]> git.ipfire.org Git - thirdparty/Python/cpython.git/log
thirdparty/Python/cpython.git
7 days agogh-141615: Check stdin instead of stdout for use_rawinput in pdb (#141616)
Tian Gao [Thu, 20 Nov 2025 02:41:25 +0000 (18:41 -0800)] 
gh-141615: Check stdin instead of stdout for use_rawinput in pdb (#141616)

7 days ago[Docs] Fix typo in bdb: `is_skipped_line` to `is_skipped_module` (#141771)
Guo Ci [Thu, 20 Nov 2025 02:39:54 +0000 (21:39 -0500)] 
[Docs] Fix typo in bdb: `is_skipped_line` to `is_skipped_module` (#141771)

7 days agoSimplify setting CLI options for WASI builds (GH-141769)
Brett Cannon [Wed, 19 Nov 2025 23:57:44 +0000 (15:57 -0800)] 
Simplify setting CLI options for WASI builds (GH-141769)

This introduces a Wasmtime configuration file to get some CLI options out of the code for easier manipulation. It also allows for easier tweaking after the Makefile is generated.

As well, cut back on the flexibility of specifying HOSTRUNNER for simpler code. The flexibility was never used and so it didn't make sense to keep it around.

7 days agoRemove GvR from '_stdauthor' in Doc/conf.py (#141765)
Guido van Rossum [Wed, 19 Nov 2025 20:30:01 +0000 (12:30 -0800)] 
Remove GvR from '_stdauthor' in Doc/conf.py (#141765)

Co-authored-by: Guido van Rossum <gvanrossum@microsoft.com>
7 days agoMove the NEWS entry for gh-141659 to the correct section (GH-141744)
Ayappan Perumal [Wed, 19 Nov 2025 20:12:10 +0000 (01:42 +0530)] 
Move the NEWS entry for gh-141659 to the correct section (GH-141744)

8 days agogh-141004: Document `Py_LOCAL` and `Py_LOCAL_INLINE` (GH-141725)
Stan Ulbrych [Wed, 19 Nov 2025 18:32:18 +0000 (18:32 +0000)] 
gh-141004: Document `Py_LOCAL` and `Py_LOCAL_INLINE` (GH-141725)

8 days agogh-141721: Improve docstring for LastUpdatedOrderedDict example (gh141724)
Luciano Ramalho [Wed, 19 Nov 2025 18:22:27 +0000 (15:22 -0300)] 
gh-141721: Improve docstring for LastUpdatedOrderedDict example (gh141724)

8 days agogh-141172: Update to wasi-sdk-29 (GH-141684)
Alex Crichton [Wed, 19 Nov 2025 18:03:30 +0000 (12:03 -0600)] 
gh-141172: Update to wasi-sdk-29 (GH-141684)

This commit updates CI and configuration from wasi-sdk-25 to wasi-sdk-29
which was released recently. This notably includes stubs for pthreads
which all return errors, so some adjustment in logic is necessary to
retain knowledge that WASI cannot yet spawn threads for example.

This additionally increases the wasm stack allowance to 32MiB from 16MiB
to accomodate the `test_recursive_pickle` test in the
`test_functools.py` file. It looks like the Clang/LLVM update that
happened in wasi-sdk-29 relative to wasi-sdk-25 is likely the cause of
this where presumably functions have more locals than before and/or a
slightly adjusted stack space requirement which overflows the stack.

8 days agogh-41779: Allow defining the __dict__ and __weakref__ __slots__ for any class (GH...
Serhiy Storchaka [Wed, 19 Nov 2025 17:11:37 +0000 (19:11 +0200)] 
gh-41779: Allow defining the __dict__ and __weakref__ __slots__ for any class (GH-141755)

8 days agoGH-140638: Add a GC "duration" stat (GH-141720)
Brandt Bucher [Wed, 19 Nov 2025 16:51:39 +0000 (08:51 -0800)] 
GH-140638: Add a GC "duration" stat (GH-141720)

8 days agogh-138072: Small clarifications and phrasing improvements to asyncio HOWTO (#138073)
Alexander Nordin [Wed, 19 Nov 2025 14:16:03 +0000 (06:16 -0800)] 
gh-138072: Small clarifications and phrasing improvements to asyncio HOWTO (#138073)

* - Small clarifications and phrasing improvements

* nit

* Apply suggestions from code review

Co-authored-by: Peter Bierma <zintensitydev@gmail.com>
* clarify event loops when multi threading.

* nit

* Update Doc/howto/a-conceptual-overview-of-asyncio.rst

Co-authored-by: Peter Bierma <zintensitydev@gmail.com>
* nit

* nit

* phrasing for threads & event loops.

* revert changes to event-loop/thread discussion.

* sentence case consistencty.

* slight re-arrange.

* Sentence case consistency.

