Brett Cannon [Wed, 7 May 2025 19:10:53 +0000 (12:10 -0700)]
GH-133600: Refactor `wasi.py` into `wasi/__main__.py`
Along the way, also move related files into their appropriate directories and update `.github/CODEOWNERS`.
gh-69426: HTMLParser: only unescape properly terminated character entities in attribute values (GH-95215)
According to the HTML5 spec, named character references in attribute values
should only be processed if they are not followed by an ASCII alphanumeric,
or an equals sign.
T. Wouters [Tue, 6 May 2025 11:23:10 +0000 (13:23 +0200)]
gh-132917: Fix data race detected by tsan (#133508)
Fix data race detected by tsan
(https://github.com/python/cpython/actions/runs/14857021107/job/41712717208?pr=133502):
young.count can be modified by other threads even while the gcstate is
locked.
This is the simplest fix to (potentially) unblock beta 1, although this
particular code path seems like it could just be an atomic swap followed by
an atomic add, without having the lock at all.
Emma Smith [Tue, 6 May 2025 00:38:08 +0000 (17:38 -0700)]
gh-132983: Add the `compression.zstd` pacakge and tests (#133365)
Co-authored-by: Adam Turner <9087854+AA-Turner@users.noreply.github.com> Co-authored-by: Gregory P. Smith <greg@krypto.org> Co-authored-by: Tomas R. <tomas.roun8@gmail.com> Co-authored-by: Rogdham <contact@rogdham.net>
Eric Snow [Mon, 5 May 2025 23:25:26 +0000 (17:25 -0600)]
gh-132775: Do Not Set __name__ to __main__ With _PyPickle_GetXIData() (gh-133472)
This is a follow-up to gh-133107. I realized that we could end up with an
infinite recursion if we try to run a function from __main__ in a subinterpreter.
"Stateless" code is a function or code object which does not rely on external state or internal state.
It may rely on arguments and builtins, but not globals or a closure. I've left a comment in
pycore_code.h that provides more detail.
We also add _PyFunction_VerifyStateless(). The new functions will be used in several later changes
that facilitate "sharing" functions and code objects between interpreters.
gh-132917: Check resident set size (RSS) before GC trigger. (gh-133399)
For the free-threaded build, check the process resident set size (RSS)
increase before triggering a full automatic garbage collection. If the RSS
has not increased 10% since the last collection then it is deferred.
Add *resolve_host* keyword-only argument to `url2pathname()`, defaulting to
false. When set to true, we call `socket.gethostbyname()` to resolve the
URL hostname.
Co-authored-by: Bénédikt Tran <10796600+picnixz@users.noreply.github.com> Co-authored-by: Adam Turner <9087854+AA-Turner@users.noreply.github.com> Co-authored-by: Steve Dower <steve.dower@microsoft.com>
Clang-cl detects that on 32-bit builds the variable is always smaller than the value. But since the same code is used for other architectures, we can't just _fix_ it. This cast avoids the tautological-constant-out-of-range-compare warning.
gh-61103: don't use C _Complex types to implement F/D/G in ctypes (GH-133237)
According to the C standard, the memory representation of _Complex types
is equivalent to 2-element arrays. Unlike _Complex, arrays are always available.
- drop _complex.h header
- use appropriate real arrays to replace complex types
Co-authored-by: Lisandro Dalcin <dalcinl@gmail.com> Co-authored-by: Petr Viktorin <encukou@gmail.com> Co-authored-by: Stan Ulbrych <89152624+StanFromIreland@users.noreply.github.com>
either line could be where the inspection finds the foo()
function as after ready is sent, the process may not have made progress
onto the next line yet. "solve" by putting the statements on the same
line.
mpage [Mon, 5 May 2025 04:00:11 +0000 (21:00 -0700)]
gh-133371: Don't optimize `LOAD_FAST` instructions whose local is killed by `DELETE_FAST` (#133383)
In certain cases it's possible for locals loaded by `LOAD_FAST` instructions
to be on the stack when the local is killed by `DEL_FAST`. These `LOAD_FAST`
instructions should not be optimized into `LOAD_FAST_BORROW` as the strong
reference in the frame is killed while there is still a reference on the stack.
Samuel [Sun, 4 May 2025 23:03:38 +0000 (00:03 +0100)]
gh-131421: fix ASDL grammar for `Dict` to have an `expr?*` keys field (#131419)
In the `ast` documentation for Python:
* https://docs.python.org/3/library/ast.html#ast.Dict
it is made clear that:
> When doing dictionary unpacking using dictionary literals the expression to be expanded goes in the values list, with a `None` at the corresponding position in `keys`.
Hence, `keys` is really a `expr?*` and *not* a `expr*`.
gh-109700: Improve stress tests for interpreter creation (GH-109946)
* Ensure that destructors are called in the test that created interpreters, not after finishing it.
* Try to create/run interpreters in threads simultaneously.
* Mark tests that requires over 6GB of memory with bigmemtest.
As noted on the issue, making get_annotate_function() support both types and
mappings is problematic because one object may be both. So let's add a new one
that works with any mapping.
This leaves get_annotate_function() not very useful, so remove it.
gh-81793: Always call linkat() from os.link(), if available (GH-132517)
This fixes os.link() on platforms (like Linux and OpenIndiana) where the
system link() function does not follow symlinks.
* On Linux, it now follows symlinks by default and if
follow_symlinks=True is specified.
* On Windows, it now raises error if follow_symlinks=True is passed.
* On macOS, it now raises error if follow_symlinks=False is passed and
the system linkat() function is not available at runtime.
* On other platforms, it now raises error if follow_symlinks is passed
with a value that does not match the system link() function behavior
if if the behavior is not known.
Co-authored-by: Joachim Henke <37883863+jo-he@users.noreply.github.com> Co-authored-by: Thomas Kluyver <takowl@gmail.com>
gh-122559: Synchronize C and Python implementation of the io module about pickling (GH-122628)
In the C implementation, remove __reduce__ and __reduce_ex__ methods
that always raise TypeError and restore __getstate__ methods that always
raise TypeErrori.
This restores fine details of the pre-3.12 behavior and unifies
both implementations.