21 19:52:36 +0200 (Wed, 21 Jun 2006) | 3 lines Make test_fcntl aware of netbsd3. ........ r47059 | georg.brandl | 2006-06-21 19:53:17 +0200 (Wed, 21 Jun 2006) | 3 lines Patch #1509001: expected skips for netbsd3. ........ r47060 | gerhard.haering | 2006-06-21 22:55:04 +0200 (Wed, 21 Jun 2006) | 2 lines Removed call to enable_callback_tracebacks that slipped in by accident. ........ r47061 | armin.rigo | 2006-06-21 23:58:50 +0200 (Wed, 21 Jun 2006) | 13 lines Fix for an obscure bug introduced by revs 46806 and 46808, with a test. The problem of checking too eagerly for recursive calls is the following: if a RuntimeError is caused by recursion, and if code needs to normalize it immediately (as in the 2nd test), then PyErr_NormalizeException() needs a call to the RuntimeError class to instantiate it, and this hits the recursion limit again... causing PyErr_NormalizeException() to never finish. Moved this particular recursion check to slot_tp_call(), which is not involved in instantiating built-in exceptions. Backport candidate. ........ r47064 | neal.norwitz | 2006-06-22 08:30:50 +0200 (Thu, 22 Jun 2006) | 3 lines Copy the wsgiref package during make install. ........ r47065 | neal.norwitz | 2006-06-22 08:35:30 +0200 (Thu, 22 Jun 2006) | 1 line Reset the doc date to today for the automatic doc builds ........ r47067 | andrew.kuchling | 2006-06-22 15:10:23 +0200 (Thu, 22 Jun 2006) | 1 line Mention how to suppress warnings ........ r47069 | georg.brandl | 2006-06-22 16:46:17 +0200 (Thu, 22 Jun 2006) | 3 lines Set lineno correctly on list, tuple and dict literals. ........ r47070 | georg.brandl | 2006-06-22 16:46:46 +0200 (Thu, 22 Jun 2006) | 4 lines Test for correct compilation of try-except-finally stmt. Test for correct lineno on list, tuple, dict literals. ........ r47071 | fred.drake | 2006-06-22 17:50:08 +0200 (Thu, 22 Jun 2006) | 1 line fix markup nit ........ r47072 | brett.cannon | 2006-06-22 18:49:14 +0200 (Thu, 22 Jun 2006) | 6 lines 'warning's was improperly requiring that a command-line Warning category be both a subclass of Warning and a subclass of types.ClassType. The latter is no longer true thanks to new-style exceptions. Closes bug #1510580. Thanks to AMK for the test. ........ r47073 | ronald.oussoren | 2006-06-22 20:33:54 +0200 (Thu, 22 Jun 2006) | 3 lines MacOSX: Add a message to the first screen of the installer that tells users how to avoid updates to their shell profile. ........ r47074 | georg.brandl | 2006-06-22 21:02:18 +0200 (Thu, 22 Jun 2006) | 3 lines Fix my name ;) ........ r47075 | thomas.heller | 2006-06-22 21:07:36 +0200 (Thu, 22 Jun 2006) | 2 lines Small fixes, mostly in the markup. ........ r47076 | peter.astrand | 2006-06-22 22:06:46 +0200 (Thu, 22 Jun 2006) | 1 line Make it possible to run test_subprocess.py on Python 2.2, which lacks test_support.is_resource_enabled. ........ r47077 | peter.astrand | 2006-06-22 22:21:26 +0200 (Thu, 22 Jun 2006) | 1 line Applied patch #1506758: Prevent MemoryErrors with large MAXFD. ........ r47079 | neal.norwitz | 2006-06-23 05:32:44 +0200 (Fri, 23 Jun 2006) | 1 line Fix refleak ........ r47080 | fred.drake | 2006-06-23 08:03:45 +0200 (Fri, 23 Jun 2006) | 9 lines - SF bug #853506: IP6 address parsing in sgmllib ('[' and ']' were not accepted in unquoted attribute values) - cleaned up tests of character and entity reference decoding so the tests cover the documented relationships among handle_charref, handle_entityref, convert_charref, convert_codepoint, and convert_entityref, without bringing up Unicode issues that sgmllib cannot be involved in ........ r47085 | andrew.kuchling | 2006-06-23 21:23:40 +0200 (Fri, 23 Jun 2006) | 11 lines Fit Makefile for the Python doc environment better; this is a step toward including the howtos in the build process. * Put LaTeX output in ../paper-<whatever>/. * Put HTML output in ../html/ * Explain some of the Makefile variables * Remove some cruft dating to my environment (e.g. the 'web' target) This makefile isn't currently invoked by the documentation build process, so these changes won't destabilize anything. ........ r47086 | hyeshik.chang | 2006-06-23 23:16:18 +0200 (Fri, 23 Jun 2006) | 5 lines Bug #1511381: codec_getstreamcodec() in codec.c is corrected to omit a default "error" argument for NULL pointer. This allows the parser to take a codec from cjkcodecs again. (Reported by Taewook Kang and reviewed by Walter Doerwald) ........ r47091 | ronald.oussoren | 2006-06-25 22:44:16 +0200 (Sun, 25 Jun 2006) | 6 lines Workaround for bug #1512124 Without this patch IDLE will get unresponsive when you open the debugger window on OSX. This is both using the system Tcl/Tk on Tiger as the latest universal download from tk-components.sf.net. ........ r47092 | ronald.oussoren | 2006-06-25 23:14:19 +0200 (Sun, 25 Jun 2006) | 3 lines Drop the calldll demo's for macos, calldll isn't present anymore, no need to keep the demo's around. ........ r47093 | ronald.oussoren | 2006-06-25 23:15:58 +0200 (Sun, 25 Jun 2006) | 3 lines Use a path without a double slash to compile the .py files after installation (macosx, binary installer). This fixes bug #1508369 for python 2.5. ........ r47094 | ronald.oussoren | 2006-06-25 23:19:06 +0200 (Sun, 25 Jun 2006) | 3 lines Also install the .egg-info files in Lib. This will cause wsgiref.egg-info to be installed. ........ r47097 | andrew.kuchling | 2006-06-26 14:40:02 +0200 (Mon, 26 Jun 2006) | 1 line [Bug #1511998] Various comments from Nick Coghlan; thanks! ........ r47098 | andrew.kuchling | 2006-06-26 14:43:43 +0200 (Mon, 26 Jun 2006) | 1 line Describe workaround for PyRange_New()'s removal ........ r47099 | andrew.kuchling | 2006-06-26 15:08:24 +0200 (Mon, 26 Jun 2006) | 5 lines [Bug #1512163] Fix typo. This change will probably break tests on FreeBSD buildbots, but I'll check in a fix for that next. ........ r47100 | andrew.kuchling | 2006-06-26 15:12:16 +0200 (Mon, 26 Jun 2006) | 9 lines [Bug #1512163] Use one set of locking methods, lockf(); remove the flock() calls. On FreeBSD, the two methods lockf() and flock() end up using the same mechanism and the second one fails. A Linux man page claims that the two methods are orthogonal (so locks acquired one way don't interact with locks acquired the other way) but that clearly must be false. ........ r47101 | andrew.kuchling | 2006-06-26 15:23:10 +0200 (Mon, 26 Jun 2006) | 5 lines Add a test for a conflicting lock. On slow machines, maybe the time intervals (2 sec, 0.5 sec) will be too tight. I'll see how the buildbots like it. ........ r47103 | andrew.kuchling | 2006-06-26 16:33:24 +0200 (Mon, 26 Jun 2006) | 1 line Windows doesn't have os.fork(). I'll just disable this test for now ........ r47106 | andrew.kuchling | 2006-06-26 19:00:35 +0200 (Mon, 26 Jun 2006) | 9 lines Attempt to fix build failure on OS X and Debian alpha; the symptom is consistent with os.wait() returning immediately because some other subprocess had previously exited; the test suite then immediately tries to lock the mailbox and gets an error saying it's already locked. To fix this, do a waitpid() so the test suite only continues once the intended child process has exited. ........ r47113 | neal.norwitz | 2006-06-27 06:06:46 +0200 (Tue, 27 Jun 2006) | 1 line Ignore some more warnings in the dynamic linker on an older gentoo ........ r47114 | neal.norwitz | 2006-06-27 06:09:13 +0200 (Tue, 27 Jun 2006) | 6 lines Instead of doing a make test, run the regression tests out of the installed copy. This will hopefully catch problems where directories are added under Lib/ but not to Makefile.pre.in. This breaks out the 2 runs of the test suite with and without -O which is also nicer. ........ r47115 | neal.norwitz | 2006-06-27 06:12:58 +0200 (Tue, 27 Jun 2006) | 5 lines Fix SF bug #1513032, 'make install' failure on FreeBSD 5.3. No need to install lib-old, it's empty in 2.5. ........ r47116 | neal.norwitz | 2006-06-27 06:23:06 +0200 (Tue, 27 Jun 2006) | 1 line Test unimportant change to verify buildbot does not try to build ........ r47117 | neal.norwitz | 2006-06-27 06:26:30 +0200 (Tue, 27 Jun 2006) | 1 line Try again: test unimportant change to verify buildbot does not try to build ........ r47118 | neal.norwitz | 2006-06-27 06:28:56 +0200 (Tue, 27 Jun 2006) | 1 line Verify buildbot picks up these changes (really needs testing after last change to Makefile.pre.in) ........ r47121 | vinay.sajip | 2006-06-27 09:34:37 +0200 (Tue, 27 Jun 2006) | 1 line Removed buggy exception handling in doRollover of rotating file handlers. Exceptions now propagate to caller. ........ r47123 | ronald.oussoren | 2006-06-27 12:08:25 +0200 (Tue, 27 Jun 2006) | 3 lines MacOSX: fix rather dumb buglet that made it impossible to create extensions on OSX 10.3 when using a binary distribution build on 10.4. ........ r47125 | tim.peters | 2006-06-27 13:52:49 +0200 (Tue, 27 Jun 2006) | 2 lines Whitespace normalization. ........ r47128 | ronald.oussoren | 2006-06-27 14:53:52 +0200 (Tue, 27 Jun 2006) | 8 lines Use staticly build copies of zlib and bzip2 to build the OSX installer, that way the resulting binaries have a better change of running on 10.3. This patch also updates the search logic for sleepycat db3/4, without this patch you cannot use a sleepycat build with a non-standard prefix; with this you can (at least on OSX) if you add the prefix to CPPFLAGS/LDFLAGS at configure-time. This change is needed to build the binary installer for OSX. ........ r47131 | ronald.oussoren | 2006-06-27 17:45:32 +0200 (Tue, 27 Jun 2006) | 5 lines macosx: Install a libpython2.5.a inside the framework as a symlink to the actual dylib at the root of the framework, that way tools that expect a unix-like install (python-config, but more importantly external products like mod_python) work correctly. ........ r47137 | neal.norwitz | 2006-06-28 07:03:22 +0200 (Wed, 28 Jun 2006) | 4 lines According to the man pages on Gentoo Linux and Tru64, EACCES or EAGAIN can be returned if fcntl (lockf) fails. This fixes the test failure on Tru64 by checking for either error rather than just EAGAIN. ........ r47139 | neal.norwitz | 2006-06-28 08:28:31 +0200 (Wed, 28 Jun 2006) | 5 lines Fix bug #1512695: cPickle.loads could crash if it was interrupted with a KeyboardInterrupt since PyTuple_Pack was passed a NULL. Will backport. ........ r47142 | nick.coghlan | 2006-06-28 12:41:47 +0200 (Wed, 28 Jun 2006) | 1 line Make full module name available as __module_name__ even when __name__ is set to something else (like '__main__') ........ r47143 | armin.rigo | 2006-06-28 12:49:51 +0200 (Wed, 28 Jun 2006) | 2 lines A couple of crashers of the "won't fix" kind. ........ r47147 | andrew.kuchling | 2006-06-28 16:25:20 +0200 (Wed, 28 Jun 2006) | 1 line [Bug #1508766] Add docs for uuid module; docs written by George Yoshida, with minor rearrangements by me. ........ r47148 | andrew.kuchling | 2006-06-28 16:27:21 +0200 (Wed, 28 Jun 2006) | 1 line [Bug #1508766] Add docs for uuid module; this puts the module in the 'Internet Protocols' section. Arguably this module could also have gone in the chapters on strings or encodings, maybe even the crypto chapter. Fred, please move if you see fit. ........ r47151 | georg.brandl | 2006-06-28 22:23:25 +0200 (Wed, 28 Jun 2006) | 3 lines Fix end_fill(). ........ r47153 | trent.mick | 2006-06-28 22:30:41 +0200 (Wed, 28 Jun 2006) | 2 lines Mention the expat upgrade and pyexpat fix I put in 2.5b1. ........ r47154 | fred.drake | 2006-06-29 02:51:53 +0200 (Thu, 29 Jun 2006) | 6 lines SF bug #1504333: sgmlib should allow angle brackets in quoted values (modified patch by Sam Ruby; changed to use separate REs for start and end tags to reduce matching cost for end tags; extended tests; updated to avoid breaking previous changes to support IPv6 addresses in unquoted attribute values) ........ r47156 | fred.drake | 2006-06-29 04:57:48 +0200 (Thu, 29 Jun 2006) | 1 line document recent bugfixes in sgmllib ........ r47158 | neal.norwitz | 2006-06-29 06:10:08 +0200 (Thu, 29 Jun 2006) | 10 lines Add new utility function, reap_children(), to test_support. This should be called at the end of each test that spawns children (perhaps it should be called from regrtest instead?). This will hopefully prevent some of the unexplained failures in the buildbots (hppa and alpha) during tests that spawn children. The problems were not reproducible. There were many zombies that remained at the end of several tests. In the worst case, this shouldn't cause any more problems, though it may not help either. Time will tell. ........ r47159 | neal.norwitz | 2006-06-29 07:48:14 +0200 (Thu, 29 Jun 2006) | 5 lines This should fix the buildbot failure on s/390 which can't connect to gmail.org. It makes the error message consistent and always sends to stderr. It would be much better for all the networking tests to hit only python.org. ........ r47161 | thomas.heller | 2006-06-29 20:34:15 +0200 (Thu, 29 Jun 2006) | 3 lines Protect the thread api calls in the _ctypes extension module within #ifdef WITH_THREADS/#endif blocks. Found by Sam Rushing. ........ r47162 | martin.v.loewis | 2006-06-29 20:58:44 +0200 (Thu, 29 Jun 2006) | 2 lines Patch #1509163: MS Toolkit Compiler no longer available ........ r47163 | skip.montanaro | 2006-06-29 21:20:09 +0200 (Thu, 29 Jun 2006) | 1 line add string methods to index ........ r47164 | vinay.sajip | 2006-06-30 02:13:08 +0200 (Fri, 30 Jun 2006) | 1 line Fixed bug in fileConfig() which failed to clear logging._handlerList ........ r47166 | tim.peters | 2006-06-30 08:18:39 +0200 (Fri, 30 Jun 2006) | 2 lines Whitespace normalization. ........ r47170 | neal.norwitz | 2006-06-30 09:32:16 +0200 (Fri, 30 Jun 2006) | 1 line Silence compiler warning ........ r47171 | neal.norwitz | 2006-06-30 09:32:46 +0200 (Fri, 30 Jun 2006) | 1 line Another problem reported by Coverity. Backport candidate. ........ r47175 | thomas.heller | 2006-06-30 19:44:54 +0200 (Fri, 30 Jun 2006) | 2 lines Revert the use of PY_FORMAT_SIZE_T in PyErr_Format. ........ r47176 | tim.peters | 2006-06-30 20:34:51 +0200 (Fri, 30 Jun 2006) | 2 lines Remove now-unused fidding with PY_FORMAT_SIZE_T. ........ r47177 | georg.brandl | 2006-06-30 20:47:56 +0200 (Fri, 30 Jun 2006) | 3 lines Document decorator usage of property. ........ r47181 | fred.drake | 2006-06-30 21:29:25 +0200 (Fri, 30 Jun 2006) | 4 lines - consistency nit: always include "()" in \function and \method (*should* be done by the presentation, but that requires changes all over) - avoid spreading the __name meme ........ r47188 | vinay.sajip | 2006-07-01 12:45:20 +0200 (Sat, 01 Jul 2006) | 1 line Added entry for fileConfig() bugfix. ........ r47189 | vinay.sajip | 2006-07-01 12:47:20 +0200 (Sat, 01 Jul 2006) | 1 line Added duplicate call to fileConfig() to ensure that it cleans up after itself correctly. ........ r47190 | martin.v.loewis | 2006-07-01 17:33:37 +0200 (Sat, 01 Jul 2006) | 2 lines Release all forwarded functions in .close. Fixes #1513223. ........ r47191 | fred.drake | 2006-07-01 18:28:20 +0200 (Sat, 01 Jul 2006) | 7 lines SF bug #1296433 (Expat bug #1515266): Unchecked calls to character data handler would cause a segfault. This merges in Expat's lib/xmlparse.c revisions 1.154 and 1.155, which fix this and a closely related problem (the later does not affect Python). Moved the crasher test to the tests for xml.parsers.expat. ........ r47197 | gerhard.haering | 2006-07-02 19:48:30 +0200 (Sun, 02 Jul 2006) | 4 lines The sqlite3 module did cut off data from the SQLite database at the first null character before sending it to a custom converter. This has been fixed now. ........ r47198 | martin.v.loewis | 2006-07-02 20:44:00 +0200 (Sun, 02 Jul 2006) | 1 line Correct arithmetic in access on Win32. Fixes #1513646. ........ r47203 | thomas.heller | 2006-07-03 09:58:09 +0200 (Mon, 03 Jul 2006) | 1 line Cleanup: Remove commented out code. ........ r47204 | thomas.heller | 2006-07-03 09:59:50 +0200 (Mon, 03 Jul 2006) | 1 line Don't run the doctests with Python 2.3 because it doesn't have the ELLIPSIS flag. ........ r47205 | thomas.heller | 2006-07-03 10:04:05 +0200 (Mon, 03 Jul 2006) | 7 lines Fixes so that _ctypes can be compiled with the MingW compiler. It seems that the definition of '__attribute__(x)' was responsible for the compiler ignoring the '__fastcall' attribute on the ffi_closure_SYSV function in libffi_msvc/ffi.c, took me quite some time to figure this out. ........ r47206 | thomas.heller | 2006-07-03 10:08:14 +0200 (Mon, 03 Jul 2006) | 11 lines Add a new function uses_seh() to the _ctypes extension module. This will return True if Windows Structured Exception handling (SEH) is used when calling functions, False otherwise. Currently, only MSVC supports SEH. Fix the test so that it doesn't crash when run with MingW compiled _ctypes. Note that two tests are still failing when mingw is used, I suspect structure layout differences and function calling conventions between MSVC and MingW. ........ r47207 | tim.peters | 2006-07-03 10:23:19 +0200 (Mon, 03 Jul 2006) | 2 lines Whitespace normalization. ........ r47208 | martin.v.loewis | 2006-07-03 11:44:00 +0200 (Mon, 03 Jul 2006) | 3 lines Only setup canvas when it is first created. Fixes #1514703 ........ r47209 | martin.v.loewis | 2006-07-03 12:05:30 +0200 (Mon, 03 Jul 2006) | 3 lines Reimplement turtle.circle using a polyline, to allow correct filling of arcs. Also fixes #1514693. ........ r47210 | martin.v.loewis | 2006-07-03 12:19:49 +0200 (Mon, 03 Jul 2006) | 3 lines Bug #1514693: Update turtle's heading when switching between degrees and radians. ........ r47211 | martin.v.loewis | 2006-07-03 13:12:06 +0200 (Mon, 03 Jul 2006) | 2 lines Document functions added in 2.3 and 2.5. ........ r47212 | martin.v.loewis | 2006-07-03 14:19:50 +0200 (Mon, 03 Jul 2006) | 3 lines Bug #1417699: Reject locale-specific decimal point in float() and atof(). ........ r47213 | martin.v.loewis | 2006-07-03 14:28:58 +0200 (Mon, 03 Jul 2006) | 3 lines Bug #1267547: Put proper recursive setup.py call into the spec file generated by bdist_rpm. ........ r47215 | martin.v.loewis | 2006-07-03 15:01:35 +0200 (Mon, 03 Jul 2006) | 3 lines Patch #825417: Fix timeout processing in expect, read_until. Will backport to 2.4. ........ r47218 | martin.v.loewis | 2006-07-03 15:47:40 +0200 (Mon, 03 Jul 2006) | 2 lines Put method-wrappers into trashcan. Fixes #927248. ........ r47219 | andrew.kuchling | 2006-07-03 16:07:30 +0200 (Mon, 03 Jul 2006) | 1 line [Bug #1515932] Clarify description of slice assignment ........ r47220 | andrew.kuchling | 2006-07-03 16:16:09 +0200 (Mon, 03 Jul 2006) | 4 lines [Bug #1511911] Clarify description of optional arguments to sorted() by improving the xref to the section on lists, and by copying the explanations of the arguments (with a slight modification). ........ r47223 | kristjan.jonsson | 2006-07-03 16:59:05 +0200 (Mon, 03 Jul 2006) | 1 line Fix build problems with the platform SDK on windows. It is not sufficient to test for the C compiler version when determining if we have the secure CRT from microsoft. Must test with an undocumented macro, __STDC_SECURE_LIB__ too. ........ r47224 | ronald.oussoren | 2006-07-04 14:30:22 +0200 (Tue, 04 Jul 2006) | 7 lines Sync the darwin/x86 port libffi with the copy in PyObjC. This fixes a number of bugs in that port. The most annoying ones were due to some subtle differences between the document ABI and the actual implementation :-( (there are no python unittests that fail without this patch, but without it some of libffi's unittests fail). ........ r47234 | georg.brandl | 2006-07-05 10:21:00 +0200 (Wed, 05 Jul 2006) | 3 lines Remove remaining references to OverflowWarning. ........ r47236 | thomas.heller | 2006-07-05 11:13:56 +0200 (Wed, 05 Jul 2006) | 3 lines Fix the bitfield test when _ctypes is compiled with MingW. Structures containing bitfields may have different layout on MSVC and MingW . ........ r47237 | thomas.wouters | 2006-07-05 13:03:49 +0200 (Wed, 05 Jul 2006) | 15 lines Fix bug in passing tuples to string.Template. All other values (with working str() or repr()) would work, just not multi-value tuples. Probably not a backport candidate, since it changes the behaviour of passing a single-element tuple: >>> string.Template("$foo").substitute(dict(foo=(1,))) '(1,)' versus '1' ........ r47241 | georg.brandl | 2006-07-05 16:18:45 +0200 (Wed, 05 Jul 2006) | 2 lines Patch #1517490: fix glitches in filter() docs. ........ r47244 | georg.brandl | 2006-07-05 17:50:05 +0200 (Wed, 05 Jul 2006) | 2 lines no need to elaborate "string". ........ r47251 | neal.norwitz | 2006-07-06 06:28:59 +0200 (Thu, 06 Jul 2006) | 3 lines Fix refleaks reported by Shane Hathaway in SF patch #1515361. This change contains only the changes related to leaking the copy variable. ........ r47253 | fred.drake | 2006-07-06 07:13:22 +0200 (Thu, 06 Jul 2006) | 4 lines - back out Expat change; the final fix to Expat will be different - change the pyexpat wrapper to not be so sensitive to this detail of the Expat implementation (the ex-crasher test still passes) ........ r47257 | neal.norwitz | 2006-07-06 08:45:08 +0200 (Thu, 06 Jul 2006) | 1 line Add a NEWS entry for a recent pyexpat fix ........ r47258 | martin.v.loewis | 2006-07-06 08:55:58 +0200 (Thu, 06 Jul 2006) | 2 lines Add sqlite3.dll to the DLLs component, not to the TkDLLs component. Fixes #1517388. ........ r47259 | martin.v.loewis | 2006-07-06 09:05:21 +0200 (Thu, 06 Jul 2006) | 1 line Properly quote compileall and Lib paths in case TARGETDIR has a space. ........ r47260 | thomas.heller | 2006-07-06 09:50:18 +0200 (Thu, 06 Jul 2006) | 5 lines Revert the change done in svn revision 47206: Add a new function uses_seh() to the _ctypes extension module. This will return True if Windows Structured Exception handling (SEH) is used when calling functions, False otherwise. ........ r47261 | armin.rigo | 2006-07-06 09:58:18 +0200 (Thu, 06 Jul 2006) | 3 lines A couple of examples about how to attack the fact that _PyType_Lookup() returns a borrowed ref. Many of the calls are open to attack. ........ r47262 | thomas.heller | 2006-07-06 10:28:14 +0200 (Thu, 06 Jul 2006) | 2 lines The test that calls a function with invalid arguments and catches the resulting Windows access violation will not be run by default. ........ r47263 | thomas.heller | 2006-07-06 10:48:35 +0200 (Thu, 06 Jul 2006) | 5 lines Patch #1517790: It is now possible to use custom objects in the ctypes foreign function argtypes sequence as long as they provide a from_param method, no longer is it required that the object is a ctypes type. ........ r47264 | thomas.heller | 2006-07-06 10:58:40 +0200 (Thu, 06 Jul 2006) | 2 lines Document the Struture and Union constructors. ........ r47265 | thomas.heller | 2006-07-06 11:11:22 +0200 (Thu, 06 Jul 2006) | 2 lines Document the changes in svn revision 47263, from patch #1517790. ........ r47267 | ronald.oussoren | 2006-07-06 12:13:35 +0200 (Thu, 06 Jul 2006) | 7 lines This patch solves the problem Skip was seeing with zlib, this patch ensures that configure uses similar compiler flags as setup.py when doing the zlib test. Without this patch configure would use the first shared library on the linker path, with this patch it uses the first shared or static library on that path just like setup.py. ........ r47268 | thomas.wouters | 2006-07-06 12:48:28 +0200 (Thu, 06 Jul 2006) | 4 lines NEWS entry for r47267: fixing configure's zlib probing. ........ r47269 | fredrik.lundh | 2006-07-06 14:29:24 +0200 (Thu, 06 Jul 2006) | 3 lines added XMLParser alias for cElementTree compatibility ........ r47271 | nick.coghlan | 2006-07-06 14:53:04 +0200 (Thu, 06 Jul 2006) | 1 line Revert the __module_name__ changes made in rev 47142. We'll revisit this in Python 2.6 ........ r47272 | nick.coghlan | 2006-07-06 15:04:56 +0200 (Thu, 06 Jul 2006) | 1 line Update the tutorial section on relative imports ........ r47273 | nick.coghlan | 2006-07-06 15:35:27 +0200 (Thu, 06 Jul 2006) | 1 line Ignore ImportWarning by default ........ r47274 | nick.coghlan | 2006-07-06 15:41:34 +0200 (Thu, 06 Jul 2006) | 1 line Cover ImportWarning, PendingDeprecationWarning and simplefilter() in the warnings module docs ........ r47275 | nick.coghlan | 2006-07-06 15:47:18 +0200 (Thu, 06 Jul 2006) | 1 line Add NEWS entries for the ImportWarning change and documentation update ........ r47276 | andrew.kuchling | 2006-07-06 15:57:28 +0200 (Thu, 06 Jul 2006) | 1 line ImportWarning is now silent by default ........ r47277 | thomas.heller | 2006-07-06 17:06:05 +0200 (Thu, 06 Jul 2006) | 2 lines Document the correct return type of PyLong_AsUnsignedLongLongMask. ........ r47278 | hyeshik.chang | 2006-07-06 17:21:52 +0200 (Thu, 06 Jul 2006) | 2 lines Add a testcase for r47086 which fixed a bug in codec_getstreamcodec(). ........ r47279 | hyeshik.chang | 2006-07-06 17:39:24 +0200 (Thu, 06 Jul 2006) | 3 lines Test using all CJK encodings for the testcases which don't require specific encodings. ........ r47280 | martin.v.loewis | 2006-07-06 21:28:03 +0200 (Thu, 06 Jul 2006) | 2 lines Properly generate logical file ids. Fixes #1515998. Also correct typo in Control.mapping. ........ r47287 | neal.norwitz | 2006-07-07 08:03:15 +0200 (Fri, 07 Jul 2006) | 17 lines Restore rev 47014: The hppa ubuntu box sometimes hangs forever in these tests. My guess is that the wait is failing for some reason. Use WNOHANG, so we won't wait until the buildbot kills the test suite. I haven't been able to reproduce the failure, so I'm not sure if this will help or not. Hopefully, this change will cause the test to fail, rather than hang. That will be better since we will get the rest of the test results. It may also help us debug the real problem. *** The reason this originally failed was because there were many zombie children outstanding before rev 47158 cleaned them up. There are still hangs in test_subprocess that need to be addressed, but that will take more work. This should close some holes. ........ r47289 | georg.brandl | 2006-07-07 10:15:12 +0200 (Fri, 07 Jul 2006) | 3 lines Fix RFC number. ........ r50489 | neal.norwitz | 2006-07-08 07:31:37 +0200 (Sat, 08 Jul 2006) | 1 line Fix SF bug #1519018: 'as' is now validated properly in import statements ........ r50490 | georg.brandl | 2006-07-08 14:15:27 +0200 (Sat, 08 Jul 2006) | 3 lines Add an additional test for bug #1519018. ........ r50491 | tim.peters | 2006-07-08 21:55:05 +0200 (Sat, 08 Jul 2006) | 2 lines Whitespace normalization. ........ r50493 | neil.schemenauer | 2006-07-09 18:16:34 +0200 (Sun, 09 Jul 2006) | 2 lines Fix AST compiler bug #1501934: incorrect LOAD/STORE_GLOBAL generation. ........ r50495 | neil.schemenauer | 2006-07-09 23:19:29 +0200 (Sun, 09 Jul 2006) | 2 lines Fix SF bug 1441486: bad unary minus folding in compiler. ........ r50497 | neal.norwitz | 2006-07-10 00:14:42 +0200 (Mon, 10 Jul 2006) | 4 lines On 64 bit systems, int literals that use less than 64 bits are now ints rather than longs. This also fixes the test for eval(-sys.maxint - 1). ........ r50500 | neal.norwitz | 2006-07-10 02:04:44 +0200 (Mon, 10 Jul 2006) | 4 lines Bug #1512814, Fix incorrect lineno's when code at module scope started after line 256. ........ r50501 | neal.norwitz | 2006-07-10 02:05:34 +0200 (Mon, 10 Jul 2006) | 1 line Fix doco. Backport candidate. ........ r50503 | neal.norwitz | 2006-07-10 02:23:17 +0200 (Mon, 10 Jul 2006) | 5 lines Part of SF patch #1484695. This removes dead code. The chksum was already verified in .frombuf() on the lines above. If there was a problem an exception is raised, so there was no way this condition could have been true. ........ r50504 | neal.norwitz | 2006-07-10 03:18:57 +0200 (Mon, 10 Jul 2006) | 3 lines Patch #1516912: improve Modules support for OpenVMS. ........ r50506 | neal.norwitz | 2006-07-10 04:36:41 +0200 (Mon, 10 Jul 2006) | 7 lines Patch #1504046: Add documentation for xml.etree. /F wrote the text docs, Englebert Gruber massaged it to latex and I did some more massaging to try and improve the consistency and fix some name mismatches between the declaration and text. ........ r50509 | martin.v.loewis | 2006-07-10 09:23:48 +0200 (Mon, 10 Jul 2006) | 2 lines Introduce DISTUTILS_USE_SDK as a flag to determine whether the SDK environment should be used. Fixes #1508010. ........ r50510 | martin.v.loewis | 2006-07-10 09:26:41 +0200 (Mon, 10 Jul 2006) | 1 line Change error message to indicate that VS2003 is necessary to build extension modules, not the .NET SDK. ........ r50511 | martin.v.loewis | 2006-07-10 09:29:41 +0200 (Mon, 10 Jul 2006) | 1 line Add svn:ignore. ........ r50512 | anthony.baxter | 2006-07-10 09:41:04 +0200 (Mon, 10 Jul 2006) | 1 line preparing for 2.5b2 ........ r50513 | thomas.heller | 2006-07-10 11:10:28 +0200 (Mon, 10 Jul 2006) | 2 lines Fix bug #1518190: accept any integer or long value in the ctypes.c_void_p constructor. ........ r50514 | thomas.heller | 2006-07-10 11:31:06 +0200 (Mon, 10 Jul 2006) | 3 lines Fixed a segfault when ctypes.wintypes were imported on non-Windows machines. ........ r50516 | thomas.heller | 2006-07-10 13:11:10 +0200 (Mon, 10 Jul 2006) | 3 lines Assigning None to pointer type structure fields possible overwrote wrong fields. ........ r50517 | thomas.heller | 2006-07-10 13:17:37 +0200 (Mon, 10 Jul 2006) | 5 lines Moved the ctypes news entries from the 'Library' section into the 'Extension Modules' section where they belong, probably. This destroyes the original order of the news entries, don't know if that is important or not. ........ r50526 | phillip.eby | 2006-07-10 21:03:29 +0200 (Mon, 10 Jul 2006) | 2 lines Fix SF#1516184 and add a test to prevent regression. ........ r50528 | phillip.eby | 2006-07-10 21:18:35 +0200 (Mon, 10 Jul 2006) | 2 lines Fix SF#1457312: bad socket error handling in distutils "upload" command. ........ r50537 | peter.astrand | 2006-07-10 22:39:49 +0200 (Mon, 10 Jul 2006) | 1 line Make it possible to run test_subprocess.py with Python 2.2, which lacks test_support.reap_children(). ........ r50541 | tim.peters | 2006-07-10 23:08:24 +0200 (Mon, 10 Jul 2006) | 5 lines After approval from Anthony, merge the tim-current_frames branch into the trunk. This adds a new sys._current_frames() function, which returns a dict mapping thread id to topmost thread stack frame. ........ r50542 | tim.peters | 2006-07-10 23:11:49 +0200 (Mon, 10 Jul 2006) | 2 lines Whitespace normalization. ........ r50553 | martin.v.loewis | 2006-07-11 00:11:28 +0200 (Tue, 11 Jul 2006) | 4 lines Patch #1519566: Remove unused _tofill member. Make begin_fill idempotent. Update demo2 to demonstrate filling of concave shapes. ........ r50567 | anthony.baxter | 2006-07-11 04:04:09 +0200 (Tue, 11 Jul 2006) | 4 lines #1494314: Fix a regression with high-numbered sockets in 2.4.3. This means that select() on sockets > FD_SETSIZE (typically 1024) work again. The patch makes sockets use poll() internally where available. ........ r50568 | tim.peters | 2006-07-11 04:17:48 +0200 (Tue, 11 Jul 2006) | 2 lines Whitespace normalization. ........ r50575 | thomas.heller | 2006-07-11 18:42:05 +0200 (Tue, 11 Jul 2006) | 1 line Add missing Py_DECREF. ........ r50576 | thomas.heller | 2006-07-11 18:44:25 +0200 (Tue, 11 Jul 2006) | 1 line Add missing Py_DECREFs. ........ r50579 | andrew.kuchling | 2006-07-11 19:20:16 +0200 (Tue, 11 Jul 2006) | 1 line Bump version number; add sys._current_frames ........ r50582 | thomas.heller | 2006-07-11 20:28:35 +0200 (Tue, 11 Jul 2006) | 3 lines When a foreign function is retrived by calling __getitem__ on a ctypes library instance, do not set it as attribute. ........ r50583 | thomas.heller | 2006-07-11 20:40:50 +0200 (Tue, 11 Jul 2006) | 2 lines Change the ctypes version number to 1.0.0. ........ r50597 | neal.norwitz | 2006-07-12 07:26:17 +0200 (Wed, 12 Jul 2006) | 3 lines Bug #1520864: unpacking singleton tuples in for loop (for x, in) work again. ........ r50598 | neal.norwitz | 2006-07-12 07:26:35 +0200 (Wed, 12 Jul 2006) | 1 line Fix function name in error msg ........ r50599 | neal.norwitz | 2006-07-12 07:27:46 +0200 (Wed, 12 Jul 2006) | 4 lines Fix uninitialized memory read reported by Valgrind when running doctest. This could happen if size == 0. ........ r50600 | neal.norwitz | 2006-07-12 09:28:29 +0200 (Wed, 12 Jul 2006) | 1 line Actually change the MAGIC #. Create a new section for 2.5c1 and mention the impact of changing the MAGIC #. ........ r50601 | thomas.heller | 2006-07-12 10:43:47 +0200 (Wed, 12 Jul 2006) | 3 lines Fix #1467450: ctypes now uses RTLD_GLOBAL by default on OSX 10.3 to load shared libraries. ........ r50604 | thomas.heller | 2006-07-12 16:25:18 +0200 (Wed, 12 Jul 2006) | 3 lines Fix the wrong description of LibraryLoader.LoadLibrary, and document the DEFAULT_MODE constant. ........ r50607 | georg.brandl | 2006-07-12 17:31:17 +0200 (Wed, 12 Jul 2006) | 3 lines Accept long options "--help" and "--version". ........ r50617 | thomas.heller | 2006-07-13 11:53:47 +0200 (Thu, 13 Jul 2006) | 3 lines A misspelled preprocessor symbol caused ctypes to be always compiled without thread support. Replaced WITH_THREADS with WITH_THREAD. ........ r50619 | thomas.heller | 2006-07-13 19:01:14 +0200 (Thu, 13 Jul 2006) | 3 lines Fix #1521375. When running with root priviledges, 'gcc -o /dev/null' did overwrite /dev/null. Use a temporary file instead of /dev/null. ........ r50620 | thomas.heller | 2006-07-13 19:05:13 +0200 (Thu, 13 Jul 2006) | 2 lines Fix misleading words. ........ r50622 | andrew.kuchling | 2006-07-13 19:37:26 +0200 (Thu, 13 Jul 2006) | 1 line Typo fix ........ r50629 | georg.brandl | 2006-07-14 09:12:54 +0200 (Fri, 14 Jul 2006) | 3 lines Patch #1521874: grammar errors in doanddont.tex. ........ r50630 | neal.norwitz | 2006-07-14 09:20:04 +0200 (Fri, 14 Jul 2006) | 1 line Try to improve grammar further. ........ r50631 | martin.v.loewis | 2006-07-14 11:58:55 +0200 (Fri, 14 Jul 2006) | 1 line Extend build_ssl to Win64, using VSExtComp. ........ r50632 | martin.v.loewis | 2006-07-14 14:10:09 +0200 (Fri, 14 Jul 2006) | 1 line Add debug output to analyse buildbot failure. ........ r50633 | martin.v.loewis | 2006-07-14 14:31:05 +0200 (Fri, 14 Jul 2006) | 1 line Fix Debug build of _ssl. ........ r50636 | andrew.kuchling | 2006-07-14 15:32:38 +0200 (Fri, 14 Jul 2006) | 1 line Mention new options ........ r50638 | peter.astrand | 2006-07-14 16:04:45 +0200 (Fri, 14 Jul 2006) | 1 line Bug #1223937: CalledProcessError.errno -> CalledProcessError.returncode. ........ r50640 | thomas.heller | 2006-07-14 17:01:05 +0200 (Fri, 14 Jul 2006) | 4 lines Make the prototypes of our private PyUnicode_FromWideChar and PyUnicode_AsWideChar replacement functions compatible to the official functions by using Py_ssize_t instead of int. ........ r50643 | thomas.heller | 2006-07-14 19:51:14 +0200 (Fri, 14 Jul 2006) | 3 lines Patch #1521817: The index range checking on ctypes arrays containing exactly one element is enabled again. ........ r50647 | thomas.heller | 2006-07-14 20:22:50 +0200 (Fri, 14 Jul 2006) | 2 lines Updates for the ctypes documentation. ........ r50655 | fredrik.lundh | 2006-07-14 23:45:48 +0200 (Fri, 14 Jul 2006) | 3 lines typo ........ r50664 | george.yoshida | 2006-07-15 18:03:49 +0200 (Sat, 15 Jul 2006) | 2 lines Bug #15187702 : ext/win-cookbook.html has a broken link to distutils ........ r50667 | bob.ippolito | 2006-07-15 18:53:15 +0200 (Sat, 15 Jul 2006) | 1 line Patch #1220874: Update the binhex module for Mach-O. ........ r50671 | fred.drake | 2006-07-16 03:21:20 +0200 (Sun, 16 Jul 2006) | 1 line clean up some link markup ........ r50673 | neal.norwitz | 2006-07-16 03:50:38 +0200 (Sun, 16 Jul 2006) | 4 lines Bug #1512814, Fix incorrect lineno's when code within a function had more than 255 blank lines. Byte codes need to go first, line #s second. ........ r50674 | neal.norwitz | 2006-07-16 04:00:32 +0200 (Sun, 16 Jul 2006) | 5 lines a & b were dereffed above, so they are known to be valid pointers. z is known to be NULL, nothing to DECREF. Reported by Klockwork, #107. ........ r50675 | neal.norwitz | 2006-07-16 04:02:57 +0200 (Sun, 16 Jul 2006) | 5 lines self is dereffed (and passed as first arg), so it's known to be good. func is returned from PyArg_ParseTuple and also dereffed. Reported by Klocwork, #30 (self one at least). ........ r50676 | neal.norwitz | 2006-07-16 04:05:35 +0200 (Sun, 16 Jul 2006) | 4 lines proto was dereffed above and is known to be good. No need for X. Reported by Klocwork, #39. ........ r50677 | neal.norwitz | 2006-07-16 04:15:27 +0200 (Sun, 16 Jul 2006) | 5 lines Fix memory leaks in some conditions. Reported by Klocwork #152. ........ r50678 | neal.norwitz | 2006-07-16 04:17:36 +0200 (Sun, 16 Jul 2006) | 4 lines Fix memory leak under some conditions. Reported by Klocwork, #98. ........ r50679 | neal.norwitz | 2006-07-16 04:22:30 +0200 (Sun, 16 Jul 2006) | 8 lines Use sizeof(buffer) instead of duplicating the constants to ensure they won't be wrong. The real change is to pass (bufsz - 1) to PyOS_ascii_formatd and 1 to strncat. strncat copies n+1 bytes from src (not dest). Reported by Klocwork #58. ........ r50680 | neal.norwitz | 2006-07-16 04:32:03 +0200 (Sun, 16 Jul 2006) | 5 lines Handle a NULL name properly. Reported by Klocwork #67 ........ r50681 | neal.norwitz | 2006-07-16 04:35:47 +0200 (Sun, 16 Jul 2006) | 6 lines PyFunction_SetDefaults() is documented as taking None or a tuple. A NULL would crash the PyTuple_Check(). Now make NULL return a SystemError. Reported by Klocwork #73. ........ r50683 | neal.norwitz | 2006-07-17 02:55:45 +0200 (Mon, 17 Jul 2006) | 5 lines Stop INCREFing name, then checking if it's NULL. name (f_name) should never be NULL so assert it. Fix one place where we could have passed NULL. Reported by Klocwork #66. ........ r50684 | neal.norwitz | 2006-07-17 02:57:15 +0200 (Mon, 17 Jul 2006) | 5 lines otherset is known to be non-NULL based on checks before and DECREF after. DECREF otherset rather than XDECREF in error conditions too. Reported by Klockwork #154. ........ r50685 | neal.norwitz | 2006-07-17 02:59:04 +0200 (Mon, 17 Jul 2006) | 7 lines Reported by Klocwork #151. v2 can be NULL if exception2 is NULL. I don't think that condition can happen, but I'm not sure it can't either. Now the code will protect against either being NULL. ........ r50686 | neal.norwitz | 2006-07-17 03:00:16 +0200 (Mon, 17 Jul 2006) | 1 line Add NEWS entry for a bunch of fixes due to warnings produced by Klocworks static analysis tool. ........ r50687 | fred.drake | 2006-07-17 07:47:52 +0200 (Mon, 17 Jul 2006) | 3 lines document xmlcore (still minimal; needs mention in each of the xml.* modules) SF bug #1504456 (partial) ........ r50688 | georg.brandl | 2006-07-17 15:23:46 +0200 (Mon, 17 Jul 2006) | 3 lines Remove usage of sets module (patch #1500609). ........ r50689 | georg.brandl | 2006-07-17 15:26:33 +0200 (Mon, 17 Jul 2006) | 3 lines Add missing NEWS item (#1522771) ........ r50690 | andrew.kuchling | 2006-07-17 18:47:54 +0200 (Mon, 17 Jul 2006) | 1 line Attribute more features ........ r50692 | kurt.kaiser | 2006-07-17 23:59:27 +0200 (Mon, 17 Jul 2006) | 8 lines Patch 1479219 - Tal Einat 1. 'as' highlighted as builtin in comment string on import line 2. Comments such as "#False identity" which start with a keyword immediately after the '#' character aren't colored as comments. 3. u or U beginning unicode string not correctly highlighted Closes bug 1325071 ........ r50693 | barry.warsaw | 2006-07-18 01:07:51 +0200 (Tue, 18 Jul 2006) | 16 lines decode_rfc2231(): Be more robust against buggy RFC 2231 encodings. Specifically, instead of raising a ValueError when there is a single tick in the parameter, simply return that the entire string unquoted, with None for both the charset and the language. Also, if there are more than 2 ticks in the parameter, interpret the first three parts as the standard RFC 2231 parts, then the rest of the parts as the encoded string. Test cases added. Original fewer-than-3-parts fix by Tokio Kikuchi. Resolves SF bug # 1218081. I will back port the fix and tests to Python 2.4 (email 3.0) and Python 2.3 (email 2.5). Also, bump the version number to email 4.0.1, removing the 'alpha' moniker. ........ r50695 | kurt.kaiser | 2006-07-18 06:03:16 +0200 (Tue, 18 Jul 2006) | 2 lines Rebinding Tab key was inserting 'tab' instead of 'Tab'. Bug 1179168. ........ r50696 | brett.cannon | 2006-07-18 06:41:36 +0200 (Tue, 18 Jul 2006) | 6 lines Fix bug #1520914. Starting in 2.4, time.strftime() began to check the bounds of values in the time tuple passed in. Unfortunately people came to rely on undocumented behaviour of setting unneeded values to 0, regardless of if it was within the valid range. Now those values force the value internally to the minimum value when 0 is passed in. ........ r50697 | facundo.batista | 2006-07-18 14:16:13 +0200 (Tue, 18 Jul 2006) | 1 line Comments and docs cleanups, and some little fixes, provided by Santiágo Peresón ........ r50704 | martin.v.loewis | 2006-07-18 19:46:31 +0200 (Tue, 18 Jul 2006) | 2 lines Patch #1524429: Use repr instead of backticks again. ........ r50706 | tim.peters | 2006-07-18 23:55:15 +0200 (Tue, 18 Jul 2006) | 2 lines Whitespace normalization. ........ r50708 | tim.peters | 2006-07-19 02:03:19 +0200 (Wed, 19 Jul 2006) | 18 lines SF bug 1524317: configure --without-threads fails to build Moved the code for _PyThread_CurrentFrames() up, so it's no longer in a huge "#ifdef WITH_THREAD" block (I didn't realize it /was/ in one). Changed test_sys's test_current_frames() so it passes with or without thread supported compiled in. Note that test_sys fails when Python is compiled without threads, but for an unrelated reason (the old test_exit() fails with an indirect ImportError on the `thread` module). There are also other unrelated compilation failures without threads, in extension modules (like ctypes); at least the core compiles again. Do we really support --without-threads? If so, there are several problems remaining. ........ r50713 | thomas.heller | 2006-07-19 11:09:32 +0200 (Wed, 19 Jul 2006) | 4 lines Make sure the _ctypes extension can be compiled when WITH_THREAD is not defined on Windows, even if that configuration is probably not supported at all. ........ r50715 | martin.v.loewis | 2006-07-19 19:18:32 +0200 (Wed, 19 Jul 2006) | 4 lines Revert r50706 (Whitespace normalization) and r50697: Comments and docs cleanups, and some little fixes per recommendation from Raymond Hettinger. ........ r50719 | phillip.eby | 2006-07-20 17:54:16 +0200 (Thu, 20 Jul 2006) | 4 lines Fix SF#1516184 (again) and add a test to prevent regression. (There was a problem with empty filenames still causing recursion) ........ r50720 | georg.brandl | 2006-07-20 18:28:39 +0200 (Thu, 20 Jul 2006) | 3 lines Guard for _active being None in __del__ method. ........ r50721 | vinay.sajip | 2006-07-20 18:28:39 +0200 (Thu, 20 Jul 2006) | 1 line Updated documentation for TimedRotatingFileHandler relating to how rollover files are named. The previous documentation was wrongly the same as for RotatingFileHandler. ........ r50731 | fred.drake | 2006-07-20 22:11:57 +0200 (Thu, 20 Jul 2006) | 1 line markup fix ........ r50739 | kurt.kaiser | 2006-07-21 00:22:52 +0200 (Fri, 21 Jul 2006) | 7 lines Avoid occasional failure to detect closing paren properly. Patch 1407280 Tal Einat M ParenMatch.py M NEWS.txt M CREDITS.txt ........ r50740 | vinay.sajip | 2006-07-21 01:20:12 +0200 (Fri, 21 Jul 2006) | 1 line Addressed SF#1524081 by using a dictionary to map level names to syslog priority names, rather than a string.lower(). ........ r50741 | neal.norwitz | 2006-07-21 07:29:58 +0200 (Fri, 21 Jul 2006) | 1 line Add some asserts that we got good params passed ........ r50742 | neal.norwitz | 2006-07-21 07:31:02 +0200 (Fri, 21 Jul 2006) | 5 lines Move the initialization of some pointers earlier. The problem is that if we call Py_DECREF(frame) like we do if allocating locals fails, frame_dealloc() will try to use these bogus values and crash. ........ r50743 | neal.norwitz | 2006-07-21 07:32:28 +0200 (Fri, 21 Jul 2006) | 4 lines Handle allocation failures gracefully. Found with failmalloc. Many (all?) of these could be backported. ........ r50745 | neal.norwitz | 2006-07-21 09:59:02 +0200 (Fri, 21 Jul 2006) | 1 line Speel initialise write. Tanks Anthony. ........ r50746 | neal.norwitz | 2006-07-21 09:59:47 +0200 (Fri, 21 Jul 2006) | 2 lines Handle more memory allocation failures without crashing. ........ r50754 | barry.warsaw | 2006-07-21 16:51:07 +0200 (Fri, 21 Jul 2006) | 23 lines More RFC 2231 improvements for the email 4.0 package. As Mark Sapiro rightly points out there are really two types of continued headers defined in this RFC (i.e. "encoded" parameters with the form "name*0*=" and unencoded parameters with the form "name*0="), but we were were handling them both the same way and that isn't correct. This patch should be much more RFC compliant in that only encoded params are %-decoded and the charset/language information is only extract if there are any encoded params in the segments. If there are no encoded params then the RFC says that there will be no charset/language parts. Note however that this will change the return value for Message.get_param() in some cases. For example, whereas before if you had all unencoded param continuations you would have still gotten a 3-tuple back from this method (with charset and language == None), you will now get just a string. I don't believe this is a backward incompatible change though because the documentation for this method already indicates that either return value is possible and that you must do an isinstance(val, tuple) check to discriminate between the two. (Yeah that API kind of sucks but we can't change /that/ without breaking code.) Test cases, some documentation updates, and a NEWS item accompany this patch. ........ r50759 | georg.brandl | 2006-07-21 19:36:31 +0200 (Fri, 21 Jul 2006) | 3 lines Fix check for empty list (vs. None). ........ r50771 | brett.cannon | 2006-07-22 00:44:07 +0200 (Sat, 22 Jul 2006) | 2 lines Remove an XXX marker in a comment. ........ r50773 | neal.norwitz | 2006-07-22 18:20:49 +0200 (Sat, 22 Jul 2006) | 1 line Fix more memory allocation issues found with failmalloc. ........ r50774 | neal.norwitz | 2006-07-22 19:00:57 +0200 (Sat, 22 Jul 2006) | 1 line Don't fail if the directory already exists ........ r50775 | greg.ward | 2006-07-23 04:25:53 +0200 (Sun, 23 Jul 2006) | 6 lines Be a lot smarter about whether this test passes: instead of assuming that a 2.93 sec audio file will always take 3.1 sec (as it did on the hardware I had when I first wrote the test), expect that it will take 2.93 sec +/- 10%, and only fail if it's outside of that range. Compute the expected ........ r50776 | kurt.kaiser | 2006-07-23 06:19:49 +0200 (Sun, 23 Jul 2006) | 2 lines Tooltips failed on new-syle class __init__ args. Bug 1027566 Loren Guthrie ........ r50777 | neal.norwitz | 2006-07-23 09:50:36 +0200 (Sun, 23 Jul 2006) | 1 line Handle more mem alloc issues found with failmalloc ........ r50778 | neal.norwitz | 2006-07-23 09:51:58 +0200 (Sun, 23 Jul 2006) | 5 lines If the for loop isn't entered, entryblock will be NULL. If passed to stackdepth_walk it will be dereffed. Not sure if I found with failmalloc or Klockwork #55. ........ r50779 | neal.norwitz | 2006-07-23 09:53:14 +0200 (Sun, 23 Jul 2006) | 4 lines Move the initialization of size_a down below the check for a being NULL. Reported by Klocwork #106 ........ r50780 | neal.norwitz | 2006-07-23 09:55:55 +0200 (Sun, 23 Jul 2006) | 9 lines Check the allocation of b_objects and return if there was a failure. Also fix a few memory leaks in other failure scenarios. It seems that if b_objects == Py_None, we will have an extra ref to b_objects. Add XXX comment so hopefully someone documents why the else isn't necessary or adds it in. Reported by Klocwork #20 ........ r50781 | neal.norwitz | 2006-07-23 09:57:11 +0200 (Sun, 23 Jul 2006) | 2 lines Fix memory leaks spotted by Klocwork #37. ........ r50782 | neal.norwitz | 2006-07-23 09:59:00 +0200 (Sun, 23 Jul 2006) | 5 lines nextlink can be NULL if teedataobject_new fails, so use XINCREF. Ensure that dataobj is never NULL. Reported by Klocwork #102 ........ r50783 | neal.norwitz | 2006-07-23 10:01:43 +0200 (Sun, 23 Jul 2006) | 8 lines Ensure we don't write beyond errText. I think I got this right, but it definitely could use some review to ensure I'm not off by one and there's no possible overflow/wrap-around of bytes_left. Reported by Klocwork #1. Fix a problem if there is a failure allocating self->db. Found with failmalloc. ........ r50784 | ronald.oussoren | 2006-07-23 11:41:09 +0200 (Sun, 23 Jul 2006) | 3 lines Without this patch CMD-W won't close EditorWindows on MacOS X. This solves part of bug #1517990. ........ r50785 | ronald.oussoren | 2006-07-23 11:46:11 +0200 (Sun, 23 Jul 2006) | 5 lines Fix for bug #1517996: Class and Path browsers show Tk menu This patch replaces the menubar that is used by AquaTk for windows without a menubar of their own by one that is more appropriate for IDLE. ........ r50786 | andrew.macintyre | 2006-07-23 14:57:02 +0200 (Sun, 23 Jul 2006) | 2 lines Build updates for OS/2 EMX port ........ r50787 | andrew.macintyre | 2006-07-23 15:00:04 +0200 (Sun, 23 Jul 2006) | 3 lines bugfix: PyThread_start_new_thread() returns the thread ID, not a flag; will backport. ........ r50789 | andrew.macintyre | 2006-07-23 15:04:00 +0200 (Sun, 23 Jul 2006) | 2 lines Get mailbox module working on OS/2 EMX port. ........ r50791 | greg.ward | 2006-07-23 18:05:51 +0200 (Sun, 23 Jul 2006) | 1 line Resync optparse with Optik 1.5.3: minor tweaks for/to tests. ........ r50794 | martin.v.loewis | 2006-07-24 07:05:22 +0200 (Mon, 24 Jul 2006) | 2 lines Update list of unsupported systems. Fixes #1510853. ........ r50795 | martin.v.loewis | 2006-07-24 12:26:33 +0200 (Mon, 24 Jul 2006) | 1 line Patch #1448199: Release GIL around ConnectRegistry. ........ r50796 | martin.v.loewis | 2006-07-24 13:54:53 +0200 (Mon, 24 Jul 2006) | 3 lines Patch #1232023: Don't include empty path component from registry, so that the current directory does not get added to sys.path. Also fixes #1526785. ........ r50797 | martin.v.loewis | 2006-07-24 14:54:17 +0200 (Mon, 24 Jul 2006) | 3 lines Bug #1524310: Properly report errors from FindNextFile in os.listdir. Will backport to 2.4. ........ r50800 | georg.brandl | 2006-07-24 15:28:57 +0200 (Mon, 24 Jul 2006) | 7 lines Patch #1523356: fix determining include dirs in python-config. Also don't install "python-config" when doing altinstall, but always install "python-config2.x" and make a link to it like with the main executable. ........ r50802 | georg.brandl | 2006-07-24 15:46:47 +0200 (Mon, 24 Jul 2006) | 3 lines Patch #1527744: right order of includes in order to have HAVE_CONIO_H defined properly. ........ r50803 | georg.brandl | 2006-07-24 16:09:56 +0200 (Mon, 24 Jul 2006) | 3 lines Patch #1515343: Fix printing of deprecated string exceptions with a value in the traceback module. ........ r50804 | kurt.kaiser | 2006-07-24 19:13:23 +0200 (Mon, 24 Jul 2006) | 7 lines EditorWindow failed when used stand-alone if sys.ps1 not set. Bug 1010370 Dave Florek M EditorWindow.py M PyShell.py M NEWS.txt ........ r50805 | kurt.kaiser | 2006-07-24 20:05:51 +0200 (Mon, 24 Jul 2006) | 6 lines - EditorWindow.test() was failing. Bug 1417598 M EditorWindow.py M ScriptBinding.py M NEWS.txt ........ r50808 | georg.brandl | 2006-07-24 22:11:35 +0200 (Mon, 24 Jul 2006) | 3 lines Repair accidental NameError. ........ r50809 | tim.peters | 2006-07-24 23:02:15 +0200 (Mon, 24 Jul 2006) | 2 lines Whitespace normalization. ........ r50810 | greg.ward | 2006-07-25 04:11:12 +0200 (Tue, 25 Jul 2006) | 3 lines Don't use standard assert: want tests to fail even when run with -O. Delete cruft. ........ r50811 | tim.peters | 2006-07-25 06:07:22 +0200 (Tue, 25 Jul 2006) | 10 lines current_frames_with_threads(): There's actually no way to guess /which/ line the spawned thread is in at the time sys._current_frames() is called: we know it finished enter_g.set(), but can't know whether the instruction counter has advanced to the following leave_g.wait(). The latter is overwhelming most likely, but not guaranteed, and I see that the "x86 Ubuntu dapper (icc) trunk" buildbot found it on the other line once. Changed the test so it passes in either case. ........ r50815 | martin.v.loewis | 2006-07-25 11:53:12 +0200 (Tue, 25 Jul 2006) | 2 lines Bug #1525817: Don't truncate short lines in IDLE's tool tips. ........ r50816 | martin.v.loewis | 2006-07-25 12:05:47 +0200 (Tue, 25 Jul 2006) | 3 lines Bug #978833: Really close underlying socket in _socketobject.close. Will backport to 2.4. ........ r50817 | martin.v.loewis | 2006-07-25 12:11:14 +0200 (Tue, 25 Jul 2006) | 1 line Revert incomplete checkin. ........ r50819 | georg.brandl | 2006-07-25 12:22:34 +0200 (Tue, 25 Jul 2006) | 4 lines Patch #1525766: correctly pass onerror arg to recursive calls of pkg.walk_packages. Also improve the docstrings. ........ r50825 | brett.cannon | 2006-07-25 19:32:20 +0200 (Tue, 25 Jul 2006) | 2 lines Add comment for changes to test_ossaudiodev. ........ r50826 | brett.cannon | 2006-07-25 19:34:36 +0200 (Tue, 25 Jul 2006) | 3 lines Fix a bug in the messages for an assert failure where not enough arguments to a string were being converted in the format. ........ r50828 | armin.rigo | 2006-07-25 20:09:57 +0200 (Tue, 25 Jul 2006) | 2 lines Document why is and is not a good way to fix the gc_inspection crasher. ........ r50829 | armin.rigo | 2006-07-25 20:11:07 +0200 (Tue, 25 Jul 2006) | 5 lines Added another crasher, which hit me today (I was not intentionally writing such code, of course, but it took some gdb time to figure out what my bug was). ........ r50830 | armin.rigo | 2006-07-25 20:38:39 +0200 (Tue, 25 Jul 2006) | 3 lines Document the crashers that will not go away soon as "won't fix", and explain why. ........ r50831 | ronald.oussoren | 2006-07-25 21:13:35 +0200 (Tue, 25 Jul 2006) | 3 lines Install the compatibility symlink to libpython.a on OSX using 'ln -sf' instead of 'ln -s', this avoid problems when reinstalling python. ........ r50832 | ronald.oussoren | 2006-07-25 21:20:54 +0200 (Tue, 25 Jul 2006) | 7 lines Fix for bug #1525447 (renaming to MacOSmodule.c would also work, but not without causing problems for anyone that is on a case-insensitive filesystem). Setup.py tries to compile the MacOS extension from MacOSmodule.c, while the actual file is named macosmodule.c. This is no problem on the (default) case-insensitive filesystem, but doesn't work on case-sensitive filesystems. ........ r50833 | ronald.oussoren | 2006-07-25 22:28:55 +0200 (Tue, 25 Jul 2006) | 7 lines Fix bug #1517990: IDLE keybindings on OSX This adds a new key definition for OSX, which is slightly different from the classic mac definition. Also add NEWS item for a couple of bugfixes I added recently. ........ r50834 | tim.peters | 2006-07-26 00:30:24 +0200 (Wed, 26 Jul 2006) | 2 lines Whitespace normalization. ........ r50839 | neal.norwitz | 2006-07-26 06:00:18 +0200 (Wed, 26 Jul 2006) | 1 line Hmm, only python2.x is installed, not plain python. Did that change recently? ........ r50840 | barry.warsaw | 2006-07-26 07:54:46 +0200 (Wed, 26 Jul 2006) | 6 lines Forward port some fixes that were in email 2.5 but for some reason didn't make it into email 4.0. Specifically, in Message.get_content_charset(), handle RFC 2231 headers that contain an encoding not known to Python, or a character in the data that isn't in the charset encoding. Also forward port the appropriate unit tests. ........ r50841 | georg.brandl | 2006-07-26 09:23:32 +0200 (Wed, 26 Jul 2006) | 3 lines NEWS entry for #1525766. ........ r50842 | georg.brandl | 2006-07-26 09:40:17 +0200 (Wed, 26 Jul 2006) | 3 lines Bug #1459963: properly capitalize HTTP header names. ........ r50843 | georg.brandl | 2006-07-26 10:03:10 +0200 (Wed, 26 Jul 2006) | 6 lines Part of bug #1523610: fix miscalculation of buffer length. Also add a guard against NULL in converttuple and add a test case (that previously would have crashed). ........ r50844 | martin.v.loewis | 2006-07-26 14:12:56 +0200 (Wed, 26 Jul 2006) | 3 lines Bug #978833: Really close underlying socket in _socketobject.close. Fix httplib.HTTPConnection.getresponse to not close the socket if it is still needed for the response. ........ r50845 | andrew.kuchling | 2006-07-26 19:16:52 +0200 (Wed, 26 Jul 2006) | 1 line [Bug #1471938] Fix build problem on Solaris 8 by conditionalizing the use of mvwgetnstr(); it was conditionalized a few lines below. Fix from Paul Eggert. I also tried out the STRICT_SYSV_CURSES case and am therefore removing the 'untested' comment. ........ r50846 | andrew.kuchling | 2006-07-26 19:18:01 +0200 (Wed, 26 Jul 2006) | 1 line Correct error message ........ r50847 | andrew.kuchling | 2006-07-26 19:19:39 +0200 (Wed, 26 Jul 2006) | 1 line Minor grammar fix ........ r50848 | andrew.kuchling | 2006-07-26 19:22:21 +0200 (Wed, 26 Jul 2006) | 1 line Put news item in right section ........ r50850 | andrew.kuchling | 2006-07-26 20:03:12 +0200 (Wed, 26 Jul 2006) | 1 line Use sys.exc_info() ........ r50851 | andrew.kuchling | 2006-07-26 20:15:45 +0200 (Wed, 26 Jul 2006) | 1 line Use sys.exc_info() ........ r50852 | phillip.eby | 2006-07-26 21:48:27 +0200 (Wed, 26 Jul 2006) | 4 lines Allow the 'onerror' argument to walk_packages() to catch any Exception, not just ImportError. This allows documentation tools to better skip unimportable packages. ........ r50854 | tim.peters | 2006-07-27 01:23:15 +0200 (Thu, 27 Jul 2006) | 2 lines Whitespace normalization. ........ r50855 | tim.peters | 2006-07-27 03:14:53 +0200 (Thu, 27 Jul 2006) | 21 lines Bug #1521947: possible bug in mystrtol.c with recent gcc. In general, C doesn't define anything about what happens when an operation on a signed integral type overflows, and PyOS_strtol() did several formally undefined things of that nature on signed longs. Some version of gcc apparently tries to exploit that now, and PyOS_strtol() could fail to detect overflow then. Tried to repair all that, although it seems at least as likely to me that we'll get screwed by bad platform definitions for LONG_MIN and/or LONG_MAX now. For that reason, I don't recommend backporting this. Note that I have no box on which this makes a lick of difference -- can't really test it, except to note that it didn't break anything on my boxes. Silent change: PyOS_strtol() used to return the hard-coded 0x7fffffff in case of overflow. Now it returns LONG_MAX. They're the same only on 32-bit boxes (although C doesn't guarantee that either ...). ........ r50856 | neal.norwitz | 2006-07-27 05:51:58 +0200 (Thu, 27 Jul 2006) | 6 lines Don't kill a normal instance of python running on windows when checking to kill a cygwin instance. build\\python.exe was matching a normal windows instance. Prefix that with a \\ to ensure build is a directory and not PCbuild. As discussed on python-dev. ........ r50857 | neal.norwitz | 2006-07-27 05:55:39 +0200 (Thu, 27 Jul 2006) | 5 lines Closure can't be NULL at this point since we know it's a tuple. Reported by Klocwork # 74. ........ r50858 | neal.norwitz | 2006-07-27 06:04:50 +0200 (Thu, 27 Jul 2006) | 1 line No functional change. Add comment and assert to describe why there cannot be overflow which was reported by Klocwork. Discussed on python-dev ........ r50859 | martin.v.loewis | 2006-07-27 08:38:16 +0200 (Thu, 27 Jul 2006) | 3 lines Bump distutils version to 2.5, as several new features have been introduced since 2.4. ........ r50860 | andrew.kuchling | 2006-07-27 14:18:20 +0200 (Thu, 27 Jul 2006) | 1 line Reformat docstring; fix typo ........ r50861 | georg.brandl | 2006-07-27 17:05:36 +0200 (Thu, 27 Jul 2006) | 6 lines Add test_main() methods. These three tests were never run by regrtest.py. We really need a simpler testing framework. ........ r50862 | tim.peters | 2006-07-27 17:09:20 +0200 (Thu, 27 Jul 2006) | 2 lines News for patch #1529686. ........ r50863 | tim.peters | 2006-07-27 17:11:00 +0200 (Thu, 27 Jul 2006) | 2 lines Whitespace normalization. ........ r50864 | georg.brandl | 2006-07-27 17:38:33 +0200 (Thu, 27 Jul 2006) | 3 lines Amend news entry. ........ r50865 | georg.brandl | 2006-07-27 18:08:15 +0200 (Thu, 27 Jul 2006) | 3 lines Make uuid test suite pass on this box by requesting output with LC_ALL=C. ........ r50866 | andrew.kuchling | 2006-07-27 20:37:33 +0200 (Thu, 27 Jul 2006) | 1 line Add example ........ r50867 | thomas.heller | 2006-07-27 20:39:55 +0200 (Thu, 27 Jul 2006) | 9 lines Remove code that is no longer used (ctypes.com). Fix the DllGetClassObject and DllCanUnloadNow so that they forward the call to the comtypes.server.inprocserver module. The latter was never documented, never used by published code, and didn't work anyway, so I think it does not deserve a NEWS entry (but I might be wrong). ........ r50868 | andrew.kuchling | 2006-07-27 20:41:21 +0200 (Thu, 27 Jul 2006) | 1 line Typo fix ('publically' is rare, poss. non-standard) ........ r50869 | andrew.kuchling | 2006-07-27 20:42:41 +0200 (Thu, 27 Jul 2006) | 1 line Add missing word ........ r50870 | andrew.kuchling | 2006-07-27 20:44:10 +0200 (Thu, 27 Jul 2006) | 1 line Repair typos ........ r50872 | andrew.kuchling | 2006-07-27 20:53:33 +0200 (Thu, 27 Jul 2006) | 1 line Update URL; add example ........ r50873 | andrew.kuchling | 2006-07-27 21:07:29 +0200 (Thu, 27 Jul 2006) | 1 line Add punctuation mark; add some examples ........ r50874 | andrew.kuchling | 2006-07-27 21:11:07 +0200 (Thu, 27 Jul 2006) | 1 line Mention base64 module; rewrite last sentence to be more positive ........ r50875 | andrew.kuchling | 2006-07-27 21:12:49 +0200 (Thu, 27 Jul 2006) | 1 line If binhex is higher-level than binascii, it should come first in the chapter ........ r50876 | tim.peters | 2006-07-27 22:47:24 +0200 (Thu, 27 Jul 2006) | 28 lines check_node(): stop spraying mystery output to stderr. When a node number disagrees, keep track of all sources & the node numbers they reported, and stick all that in the error message. Changed all callers to supply a non-empty "source" argument; made the "source" argument non-optional. On my box, test_uuid still fails, but with the less confusing output: AssertionError: different sources disagree on node: from source 'getnode1', node was 00038a000015 from source 'getnode2', node was 00038a000015 from source 'ipconfig', node was 001111b2b7bf Only the last one appears to be correct; e.g., C:\Code\python\PCbuild>getmac Physical Address Transport Name =================== ========================================================== 00-11-11-B2-B7-BF \Device\Tcpip_{190FB163-5AFD-4483-86A1-2FE16AC61FF1} 62-A1-AC-6C-FD-BE \Device\Tcpip_{8F77DF5A-EA3D-4F1D-975E-D472CEE6438A} E2-1F-01-C6-5D-88 \Device\Tcpip_{CD18F76B-2EF3-409F-9B8A-6481EE70A1E4} I can't find anything on my box with MAC 00-03-8a-00-00-15, and am not clear on where that comes from. ........ r50878 | andrew.kuchling | 2006-07-28 00:40:05 +0200 (Fri, 28 Jul 2006) | 1 line Reword paragraph ........ r50879 | andrew.kuchling | 2006-07-28 00:49:38 +0200 (Fri, 28 Jul 2006) | 1 line Add example ........ r50880 | andrew.kuchling | 2006-07-28 00:49:54 +0200 (Fri, 28 Jul 2006) | 1 line Add example ........ r50881 | barry.warsaw | 2006-07-28 01:43:15 +0200 (Fri, 28 Jul 2006) | 27 lines Patch #1520294: Support for getset and member descriptors in types.py, inspect.py, and pydoc.py. Specifically, this allows for querying the type of an object against these built-in C types and more importantly, for getting their docstrings printed in the interactive interpreter's help() function. This patch includes a new built-in module called _types which provides definitions of getset and member descriptors for use by the types.py module. These types are exposed as types.GetSetDescriptorType and types.MemberDescriptorType. Query functions are provided as inspect.isgetsetdescriptor() and inspect.ismemberdescriptor(). The implementations of these are robust enough to work with Python implementations other than CPython, which may not have these fundamental types. The patch also includes documentation and test suite updates. I commit these changes now under these guiding principles: 1. Silence is assent. The release manager has not said "no", and of the few people that cared enough to respond to the thread, the worst vote was "0". 2. It's easier to ask for forgiveness than permission. 3. It's so dang easy to revert stuff in svn, that you could view this as a forcing function. :) Windows build patches will follow. ........ r50882 | tim.peters | 2006-07-28 01:44:37 +0200 (Fri, 28 Jul 2006) | 4 lines Bug #1529297: The rewrite of doctest for Python 2.4 unintentionally lost that tests are sorted by name before being run. ``DocTestFinder`` has been changed to sort the list of tests it returns. ........ r50883 | tim.peters | 2006-07-28 01:45:48 +0200 (Fri, 28 Jul 2006) | 2 lines Whitespace normalization. ........ r50884 | tim.peters | 2006-07-28 01:46:36 +0200 (Fri, 28 Jul 2006) | 2 lines Add missing svn:eol-style property to text files. ........ r50885 | barry.warsaw | 2006-07-28 01:50:40 +0200 (Fri, 28 Jul 2006) | 4 lines Enable the building of the _types module on Windows. Note that this has only been tested for VS 2003 since that's all I have. ........ r50887 | tim.peters | 2006-07-28 02:23:15 +0200 (Fri, 28 Jul 2006) | 7 lines defdict_reduce(): Plug leaks. We didn't notice these before because test_defaultdict didn't actually do anything before Georg fixed that earlier today. Neal's next refleak run then showed test_defaultdict leaking 9 references on each run. That's repaired by this checkin. ........ r50888 | tim.peters | 2006-07-28 02:30:00 +0200 (Fri, 28 Jul 2006) | 2 lines News about the repaired memory leak in defaultdict. ........ r50889 | gregory.p.smith | 2006-07-28 03:35:25 +0200 (Fri, 28 Jul 2006) | 7 lines - pybsddb Bug #1527939: bsddb module DBEnv dbremove and dbrename methods now allow their database parameter to be None as the sleepycat API allows. Also adds an appropriate test case for DBEnv.dbrename and dbremove. ........ r50895 | neal.norwitz | 2006-07-28 06:22:34 +0200 (Fri, 28 Jul 2006) | 1 line Ensure the actual number matches the expected count ........ r50896 | tim.peters | 2006-07-28 06:51:59 +0200 (Fri, 28 Jul 2006) | 6 lines Live with that "the hardware address" is an ill-defined concept, and that different ways of trying to find "the hardware address" may return different results. Certainly true on both of my Windows boxes, and in different ways (see whining on python-dev). ........ r50897 | neal.norwitz | 2006-07-28 09:21:27 +0200 (Fri, 28 Jul 2006) | 3 lines Try to find the MAC addr on various flavours of Unix. This seems hopeless. The reduces the test_uuid failures, but there's still another method failing. ........ r50898 | martin.v.loewis | 2006-07-28 09:45:49 +0200 (Fri, 28 Jul 2006) | 2 lines Add UUID for upcoming 2.5b3. ........ r50899 | matt.fleming | 2006-07-28 13:27:27 +0200 (Fri, 28 Jul 2006) | 3 lines Allow socketmodule to compile on NetBSD -current, whose bluetooth API differs from both Linux and FreeBSD. Accepted by Neal Norwitz. ........ r50900 | andrew.kuchling | 2006-07-28 14:07:12 +0200 (Fri, 28 Jul 2006) | 1 line [Patch #1529811] Correction to description of r|* mode ........ r50901 | andrew.kuchling | 2006-07-28 14:18:22 +0200 (Fri, 28 Jul 2006) | 1 line Typo fix ........ r50902 | andrew.kuchling | 2006-07-28 14:32:43 +0200 (Fri, 28 Jul 2006) | 1 line Add example ........ r50903 | andrew.kuchling | 2006-07-28 14:33:19 +0200 (Fri, 28 Jul 2006) | 1 line Add example ........ r50904 | andrew.kuchling | 2006-07-28 14:45:55 +0200 (Fri, 28 Jul 2006) | 1 line Don't overwrite built-in name; add some blank lines for readability ........ r50905 | andrew.kuchling | 2006-07-28 14:48:07 +0200 (Fri, 28 Jul 2006) | 1 line Add example. Should I propagate this example to all the other DBM-ish modules, too? ........ r50912 | georg.brandl | 2006-07-28 20:31:39 +0200 (Fri, 28 Jul 2006) | 3 lines Patch #1529686: also run test_email_codecs with regrtest.py. ........ r50913 | georg.brandl | 2006-07-28 20:36:01 +0200 (Fri, 28 Jul 2006) | 3 lines Fix spelling. ........ r50915 | thomas.heller | 2006-07-28 21:42:40 +0200 (Fri, 28 Jul 2006) | 3 lines Remove a useless XXX comment. Cosmetic changes to the code so that the #ifdef _UNICODE block doesn't mess emacs code formatting. ........ r50916 | phillip.eby | 2006-07-28 23:12:07 +0200 (Fri, 28 Jul 2006) | 5 lines Bug #1529871: The speed enhancement patch #921466 broke Python's compliance with PEP 302. This was fixed by adding an ``imp.NullImporter`` type that is used in ``sys.path_importer_cache`` to cache non-directory paths and avoid excessive filesystem operations during imports. ........ r50917 | phillip.eby | 2006-07-28 23:31:54 +0200 (Fri, 28 Jul 2006) | 2 lines Fix svn merge spew. ........ r50918 | thomas.heller | 2006-07-28 23:43:20 +0200 (Fri, 28 Jul 2006) | 4 lines Patch #1529514: More openbsd platforms for ctypes. Regenerated Modules/_ctypes/libffi/configure with autoconf 2.59. Approved by Neal. ........ r50922 | georg.brandl | 2006-07-29 10:51:21 +0200 (Sat, 29 Jul 2006) | 2 lines Bug #835255: The "closure" argument to new.function() is now documented. ........ r50924 | georg.brandl | 2006-07-29 11:33:26 +0200 (Sat, 29 Jul 2006) | 3 lines Bug #1441397: The compiler module now recognizes module and function docstrings correctly as it did in Python 2.4. ........ r50925 | georg.brandl | 2006-07-29 12:25:46 +0200 (Sat, 29 Jul 2006) | 4 lines Revert rev 42617, it was introduced to work around bug #1441397. test_compiler now passes again. ........ r50926 | fred.drake | 2006-07-29 15:22:49 +0200 (Sat, 29 Jul 2006) | 1 line update target version number ........ r50927 | andrew.kuchling | 2006-07-29 15:56:48 +0200 (Sat, 29 Jul 2006) | 1 line Add example ........ r50928 | andrew.kuchling | 2006-07-29 16:04:47 +0200 (Sat, 29 Jul 2006) | 1 line Update URL ........ r50930 | andrew.kuchling | 2006-07-29 16:08:15 +0200 (Sat, 29 Jul 2006) | 1 line Reword paragraph to match the order of the subsequent sections ........ r50931 | andrew.kuchling | 2006-07-29 16:21:15 +0200 (Sat, 29 Jul 2006) | 1 line [Bug #1529157] Mention raw_input() and input(); while I'm at it, reword the description a bit ........ r50932 | andrew.kuchling | 2006-07-29 16:42:48 +0200 (Sat, 29 Jul 2006) | 1 line [Bug #1519571] Document some missing functions: setup(), title(), done() ........ r50933 | andrew.kuchling | 2006-07-29 16:43:55 +0200 (Sat, 29 Jul 2006) | 1 line Fix docstring punctuation ........ r50934 | andrew.kuchling | 2006-07-29 17:10:32 +0200 (Sat, 29 Jul 2006) | 1 line [Bug #1414697] Change docstring of set/frozenset types to specify that the contents are unique. Raymond, please feel free to edit or revert. ........ r50935 | andrew.kuchling | 2006-07-29 17:35:21 +0200 (Sat, 29 Jul 2006) | 1 line [Bug #1530382] Document SSL.server(), .issuer() methods ........ r50936 | andrew.kuchling | 2006-07-29 17:42:46 +0200 (Sat, 29 Jul 2006) | 1 line Typo fix ........ r50937 | andrew.kuchling | 2006-07-29 17:43:13 +0200 (Sat, 29 Jul 2006) | 1 line Tweak wording ........ r50938 | matt.fleming | 2006-07-29 17:55:30 +0200 (Sat, 29 Jul 2006) | 2 lines Fix typo ........ r50939 | andrew.kuchling | 2006-07-29 17:57:08 +0200 (Sat, 29 Jul 2006) | 6 lines [Bug #1528258] Mention that the 'data' argument can be None. The constructor docs referred the reader to the add_data() method's docs, but they weren't very helpful. I've simply copied an earlier explanation of 'data' that's more useful. ........ r50940 | andrew.kuchling | 2006-07-29 18:08:40 +0200 (Sat, 29 Jul 2006) | 1 line Set bug/patch count. Take a bow, everyone! ........ r50941 | fred.drake | 2006-07-29 18:56:15 +0200 (Sat, 29 Jul 2006) | 18 lines expunge the xmlcore changes: 41667, 41668 - initial switch to xmlcore 47044 - mention of xmlcore in What's New 50687 - mention of xmlcore in the library reference re-apply xmlcore changes to xml: 41674 - line ending changes (re-applied manually), directory props 41677 - add cElementTree wrapper 41678 - PSF licensing for etree 41812 - whitespace normalization 42724 - fix svn:eol-style settings 43681, 43682 - remove Python version-compatibility cruft from minidom 46773 - fix encoding of \r\n\t in attr values in saxutils 47269 - added XMLParser alias for cElementTree compatibility additional tests were added in Lib/test/test_sax.py that failed with the xmlcore changes; these relate to SF bugs #1511497, #1513611 ........ r50942 | andrew.kuchling | 2006-07-29 20:14:07 +0200 (Sat, 29 Jul 2006) | 17 lines Reorganize the docs for 'file' and 'open()' after some discussion with Fred. We want to encourage users to write open() when opening a file, but open() was described with a single paragraph and 'file' had lots of explanation of the mode and bufsize arguments. I've shrunk the description of 'file' to cross-reference to the 'File objects' section, and to open() for an explanation of the arguments. open() now has all the paragraphs about the mode string. The bufsize argument was moved up so that it isn't buried at the end; now there's 1 paragraph on mode, 1 on bufsize, and then 3 more on mode. Various other edits and rearrangements were made in the process. It's probably best to read the final text and not to try to make sense of the diffs. ........ r50943 | fred.drake | 2006-07-29 20:19:19 +0200 (Sat, 29 Jul 2006) | 1 line restore test un-intentionally removed in the xmlcore purge (revision 50941) ........ r50944 | fred.drake | 2006-07-29 20:33:29 +0200 (Sat, 29 Jul 2006) | 3 lines make the reference to older versions of the documentation a link to the right page on python.org ........ r50945 | fred.drake | 2006-07-29 21:09:01 +0200 (Sat, 29 Jul 2006) | 1 line document the footnote usage pattern ........ r50947 | fred.drake | 2006-07-29 21:14:10 +0200 (Sat, 29 Jul 2006) | 1 line emphasize and oddball nuance of LaTeX comment syntax ........ r50948 | andrew.kuchling | 2006-07-29 21:24:04 +0200 (Sat, 29 Jul 2006) | 1 line [Patch #1490989 from Skip Montanaro] Mention debugging builds in the API documentation. I've changed Skip's patch to point to Misc/SpecialBuilds and fiddled with the markup a bit. ........ r50949 | neal.norwitz | 2006-07-29 21:29:35 +0200 (Sat, 29 Jul 2006) | 6 lines Disable these tests until they are reliable across platforms. These problems may mask more important, real problems. One or both methods are known to fail on: Solaris, OpenBSD, Debian, Ubuntu. They pass on Windows and some Linux boxes. ........ r50950 | andrew.kuchling | 2006-07-29 21:50:37 +0200 (Sat, 29 Jul 2006) | 1 line [Patch #1068277] Clarify that os.path.exists() can return False depending on permissions. Fred approved committing this patch in December 2004! ........ r50952 | fred.drake | 2006-07-29 22:04:42 +0200 (Sat, 29 Jul 2006) | 6 lines SF bug #1193966: Weakref types documentation misplaced The information about supporting weakrefs with types defined in C extensions is moved to the Extending & Embedding manual. Py_TPFLAGS_HAVE_WEAKREFS is no longer mentioned since it is part of Py_TPFLAGS_DEFAULT. ........ r50953 | skip.montanaro | 2006-07-29 22:06:05 +0200 (Sat, 29 Jul 2006) | 4 lines Add a comment to the csv reader documentation that explains why the treatment of newlines changed in 2.5. Pulled almost verbatim from a comment by Andrew McNamara in <http://python.org/sf/1465014>. ........ r50954 | neal.norwitz | 2006-07-29 22:20:52 +0200 (Sat, 29 Jul 2006) | 3 lines If the executable doesn't exist, there's no reason to try to start it. This prevents garbage about command not found being printed on Solaris. ........ r50955 | fred.drake | 2006-07-29 22:21:25 +0200 (Sat, 29 Jul 2006) | 1 line fix minor markup error that introduced extra punctuation ........ r50957 | neal.norwitz | 2006-07-29 22:37:08 +0200 (Sat, 29 Jul 2006) | 3 lines Disable test_getnode too, since this is also unreliable. ........ r50958 | andrew.kuchling | 2006-07-29 23:27:12 +0200 (Sat, 29 Jul 2006) | 1 line Follow TeX's conventions for hyphens ........ r50959 | andrew.kuchling | 2006-07-29 23:30:21 +0200 (Sat, 29 Jul 2006) | 1 line Fix case for 'Unix' ........ r50960 | fred.drake | 2006-07-30 01:34:57 +0200 (Sun, 30 Jul 2006) | 1 line markup cleanups ........ r50961 | andrew.kuchling | 2006-07-30 02:27:34 +0200 (Sun, 30 Jul 2006) | 1 line Minor typo fixes ........ r50962 | andrew.kuchling | 2006-07-30 02:37:56 +0200 (Sun, 30 Jul 2006) | 1 line [Bug #793553] Correct description of keyword arguments for SSL authentication ........ r50963 | tim.peters | 2006-07-30 02:58:15 +0200 (Sun, 30 Jul 2006) | 2 lines Whitespace normalization. ........ r50964 | fred.drake | 2006-07-30 05:03:43 +0200 (Sun, 30 Jul 2006) | 1 line lots of markup nits, most commonly Unix/unix --> \UNIX ........ r50965 | fred.drake | 2006-07-30 07:41:28 +0200 (Sun, 30 Jul 2006) | 1 line update information on wxPython, from Robin Dunn ........ r50966 | fred.drake | 2006-07-30 07:49:49 +0200 (Sun, 30 Jul 2006) | 4 lines remove possibly-outdated comment on what GUI toolkit is most commonly used; it is hard to know whether this is right, and it does not add valuable reference information at any rate ........ r50967 | fred.drake | 2006-07-30 07:55:39 +0200 (Sun, 30 Jul 2006) | 3 lines - remove yet another reference to how commonly Tkinter is (thought to be) used - fix an internal section reference ........ r50968 | neal.norwitz | 2006-07-30 08:53:31 +0200 (Sun, 30 Jul 2006) | 4 lines Patch #1531113: Fix augmented assignment with yield expressions. Also fix a SystemError when trying to assign to yield expressions. ........ r50969 | neal.norwitz | 2006-07-30 08:55:48 +0200 (Sun, 30 Jul 2006) | 5 lines Add PyErr_WarnEx() so C code can pass the stacklevel to warnings.warn(). This provides the proper warning for struct.pack(). PyErr_Warn() is now deprecated in favor of PyErr_WarnEx(). As mentioned by Tim Peters on python-dev. ........ r50970 | neal.norwitz | 2006-07-30 08:57:04 +0200 (Sun, 30 Jul 2006) | 3 lines Bug #1515471: string.replace() accepts character buffers again. Pass the char* and size around rather than PyObject's. ........ r50971 | neal.norwitz | 2006-07-30 08:59:13 +0200 (Sun, 30 Jul 2006) | 1 line Whitespace normalization ........ r50973 | georg.brandl | 2006-07-30 12:53:32 +0200 (Sun, 30 Jul 2006) | 3 lines Clarify that __op__ methods must return NotImplemented if they don't support the operation. ........ r50974 | georg.brandl | 2006-07-30 13:07:23 +0200 (Sun, 30 Jul 2006) | 3 lines Bug #1002398: The documentation for os.path.sameopenfile now correctly refers to file descriptors, not file objects. ........ r50977 | martin.v.loewis | 2006-07-30 15:00:31 +0200 (Sun, 30 Jul 2006) | 3 lines Don't copy directory stat times in shutil.copytree on Windows Fixes #1525866. ........ r50978 | martin.v.loewis | 2006-07-30 15:14:05 +0200 (Sun, 30 Jul 2006) | 3 lines Base __version__ on sys.version_info, as distutils is no longer maintained separatedly. ........ r50979 | martin.v.loewis | 2006-07-30 15:27:31 +0200 (Sun, 30 Jul 2006) | 3 lines Mention Cygwin in distutils error message about a missing VS 2003. Fixes #1257728. ........ r50982 | martin.v.loewis | 2006-07-30 16:09:47 +0200 (Sun, 30 Jul 2006) | 5 lines Drop usage of test -e in configure as it is not portable. Fixes #1439538 Will backport to 2.4 Also regenerate pyconfig.h.in. ........ r50984 | georg.brandl | 2006-07-30 18:20:10 +0200 (Sun, 30 Jul 2006) | 3 lines Fix makefile changes for python-config. ........ r50985 | george.yoshida | 2006-07-30 18:37:37 +0200 (Sun, 30 Jul 2006) | 2 lines Rename struct.pack_to to struct.pack_into as changed in revision 46642. ........ r50986 | george.yoshida | 2006-07-30 18:41:30 +0200 (Sun, 30 Jul 2006) | 2 lines Typo fix ........ r50987 | neal.norwitz | 2006-07-30 21:18:13 +0200 (Sun, 30 Jul 2006) | 1 line Add some asserts and update comments ........ r50988 | neal.norwitz | 2006-07-30 21:18:38 +0200 (Sun, 30 Jul 2006) | 1 line Verify that the signal handlers were really called ........ r50989 | neal.norwitz | 2006-07-30 21:20:42 +0200 (Sun, 30 Jul 2006) | 3 lines Try to prevent hangs on Tru64/Alpha buildbot. I'm not certain this will help and may need to be reverted if it causes problems. ........ r50990 | georg.brandl | 2006-07-30 22:18:51 +0200 (Sun, 30 Jul 2006) | 2 lines Bug #1531349: right <-> left glitch in __rop__ description. ........ r50992 | tim.peters | 2006-07-31 03:46:03 +0200 (Mon, 31 Jul 2006) | 2 lines Whitespace normalization. ........ r50993 | andrew.mcnamara | 2006-07-31 04:27:48 +0200 (Mon, 31 Jul 2006) | 2 lines Redo the comment about the 2.5 change in quoted-newline handling. ........ r50994 | tim.peters | 2006-07-31 04:40:23 +0200 (Mon, 31 Jul 2006) | 10 lines ZipFile.close(): Killed one of the struct.pack deprecation warnings on Win32. Also added an XXX about the line: pos3 = self.fp.tell() `pos3` is never referenced, and I have no idea what the code intended to do instead. ........ r50996 | tim.peters | 2006-07-31 04:53:03 +0200 (Mon, 31 Jul 2006) | 8 lines ZipFile.close(): Kill the other struct.pack deprecation warning on Windows. Afraid I can't detect a pattern to when the pack formats decide to use a signed or unsigned format code -- appears nearly arbitrary to my eyes. So I left all the pack formats alone and changed the special-case data values instead. ........ r50997 | skip.montanaro | 2006-07-31 05:09:45 +0200 (Mon, 31 Jul 2006) | 1 line minor tweaks ........ r50998 | skip.montanaro | 2006-07-31 05:11:11 +0200 (Mon, 31 Jul 2006) | 1 line minor tweaks ........ r50999 | andrew.kuchling | 2006-07-31 14:20:24 +0200 (Mon, 31 Jul 2006) | 1 line Add refcounts for PyErr_WarnEx ........ r51000 | andrew.kuchling | 2006-07-31 14:39:05 +0200 (Mon, 31 Jul 2006) | 9 lines Document PyErr_WarnEx. (Bad Neal! No biscuit!) Is the explanation of the 'stacklevel' parameter clear? Please feel free to edit it. I don't have LaTeX installed on this machine, so haven't verified that the markup is correct. Will check tonight, or maybe the automatic doc build will tell me. ........ r51001 | andrew.kuchling | 2006-07-31 14:52:26 +0200 (Mon, 31 Jul 2006) | 1 line Add PyErr_WarnEx() ........ r51002 | andrew.kuchling | 2006-07-31 15:18:27 +0200 (Mon, 31 Jul 2006) | 1 line Mention csv newline changes ........ r51003 | andrew.kuchling | 2006-07-31 17:22:58 +0200 (Mon, 31 Jul 2006) | 1 line Typo fix ........ r51004 | andrew.kuchling | 2006-07-31 17:23:43 +0200 (Mon, 31 Jul 2006) | 1 line Remove reference to notation ........ r51005 | georg.brandl | 2006-07-31 18:00:34 +0200 (Mon, 31 Jul 2006) | 3 lines Fix function name. ........ r51006 | andrew.kuchling | 2006-07-31 18:10:24 +0200 (Mon, 31 Jul 2006) | 1 line [Bug #1514540] Instead of putting the standard types in a section, put them in a chapter of their own. This means string methods will now show up in the ToC. (Should the types come before or after the functions+exceptions+constants chapter? I've put them after, for now.) ........ r51007 | andrew.kuchling | 2006-07-31 18:22:05 +0200 (Mon, 31 Jul 2006) | 1 line [Bug #848556] Remove \d* from second alternative to avoid exponential case when repeating match ........ r51008 | andrew.kuchling | 2006-07-31 18:27:57 +0200 (Mon, 31 Jul 2006) | 1 line Update list of files; fix a typo ........ r51013 | andrew.kuchling | 2006-08-01 18:24:30 +0200 (Tue, 01 Aug 2006) | 1 line typo fix ........ r51018 | thomas.heller | 2006-08-01 18:54:43 +0200 (Tue, 01 Aug 2006) | 2 lines Fix a potential segfault and various potentail refcount leaks in the cast() function. ........ r51020 | thomas.heller | 2006-08-01 19:46:10 +0200 (Tue, 01 Aug 2006) | 1 line Minimal useful docstring for CopyComPointer. ........ r51021 | andrew.kuchling | 2006-08-01 20:16:15 +0200 (Tue, 01 Aug 2006) | 8 lines [Patch #1520905] Attempt to suppress core file created by test_subprocess.py. Patch by Douglas Greiman. The test_run_abort() testcase produces a core file on Unix systems, even though the test is successful. This can be confusing or alarming to someone who runs 'make test' and then finds that the Python interpreter apparently crashed. ........ r51023 | georg.brandl | 2006-08-01 20:49:24 +0200 (Tue, 01 Aug 2006) | 3 lines os.urandom no longer masks unrelated exceptions like SystemExit or KeyboardInterrupt. ........ r51025 | thomas.heller | 2006-08-01 21:14:15 +0200 (Tue, 01 Aug 2006) | 2 lines Speed up PyType_stgdict and PyObject_stgdict. ........ r51027 | ronald.oussoren | 2006-08-01 22:30:31 +0200 (Tue, 01 Aug 2006) | 3 lines Make sure the postinstall action that optionally updates the user's profile on MacOS X actually works correctly in all cases. ........ r51028 | ronald.oussoren | 2006-08-01 23:00:57 +0200 (Tue, 01 Aug 2006) | 4 lines This fixes bug #1527397: PythonLauncher runs scripts with the wrong working directory. It also fixes a bug where PythonLauncher failed to launch scripts when the scriptname (or the path to the script) contains quotes. ........ r51031 | tim.peters | 2006-08-02 05:27:46 +0200 (Wed, 02 Aug 2006) | 2 lines Whitespace normalization. ........ r51032 | tim.peters | 2006-08-02 06:12:36 +0200 (Wed, 02 Aug 2006) | 19 lines Try to squash struct.pack warnings on the "amd64 gentoo trunk" buildbot (& possibly other 64-bit boxes) during test_gzip. The native zlib crc32 function returns an unsigned 32-bit integer, which the Python wrapper implicitly casts to C long. Therefore the same crc can "look negative" on a 32-bit box but "look positive" on a 64-bit box. This patch papers over that platform difference when writing the crc to file. It may be better to change the Python wrapper, either to make the result "look positive" on all platforms (which means it may have to return a Python long at times on a 32-bit box), or to keep the sign the same across boxes. But that would be a visible change in what users see, while the current hack changes no visible behavior (well, apart from stopping the struct deprecation warning). Note that the module-level write32() function is no longer used. ........ r51033 | neal.norwitz | 2006-08-02 06:27:11 +0200 (Wed, 02 Aug 2006) | 4 lines Prevent memory leak on error. Reported by Klocwork #36 ........ r51034 | tim.peters | 2006-08-02 07:20:08 +0200 (Wed, 02 Aug 2006) | 9 lines _Stream.close(): Try to kill struct.pack() warnings when writing the crc to file on the "PPC64 Debian trunk" buildbot when running test_tarfile. This is again a case where the native zlib crc is an unsigned 32-bit int, but the Python wrapper implicitly casts it to signed C long, so that "the sign bit looks different" on different platforms. ........ r51035 | ronald.oussoren | 2006-08-02 08:10:10 +0200 (Wed, 02 Aug 2006) | 2 lines Updated documentation for the script that builds the OSX installer. ........ r51036 | neal.norwitz | 2006-08-02 08:14:22 +0200 (Wed, 02 Aug 2006) | 2 lines _PyWeakref_GetWeakrefCount() now returns a Py_ssize_t instead of long. ........ r51037 | neal.norwitz | 2006-08-02 08:15:10 +0200 (Wed, 02 Aug 2006) | 1 line v is already checked for NULL, so just DECREF it ........ r51038 | neal.norwitz | 2006-08-02 08:19:19 +0200 (Wed, 02 Aug 2006) | 1 line Let us know when there was a problem and the child had to kill the parent ........ r51039 | neal.norwitz | 2006-08-02 08:46:21 +0200 (Wed, 02 Aug 2006) | 5 lines Patch #1519025 and bug #926423: If a KeyboardInterrupt occurs during a socket operation on a socket with a timeout, the exception will be caught correctly. Previously, the exception was not caught. ........ r51040 | neal.norwitz | 2006-08-02 09:09:32 +0200 (Wed, 02 Aug 2006) | 1 line Add some explanation about Klocwork and Coverity static analysis ........ r51041 | anthony.baxter | 2006-08-02 09:43:09 +0200 (Wed, 02 Aug 2006) | 1 line pre-release machinations ........ r51043 | thomas.heller | 2006-08-02 13:35:31 +0200 (Wed, 02 Aug 2006) | 4 lines A few nore words about what ctypes does. Document that using the wrong calling convention can also raise 'ValueError: Procedure called with the wrong number of arguments'. ........ r51045 | thomas.heller | 2006-08-02 14:00:13 +0200 (Wed, 02 Aug 2006) | 1 line Fix a mistake. ........ r51046 | martin.v.loewis | 2006-08-02 15:53:55 +0200 (Wed, 02 Aug 2006) | 3 lines Correction of patch #1455898: In the mbcs decoder, set final=False for stream decoder, but final=True for the decode function. ........ r51049 | tim.peters | 2006-08-02 20:19:35 +0200 (Wed, 02 Aug 2006) | 2 lines Add missing svn:eol-style property to text files. ........ r51079 | neal.norwitz | 2006-08-04 06:50:21 +0200 (Fri, 04 Aug 2006) | 3 lines Bug #1531405, format_exception no longer raises an exception if str(exception) raised an exception. ........ r51080 | neal.norwitz | 2006-08-04 06:58:47 +0200 (Fri, 04 Aug 2006) | 11 lines Bug #1191458: tracing over for loops now produces a line event on each iteration. I'm not positive this is the best way to handle this. I'm also not sure that there aren't other cases where the lnotab is generated incorrectly. It would be great if people that use pdb or tracing could test heavily. Also: * Remove dead/duplicated code that wasn't used/necessary because we already handled the docstring prior to entering the loop. * add some debugging code into the compiler (#if 0'd out). ........ r51081 | neal.norwitz | 2006-08-04 07:09:28 +0200 (Fri, 04 Aug 2006) | 4 lines Bug #1333982: string/number constants were inappropriately stored in the byte code and co_consts even if they were not used, ie immediately popped off the stack. ........ r51082 | neal.norwitz | 2006-08-04 07:12:19 +0200 (Fri, 04 Aug 2006) | 1 line There were really two issues ........ r51084 | fred.drake | 2006-08-04 07:17:21 +0200 (Fri, 04 Aug 2006) | 1 line SF patch #1534048 (bug #1531003): fix typo in error message ........ r51085 | gregory.p.smith | 2006-08-04 07:17:47 +0200 (Fri, 04 Aug 2006) | 3 lines fix typos ........ r51087 | georg.brandl | 2006-08-04 08:03:53 +0200 (Fri, 04 Aug 2006) | 3 lines Fix bug caused by first decrefing, then increfing. ........ r51109 | neil.schemenauer | 2006-08-04 18:20:30 +0200 (Fri, 04 Aug 2006) | 5 lines Fix the 'compiler' package to generate correct code for MAKE_CLOSURE. In the 2.5 development cycle, MAKE_CLOSURE as changed to take free variables as a tuple rather than as individual items on the stack. Closes patch #1534084. ........ r51110 | georg.brandl | 2006-08-04 20:03:37 +0200 (Fri, 04 Aug 2006) | 3 lines Change fix for segfaulting property(), add a NEWS entry and a test. ........ r51111 | georg.brandl | 2006-08-04 20:07:34 +0200 (Fri, 04 Aug 2006) | 3 lines Better fix for bug #1531405, not executing str(value) twice. ........ r51112 | thomas.heller | 2006-08-04 20:17:40 +0200 (Fri, 04 Aug 2006) | 1 line On Windows, make PyErr_Warn an exported function again. ........ r51113 | thomas.heller | 2006-08-04 20:57:34 +0200 (Fri, 04 Aug 2006) | 4 lines Fix #1530448 - fix ctypes build failure on solaris 10. The '-mimpure-text' linker flag is required when linking _ctypes.so. ........ r51114 | thomas.heller | 2006-08-04 21:49:31 +0200 (Fri, 04 Aug 2006) | 3 lines Fix #1534738: win32 debug version of _msi must be _msi_d.pyd, not _msi.pyd. Fix the name of the pdb file as well. ........ r51115 | andrew.kuchling | 2006-08-04 22:37:43 +0200 (Fri, 04 Aug 2006) | 1 line Typo fixes ........ r51116 | andrew.kuchling | 2006-08-04 23:10:03 +0200 (Fri, 04 Aug 2006) | 1 line Fix mangled sentence ........ r51118 | tim.peters | 2006-08-05 00:00:35 +0200 (Sat, 05 Aug 2006) | 2 lines Whitespace normalization. ........ r51119 | bob.ippolito | 2006-08-05 01:59:21 +0200 (Sat, 05 Aug 2006) | 5 lines Fix #1530559, struct.pack raises TypeError where it used to convert. Passing float arguments to struct.pack when integers are expected now triggers a DeprecationWarning. ........ r51123 | georg.brandl | 2006-08-05 08:10:54 +0200 (Sat, 05 Aug 2006) | 3 lines Patch #1534922: correct and enhance unittest docs. ........ r51126 | georg.brandl | 2006-08-06 09:06:33 +0200 (Sun, 06 Aug 2006) | 2 lines Bug #1535182: really test the xreadlines() method of bz2 objects. ........ r51128 | georg.brandl | 2006-08-06 09:26:21 +0200 (Sun, 06 Aug 2006) | 4 lines Bug #1535081: A leading underscore has been added to the names of the md5 and sha modules, so add it in Modules/Setup.dist too. ........ r51129 | georg.brandl | 2006-08-06 10:23:54 +0200 (Sun, 06 Aug 2006) | 3 lines Bug #1535165: fixed a segfault in input() and raw_input() when sys.stdin is closed. ........ r51131 | georg.brandl | 2006-08-06 11:17:16 +0200 (Sun, 06 Aug 2006) | 2 lines Don't produce output in test_builtin. ........ r51133 | andrew.macintyre | 2006-08-06 14:37:03 +0200 (Sun, 06 Aug 2006) | 4 lines test_threading now skips testing alternate thread stack sizes on platforms that don't support changing thread stack size. ........ r51134 | andrew.kuchling | 2006-08-07 00:07:04 +0200 (Mon, 07 Aug 2006) | 2 lines [Patch #1464056] Ensure that we use the panelw library when linking with ncursesw. Once I see how the buildbots react, I'll backport this to 2.4. ........ r51137 | georg.brandl | 2006-08-08 13:52:34 +0200 (Tue, 08 Aug 2006) | 3 lines webbrowser: Silence stderr output if no gconftool or gnome browser found ........ r51138 | georg.brandl | 2006-08-08 13:56:21 +0200 (Tue, 08 Aug 2006) | 7 lines Remove "non-mapping" and "non-sequence" from TypeErrors raised by PyMapping_Size and PySequence_Size. Because len() tries first sequence, then mapping size, it will always raise a "non-mapping object has no len" error which is confusing. ........ r51139 | thomas.heller | 2006-08-08 19:37:00 +0200 (Tue, 08 Aug 2006) | 3 lines memcmp() can return values other than -1, 0, and +1 but tp_compare must not. ........ r51140 | thomas.heller | 2006-08-08 19:39:20 +0200 (Tue, 08 Aug 2006) | 1 line Remove accidently committed, duplicated test. ........ r51147 | andrew.kuchling | 2006-08-08 20:50:14 +0200 (Tue, 08 Aug 2006) | 1 line Reword paragraph to clarify ........ r51148 | andrew.kuchling | 2006-08-08 20:56:08 +0200 (Tue, 08 Aug 2006) | 1 line Move obmalloc item into C API section ........ r51149 | andrew.kuchling | 2006-08-08 21:00:14 +0200 (Tue, 08 Aug 2006) | 1 line 'Other changes' section now has only one item; move the item elsewhere and remove the section ........ r51150 | andrew.kuchling | 2006-08-08 21:00:34 +0200 (Tue, 08 Aug 2006) | 1 line Bump version number ........ r51151 | georg.brandl | 2006-08-08 22:11:22 +0200 (Tue, 08 Aug 2006) | 2 lines Bug #1536828: typo: TypeType should have been StringType. ........ r51153 | georg.brandl | 2006-08-08 22:13:13 +0200 (Tue, 08 Aug 2006) | 2 lines Bug #1536660: separate two words. ........ r51155 | georg.brandl | 2006-08-08 22:48:10 +0200 (Tue, 08 Aug 2006) | 3 lines ``str`` is now the same object as ``types.StringType``. ........ r51156 | tim.peters | 2006-08-09 02:52:26 +0200 (Wed, 09 Aug 2006) | 2 lines Whitespace normalization. ........ r51158 | georg.brandl | 2006-08-09 09:03:22 +0200 (Wed, 09 Aug 2006) | 4 lines Introduce an upper bound on tuple nesting depth in C argument format strings; fixes rest of #1523610. ........ r51160 | martin.v.loewis | 2006-08-09 09:57:39 +0200 (Wed, 09 Aug 2006) | 4 lines __hash__ may now return long int; the final hash value is obtained by invoking hash on the long int. Fixes #1536021. ........ r51168 | andrew.kuchling | 2006-08-09 15:03:41 +0200 (Wed, 09 Aug 2006) | 1 line [Bug #1536021] Mention __hash__ change ........ r51169 | andrew.kuchling | 2006-08-09 15:57:05 +0200 (Wed, 09 Aug 2006) | 1 line [Patch #1534027] Add notes on locale module changes ........ r51170 | andrew.kuchling | 2006-08-09 16:05:35 +0200 (Wed, 09 Aug 2006) | 1 line Add missing 'self' parameters ........ r51171 | andrew.kuchling | 2006-08-09 16:06:19 +0200 (Wed, 09 Aug 2006) | 1 line Reindent code ........ r51172 | armin.rigo | 2006-08-09 16:55:26 +0200 (Wed, 09 Aug 2006) | 2 lines Fix and test for an infinite C recursion. ........ r51173 | ronald.oussoren | 2006-08-09 16:56:33 +0200 (Wed, 09 Aug 2006) | 2 lines It's unlikely that future versions will require _POSIX_C_SOURCE ........ r51178 | armin.rigo | 2006-08-09 17:37:26 +0200 (Wed, 09 Aug 2006) | 2 lines Concatenation on a long string breaks (SF #1526585). ........ r51180 | kurt.kaiser | 2006-08-09 18:46:15 +0200 (Wed, 09 Aug 2006) | 8 lines 1. When used w/o subprocess, all exceptions were preceeded by an error message claiming they were IDLE internal errors (since 1.2a1). 2. Add Ronald Oussoren to CREDITS M NEWS.txt M PyShell.py M CREDITS.txt ........ r51181 | kurt.kaiser | 2006-08-09 19:47:15 +0200 (Wed, 09 Aug 2006) | 4 lines As a slight enhancement to the previous checkin, improve the internal error reporting by moving message to IDLE console. ........ r51182 | andrew.kuchling | 2006-08-09 20:23:14 +0200 (Wed, 09 Aug 2006) | 1 line Typo fix ........ r51183 | kurt.kaiser | 2006-08-09 22:34:46 +0200 (Wed, 09 Aug 2006) | 2 lines ToggleTab dialog was setting indent to 8 even if cancelled (since 1.2a1). ........ r51184 | martin.v.loewis | 2006-08-10 01:42:18 +0200 (Thu, 10 Aug 2006) | 2 lines Add some commentary on -mimpure-text. ........ r51185 | tim.peters | 2006-08-10 02:58:49 +0200 (Thu, 10 Aug 2006) | 2 lines Add missing svn:eol-style property to text files. ........ r51186 | kurt.kaiser | 2006-08-10 03:41:17 +0200 (Thu, 10 Aug 2006) | 2 lines Changing tokenize (39046) to detect dedent broke tabnanny check (since 1.2a1) ........ r51187 | tim.peters | 2006-08-10 05:01:26 +0200 (Thu, 10 Aug 2006) | 13 lines test_copytree_simple(): This was leaving behind two new temp directories each time it ran, at least on Windows. Several changes: explicitly closed all files; wrapped long lines; stopped suppressing errors when removing a file or directory fails (removing /shouldn't/ fail!); and changed what appeared to be incorrect usage of os.removedirs() (that doesn't remove empty directories at and /under/ the given path, instead it must be given an empty leaf directory and then deletes empty directories moving /up/ the path -- could be that the conceptually simpler shutil.rmtree() was really actually intended here). ........