* tweak language. non-negligible

---------

Co-authored-by: Peter Bierma <zintensitydev@gmail.com>
8 days agogh-141570: can_colorize: Expect fileno() to raise OSError, as documented (#141716)
Miro Hrončok [Wed, 19 Nov 2025 14:02:52 +0000 (15:02 +0100)] 
gh-141570: can_colorize: Expect fileno() to raise OSError, as documented (#141716)

In Fedora, we've been given a slightly incomplete reproducer for a problematic
Python 3.14 color-related change in argparse that leads to an exception when
Python is used from mod_wsgi: https://bugzilla.redhat.com/2414940

mod_wsgi replaces sys.stdout with a custom object that raises OSError on .fileno():

https://github.com/GrahamDumpleton/mod_wsgi/blob/8460dbfcd5c7108892b3cde9fab7cbc1caa27886/src/server/wsgi_logger.c#L434-L440

This should be supported, as the documentation of fileno explicitly says:

> An OSError is raised if the IO object does not use a file descriptor.

https://docs.python.org/3.14/library/io.html#io.IOBase.fileno

The previously expected exception inherits from OSError,
so it is still expected.

Fixes https://github.com/python/cpython/issues/141570

Co-authored-by: Cody Maloney <cmaloney@users.noreply.github.com>
Co-authored-by: Victor Stinner <vstinner@python.org>
8 days agogh-135953: Avoid BytesWarning when sampling profiler tests fail (GH-141719)
Petr Viktorin [Wed, 19 Nov 2025 13:43:42 +0000 (14:43 +0100)] 
gh-135953: Avoid BytesWarning when sampling profiler tests fail (GH-141719)

With `-bb`, this failed on BytesWarning: str() on a bytes instance

8 days agogh-140875: Fix handling of unclosed charrefs before EOF in HTMLParser (GH-140904)
Serhiy Storchaka [Wed, 19 Nov 2025 11:55:10 +0000 (13:55 +0200)] 
gh-140875: Fix handling of unclosed charrefs before EOF in HTMLParser (GH-140904)

8 days agogh-141726: Add PyDict_SetDefaultRef() to the Stable ABI (#141727)
da-woods [Wed, 19 Nov 2025 11:38:10 +0000 (11:38 +0000)] 
gh-141726: Add PyDict_SetDefaultRef() to the Stable ABI (#141727)

Co-authored-by: Stan Ulbrych <89152624+StanFromIreland@users.noreply.github.com>
8 days agoCorrect class name from PullDom to PullDOM (#141207)
Guo Ci [Wed, 19 Nov 2025 10:30:53 +0000 (05:30 -0500)] 
Correct class name from PullDom to PullDOM (#141207)

8 days agoGH-139653: Only raise an exception (or fatal error) when the stack pointer is about...
Mark Shannon [Wed, 19 Nov 2025 10:16:24 +0000 (10:16 +0000)] 
GH-139653: Only raise an exception (or fatal error) when the stack pointer is about to overflow the stack. (GH-141711)

Only raises if the stack pointer is both below the limit *and* above the stack base.
This prevents false positives for user-space threads, as the stack pointer will be outside those bounds
if the stack has been swapped.

8 days agoMerge branch 'main' of https://github.com/python/cpython
Hugo van Kemenade [Wed, 19 Nov 2025 09:42:45 +0000 (11:42 +0200)] 
Merge branch 'main' of https://github.com/python/cpython

8 days agoPost 3.15.0a2
Hugo van Kemenade [Wed, 19 Nov 2025 09:42:16 +0000 (11:42 +0200)] 
Post 3.15.0a2

8 days agogh-141659: Fix bad file descriptor error in subprocess on AIX (GH-141660)
Ayappan Perumal [Wed, 19 Nov 2025 07:37:09 +0000 (13:07 +0530)] 
gh-141659: Fix bad file descriptor error in subprocess on AIX (GH-141660)

/proc/self does not exist on AIX.

8 days agogh-141692: Add a slice-specific lib folder to iOS XCframeworks. (#141693)
Russell Keith-Magee [Wed, 19 Nov 2025 01:48:51 +0000 (09:48 +0800)] 
gh-141692: Add a slice-specific lib folder to iOS XCframeworks. (#141693)

Modifies the iOS XCframework to include a lib folder for each slice that
contains a symlinked version of the libPython dynamic library.

8 days agogh-139103: fix free-threading `dataclass.__init__` perf issue (gh-141596)
Edward Xu [Wed, 19 Nov 2025 00:57:59 +0000 (08:57 +0800)] 
gh-139103: fix free-threading `dataclass.__init__` perf issue (gh-141596)

The dataclasses `__init__` function is generated dynamically by a call to `exec()` and so doesn't have deferred reference counting enabled. Enable deferred reference counting on functions when assigned as an attribute to type objects to avoid reference count contention when creating dataclass instances.

8 days agogh-140381: Increase slow_fibonacci call frequency in test_profiling (#140673)
Thierry Martos [Wed, 19 Nov 2025 00:01:09 +0000 (16:01 -0800)] 
gh-140381: Increase slow_fibonacci call frequency in test_profiling (#140673)

8 days agogh-141691: Apply ruff rules to Apple folder. (#141694)
Russell Keith-Magee [Tue, 18 Nov 2025 22:39:21 +0000 (06:39 +0800)] 
gh-141691: Apply ruff rules to Apple folder. (#141694)

Add ruff configuration to run over the Apple build tooling and testbed script.

