* Replace Py_FatalError() calls with regular SystemError exceptions.
* compiler_exit_scope() calls _PyErr_WriteUnraisableMsg() to log the
PySequence_DelItem() failure.
* compiler_unit_check() uses _PyMem_IsPtrFreed().
* compiler_make_closure(): remove "(reftype == FREE)" comment since
reftype can also be LOCAL or GLOBAL_EXPLICIT.
When Python is built in debug mode (with C assertions), calling a
type slot like sq_length (__len__() in Python) now fails with a fatal
error if the slot succeeded with an exception set, or failed with no
exception set. The error message contains the slot, the type name,
and the current exception (if an exception is set).
* Check the result of all slots using _Py_CheckSlotResult().
* No longer pass op_name to ternary_op() in release mode.
* Replace operator with dunder Python method name in error messages.
For example, replace "*" with "__mul__".
* Fix compiler_exit_scope() when an exception is set.
* Fix bytearray.extend() when an exception is set: don't call
bytearray_setslice() with an exception set.
Starting stack viewer when user code is running, including when Debugger is active, hangs or crashes IDLE.
Co-authored-by: Terry Jan Reedy <tjreedy@udel.edu>
Flag members are now divided by one-bit verses multi-bit, with multi-bit being treated as aliases. Iterating over a flag only returns the contained single-bit flags.
Iterating, repr(), and str() show members in definition order.
When constructing combined-member flags, any extra integer values are either discarded (CONFORM), turned into ints (EJECT) or treated as errors (STRICT). Flag classes can specify which of those three behaviors is desired:
>>> class Test(Flag, boundary=CONFORM):
... ONE = 1
... TWO = 2
...
>>> Test(5)
<Test.ONE: 1>
Besides the three above behaviors, there is also KEEP, which should not be used unless necessary -- for example, _convert_ specifies KEEP as there are flag sets in the stdlib that are incomplete and/or inconsistent (e.g. ssl.Options). KEEP will, as the name suggests, keep all bits; however, iterating over a flag with extra bits will only return the canonical flags contained, not the extra bits.
Iteration is now in member definition order. If member definition order
matches increasing value order, then a more efficient method of flag
decomposition is used; otherwise, sort() is called on the results of
that method to get definition order.
``re`` module:
repr() has been modified to support as closely as possible its previous
output; the big difference is that inverted flags cannot be output as
before because the inversion operation now always returns the comparable
positive result; i.e.
re.A|re.I|re.M|re.S is ~(re.L|re.U|re.S|re.T|re.DEBUG)
in both of the above terms, the ``value`` is 282.
re's tests have been updated to reflect the modifications to repr().
Co-authored-by: Éric Araujo <merwok@netwok.org>
Co-authored-by: Terry Jan Reedy <tjreedy@udel.edu>
Co-authored-by: Tal Einat <532281+taleinat@users.noreply.github.com>
Add --with-wheel-pkg-dir=PATH option to the ./configure script. If
specified, the ensurepip module looks for setuptools and pip wheel
packages in this directory: if both are present, these wheel packages
are used instead of ensurepip bundled wheel packages.
Some Linux distribution packaging policies recommend against bundling
dependencies. For example, Fedora installs wheel packages in the
/usr/share/python-wheels/ directory and don't install the
ensurepip._bundled package.
ensurepip: Remove unused runpy import.
The Py_FatalError() function and the faulthandler module now dump the
list of extension modules on a fatal error.
Add _Py_DumpExtensionModules() and _PyModule_IsExtension() internal
functions.
When trying to extract the error line for the error message there
are two distinct cases:
1. The input comes from a file, which means that we can extract the
error line by using `PyErr_ProgramTextObject` and which we already
do.
2. The input does not come from a file, at which point we need to get
the source code from the tokenizer:
* If the tokenizer's current line number is the same with the line
of the error, we get the line from `tok->buf` and we're ready.
* Else, we can extract the error line from the source code in the
following two ways:
* If the input comes from a string we have all the input
in `tok->str` and we can extract the error line from it.
* If the input comes from stdin, i.e. the interactive prompt, we
do not have access to the previous line. That's why a new
field `tok->stdin_content` is added which holds the whole input for the
current (multiline) statement or expression. We can then extract the
error line from `tok->stdin_content` like we do in the string case above.
