Instead of surprise crashes and memory corruption, we now hang threads that attempt to re-enter the Python interpreter after Python runtime finalization has started. These are typically daemon threads (our long standing mis-feature) but could also be threads spawned by extension modules that then try to call into Python. This marks the `PyThread_exit_thread` public C API as deprecated as there is no plausible safe way to accomplish that on any supported platform in the face of things like C++ code with finalizers anywhere on a thread's stack. Doing this was the least bad option.
Co-authored-by: Gregory P. Smith <greg@krypto.org>
This is actually an upstream problem in curses, and has been reported
to them already:
https://lists.gnu.org/archive/html/bug-ncurses/2024-09/msg00101.html
This is a nice workaround in the meantime to prevent the segfault.
Co-authored-by: Bénédikt Tran <10796600+picnixz@users.noreply.github.com>
Resolve a memory leak introduced in CPython 3.10's :mod:`ssl` when the :attr:`ssl.SSLSocket.session` property was accessed. Speeds up read and write access to said property by no longer unnecessarily cloning session objects via serialization.
Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com>
Co-authored-by: Peter Bierma <zintensitydev@gmail.com>
Co-authored-by: Gregory P. Smith <greg@krypto.org>
Co-authored-by: Antoine Pitrou <antoine@python.org>
Instead of be limited just by the size of addressable memory (2**63
bytes), Python integers are now also limited by the number of bits, so
the number of bit now always fit in a 64-bit integer.
Both limits are much larger than what might be available in practice,
so it doesn't affect users.
_PyLong_NumBits() and _PyLong_Frexp() are now always successful.
* The module state now stores a strong reference to the Placeholder
singleton.
* Use a regular dealloc function.
* Add Py_TPFLAGS_HAVE_GC flag and a traverse function to help the GC
to collect the type when a _functools extension is unloaded.
* Add NEWS.d entry
* Allow ISO-8601 24:00 alternative to midnight on datetime.time.fromisoformat()
* Allow ISO-8601 24:00 alternative to midnight on datetime.datetime.fromisoformat()
* Add NEWS.d entry
* Improve error message when hour is 24 and minute/second/microsecond is not 0
* Add tests for 24:00 fromisoformat
* Remove duplicate call to days_in_month() by storing in variable
* Add Python implementation
* Fix Lint
* Fix differing error msg in datetime.fromisoformat implementations when 24hrs has non-zero time component(s)
* Fix using time components inside tzinfo in Python implementation
* Don't parse tzinfo in C implementation when invalid iso midnight
* Remove duplicated variable in datetime test assertion line
* Add self to acknowledgements
* Remove duplicate NEWS entry
* Linting
* Add missing test case for when wrapping the year makes it invalid (too large)
Use a `_PyStackRef` and defer the reference to `f_funcobj` when
possible. This avoids some reference count contention in the common case
of executing the same code object from multiple threads concurrently in
the free-threaded build.
- If setting `_fields_` fails, e.g. with AttributeError, don't set the attribute in `__dict__`
- Document the “finalization” behaviour
- Beef up tests: add `getattr`, test Union as well as Structure
- Put common functionality in a common function
Co-authored-by: Peter Bierma <zintensitydev@gmail.com>
Multiple places in the I/O stack optimize common cases by using the
information from stat. Currently individual members are extracted from
the stat and stored into the fileio struct. Refactor the code to store
the whole stat struct instead.
Parallels the changes to _io. The `stat` Python object doesn't allow
changing members, so rather than modifying estimated_size, just clear
the value.
Buildbot failure on Windows 10 with MSC v.1916 64 bit (AMD64):
FAIL: testFmod (test.test_math.MathTests.testFmod)
----------------------------------------------------------------------
Traceback (most recent call last):
File "D:\buildarea\3.x.bolen-windows10\build\Lib\test\test_math.py", line 605, in testFmod
self.ftest('fmod(-10, 1)', math.fmod(-10, 1), -0.0)
~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "D:\buildarea\3.x.bolen-windows10\build\Lib\test\test_math.py", line 258, in ftest
self.fail("{}: {}".format(name, failure))
~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AssertionError: fmod(-10, 1): expected -0.0, got 0.0 (zero has wrong sign)
Here Windows loose sign of the result; if y is nonzero, the result
should have the same sign as x.
This amends commit 28aea5d07d.
Use a `_PyStackRef` and defer the reference to `f_executable` when
possible. This avoids some reference count contention in the common case
of executing the same code object from multiple threads concurrently in
the free-threaded build.
POSIX allows errno to be negative.
Even though all currently supported platforms have non-negative errno,
relying on a quirk like that would make Python less portable.
* Raise PicklingError instead of UnicodeEncodeError, ValueError
and AttributeError in both implementations.