9 days agoGH-141686: Break cycles created by JSONEncoder.iterencode (GH-141687)
Brandt Bucher [Tue, 18 Nov 2025 17:51:18 +0000 (09:51 -0800)] 
GH-141686: Break cycles created by JSONEncoder.iterencode (GH-141687)

9 days agogh-42400: Fix buffer overflow in _Py_wrealpath() for very long paths (#141529)
Shamil [Tue, 18 Nov 2025 16:34:58 +0000 (19:34 +0300)] 
gh-42400: Fix buffer overflow in _Py_wrealpath() for very long paths (#141529)

Co-authored-by: Victor Stinner <vstinner@python.org>
9 days agogh-141070: Add PyUnstable_Object_Dump() function (#141072)
Victor Stinner [Tue, 18 Nov 2025 16:13:13 +0000 (17:13 +0100)] 
gh-141070: Add PyUnstable_Object_Dump() function (#141072)

* Promote _PyObject_Dump() as a public function.
* Keep _PyObject_Dump() alias to PyUnstable_Object_Dump()
  for backward compatibility.
* Replace _PyObject_Dump() with PyUnstable_Object_Dump().

Co-authored-by: Peter Bierma <zintensitydev@gmail.com>
Co-authored-by: Kumar Aditya <kumaraditya@python.org>
Co-authored-by: Petr Viktorin <encukou@gmail.com>
9 days agogh-138189: Link references to type slots (GH-141410)
Petr Viktorin [Tue, 18 Nov 2025 15:33:52 +0000 (16:33 +0100)] 
gh-138189: Link references to type slots (GH-141410)

Link references to type slots

9 days agogh-135953: Refactor test_sampling_profiler into multiple files (#141689)
Pablo Galindo Salgado [Tue, 18 Nov 2025 15:14:16 +0000 (15:14 +0000)] 
gh-135953: Refactor test_sampling_profiler into multiple files (#141689)

9 days agoPython 3.15.0a2 v3.15.0a2
Hugo van Kemenade [Tue, 18 Nov 2025 14:50:49 +0000 (16:50 +0200)] 
Python 3.15.0a2

9 days agogh-133879: Copyedit "What's new in Python 3.15" (#141717)
Hugo van Kemenade [Tue, 18 Nov 2025 14:42:13 +0000 (16:42 +0200)] 
gh-133879: Copyedit "What's new in Python 3.15" (#141717)

9 days agoGH-139109: Support switch/case dispatch with the tracing interpreter. (GH-141703)
Mark Shannon [Tue, 18 Nov 2025 13:31:48 +0000 (13:31 +0000)] 
GH-139109: Support switch/case dispatch with the tracing interpreter. (GH-141703)

9 days agoAdd missing backticks in os and decimal docs (#141699)
Mariusz Felisiak [Tue, 18 Nov 2025 10:32:15 +0000 (11:32 +0100)] 
Add missing backticks in os and decimal docs (#141699)

9 days agogh-141004: Document Py_HUGE_VAL/IS_FINITE/IS_INFINITE/IS_NAN (#141544)
Sergey B Kirpichev [Tue, 18 Nov 2025 09:17:37 +0000 (12:17 +0300)] 
gh-141004: Document Py_HUGE_VAL/IS_FINITE/IS_INFINITE/IS_NAN (#141544)

Co-authored-by: Victor Stinner <vstinner@python.org>
Co-authored-by: Stan Ulbrych <89152624+StanFromIreland@users.noreply.github.com>
9 days agogh-129005: Remove copies from _pyio using take_bytes (#141539)
Cody Maloney [Tue, 18 Nov 2025 09:10:32 +0000 (01:10 -0800)] 
gh-129005: Remove copies from _pyio using take_bytes (#141539)

Memory usage now matches that of _io for large files.

9 days agogh-140729: Fix subprocess handling in test_process_pool_executor_pickle (#141688)
Pablo Galindo Salgado [Tue, 18 Nov 2025 02:26:40 +0000 (02:26 +0000)] 
gh-140729: Fix subprocess handling in test_process_pool_executor_pickle (#141688)

9 days agogh-141004: Document `Py_MEMCPY` (GH-141676)
Stan Ulbrych [Mon, 17 Nov 2025 22:52:13 +0000 (22:52 +0000)] 
gh-141004: Document `Py_MEMCPY` (GH-141676)

10 days agogh-141004: Document `PyODict*` APIs (GH-141136)
Peter Bierma [Mon, 17 Nov 2025 19:40:47 +0000 (14:40 -0500)] 
gh-141004: Document `PyODict*` APIs (GH-141136)

10 days agogh-140578: Delete unnecessary NEWS entry (#141427)
Louis [Mon, 17 Nov 2025 17:47:00 +0000 (18:47 +0100)] 
gh-140578: Delete unnecessary NEWS entry (#141427)

Co-authored-by: Hugo van Kemenade <1324225+hugovk@users.noreply.github.com>
10 days agogh-141497: Make ipaddress.IP{v4,v6}Network.hosts() always returning an iterator ...
Krishna Chaitanya [Mon, 17 Nov 2025 17:29:06 +0000 (22:59 +0530)] 
gh-141497: Make ipaddress.IP{v4,v6}Network.hosts() always returning an iterator (GH-141547)

10 days agogh-135953: Simplify GC markers in the tachyon profiler (#141666)
Pablo Galindo Salgado [Mon, 17 Nov 2025 16:32:08 +0000 (16:32 +0000)] 
gh-135953: Simplify GC markers in the tachyon profiler (#141666)

10 days agogh-130160: Add anchors to CLI Usage section for `cmdline` (#133182)
Semyon Moroz [Mon, 17 Nov 2025 14:51:21 +0000 (18:51 +0400)] 
gh-130160: Add anchors to CLI Usage section for `cmdline` (#133182)

10 days agogh-130693: Support more options for search in tkinter.Text (GH-130848)
R.C.M [Mon, 17 Nov 2025 14:42:26 +0000 (09:42 -0500)] 
gh-130693: Support more options for search in tkinter.Text (GH-130848)

* Add parameters nolinestop and strictlimits in the tkinter.Text.search() method.
* Add the tkinter.Text.search_all() method.
* Add more tests for tkinter.Text.search().
* stopindex is now only ignored if it is None.