Co-authored-by: Pablo Galindo <Pablogsal@gmail.com>
Before, using the * operator to repeat a bytearray would copy data from the start of
the internal buffer (ob_bytes) and not from the start of the actual data (ob_start).
`type.__new__` calls `__set_name__` and `__init_subclass__`, which means
that any work metaclasses do after calling `super().__new__()` will not
be available to those two methods. In particular, `Enum` classes that
want to make use of `__init_subclass__` will not see any members.
Almost all customization is therefore moved to before the
`type.__new__()` call, including changing all members to a proto member
descriptor with a `__set_name__` that will do the final conversion of a
member to be an instance of the `Enum` class.
According to [bpo-42874](), some versions of grep do not support the `-q` and `-E` options. Although both options are used elsewhere in the configure script, this particular bit of validation can be achieved without them,
so there's no real harm in using a grep call with no flags.
Would be good to get some people taking advantage of the `--with-tzpath` arguments in the wild to try this out.. Local testing seems to indicate that this does the same thing, but I don't know that we have any buildbots using this option. Maybe @pablogsal?
[bpo-42874]():
Fix the _PyUnicode_FromId() function (_Py_IDENTIFIER(var) API) when
Py_Initialize() / Py_Finalize() is called multiple times:
preserve _PyRuntime.unicode_ids.next_index value.
Use _PyRuntimeState_INIT macro instead memset(0) to reset
_PyRuntimeState members to zero.
If __repr__ uses instance attributes, as normal, and one steps
through the __init__ method, debugger may try to get repr before
the instance attributes exist. reprlib.repr handles the error.
The distutils bdist_wininst command deprecated in Python 3.8 has been
removed. The distutils bidst_wheel command is now recommended to
distribute binary packages on Windows.
* Remove Lib/distutils/command/bdist_wininst.py
* Remove PC/bdist_wininst/ project
* Remove Lib/distutils/command/wininst-*.exe programs
* Remove all references to bdist_wininst
Convert the 6 CJK codec extension modules (_codecs_cn, _codecs_hk,
_codecs_iso2022, _codecs_jp, _codecs_kr and _codecs_tw) to the
multiphase initialization API (PEP 489).
Remove getmultibytecodec() local cache: always import
_multibytecodec. It should be uncommon to get a codec. For example,
this function is only called once per CJK codec module.
Fix a reference leak in register_maps() error path.
* Add test for frame.f_lineno with/without tracing.
* Make sure that frame.f_lineno is correct regardless of whether frame.f_trace is set.
* Update importlib
* Add NEWS
Ref. [SQLite 3.7.15 changelog](https://sqlite.org/changes.html#version_3_7_15):
_"Avoid invoking the sqlite3_trace() callback multiple times when a statement is automatically reprepared due to SQLITE_SCHEMA errors."_
A compiler that doesn't define `__has_builtin` will error out when it is
used on the same line as the check for it.
Automerge-Triggered-By: GH:ronaldoussoren
In is_typing_name(), va_end() is not always called before the
function returns. It is undefined behavior to call va_start()
without also calling va_end().
* Add test capturing missed expectation with uname_result._replace.
* bpo-42163: Override uname_result._make to allow uname_result._replace to work (for everything but 'processor'.
* Replace hard-coded length with one derived from the definition.
* Add test capturing missed expectation with copy/deepcopy on namedtuple (bpo-42189).
* bpo-42189: Exclude processor parameter when constructing uname_result.
* In _make, rely on __new__ to strip processor.
* Add blurb.
* iter is not necessary here.
* Rely on num_fields in __new__
* Add test for slices on uname
* Add test for copy and pickle.
Co-authored-by: Serhiy Storchaka <storchaka@gmail.com>
* import pickle
* Fix equality test after pickling.
* Simply rely on __reduce__ for pickling.
Co-authored-by: Serhiy Storchaka <storchaka@gmail.com>
* bpo-42382: In importlib.metadata, `EntryPoint` objects now expose a `.dist` object referencing the `Distribution` when constructed from a `Distribution`.
Also, sync importlib_metadata 3.3:
- Add support for package discovery under package normalization rules.
- The object returned by `metadata()` now has a formally-defined protocol called `PackageMetadata` with declared support for the `.get_all()` method.