* Chain the original exception to the pickle-specific one as __context__.
* Include the error message of ImportError and some AttributeError in
the PicklingError error message.
* Unify error messages between Python and C implementations.
* Refer to documented __reduce__ and __newobj__ callables instead of
internal methods (e.g. save_reduce()) or pickle opcodes (e.g. NEWOBJ).
* Include more details in error messages (what expected, what got).
* Avoid including a potentially long repr of an arbitrary object in
error messages.
* protect macros expansion via `do { ... } while (0)` constructions in `_decimal.c`
* Use public macro `Py_UNUSED`
This replaces the usages of the `UNUSED` macro which
was not consistent with the `Py_UNUSED` macro itself.
In addition, this amends the parameter names so that
they match their semantic meanings.
* Remove redundant `PyCFunction` casts
Add PyConfig_Get(), PyConfig_GetInt(), PyConfig_Set() and
PyConfig_Names() functions to get and set the current runtime Python
configuration.
Add visibility and "sys spec" to config and preconfig specifications.
_PyConfig_AsDict() now converts PyConfig.xoptions as a dictionary.
Co-authored-by: Bénédikt Tran <10796600+picnixz@users.noreply.github.com>
Switch more _Py_IsImmortal(...) assertions to _Py_IsImmortalLoose(...)
The remaining calls to _Py_IsImmortal are in free-threaded-only code,
initialization of core objects, tests, and guards that fall back to
code that works with mortal objects.
This checks are redundant in normal circumstances and can only work if
the extension registry was intentionally broken.
* The Python implementation now raises exception for the extension code
with false boolean value.
* Simplify the C code. RuntimeError is now raised in explicit checks.
* Add many tests.
This replaces the existing hashlib Blake2 module with a single implementation that uses HACL\*'s Blake2b/Blake2s implementations. We added support for all the modes exposed by the Python API, including tree hashing, leaf nodes, and so on. We ported and merged all of these changes upstream in HACL\*, added test vectors based on Python's existing implementation, and exposed everything needed for hashlib.
This was joint work done with @R1kM.
See the PR for much discussion and benchmarking details. TL;DR: On many systems, 8-50% faster (!) than `libb2`, on some systems it appeared 10-20% slower than `libb2`.
There were a still a number of gaps in the tests, including not looking
at all the builtin types and not checking wrappers in subinterpreters
that weren't in the main interpreter. This fixes all that.
I considered incorporating the names of the PyTypeObject fields
(a la gh-122866), but figured doing so doesn't add much value.
* Parameters after the var-positional parameter are now keyword-only
instead of positional-or-keyword.
* Correctly calculate min_kw_only.
* Raise errors for invalid combinations of the var-positional parameter
with "*", "/" and deprecation markers.
Return -1 and set an exception on error; return 0 if the iterator is
exhausted, and return 1 if the next item was fetched successfully.
Prefer this API to PyIter_Next(), which requires the caller to use
PyErr_Occurred() to differentiate between iterator exhaustion and errors.
Co-authered-by: Irit Katriel <iritkatriel@yahoo.com>
Fix _PyArg_UnpackKeywordsWithVararg for the case when argument for
positional-or-keyword parameter is passed by keyword.
There was only one such case in the stdlib -- the TypeVar constructor.
Fix PyEval_GetLocals() to avoid SystemError ("bad argument to
internal function"). Don't redefine the 'ret' variable in the if
block.
Add an unit test on PyEval_GetLocals().
As per C11 DR#471, ctanh (0 + i NaN) and ctanh (0 + i Inf) should return
0 + i NaN (with "invalid" exception in the second case). This has
corresponding implications for ctan(z), as its errors and special cases
are handled as if the operation is implemented by -i*ctanh(i*z).
This patch fixes cmath's code to do same.
Glibs patch: https://sourceware.org/git/?p=glibc.git;a=commitdiff;h=d15e83c5f5231d971472b5ffc9219d54056ca0f1
As per C11 DR#471 (adjusted resolution accepted for C17), cacosh (0 +
iNaN) should return NaN ± i pi/2, not NaN + iNaN. This patch
fixes cmath's code to do same.
This flag was added as an escape hatch in gh-91401 and backported to
Python 3.10. The flag broke at some point between its addition and now.
As there is currently no publicly known environments that require this,
remove it rather than work on fixing it.
This leaves the flag in the subprocess module to not break code which
may have used / checked the flag itself.
discussion: https://discuss.python.org/t/subprocess-use-vfork-escape-hatch-broken-fix-or-remove/56915/2
Serializing objects with complex __qualname__ (such as unbound methods and
nested classes) by name no longer involves serializing parent objects by value
in pickle protocols < 4.