10 days agoGH-139914: Handle stack growth direction on HPPA (GH-140028)
Stefano Rivera [Mon, 17 Nov 2025 13:41:22 +0000 (05:41 -0800)] 
GH-139914: Handle stack growth direction on HPPA (GH-140028)

Adapted from a patch for Python 3.14 submitted to the Debian BTS by John
https://bugs.debian.org/1105111#20

Co-authored-by: John David Anglin <dave.anglin@bell.net>
10 days agoGH-140643: Add `<native>` and `<GC>` frames to the sampling profiler (#141108)
Brandt Bucher [Mon, 17 Nov 2025 13:39:00 +0000 (05:39 -0800)] 
GH-140643: Add `<native>` and `<GC>` frames to the sampling profiler (#141108)

- Introduce a new field in the GC state to store the frame that initiated garbage collection.
- Update RemoteUnwinder to include options for including "<native>" and "<GC>" frames in the stack trace.
- Modify the sampling profiler to accept parameters for controlling the inclusion of native and GC frames.
- Enhance the stack collector to properly format and append these frames during profiling.
- Add tests to verify the correct behavior of the profiler with respect to native and GC frames, including options to exclude them.

Co-authored-by: Pablo Galindo Salgado <pablogsal@gmail.com>
10 days agogh-135953: Add GIL contention markers to sampling profiler Gecko format (#139485)
Pablo Galindo Salgado [Mon, 17 Nov 2025 12:46:26 +0000 (12:46 +0000)] 
gh-135953: Add GIL contention markers to sampling profiler Gecko format (#139485)

This commit enhances the Gecko format reporter in the sampling profiler
to include markers for GIL acquisition events.

10 days agogh-140729: Add __mp_main__ as a duplicate for __main__ for pickle to work (#140735)
yihong [Mon, 17 Nov 2025 12:43:14 +0000 (20:43 +0800)] 
gh-140729: Add __mp_main__ as a duplicate for __main__ for pickle to work (#140735)

10 days agoDocs: Fix typo in socketserver documentation (#140956)
Thomas Ballard [Mon, 17 Nov 2025 11:47:28 +0000 (06:47 -0500)] 
Docs: Fix typo in socketserver documentation (#140956)

10 days agogh-62480: De-personalize "Coping with mutable arguments" section in `unittest.mock...
dereckduran [Mon, 17 Nov 2025 11:44:44 +0000 (03:44 -0800)] 
gh-62480: De-personalize "Coping with mutable arguments" section in `unittest.mock` examples (#141323)

10 days agogh-141650: Fix typo in `xml.sax.saxutils.unescape` documentation (#141652)
Tamzin Hadasa Kelly [Mon, 17 Nov 2025 11:35:01 +0000 (18:35 +0700)] 
gh-141650: Fix typo in `xml.sax.saxutils.unescape` documentation (#141652)

10 days agogh-141018: Update `.exe`, `.dll`, `.rtf` and `.jpg` mime types in `mimetypes` (#141023)
Stan Ulbrych [Mon, 17 Nov 2025 11:32:00 +0000 (11:32 +0000)] 
gh-141018: Update `.exe`, `.dll`, `.rtf` and `.jpg` mime types in `mimetypes` (#141023)

10 days agofix memory order of `_Py_atomic_store_uint_release` (#141562)
SubbaraoGarlapati [Mon, 17 Nov 2025 11:23:12 +0000 (06:23 -0500)] 
fix memory order of `_Py_atomic_store_uint_release` (#141562)

10 days agogh-136057: Allow step and next to step over for loops (#136160)
Tian Gao [Sun, 16 Nov 2025 21:57:07 +0000 (13:57 -0800)] 
gh-136057: Allow step and next to step over for loops (#136160)

11 days agogh-141004: Document missing `PyCFunction*` and `PyCMethod*` APIs (GH-141253)
Peter Bierma [Sun, 16 Nov 2025 19:25:50 +0000 (14:25 -0500)] 
gh-141004: Document missing `PyCFunction*` and `PyCMethod*` APIs (GH-141253)

Co-authored-by: Bénédikt Tran <10796600+picnixz@users.noreply.github.com>
11 days agogh-125115 : Refactor the pdb parsing issue so positional arguments can pass through...
Tian Gao [Sun, 16 Nov 2025 18:58:28 +0000 (10:58 -0800)] 
gh-125115 : Refactor the pdb parsing issue so positional arguments can pass through (#140933)

11 days agogh-141004: Document the `PyPickleBuffer_*` C API (GH-141630)
Stan Ulbrych [Sun, 16 Nov 2025 18:50:54 +0000 (18:50 +0000)] 
gh-141004: Document the `PyPickleBuffer_*` C API (GH-141630)

