* gh-112529: Remove PyGC_Head from object pre-header in free-threaded build
This avoids allocating space for PyGC_Head in the free-threaded build.
The GC implementation for free-threaded CPython does not use the
PyGC_Head structure.
* The trashcan mechanism uses the `ob_tid` field instead of `_gc_prev`
in the free-threaded build.
* The GDB libpython.py file now determines the offset of the managed
dict field based on whether the running process is a free-threaded
build. Those are identified by the `ob_ref_local` field in PyObject.
* Fixes `_PySys_GetSizeOf()` which incorrectly incorrectly included the
size of `PyGC_Head` in the size of static `PyTypeObject`.
The free-threaded build's GC implementation is non-generational, but was
scheduled as if it were collecting a young generation leading to
quadratic behavior. This increases the minimum threshold and scales it
to the number of live objects as we do for the old generation in the
default build.
Note that the scheduling is still not thread-safe without the GIL. Those
changes will come in later PRs.
A few tests, like "test_sneaky_frame_object" rely on prompt scheduling
of the GC. For now, to keep that test passing, we disable the scaled
threshold after calls like `gc.set_threshold(1, 0, 0)`.
---------
Co-authored-by: Kirill Podoprigora <kirill.bast9@mail.ru>
Co-authored-by: Terry Jan Reedy <tjreedy@udel.edu>
Co-authored-by: Erlend E. Aasland <erlend.aasland@protonmail.com>
Improve misleading TCP server docs and example.
socket.recv(), as documented by the Python reference documentation,
returns at most `bufsize` bytes, and the underlying TCP protocol means
there is no guaranteed correspondence between what is sent by the client
and what is received by the server.
This conflation could mislead readers into thinking that TCP is
datagram-based or has similar semantics, which will likely appear to
work for simple cases, but introduce difficult to reproduce bugs.
Prior to gh-114269, the iterator returned by ElementTree.iterparse was
initialized with the root attribute as None. This restores the previous
behavior.
PyObject_GetBuffer() now raises a SystemError if called with
PyBUF_READ or PyBUF_WRITE as flags. These flags should
only be used with the PyMemoryView_* C API.
The text clearly seems to be referencing `TestFuncAcceptsSequencesMixin`,
for which no target is available. Name the class properly and suppress
the dangling reference.
Methods like `full_match()`, `glob()`, etc, are difficult to make work with
byte paths, and it's not worth the effort. This patch makes `PurePathBase`
raise `TypeError` when given non-`str` path segments.
Raise `UnsupportedOperation` directly, rather than via an `_unsupported()`
helper, to give human readers and IDEs/typecheckers/etc a bigger hint that
these methods are abstract.
Add a configure define for HAVE_PTHREAD_COND_TIMEDWAIT_RELATIVE_NP and
replaces pthread_cond_timedwait() with pthread_cond_timedwait_relative_np()
for relative time when supported in semaphore waiting logic.
Return files and directories from `pathlib.Path.glob()` if the pattern ends
with `**`. This is more compatible with `PurePath.full_match()` and with
other glob implementations such as bash and `glob.glob()`. Users can add a
trailing slash to match only directories.
In my previous patch I added a `FutureWarning` with the intention of fixing
this in Python 3.15. Upon further reflection I think this was an
unnecessarily cautious remedy to a clear bug.
Zachary Ware
Donghee Na
Sam Gross
Oleg Iarygin
Mark Shannon
He Weidong
Nikita Sobolev
Skip Montanaro
Nicholas Hollander
Kirill Podoprigora
Guido van Rossum
Anders Kaseorg
Ayappan Perumal
Tomas R
dependabot[bot]
Christophe Nanteuil
technillogue
srinivasan
Pradyot Ranjan
Hugo van Kemenade
Aidan Holm
Jamie Phan
Bradley Reynolds
Albert Zeyer
Nachtalb
Tian Gao
Alex Waygood
Serhiy Storchaka
Daniel Hollas
Erlend E. Aasland
Brett Cannon
Barney Gale
Diego Russo
Brandt Bucher
Matt Prodani