* gh-120974: Make _asyncio._leave_task atomic in the free-threaded build
Update `_PyDict_DelItemIf` to allow for an argument to be passed to the
predicate.
This refactors asyncio to use the common freelist helper functions and
macros. As a side effect, the freelist for _asyncio.Future is now
re-enabled in the free-threaded build.
* Switch PyUnicode_InternInPlace to _PyUnicode_InternMortal, clarify docs
* Document immortality in some functions that take `const char *`
This is PyUnicode_InternFromString;
PyDict_SetItemString, PyObject_SetAttrString;
PyObject_DelAttrString; PyUnicode_InternFromString;
and the PyModule_Add convenience functions.
Always point out a non-immortalizing alternative.
* Don't immortalize user-provided attr names in _ctypes
Any cross-interpreter mechanism for passing objects between interpreters must be very careful to respect isolation, even when the object is effectively immutable (e.g. int, str). Here this especially relates to when an interpreter sends one of its objects, and then is destroyed while the inter-interpreter machinery (e.g. queue) still holds a reference to the object.
When I added interpreters.Queue, I dealt with that case (using an atexit hook) by silently removing all items from the queue that were added by the finalizing interpreter.
Later, while working on concurrent.futures.InterpreterPoolExecutor (gh-116430), I noticed it was somewhat surprising when items were silently removed from the queue when the originating interpreter was destroyed. (See my comment on that PR.)
It took me a little while to realize what was going on. I expect that users, which much less context than I have, would experience the same pain.
My approach, here, to improving the situation is to give users three options:
1. return a singleton (interpreters.queues.UNBOUND) from Queue.get() in place of each removed item
2. raise an exception (interpreters.queues.ItemInterpreterDestroyed) from Queue.get() in place of each removed item
3. existing behavior: silently remove each item (i.e. Queue.get() skips each one)
The default will now be (1), but users can still explicitly opt in any of them, including to the silent removal behavior.
The behavior for each item may be set with the corresponding Queue.put() call. and a queue-wide default may be set when the queue is created. (This is the same as I did for "synconly".)
On POSIX systems, excluding macOS framework installs, the lib directory
for the free-threaded build now includes a "t" suffix to avoid conflicts
with a co-located default build installation.
This makes select.poll() and kqueue() objects thread-safe in the
free-threaded build. Note that calling close() concurrently with other
functions is still not thread-safe due to races on file descriptors
(gh-121544).
The `_PySeqLock_EndRead` function needs an acquire fence to ensure that
the load of the sequence happens after any loads within the read side
critical section. The missing fence can trigger bugs on macOS arm64.
Additionally, we need a release fence in `_PySeqLock_LockWrite` to
ensure that the sequence update is visible before any modifications to
the cache entry.
Make error message for index() methods consistent
Remove the repr of the searched value (which can be arbitrary large)
from ValueError messages for list.index(), range.index(), deque.index(),
deque.remove() and ShareableList.index(). Make the error messages
consistent with error messages for other index() and remove()
methods.
Sometimes a large file is truncated (test_largefile). While
estimated_size is used as a estimate (the read will stil get the number
of bytes in the file), that it is much larger than the actual size of
data can result in a significant over allocation and sometimes lead to
a MemoryError / running out of memory.
This brings the C implementation to match the Python _pyio
implementation.
This reduces the system call count of a simple program[0] that reads all
the `.rst` files in Doc by over 10% (5706 -> 4734 system calls on my
linux system, 5813 -> 4875 on my macOS)
This reduces the number of `fstat()` calls always and seek calls most
the time. Stat was always called twice, once at open (to error early on
directories), and a second time to get the size of the file to be able
to read the whole file in one read. Now the size is cached with the
first call.
The code keeps an optimization that if the user had previously read a
lot of data, the current position is subtracted from the number of bytes
to read. That is somewhat expensive so only do it on larger files,
otherwise just try and read the extra bytes and resize the PyBytes as
needeed.
I built a little test program to validate the behavior + assumptions
around relative costs and then ran it under `strace` to get a log of the
system calls. Full samples below[1].
After the changes, this is everything in one `filename.read_text()`:
```python3
openat(AT_FDCWD, "cpython/Doc/howto/clinic.rst", O_RDONLY|O_CLOEXEC) = 3`
fstat(3, {st_mode=S_IFREG|0644, st_size=343, ...}) = 0`
ioctl(3, TCGETS, 0x7ffdfac04b40) = -1 ENOTTY (Inappropriate ioctl for device)
lseek(3, 0, SEEK_CUR) = 0
read(3, ":orphan:\n\n.. This page is retain"..., 344) = 343
read(3, "", 1) = 0
close(3) = 0
```
This does make some tradeoffs
1. If the file size changes between open() and readall(), this will
still get all the data but might have more read calls.