Co-authored-by: Peter Bierma <zintensitydev@gmail.com>
11 days agogh-139109: JIT _EXIT_TRACE to ENTER_EXECUTOR rather than _DEOPT (GH-141573)
Ken Jin [Sat, 15 Nov 2025 20:19:41 +0000 (04:19 +0800)] 
gh-139109: JIT _EXIT_TRACE to ENTER_EXECUTOR rather than _DEOPT (GH-141573)

12 days agogh-141553: Fix incorrect function signatures in `_testmultiphase` (#141554)
Shamil [Sat, 15 Nov 2025 17:46:54 +0000 (20:46 +0300)] 
gh-141553: Fix incorrect function signatures in `_testmultiphase` (#141554)

12 days agogh-140458: `xmlrpc.client` raises Fault, does not returns it. (GH-140759)
Yongzi Li [Sat, 15 Nov 2025 16:14:23 +0000 (00:14 +0800)] 
gh-140458: `xmlrpc.client` raises Fault, does not returns it. (GH-140759)

12 days agogh-141579: Fix perf_jit backend in sys.activate_stack_trampoline() (#141580)
Pablo Galindo Salgado [Sat, 15 Nov 2025 02:23:54 +0000 (02:23 +0000)] 
gh-141579: Fix perf_jit backend in sys.activate_stack_trampoline() (#141580)

12 days agogh-136442: Fix unittest to return exit code 5 when setUpClass raises an exception...
Takuya UESHIN [Sat, 15 Nov 2025 00:59:51 +0000 (16:59 -0800)] 
gh-136442: Fix unittest to return exit code 5 when setUpClass raises an exception (#136487)

12 days agoGH-90344: replace single-call `io.IncrementalNewlineDecoder` usage with non-increment...
Guo Ci [Sat, 15 Nov 2025 00:13:37 +0000 (19:13 -0500)] 
GH-90344: replace single-call `io.IncrementalNewlineDecoder` usage with non-incremental newline decoders (GH-30276)

Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com>
Co-authored-by: Brett Cannon <brett@python.org>
12 days agogh-140938: Raise ValueError for infinite inputs to stdev/pstdev (GH-141531)
Raymond Hettinger [Fri, 14 Nov 2025 23:25:45 +0000 (17:25 -0600)] 
gh-140938: Raise ValueError for infinite inputs to stdev/pstdev (GH-141531)

Raise ValueError for infinite inputs to stdev/pstdev

---

Co-authored-by: Gregory P. Smith <68491+gpshead@users.noreply.github.com>
Co-authored-by: Bénédikt Tran <10796600+picnixz@users.noreply.github.com>
12 days agogh-141367: Use CALL_LIST_APPEND instruction only for lists, not for list subclasses...
Mikhail Efimov [Fri, 14 Nov 2025 21:38:39 +0000 (00:38 +0300)] 
gh-141367: Use CALL_LIST_APPEND instruction only for lists, not for list subclasses (GH-141398)

Co-authored-by: Ken Jin <kenjin4096@gmail.com>
13 days agogh-114203: skip locking if object is already locked by two-mutex critical section...
Kumar Aditya [Fri, 14 Nov 2025 17:47:59 +0000 (23:17 +0530)] 
gh-114203: skip locking if object is already locked by two-mutex critical section (#141476)

13 days agogh-141488: Add `Py_` prefix to Include/datetime.h macros (#141493)
Locked-chess-official [Fri, 14 Nov 2025 15:52:14 +0000 (23:52 +0800)] 
gh-141488: Add `Py_` prefix to Include/datetime.h macros (#141493)

13 days agoAdd PyManager troubleshooting steps for direct launch of script files (GH-141530)
Steve Dower [Fri, 14 Nov 2025 15:23:01 +0000 (15:23 +0000)] 
Add PyManager troubleshooting steps for direct launch of script files (GH-141530)

13 days agogh-116146: Avoid empty braces in _testembed.c (GH-141556)
Petr Viktorin [Fri, 14 Nov 2025 15:05:42 +0000 (16:05 +0100)] 
gh-116146: Avoid empty braces in _testembed.c (GH-141556)

13 days agogh-139653: Remove assertions in _Py_InitializeRecursionLimits() (#141551)
Victor Stinner [Fri, 14 Nov 2025 14:56:37 +0000 (15:56 +0100)] 
gh-139653: Remove assertions in _Py_InitializeRecursionLimits() (#141551)

These checks were invalid and failed randomly on FreeBSD
and Alpine Linux.

13 days agoGH-141509: Fix warning about remaining subinterpreters (GH-141528)
Sergey Miryanov [Fri, 14 Nov 2025 14:55:04 +0000 (19:55 +0500)] 
GH-141509: Fix warning about remaining subinterpreters (GH-141528)

Co-authored-by: Peter Bierma <zintensitydev@gmail.com>
13 days agoGH-141312: Allow only integers to longrangeiter_setstate state (GH-141317)
Sergey Miryanov [Fri, 14 Nov 2025 14:52:01 +0000 (19:52 +0500)] 
GH-141312: Allow only integers to longrangeiter_setstate state (GH-141317)

This fixes an assertion error when the new computed start is not an integer.