2006-08-11 11:57:12 -03:00
requiring allocation in multiples of the system memory page size\n\
- platform documentation should be referred to for more information\n\
(4 KiB pages are common; using multiples of 4096 for the stack size is\n\
Merged revisions 46753-51188 via svnmerge from svn+ssh://pythondev@svn.python.org/python/trunk ........ r46755 | brett.cannon | 2006-06-08 18:23:04 +0200 (Thu, 08 Jun 2006) | 4 lines Make binascii.hexlify() use s# for its arguments instead of t# to actually match its documentation stating it accepts any read-only buffer. ........ r46757 | brett.cannon | 2006-06-08 19:00:45 +0200 (Thu, 08 Jun 2006) | 8 lines Buffer objects would return the read or write buffer for a wrapped object when the char buffer was requested. Now it actually returns the char buffer if available or raises a TypeError if it isn't (as is raised for the other buffer types if they are not present but requested). Not a backport candidate since it does change semantics of the buffer object (although it could be argued this is enough of a bug to bother backporting). ........ r46760 | andrew.kuchling | 2006-06-09 03:10:17 +0200 (Fri, 09 Jun 2006) | 1 line Update functools section ........ r46762 | tim.peters | 2006-06-09 04:11:02 +0200 (Fri, 09 Jun 2006) | 6 lines Whitespace normalization. Since test_file is implicated in mysterious test failures when followed by test_optparse, if I had any brains I'd look at the checkin that last changed test_file ;-) ........ r46763 | tim.peters | 2006-06-09 05:09:42 +0200 (Fri, 09 Jun 2006) | 5 lines To boost morale :-), force test_optparse to run immediately after test_file until we can figure out how to fix it. (See python-dev; at the moment we don't even know which checkin caused the problem.) ........ r46764 | tim.peters | 2006-06-09 05:51:41 +0200 (Fri, 09 Jun 2006) | 6 lines AutoFileTests.tearDown(): Removed mysterious undocumented try/except. Remove TESTFN. Throughout: used open() instead of file(), and wrapped long lines. ........ r46765 | tim.peters | 2006-06-09 06:02:06 +0200 (Fri, 09 Jun 2006) | 8 lines testUnicodeOpen(): I have no idea why, but making this test clean up after itself appears to fix the test failures when test_optparse follows test_file. test_main(): Get rid of TESTFN no matter what. That's also enough to fix the mystery failures. Doesn't hurt to fix them twice :-) ........ r46766 | tim.peters | 2006-06-09 07:12:40 +0200 (Fri, 09 Jun 2006) | 6 lines Remove the temporary hack to force test_optparse to run immediately after test_file. At least 8 buildbot boxes passed since the underlying problem got fixed, and they all failed before the fix, so there's no point to this anymore. ........ r46767 | neal.norwitz | 2006-06-09 07:54:18 +0200 (Fri, 09 Jun 2006) | 1 line Fix grammar and reflow ........ r46769 | andrew.kuchling | 2006-06-09 12:22:35 +0200 (Fri, 09 Jun 2006) | 1 line Markup fix ........ r46773 | andrew.kuchling | 2006-06-09 15:15:57 +0200 (Fri, 09 Jun 2006) | 1 line [Bug #1472827] Make saxutils.XMLGenerator handle \r\n\t in attribute values by escaping them properly. 2.4 bugfix candidate. ........ r46778 | kristjan.jonsson | 2006-06-09 18:28:01 +0200 (Fri, 09 Jun 2006) | 2 lines Turn off warning about deprecated CRT functions on for VisualStudio .NET 2005. Make the definition #ARRAYSIZE conditional. VisualStudio .NET 2005 already has it defined using a better gimmick. ........ r46779 | phillip.eby | 2006-06-09 18:40:18 +0200 (Fri, 09 Jun 2006) | 2 lines Import wsgiref into the stdlib, as of the external version 0.1-r2181. ........ r46783 | andrew.kuchling | 2006-06-09 18:44:40 +0200 (Fri, 09 Jun 2006) | 1 line Add note about XMLGenerator bugfix ........ r46784 | andrew.kuchling | 2006-06-09 18:46:51 +0200 (Fri, 09 Jun 2006) | 1 line Add note about wsgiref ........ r46785 | brett.cannon | 2006-06-09 19:05:48 +0200 (Fri, 09 Jun 2006) | 2 lines Fix inconsistency in naming within an enum. ........ r46787 | tim.peters | 2006-06-09 19:47:00 +0200 (Fri, 09 Jun 2006) | 2 lines Whitespace normalization. ........ r46792 | georg.brandl | 2006-06-09 20:29:52 +0200 (Fri, 09 Jun 2006) | 3 lines Test file.__exit__. ........ r46794 | brett.cannon | 2006-06-09 20:40:46 +0200 (Fri, 09 Jun 2006) | 2 lines svn:ignore .pyc and .pyo files. ........ r46795 | georg.brandl | 2006-06-09 20:45:48 +0200 (Fri, 09 Jun 2006) | 3 lines RFE #1491485: str/unicode.endswith()/startswith() now accept a tuple as first argument. ........ r46798 | andrew.kuchling | 2006-06-09 21:03:16 +0200 (Fri, 09 Jun 2006) | 1 line Describe startswith()/endswiith() change; add reminder about wsgiref ........ r46799 | tim.peters | 2006-06-09 21:24:44 +0200 (Fri, 09 Jun 2006) | 11 lines Implementing a happy idea from Georg Brandl: make runtest() try to clean up files and directories the tests often leave behind by mistake. This is the first time in history I don't have a bogus "db_home" directory after running the tests ;-) Also worked on runtest's docstring, to say something about all the arguments, and to document the non-obvious return values. New functions runtest_inner() and cleanup_test_droppings() in support of the above. ........ r46800 | andrew.kuchling | 2006-06-09 21:43:25 +0200 (Fri, 09 Jun 2006) | 1 line Remove unused variable ........ r46801 | andrew.kuchling | 2006-06-09 21:56:05 +0200 (Fri, 09 Jun 2006) | 1 line Add some wsgiref text ........ r46803 | thomas.heller | 2006-06-09 21:59:11 +0200 (Fri, 09 Jun 2006) | 1 line set eol-style svn property ........ r46804 | thomas.heller | 2006-06-09 22:01:01 +0200 (Fri, 09 Jun 2006) | 1 line set eol-style svn property ........ r46805 | georg.brandl | 2006-06-09 22:43:48 +0200 (Fri, 09 Jun 2006) | 3 lines Make use of new str.startswith/endswith semantics. Occurences in email and compiler were ignored due to backwards compat requirements. ........ r46806 | brett.cannon | 2006-06-10 00:31:23 +0200 (Sat, 10 Jun 2006) | 4 lines An object with __call__ as an attribute, when called, will have that attribute checked for __call__ itself, and will continue to look until it finds an object without the attribute. This can lead to an infinite recursion. Closes bug #532646, again. Will be backported. ........ r46808 | brett.cannon | 2006-06-10 00:45:54 +0200 (Sat, 10 Jun 2006) | 2 lines Fix bug introduced in rev. 46806 by not having variable declaration at the top of a block. ........ r46812 | georg.brandl | 2006-06-10 08:40:50 +0200 (Sat, 10 Jun 2006) | 4 lines Apply perky's fix for #1503157: "/".join([u"", u""]) raising OverflowError. Also improve error message on overflow. ........ r46817 | martin.v.loewis | 2006-06-10 10:14:03 +0200 (Sat, 10 Jun 2006) | 2 lines Port cygwin kill_python changes from 2.4 branch. ........ r46818 | armin.rigo | 2006-06-10 12:57:40 +0200 (Sat, 10 Jun 2006) | 4 lines SF bug #1503294. PyThreadState_GET() complains if the tstate is NULL, but only in debug mode. ........ r46819 | martin.v.loewis | 2006-06-10 14:23:46 +0200 (Sat, 10 Jun 2006) | 4 lines Patch #1495999: Part two of Windows CE changes. - update header checks, using autoconf - provide dummies for getenv, environ, and GetVersion - adjust MSC_VER check in socketmodule.c ........ r46820 | skip.montanaro | 2006-06-10 16:09:11 +0200 (Sat, 10 Jun 2006) | 1 line document the class, not its initializer ........ r46821 | greg.ward | 2006-06-10 18:40:01 +0200 (Sat, 10 Jun 2006) | 4 lines Sync with Optik docs (rev 518): * restore "Extending optparse" section * document ALWAYS_TYPED_ACTIONS (SF #1449311) ........ r46824 | thomas.heller | 2006-06-10 21:51:46 +0200 (Sat, 10 Jun 2006) | 8 lines Upgrade to ctypes version 0.9.9.7. Summary of changes: - support for 'variable sized' data - support for anonymous structure/union fields - fix severe bug with certain arrays or structures containing more than 256 fields ........ r46825 | thomas.heller | 2006-06-10 21:55:36 +0200 (Sat, 10 Jun 2006) | 8 lines Upgrade to ctypes version 0.9.9.7. Summary of changes: - support for 'variable sized' data - support for anonymous structure/union fields - fix severe bug with certain arrays or structures containing more than 256 fields ........ r46826 | fred.drake | 2006-06-10 22:01:34 +0200 (Sat, 10 Jun 2006) | 4 lines SF patch #1303595: improve description of __builtins__, explaining how it varies between __main__ and other modules, and strongly suggest not touching it but using __builtin__ if absolutely necessary ........ r46827 | fred.drake | 2006-06-10 22:02:58 +0200 (Sat, 10 Jun 2006) | 1 line credit for SF patch #1303595 ........ r46831 | thomas.heller | 2006-06-10 22:29:34 +0200 (Sat, 10 Jun 2006) | 2 lines New docs for ctypes. ........ r46834 | thomas.heller | 2006-06-10 23:07:19 +0200 (Sat, 10 Jun 2006) | 1 line Fix a wrong printf format. ........ r46835 | thomas.heller | 2006-06-10 23:17:58 +0200 (Sat, 10 Jun 2006) | 1 line Fix the second occurrence of the problematic printf format. ........ r46837 | thomas.heller | 2006-06-10 23:56:03 +0200 (Sat, 10 Jun 2006) | 1 line Don't use C++ comment. ........ r46838 | thomas.heller | 2006-06-11 00:01:50 +0200 (Sun, 11 Jun 2006) | 1 line Handle failure of PyMem_Realloc. ........ r46839 | skip.montanaro | 2006-06-11 00:38:13 +0200 (Sun, 11 Jun 2006) | 2 lines Suppress warning on MacOSX about possible use before set of proc. ........ r46840 | tim.peters | 2006-06-11 00:51:45 +0200 (Sun, 11 Jun 2006) | 8 lines shuffle() doscstring: Removed warning about sequence length versus generator period. While this was a real weakness of the older WH generator for lists with just a few dozen elements, and so could potentially bite the naive ;-), the Twister should show excellent behavior up to at least 600 elements. Module docstring: reflowed some jarringly short lines. ........ r46844 | greg.ward | 2006-06-11 02:40:49 +0200 (Sun, 11 Jun 2006) | 4 lines Bug #1361643: fix textwrap.dedent() so it handles tabs appropriately, i.e. do *not* expand tabs, but treat them as whitespace that is not equivalent to spaces. Add a couple of test cases. Clarify docs. ........ r46850 | neal.norwitz | 2006-06-11 07:44:18 +0200 (Sun, 11 Jun 2006) | 5 lines Fix Coverity # 146. newDBSequenceObject would deref dbobj, so it can't be NULL. We know it's not NULL from the ParseTuple and DbObject_Check will verify it's not NULL. ........ r46851 | neal.norwitz | 2006-06-11 07:45:25 +0200 (Sun, 11 Jun 2006) | 4 lines Wrap some long lines Top/Bottom factor out some common expressions Add a XXX comment about widing offset. ........ r46852 | neal.norwitz | 2006-06-11 07:45:47 +0200 (Sun, 11 Jun 2006) | 1 line Add versionadded to doc ........ r46853 | neal.norwitz | 2006-06-11 07:47:14 +0200 (Sun, 11 Jun 2006) | 3 lines Update doc to make it agree with code. Bottom factor out some common code. ........ r46854 | neal.norwitz | 2006-06-11 07:48:14 +0200 (Sun, 11 Jun 2006) | 3 lines f_code can't be NULL based on Frame_New and other code that derefs it. So there doesn't seem to be much point to checking here. ........ r46855 | neal.norwitz | 2006-06-11 09:26:27 +0200 (Sun, 11 Jun 2006) | 1 line Fix errors found by pychecker ........ r46856 | neal.norwitz | 2006-06-11 09:26:50 +0200 (Sun, 11 Jun 2006) | 1 line warnings was imported at module scope, no need to import again ........ r46857 | neal.norwitz | 2006-06-11 09:27:56 +0200 (Sun, 11 Jun 2006) | 5 lines Fix errors found by pychecker. I think these changes are correct, but I'm not sure. Could someone who knows how this module works test it? It can at least start on the cmd line. ........ r46858 | neal.norwitz | 2006-06-11 10:35:14 +0200 (Sun, 11 Jun 2006) | 1 line Fix errors found by pychecker ........ r46859 | ronald.oussoren | 2006-06-11 16:33:36 +0200 (Sun, 11 Jun 2006) | 4 lines This patch improves the L&F of IDLE on OSX. The changes are conditionalized on being in an IDLE.app bundle on darwin. This does a slight reorganisation of the menus and adds support for file-open events. ........ r46860 | greg.ward | 2006-06-11 16:42:41 +0200 (Sun, 11 Jun 2006) | 1 line SF #1366250: optparse docs: fix inconsistency in variable name; minor tweaks. ........ r46861 | greg.ward | 2006-06-11 18:24:11 +0200 (Sun, 11 Jun 2006) | 3 lines Bug #1498146: fix optparse to handle Unicode strings in option help, description, and epilog. ........ r46862 | thomas.heller | 2006-06-11 19:04:22 +0200 (Sun, 11 Jun 2006) | 2 lines Release the GIL during COM method calls, to avoid deadlocks in Python coded COM objects. ........ r46863 | tim.peters | 2006-06-11 21:42:51 +0200 (Sun, 11 Jun 2006) | 2 lines Whitespace normalization. ........ r46864 | tim.peters | 2006-06-11 21:43:49 +0200 (Sun, 11 Jun 2006) | 2 lines Add missing svn:eol-style property to text files. ........ r46865 | ronald.oussoren | 2006-06-11 21:45:57 +0200 (Sun, 11 Jun 2006) | 2 lines Remove message about using make frameworkinstall, that's no longer necesssary ........ r46866 | ronald.oussoren | 2006-06-11 22:23:29 +0200 (Sun, 11 Jun 2006) | 2 lines Use configure to substitute the correct prefix instead of hardcoding ........ r46867 | ronald.oussoren | 2006-06-11 22:24:45 +0200 (Sun, 11 Jun 2006) | 4 lines - Change fixapplepython23.py to ensure that it will run with /usr/bin/python on intel macs. - Fix some minor problems in the installer for OSX ........ r46868 | neal.norwitz | 2006-06-11 22:25:56 +0200 (Sun, 11 Jun 2006) | 5 lines Try to fix several networking tests. The problem is that if hosts have a search path setup, some of these hosts resolve to the wrong address. By appending a period to the hostname, the hostname should only resolve to what we want it to resolve to. Hopefully this doesn't break different bots. ........ r46869 | neal.norwitz | 2006-06-11 22:42:02 +0200 (Sun, 11 Jun 2006) | 7 lines Try to fix another networking test. The problem is that if hosts have a search path setup, some of these hosts resolve to the wrong address. By appending a period to the hostname, the hostname should only resolve to what we want it to resolve to. Hopefully this doesn't break different bots. Also add more info to failure message to aid debugging test failure. ........ r46870 | neal.norwitz | 2006-06-11 22:46:46 +0200 (Sun, 11 Jun 2006) | 4 lines Fix test on PPC64 buildbot. It raised an IOError (really an URLError which derives from an IOError). That seems valid. Env Error includes both OSError and IOError, so this seems like a reasonable fix. ........ r46871 | tim.peters | 2006-06-11 22:52:59 +0200 (Sun, 11 Jun 2006) | 10 lines compare_generic_iter(): Fixed the failure of test_wsgiref's testFileWrapper when running with -O. test_simple_validation_error still fails under -O. That appears to be because wsgiref's validate.py uses `assert` statements all over the place to check arguments for sanity. That should all be changed (it's not a logical error in the software if a user passes bogus arguments, so this isn't a reasonable use for `assert` -- checking external preconditions should generally raise ValueError or TypeError instead, as appropriate). ........ r46872 | neal.norwitz | 2006-06-11 23:38:38 +0200 (Sun, 11 Jun 2006) | 1 line Get test to pass on S/390. Shout if you think this change is incorrect. ........ r46873 | neal.norwitz | 2006-06-12 04:05:55 +0200 (Mon, 12 Jun 2006) | 1 line Cleanup Py_ssize_t a little (get rid of second #ifdef) ........ r46874 | neal.norwitz | 2006-06-12 04:06:17 +0200 (Mon, 12 Jun 2006) | 1 line Fix some Py_ssize_t issues ........ r46875 | neal.norwitz | 2006-06-12 04:06:42 +0200 (Mon, 12 Jun 2006) | 1 line Fix some Py_ssize_t issues ........ r46876 | neal.norwitz | 2006-06-12 04:07:24 +0200 (Mon, 12 Jun 2006) | 2 lines Cleanup: Remove import of types to get StringTypes, we can just use basestring. ........ r46877 | neal.norwitz | 2006-06-12 04:07:57 +0200 (Mon, 12 Jun 2006) | 1 line Don't truncate if size_t is bigger than uint ........ r46878 | neal.norwitz | 2006-06-12 04:08:41 +0200 (Mon, 12 Jun 2006) | 1 line Don't leak the list object if there's an error allocating the item storage. Backport candidate ........ r46879 | neal.norwitz | 2006-06-12 04:09:03 +0200 (Mon, 12 Jun 2006) | 1 line Fix typo. Backport if anyone cares. :-) ........ r46880 | neal.norwitz | 2006-06-12 04:09:34 +0200 (Mon, 12 Jun 2006) | 1 line Fix indentation of case and a Py_ssize_t issue. ........ r46881 | neal.norwitz | 2006-06-12 04:11:18 +0200 (Mon, 12 Jun 2006) | 3 lines Get rid of f_restricted too. Doc the other 4 ints that were already removed at the NeedForSpeed sprint. ........ r46882 | neal.norwitz | 2006-06-12 04:13:21 +0200 (Mon, 12 Jun 2006) | 1 line Fix the socket tests so they can be run concurrently. Backport candidate ........ r46883 | neal.norwitz | 2006-06-12 04:16:10 +0200 (Mon, 12 Jun 2006) | 1 line i and j are initialized below when used. No need to do it twice ........ r46884 | neal.norwitz | 2006-06-12 05:05:03 +0200 (Mon, 12 Jun 2006) | 1 line Remove unused import ........ r46885 | neal.norwitz | 2006-06-12 05:05:40 +0200 (Mon, 12 Jun 2006) | 1 line Impl ssize_t ........ r46886 | neal.norwitz | 2006-06-12 05:33:09 +0200 (Mon, 12 Jun 2006) | 6 lines Patch #1503046, Conditional compilation of zlib.(de)compressobj.copy copy is only in newer versions of zlib. This should allow zlibmodule to work with older versions like the Tru64 buildbot. ........ r46887 | phillip.eby | 2006-06-12 06:04:32 +0200 (Mon, 12 Jun 2006) | 2 lines Sync w/external release 0.1.2. Please see PEP 360 before making changes to external packages. ........ r46888 | martin.v.loewis | 2006-06-12 06:26:31 +0200 (Mon, 12 Jun 2006) | 2 lines Get rid of function pointer cast. ........ r46889 | thomas.heller | 2006-06-12 08:05:57 +0200 (Mon, 12 Jun 2006) | 3 lines I don't know how that happend, but the entire file contents was duplicated. Thanks to Simon Percivall for the heads up. ........ r46890 | nick.coghlan | 2006-06-12 10:19:37 +0200 (Mon, 12 Jun 2006) | 1 line Fix site module docstring to match the code ........ r46891 | nick.coghlan | 2006-06-12 10:23:02 +0200 (Mon, 12 Jun 2006) | 1 line Fix site module docstring to match the code for Mac OSX, too ........ r46892 | nick.coghlan | 2006-06-12 10:27:13 +0200 (Mon, 12 Jun 2006) | 1 line The site module documentation also described the Windows behaviour incorrectly. ........ r46893 | nick.coghlan | 2006-06-12 12:17:11 +0200 (Mon, 12 Jun 2006) | 1 line Make the -m switch conform to the documentation of sys.path by behaving like the -c switch ........ r46894 | kristjan.jonsson | 2006-06-12 17:45:12 +0200 (Mon, 12 Jun 2006) | 2 lines Fix the CRT argument error handling for VisualStudio .NET 2005. Install a CRT error handler and disable the assertion for debug builds. This causes CRT to set errno to EINVAL. This update fixes crash cases in the test suite where the default CRT error handler would cause process exit. ........ r46899 | thomas.heller | 2006-06-12 22:56:48 +0200 (Mon, 12 Jun 2006) | 1 line Add pep-291 compatibility markers. ........ r46901 | ka-ping.yee | 2006-06-13 01:47:52 +0200 (Tue, 13 Jun 2006) | 5 lines Add the uuid module. This module has been tested so far on Windows XP (Python 2.4 and 2.5a2), Mac OS X (Python 2.3, 2.4, and 2.5a2), and Linux (Python 2.4 and 2.5a2). ........ r46902 | tim.peters | 2006-06-13 02:30:01 +0200 (Tue, 13 Jun 2006) | 2 lines Whitespace normalization. ........ r46903 | tim.peters | 2006-06-13 02:30:50 +0200 (Tue, 13 Jun 2006) | 2 lines Added missing svn:eol-style property to text files. ........ r46905 | tim.peters | 2006-06-13 05:30:07 +0200 (Tue, 13 Jun 2006) | 5 lines get_matching_blocks(): rewrote code & comments so they match; added more comments about why it's this way at all; and removed what looked like needless expense (sorting (i, j, k) triples directly should give exactly the same order as sorting (i, (i, j, k)) pairs). ........ r46906 | neal.norwitz | 2006-06-13 06:08:53 +0200 (Tue, 13 Jun 2006) | 1 line Don't fail if another process is listening on our port. ........ r46908 | neal.norwitz | 2006-06-13 10:28:19 +0200 (Tue, 13 Jun 2006) | 2 lines Initialize the type object so pychecker can't crash the interpreter. ........ r46909 | neal.norwitz | 2006-06-13 10:41:06 +0200 (Tue, 13 Jun 2006) | 1 line Verify the crash due to EncodingMap not initialized does not return ........ r46910 | thomas.heller | 2006-06-13 10:56:14 +0200 (Tue, 13 Jun 2006) | 3 lines Add some windows datatypes that were missing from this file, and add the aliases defined in windows header files for the structures. ........ r46911 | thomas.heller | 2006-06-13 11:40:14 +0200 (Tue, 13 Jun 2006) | 3 lines Add back WCHAR, UINT, DOUBLE, _LARGE_INTEGER, _ULARGE_INTEGER. VARIANT_BOOL is a special _ctypes data type, not c_short. ........ r46912 | ronald.oussoren | 2006-06-13 13:19:56 +0200 (Tue, 13 Jun 2006) | 4 lines Linecache contains support for PEP302 loaders, but fails to deal with loaders that return None to indicate that the module is valid but no source is available. This patch fixes that. ........ r46913 | andrew.kuchling | 2006-06-13 13:57:04 +0200 (Tue, 13 Jun 2006) | 1 line Mention uuid module ........ r46915 | walter.doerwald | 2006-06-13 14:02:12 +0200 (Tue, 13 Jun 2006) | 2 lines Fix passing errors to the encoder and decoder functions. ........ r46917 | walter.doerwald | 2006-06-13 14:04:43 +0200 (Tue, 13 Jun 2006) | 3 lines errors is an attribute in the incremental decoder not an argument. ........ r46919 | andrew.macintyre | 2006-06-13 17:04:24 +0200 (Tue, 13 Jun 2006) | 11 lines Patch #1454481: Make thread stack size runtime tunable. Heavily revised, comprising revisions: 46640 - original trunk revision (backed out in r46655) 46647 - markup fix (backed out in r46655) 46692:46918 merged from branch aimacintyre-sf1454481 branch tested on buildbots (Windows buildbots had problems not related to these changes). ........ r46920 | brett.cannon | 2006-06-13 18:06:55 +0200 (Tue, 13 Jun 2006) | 2 lines Remove unused variable. ........ r46921 | andrew.kuchling | 2006-06-13 18:41:41 +0200 (Tue, 13 Jun 2006) | 1 line Add ability to set stack size ........ r46923 | marc-andre.lemburg | 2006-06-13 19:04:26 +0200 (Tue, 13 Jun 2006) | 2 lines Update pybench to version 2.0. ........ r46924 | marc-andre.lemburg | 2006-06-13 19:07:14 +0200 (Tue, 13 Jun 2006) | 2 lines Revert wrong svn copy. ........ r46925 | andrew.macintyre | 2006-06-13 19:14:36 +0200 (Tue, 13 Jun 2006) | 2 lines fix exception usage ........ r46927 | tim.peters | 2006-06-13 20:37:07 +0200 (Tue, 13 Jun 2006) | 2 lines Whitespace normalization. ........ r46928 | marc-andre.lemburg | 2006-06-13 20:56:56 +0200 (Tue, 13 Jun 2006) | 9 lines Updated to pybench 2.0. See svn.python.org/external/pybench-2.0 for the original import of that version. Note that platform.py was not copied over from pybench-2.0 since it is already part of Python 2.5. ........ r46929 | andrew.macintyre | 2006-06-13 21:02:35 +0200 (Tue, 13 Jun 2006) | 5 lines Increase the small thread stack size to get the test to pass reliably on the one buildbot that insists on more than 32kB of thread stack. ........ r46930 | marc-andre.lemburg | 2006-06-13 21:20:07 +0200 (Tue, 13 Jun 2006) | 2 lines Whitespace normalization. ........ r46931 | thomas.heller | 2006-06-13 22:18:43 +0200 (Tue, 13 Jun 2006) | 2 lines More docs for ctypes. ........ r46932 | brett.cannon | 2006-06-13 23:34:24 +0200 (Tue, 13 Jun 2006) | 2 lines Ignore .pyc and .pyo files in Pybench. ........ r46933 | brett.cannon | 2006-06-13 23:46:41 +0200 (Tue, 13 Jun 2006) | 7 lines If a classic class defined a __coerce__() method that just returned its two arguments in reverse, the interpreter would infinitely recourse trying to get a coercion that worked. So put in a recursion check after a coercion is made and the next call to attempt to use the coerced values. Fixes bug #992017 and closes crashers/coerce.py . ........ r46936 | gerhard.haering | 2006-06-14 00:24:47 +0200 (Wed, 14 Jun 2006) | 3 lines Merged changes from external pysqlite 2.3.0 release. Documentation updates will follow in a few hours at the latest. Then we should be ready for beta1. ........ r46937 | brett.cannon | 2006-06-14 00:26:13 +0200 (Wed, 14 Jun 2006) | 2 lines Missed test for rev. 46933; infinite recursion from __coerce__() returning its arguments reversed. ........ r46938 | gerhard.haering | 2006-06-14 00:53:48 +0200 (Wed, 14 Jun 2006) | 2 lines Updated documentation for pysqlite 2.3.0 API. ........ r46939 | tim.peters | 2006-06-14 06:09:25 +0200 (Wed, 14 Jun 2006) | 10 lines SequenceMatcher.get_matching_blocks(): This now guarantees that adjacent triples in the result list describe non-adjacent matching blocks. That's _nice_ to have, and Guido said he wanted it. Not a bugfix candidate: Guido or not ;-), this changes visible endcase semantics (note that some tests had to change), and nothing about this was documented before. Since it was working as designed, and behavior was consistent with the docs, it wasn't "a bug". ........ r46940 | tim.peters | 2006-06-14 06:13:00 +0200 (Wed, 14 Jun 2006) | 2 lines Repaired typo in new comment. ........ r46941 | tim.peters | 2006-06-14 06:15:27 +0200 (Wed, 14 Jun 2006) | 2 lines Whitespace normalization. ........ r46942 | fred.drake | 2006-06-14 06:25:02 +0200 (Wed, 14 Jun 2006) | 3 lines - make some disabled tests run what they intend when enabled - remove some over-zealous triple-quoting ........ r46943 | fred.drake | 2006-06-14 07:04:47 +0200 (Wed, 14 Jun 2006) | 3 lines add tests for two cases that are handled correctly in the current code, but that SF patch 1504676 as written mis-handles ........ r46944 | fred.drake | 2006-06-14 07:15:51 +0200 (Wed, 14 Jun 2006) | 1 line explain an XXX in more detail ........ r46945 | martin.v.loewis | 2006-06-14 07:21:04 +0200 (Wed, 14 Jun 2006) | 1 line Patch #1455898: Incremental mode for "mbcs" codec. ........ r46946 | georg.brandl | 2006-06-14 08:08:31 +0200 (Wed, 14 Jun 2006) | 3 lines Bug #1339007: Shelf objects now don't raise an exception in their __del__ method when initialization failed. ........ r46948 | thomas.heller | 2006-06-14 08:18:15 +0200 (Wed, 14 Jun 2006) | 1 line Fix docstring. ........ r46949 | georg.brandl | 2006-06-14 08:29:07 +0200 (Wed, 14 Jun 2006) | 2 lines Bug #1501122: mention __gt__ &co in description of comparison order. ........ r46951 | thomas.heller | 2006-06-14 09:08:38 +0200 (Wed, 14 Jun 2006) | 1 line Write more docs. ........ r46952 | georg.brandl | 2006-06-14 10:31:39 +0200 (Wed, 14 Jun 2006) | 3 lines Bug #1153163: describe __add__ vs __radd__ behavior when adding objects of same type/of subclasses of the other. ........ r46954 | georg.brandl | 2006-06-14 10:42:11 +0200 (Wed, 14 Jun 2006) | 3 lines Bug #1202018: add some common mime.types locations. ........ r46955 | georg.brandl | 2006-06-14 10:50:03 +0200 (Wed, 14 Jun 2006) | 3 lines Bug #1117556: SimpleHTTPServer now tries to find and use the system's mime.types file for determining MIME types. ........ r46957 | thomas.heller | 2006-06-14 11:09:08 +0200 (Wed, 14 Jun 2006) | 1 line Document paramflags. ........ r46958 | thomas.heller | 2006-06-14 11:20:11 +0200 (Wed, 14 Jun 2006) | 1 line Add an __all__ list, since this module does 'from ctypes import *'. ........ r46959 | andrew.kuchling | 2006-06-14 15:59:15 +0200 (Wed, 14 Jun 2006) | 1 line Add item ........ r46961 | georg.brandl | 2006-06-14 18:46:43 +0200 (Wed, 14 Jun 2006) | 3 lines Bug #805015: doc error in PyUnicode_FromEncodedObject. ........ r46962 | gerhard.haering | 2006-06-15 00:28:37 +0200 (Thu, 15 Jun 2006) | 10 lines - Added version checks in C code to make sure we don't trigger bugs in older SQLite versions. - Added version checks in test suite so that we don't execute tests that we know will fail with older (buggy) SQLite versions. Now, all tests should run against all SQLite versions from 3.0.8 until 3.3.6 (latest one now). The sqlite3 module can be built against all these SQLite versions and the sqlite3 module does its best to not trigger bugs in SQLite, but using SQLite 3.3.3 or later is recommended. ........ r46963 | tim.peters | 2006-06-15 00:38:13 +0200 (Thu, 15 Jun 2006) | 2 lines Whitespace normalization. ........ r46964 | neal.norwitz | 2006-06-15 06:54:29 +0200 (Thu, 15 Jun 2006) | 9 lines Speculative checkin (requires approval of Gerhard Haering) This backs out the test changes in 46962 which prevented crashes by not running the tests via a version check. All the version checks added in that rev were removed from the tests. Code was added to the error handler in connection.c that seems to work with older versions of sqlite including 3.1.3. ........ r46965 | neal.norwitz | 2006-06-15 07:55:49 +0200 (Thu, 15 Jun 2006) | 1 line Try to narrow window of failure on slow/busy boxes (ppc64 buildbot) ........ r46966 | martin.v.loewis | 2006-06-15 08:45:05 +0200 (Thu, 15 Jun 2006) | 2 lines Make import/lookup of mbcs fail on non-Windows systems. ........ r46967 | ronald.oussoren | 2006-06-15 10:14:18 +0200 (Thu, 15 Jun 2006) | 2 lines Patch #1446489 (zipfile: support for ZIP64) ........ r46968 | neal.norwitz | 2006-06-15 10:16:44 +0200 (Thu, 15 Jun 2006) | 6 lines Re-revert this change. Install the version check and don't run the test until Gerhard has time to fully debug the issue. This affects versions before 3.2.1 (possibly only versions earlier than 3.1.3). Based on discussion on python-checkins. ........ r46969 | gregory.p.smith | 2006-06-15 10:52:32 +0200 (Thu, 15 Jun 2006) | 6 lines - bsddb: multithreaded DB access using the simple bsddb module interface now works reliably. It has been updated to use automatic BerkeleyDB deadlock detection and the bsddb.dbutils.DeadlockWrap wrapper to retry database calls that would previously deadlock. [SF python bug #775414] ........ r46970 | gregory.p.smith | 2006-06-15 11:23:52 +0200 (Thu, 15 Jun 2006) | 2 lines minor documentation cleanup. mention the bsddb.db interface explicitly by name. ........ r46971 | neal.norwitz | 2006-06-15 11:57:03 +0200 (Thu, 15 Jun 2006) | 5 lines Steal the trick from test_compiler to print out a slow msg. This will hopefully get the buildbots to pass. Not sure this test will be feasible or even work. But everything is red now, so it can't get much worse. ........ r46972 | neal.norwitz | 2006-06-15 12:24:49 +0200 (Thu, 15 Jun 2006) | 1 line Print some more info to get an idea of how much longer the test will last ........ r46981 | tim.peters | 2006-06-15 20:04:40 +0200 (Thu, 15 Jun 2006) | 6 lines Try to reduce the extreme peak memory and disk-space use of this test. It probably still requires more disk space than most buildbots have, and in any case is still so intrusive that if we don't find another way to test this I'm taking my buildbot offline permanently ;-) ........ r46982 | tim.peters | 2006-06-15 20:06:29 +0200 (Thu, 15 Jun 2006) | 2 lines Whitespace normalization. ........ r46983 | tim.peters | 2006-06-15 20:07:28 +0200 (Thu, 15 Jun 2006) | 2 lines Add missing svn:eol-style property to text files. ........ r46984 | tim.peters | 2006-06-15 20:38:19 +0200 (Thu, 15 Jun 2006) | 2 lines Oops -- I introduced an off-by-6436159488 error. ........ r46990 | neal.norwitz | 2006-06-16 06:30:34 +0200 (Fri, 16 Jun 2006) | 1 line Disable this test until we can determine what to do about it ........ r46991 | neal.norwitz | 2006-06-16 06:31:06 +0200 (Fri, 16 Jun 2006) | 1 line Param name is dir, not directory. Update docstring. Backport candidate ........ r46992 | neal.norwitz | 2006-06-16 06:31:28 +0200 (Fri, 16 Jun 2006) | 1 line Add missing period in comment. ........ r46993 | neal.norwitz | 2006-06-16 06:32:43 +0200 (Fri, 16 Jun 2006) | 1 line Fix whitespace, there are memory leaks in this module. ........ r46995 | fred.drake | 2006-06-17 01:45:06 +0200 (Sat, 17 Jun 2006) | 3 lines SF patch 1504676: Make sgmllib char and entity references pluggable (implementation/tests contributed by Sam Ruby) ........ r46996 | fred.drake | 2006-06-17 03:07:54 +0200 (Sat, 17 Jun 2006) | 1 line fix change that broke the htmllib tests ........ r46998 | martin.v.loewis | 2006-06-17 11:15:14 +0200 (Sat, 17 Jun 2006) | 3 lines Patch #763580: Add name and value arguments to Tkinter variable classes. ........ r46999 | martin.v.loewis | 2006-06-17 11:20:41 +0200 (Sat, 17 Jun 2006) | 2 lines Patch #1096231: Add default argument to wm_iconbitmap. ........ r47000 | martin.v.loewis | 2006-06-17 11:25:15 +0200 (Sat, 17 Jun 2006) | 2 lines Patch #1494750: Destroy master after deleting children. ........ r47003 | george.yoshida | 2006-06-17 18:31:52 +0200 (Sat, 17 Jun 2006) | 2 lines markup fix ........ r47005 | george.yoshida | 2006-06-17 18:39:13 +0200 (Sat, 17 Jun 2006) | 4 lines Update url. Old url returned status code:301 Moved permanently. ........ r47007 | martin.v.loewis | 2006-06-17 20:44:27 +0200 (Sat, 17 Jun 2006) | 2 lines Patch #812986: Update the canvas even if not tracing. ........ r47008 | martin.v.loewis | 2006-06-17 21:03:26 +0200 (Sat, 17 Jun 2006) | 2 lines Patch #815924: Restore ability to pass type= and icon= ........ r47009 | neal.norwitz | 2006-06-18 00:37:45 +0200 (Sun, 18 Jun 2006) | 1 line Fix typo in docstring ........ r47010 | neal.norwitz | 2006-06-18 00:38:15 +0200 (Sun, 18 Jun 2006) | 1 line Fix memory leak reported by valgrind while running test_subprocess ........ r47011 | fred.drake | 2006-06-18 04:57:35 +0200 (Sun, 18 Jun 2006) | 1 line remove unnecessary markup ........ r47013 | neal.norwitz | 2006-06-18 21:35:01 +0200 (Sun, 18 Jun 2006) | 7 lines Prevent spurious leaks when running regrtest.py -R. There may be more issues that crop up from time to time, but this change seems to have been pretty stable (no spurious warnings) for about a week. Other modules which use threads may require similar use of threading_setup/threading_cleanup from test_support. ........ r47014 | neal.norwitz | 2006-06-18 21:37:40 +0200 (Sun, 18 Jun 2006) | 9 lines The hppa ubuntu box sometimes hangs forever in these tests. My guess is that the wait is failing for some reason. Use WNOHANG, so we won't wait until the buildbot kills the test suite. I haven't been able to reproduce the failure, so I'm not sure if this will help or not. Hopefully, this change will cause the test to fail, rather than hang. That will be better since we will get the rest of the test results. It may also help us debug the real problem. ........ r47015 | neal.norwitz | 2006-06-18 22:10:24 +0200 (Sun, 18 Jun 2006) | 1 line Revert 47014 until it is more robust ........ r47016 | thomas.heller | 2006-06-18 23:27:04 +0200 (Sun, 18 Jun 2006) | 6 lines Fix typos. Fix doctest example. Mention in the tutorial that 'errcheck' is explained in the ref manual. Use better wording in some places. Remoce code examples that shouldn't be in the tutorial. Remove some XXX notices. ........ r47017 | georg.brandl | 2006-06-19 00:17:29 +0200 (Mon, 19 Jun 2006) | 3 lines Patch #1507676: improve exception messages in abstract.c, object.c and typeobject.c. ........ r47018 | neal.norwitz | 2006-06-19 07:40:44 +0200 (Mon, 19 Jun 2006) | 1 line Use Py_ssize_t ........ r47019 | georg.brandl | 2006-06-19 08:35:54 +0200 (Mon, 19 Jun 2006) | 3 lines Add news entry about error msg improvement. ........ r47020 | thomas.heller | 2006-06-19 09:07:49 +0200 (Mon, 19 Jun 2006) | 2 lines Try to repair the failing test on the OpenBSD buildbot. Trial and error... ........ r47021 | tim.peters | 2006-06-19 09:45:16 +0200 (Mon, 19 Jun 2006) | 2 lines Whitespace normalization. ........ r47022 | walter.doerwald | 2006-06-19 10:07:50 +0200 (Mon, 19 Jun 2006) | 4 lines Patch #1506645: add Python wrappers for the curses functions is_term_resized, resize_term and resizeterm. This uses three separate configure checks (one for each function). ........ r47023 | walter.doerwald | 2006-06-19 10:14:09 +0200 (Mon, 19 Jun 2006) | 2 lines Make check order match in configure and configure.in. ........ r47024 | tim.peters | 2006-06-19 10:14:28 +0200 (Mon, 19 Jun 2006) | 3 lines Repair KeyError when running test_threaded_import under -R, as reported by Neal on python-dev. ........ r47025 | thomas.heller | 2006-06-19 10:32:46 +0200 (Mon, 19 Jun 2006) | 3 lines Next try to fix the OpenBSD buildbot tests: Use ctypes.util.find_library to locate the C runtime library on platforms where is returns useful results. ........ r47026 | tim.peters | 2006-06-19 11:09:44 +0200 (Mon, 19 Jun 2006) | 13 lines TestHelp.make_parser(): This was making a permanent change to os.environ (setting envar COLUMNS), which at least caused test_float_default() to fail if the tests were run more than once. This repairs the test_optparse -R failures Neal reported on python-dev. It also explains some seemingly bizarre test_optparse failures we saw a couple weeks ago on the buildbots, when test_optparse failed due to test_file failing to clean up after itself, and then test_optparse failed in an entirely different way when regrtest's -w option ran test_optparse a second time. It's now obvious that make_parser() permanently changing os.environ was responsible for the second half of that. ........ r47027 | anthony.baxter | 2006-06-19 14:04:15 +0200 (Mon, 19 Jun 2006) | 2 lines Preparing for 2.5b1. ........ r47029 | fred.drake | 2006-06-19 19:31:16 +0200 (Mon, 19 Jun 2006) | 1 line remove non-working document formats from edist ........ r47030 | gerhard.haering | 2006-06-19 23:17:35 +0200 (Mon, 19 Jun 2006) | 5 lines Fixed a memory leak that was introduced with incorrect usage of the Python weak reference API in pysqlite 2.2.1. Bumbed pysqlite version number to upcoming pysqlite 2.3.1 release. ........ r47032 | ka-ping.yee | 2006-06-20 00:49:36 +0200 (Tue, 20 Jun 2006) | 2 lines Remove Python 2.3 compatibility comment. ........ r47033 | trent.mick | 2006-06-20 01:21:25 +0200 (Tue, 20 Jun 2006) | 2 lines Upgrade pyexpat to expat 2.0.0 (http://python.org/sf/1462338). ........ r47034 | trent.mick | 2006-06-20 01:57:41 +0200 (Tue, 20 Jun 2006) | 3 lines [ 1295808 ] expat symbols should be namespaced in pyexpat (http://python.org/sf/1295808) ........ r47039 | andrew.kuchling | 2006-06-20 13:52:16 +0200 (Tue, 20 Jun 2006) | 1 line Uncomment wsgiref section ........ r47040 | andrew.kuchling | 2006-06-20 14:15:09 +0200 (Tue, 20 Jun 2006) | 1 line Add four library items ........ r47041 | andrew.kuchling | 2006-06-20 14:19:54 +0200 (Tue, 20 Jun 2006) | 1 line Terminology and typography fixes ........ r47042 | andrew.kuchling | 2006-06-20 15:05:12 +0200 (Tue, 20 Jun 2006) | 1 line Add introductory paragraphs summarizing the release; minor edits ........ r47043 | andrew.kuchling | 2006-06-20 15:11:29 +0200 (Tue, 20 Jun 2006) | 1 line Minor edits and rearrangements; markup fix ........ r47044 | andrew.kuchling | 2006-06-20 15:20:30 +0200 (Tue, 20 Jun 2006) | 1 line [Bug #1504456] Mention xml -> xmlcore change ........ r47047 | brett.cannon | 2006-06-20 19:30:26 +0200 (Tue, 20 Jun 2006) | 2 lines Raise TestSkipped when the test socket connection is refused. ........ r47049 | brett.cannon | 2006-06-20 21:20:17 +0200 (Tue, 20 Jun 2006) | 2 lines Fix typo of exception name. ........ r47053 | brett.cannon | 2006-06-21 18:57:57 +0200 (Wed, 21 Jun 2006) | 5 lines At the C level, tuple arguments are passed in directly to the exception constructor, meaning it is treated as *args, not as a single argument. This means using the 'message' attribute won't work (until Py3K comes around), and so one must grab from 'arg' to get the error number. ........ r47054 | andrew.kuchling | 2006-06-21 19:10:18 +0200 (Wed, 21 Jun 2006) | 1 line Link to LibRef module documentation ........ r47055 | andrew.kuchling | 2006-06-21 19:17:10 +0200 (Wed, 21 Jun 2006) | 1 line Note some of Barry's work ........ r47056 | andrew.kuchling | 2006-06-21 19:17:28 +0200 (Wed, 21 Jun 2006) | 1 line Bump version ........ r47057 | georg.brandl | 2006-06-21 19:45:17 +0200 (Wed, 21 Jun 2006) | 3 lines fix [ 1509132 ] compiler module builds incorrect AST for TryExceptFinally ........ r47058 | georg.brandl | 2006-06-21 19:52:36 +0200 (Wed, 21 Jun 2006) | 3 lines Make test_fcntl aware of netbsd3. ........ r47059 | georg.brandl | 2006-06-21 19:53:17 +0200 (Wed, 21 Jun 2006) | 3 lines Patch #1509001: expected skips for netbsd3. ........ r47060 | gerhard.haering | 2006-06-21 22:55:04 +0200 (Wed, 21 Jun 2006) | 2 lines Removed call to enable_callback_tracebacks that slipped in by accident. ........ r47061 | armin.rigo | 2006-06-21 23:58:50 +0200 (Wed, 21 Jun 2006) | 13 lines Fix for an obscure bug introduced by revs 46806 and 46808, with a test. The problem of checking too eagerly for recursive calls is the following: if a RuntimeError is caused by recursion, and if code needs to normalize it immediately (as in the 2nd test), then PyErr_NormalizeException() needs a call to the RuntimeError class to instantiate it, and this hits the recursion limit again... causing PyErr_NormalizeException() to never finish. Moved this particular recursion check to slot_tp_call(), which is not involved in instantiating built-in exceptions. Backport candidate. ........ r47064 | neal.norwitz | 2006-06-22 08:30:50 +0200 (Thu, 22 Jun 2006) | 3 lines Copy the wsgiref package during make install. ........ r47065 | neal.norwitz | 2006-06-22 08:35:30 +0200 (Thu, 22 Jun 2006) | 1 line Reset the doc date to today for the automatic doc builds ........ r47067 | andrew.kuchling | 2006-06-22 15:10:23 +0200 (Thu, 22 Jun 2006) | 1 line Mention how to suppress warnings ........ r47069 | georg.brandl | 2006-06-22 16:46:17 +0200 (Thu, 22 Jun 2006) | 3 lines Set lineno correctly on list, tuple and dict literals. ........ r47070 | georg.brandl | 2006-06-22 16:46:46 +0200 (Thu, 22 Jun 2006) | 4 lines Test for correct compilation of try-except-finally stmt. Test for correct lineno on list, tuple, dict literals. ........ r47071 | fred.drake | 2006-06-22 17:50:08 +0200 (Thu, 22 Jun 2006) | 1 line fix markup nit ........ r47072 | brett.cannon | 2006-06-22 18:49:14 +0200 (Thu, 22 Jun 2006) | 6 lines 'warning's was improperly requiring that a command-line Warning category be both a subclass of Warning and a subclass of types.ClassType. The latter is no longer true thanks to new-style exceptions. Closes bug #1510580. Thanks to AMK for the test. ........ r47073 | ronald.oussoren | 2006-06-22 20:33:54 +0200 (Thu, 22 Jun 2006) | 3 lines MacOSX: Add a message to the first screen of the installer that tells users how to avoid updates to their shell profile. ........ r47074 | georg.brandl | 2006-06-22 21:02:18 +0200 (Thu, 22 Jun 2006) | 3 lines Fix my name ;) ........ r47075 | thomas.heller | 2006-06-22 21:07:36 +0200 (Thu, 22 Jun 2006) | 2 lines Small fixes, mostly in the markup. ........ r47076 | peter.astrand | 2006-06-22 22:06:46 +0200 (Thu, 22 Jun 2006) | 1 line Make it possible to run test_subprocess.py on Python 2.2, which lacks test_support.is_resource_enabled. ........ r47077 | peter.astrand | 2006-06-22 22:21:26 +0200 (Thu, 22 Jun 2006) | 1 line Applied patch #1506758: Prevent MemoryErrors with large MAXFD. ........ r47079 | neal.norwitz | 2006-06-23 05:32:44 +0200 (Fri, 23 Jun 2006) | 1 line Fix refleak ........ r47080 | fred.drake | 2006-06-23 08:03:45 +0200 (Fri, 23 Jun 2006) | 9 lines - SF bug #853506: IP6 address parsing in sgmllib ('[' and ']' were not accepted in unquoted attribute values) - cleaned up tests of character and entity reference decoding so the tests cover the documented relationships among handle_charref, handle_entityref, convert_charref, convert_codepoint, and convert_entityref, without bringing up Unicode issues that sgmllib cannot be involved in ........ r47085 | andrew.kuchling | 2006-06-23 21:23:40 +0200 (Fri, 23 Jun 2006) | 11 lines Fit Makefile for the Python doc environment better; this is a step toward including the howtos in the build process. * Put LaTeX output in ../paper-<whatever>/. * Put HTML output in ../html/ * Explain some of the Makefile variables * Remove some cruft dating to my environment (e.g. the 'web' target) This makefile isn't currently invoked by the documentation build process, so these changes won't destabilize anything. ........ r47086 | hyeshik.chang | 2006-06-23 23:16:18 +0200 (Fri, 23 Jun 2006) | 5 lines Bug #1511381: codec_getstreamcodec() in codec.c is corrected to omit a default "error" argument for NULL pointer. This allows the parser to take a codec from cjkcodecs again. (Reported by Taewook Kang and reviewed by Walter Doerwald) ........ r47091 | ronald.oussoren | 2006-06-25 22:44:16 +0200 (Sun, 25 Jun 2006) | 6 lines Workaround for bug #1512124 Without this patch IDLE will get unresponsive when you open the debugger window on OSX. This is both using the system Tcl/Tk on Tiger as the latest universal download from tk-components.sf.net. ........ r47092 | ronald.oussoren | 2006-06-25 23:14:19 +0200 (Sun, 25 Jun 2006) | 3 lines Drop the calldll demo's for macos, calldll isn't present anymore, no need to keep the demo's around. ........ r47093 | ronald.oussoren | 2006-06-25 23:15:58 +0200 (Sun, 25 Jun 2006) | 3 lines Use a path without a double slash to compile the .py files after installation (macosx, binary installer). This fixes bug #1508369 for python 2.5. ........ r47094 | ronald.oussoren | 2006-06-25 23:19:06 +0200 (Sun, 25 Jun 2006) | 3 lines Also install the .egg-info files in Lib. This will cause wsgiref.egg-info to be installed. ........ r47097 | andrew.kuchling | 2006-06-26 14:40:02 +0200 (Mon, 26 Jun 2006) | 1 line [Bug #1511998] Various comments from Nick Coghlan; thanks! ........ r47098 | andrew.kuchling | 2006-06-26 14:43:43 +0200 (Mon, 26 Jun 2006) | 1 line Describe workaround for PyRange_New()'s removal ........ r47099 | andrew.kuchling | 2006-06-26 15:08:24 +0200 (Mon, 26 Jun 2006) | 5 lines [Bug #1512163] Fix typo. This change will probably break tests on FreeBSD buildbots, but I'll check in a fix for that next. ........ r47100 | andrew.kuchling | 2006-06-26 15:12:16 +0200 (Mon, 26 Jun 2006) | 9 lines [Bug #1512163] Use one set of locking methods, lockf(); remove the flock() calls. On FreeBSD, the two methods lockf() and flock() end up using the same mechanism and the second one fails. A Linux man page claims that the two methods are orthogonal (so locks acquired one way don't interact with locks acquired the other way) but that clearly must be false. ........ r47101 | andrew.kuchling | 2006-06-26 15:23:10 +0200 (Mon, 26 Jun 2006) | 5 lines Add a test for a conflicting lock. On slow machines, maybe the time intervals (2 sec, 0.5 sec) will be too tight. I'll see how the buildbots like it. ........ r47103 | andrew.kuchling | 2006-06-26 16:33:24 +0200 (Mon, 26 Jun 2006) | 1 line Windows doesn't have os.fork(). I'll just disable this test for now ........ r47106 | andrew.kuchling | 2006-06-26 19:00:35 +0200 (Mon, 26 Jun 2006) | 9 lines Attempt to fix build failure on OS X and Debian alpha; the symptom is consistent with os.wait() returning immediately because some other subprocess had previously exited; the test suite then immediately tries to lock the mailbox and gets an error saying it's already locked. To fix this, do a waitpid() so the test suite only continues once the intended child process has exited. ........ r47113 | neal.norwitz | 2006-06-27 06:06:46 +0200 (Tue, 27 Jun 2006) | 1 line Ignore some more warnings in the dynamic linker on an older gentoo ........ r47114 | neal.norwitz | 2006-06-27 06:09:13 +0200 (Tue, 27 Jun 2006) | 6 lines Instead of doing a make test, run the regression tests out of the installed copy. This will hopefully catch problems where directories are added under Lib/ but not to Makefile.pre.in. This breaks out the 2 runs of the test suite with and without -O which is also nicer. ........ r47115 | neal.norwitz | 2006-06-27 06:12:58 +0200 (Tue, 27 Jun 2006) | 5 lines Fix SF bug #1513032, 'make install' failure on FreeBSD 5.3. No need to install lib-old, it's empty in 2.5. ........ r47116 | neal.norwitz | 2006-06-27 06:23:06 +0200 (Tue, 27 Jun 2006) | 1 line Test unimportant change to verify buildbot does not try to build ........ r47117 | neal.norwitz | 2006-06-27 06:26:30 +0200 (Tue, 27 Jun 2006) | 1 line Try again: test unimportant change to verify buildbot does not try to build ........ r47118 | neal.norwitz | 2006-06-27 06:28:56 +0200 (Tue, 27 Jun 2006) | 1 line Verify buildbot picks up these changes (really needs testing after last change to Makefile.pre.in) ........ r47121 | vinay.sajip | 2006-06-27 09:34:37 +0200 (Tue, 27 Jun 2006) | 1 line Removed buggy exception handling in doRollover of rotating file handlers. Exceptions now propagate to caller. ........ r47123 | ronald.oussoren | 2006-06-27 12:08:25 +0200 (Tue, 27 Jun 2006) | 3 lines MacOSX: fix rather dumb buglet that made it impossible to create extensions on OSX 10.3 when using a binary distribution build on 10.4. ........ r47125 | tim.peters | 2006-06-27 13:52:49 +0200 (Tue, 27 Jun 2006) | 2 lines Whitespace normalization. ........ r47128 | ronald.oussoren | 2006-06-27 14:53:52 +0200 (Tue, 27 Jun 2006) | 8 lines Use staticly build copies of zlib and bzip2 to build the OSX installer, that way the resulting binaries have a better change of running on 10.3. This patch also updates the search logic for sleepycat db3/4, without this patch you cannot use a sleepycat build with a non-standard prefix; with this you can (at least on OSX) if you add the prefix to CPPFLAGS/LDFLAGS at configure-time. This change is needed to build the binary installer for OSX. ........ r47131 | ronald.oussoren | 2006-06-27 17:45:32 +0200 (Tue, 27 Jun 2006) | 5 lines macosx: Install a libpython2.5.a inside the framework as a symlink to the actual dylib at the root of the framework, that way tools that expect a unix-like install (python-config, but more importantly external products like mod_python) work correctly. ........ r47137 | neal.norwitz | 2006-06-28 07:03:22 +0200 (Wed, 28 Jun 2006) | 4 lines According to the man pages on Gentoo Linux and Tru64, EACCES or EAGAIN can be returned if fcntl (lockf) fails. This fixes the test failure on Tru64 by checking for either error rather than just EAGAIN. ........ r47139 | neal.norwitz | 2006-06-28 08:28:31 +0200 (Wed, 28 Jun 2006) | 5 lines Fix bug #1512695: cPickle.loads could crash if it was interrupted with a KeyboardInterrupt since PyTuple_Pack was passed a NULL. Will backport. ........ r47142 | nick.coghlan | 2006-06-28 12:41:47 +0200 (Wed, 28 Jun 2006) | 1 line Make full module name available as __module_name__ even when __name__ is set to something else (like '__main__') ........ r47143 | armin.rigo | 2006-06-28 12:49:51 +0200 (Wed, 28 Jun 2006) | 2 lines A couple of crashers of the "won't fix" kind. ........ r47147 | andrew.kuchling | 2006-06-28 16:25:20 +0200 (Wed, 28 Jun 2006) | 1 line [Bug #1508766] Add docs for uuid module; docs written by George Yoshida, with minor rearrangements by me. ........ r47148 | andrew.kuchling | 2006-06-28 16:27:21 +0200 (Wed, 28 Jun 2006) | 1 line [Bug #1508766] Add docs for uuid module; this puts the module in the 'Internet Protocols' section. Arguably this module could also have gone in the chapters on strings or encodings, maybe even the crypto chapter. Fred, please move if you see fit. ........ r47151 | georg.brandl | 2006-06-28 22:23:25 +0200 (Wed, 28 Jun 2006) | 3 lines Fix end_fill(). ........ r47153 | trent.mick | 2006-06-28 22:30:41 +0200 (Wed, 28 Jun 2006) | 2 lines Mention the expat upgrade and pyexpat fix I put in 2.5b1. ........ r47154 | fred.drake | 2006-06-29 02:51:53 +0200 (Thu, 29 Jun 2006) | 6 lines SF bug #1504333: sgmlib should allow angle brackets in quoted values (modified patch by Sam Ruby; changed to use separate REs for start and end tags to reduce matching cost for end tags; extended tests; updated to avoid breaking previous changes to support IPv6 addresses in unquoted attribute values) ........ r47156 | fred.drake | 2006-06-29 04:57:48 +0200 (Thu, 29 Jun 2006) | 1 line document recent bugfixes in sgmllib ........ r47158 | neal.norwitz | 2006-06-29 06:10:08 +0200 (Thu, 29 Jun 2006) | 10 lines Add new utility function, reap_children(), to test_support. This should be called at the end of each test that spawns children (perhaps it should be called from regrtest instead?). This will hopefully prevent some of the unexplained failures in the buildbots (hppa and alpha) during tests that spawn children. The problems were not reproducible. There were many zombies that remained at the end of several tests. In the worst case, this shouldn't cause any more problems, though it may not help either. Time will tell. ........ r47159 | neal.norwitz | 2006-06-29 07:48:14 +0200 (Thu, 29 Jun 2006) | 5 lines This should fix the buildbot failure on s/390 which can't connect to gmail.org. It makes the error message consistent and always sends to stderr. It would be much better for all the networking tests to hit only python.org. ........ r47161 | thomas.heller | 2006-06-29 20:34:15 +0200 (Thu, 29 Jun 2006) | 3 lines Protect the thread api calls in the _ctypes extension module within #ifdef WITH_THREADS/#endif blocks. Found by Sam Rushing. ........ r47162 | martin.v.loewis | 2006-06-29 20:58:44 +0200 (Thu, 29 Jun 2006) | 2 lines Patch #1509163: MS Toolkit Compiler no longer available ........ r47163 | skip.montanaro | 2006-06-29 21:20:09 +0200 (Thu, 29 Jun 2006) | 1 line add string methods to index ........ r47164 | vinay.sajip | 2006-06-30 02:13:08 +0200 (Fri, 30 Jun 2006) | 1 line Fixed bug in fileConfig() which failed to clear logging._handlerList ........ r47166 | tim.peters | 2006-06-30 08:18:39 +0200 (Fri, 30 Jun 2006) | 2 lines Whitespace normalization. ........ r47170 | neal.norwitz | 2006-06-30 09:32:16 +0200 (Fri, 30 Jun 2006) | 1 line Silence compiler warning ........ r47171 | neal.norwitz | 2006-06-30 09:32:46 +0200 (Fri, 30 Jun 2006) | 1 line Another problem reported by Coverity. Backport candidate. ........ r47175 | thomas.heller | 2006-06-30 19:44:54 +0200 (Fri, 30 Jun 2006) | 2 lines Revert the use of PY_FORMAT_SIZE_T in PyErr_Format. ........ r47176 | tim.peters | 2006-06-30 20:34:51 +0200 (Fri, 30 Jun 2006) | 2 lines Remove now-unused fidding with PY_FORMAT_SIZE_T. ........ r47177 | georg.brandl | 2006-06-30 20:47:56 +0200 (Fri, 30 Jun 2006) | 3 lines Document decorator usage of property. ........ r47181 | fred.drake | 2006-06-30 21:29:25 +0200 (Fri, 30 Jun 2006) | 4 lines - consistency nit: always include "()" in \function and \method (*should* be done by the presentation, but that requires changes all over) - avoid spreading the __name meme ........ r47188 | vinay.sajip | 2006-07-01 12:45:20 +0200 (Sat, 01 Jul 2006) | 1 line Added entry for fileConfig() bugfix. ........ r47189 | vinay.sajip | 2006-07-01 12:47:20 +0200 (Sat, 01 Jul 2006) | 1 line Added duplicate call to fileConfig() to ensure that it cleans up after itself correctly. ........ r47190 | martin.v.loewis | 2006-07-01 17:33:37 +0200 (Sat, 01 Jul 2006) | 2 lines Release all forwarded functions in .close. Fixes #1513223. ........ r47191 | fred.drake | 2006-07-01 18:28:20 +0200 (Sat, 01 Jul 2006) | 7 lines SF bug #1296433 (Expat bug #1515266): Unchecked calls to character data handler would cause a segfault. This merges in Expat's lib/xmlparse.c revisions 1.154 and 1.155, which fix this and a closely related problem (the later does not affect Python). Moved the crasher test to the tests for xml.parsers.expat. ........ r47197 | gerhard.haering | 2006-07-02 19:48:30 +0200 (Sun, 02 Jul 2006) | 4 lines The sqlite3 module did cut off data from the SQLite database at the first null character before sending it to a custom converter. This has been fixed now. ........ r47198 | martin.v.loewis | 2006-07-02 20:44:00 +0200 (Sun, 02 Jul 2006) | 1 line Correct arithmetic in access on Win32. Fixes #1513646. ........ r47203 | thomas.heller | 2006-07-03 09:58:09 +0200 (Mon, 03 Jul 2006) | 1 line Cleanup: Remove commented out code. ........ r47204 | thomas.heller | 2006-07-03 09:59:50 +0200 (Mon, 03 Jul 2006) | 1 line Don't run the doctests with Python 2.3 because it doesn't have the ELLIPSIS flag. ........ r47205 | thomas.heller | 2006-07-03 10:04:05 +0200 (Mon, 03 Jul 2006) | 7 lines Fixes so that _ctypes can be compiled with the MingW compiler. It seems that the definition of '__attribute__(x)' was responsible for the compiler ignoring the '__fastcall' attribute on the ffi_closure_SYSV function in libffi_msvc/ffi.c, took me quite some time to figure this out. ........ r47206 | thomas.heller | 2006-07-03 10:08:14 +0200 (Mon, 03 Jul 2006) | 11 lines Add a new function uses_seh() to the _ctypes extension module. This will return True if Windows Structured Exception handling (SEH) is used when calling functions, False otherwise. Currently, only MSVC supports SEH. Fix the test so that it doesn't crash when run with MingW compiled _ctypes. Note that two tests are still failing when mingw is used, I suspect structure layout differences and function calling conventions between MSVC and MingW. ........ r47207 | tim.peters | 2006-07-03 10:23:19 +0200 (Mon, 03 Jul 2006) | 2 lines Whitespace normalization. ........ r47208 | martin.v.loewis | 2006-07-03 11:44:00 +0200 (Mon, 03 Jul 2006) | 3 lines Only setup canvas when it is first created. Fixes #1514703 ........ r47209 | martin.v.loewis | 2006-07-03 12:05:30 +0200 (Mon, 03 Jul 2006) | 3 lines Reimplement turtle.circle using a polyline, to allow correct filling of arcs. Also fixes #1514693. ........ r47210 | martin.v.loewis | 2006-07-03 12:19:49 +0200 (Mon, 03 Jul 2006) | 3 lines Bug #1514693: Update turtle's heading when switching between degrees and radians. ........ r47211 | martin.v.loewis | 2006-07-03 13:12:06 +0200 (Mon, 03 Jul 2006) | 2 lines Document functions added in 2.3 and 2.5. ........ r47212 | martin.v.loewis | 2006-07-03 14:19:50 +0200 (Mon, 03 Jul 2006) | 3 lines Bug #1417699: Reject locale-specific decimal point in float() and atof(). ........ r47213 | martin.v.loewis | 2006-07-03 14:28:58 +0200 (Mon, 03 Jul 2006) | 3 lines Bug #1267547: Put proper recursive setup.py call into the spec file generated by bdist_rpm. ........ r47215 | martin.v.loewis | 2006-07-03 15:01:35 +0200 (Mon, 03 Jul 2006) | 3 lines Patch #825417: Fix timeout processing in expect, read_until. Will backport to 2.4. ........ r47218 | martin.v.loewis | 2006-07-03 15:47:40 +0200 (Mon, 03 Jul 2006) | 2 lines Put method-wrappers into trashcan. Fixes #927248. ........ r47219 | andrew.kuchling | 2006-07-03 16:07:30 +0200 (Mon, 03 Jul 2006) | 1 line [Bug #1515932] Clarify description of slice assignment ........ r47220 | andrew.kuchling | 2006-07-03 16:16:09 +0200 (Mon, 03 Jul 2006) | 4 lines [Bug #1511911] Clarify description of optional arguments to sorted() by improving the xref to the section on lists, and by copying the explanations of the arguments (with a slight modification). ........ r47223 | kristjan.jonsson | 2006-07-03 16:59:05 +0200 (Mon, 03 Jul 2006) | 1 line Fix build problems with the platform SDK on windows. It is not sufficient to test for the C compiler version when determining if we have the secure CRT from microsoft. Must test with an undocumented macro, __STDC_SECURE_LIB__ too. ........ r47224 | ronald.oussoren | 2006-07-04 14:30:22 +0200 (Tue, 04 Jul 2006) | 7 lines Sync the darwin/x86 port libffi with the copy in PyObjC. This fixes a number of bugs in that port. The most annoying ones were due to some subtle differences between the document ABI and the actual implementation :-( (there are no python unittests that fail without this patch, but without it some of libffi's unittests fail). ........ r47234 | georg.brandl | 2006-07-05 10:21:00 +0200 (Wed, 05 Jul 2006) | 3 lines Remove remaining references to OverflowWarning. ........ r47236 | thomas.heller | 2006-07-05 11:13:56 +0200 (Wed, 05 Jul 2006) | 3 lines Fix the bitfield test when _ctypes is compiled with MingW. Structures containing bitfields may have different layout on MSVC and MingW . ........ r47237 | thomas.wouters | 2006-07-05 13:03:49 +0200 (Wed, 05 Jul 2006) | 15 lines Fix bug in passing tuples to string.Template. All other values (with working str() or repr()) would work, just not multi-value tuples. Probably not a backport candidate, since it changes the behaviour of passing a single-element tuple: >>> string.Template("$foo").substitute(dict(foo=(1,))) '(1,)' versus '1' ........ r47241 | georg.brandl | 2006-07-05 16:18:45 +0200 (Wed, 05 Jul 2006) | 2 lines Patch #1517490: fix glitches in filter() docs. ........ r47244 | georg.brandl | 2006-07-05 17:50:05 +0200 (Wed, 05 Jul 2006) | 2 lines no need to elaborate "string". ........ r47251 | neal.norwitz | 2006-07-06 06:28:59 +0200 (Thu, 06 Jul 2006) | 3 lines Fix refleaks reported by Shane Hathaway in SF patch #1515361. This change contains only the changes related to leaking the copy variable. ........ r47253 | fred.drake | 2006-07-06 07:13:22 +0200 (Thu, 06 Jul 2006) | 4 lines - back out Expat change; the final fix to Expat will be different - change the pyexpat wrapper to not be so sensitive to this detail of the Expat implementation (the ex-crasher test still passes) ........ r47257 | neal.norwitz | 2006-07-06 08:45:08 +0200 (Thu, 06 Jul 2006) | 1 line Add a NEWS entry for a recent pyexpat fix ........ r47258 | martin.v.loewis | 2006-07-06 08:55:58 +0200 (Thu, 06 Jul 2006) | 2 lines Add sqlite3.dll to the DLLs component, not to the TkDLLs component. Fixes #1517388. ........ r47259 | martin.v.loewis | 2006-07-06 09:05:21 +0200 (Thu, 06 Jul 2006) | 1 line Properly quote compileall and Lib paths in case TARGETDIR has a space. ........ r47260 | thomas.heller | 2006-07-06 09:50:18 +0200 (Thu, 06 Jul 2006) | 5 lines Revert the change done in svn revision 47206: Add a new function uses_seh() to the _ctypes extension module. This will return True if Windows Structured Exception handling (SEH) is used when calling functions, False otherwise. ........ r47261 | armin.rigo | 2006-07-06 09:58:18 +0200 (Thu, 06 Jul 2006) | 3 lines A couple of examples about how to attack the fact that _PyType_Lookup() returns a borrowed ref. Many of the calls are open to attack. ........ r47262 | thomas.heller | 2006-07-06 10:28:14 +0200 (Thu, 06 Jul 2006) | 2 lines The test that calls a function with invalid arguments and catches the resulting Windows access violation will not be run by default. ........ r47263 | thomas.heller | 2006-07-06 10:48:35 +0200 (Thu, 06 Jul 2006) | 5 lines Patch #1517790: It is now possible to use custom objects in the ctypes foreign function argtypes sequence as long as they provide a from_param method, no longer is it required that the object is a ctypes type. ........ r47264 | thomas.heller | 2006-07-06 10:58:40 +0200 (Thu, 06 Jul 2006) | 2 lines Document the Struture and Union constructors. ........ r47265 | thomas.heller | 2006-07-06 11:11:22 +0200 (Thu, 06 Jul 2006) | 2 lines Document the changes in svn revision 47263, from patch #1517790. ........ r47267 | ronald.oussoren | 2006-07-06 12:13:35 +0200 (Thu, 06 Jul 2006) | 7 lines This patch solves the problem Skip was seeing with zlib, this patch ensures that configure uses similar compiler flags as setup.py when doing the zlib test. Without this patch configure would use the first shared library on the linker path, with this patch it uses the first shared or static library on that path just like setup.py. ........ r47268 | thomas.wouters | 2006-07-06 12:48:28 +0200 (Thu, 06 Jul 2006) | 4 lines NEWS entry for r47267: fixing configure's zlib probing. ........ r47269 | fredrik.lundh | 2006-07-06 14:29:24 +0200 (Thu, 06 Jul 2006) | 3 lines added XMLParser alias for cElementTree compatibility ........ r47271 | nick.coghlan | 2006-07-06 14:53:04 +0200 (Thu, 06 Jul 2006) | 1 line Revert the __module_name__ changes made in rev 47142. We'll revisit this in Python 2.6 ........ r47272 | nick.coghlan | 2006-07-06 15:04:56 +0200 (Thu, 06 Jul 2006) | 1 line Update the tutorial section on relative imports ........ r47273 | nick.coghlan | 2006-07-06 15:35:27 +0200 (Thu, 06 Jul 2006) | 1 line Ignore ImportWarning by default ........ r47274 | nick.coghlan | 2006-07-06 15:41:34 +0200 (Thu, 06 Jul 2006) | 1 line Cover ImportWarning, PendingDeprecationWarning and simplefilter() in the warnings module docs ........ r47275 | nick.coghlan | 2006-07-06 15:47:18 +0200 (Thu, 06 Jul 2006) | 1 line Add NEWS entries for the ImportWarning change and documentation update ........ r47276 | andrew.kuchling | 2006-07-06 15:57:28 +0200 (Thu, 06 Jul 2006) | 1 line ImportWarning is now silent by default ........ r47277 | thomas.heller | 2006-07-06 17:06:05 +0200 (Thu, 06 Jul 2006) | 2 lines Document the correct return type of PyLong_AsUnsignedLongLongMask. ........ r47278 | hyeshik.chang | 2006-07-06 17:21:52 +0200 (Thu, 06 Jul 2006) | 2 lines Add a testcase for r47086 which fixed a bug in codec_getstreamcodec(). ........ r47279 | hyeshik.chang | 2006-07-06 17:39:24 +0200 (Thu, 06 Jul 2006) | 3 lines Test using all CJK encodings for the testcases which don't require specific encodings. ........ r47280 | martin.v.loewis | 2006-07-06 21:28:03 +0200 (Thu, 06 Jul 2006) | 2 lines Properly generate logical file ids. Fixes #1515998. Also correct typo in Control.mapping. ........ r47287 | neal.norwitz | 2006-07-07 08:03:15 +0200 (Fri, 07 Jul 2006) | 17 lines Restore rev 47014: The hppa ubuntu box sometimes hangs forever in these tests. My guess is that the wait is failing for some reason. Use WNOHANG, so we won't wait until the buildbot kills the test suite. I haven't been able to reproduce the failure, so I'm not sure if this will help or not. Hopefully, this change will cause the test to fail, rather than hang. That will be better since we will get the rest of the test results. It may also help us debug the real problem. *** The reason this originally failed was because there were many zombie children outstanding before rev 47158 cleaned them up. There are still hangs in test_subprocess that need to be addressed, but that will take more work. This should close some holes. ........ r47289 | georg.brandl | 2006-07-07 10:15:12 +0200 (Fri, 07 Jul 2006) | 3 lines Fix RFC number. ........ r50489 | neal.norwitz | 2006-07-08 07:31:37 +0200 (Sat, 08 Jul 2006) | 1 line Fix SF bug #1519018: 'as' is now validated properly in import statements ........ r50490 | georg.brandl | 2006-07-08 14:15:27 +0200 (Sat, 08 Jul 2006) | 3 lines Add an additional test for bug #1519018. ........ r50491 | tim.peters | 2006-07-08 21:55:05 +0200 (Sat, 08 Jul 2006) | 2 lines Whitespace normalization. ........ r50493 | neil.schemenauer | 2006-07-09 18:16:34 +0200 (Sun, 09 Jul 2006) | 2 lines Fix AST compiler bug #1501934: incorrect LOAD/STORE_GLOBAL generation. ........ r50495 | neil.schemenauer | 2006-07-09 23:19:29 +0200 (Sun, 09 Jul 2006) | 2 lines Fix SF bug 1441486: bad unary minus folding in compiler. ........ r50497 | neal.norwitz | 2006-07-10 00:14:42 +0200 (Mon, 10 Jul 2006) | 4 lines On 64 bit systems, int literals that use less than 64 bits are now ints rather than longs. This also fixes the test for eval(-sys.maxint - 1). ........ r50500 | neal.norwitz | 2006-07-10 02:04:44 +0200 (Mon, 10 Jul 2006) | 4 lines Bug #1512814, Fix incorrect lineno's when code at module scope started after line 256. ........ r50501 | neal.norwitz | 2006-07-10 02:05:34 +0200 (Mon, 10 Jul 2006) | 1 line Fix doco. Backport candidate. ........ r50503 | neal.norwitz | 2006-07-10 02:23:17 +0200 (Mon, 10 Jul 2006) | 5 lines Part of SF patch #1484695. This removes dead code. The chksum was already verified in .frombuf() on the lines above. If there was a problem an exception is raised, so there was no way this condition could have been true. ........ r50504 | neal.norwitz | 2006-07-10 03:18:57 +0200 (Mon, 10 Jul 2006) | 3 lines Patch #1516912: improve Modules support for OpenVMS. ........ r50506 | neal.norwitz | 2006-07-10 04:36:41 +0200 (Mon, 10 Jul 2006) | 7 lines Patch #1504046: Add documentation for xml.etree. /F wrote the text docs, Englebert Gruber massaged it to latex and I did some more massaging to try and improve the consistency and fix some name mismatches between the declaration and text. ........ r50509 | martin.v.loewis | 2006-07-10 09:23:48 +0200 (Mon, 10 Jul 2006) | 2 lines Introduce DISTUTILS_USE_SDK as a flag to determine whether the SDK environment should be used. Fixes #1508010. ........ r50510 | martin.v.loewis | 2006-07-10 09:26:41 +0200 (Mon, 10 Jul 2006) | 1 line Change error message to indicate that VS2003 is necessary to build extension modules, not the .NET SDK. ........ r50511 | martin.v.loewis | 2006-07-10 09:29:41 +0200 (Mon, 10 Jul 2006) | 1 line Add svn:ignore. ........ r50512 | anthony.baxter | 2006-07-10 09:41:04 +0200 (Mon, 10 Jul 2006) | 1 line preparing for 2.5b2 ........ r50513 | thomas.heller | 2006-07-10 11:10:28 +0200 (Mon, 10 Jul 2006) | 2 lines Fix bug #1518190: accept any integer or long value in the ctypes.c_void_p constructor. ........ r50514 | thomas.heller | 2006-07-10 11:31:06 +0200 (Mon, 10 Jul 2006) | 3 lines Fixed a segfault when ctypes.wintypes were imported on non-Windows machines. ........ r50516 | thomas.heller | 2006-07-10 13:11:10 +0200 (Mon, 10 Jul 2006) | 3 lines Assigning None to pointer type structure fields possible overwrote wrong fields. ........ r50517 | thomas.heller | 2006-07-10 13:17:37 +0200 (Mon, 10 Jul 2006) | 5 lines Moved the ctypes news entries from the 'Library' section into the 'Extension Modules' section where they belong, probably. This destroyes the original order of the news entries, don't know if that is important or not. ........ r50526 | phillip.eby | 2006-07-10 21:03:29 +0200 (Mon, 10 Jul 2006) | 2 lines Fix SF#1516184 and add a test to prevent regression. ........ r50528 | phillip.eby | 2006-07-10 21:18:35 +0200 (Mon, 10 Jul 2006) | 2 lines Fix SF#1457312: bad socket error handling in distutils "upload" command. ........ r50537 | peter.astrand | 2006-07-10 22:39:49 +0200 (Mon, 10 Jul 2006) | 1 line Make it possible to run test_subprocess.py with Python 2.2, which lacks test_support.reap_children(). ........ r50541 | tim.peters | 2006-07-10 23:08:24 +0200 (Mon, 10 Jul 2006) | 5 lines After approval from Anthony, merge the tim-current_frames branch into the trunk. This adds a new sys._current_frames() function, which returns a dict mapping thread id to topmost thread stack frame. ........ r50542 | tim.peters | 2006-07-10 23:11:49 +0200 (Mon, 10 Jul 2006) | 2 lines Whitespace normalization. ........ r50553 | martin.v.loewis | 2006-07-11 00:11:28 +0200 (Tue, 11 Jul 2006) | 4 lines Patch #1519566: Remove unused _tofill member. Make begin_fill idempotent. Update demo2 to demonstrate filling of concave shapes. ........ r50567 | anthony.baxter | 2006-07-11 04:04:09 +0200 (Tue, 11 Jul 2006) | 4 lines #1494314: Fix a regression with high-numbered sockets in 2.4.3. This means that select() on sockets > FD_SETSIZE (typically 1024) work again. The patch makes sockets use poll() internally where available. ........ r50568 | tim.peters | 2006-07-11 04:17:48 +0200 (Tue, 11 Jul 2006) | 2 lines Whitespace normalization. ........ r50575 | thomas.heller | 2006-07-11 18:42:05 +0200 (Tue, 11 Jul 2006) | 1 line Add missing Py_DECREF. ........ r50576 | thomas.heller | 2006-07-11 18:44:25 +0200 (Tue, 11 Jul 2006) | 1 line Add missing Py_DECREFs. ........ r50579 | andrew.kuchling | 2006-07-11 19:20:16 +0200 (Tue, 11 Jul 2006) | 1 line Bump version number; add sys._current_frames ........ r50582 | thomas.heller | 2006-07-11 20:28:35 +0200 (Tue, 11 Jul 2006) | 3 lines When a foreign function is retrived by calling __getitem__ on a ctypes library instance, do not set it as attribute. ........ r50583 | thomas.heller | 2006-07-11 20:40:50 +0200 (Tue, 11 Jul 2006) | 2 lines Change the ctypes version number to 1.0.0. ........ r50597 | neal.norwitz | 2006-07-12 07:26:17 +0200 (Wed, 12 Jul 2006) | 3 lines Bug #1520864: unpacking singleton tuples in for loop (for x, in) work again. ........ r50598 | neal.norwitz | 2006-07-12 07:26:35 +0200 (Wed, 12 Jul 2006) | 1 line Fix function name in error msg ........ r50599 | neal.norwitz | 2006-07-12 07:27:46 +0200 (Wed, 12 Jul 2006) | 4 lines Fix uninitialized memory read reported by Valgrind when running doctest. This could happen if size == 0. ........ r50600 | neal.norwitz | 2006-07-12 09:28:29 +0200 (Wed, 12 Jul 2006) | 1 line Actually change the MAGIC #. Create a new section for 2.5c1 and mention the impact of changing the MAGIC #. ........ r50601 | thomas.heller | 2006-07-12 10:43:47 +0200 (Wed, 12 Jul 2006) | 3 lines Fix #1467450: ctypes now uses RTLD_GLOBAL by default on OSX 10.3 to load shared libraries. ........ r50604 | thomas.heller | 2006-07-12 16:25:18 +0200 (Wed, 12 Jul 2006) | 3 lines Fix the wrong description of LibraryLoader.LoadLibrary, and document the DEFAULT_MODE constant. ........ r50607 | georg.brandl | 2006-07-12 17:31:17 +0200 (Wed, 12 Jul 2006) | 3 lines Accept long options "--help" and "--version". ........ r50617 | thomas.heller | 2006-07-13 11:53:47 +0200 (Thu, 13 Jul 2006) | 3 lines A misspelled preprocessor symbol caused ctypes to be always compiled without thread support. Replaced WITH_THREADS with WITH_THREAD. ........ r50619 | thomas.heller | 2006-07-13 19:01:14 +0200 (Thu, 13 Jul 2006) | 3 lines Fix #1521375. When running with root priviledges, 'gcc -o /dev/null' did overwrite /dev/null. Use a temporary file instead of /dev/null. ........ r50620 | thomas.heller | 2006-07-13 19:05:13 +0200 (Thu, 13 Jul 2006) | 2 lines Fix misleading words. ........ r50622 | andrew.kuchling | 2006-07-13 19:37:26 +0200 (Thu, 13 Jul 2006) | 1 line Typo fix ........ r50629 | georg.brandl | 2006-07-14 09:12:54 +0200 (Fri, 14 Jul 2006) | 3 lines Patch #1521874: grammar errors in doanddont.tex. ........ r50630 | neal.norwitz | 2006-07-14 09:20:04 +0200 (Fri, 14 Jul 2006) | 1 line Try to improve grammar further. ........ r50631 | martin.v.loewis | 2006-07-14 11:58:55 +0200 (Fri, 14 Jul 2006) | 1 line Extend build_ssl to Win64, using VSExtComp. ........ r50632 | martin.v.loewis | 2006-07-14 14:10:09 +0200 (Fri, 14 Jul 2006) | 1 line Add debug output to analyse buildbot failure. ........ r50633 | martin.v.loewis | 2006-07-14 14:31:05 +0200 (Fri, 14 Jul 2006) | 1 line Fix Debug build of _ssl. ........ r50636 | andrew.kuchling | 2006-07-14 15:32:38 +0200 (Fri, 14 Jul 2006) | 1 line Mention new options ........ r50638 | peter.astrand | 2006-07-14 16:04:45 +0200 (Fri, 14 Jul 2006) | 1 line Bug #1223937: CalledProcessError.errno -> CalledProcessError.returncode. ........ r50640 | thomas.heller | 2006-07-14 17:01:05 +0200 (Fri, 14 Jul 2006) | 4 lines Make the prototypes of our private PyUnicode_FromWideChar and PyUnicode_AsWideChar replacement functions compatible to the official functions by using Py_ssize_t instead of int. ........ r50643 | thomas.heller | 2006-07-14 19:51:14 +0200 (Fri, 14 Jul 2006) | 3 lines Patch #1521817: The index range checking on ctypes arrays containing exactly one element is enabled again. ........ r50647 | thomas.heller | 2006-07-14 20:22:50 +0200 (Fri, 14 Jul 2006) | 2 lines Updates for the ctypes documentation. ........ r50655 | fredrik.lundh | 2006-07-14 23:45:48 +0200 (Fri, 14 Jul 2006) | 3 lines typo ........ r50664 | george.yoshida | 2006-07-15 18:03:49 +0200 (Sat, 15 Jul 2006) | 2 lines Bug #15187702 : ext/win-cookbook.html has a broken link to distutils ........ r50667 | bob.ippolito | 2006-07-15 18:53:15 +0200 (Sat, 15 Jul 2006) | 1 line Patch #1220874: Update the binhex module for Mach-O. ........ r50671 | fred.drake | 2006-07-16 03:21:20 +0200 (Sun, 16 Jul 2006) | 1 line clean up some link markup ........ r50673 | neal.norwitz | 2006-07-16 03:50:38 +0200 (Sun, 16 Jul 2006) | 4 lines Bug #1512814, Fix incorrect lineno's when code within a function had more than 255 blank lines. Byte codes need to go first, line #s second. ........ r50674 | neal.norwitz | 2006-07-16 04:00:32 +0200 (Sun, 16 Jul 2006) | 5 lines a & b were dereffed above, so they are known to be valid pointers. z is known to be NULL, nothing to DECREF. Reported by Klockwork, #107. ........ r50675 | neal.norwitz | 2006-07-16 04:02:57 +0200 (Sun, 16 Jul 2006) | 5 lines self is dereffed (and passed as first arg), so it's known to be good. func is returned from PyArg_ParseTuple and also dereffed. Reported by Klocwork, #30 (self one at least). ........ r50676 | neal.norwitz | 2006-07-16 04:05:35 +0200 (Sun, 16 Jul 2006) | 4 lines proto was dereffed above and is known to be good. No need for X. Reported by Klocwork, #39. ........ r50677 | neal.norwitz | 2006-07-16 04:15:27 +0200 (Sun, 16 Jul 2006) | 5 lines Fix memory leaks in some conditions. Reported by Klocwork #152. ........ r50678 | neal.norwitz | 2006-07-16 04:17:36 +0200 (Sun, 16 Jul 2006) | 4 lines Fix memory leak under some conditions. Reported by Klocwork, #98. ........ r50679 | neal.norwitz | 2006-07-16 04:22:30 +0200 (Sun, 16 Jul 2006) | 8 lines Use sizeof(buffer) instead of duplicating the constants to ensure they won't be wrong. The real change is to pass (bufsz - 1) to PyOS_ascii_formatd and 1 to strncat. strncat copies n+1 bytes from src (not dest). Reported by Klocwork #58. ........ r50680 | neal.norwitz | 2006-07-16 04:32:03 +0200 (Sun, 16 Jul 2006) | 5 lines Handle a NULL name properly. Reported by Klocwork #67 ........ r50681 | neal.norwitz | 2006-07-16 04:35:47 +0200 (Sun, 16 Jul 2006) | 6 lines PyFunction_SetDefaults() is documented as taking None or a tuple. A NULL would crash the PyTuple_Check(). Now make NULL return a SystemError. Reported by Klocwork #73. ........ r50683 | neal.norwitz | 2006-07-17 02:55:45 +0200 (Mon, 17 Jul 2006) | 5 lines Stop INCREFing name, then checking if it's NULL. name (f_name) should never be NULL so assert it. Fix one place where we could have passed NULL. Reported by Klocwork #66. ........ r50684 | neal.norwitz | 2006-07-17 02:57:15 +0200 (Mon, 17 Jul 2006) | 5 lines otherset is known to be non-NULL based on checks before and DECREF after. DECREF otherset rather than XDECREF in error conditions too. Reported by Klockwork #154. ........ r50685 | neal.norwitz | 2006-07-17 02:59:04 +0200 (Mon, 17 Jul 2006) | 7 lines Reported by Klocwork #151. v2 can be NULL if exception2 is NULL. I don't think that condition can happen, but I'm not sure it can't either. Now the code will protect against either being NULL. ........ r50686 | neal.norwitz | 2006-07-17 03:00:16 +0200 (Mon, 17 Jul 2006) | 1 line Add NEWS entry for a bunch of fixes due to warnings produced by Klocworks static analysis tool. ........ r50687 | fred.drake | 2006-07-17 07:47:52 +0200 (Mon, 17 Jul 2006) | 3 lines document xmlcore (still minimal; needs mention in each of the xml.* modules) SF bug #1504456 (partial) ........ r50688 | georg.brandl | 2006-07-17 15:23:46 +0200 (Mon, 17 Jul 2006) | 3 lines Remove usage of sets module (patch #1500609). ........ r50689 | georg.brandl | 2006-07-17 15:26:33 +0200 (Mon, 17 Jul 2006) | 3 lines Add missing NEWS item (#1522771) ........ r50690 | andrew.kuchling | 2006-07-17 18:47:54 +0200 (Mon, 17 Jul 2006) | 1 line Attribute more features ........ r50692 | kurt.kaiser | 2006-07-17 23:59:27 +0200 (Mon, 17 Jul 2006) | 8 lines Patch 1479219 - Tal Einat 1. 'as' highlighted as builtin in comment string on import line 2. Comments such as "#False identity" which start with a keyword immediately after the '#' character aren't colored as comments. 3. u or U beginning unicode string not correctly highlighted Closes bug 1325071 ........ r50693 | barry.warsaw | 2006-07-18 01:07:51 +0200 (Tue, 18 Jul 2006) | 16 lines decode_rfc2231(): Be more robust against buggy RFC 2231 encodings. Specifically, instead of raising a ValueError when there is a single tick in the parameter, simply return that the entire string unquoted, with None for both the charset and the language. Also, if there are more than 2 ticks in the parameter, interpret the first three parts as the standard RFC 2231 parts, then the rest of the parts as the encoded string. Test cases added. Original fewer-than-3-parts fix by Tokio Kikuchi. Resolves SF bug # 1218081. I will back port the fix and tests to Python 2.4 (email 3.0) and Python 2.3 (email 2.5). Also, bump the version number to email 4.0.1, removing the 'alpha' moniker. ........ r50695 | kurt.kaiser | 2006-07-18 06:03:16 +0200 (Tue, 18 Jul 2006) | 2 lines Rebinding Tab key was inserting 'tab' instead of 'Tab'. Bug 1179168. ........ r50696 | brett.cannon | 2006-07-18 06:41:36 +0200 (Tue, 18 Jul 2006) | 6 lines Fix bug #1520914. Starting in 2.4, time.strftime() began to check the bounds of values in the time tuple passed in. Unfortunately people came to rely on undocumented behaviour of setting unneeded values to 0, regardless of if it was within the valid range. Now those values force the value internally to the minimum value when 0 is passed in. ........ r50697 | facundo.batista | 2006-07-18 14:16:13 +0200 (Tue, 18 Jul 2006) | 1 line Comments and docs cleanups, and some little fixes, provided by Santiágo Peresón ........ r50704 | martin.v.loewis | 2006-07-18 19:46:31 +0200 (Tue, 18 Jul 2006) | 2 lines Patch #1524429: Use repr instead of backticks again. ........ r50706 | tim.peters | 2006-07-18 23:55:15 +0200 (Tue, 18 Jul 2006) | 2 lines Whitespace normalization. ........ r50708 | tim.peters | 2006-07-19 02:03:19 +0200 (Wed, 19 Jul 2006) | 18 lines SF bug 1524317: configure --without-threads fails to build Moved the code for _PyThread_CurrentFrames() up, so it's no longer in a huge "#ifdef WITH_THREAD" block (I didn't realize it /was/ in one). Changed test_sys's test_current_frames() so it passes with or without thread supported compiled in. Note that test_sys fails when Python is compiled without threads, but for an unrelated reason (the old test_exit() fails with an indirect ImportError on the `thread` module). There are also other unrelated compilation failures without threads, in extension modules (like ctypes); at least the core compiles again. Do we really support --without-threads? If so, there are several problems remaining. ........ r50713 | thomas.heller | 2006-07-19 11:09:32 +0200 (Wed, 19 Jul 2006) | 4 lines Make sure the _ctypes extension can be compiled when WITH_THREAD is not defined on Windows, even if that configuration is probably not supported at all. ........ r50715 | martin.v.loewis | 2006-07-19 19:18:32 +0200 (Wed, 19 Jul 2006) | 4 lines Revert r50706 (Whitespace normalization) and r50697: Comments and docs cleanups, and some little fixes per recommendation from Raymond Hettinger. ........ r50719 | phillip.eby | 2006-07-20 17:54:16 +0200 (Thu, 20 Jul 2006) | 4 lines Fix SF#1516184 (again) and add a test to prevent regression. (There was a problem with empty filenames still causing recursion) ........ r50720 | georg.brandl | 2006-07-20 18:28:39 +0200 (Thu, 20 Jul 2006) | 3 lines Guard for _active being None in __del__ method. ........ r50721 | vinay.sajip | 2006-07-20 18:28:39 +0200 (Thu, 20 Jul 2006) | 1 line Updated documentation for TimedRotatingFileHandler relating to how rollover files are named. The previous documentation was wrongly the same as for RotatingFileHandler. ........ r50731 | fred.drake | 2006-07-20 22:11:57 +0200 (Thu, 20 Jul 2006) | 1 line markup fix ........ r50739 | kurt.kaiser | 2006-07-21 00:22:52 +0200 (Fri, 21 Jul 2006) | 7 lines Avoid occasional failure to detect closing paren properly. Patch 1407280 Tal Einat M ParenMatch.py M NEWS.txt M CREDITS.txt ........ r50740 | vinay.sajip | 2006-07-21 01:20:12 +0200 (Fri, 21 Jul 2006) | 1 line Addressed SF#1524081 by using a dictionary to map level names to syslog priority names, rather than a string.lower(). ........ r50741 | neal.norwitz | 2006-07-21 07:29:58 +0200 (Fri, 21 Jul 2006) | 1 line Add some asserts that we got good params passed ........ r50742 | neal.norwitz | 2006-07-21 07:31:02 +0200 (Fri, 21 Jul 2006) | 5 lines Move the initialization of some pointers earlier. The problem is that if we call Py_DECREF(frame) like we do if allocating locals fails, frame_dealloc() will try to use these bogus values and crash. ........ r50743 | neal.norwitz | 2006-07-21 07:32:28 +0200 (Fri, 21 Jul 2006) | 4 lines Handle allocation failures gracefully. Found with failmalloc. Many (all?) of these could be backported. ........ r50745 | neal.norwitz | 2006-07-21 09:59:02 +0200 (Fri, 21 Jul 2006) | 1 line Speel initialise write. Tanks Anthony. ........ r50746 | neal.norwitz | 2006-07-21 09:59:47 +0200 (Fri, 21 Jul 2006) | 2 lines Handle more memory allocation failures without crashing. ........ r50754 | barry.warsaw | 2006-07-21 16:51:07 +0200 (Fri, 21 Jul 2006) | 23 lines More RFC 2231 improvements for the email 4.0 package. As Mark Sapiro rightly points out there are really two types of continued headers defined in this RFC (i.e. "encoded" parameters with the form "name*0*=" and unencoded parameters with the form "name*0="), but we were were handling them both the same way and that isn't correct. This patch should be much more RFC compliant in that only encoded params are %-decoded and the charset/language information is only extract if there are any encoded params in the segments. If there are no encoded params then the RFC says that there will be no charset/language parts. Note however that this will change the return value for Message.get_param() in some cases. For example, whereas before if you had all unencoded param continuations you would have still gotten a 3-tuple back from this method (with charset and language == None), you will now get just a string. I don't believe this is a backward incompatible change though because the documentation for this method already indicates that either return value is possible and that you must do an isinstance(val, tuple) check to discriminate between the two. (Yeah that API kind of sucks but we can't change /that/ without breaking code.) Test cases, some documentation updates, and a NEWS item accompany this patch. ........ r50759 | georg.brandl | 2006-07-21 19:36:31 +0200 (Fri, 21 Jul 2006) | 3 lines Fix check for empty list (vs. None). ........ r50771 | brett.cannon | 2006-07-22 00:44:07 +0200 (Sat, 22 Jul 2006) | 2 lines Remove an XXX marker in a comment. ........ r50773 | neal.norwitz | 2006-07-22 18:20:49 +0200 (Sat, 22 Jul 2006) | 1 line Fix more memory allocation issues found with failmalloc. ........ r50774 | neal.norwitz | 2006-07-22 19:00:57 +0200 (Sat, 22 Jul 2006) | 1 line Don't fail if the directory already exists ........ r50775 | greg.ward | 2006-07-23 04:25:53 +0200 (Sun, 23 Jul 2006) | 6 lines Be a lot smarter about whether this test passes: instead of assuming that a 2.93 sec audio file will always take 3.1 sec (as it did on the hardware I had when I first wrote the test), expect that it will take 2.93 sec +/- 10%, and only fail if it's outside of that range. Compute the expected ........ r50776 | kurt.kaiser | 2006-07-23 06:19:49 +0200 (Sun, 23 Jul 2006) | 2 lines Tooltips failed on new-syle class __init__ args. Bug 1027566 Loren Guthrie ........ r50777 | neal.norwitz | 2006-07-23 09:50:36 +0200 (Sun, 23 Jul 2006) | 1 line Handle more mem alloc issues found with failmalloc ........ r50778 | neal.norwitz | 2006-07-23 09:51:58 +0200 (Sun, 23 Jul 2006) | 5 lines If the for loop isn't entered, entryblock will be NULL. If passed to stackdepth_walk it will be dereffed. Not sure if I found with failmalloc or Klockwork #55. ........ r50779 | neal.norwitz | 2006-07-23 09:53:14 +0200 (Sun, 23 Jul 2006) | 4 lines Move the initialization of size_a down below the check for a being NULL. Reported by Klocwork #106 ........ r50780 | neal.norwitz | 2006-07-23 09:55:55 +0200 (Sun, 23 Jul 2006) | 9 lines Check the allocation of b_objects and return if there was a failure. Also fix a few memory leaks in other failure scenarios. It seems that if b_objects == Py_None, we will have an extra ref to b_objects. Add XXX comment so hopefully someone documents why the else isn't necessary or adds it in. Reported by Klocwork #20 ........ r50781 | neal.norwitz | 2006-07-23 09:57:11 +0200 (Sun, 23 Jul 2006) | 2 lines Fix memory leaks spotted by Klocwork #37. ........ r50782 | neal.norwitz | 2006-07-23 09:59:00 +0200 (Sun, 23 Jul 2006) | 5 lines nextlink can be NULL if teedataobject_new fails, so use XINCREF. Ensure that dataobj is never NULL. Reported by Klocwork #102 ........ r50783 | neal.norwitz | 2006-07-23 10:01:43 +0200 (Sun, 23 Jul 2006) | 8 lines Ensure we don't write beyond errText. I think I got this right, but it definitely could use some review to ensure I'm not off by one and there's no possible overflow/wrap-around of bytes_left. Reported by Klocwork #1. Fix a problem if there is a failure allocating self->db. Found with failmalloc. ........ r50784 | ronald.oussoren | 2006-07-23 11:41:09 +0200 (Sun, 23 Jul 2006) | 3 lines Without this patch CMD-W won't close EditorWindows on MacOS X. This solves part of bug #1517990. ........ r50785 | ronald.oussoren | 2006-07-23 11:46:11 +0200 (Sun, 23 Jul 2006) | 5 lines Fix for bug #1517996: Class and Path browsers show Tk menu This patch replaces the menubar that is used by AquaTk for windows without a menubar of their own by one that is more appropriate for IDLE. ........ r50786 | andrew.macintyre | 2006-07-23 14:57:02 +0200 (Sun, 23 Jul 2006) | 2 lines Build updates for OS/2 EMX port ........ r50787 | andrew.macintyre | 2006-07-23 15:00:04 +0200 (Sun, 23 Jul 2006) | 3 lines bugfix: PyThread_start_new_thread() returns the thread ID, not a flag; will backport. ........ r50789 | andrew.macintyre | 2006-07-23 15:04:00 +0200 (Sun, 23 Jul 2006) | 2 lines Get mailbox module working on OS/2 EMX port. ........ r50791 | greg.ward | 2006-07-23 18:05:51 +0200 (Sun, 23 Jul 2006) | 1 line Resync optparse with Optik 1.5.3: minor tweaks for/to tests. ........ r50794 | martin.v.loewis | 2006-07-24 07:05:22 +0200 (Mon, 24 Jul 2006) | 2 lines Update list of unsupported systems. Fixes #1510853. ........ r50795 | martin.v.loewis | 2006-07-24 12:26:33 +0200 (Mon, 24 Jul 2006) | 1 line Patch #1448199: Release GIL around ConnectRegistry. ........ r50796 | martin.v.loewis | 2006-07-24 13:54:53 +0200 (Mon, 24 Jul 2006) | 3 lines Patch #1232023: Don't include empty path component from registry, so that the current directory does not get added to sys.path. Also fixes #1526785. ........ r50797 | martin.v.loewis | 2006-07-24 14:54:17 +0200 (Mon, 24 Jul 2006) | 3 lines Bug #1524310: Properly report errors from FindNextFile in os.listdir. Will backport to 2.4. ........ r50800 | georg.brandl | 2006-07-24 15:28:57 +0200 (Mon, 24 Jul 2006) | 7 lines Patch #1523356: fix determining include dirs in python-config. Also don't install "python-config" when doing altinstall, but always install "python-config2.x" and make a link to it like with the main executable. ........ r50802 | georg.brandl | 2006-07-24 15:46:47 +0200 (Mon, 24 Jul 2006) | 3 lines Patch #1527744: right order of includes in order to have HAVE_CONIO_H defined properly. ........ r50803 | georg.brandl | 2006-07-24 16:09:56 +0200 (Mon, 24 Jul 2006) | 3 lines Patch #1515343: Fix printing of deprecated string exceptions with a value in the traceback module. ........ r50804 | kurt.kaiser | 2006-07-24 19:13:23 +0200 (Mon, 24 Jul 2006) | 7 lines EditorWindow failed when used stand-alone if sys.ps1 not set. Bug 1010370 Dave Florek M EditorWindow.py M PyShell.py M NEWS.txt ........ r50805 | kurt.kaiser | 2006-07-24 20:05:51 +0200 (Mon, 24 Jul 2006) | 6 lines - EditorWindow.test() was failing. Bug 1417598 M EditorWindow.py M ScriptBinding.py M NEWS.txt ........ r50808 | georg.brandl | 2006-07-24 22:11:35 +0200 (Mon, 24 Jul 2006) | 3 lines Repair accidental NameError. ........ r50809 | tim.peters | 2006-07-24 23:02:15 +0200 (Mon, 24 Jul 2006) | 2 lines Whitespace normalization. ........ r50810 | greg.ward | 2006-07-25 04:11:12 +0200 (Tue, 25 Jul 2006) | 3 lines Don't use standard assert: want tests to fail even when run with -O. Delete cruft. ........ r50811 | tim.peters | 2006-07-25 06:07:22 +0200 (Tue, 25 Jul 2006) | 10 lines current_frames_with_threads(): There's actually no way to guess /which/ line the spawned thread is in at the time sys._current_frames() is called: we know it finished enter_g.set(), but can't know whether the instruction counter has advanced to the following leave_g.wait(). The latter is overwhelming most likely, but not guaranteed, and I see that the "x86 Ubuntu dapper (icc) trunk" buildbot found it on the other line once. Changed the test so it passes in either case. ........ r50815 | martin.v.loewis | 2006-07-25 11:53:12 +0200 (Tue, 25 Jul 2006) | 2 lines Bug #1525817: Don't truncate short lines in IDLE's tool tips. ........ r50816 | martin.v.loewis | 2006-07-25 12:05:47 +0200 (Tue, 25 Jul 2006) | 3 lines Bug #978833: Really close underlying socket in _socketobject.close. Will backport to 2.4. ........ r50817 | martin.v.loewis | 2006-07-25 12:11:14 +0200 (Tue, 25 Jul 2006) | 1 line Revert incomplete checkin. ........ r50819 | georg.brandl | 2006-07-25 12:22:34 +0200 (Tue, 25 Jul 2006) | 4 lines Patch #1525766: correctly pass onerror arg to recursive calls of pkg.walk_packages. Also improve the docstrings. ........ r50825 | brett.cannon | 2006-07-25 19:32:20 +0200 (Tue, 25 Jul 2006) | 2 lines Add comment for changes to test_ossaudiodev. ........ r50826 | brett.cannon | 2006-07-25 19:34:36 +0200 (Tue, 25 Jul 2006) | 3 lines Fix a bug in the messages for an assert failure where not enough arguments to a string were being converted in the format. ........ r50828 | armin.rigo | 2006-07-25 20:09:57 +0200 (Tue, 25 Jul 2006) | 2 lines Document why is and is not a good way to fix the gc_inspection crasher. ........ r50829 | armin.rigo | 2006-07-25 20:11:07 +0200 (Tue, 25 Jul 2006) | 5 lines Added another crasher, which hit me today (I was not intentionally writing such code, of course, but it took some gdb time to figure out what my bug was). ........ r50830 | armin.rigo | 2006-07-25 20:38:39 +0200 (Tue, 25 Jul 2006) | 3 lines Document the crashers that will not go away soon as "won't fix", and explain why. ........ r50831 | ronald.oussoren | 2006-07-25 21:13:35 +0200 (Tue, 25 Jul 2006) | 3 lines Install the compatibility symlink to libpython.a on OSX using 'ln -sf' instead of 'ln -s', this avoid problems when reinstalling python. ........ r50832 | ronald.oussoren | 2006-07-25 21:20:54 +0200 (Tue, 25 Jul 2006) | 7 lines Fix for bug #1525447 (renaming to MacOSmodule.c would also work, but not without causing problems for anyone that is on a case-insensitive filesystem). Setup.py tries to compile the MacOS extension from MacOSmodule.c, while the actual file is named macosmodule.c. This is no problem on the (default) case-insensitive filesystem, but doesn't work on case-sensitive filesystems. ........ r50833 | ronald.oussoren | 2006-07-25 22:28:55 +0200 (Tue, 25 Jul 2006) | 7 lines Fix bug #1517990: IDLE keybindings on OSX This adds a new key definition for OSX, which is slightly different from the classic mac definition. Also add NEWS item for a couple of bugfixes I added recently. ........ r50834 | tim.peters | 2006-07-26 00:30:24 +0200 (Wed, 26 Jul 2006) | 2 lines Whitespace normalization. ........ r50839 | neal.norwitz | 2006-07-26 06:00:18 +0200 (Wed, 26 Jul 2006) | 1 line Hmm, only python2.x is installed, not plain python. Did that change recently? ........ r50840 | barry.warsaw | 2006-07-26 07:54:46 +0200 (Wed, 26 Jul 2006) | 6 lines Forward port some fixes that were in email 2.5 but for some reason didn't make it into email 4.0. Specifically, in Message.get_content_charset(), handle RFC 2231 headers that contain an encoding not known to Python, or a character in the data that isn't in the charset encoding. Also forward port the appropriate unit tests. ........ r50841 | georg.brandl | 2006-07-26 09:23:32 +0200 (Wed, 26 Jul 2006) | 3 lines NEWS entry for #1525766. ........ r50842 | georg.brandl | 2006-07-26 09:40:17 +0200 (Wed, 26 Jul 2006) | 3 lines Bug #1459963: properly capitalize HTTP header names. ........ r50843 | georg.brandl | 2006-07-26 10:03:10 +0200 (Wed, 26 Jul 2006) | 6 lines Part of bug #1523610: fix miscalculation of buffer length. Also add a guard against NULL in converttuple and add a test case (that previously would have crashed). ........ r50844 | martin.v.loewis | 2006-07-26 14:12:56 +0200 (Wed, 26 Jul 2006) | 3 lines Bug #978833: Really close underlying socket in _socketobject.close. Fix httplib.HTTPConnection.getresponse to not close the socket if it is still needed for the response. ........ r50845 | andrew.kuchling | 2006-07-26 19:16:52 +0200 (Wed, 26 Jul 2006) | 1 line [Bug #1471938] Fix build problem on Solaris 8 by conditionalizing the use of mvwgetnstr(); it was conditionalized a few lines below. Fix from Paul Eggert. I also tried out the STRICT_SYSV_CURSES case and am therefore removing the 'untested' comment. ........ r50846 | andrew.kuchling | 2006-07-26 19:18:01 +0200 (Wed, 26 Jul 2006) | 1 line Correct error message ........ r50847 | andrew.kuchling | 2006-07-26 19:19:39 +0200 (Wed, 26 Jul 2006) | 1 line Minor grammar fix ........ r50848 | andrew.kuchling | 2006-07-26 19:22:21 +0200 (Wed, 26 Jul 2006) | 1 line Put news item in right section ........ r50850 | andrew.kuchling | 2006-07-26 20:03:12 +0200 (Wed, 26 Jul 2006) | 1 line Use sys.exc_info() ........ r50851 | andrew.kuchling | 2006-07-26 20:15:45 +0200 (Wed, 26 Jul 2006) | 1 line Use sys.exc_info() ........ r50852 | phillip.eby | 2006-07-26 21:48:27 +0200 (Wed, 26 Jul 2006) | 4 lines Allow the 'onerror' argument to walk_packages() to catch any Exception, not just ImportError. This allows documentation tools to better skip unimportable packages. ........ r50854 | tim.peters | 2006-07-27 01:23:15 +0200 (Thu, 27 Jul 2006) | 2 lines Whitespace normalization. ........ r50855 | tim.peters | 2006-07-27 03:14:53 +0200 (Thu, 27 Jul 2006) | 21 lines Bug #1521947: possible bug in mystrtol.c with recent gcc. In general, C doesn't define anything about what happens when an operation on a signed integral type overflows, and PyOS_strtol() did several formally undefined things of that nature on signed longs. Some version of gcc apparently tries to exploit that now, and PyOS_strtol() could fail to detect overflow then. Tried to repair all that, although it seems at least as likely to me that we'll get screwed by bad platform definitions for LONG_MIN and/or LONG_MAX now. For that reason, I don't recommend backporting this. Note that I have no box on which this makes a lick of difference -- can't really test it, except to note that it didn't break anything on my boxes. Silent change: PyOS_strtol() used to return the hard-coded 0x7fffffff in case of overflow. Now it returns LONG_MAX. They're the same only on 32-bit boxes (although C doesn't guarantee that either ...). ........ r50856 | neal.norwitz | 2006-07-27 05:51:58 +0200 (Thu, 27 Jul 2006) | 6 lines Don't kill a normal instance of python running on windows when checking to kill a cygwin instance. build\\python.exe was matching a normal windows instance. Prefix that with a \\ to ensure build is a directory and not PCbuild. As discussed on python-dev. ........ r50857 | neal.norwitz | 2006-07-27 05:55:39 +0200 (Thu, 27 Jul 2006) | 5 lines Closure can't be NULL at this point since we know it's a tuple. Reported by Klocwork # 74. ........ r50858 | neal.norwitz | 2006-07-27 06:04:50 +0200 (Thu, 27 Jul 2006) | 1 line No functional change. Add comment and assert to describe why there cannot be overflow which was reported by Klocwork. Discussed on python-dev ........ r50859 | martin.v.loewis | 2006-07-27 08:38:16 +0200 (Thu, 27 Jul 2006) | 3 lines Bump distutils version to 2.5, as several new features have been introduced since 2.4. ........ r50860 | andrew.kuchling | 2006-07-27 14:18:20 +0200 (Thu, 27 Jul 2006) | 1 line Reformat docstring; fix typo ........ r50861 | georg.brandl | 2006-07-27 17:05:36 +0200 (Thu, 27 Jul 2006) | 6 lines Add test_main() methods. These three tests were never run by regrtest.py. We really need a simpler testing framework. ........ r50862 | tim.peters | 2006-07-27 17:09:20 +0200 (Thu, 27 Jul 2006) | 2 lines News for patch #1529686. ........ r50863 | tim.peters | 2006-07-27 17:11:00 +0200 (Thu, 27 Jul 2006) | 2 lines Whitespace normalization. ........ r50864 | georg.brandl | 2006-07-27 17:38:33 +0200 (Thu, 27 Jul 2006) | 3 lines Amend news entry. ........ r50865 | georg.brandl | 2006-07-27 18:08:15 +0200 (Thu, 27 Jul 2006) | 3 lines Make uuid test suite pass on this box by requesting output with LC_ALL=C. ........ r50866 | andrew.kuchling | 2006-07-27 20:37:33 +0200 (Thu, 27 Jul 2006) | 1 line Add example ........ r50867 | thomas.heller | 2006-07-27 20:39:55 +0200 (Thu, 27 Jul 2006) | 9 lines Remove code that is no longer used (ctypes.com). Fix the DllGetClassObject and DllCanUnloadNow so that they forward the call to the comtypes.server.inprocserver module. The latter was never documented, never used by published code, and didn't work anyway, so I think it does not deserve a NEWS entry (but I might be wrong). ........ r50868 | andrew.kuchling | 2006-07-27 20:41:21 +0200 (Thu, 27 Jul 2006) | 1 line Typo fix ('publically' is rare, poss. non-standard) ........ r50869 | andrew.kuchling | 2006-07-27 20:42:41 +0200 (Thu, 27 Jul 2006) | 1 line Add missing word ........ r50870 | andrew.kuchling | 2006-07-27 20:44:10 +0200 (Thu, 27 Jul 2006) | 1 line Repair typos ........ r50872 | andrew.kuchling | 2006-07-27 20:53:33 +0200 (Thu, 27 Jul 2006) | 1 line Update URL; add example ........ r50873 | andrew.kuchling | 2006-07-27 21:07:29 +0200 (Thu, 27 Jul 2006) | 1 line Add punctuation mark; add some examples ........ r50874 | andrew.kuchling | 2006-07-27 21:11:07 +0200 (Thu, 27 Jul 2006) | 1 line Mention base64 module; rewrite last sentence to be more positive ........ r50875 | andrew.kuchling | 2006-07-27 21:12:49 +0200 (Thu, 27 Jul 2006) | 1 line If binhex is higher-level than binascii, it should come first in the chapter ........ r50876 | tim.peters | 2006-07-27 22:47:24 +0200 (Thu, 27 Jul 2006) | 28 lines check_node(): stop spraying mystery output to stderr. When a node number disagrees, keep track of all sources & the node numbers they reported, and stick all that in the error message. Changed all callers to supply a non-empty "source" argument; made the "source" argument non-optional. On my box, test_uuid still fails, but with the less confusing output: AssertionError: different sources disagree on node: from source 'getnode1', node was 00038a000015 from source 'getnode2', node was 00038a000015 from source 'ipconfig', node was 001111b2b7bf Only the last one appears to be correct; e.g., C:\Code\python\PCbuild>getmac Physical Address Transport Name =================== ========================================================== 00-11-11-B2-B7-BF \Device\Tcpip_{190FB163-5AFD-4483-86A1-2FE16AC61FF1} 62-A1-AC-6C-FD-BE \Device\Tcpip_{8F77DF5A-EA3D-4F1D-975E-D472CEE6438A} E2-1F-01-C6-5D-88 \Device\Tcpip_{CD18F76B-2EF3-409F-9B8A-6481EE70A1E4} I can't find anything on my box with MAC 00-03-8a-00-00-15, and am not clear on where that comes from. ........ r50878 | andrew.kuchling | 2006-07-28 00:40:05 +0200 (Fri, 28 Jul 2006) | 1 line Reword paragraph ........ r50879 | andrew.kuchling | 2006-07-28 00:49:38 +0200 (Fri, 28 Jul 2006) | 1 line Add example ........ r50880 | andrew.kuchling | 2006-07-28 00:49:54 +0200 (Fri, 28 Jul 2006) | 1 line Add example ........ r50881 | barry.warsaw | 2006-07-28 01:43:15 +0200 (Fri, 28 Jul 2006) | 27 lines Patch #1520294: Support for getset and member descriptors in types.py, inspect.py, and pydoc.py. Specifically, this allows for querying the type of an object against these built-in C types and more importantly, for getting their docstrings printed in the interactive interpreter's help() function. This patch includes a new built-in module called _types which provides definitions of getset and member descriptors for use by the types.py module. These types are exposed as types.GetSetDescriptorType and types.MemberDescriptorType. Query functions are provided as inspect.isgetsetdescriptor() and inspect.ismemberdescriptor(). The implementations of these are robust enough to work with Python implementations other than CPython, which may not have these fundamental types. The patch also includes documentation and test suite updates. I commit these changes now under these guiding principles: 1. Silence is assent. The release manager has not said "no", and of the few people that cared enough to respond to the thread, the worst vote was "0". 2. It's easier to ask for forgiveness than permission. 3. It's so dang easy to revert stuff in svn, that you could view this as a forcing function. :) Windows build patches will follow. ........ r50882 | tim.peters | 2006-07-28 01:44:37 +0200 (Fri, 28 Jul 2006) | 4 lines Bug #1529297: The rewrite of doctest for Python 2.4 unintentionally lost that tests are sorted by name before being run. ``DocTestFinder`` has been changed to sort the list of tests it returns. ........ r50883 | tim.peters | 2006-07-28 01:45:48 +0200 (Fri, 28 Jul 2006) | 2 lines Whitespace normalization. ........ r50884 | tim.peters | 2006-07-28 01:46:36 +0200 (Fri, 28 Jul 2006) | 2 lines Add missing svn:eol-style property to text files. ........ r50885 | barry.warsaw | 2006-07-28 01:50:40 +0200 (Fri, 28 Jul 2006) | 4 lines Enable the building of the _types module on Windows. Note that this has only been tested for VS 2003 since that's all I have. ........ r50887 | tim.peters | 2006-07-28 02:23:15 +0200 (Fri, 28 Jul 2006) | 7 lines defdict_reduce(): Plug leaks. We didn't notice these before because test_defaultdict didn't actually do anything before Georg fixed that earlier today. Neal's next refleak run then showed test_defaultdict leaking 9 references on each run. That's repaired by this checkin. ........ r50888 | tim.peters | 2006-07-28 02:30:00 +0200 (Fri, 28 Jul 2006) | 2 lines News about the repaired memory leak in defaultdict. ........ r50889 | gregory.p.smith | 2006-07-28 03:35:25 +0200 (Fri, 28 Jul 2006) | 7 lines - pybsddb Bug #1527939: bsddb module DBEnv dbremove and dbrename methods now allow their database parameter to be None as the sleepycat API allows. Also adds an appropriate test case for DBEnv.dbrename and dbremove. ........ r50895 | neal.norwitz | 2006-07-28 06:22:34 +0200 (Fri, 28 Jul 2006) | 1 line Ensure the actual number matches the expected count ........ r50896 | tim.peters | 2006-07-28 06:51:59 +0200 (Fri, 28 Jul 2006) | 6 lines Live with that "the hardware address" is an ill-defined concept, and that different ways of trying to find "the hardware address" may return different results. Certainly true on both of my Windows boxes, and in different ways (see whining on python-dev). ........ r50897 | neal.norwitz | 2006-07-28 09:21:27 +0200 (Fri, 28 Jul 2006) | 3 lines Try to find the MAC addr on various flavours of Unix. This seems hopeless. The reduces the test_uuid failures, but there's still another method failing. ........ r50898 | martin.v.loewis | 2006-07-28 09:45:49 +0200 (Fri, 28 Jul 2006) | 2 lines Add UUID for upcoming 2.5b3. ........ r50899 | matt.fleming | 2006-07-28 13:27:27 +0200 (Fri, 28 Jul 2006) | 3 lines Allow socketmodule to compile on NetBSD -current, whose bluetooth API differs from both Linux and FreeBSD. Accepted by Neal Norwitz. ........ r50900 | andrew.kuchling | 2006-07-28 14:07:12 +0200 (Fri, 28 Jul 2006) | 1 line [Patch #1529811] Correction to description of r|* mode ........ r50901 | andrew.kuchling | 2006-07-28 14:18:22 +0200 (Fri, 28 Jul 2006) | 1 line Typo fix ........ r50902 | andrew.kuchling | 2006-07-28 14:32:43 +0200 (Fri, 28 Jul 2006) | 1 line Add example ........ r50903 | andrew.kuchling | 2006-07-28 14:33:19 +0200 (Fri, 28 Jul 2006) | 1 line Add example ........ r50904 | andrew.kuchling | 2006-07-28 14:45:55 +0200 (Fri, 28 Jul 2006) | 1 line Don't overwrite built-in name; add some blank lines for readability ........ r50905 | andrew.kuchling | 2006-07-28 14:48:07 +0200 (Fri, 28 Jul 2006) | 1 line Add example. Should I propagate this example to all the other DBM-ish modules, too? ........ r50912 | georg.brandl | 2006-07-28 20:31:39 +0200 (Fri, 28 Jul 2006) | 3 lines Patch #1529686: also run test_email_codecs with regrtest.py. ........ r50913 | georg.brandl | 2006-07-28 20:36:01 +0200 (Fri, 28 Jul 2006) | 3 lines Fix spelling. ........ r50915 | thomas.heller | 2006-07-28 21:42:40 +0200 (Fri, 28 Jul 2006) | 3 lines Remove a useless XXX comment. Cosmetic changes to the code so that the #ifdef _UNICODE block doesn't mess emacs code formatting. ........ r50916 | phillip.eby | 2006-07-28 23:12:07 +0200 (Fri, 28 Jul 2006) | 5 lines Bug #1529871: The speed enhancement patch #921466 broke Python's compliance with PEP 302. This was fixed by adding an ``imp.NullImporter`` type that is used in ``sys.path_importer_cache`` to cache non-directory paths and avoid excessive filesystem operations during imports. ........ r50917 | phillip.eby | 2006-07-28 23:31:54 +0200 (Fri, 28 Jul 2006) | 2 lines Fix svn merge spew. ........ r50918 | thomas.heller | 2006-07-28 23:43:20 +0200 (Fri, 28 Jul 2006) | 4 lines Patch #1529514: More openbsd platforms for ctypes. Regenerated Modules/_ctypes/libffi/configure with autoconf 2.59. Approved by Neal. ........ r50922 | georg.brandl | 2006-07-29 10:51:21 +0200 (Sat, 29 Jul 2006) | 2 lines Bug #835255: The "closure" argument to new.function() is now documented. ........ r50924 | georg.brandl | 2006-07-29 11:33:26 +0200 (Sat, 29 Jul 2006) | 3 lines Bug #1441397: The compiler module now recognizes module and function docstrings correctly as it did in Python 2.4. ........ r50925 | georg.brandl | 2006-07-29 12:25:46 +0200 (Sat, 29 Jul 2006) | 4 lines Revert rev 42617, it was introduced to work around bug #1441397. test_compiler now passes again. ........ r50926 | fred.drake | 2006-07-29 15:22:49 +0200 (Sat, 29 Jul 2006) | 1 line update target version number ........ r50927 | andrew.kuchling | 2006-07-29 15:56:48 +0200 (Sat, 29 Jul 2006) | 1 line Add example ........ r50928 | andrew.kuchling | 2006-07-29 16:04:47 +0200 (Sat, 29 Jul 2006) | 1 line Update URL ........ r50930 | andrew.kuchling | 2006-07-29 16:08:15 +0200 (Sat, 29 Jul 2006) | 1 line Reword paragraph to match the order of the subsequent sections ........ r50931 | andrew.kuchling | 2006-07-29 16:21:15 +0200 (Sat, 29 Jul 2006) | 1 line [Bug #1529157] Mention raw_input() and input(); while I'm at it, reword the description a bit ........ r50932 | andrew.kuchling | 2006-07-29 16:42:48 +0200 (Sat, 29 Jul 2006) | 1 line [Bug #1519571] Document some missing functions: setup(), title(), done() ........ r50933 | andrew.kuchling | 2006-07-29 16:43:55 +0200 (Sat, 29 Jul 2006) | 1 line Fix docstring punctuation ........ r50934 | andrew.kuchling | 2006-07-29 17:10:32 +0200 (Sat, 29 Jul 2006) | 1 line [Bug #1414697] Change docstring of set/frozenset types to specify that the contents are unique. Raymond, please feel free to edit or revert. ........ r50935 | andrew.kuchling | 2006-07-29 17:35:21 +0200 (Sat, 29 Jul 2006) | 1 line [Bug #1530382] Document SSL.server(), .issuer() methods ........ r50936 | andrew.kuchling | 2006-07-29 17:42:46 +0200 (Sat, 29 Jul 2006) | 1 line Typo fix ........ r50937 | andrew.kuchling | 2006-07-29 17:43:13 +0200 (Sat, 29 Jul 2006) | 1 line Tweak wording ........ r50938 | matt.fleming | 2006-07-29 17:55:30 +0200 (Sat, 29 Jul 2006) | 2 lines Fix typo ........ r50939 | andrew.kuchling | 2006-07-29 17:57:08 +0200 (Sat, 29 Jul 2006) | 6 lines [Bug #1528258] Mention that the 'data' argument can be None. The constructor docs referred the reader to the add_data() method's docs, but they weren't very helpful. I've simply copied an earlier explanation of 'data' that's more useful. ........ r50940 | andrew.kuchling | 2006-07-29 18:08:40 +0200 (Sat, 29 Jul 2006) | 1 line Set bug/patch count. Take a bow, everyone! ........ r50941 | fred.drake | 2006-07-29 18:56:15 +0200 (Sat, 29 Jul 2006) | 18 lines expunge the xmlcore changes: 41667, 41668 - initial switch to xmlcore 47044 - mention of xmlcore in What's New 50687 - mention of xmlcore in the library reference re-apply xmlcore changes to xml: 41674 - line ending changes (re-applied manually), directory props 41677 - add cElementTree wrapper 41678 - PSF licensing for etree 41812 - whitespace normalization 42724 - fix svn:eol-style settings 43681, 43682 - remove Python version-compatibility cruft from minidom 46773 - fix encoding of \r\n\t in attr values in saxutils 47269 - added XMLParser alias for cElementTree compatibility additional tests were added in Lib/test/test_sax.py that failed with the xmlcore changes; these relate to SF bugs #1511497, #1513611 ........ r50942 | andrew.kuchling | 2006-07-29 20:14:07 +0200 (Sat, 29 Jul 2006) | 17 lines Reorganize the docs for 'file' and 'open()' after some discussion with Fred. We want to encourage users to write open() when opening a file, but open() was described with a single paragraph and 'file' had lots of explanation of the mode and bufsize arguments. I've shrunk the description of 'file' to cross-reference to the 'File objects' section, and to open() for an explanation of the arguments. open() now has all the paragraphs about the mode string. The bufsize argument was moved up so that it isn't buried at the end; now there's 1 paragraph on mode, 1 on bufsize, and then 3 more on mode. Various other edits and rearrangements were made in the process. It's probably best to read the final text and not to try to make sense of the diffs. ........ r50943 | fred.drake | 2006-07-29 20:19:19 +0200 (Sat, 29 Jul 2006) | 1 line restore test un-intentionally removed in the xmlcore purge (revision 50941) ........ r50944 | fred.drake | 2006-07-29 20:33:29 +0200 (Sat, 29 Jul 2006) | 3 lines make the reference to older versions of the documentation a link to the right page on python.org ........ r50945 | fred.drake | 2006-07-29 21:09:01 +0200 (Sat, 29 Jul 2006) | 1 line document the footnote usage pattern ........ r50947 | fred.drake | 2006-07-29 21:14:10 +0200 (Sat, 29 Jul 2006) | 1 line emphasize and oddball nuance of LaTeX comment syntax ........ r50948 | andrew.kuchling | 2006-07-29 21:24:04 +0200 (Sat, 29 Jul 2006) | 1 line [Patch #1490989 from Skip Montanaro] Mention debugging builds in the API documentation. I've changed Skip's patch to point to Misc/SpecialBuilds and fiddled with the markup a bit. ........ r50949 | neal.norwitz | 2006-07-29 21:29:35 +0200 (Sat, 29 Jul 2006) | 6 lines Disable these tests until they are reliable across platforms. These problems may mask more important, real problems. One or both methods are known to fail on: Solaris, OpenBSD, Debian, Ubuntu. They pass on Windows and some Linux boxes. ........ r50950 | andrew.kuchling | 2006-07-29 21:50:37 +0200 (Sat, 29 Jul 2006) | 1 line [Patch #1068277] Clarify that os.path.exists() can return False depending on permissions. Fred approved committing this patch in December 2004! ........ r50952 | fred.drake | 2006-07-29 22:04:42 +0200 (Sat, 29 Jul 2006) | 6 lines SF bug #1193966: Weakref types documentation misplaced The information about supporting weakrefs with types defined in C extensions is moved to the Extending & Embedding manual. Py_TPFLAGS_HAVE_WEAKREFS is no longer mentioned since it is part of Py_TPFLAGS_DEFAULT. ........ r50953 | skip.montanaro | 2006-07-29 22:06:05 +0200 (Sat, 29 Jul 2006) | 4 lines Add a comment to the csv reader documentation that explains why the treatment of newlines changed in 2.5. Pulled almost verbatim from a comment by Andrew McNamara in <http://python.org/sf/1465014>. ........ r50954 | neal.norwitz | 2006-07-29 22:20:52 +0200 (Sat, 29 Jul 2006) | 3 lines If the executable doesn't exist, there's no reason to try to start it. This prevents garbage about command not found being printed on Solaris. ........ r50955 | fred.drake | 2006-07-29 22:21:25 +0200 (Sat, 29 Jul 2006) | 1 line fix minor markup error that introduced extra punctuation ........ r50957 | neal.norwitz | 2006-07-29 22:37:08 +0200 (Sat, 29 Jul 2006) | 3 lines Disable test_getnode too, since this is also unreliable. ........ r50958 | andrew.kuchling | 2006-07-29 23:27:12 +0200 (Sat, 29 Jul 2006) | 1 line Follow TeX's conventions for hyphens ........ r50959 | andrew.kuchling | 2006-07-29 23:30:21 +0200 (Sat, 29 Jul 2006) | 1 line Fix case for 'Unix' ........ r50960 | fred.drake | 2006-07-30 01:34:57 +0200 (Sun, 30 Jul 2006) | 1 line markup cleanups ........ r50961 | andrew.kuchling | 2006-07-30 02:27:34 +0200 (Sun, 30 Jul 2006) | 1 line Minor typo fixes ........ r50962 | andrew.kuchling | 2006-07-30 02:37:56 +0200 (Sun, 30 Jul 2006) | 1 line [Bug #793553] Correct description of keyword arguments for SSL authentication ........ r50963 | tim.peters | 2006-07-30 02:58:15 +0200 (Sun, 30 Jul 2006) | 2 lines Whitespace normalization. ........ r50964 | fred.drake | 2006-07-30 05:03:43 +0200 (Sun, 30 Jul 2006) | 1 line lots of markup nits, most commonly Unix/unix --> \UNIX ........ r50965 | fred.drake | 2006-07-30 07:41:28 +0200 (Sun, 30 Jul 2006) | 1 line update information on wxPython, from Robin Dunn ........ r50966 | fred.drake | 2006-07-30 07:49:49 +0200 (Sun, 30 Jul 2006) | 4 lines remove possibly-outdated comment on what GUI toolkit is most commonly used; it is hard to know whether this is right, and it does not add valuable reference information at any rate ........ r50967 | fred.drake | 2006-07-30 07:55:39 +0200 (Sun, 30 Jul 2006) | 3 lines - remove yet another reference to how commonly Tkinter is (thought to be) used - fix an internal section reference ........ r50968 | neal.norwitz | 2006-07-30 08:53:31 +0200 (Sun, 30 Jul 2006) | 4 lines Patch #1531113: Fix augmented assignment with yield expressions. Also fix a SystemError when trying to assign to yield expressions. ........ r50969 | neal.norwitz | 2006-07-30 08:55:48 +0200 (Sun, 30 Jul 2006) | 5 lines Add PyErr_WarnEx() so C code can pass the stacklevel to warnings.warn(). This provides the proper warning for struct.pack(). PyErr_Warn() is now deprecated in favor of PyErr_WarnEx(). As mentioned by Tim Peters on python-dev. ........ r50970 | neal.norwitz | 2006-07-30 08:57:04 +0200 (Sun, 30 Jul 2006) | 3 lines Bug #1515471: string.replace() accepts character buffers again. Pass the char* and size around rather than PyObject's. ........ r50971 | neal.norwitz | 2006-07-30 08:59:13 +0200 (Sun, 30 Jul 2006) | 1 line Whitespace normalization ........ r50973 | georg.brandl | 2006-07-30 12:53:32 +0200 (Sun, 30 Jul 2006) | 3 lines Clarify that __op__ methods must return NotImplemented if they don't support the operation. ........ r50974 | georg.brandl | 2006-07-30 13:07:23 +0200 (Sun, 30 Jul 2006) | 3 lines Bug #1002398: The documentation for os.path.sameopenfile now correctly refers to file descriptors, not file objects. ........ r50977 | martin.v.loewis | 2006-07-30 15:00:31 +0200 (Sun, 30 Jul 2006) | 3 lines Don't copy directory stat times in shutil.copytree on Windows Fixes #1525866. ........ r50978 | martin.v.loewis | 2006-07-30 15:14:05 +0200 (Sun, 30 Jul 2006) | 3 lines Base __version__ on sys.version_info, as distutils is no longer maintained separatedly. ........ r50979 | martin.v.loewis | 2006-07-30 15:27:31 +0200 (Sun, 30 Jul 2006) | 3 lines Mention Cygwin in distutils error message about a missing VS 2003. Fixes #1257728. ........ r50982 | martin.v.loewis | 2006-07-30 16:09:47 +0200 (Sun, 30 Jul 2006) | 5 lines Drop usage of test -e in configure as it is not portable. Fixes #1439538 Will backport to 2.4 Also regenerate pyconfig.h.in. ........ r50984 | georg.brandl | 2006-07-30 18:20:10 +0200 (Sun, 30 Jul 2006) | 3 lines Fix makefile changes for python-config. ........ r50985 | george.yoshida | 2006-07-30 18:37:37 +0200 (Sun, 30 Jul 2006) | 2 lines Rename struct.pack_to to struct.pack_into as changed in revision 46642. ........ r50986 | george.yoshida | 2006-07-30 18:41:30 +0200 (Sun, 30 Jul 2006) | 2 lines Typo fix ........ r50987 | neal.norwitz | 2006-07-30 21:18:13 +0200 (Sun, 30 Jul 2006) | 1 line Add some asserts and update comments ........ r50988 | neal.norwitz | 2006-07-30 21:18:38 +0200 (Sun, 30 Jul 2006) | 1 line Verify that the signal handlers were really called ........ r50989 | neal.norwitz | 2006-07-30 21:20:42 +0200 (Sun, 30 Jul 2006) | 3 lines Try to prevent hangs on Tru64/Alpha buildbot. I'm not certain this will help and may need to be reverted if it causes problems. ........ r50990 | georg.brandl | 2006-07-30 22:18:51 +0200 (Sun, 30 Jul 2006) | 2 lines Bug #1531349: right <-> left glitch in __rop__ description. ........ r50992 | tim.peters | 2006-07-31 03:46:03 +0200 (Mon, 31 Jul 2006) | 2 lines Whitespace normalization. ........ r50993 | andrew.mcnamara | 2006-07-31 04:27:48 +0200 (Mon, 31 Jul 2006) | 2 lines Redo the comment about the 2.5 change in quoted-newline handling. ........ r50994 | tim.peters | 2006-07-31 04:40:23 +0200 (Mon, 31 Jul 2006) | 10 lines ZipFile.close(): Killed one of the struct.pack deprecation warnings on Win32. Also added an XXX about the line: pos3 = self.fp.tell() `pos3` is never referenced, and I have no idea what the code intended to do instead. ........ r50996 | tim.peters | 2006-07-31 04:53:03 +0200 (Mon, 31 Jul 2006) | 8 lines ZipFile.close(): Kill the other struct.pack deprecation warning on Windows. Afraid I can't detect a pattern to when the pack formats decide to use a signed or unsigned format code -- appears nearly arbitrary to my eyes. So I left all the pack formats alone and changed the special-case data values instead. ........ r50997 | skip.montanaro | 2006-07-31 05:09:45 +0200 (Mon, 31 Jul 2006) | 1 line minor tweaks ........ r50998 | skip.montanaro | 2006-07-31 05:11:11 +0200 (Mon, 31 Jul 2006) | 1 line minor tweaks ........ r50999 | andrew.kuchling | 2006-07-31 14:20:24 +0200 (Mon, 31 Jul 2006) | 1 line Add refcounts for PyErr_WarnEx ........ r51000 | andrew.kuchling | 2006-07-31 14:39:05 +0200 (Mon, 31 Jul 2006) | 9 lines Document PyErr_WarnEx. (Bad Neal! No biscuit!) Is the explanation of the 'stacklevel' parameter clear? Please feel free to edit it. I don't have LaTeX installed on this machine, so haven't verified that the markup is correct. Will check tonight, or maybe the automatic doc build will tell me. ........ r51001 | andrew.kuchling | 2006-07-31 14:52:26 +0200 (Mon, 31 Jul 2006) | 1 line Add PyErr_WarnEx() ........ r51002 | andrew.kuchling | 2006-07-31 15:18:27 +0200 (Mon, 31 Jul 2006) | 1 line Mention csv newline changes ........ r51003 | andrew.kuchling | 2006-07-31 17:22:58 +0200 (Mon, 31 Jul 2006) | 1 line Typo fix ........ r51004 | andrew.kuchling | 2006-07-31 17:23:43 +0200 (Mon, 31 Jul 2006) | 1 line Remove reference to notation ........ r51005 | georg.brandl | 2006-07-31 18:00:34 +0200 (Mon, 31 Jul 2006) | 3 lines Fix function name. ........ r51006 | andrew.kuchling | 2006-07-31 18:10:24 +0200 (Mon, 31 Jul 2006) | 1 line [Bug #1514540] Instead of putting the standard types in a section, put them in a chapter of their own. This means string methods will now show up in the ToC. (Should the types come before or after the functions+exceptions+constants chapter? I've put them after, for now.) ........ r51007 | andrew.kuchling | 2006-07-31 18:22:05 +0200 (Mon, 31 Jul 2006) | 1 line [Bug #848556] Remove \d* from second alternative to avoid exponential case when repeating match ........ r51008 | andrew.kuchling | 2006-07-31 18:27:57 +0200 (Mon, 31 Jul 2006) | 1 line Update list of files; fix a typo ........ r51013 | andrew.kuchling | 2006-08-01 18:24:30 +0200 (Tue, 01 Aug 2006) | 1 line typo fix ........ r51018 | thomas.heller | 2006-08-01 18:54:43 +0200 (Tue, 01 Aug 2006) | 2 lines Fix a potential segfault and various potentail refcount leaks in the cast() function. ........ r51020 | thomas.heller | 2006-08-01 19:46:10 +0200 (Tue, 01 Aug 2006) | 1 line Minimal useful docstring for CopyComPointer. ........ r51021 | andrew.kuchling | 2006-08-01 20:16:15 +0200 (Tue, 01 Aug 2006) | 8 lines [Patch #1520905] Attempt to suppress core file created by test_subprocess.py. Patch by Douglas Greiman. The test_run_abort() testcase produces a core file on Unix systems, even though the test is successful. This can be confusing or alarming to someone who runs 'make test' and then finds that the Python interpreter apparently crashed. ........ r51023 | georg.brandl | 2006-08-01 20:49:24 +0200 (Tue, 01 Aug 2006) | 3 lines os.urandom no longer masks unrelated exceptions like SystemExit or KeyboardInterrupt. ........ r51025 | thomas.heller | 2006-08-01 21:14:15 +0200 (Tue, 01 Aug 2006) | 2 lines Speed up PyType_stgdict and PyObject_stgdict. ........ r51027 | ronald.oussoren | 2006-08-01 22:30:31 +0200 (Tue, 01 Aug 2006) | 3 lines Make sure the postinstall action that optionally updates the user's profile on MacOS X actually works correctly in all cases. ........ r51028 | ronald.oussoren | 2006-08-01 23:00:57 +0200 (Tue, 01 Aug 2006) | 4 lines This fixes bug #1527397: PythonLauncher runs scripts with the wrong working directory. It also fixes a bug where PythonLauncher failed to launch scripts when the scriptname (or the path to the script) contains quotes. ........ r51031 | tim.peters | 2006-08-02 05:27:46 +0200 (Wed, 02 Aug 2006) | 2 lines Whitespace normalization. ........ r51032 | tim.peters | 2006-08-02 06:12:36 +0200 (Wed, 02 Aug 2006) | 19 lines Try to squash struct.pack warnings on the "amd64 gentoo trunk" buildbot (& possibly other 64-bit boxes) during test_gzip. The native zlib crc32 function returns an unsigned 32-bit integer, which the Python wrapper implicitly casts to C long. Therefore the same crc can "look negative" on a 32-bit box but "look positive" on a 64-bit box. This patch papers over that platform difference when writing the crc to file. It may be better to change the Python wrapper, either to make the result "look positive" on all platforms (which means it may have to return a Python long at times on a 32-bit box), or to keep the sign the same across boxes. But that would be a visible change in what users see, while the current hack changes no visible behavior (well, apart from stopping the struct deprecation warning). Note that the module-level write32() function is no longer used. ........ r51033 | neal.norwitz | 2006-08-02 06:27:11 +0200 (Wed, 02 Aug 2006) | 4 lines Prevent memory leak on error. Reported by Klocwork #36 ........ r51034 | tim.peters | 2006-08-02 07:20:08 +0200 (Wed, 02 Aug 2006) | 9 lines _Stream.close(): Try to kill struct.pack() warnings when writing the crc to file on the "PPC64 Debian trunk" buildbot when running test_tarfile. This is again a case where the native zlib crc is an unsigned 32-bit int, but the Python wrapper implicitly casts it to signed C long, so that "the sign bit looks different" on different platforms. ........ r51035 | ronald.oussoren | 2006-08-02 08:10:10 +0200 (Wed, 02 Aug 2006) | 2 lines Updated documentation for the script that builds the OSX installer. ........ r51036 | neal.norwitz | 2006-08-02 08:14:22 +0200 (Wed, 02 Aug 2006) | 2 lines _PyWeakref_GetWeakrefCount() now returns a Py_ssize_t instead of long. ........ r51037 | neal.norwitz | 2006-08-02 08:15:10 +0200 (Wed, 02 Aug 2006) | 1 line v is already checked for NULL, so just DECREF it ........ r51038 | neal.norwitz | 2006-08-02 08:19:19 +0200 (Wed, 02 Aug 2006) | 1 line Let us know when there was a problem and the child had to kill the parent ........ r51039 | neal.norwitz | 2006-08-02 08:46:21 +0200 (Wed, 02 Aug 2006) | 5 lines Patch #1519025 and bug #926423: If a KeyboardInterrupt occurs during a socket operation on a socket with a timeout, the exception will be caught correctly. Previously, the exception was not caught. ........ r51040 | neal.norwitz | 2006-08-02 09:09:32 +0200 (Wed, 02 Aug 2006) | 1 line Add some explanation about Klocwork and Coverity static analysis ........ r51041 | anthony.baxter | 2006-08-02 09:43:09 +0200 (Wed, 02 Aug 2006) | 1 line pre-release machinations ........ r51043 | thomas.heller | 2006-08-02 13:35:31 +0200 (Wed, 02 Aug 2006) | 4 lines A few nore words about what ctypes does. Document that using the wrong calling convention can also raise 'ValueError: Procedure called with the wrong number of arguments'. ........ r51045 | thomas.heller | 2006-08-02 14:00:13 +0200 (Wed, 02 Aug 2006) | 1 line Fix a mistake. ........ r51046 | martin.v.loewis | 2006-08-02 15:53:55 +0200 (Wed, 02 Aug 2006) | 3 lines Correction of patch #1455898: In the mbcs decoder, set final=False for stream decoder, but final=True for the decode function. ........ r51049 | tim.peters | 2006-08-02 20:19:35 +0200 (Wed, 02 Aug 2006) | 2 lines Add missing svn:eol-style property to text files. ........ r51079 | neal.norwitz | 2006-08-04 06:50:21 +0200 (Fri, 04 Aug 2006) | 3 lines Bug #1531405, format_exception no longer raises an exception if str(exception) raised an exception. ........ r51080 | neal.norwitz | 2006-08-04 06:58:47 +0200 (Fri, 04 Aug 2006) | 11 lines Bug #1191458: tracing over for loops now produces a line event on each iteration. I'm not positive this is the best way to handle this. I'm also not sure that there aren't other cases where the lnotab is generated incorrectly. It would be great if people that use pdb or tracing could test heavily. Also: * Remove dead/duplicated code that wasn't used/necessary because we already handled the docstring prior to entering the loop. * add some debugging code into the compiler (#if 0'd out). ........ r51081 | neal.norwitz | 2006-08-04 07:09:28 +0200 (Fri, 04 Aug 2006) | 4 lines Bug #1333982: string/number constants were inappropriately stored in the byte code and co_consts even if they were not used, ie immediately popped off the stack. ........ r51082 | neal.norwitz | 2006-08-04 07:12:19 +0200 (Fri, 04 Aug 2006) | 1 line There were really two issues ........ r51084 | fred.drake | 2006-08-04 07:17:21 +0200 (Fri, 04 Aug 2006) | 1 line SF patch #1534048 (bug #1531003): fix typo in error message ........ r51085 | gregory.p.smith | 2006-08-04 07:17:47 +0200 (Fri, 04 Aug 2006) | 3 lines fix typos ........ r51087 | georg.brandl | 2006-08-04 08:03:53 +0200 (Fri, 04 Aug 2006) | 3 lines Fix bug caused by first decrefing, then increfing. ........ r51109 | neil.schemenauer | 2006-08-04 18:20:30 +0200 (Fri, 04 Aug 2006) | 5 lines Fix the 'compiler' package to generate correct code for MAKE_CLOSURE. In the 2.5 development cycle, MAKE_CLOSURE as changed to take free variables as a tuple rather than as individual items on the stack. Closes patch #1534084. ........ r51110 | georg.brandl | 2006-08-04 20:03:37 +0200 (Fri, 04 Aug 2006) | 3 lines Change fix for segfaulting property(), add a NEWS entry and a test. ........ r51111 | georg.brandl | 2006-08-04 20:07:34 +0200 (Fri, 04 Aug 2006) | 3 lines Better fix for bug #1531405, not executing str(value) twice. ........ r51112 | thomas.heller | 2006-08-04 20:17:40 +0200 (Fri, 04 Aug 2006) | 1 line On Windows, make PyErr_Warn an exported function again. ........ r51113 | thomas.heller | 2006-08-04 20:57:34 +0200 (Fri, 04 Aug 2006) | 4 lines Fix #1530448 - fix ctypes build failure on solaris 10. The '-mimpure-text' linker flag is required when linking _ctypes.so. ........ r51114 | thomas.heller | 2006-08-04 21:49:31 +0200 (Fri, 04 Aug 2006) | 3 lines Fix #1534738: win32 debug version of _msi must be _msi_d.pyd, not _msi.pyd. Fix the name of the pdb file as well. ........ r51115 | andrew.kuchling | 2006-08-04 22:37:43 +0200 (Fri, 04 Aug 2006) | 1 line Typo fixes ........ r51116 | andrew.kuchling | 2006-08-04 23:10:03 +0200 (Fri, 04 Aug 2006) | 1 line Fix mangled sentence ........ r51118 | tim.peters | 2006-08-05 00:00:35 +0200 (Sat, 05 Aug 2006) | 2 lines Whitespace normalization. ........ r51119 | bob.ippolito | 2006-08-05 01:59:21 +0200 (Sat, 05 Aug 2006) | 5 lines Fix #1530559, struct.pack raises TypeError where it used to convert. Passing float arguments to struct.pack when integers are expected now triggers a DeprecationWarning. ........ r51123 | georg.brandl | 2006-08-05 08:10:54 +0200 (Sat, 05 Aug 2006) | 3 lines Patch #1534922: correct and enhance unittest docs. ........ r51126 | georg.brandl | 2006-08-06 09:06:33 +0200 (Sun, 06 Aug 2006) | 2 lines Bug #1535182: really test the xreadlines() method of bz2 objects. ........ r51128 | georg.brandl | 2006-08-06 09:26:21 +0200 (Sun, 06 Aug 2006) | 4 lines Bug #1535081: A leading underscore has been added to the names of the md5 and sha modules, so add it in Modules/Setup.dist too. ........ r51129 | georg.brandl | 2006-08-06 10:23:54 +0200 (Sun, 06 Aug 2006) | 3 lines Bug #1535165: fixed a segfault in input() and raw_input() when sys.stdin is closed. ........ r51131 | georg.brandl | 2006-08-06 11:17:16 +0200 (Sun, 06 Aug 2006) | 2 lines Don't produce output in test_builtin. ........ r51133 | andrew.macintyre | 2006-08-06 14:37:03 +0200 (Sun, 06 Aug 2006) | 4 lines test_threading now skips testing alternate thread stack sizes on platforms that don't support changing thread stack size. ........ r51134 | andrew.kuchling | 2006-08-07 00:07:04 +0200 (Mon, 07 Aug 2006) | 2 lines [Patch #1464056] Ensure that we use the panelw library when linking with ncursesw. Once I see how the buildbots react, I'll backport this to 2.4. ........ r51137 | georg.brandl | 2006-08-08 13:52:34 +0200 (Tue, 08 Aug 2006) | 3 lines webbrowser: Silence stderr output if no gconftool or gnome browser found ........ r51138 | georg.brandl | 2006-08-08 13:56:21 +0200 (Tue, 08 Aug 2006) | 7 lines Remove "non-mapping" and "non-sequence" from TypeErrors raised by PyMapping_Size and PySequence_Size. Because len() tries first sequence, then mapping size, it will always raise a "non-mapping object has no len" error which is confusing. ........ r51139 | thomas.heller | 2006-08-08 19:37:00 +0200 (Tue, 08 Aug 2006) | 3 lines memcmp() can return values other than -1, 0, and +1 but tp_compare must not. ........ r51140 | thomas.heller | 2006-08-08 19:39:20 +0200 (Tue, 08 Aug 2006) | 1 line Remove accidently committed, duplicated test. ........ r51147 | andrew.kuchling | 2006-08-08 20:50:14 +0200 (Tue, 08 Aug 2006) | 1 line Reword paragraph to clarify ........ r51148 | andrew.kuchling | 2006-08-08 20:56:08 +0200 (Tue, 08 Aug 2006) | 1 line Move obmalloc item into C API section ........ r51149 | andrew.kuchling | 2006-08-08 21:00:14 +0200 (Tue, 08 Aug 2006) | 1 line 'Other changes' section now has only one item; move the item elsewhere and remove the section ........ r51150 | andrew.kuchling | 2006-08-08 21:00:34 +0200 (Tue, 08 Aug 2006) | 1 line Bump version number ........ r51151 | georg.brandl | 2006-08-08 22:11:22 +0200 (Tue, 08 Aug 2006) | 2 lines Bug #1536828: typo: TypeType should have been StringType. ........ r51153 | georg.brandl | 2006-08-08 22:13:13 +0200 (Tue, 08 Aug 2006) | 2 lines Bug #1536660: separate two words. ........ r51155 | georg.brandl | 2006-08-08 22:48:10 +0200 (Tue, 08 Aug 2006) | 3 lines ``str`` is now the same object as ``types.StringType``. ........ r51156 | tim.peters | 2006-08-09 02:52:26 +0200 (Wed, 09 Aug 2006) | 2 lines Whitespace normalization. ........ r51158 | georg.brandl | 2006-08-09 09:03:22 +0200 (Wed, 09 Aug 2006) | 4 lines Introduce an upper bound on tuple nesting depth in C argument format strings; fixes rest of #1523610. ........ r51160 | martin.v.loewis | 2006-08-09 09:57:39 +0200 (Wed, 09 Aug 2006) | 4 lines __hash__ may now return long int; the final hash value is obtained by invoking hash on the long int. Fixes #1536021. ........ r51168 | andrew.kuchling | 2006-08-09 15:03:41 +0200 (Wed, 09 Aug 2006) | 1 line [Bug #1536021] Mention __hash__ change ........ r51169 | andrew.kuchling | 2006-08-09 15:57:05 +0200 (Wed, 09 Aug 2006) | 1 line [Patch #1534027] Add notes on locale module changes ........ r51170 | andrew.kuchling | 2006-08-09 16:05:35 +0200 (Wed, 09 Aug 2006) | 1 line Add missing 'self' parameters ........ r51171 | andrew.kuchling | 2006-08-09 16:06:19 +0200 (Wed, 09 Aug 2006) | 1 line Reindent code ........ r51172 | armin.rigo | 2006-08-09 16:55:26 +0200 (Wed, 09 Aug 2006) | 2 lines Fix and test for an infinite C recursion. ........ r51173 | ronald.oussoren | 2006-08-09 16:56:33 +0200 (Wed, 09 Aug 2006) | 2 lines It's unlikely that future versions will require _POSIX_C_SOURCE ........ r51178 | armin.rigo | 2006-08-09 17:37:26 +0200 (Wed, 09 Aug 2006) | 2 lines Concatenation on a long string breaks (SF #1526585). ........ r51180 | kurt.kaiser | 2006-08-09 18:46:15 +0200 (Wed, 09 Aug 2006) | 8 lines 1. When used w/o subprocess, all exceptions were preceeded by an error message claiming they were IDLE internal errors (since 1.2a1). 2. Add Ronald Oussoren to CREDITS M NEWS.txt M PyShell.py M CREDITS.txt ........ r51181 | kurt.kaiser | 2006-08-09 19:47:15 +0200 (Wed, 09 Aug 2006) | 4 lines As a slight enhancement to the previous checkin, improve the internal error reporting by moving message to IDLE console. ........ r51182 | andrew.kuchling | 2006-08-09 20:23:14 +0200 (Wed, 09 Aug 2006) | 1 line Typo fix ........ r51183 | kurt.kaiser | 2006-08-09 22:34:46 +0200 (Wed, 09 Aug 2006) | 2 lines ToggleTab dialog was setting indent to 8 even if cancelled (since 1.2a1). ........ r51184 | martin.v.loewis | 2006-08-10 01:42:18 +0200 (Thu, 10 Aug 2006) | 2 lines Add some commentary on -mimpure-text. ........ r51185 | tim.peters | 2006-08-10 02:58:49 +0200 (Thu, 10 Aug 2006) | 2 lines Add missing svn:eol-style property to text files. ........ r51186 | kurt.kaiser | 2006-08-10 03:41:17 +0200 (Thu, 10 Aug 2006) | 2 lines Changing tokenize (39046) to detect dedent broke tabnanny check (since 1.2a1) ........ r51187 | tim.peters | 2006-08-10 05:01:26 +0200 (Thu, 10 Aug 2006) | 13 lines test_copytree_simple(): This was leaving behind two new temp directories each time it ran, at least on Windows. Several changes: explicitly closed all files; wrapped long lines; stopped suppressing errors when removing a file or directory fails (removing /shouldn't/ fail!); and changed what appeared to be incorrect usage of os.removedirs() (that doesn't remove empty directories at and /under/ the given path, instead it must be given an empty leaf directory and then deletes empty directories moving /up/ the path -- could be that the conceptually simpler shutil.rmtree() was really actually intended here). ........
2006-08-11 11:57:12 -03:00
the suggested approach in the absence of more specific information).");
static int
thread_excepthook_file(PyObject *file, PyObject *exc_type, PyObject *exc_value,
PyObject *exc_traceback, PyObject *thread)
{
/* print(f"Exception in thread {thread.name}:", file=file) */
if (PyFile_WriteString("Exception in thread ", file) < 0) {
return -1;
}
PyObject *name = NULL;
if (thread != Py_None) {
if (PyObject_GetOptionalAttr(thread, &_Py_ID(name), &name) < 0) {
return -1;
}
}
if (name != NULL) {
if (PyFile_WriteObject(name, file, Py_PRINT_RAW) < 0) {
Py_DECREF(name);
return -1;
}
Py_DECREF(name);
}
else {
PyThread_ident_t ident = PyThread_get_thread_ident_ex();
PyObject *str = PyUnicode_FromFormat("%" PY_FORMAT_THREAD_IDENT_T, ident);
if (str != NULL) {
if (PyFile_WriteObject(str, file, Py_PRINT_RAW) < 0) {
Py_DECREF(str);
return -1;
}
Py_DECREF(str);
}
else {
PyErr_Clear();
if (PyFile_WriteString("<failed to get thread name>", file) < 0) {
return -1;
}
}
}
if (PyFile_WriteString(":\n", file) < 0) {
return -1;
}
/* Display the traceback */
_PyErr_Display(file, exc_type, exc_value, exc_traceback);
/* Call file.flush() */
if (_PyFile_Flush(file) < 0) {
return -1;
}
return 0;
}
PyDoc_STRVAR(ExceptHookArgs__doc__,
"ExceptHookArgs\n\
\n\
Type used to pass arguments to threading.excepthook.");
static PyStructSequence_Field ExceptHookArgs_fields[] = {
{"exc_type", "Exception type"},
{"exc_value", "Exception value"},
{"exc_traceback", "Exception traceback"},
{"thread", "Thread"},
{0}
};
static PyStructSequence_Desc ExceptHookArgs_desc = {
.name = "_thread._ExceptHookArgs",
.doc = ExceptHookArgs__doc__,
.fields = ExceptHookArgs_fields,
.n_in_sequence = 4
};
static PyObject *
thread_excepthook(PyObject *module, PyObject *args)
{
thread_module_state *state = get_thread_state(module);
if (!Py_IS_TYPE(args, state->excepthook_type)) {
PyErr_SetString(PyExc_TypeError,
"_thread.excepthook argument type "
"must be ExceptHookArgs");
return NULL;
}
/* Borrowed reference */
PyObject *exc_type = PyStructSequence_GET_ITEM(args, 0);
if (exc_type == PyExc_SystemExit) {
/* silently ignore SystemExit */
Py_RETURN_NONE;
}
/* Borrowed references */
PyObject *exc_value = PyStructSequence_GET_ITEM(args, 1);
PyObject *exc_tb = PyStructSequence_GET_ITEM(args, 2);
PyObject *thread = PyStructSequence_GET_ITEM(args, 3);
bpo-46541: Replace core use of _Py_IDENTIFIER() with statically initialized global objects. (gh-30928) We're no longer using _Py_IDENTIFIER() (or _Py_static_string()) in any core CPython code. It is still used in a number of non-builtin stdlib modules. The replacement is: PyUnicodeObject (not pointer) fields under _PyRuntimeState, statically initialized as part of _PyRuntime. A new _Py_GET_GLOBAL_IDENTIFIER() macro facilitates lookup of the fields (along with _Py_GET_GLOBAL_STRING() for non-identifier strings). https://bugs.python.org/issue46541#msg411799 explains the rationale for this change. The core of the change is in: * (new) Include/internal/pycore_global_strings.h - the declarations for the global strings, along with the macros * Include/internal/pycore_runtime_init.h - added the static initializers for the global strings * Include/internal/pycore_global_objects.h - where the struct in pycore_global_strings.h is hooked into _PyRuntimeState * Tools/scripts/generate_global_objects.py - added generation of the global string declarations and static initializers I've also added a --check flag to generate_global_objects.py (along with make check-global-objects) to check for unused global strings. That check is added to the PR CI config. The remainder of this change updates the core code to use _Py_GET_GLOBAL_IDENTIFIER() instead of _Py_IDENTIFIER() and the related _Py*Id functions (likewise for _Py_GET_GLOBAL_STRING() instead of _Py_static_string()). This includes adding a few functions where there wasn't already an alternative to _Py*Id(), replacing the _Py_Identifier * parameter with PyObject *. The following are not changed (yet): * stop using _Py_IDENTIFIER() in the stdlib modules * (maybe) get rid of _Py_IDENTIFIER(), etc. entirely -- this may not be doable as at least one package on PyPI using this (private) API * (maybe) intern the strings during runtime init https://bugs.python.org/issue46541
2022-02-08 16:39:07 -04:00
PyThreadState *tstate = _PyThreadState_GET();
PyObject *file = _PySys_GetAttr(tstate, &_Py_ID(stderr));
if (file == NULL || file == Py_None) {
if (thread == Py_None) {
/* do nothing if sys.stderr is None and thread is None */
Py_RETURN_NONE;
}
file = PyObject_GetAttrString(thread, "_stderr");
if (file == NULL) {
return NULL;
}
if (file == Py_None) {
Py_DECREF(file);
/* do nothing if sys.stderr is None and sys.stderr was None
when the thread was created */
Py_RETURN_NONE;
}
}
else {
Py_INCREF(file);
}
int res = thread_excepthook_file(file, exc_type, exc_value, exc_tb,
thread);
Py_DECREF(file);
if (res < 0) {
return NULL;
}
Py_RETURN_NONE;
}
PyDoc_STRVAR(excepthook_doc,
"_excepthook($module, (exc_type, exc_value, exc_traceback, thread), /)\n\
--\n\
\n\
Handle uncaught Thread.run() exception.");
static PyObject *
thread__is_main_interpreter(PyObject *module, PyObject *Py_UNUSED(ignored))
{
PyInterpreterState *interp = _PyInterpreterState_GET();
return PyBool_FromLong(_Py_IsMainInterpreter(interp));
}
PyDoc_STRVAR(thread__is_main_interpreter_doc,
"_is_main_interpreter($module, /)\n\
--\n\
\n\
Return True if the current interpreter is the main Python interpreter.");
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
static PyObject *
thread_shutdown(PyObject *self, PyObject *args)
{
PyThread_ident_t ident = PyThread_get_thread_ident_ex();
thread_module_state *state = get_thread_state(self);
for (;;) {
ThreadHandle *handle = NULL;
// Find a thread that's not yet finished.
HEAD_LOCK(&_PyRuntime);
struct llist_node *node;
llist_for_each_safe(node, &state->shutdown_handles) {
ThreadHandle *cur = llist_data(node, ThreadHandle, shutdown_node);
if (cur->ident != ident) {
ThreadHandle_incref(cur);
handle = cur;
break;
}
}
HEAD_UNLOCK(&_PyRuntime);
if (!handle) {
// No more threads to wait on!
break;
}
// Wait for the thread to finish. If we're interrupted, such
// as by a ctrl-c we print the error and exit early.
if (ThreadHandle_join(handle, -1) < 0) {
PyErr_WriteUnraisable(NULL);
ThreadHandle_decref(handle);
Py_RETURN_NONE;
}
ThreadHandle_decref(handle);
}
Py_RETURN_NONE;
}
PyDoc_STRVAR(shutdown_doc,
"_shutdown($module, /)\n\
--\n\
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
\n\
Wait for all non-daemon threads (other than the calling thread) to stop.");
static PyObject *
thread__make_thread_handle(PyObject *module, PyObject *identobj)
{
thread_module_state *state = get_thread_state(module);
if (!PyLong_Check(identobj)) {
PyErr_SetString(PyExc_TypeError, "ident must be an integer");
return NULL;
}
PyThread_ident_t ident = PyLong_AsUnsignedLongLong(identobj);
if (PyErr_Occurred()) {
return NULL;
}
PyThreadHandleObject *hobj =
PyThreadHandleObject_new(state->thread_handle_type);
if (hobj == NULL) {
return NULL;
}
PyMutex_Lock(&hobj->handle->mutex);
hobj->handle->ident = ident;
hobj->handle->state = THREAD_HANDLE_RUNNING;
PyMutex_Unlock(&hobj->handle->mutex);
return (PyObject*) hobj;
}
PyDoc_STRVAR(thread__make_thread_handle_doc,
"_make_thread_handle($module, ident, /)\n\
--\n\
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
\n\
Internal only. Make a thread handle for threads not spawned\n\
by the _thread or threading module.");
static PyObject *
thread__get_main_thread_ident(PyObject *module, PyObject *Py_UNUSED(ignored))
{
return PyLong_FromUnsignedLongLong(_PyRuntime.main_thread);
}
PyDoc_STRVAR(thread__get_main_thread_ident_doc,
"_get_main_thread_ident($module, /)\n\
--\n\
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
\n\
Internal only. Return a non-zero integer that uniquely identifies the main thread\n\
of the main interpreter.");
1996-12-16 20:05:22 -04:00
static PyMethodDef thread_methods[] = {
{"start_new_thread", (PyCFunction)thread_PyThread_start_new_thread,
METH_VARARGS, start_new_thread_doc},
{"start_new", (PyCFunction)thread_PyThread_start_new_thread,
METH_VARARGS, start_new_doc},
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
{"start_joinable_thread", _PyCFunction_CAST(thread_PyThread_start_joinable_thread),
METH_VARARGS | METH_KEYWORDS, start_joinable_doc},
{"daemon_threads_allowed", (PyCFunction)thread_daemon_threads_allowed,
METH_NOARGS, daemon_threads_allowed_doc},
{"allocate_lock", thread_PyThread_allocate_lock,
METH_NOARGS, allocate_lock_doc},
{"allocate", thread_PyThread_allocate_lock,
METH_NOARGS, allocate_doc},
{"exit_thread", thread_PyThread_exit_thread,
METH_NOARGS, exit_thread_doc},
{"exit", thread_PyThread_exit_thread,
METH_NOARGS, exit_doc},
{"interrupt_main", (PyCFunction)thread_PyThread_interrupt_main,
METH_VARARGS, interrupt_doc},
{"get_ident", thread_get_ident,
METH_NOARGS, get_ident_doc},
#ifdef PY_HAVE_THREAD_NATIVE_ID
{"get_native_id", thread_get_native_id,
METH_NOARGS, get_native_id_doc},
#endif
{"_count", thread__count,
METH_NOARGS, _count_doc},
{"stack_size", (PyCFunction)thread_stack_size,
METH_VARARGS, stack_size_doc},
{"_excepthook", thread_excepthook,
METH_O, excepthook_doc},
{"_is_main_interpreter", thread__is_main_interpreter,
METH_NOARGS, thread__is_main_interpreter_doc},
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
{"_shutdown", thread_shutdown,
METH_NOARGS, shutdown_doc},
{"_make_thread_handle", thread__make_thread_handle,
METH_O, thread__make_thread_handle_doc},
{"_get_main_thread_ident", thread__get_main_thread_ident,
METH_NOARGS, thread__get_main_thread_ident_doc},
{NULL, NULL} /* sentinel */
};
/* Initialization function */
static int
thread_module_exec(PyObject *module)
{
thread_module_state *state = get_thread_state(module);
PyObject *d = PyModule_GetDict(module);
// Initialize the C thread library
PyThread_init_thread();
// _ThreadHandle
state->thread_handle_type = (PyTypeObject *)PyType_FromSpec(&ThreadHandle_Type_spec);
if (state->thread_handle_type == NULL) {
return -1;
}
if (PyDict_SetItemString(d, "_ThreadHandle", (PyObject *)state->thread_handle_type) < 0) {
return -1;
}
// Lock
state->lock_type = (PyTypeObject *)PyType_FromModuleAndSpec(module, &lock_type_spec, NULL);
if (state->lock_type == NULL) {
return -1;
}
if (PyModule_AddType(module, state->lock_type) < 0) {
return -1;
}
// Old alias: lock -> LockType
if (PyDict_SetItemString(d, "LockType", (PyObject *)state->lock_type) < 0) {
return -1;
}
// RLock
PyTypeObject *rlock_type = (PyTypeObject *)PyType_FromSpec(&rlock_type_spec);
if (rlock_type == NULL) {
return -1;
}
if (PyModule_AddType(module, rlock_type) < 0) {
Py_DECREF(rlock_type);
return -1;
}
Py_DECREF(rlock_type);
// Local dummy
state->local_dummy_type = (PyTypeObject *)PyType_FromSpec(&local_dummy_type_spec);
if (state->local_dummy_type == NULL) {
return -1;
}
// Local
state->local_type = (PyTypeObject *)PyType_FromModuleAndSpec(module, &local_type_spec, NULL);
if (state->local_type == NULL) {
return -1;
}
if (PyModule_AddType(module, state->local_type) < 0) {
return -1;
}
// Add module attributes
if (PyDict_SetItemString(d, "error", ThreadError) < 0) {
return -1;
}
// _ExceptHookArgs type
state->excepthook_type = PyStructSequence_NewType(&ExceptHookArgs_desc);
if (state->excepthook_type == NULL) {
return -1;
}
if (PyModule_AddType(module, state->excepthook_type) < 0) {
return -1;
}
// TIMEOUT_MAX
double timeout_max = (double)PY_TIMEOUT_MAX * 1e-6;
double time_max = PyTime_AsSecondsDouble(PyTime_MAX);
timeout_max = Py_MIN(timeout_max, time_max);
// Round towards minus infinity
timeout_max = floor(timeout_max);
if (PyModule_Add(module, "TIMEOUT_MAX",
PyFloat_FromDouble(timeout_max)) < 0) {
return -1;
}
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
llist_init(&state->shutdown_handles);
return 0;
}
static int
thread_module_traverse(PyObject *module, visitproc visit, void *arg)
{
thread_module_state *state = get_thread_state(module);
Py_VISIT(state->excepthook_type);
Py_VISIT(state->lock_type);
Py_VISIT(state->local_type);
Py_VISIT(state->local_dummy_type);
Py_VISIT(state->thread_handle_type);
return 0;
}
static int
thread_module_clear(PyObject *module)
{
thread_module_state *state = get_thread_state(module);
Py_CLEAR(state->excepthook_type);
Py_CLEAR(state->lock_type);
Py_CLEAR(state->local_type);
Py_CLEAR(state->local_dummy_type);
Py_CLEAR(state->thread_handle_type);
gh-114271: Fix race in `Thread.join()` (#114839) There is a race between when `Thread._tstate_lock` is released[^1] in `Thread._wait_for_tstate_lock()` and when `Thread._stop()` asserts[^2] that it is unlocked. Consider the following execution involving threads A, B, and C: 1. A starts. 2. B joins A, blocking on its `_tstate_lock`. 3. C joins A, blocking on its `_tstate_lock`. 4. A finishes and releases its `_tstate_lock`. 5. B acquires A's `_tstate_lock` in `_wait_for_tstate_lock()`, releases it, but is swapped out before calling `_stop()`. 6. C is scheduled, acquires A's `_tstate_lock` in `_wait_for_tstate_lock()` but is swapped out before releasing it. 7. B is scheduled, calls `_stop()`, which asserts that A's `_tstate_lock` is not held. However, C holds it, so the assertion fails. The race can be reproduced[^3] by inserting sleeps at the appropriate points in the threading code. To do so, run the `repro_join_race.py` from the linked repo. There are two main parts to this PR: 1. `_tstate_lock` is replaced with an event that is attached to `PyThreadState`. The event is set by the runtime prior to the thread being cleared (in the same place that `_tstate_lock` was released). `Thread.join()` blocks waiting for the event to be set. 2. `_PyInterpreterState_WaitForThreads()` provides the ability to wait for all non-daemon threads to exit. To do so, an `is_daemon` predicate was added to `PyThreadState`. This field is set each time a thread is created. `threading._shutdown()` now calls into `_PyInterpreterState_WaitForThreads()` instead of waiting on `_tstate_lock`s. [^1]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1201 [^2]: https://github.com/python/cpython/blob/441affc9e7f419ef0b68f734505fa2f79fe653c7/Lib/threading.py#L1115 [^3]: https://github.com/mpage/cpython/commit/81946532792f938cd6f6ab4c4ff92a4edf61314f --------- Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com> Co-authored-by: Antoine Pitrou <antoine@python.org>
2024-03-16 09:56:30 -03:00
// Remove any remaining handles (e.g. if shutdown exited early due to
// interrupt) so that attempts to unlink the handle after our module state
// is destroyed do not crash.
clear_shutdown_handles(state);
return 0;
}
static void
thread_module_free(void *module)
{
thread_module_clear((PyObject *)module);
}
PyDoc_STRVAR(thread_doc,
"This module provides primitive operations to write multi-threaded programs.\n\
The 'threading' module provides a more convenient interface.");
static PyModuleDef_Slot thread_module_slots[] = {
{Py_mod_exec, thread_module_exec},
{Py_mod_multiple_interpreters, Py_MOD_PER_INTERPRETER_GIL_SUPPORTED},
{Py_mod_gil, Py_MOD_GIL_NOT_USED},
{0, NULL}
};
static struct PyModuleDef thread_module = {
PyModuleDef_HEAD_INIT,
.m_name = "_thread",
.m_doc = thread_doc,
.m_size = sizeof(thread_module_state),
.m_methods = thread_methods,
.m_traverse = thread_module_traverse,
.m_clear = thread_module_clear,
.m_free = thread_module_free,
.m_slots = thread_module_slots,
};
PyMODINIT_FUNC
PyInit__thread(void)
{
return PyModuleDef_Init(&thread_module);
}