* Add blurb
* Remove latent footnote.
Added --disable-test-modules option to the configure script:
don't build nor install test modules.
Patch by Xavier de Gaye, Thomas Petazzoni and Peixing Xin.
Co-Authored-By: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Co-Authored-By: Xavier de Gaye <xdegaye@gmail.com>
Make the Unicode dictionary of interned strings compatible with
subinterpreters.
Remove the INTERN_NAME_STRINGS macro in typeobject.c: names are
always now interned (even if EXPERIMENTAL_ISOLATED_SUBINTERPRETERS
macro is defined).
_PyUnicode_ClearInterned() now uses PyDict_Next() to no longer
allocate memory, to ensure that the interned dictionary is cleared.
Make the type attribute lookup cache per-interpreter.
Add private _PyType_InitCache() function, called by PyInterpreterState_New().
Continue to share next_version_tag between interpreters, since static
types are still shared by interpreters.
Remove MCACHE macro: the cache is no longer disabled if the
EXPERIMENTAL_ISOLATED_SUBINTERPRETERS macro is defined.
Make _PyUnicode_FromId() function compatible with subinterpreters.
Each interpreter now has an array of identifier objects (interned
strings decoded from UTF-8).
* Add PyInterpreterState.unicode.identifiers: array of identifiers
objects.
* Add _PyRuntimeState.unicode_ids used to allocate unique indexes
to _Py_Identifier.
* Rewrite the _Py_Identifier structure.
Microbenchmark on _PyUnicode_FromId(&PyId_a) with _Py_IDENTIFIER(a):
[ref] 2.42 ns +- 0.00 ns -> [atomic] 3.39 ns +- 0.00 ns: 1.40x slower
This change adds 1 ns per _PyUnicode_FromId() call in average.
When simple query dialogs (tkinter.simpledialog), message boxes
(tkinter.messagebox) or color choose dialog (tkinter.colorchooser)
are created without arguments master and parent, and the default
root window is not yet created, a new temporary hidden root window
will be created automatically. It will not be set as the default root
window and will be destroyed right after closing the dialog window.
It will help to use these simple dialog windows in programs which do
not need other GUI.
Previously, message boxes and color chooser created the blank root
window and left it after closing the dialog window, and query dialogs
just raised an exception.
Co-authored-by: Terry Jan Reedy <tjreedy@udel.edu>
When the modern text= spelling of the universal_newlines= parameter was added
for Python 3.7, check_output's special case around input=None was overlooked.
So it behaved differently with universal_newlines=True vs text=True. This
reconciles the behavior to be consistent and adds a test to guarantee it.
Also clarifies the existing check_output documentation.
Co-authored-by: Alexey Izbyshev <izbyshev@ispras.ru>
It was causing CI failures. the offending file came from https://github.com/python/cpython/pull/23917
```
python3 tools/rstlint.py ../Misc/NEWS.d/next/
[2] ../Misc/NEWS.d/next/Library/2020-12-23-19-43-06.bpo-42727.WH3ODh.rst:1: default role used
[2] ../Misc/NEWS.d/next/Library/2020-12-23-19-43-06.bpo-42727.WH3ODh.rst:2: default role used
2 problems with severity 2 found.
Makefile:204: recipe for target 'check' failed
```
* If parent is specified and mapped, the query widget is
centered at the center of parent. Its position and size
can be corrected so that it fits in the virtual root window.
* Otherwise it is centered at the center of the screen.
On Fedora 31 gdb is using python 3.7.9, calling `proxyval` on an instance with a dictionary fails because of the `dict.iteritems` usage. This PR changes the code to be compatible with py2 and py3.
This changed seemed small enough to not need an issue and news blurb, if one is required please let me know.
Automerge-Triggered-By: GH:benjaminp
The coding cookie (ex: "# coding: latin1") is now ignored in the
command passed to the -c command line option.
Since pymain_run_command() uses UTF-8, pass PyCF_IGNORE_COOKIE
compiler flag to the parser.
pymain_run_python() no longer propages compiler flags between
function calls.