13 days agogh-139344: Remove pending removal notice for undeprecated importlib.resources API...
Petr Viktorin [Fri, 14 Nov 2025 14:50:03 +0000 (15:50 +0100)] 
gh-139344: Remove pending removal notice for undeprecated importlib.resources API (GH-141507)

13 days agogh-122255: Synchronize warnings in C and Python implementations of the warnings modul...
Serhiy Storchaka [Fri, 14 Nov 2025 14:49:28 +0000 (16:49 +0200)] 
gh-122255: Synchronize warnings in C and Python implementations of the warnings module (GH-122824)

In the linecache module and in the Python implementation of the
warnings module, a DeprecationWarning is issued when
m.__loader__ differs from m.__spec__.loader (like in the C
implementation of the warnings module).

13 days agogh-131229: Temporarily skip `test_basic_multiple_interpreters_deleted_no_reset` ...
Peter Bierma [Fri, 14 Nov 2025 14:22:36 +0000 (09:22 -0500)] 
gh-131229: Temporarily skip `test_basic_multiple_interpreters_deleted_no_reset` (GH-141552)

This is a temporary band-aid to unblock other PRs.

Co-authored-by: Kumar Aditya <kumaraditya@python.org>
13 days agogh-141004: Document `PyOS_InterruptOccurred` (GH-141526)
Stan Ulbrych [Fri, 14 Nov 2025 11:20:36 +0000 (11:20 +0000)] 
gh-141004: Document `PyOS_InterruptOccurred` (GH-141526)

13 days agogh-141376: Remove exceptions from `make smelly` (GH-141392)
Petr Viktorin [Fri, 14 Nov 2025 10:22:18 +0000 (11:22 +0100)] 
gh-141376: Remove exceptions from `make smelly` (GH-141392)

* Don't ignore initialized data and BSS
* Remove exceptions for _init and _fini

13 days agogh-131510: Use PyUnstable_Unicode_GET_CACHED_HASH() (GH-141520)
Victor Stinner [Fri, 14 Nov 2025 10:13:24 +0000 (11:13 +0100)] 
gh-131510: Use PyUnstable_Unicode_GET_CACHED_HASH() (GH-141520)

Replace code that directly accesses PyASCIIObject.hash with
PyUnstable_Unicode_GET_CACHED_HASH().

Remove redundant "assert(PyUnicode_Check(op))" from
PyUnstable_Unicode_GET_CACHED_HASH(), _PyASCIIObject_CAST() already
implements the check.

13 days agogh-139596: Cease caching config.cache & ccache in GH Actions (#141451)
Hugo van Kemenade [Fri, 14 Nov 2025 09:59:19 +0000 (11:59 +0200)] 
gh-139596: Cease caching config.cache & ccache in GH Actions (#141451)

13 days agogh-116146: Add C-API to create module from spec and initfunc (GH-139196)
Itamar Oren [Fri, 14 Nov 2025 09:43:25 +0000 (01:43 -0800)] 
gh-116146: Add C-API to create module from spec and initfunc (GH-139196)

Co-authored-by: Kumar Aditya <kumaraditya@python.org>
Co-authored-by: Petr Viktorin <encukou@gmail.com>
Co-authored-by: Victor Stinner <vstinner@python.org>
13 days agogh-140550: Use a bool for the Py_mod_gil value (GH-141519)
Petr Viktorin [Fri, 14 Nov 2025 09:38:49 +0000 (10:38 +0100)] 
gh-140550: Use a bool for the Py_mod_gil value (GH-141519)

This needs a single bit, but was stored as a void* in the module
struct. This didn't matter due to packing, but now that there's
another bool in the struct, we can save a bit of memory by
making md_gil a bool.

Variables that changed type are renamed, to detect conflicts.

13 days agogh-140790: pdb: Initialize instance variables in Pdb.__init__ (#140791)
M Bussonnier [Fri, 14 Nov 2025 03:18:54 +0000 (04:18 +0100)] 
gh-140790: pdb: Initialize instance variables in Pdb.__init__ (#140791)

Initialize lineno, stack, curindex, curframe, currentbp, and _user_requested_quit attributes in `Pdb.__init__``.

2 weeks agogh-140601: Add ResourceWarning to iterparse when not closed (GH-140603)
Osama Abdelkader [Thu, 13 Nov 2025 19:05:28 +0000 (20:05 +0100)] 
gh-140601: Add ResourceWarning to iterparse when not closed (GH-140603)

When iterparse() opens a file by filename and is not explicitly closed,
emit a ResourceWarning to alert developers of the resource leak.

Signed-off-by: Osama Abdelkader <osama.abdelkader@gmail.com>
Co-authored-by: Serhiy Storchaka <storchaka@gmail.com>
2 weeks agogh-137969: Fix double evaluation of `ForwardRef`s which rely on globals (#140974)
dr-carlos [Thu, 13 Nov 2025 18:17:17 +0000 (04:47 +1030)] 
gh-137969: Fix double evaluation of `ForwardRef`s which rely on globals (#140974)

2 weeks agogh-139109: A new tracing JIT compiler frontend for CPython (GH-140310)
Ken Jin [Thu, 13 Nov 2025 18:08:32 +0000 (02:08 +0800)] 
gh-139109: A new tracing JIT compiler frontend for CPython (GH-140310)

