Bénédikt Tran [Mon, 31 Mar 2025 10:26:52 +0000 (12:26 +0200)]
gh-126037: fix UAF in `xml.etree.ElementTree.Element.find*` when concurrent mutations happen (#127964)
We fix a use-after-free in the `find`, `findtext` and `findall` methods of `xml.etree.ElementTree.Element`
objects that can be triggered when the tag to find implements an `__eq__` method that mutates the
element being queried.
Martin DeMello [Fri, 28 Mar 2025 23:58:17 +0000 (16:58 -0700)]
gh-131740: minor readability fix in PyUnstable_GC_VisitObjects (gh-131786)
Minor readability fix in PyUnstable_GC_VisitObjects
Replaces `if (visit_generation())` with `if (visit_generation() < 0)`,
since we are checking for the failure case, and it's confusing to have
that be implicitly `true`.
Steve Dower [Fri, 28 Mar 2025 15:07:57 +0000 (15:07 +0000)]
gh-131423: Update to OpenSSL 3.0.16. (GH-131839)
The bin tag is 3.0.16.1 because we rebuilt without uplink support to fix gh-131804.
This PR also prevents making calls that are now unsafe without uplink, and updates
the tests to property interpret these failures as unsupported.
Thomas Grainger [Thu, 27 Mar 2025 19:06:52 +0000 (19:06 +0000)]
gh-128770: raise warnings as errors in test suite - except for test_socket which still logs warnings, and internal test warnings that are now logged (#128973)
Sam Gross [Wed, 26 Mar 2025 18:38:47 +0000 (14:38 -0400)]
gh-131586: Avoid refcount contention in some "special" calls (#131588)
In the free threaded build, the `_PyObject_LookupSpecial()` call can lead to
reference count contention on the returned function object becuase it
doesn't use stackrefs. Refactor some of the callers to use
`_PyObject_MaybeCallSpecialNoArgs`, which uses stackrefs internally.
This fixes the scaling bottleneck in the "lookup_special" microbenchmark
in `ftscalingbench.py`. However, the are still some uses of
`_PyObject_LookupSpecial()` that need to be addressed in future PRs.
Sam Gross [Wed, 26 Mar 2025 16:08:20 +0000 (12:08 -0400)]
gh-123358: Use `_PyStackRef` in `LOAD_DEREF` (gh-130064)
Concurrent accesses from multiple threads to the same `cell` object did not
scale well in the free-threaded build. Use `_PyStackRef` and optimistically
avoid locking to improve scaling.
With the locks around cell reads gone, some of the free threading tests were
prone to starvation: the readers were able to run in a tight loop and the
writer threads weren't scheduled frequently enough to make timely progress.
Adjust the tests to avoid this.
Co-authored-by: Donghee Na <donghee.na@python.org>
Idan Noiman [Wed, 26 Mar 2025 14:50:37 +0000 (16:50 +0200)]
gh-131741: Add documentation for Windows version detection change in `platform` (#131742)
Document the behavior change between 3.11 & 3.12, where ``platform`` now correctly detects Windows 11 and Windows Server releases past Windows Server 2012.
Co-authored-by: Adam Turner <9087854+AA-Turner@users.noreply.github.com> Co-authored-by: Hugo van Kemenade <1324225+hugovk@users.noreply.github.com> Co-authored-by: Wulian <1055917385@qq.com>
Chris Eibl [Mon, 24 Mar 2025 16:21:57 +0000 (17:21 +0100)]
GH-131521: Uses correct build flags for zlib-ng on Windows (GH-131526)
Do not enable AdvancedVectorExtensions2 for all *.c files, so that the resulting binary can be executed on older CPUs, too. Also enable AdvancedVectorExtensions512 where necessary, and add the ClangCL flags required to enable vector extensions.
In #115684, HopedForLuck noted that `datetime.date.replace()`
documentation was confusing because it looked like it would be changing
immutable objects.
This documentation change specifies that the `replace()` methods in
`datetime` return new objects. This uses similar wording to the
documentation for `datetime.combine()`, which specifies that a new
datetime is returned. This is also similar to wording for
`string.replace()`, except `string.replace()` emphasizes that a "copy"
is returned.
Resolves #115684.
* Include reviewer comments
Thanks Privat33r-dev for the comments!
---------
Co-authored-by: Paul Ganssle <1377457+pganssle@users.noreply.github.com>
When `pathlib._os.magic_open()` is called to open a path in binary mode,
raise `ValueError` if any of the *encoding*, *errors* or *newline*
arguments are given. This matches the `open()` built-in.
Sam Gross [Mon, 24 Mar 2025 13:49:39 +0000 (09:49 -0400)]
gh-128421: Avoid TSAN warnings in `sys._current_frames()` (gh-131548)
This tells TSAN not to sanitize `PyUnstable_InterpreterFrame_GetLine()`.
There's a possible data race on the access to the frame's `instr_ptr`
if the frame is currently executing. We don't really care about the
race. In theory, we could use relaxed atomics for every access to
`instr_ptr`, but that would create more code churn and current compilers
are overly conservative with optimizations around relaxed atomic
accesses.
We also don't sanitize `_PyFrame_IsIncomplete()` because it accesses
`instr_ptr` and is called from assertions within PyFrame_GetCode().
Petr Viktorin [Mon, 24 Mar 2025 13:18:34 +0000 (14:18 +0100)]
gh-128715: Expose ctypes.CField, with info attributes (GH-128950)
- Restore max field size to sys.maxsize, as in Python 3.13 & below
- PyCField: Split out bit/byte sizes/offsets.
- Expose CField's size/offset data to Python code
- Add generic checks for all the test structs/unions, using the newly exposed attrs
Sergey Miryanov [Mon, 24 Mar 2025 12:42:10 +0000 (17:42 +0500)]
gh-131311: Fix additional memory leaks in ctypes (GH-131429)
* Visit keep in StructParam_traverse
* Decref swapped in PyCSimpleType_init
* Decref ob in make_funcptrtype_dict
* Check Pointer_item result while constructing result list in Pointer_subscript
* Fix align and size naming in PyCStructUnionType_update_stginfo
- as discussed in previous PR
Petr Viktorin [Mon, 24 Mar 2025 12:06:21 +0000 (13:06 +0100)]
gh-123909: Remove obsolete note in PyType_FromMetaclass docs (GH-131506)
Docs of the other `PyType_From*` functions link to `PyType_FromMetaclass`,
which noted that they differ for backwards compatibility reasons.
The note is no longer relevant in 3.14.
The other functions have `versionchanged` blurbs.