2. I experimented with avoiding the stat + cached result for small files
in general, but on my dev workstation at least that tended to reduce
performance compared to using the fstat().
[0]
```python3
from pathlib import Path
nlines = []
for filename in Path("cpython/Doc").glob("**/*.rst"):
nlines.append(len(filename.read_text()))
```
[1]
Before small file:
```
openat(AT_FDCWD, "cpython/Doc/howto/clinic.rst", O_RDONLY|O_CLOEXEC) = 3
fstat(3, {st_mode=S_IFREG|0644, st_size=343, ...}) = 0
ioctl(3, TCGETS, 0x7ffe52525930) = -1 ENOTTY (Inappropriate ioctl for device)
lseek(3, 0, SEEK_CUR) = 0
lseek(3, 0, SEEK_CUR) = 0
fstat(3, {st_mode=S_IFREG|0644, st_size=343, ...}) = 0
read(3, ":orphan:\n\n.. This page is retain"..., 344) = 343
read(3, "", 1) = 0
close(3) = 0
```
After small file:
```
openat(AT_FDCWD, "cpython/Doc/howto/clinic.rst", O_RDONLY|O_CLOEXEC) = 3
fstat(3, {st_mode=S_IFREG|0644, st_size=343, ...}) = 0
ioctl(3, TCGETS, 0x7ffdfac04b40) = -1 ENOTTY (Inappropriate ioctl for device)
lseek(3, 0, SEEK_CUR) = 0
read(3, ":orphan:\n\n.. This page is retain"..., 344) = 343
read(3, "", 1) = 0
close(3) = 0
```
Before large file:
```
openat(AT_FDCWD, "cpython/Doc/c-api/typeobj.rst", O_RDONLY|O_CLOEXEC) = 3
fstat(3, {st_mode=S_IFREG|0644, st_size=133104, ...}) = 0
ioctl(3, TCGETS, 0x7ffe52525930) = -1 ENOTTY (Inappropriate ioctl for device)
lseek(3, 0, SEEK_CUR) = 0
lseek(3, 0, SEEK_CUR) = 0
fstat(3, {st_mode=S_IFREG|0644, st_size=133104, ...}) = 0
read(3, ".. highlight:: c\n\n.. _type-struc"..., 133105) = 133104
read(3, "", 1) = 0
close(3) = 0
```
After large file:
```
openat(AT_FDCWD, "cpython/Doc/c-api/typeobj.rst", O_RDONLY|O_CLOEXEC) = 3
fstat(3, {st_mode=S_IFREG|0644, st_size=133104, ...}) = 0
ioctl(3, TCGETS, 0x7ffdfac04b40) = -1 ENOTTY (Inappropriate ioctl for device)
lseek(3, 0, SEEK_CUR) = 0
lseek(3, 0, SEEK_CUR) = 0
read(3, ".. highlight:: c\n\n.. _type-struc"..., 133105) = 133104
read(3, "", 1) = 0
close(3) = 0
```
Co-authored-by: Shantanu <12621235+hauntsaninja@users.noreply.github.com>
Co-authored-by: Erlend E. Aasland <erlend.aasland@protonmail.com>
Co-authored-by: Victor Stinner <vstinner@python.org>
This amends 6988ff02a5: memory allocation for
stginfo->ffi_type_pointer.elements in PyCSimpleType_init() should be
more generic (perhaps someday fmt->pffi_type->elements will be not a
two-elements array).
It should finally resolve#61103.
Co-authored-by: Victor Stinner <vstinner@python.org>
Co-authored-by: Bénédikt Tran <10796600+picnixz@users.noreply.github.com>
sobolevn
Victor Stinner
Sergey B Kirpichev
Bénédikt Tran
Jeremy Maitin-Shepard
Peter Bierma
Sam Gross
Jeffrey R. Van Voorst
Serhiy Storchaka
neonene
Malcolm Smith
Kirill Podoprigora
dgpb
Nice Zombies
TizzySaurus
Raymond Hettinger
Jason Fried
Petr Viktorin
Brian Schubert
Cody Maloney
Jakub Kulík
Irit Katriel
Russell Keith-Magee
Furkan Onder
Wenzel Jakob
Benjamin Peterson
Jérôme Duval
algonell
Adam Turner
nkinnan
Seth Michael Larson
Wulian
blhsing
Bar Harel
Jonathan Protzenko
Eric Snow
Kumar Aditya
Nico Mexis
Erlend E. Aasland
Lucas Esposito
Marc Mueller
Mikołaj Kuranowski
Mark Shannon
Łukasz Langa
mpage
Rodrigo Oliveira
Tian Gao
Alex Waygood
Ken Jin
Savannah Ostrowski
Max Muoto
AN Long