This PR changes the current JIT model from trace projection to trace recording. Benchmarking: better pyperformance (about 1.7% overall) geomean versus current https://raw.githubusercontent.com/facebookexperimental/free-threading-benchmarking/refs/heads/main/results/bm-20251108-3.15.0a1%2B-7e2bc1d-JIT/bm-20251108-vultr-x86_64-Fidget%252dSpinner-tracing_jit-3.15.0a1%2B-7e2bc1d-vs-base.svg, 100% faster Richards on the most improved benchmark versus the current JIT. Slowdown of about 10-15% on the worst benchmark versus the current JIT. **Note: the fastest version isn't the one merged, as it relies on fixing bugs in the specializing interpreter, which is left to another PR**. The speedup in the merged version is about 1.1%. https://raw.githubusercontent.com/facebookexperimental/free-threading-benchmarking/refs/heads/main/results/bm-20251112-3.15.0a1%2B-f8a764a-JIT/bm-20251112-vultr-x86_64-Fidget%252dSpinner-tracing_jit-3.15.0a1%2B-f8a764a-vs-base.svg

Stats: 50% more uops executed, 30% more traces entered the last time we ran them. It also suggests our trace lengths for a real trace recording JIT are too short, as a lot of trace too long aborts https://github.com/facebookexperimental/free-threading-benchmarking/blob/main/results/bm-20251023-3.15.0a1%2B-eb73378-CLANG%2CJIT/bm-20251023-vultr-x86_64-Fidget%252dSpinner-tracing_jit-3.15.0a1%2B-eb73378-pystats-vs-base.md .

This new JIT frontend is already able to record/execute significantly more instructions than the previous JIT frontend. In this PR, we are now able to record through custom dunders, simple object creation, generators, etc. None of these were done by the old JIT frontend. Some custom dunders uops were discovered to be broken as part of this work gh-140277

The optimizer stack space check is disabled, as it's no longer valid to deal with underflow.

Pros:
* Ignoring the generated tracer code as it's automatically created, this is only additional 1k lines of code. The maintenance burden is handled by the DSL and code generator.
* `optimizer.c` is now significantly simpler, as we don't have to do strange things to recover the bytecode from a trace.
* The new JIT frontend is able to handle a lot more control-flow than the old one.
* Tracing is very low overhead. We use the tail calling interpreter/computed goto interpreter to switch between tracing mode and non-tracing mode. I call this mechanism dual dispatch, as we have two dispatch tables dispatching to each other. Specialization is still enabled while tracing.
* Better handling of polymorphism. We leverage the specializing interpreter for this.

Cons:
* (For now) requires tail calling interpreter or computed gotos. This means no Windows JIT for now :(. Not to fret, tail calling is coming soon to Windows though https://github.com/python/cpython/pull/139962

Design:
* After each instruction, the `record_previous_inst` function/label is executed. This does as the name suggests.
* The tracing interpreter lowers bytecode to uops directly so that it can obtain "fresh" values at the point of lowering.
* The tracing version behaves nearly identical to the normal interpreter, in fact it even has specialization! This allows it to run without much of a slowdown when tracing. The actual cost of tracing is only a function call and writes to memory.
* The tracing interpreter uses the specializing interpreter's deopt to naturally form the side exit chains. This allows it to side exit chain effectively, without repeating much code. We force a re-specializing when tracing a deopt.
* The tracing interpreter can even handle goto errors/exceptions, but I chose to disable them for now as it's not tested.
* Because we do not share interpreter dispatch, there is should be no significant slowdown to the original specializing interpreter on tailcall and computed got with JIT disabled. With JIT enabled, there might be a slowdown in the form of the JIT trying to trace.
* Things that could have dynamic instruction pointer effects are guarded on. The guard deopts to a new instruction --- `_DYNAMIC_EXIT`.

2 weeks agogh-141004: Document `PyErr_RangedSyntaxLocationObject` (#141521)
Stan Ulbrych [Thu, 13 Nov 2025 17:58:47 +0000 (17:58 +0000)] 
gh-141004: Document `PyErr_RangedSyntaxLocationObject` (#141521)

PyErr_RangedSyntaxLocationObject

2 weeks agogh-140873: Add support of non-descriptor callables in functools.singledispatchmethod...
Serhiy Storchaka [Thu, 13 Nov 2025 17:48:52 +0000 (19:48 +0200)] 
gh-140873: Add support of non-descriptor callables in functools.singledispatchmethod() (GH-140884)

2 weeks agogh-139653: Add PyUnstable_ThreadState_SetStackProtection() (#139668)
Victor Stinner [Thu, 13 Nov 2025 16:30:50 +0000 (17:30 +0100)] 
gh-139653: Add PyUnstable_ThreadState_SetStackProtection() (#139668)

Add PyUnstable_ThreadState_SetStackProtection() and
PyUnstable_ThreadState_ResetStackProtection() functions
to set the stack base address and stack size of a Python
thread state.

