Co-authored-by: Ken Jin <28750310+Fidget-Spinner@users.noreply.github.com>
Co-authored-by: Brandt Bucher <brandt@python.org>
Co-authored-by: Hugo van Kemenade <1324225+hugovk@users.noreply.github.com>
* Add terminator to JIT code when halting due to invalid dependencies
* 📜🤖 Added by blurb_it.
---------
Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com>
* gh-143010: Prevent a TOCTOU issue by gh-143010: Prevent a TOCTOU issue by only calling open once
RDM: per AZero13's research the 'x' option did not exist when this code was written, This
modernization can thus drop the fd trick in _create_carefully and just use open with 'x' to achieve the same goal more securely.
Co-authored-by: sobolevn <mail@sobolevn.me>
mmapmodule: remove unreachable code in Windows error path
Remove an unreachable `return NULL` after `PyErr_SetFromWindowsErr()` in
the Windows mmap resize error path.
Signed-off-by: Yongtao Huang <yongtaoh2022@gmail.com>
* gh-70647: Better promote how to safely parse yearless dates in datetime.
Every four years people encounter this because it just isn't obvious.
This moves the footnote up to a note with a code example.
We'd love to change the default year value for datetime but doing
that could have other consequences for existing code. This documented
workaround *always* works.
* doctest code within note is bad, dedent.
* Update to match the error message.
* remove no longer referenced footnote
* ignore the warning in the doctest
* use Petr's suggestion for the docs to hide the warning processing
* cover date.strptime (3.14) as well
The use of memmove and _Py_memory_repeat were not thread-safe in the
free threading build in some cases. In theory, memmove and
_Py_memory_repeat can copy byte-by-byte instead of pointer-by-pointer,
so concurrent readers could see uninitialized data or tearing.
Additionally, we should be using "release" (or stronger) ordering to be
compliant with the C11 memory model when copying objects within a list.
This makes generator frame state transitions atomic in the free
threading build, which avoids segfaults when trying to execute
a generator from multiple threads concurrently.
There are still a few operations that aren't thread-safe and may crash
if performed concurrently on the same generator/coroutine:
* Accessing gi_yieldfrom/cr_await/ag_await
* Accessing gi_frame/cr_frame/ag_frame
* Async generator operations
If *a* is an integer, the sign of *a* is discarded in the C source code. Clarify this behavior to prevent foot guns, where a common use case might naively assume that flipping the sign will produce different sequences (e.g. for a train/test split of a synthetic data generator in machine learning).
Co-authored-by: Adam Turner <9087854+AA-Turner@users.noreply.github.com>
Now that we specialize range iteration in the interpreter for the common
case where the iterator has only one reference, there's not a
significant performance cost to making the iteration thread-safe.