In Python 2, it was possible to use `except` with a nested tuple, and occasionally natural. For example, `zope.formlib.interfaces.InputErrors` is a tuple of several exception classes, and one might reasonably think to do something like this:
try:
self.getInputValue()
return True
except (InputErrors, SomethingElse):
return False
As of Python 3.0, this raises `TypeError: catching classes that do not inherit from BaseException is not allowed` instead: one must instead either break it up into multiple `except` clauses or flatten the tuple. However, the reference documentation was never updated to match this new restriction. Make it clear that the definition is no longer recursive.
Automerge-Triggered-By: GH:ericvsmith
Now all platforms use a value for the "EXT_SUFFIX" build variable derived
from SOABI (for instance in FreeBSD, "EXT_SUFFIX" is now ".cpython-310d.so"
instead of ".so"). Previously only Linux, Mac and VxWorks were using a value
for "EXT_SUFFIX" that included "SOABI".
Co-authored-by: Pablo Galindo <pablogsal@gmail.com>
* Tkinter functions and constructors which need a default root window
raise now RuntimeError with descriptive message instead of obscure
AttributeError or NameError if it is not created yet or cannot
be created automatically.
* Add tests for all functions which use default root window.
* Fix import in the pynche script.
zipimport's _unmarshal_code swallows import errors and then _get_module_code doesn't know the cause of the error, and returns the generic, and sometimes incorrect, 'could not find...'.
Automerge-Triggered-By: GH:brettcannon
Up until now, the `multiprocessing.pool.ThreadPool` class has gone
undocumented, despite being a public class in multiprocessing that is
included in `multiprocessing.pool.__all__`.
Port the _thread extension module to the multiphase initialization
API (PEP 489) and convert its static types to heap types.
Add a traverse function to the lock type, so the garbage collector
can break reference cycles.
* Delete jump instructions that bypass empty blocks
* Add news entry
* Explicitly check for unconditional jump opcodes
Using the is_jump function results in the inclusion of instructions like
returns for which this optimization is not really valid. So, instead
explicitly check that the instruction is an unconditional jump.
* Handle conditional jumps, delete jumps gracefully
* Ensure b_nofallthrough and b_reachable are valid
* Add test for redundant jumps
* Regenerate importlib.h and edit Misc/ACKS
* Fix bad whitespace
* bpo-42644: Validate values in logging.disable()
Technically make the value of manager a property that checks and convert
values assigned to it properly. This has the side effect of making
`logging.disable` also accept strings representing the various level of
warnings.
We want to validate the type of the disable attribute at assignment
time, as it is later compared to other levels when emitting warnings and
would generate a `TypeError: '>=' not supported between ....` in a
different part of the code base, which can make it difficult to track
down.
When assigned an incorrect value; it will raise a TypeError when the
wrong type, or ValueError if an invalid str.
Co-authored-by: Andrew Svetlov <andrew.svetlov@gmail.com>
Enhance test_select.test_select(): it now takes 500 ms rather than 10
seconds.
* Use Python rather than a shell as the child process to make the
test more portable.
* Use a sleep of 50 ms per line rather than 1 second.
* Use subprocess.Popen rather than os.popen().
atexit._run_exitfuncs() now logs callback exceptions using
sys.unraisablehook, rather than logging them directly into
sys.stderr and raising the last exception.
Run GeneralTest of test_atexit in a subprocess since it calls
atexit._clear() which clears all atexit callbacks.
_PyAtExit_Fini() sets state->callbacks to NULL.
* Add _PyAtExit_Call() function and remove pyexitfunc and
pyexitmodule members of PyInterpreterState. The function
logs atexit callback errors using _PyErr_WriteUnraisableMsg().
* Add _PyAtExit_Init() and _PyAtExit_Fini() functions.
* Remove traverse, clear and free functions of the atexit module.
Co-authored-by: Dong-hee Na <donghee.na@python.org>
At Python exit, if a callback registered with atexit.register()
fails, its exception is now logged. Previously, only some exceptions
were logged, and the last exception was always silently ignored.
Add _PyAtExit_Call() function and remove
PyInterpreterState.atexit_func member. call_py_exitfuncs() now calls
directly _PyAtExit_Call().
The atexit module must now always be built as a built-in module.
patch, patch.object and create_autospec silently ignore misspelled
arguments such as autospect, auto_spec and set_spec. This can lead
to tests failing to check what they are supposed to check.
This change adds a check causing a RuntimeError if the above
functions get any of the above misspellings as arguments. It also
adds a new argument, "unsafe", which can be set to True to disable
this check.
