* Replace _Py_ALIGN_AS(V) by _Py_ALIGNED_DEF(N, T)
This is now a common façade for the various `_Alignas` alternatives,
which behave in interesting ways -- see the source comment.
The new macro (and MSVC's `__declspec(align)`) should not be used
on a variable/member declaration that includes a struct declaraton.
A workaround is to separate the struct definition.
Do that for `PyASCIIObject.state`.
* Specify minimum PyGC_Head and PyObject alignment
As documented in InternalDocs/garbage_collector.md, the garbage collector
stores flags in the least significant two bits of the _gc_prev pointer
in struct PyGC_Head. Consequently, this pointer is only capable of storing
a location that's aligned to a 4-byte boundary.
Encode this requirement using _Py_ALIGNED_DEF.
This patch fixes a segfault in m68k, which was previously investigated
by Adrian Glaubitz here:
https://lists.debian.org/debian-68k/2024/11/msg00020.htmlhttps://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1087600
Original patch (using the GCC-only Py_ALIGNED) by Finn Thain.
Co-authored-by: Finn Thain <fthain@linux-m68k.org>
Co-authored-by: Victor Stinner <vstinner@python.org>
Co-authored-by: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de>
When shutdown is called with wait=False, the executor thread keeps running
even after the ProcessPoolExecutor's state is reset. The executor then tries
to replenish the worker processes pool resulting in an error and a potential hang
when it comes across a worker that has died. Fixed the issue by having
_adjust_process_count() return without doing anything if the ProcessPoolExecutor's
state has been reset.
Added unit tests to validate two scenarios:
max_workers < num_tasks (exception)
max_workers > num_tasks (exception + hang)
Use critical sections to make heapq methods that update the heap thread-safe when the GIL is disabled.
---------
Co-authored-by: mpage <mpage@meta.com>
* Cleanup imports and update module docstring.
* Simplify detection of SIMD support.
* Correctly guard `update()` cases.
* Rewrite `py_blake2b_or_s_new` and rename it to `py_blake2_new`.
* Rewrite `blake2_blake2b_copy_locked` and `py_blake2_clear`.
* Refactor computations of `digest` and `hexdigest`.
* Simplify `py_blake2b_get_name` and `py_blake2b_get_block_size`.
* Add `hacl_get_blake2_info` to extract static BLAKE-2 information.
This new helper is used by `py_blake2b_get_digest_size`, but can
be later used to expose `key_length` more easily.
Append the defect to defects instead of to the parse tree.
Co-authored-by: Bénédikt Tran <10796600+picnixz@users.noreply.github.com>
Co-authored-by: Hugo van Kemenade <1324225+hugovk@users.noreply.github.com>
Fix Py_RETURN_NONE, Py_RETURN_TRUE and Py_RETURN_FALSE macros in the
limited C API 3.11 and older:
Don't treat Py_None, Py_True and Py_False as immortal.
We were incorrectly handling a few opcodes that leave their operands on the stack. Treat all of these conservatively; assume that they always leave operands on the stack.
On Windows, the `_PyOS_SigintEvent()` event handle is used to interrupt
the main thread when Ctrl-C is pressed. Previously, we also waited on
the event from other threads, but ignored the result. However, this can
race with interpreter shutdown because the main thread closes the handle
in `_PySignal_Fini` and threads may still be running and using mutexes
during interpreter shtudown.
Only use `_PyOS_SigintEvent()` in the main thread in parking_lot.c, like
we do in other places in the CPython codebase.
Run each example as a subtest in unit tests synthesized by
doctest.DocFileSuite() and doctest.DocTestSuite().
Add the doctest.DocTestRunner.report_skip() method.
OpenSSL and HACL*-based hash functions constructors now support both `data` and `string` parameters.
Previously these constructor functions inconsistently supported sometimes `data` and sometimes `string`,
while the documentation expected `data` to be given in all cases.