Co-authored-by: Petr Viktorin <encukou@gmail.com>
2 weeks agogh-141004: Document `PyCode_Optimize` (GH-141378)
Peter Bierma [Thu, 13 Nov 2025 15:07:57 +0000 (10:07 -0500)] 
gh-141004: Document `PyCode_Optimize` (GH-141378)

2 weeks agogh-141004: Document C APIs for dictionary keys, values, and items (GH-141009)
Peter Bierma [Thu, 13 Nov 2025 15:02:21 +0000 (10:02 -0500)] 
gh-141004: Document C APIs for dictionary keys, values, and items (GH-141009)

Co-authored-by: Petr Viktorin <encukou@gmail.com>
2 weeks agogh-137959: Fix `TIER1_TO_TIER2` macro name in JIT InternalDocs (GH-141496)
Mikhail Efimov [Thu, 13 Nov 2025 14:44:40 +0000 (17:44 +0300)] 
gh-137959: Fix `TIER1_TO_TIER2` macro name in JIT InternalDocs (GH-141496)

JIT InternalDocs fix

2 weeks agogh-139871: Add `bytearray.take_bytes([n])` to efficiently extract `bytes` (GH-140128)
Cody Maloney [Thu, 13 Nov 2025 13:19:44 +0000 (05:19 -0800)] 
gh-139871: Add `bytearray.take_bytes([n])` to efficiently extract `bytes` (GH-140128)

Update `bytearray` to contain a `bytes` and provide a zero-copy path to
"extract" the `bytes`. This allows making several code paths more efficient.

This does not move any codepaths to make use of this new API. The documentation
changes include common code patterns which can be made more efficient with
this API.

---

When just changing `bytearray` to contain `bytes` I ran pyperformance on a
`--with-lto --enable-optimizations --with-static-libpython` build and don't see
any major speedups or slowdowns with this; all seems to be in the noise of
my machine (Generally changes under 5% or benchmarks that don't touch
bytes/bytearray).

Co-authored-by: Victor Stinner <vstinner@python.org>
Co-authored-by: Maurycy Pawłowski-Wieroński <5383+maurycy@users.noreply.github.com>
2 weeks agogh-140601: Refactor ElementTree.iterparse() tests (GH-141499)
Serhiy Storchaka [Thu, 13 Nov 2025 11:37:01 +0000 (13:37 +0200)] 
gh-140601: Refactor ElementTree.iterparse() tests (GH-141499)

Split existing tests on smaller methods and move them to separate class.
Rename variable "content" to "it".
Use BytesIO instead of StringIO.
Add few more tests.

2 weeks agogh-135801: Add the module parameter to compile() etc (GH-139652)
Serhiy Storchaka [Thu, 13 Nov 2025 11:21:32 +0000 (13:21 +0200)] 
gh-135801: Add the module parameter to compile() etc (GH-139652)

Many functions related to compiling or parsing Python code, such as
compile(), ast.parse(), symtable.symtable(),
and importlib.abc.InspectLoader.source_to_code() now allow to pass
the module name used when filtering syntax warnings.

2 weeks agogh-140260: fix data race in `_struct` module initialization with subinterpreters...
Shamil [Thu, 13 Nov 2025 11:01:31 +0000 (14:01 +0300)] 
gh-140260: fix data race in `_struct` module initialization with subinterpreters (#140909)

2 weeks agogh-137109: refactor warning about threads when forking (#141438)
Gregory P. Smith [Thu, 13 Nov 2025 02:15:16 +0000 (18:15 -0800)] 
gh-137109: refactor warning about threads when forking (#141438)

* gh-137109: refactor warning about threads when forking

This splits the OS API specific functionality to get the number of threads out
from the fallback Python method and warning raising code itself.  This way the
OS APIs can be queried before we've run
`os.register_at_fork(after_in_parent=...)` registered functions which
themselves may (re)start threads that would otherwise be detected.

This is best effort.  If the OS APIs are either unavailable or fail, the
warning generating code still falls back to looking at the Python threading
state after the CPython interpreter world has been restarted and the
after_in_parent calls have been made.  The common case for most Linux and macOS
environments should work today.

This also lines up with the existing TODO refactoring, we may choose to expose
this API to get the number of OS threads in the `os` module in the future.

* NEWS entry

* avoid "function-prototype" compiler warning?

2 weeks agogh-141004: Document `PyRun_InteractiveOneObject` (GH-141405)
Peter Bierma [Wed, 12 Nov 2025 22:52:56 +0000 (17:52 -0500)] 
gh-141004: Document `PyRun_InteractiveOneObject` (GH-141405)

2 weeks agogh-124111: Fix TCL 9 thread detection (GH-128103)
Michael Cho [Wed, 12 Nov 2025 22:16:58 +0000 (17:16 -0500)] 
gh-124111: Fix TCL 9 thread detection (GH-128103)

2 weeks agogh-141442: Add escaping to iOS testbed arguments (#141443)
Russell Keith-Magee [Wed, 12 Nov 2025 21:41:26 +0000 (05:41 +0800)] 
gh-141442: Add escaping to iOS testbed arguments (#141443)

Xcode concatenates the test argument array, losing quoting in the process.

2 weeks agogh-140936: Fix JIT assertion crash at finalization if some generator is alive (GH...
Mikhail Efimov [Wed, 12 Nov 2025 19:04:02 +0000 (22:04 +0300)] 
gh-140936: Fix JIT assertion crash at finalization if some generator is alive (GH-140969)

2 weeks agoAdd details about JIT build infrastructure and updating dependencies to `Tools/jit...
Savannah Ostrowski [Wed, 12 Nov 2025 18:26:50 +0000 (10:26 -0800)] 
Add details about JIT build infrastructure and updating dependencies to `Tools/jit` (#141167)