Also add "!r" to format specifiers in added error messages.
Add positional only args support to lib2to3 pgen2.
This adds 3.8's PEP-570 support to lib2to3's pgen2. lib2to3, while
being deprecated is still used by things to parse all versions of Python
code today. We need it to support parsing modern 3.8 and 3.9 constructs.
Also add tests for complex *expr and **expr's.
This is invalid in C99 and later and is an error with some compilers
(e.g. clang in Xcode 12), and can thus cause configure checks to
produce incorrect results.
This allows easier Enum construction in unusual cases, such as including dynamic member definitions into a class definition:
# created dynamically
foo_defines = {'FOO_CAT': 'aloof', 'BAR_DOG': 'friendly', 'FOO_HORSE': 'big'}
class Foo(Enum):
vars().update({
k: v
for k, v in foo_defines.items()
if k.startswith('FOO_')
})
def upper(self):
# example method
return self.value.upper()
The default for auto() is to return an integer, which doesn't work for `StrEnum`. The new `_generate_next_value_` for `StrEnum` returns the member name, lower cased.
This is a follow-up to
4662fa9bfe.
That original commit expanded guards against misspelling assertions on
mocks. This follow-up updates the documentation and improves the error
message by pointing out the potential cause and solution.
Automerge-Triggered-By: GH:gpshead
Export the Py_FrozenMain() function: fix a Python 3.9.0 regression.
Python 3.9 uses -fvisibility=hidden and the function was not exported
explicitly and so not exported.
Add also Py_FrozenMain to the stable ABI on Windows.
When creating an Enum, type.__new__ calls __init_subclass__, but at that point the members have not been added.
This patch suppresses the initial call, then manually calls the ancestor __init_subclass__ before returning the new Enum class.
- Copy existing xxlimited to xxlimited53 (named for the limited API version it uses)
- Build both modules, both in debug and release
- Test both modules
Fix encoding name when running a ".pyc" file on Windows:
PyRun_SimpleFileExFlags() now uses the correct encoding to decode the
filename.
* Add pyrun_file() subfunction.
* Add pyrun_simple_file() subfunction.
* PyRun_SimpleFileExFlags() now calls _Py_fopen_obj() rather than
_Py_fopen().
Remove obsolete workaround that prevented running files with
shortcuts when using new universal2 installers built on macOS 11.
Ignore buggy 2nd run_module_event call.
Use `_PyArg_NoKeywords` instead of `_PyArg_NoKwnames` when checking the `kwds` tuple when creating `GenericAlias`. This fixes an interpreter crash when passing in keyword arguments to `GenericAlias`'s constructor.
Needs backport to 3.9.
Automerge-Triggered-By: GH:gvanrossum
Several built-in and standard library types now ensure that their internal result tuples are always tracked by the garbage collector:
- collections.OrderedDict.items
- dict.items
- enumerate
- functools.reduce
- itertools.combinations
- itertools.combinations_with_replacement
- itertools.permutations
- itertools.product
- itertools.zip_longest
- zip
Previously, they could have become untracked by a prior garbage collection.
Fix test_asyncio.test_call_later() race condition: don't measure
asyncio performance in the call_later() unit test. The test failed
randomly on the CI.
macOS releases numbering has changed as of macOS 11 Big Sur. Previously, major releases were of the form 10.x, 10.x+1, 10.x+2, etc; as of Big Sur, they are now x, x+1, etc, so, for example, 10.15, 10.15.1, ..., 10.15.7, 11, 11.0.1, 11.1, ..., 12, 12.1, etc. Allow Python to build with single-digit deployment target values. Patch provided by FX Coudert.
Add platform.freedesktop_os_release() function to parse freedesktop.org
os-release files.
Signed-off-by: Christian Heimes <christian@python.org>
Co-authored-by: Victor Stinner <vstinner@python.org>
* bpo-42406: Fix whichmodule() with multiprocessing
Signed-off-by: Renato L. de F. Cunha <renatoc@br.ibm.com>
Co-authored-by: Gregory P. Smith <greg@krypto.org>
Previously on VxWorks compiling socket extension module needs the libnet to link. Now VxWorks has moved the replied functions to libc. So removing libnet from setup.py.