Barney Gale [Sat, 27 Jan 2024 19:59:51 +0000 (19:59 +0000)]
GH-79634: Speed up pathlib globbing by removing `joinpath()` call. (#114623)
Remove `self.joinpath('')` call that should have been removed in 6313cdde.
This makes `PathBase.glob('')` yield itself *without* adding a trailing slash. It's hard to say whether this is more or less correct, but at least everything else is faster, and there's no behaviour change in the public classes where empty glob patterns are disallowed.
Neil Schemenauer [Sat, 27 Jan 2024 03:38:14 +0000 (19:38 -0800)]
gh-113055: Use pointer for interp->obmalloc state (gh-113412)
For interpreters that share state with the main interpreter, this points
to the same static memory structure. For interpreters with their own
obmalloc state, it is heap allocated. Add free_obmalloc_arenas() which
will free the obmalloc arenas and radix tree structures for interpreters
with their own obmalloc state.
Co-authored-by: Eric Snow <ericsnowcurrently@gmail.com>
Barney Gale [Fri, 26 Jan 2024 22:29:28 +0000 (22:29 +0000)]
pathlib tests: annotate tests needing symlinks with decorator (#114625)
Add `@needs_symlinks` decorator for tests that require symlink support in
the path class.
Also add `@needs_windows` and `@needs_posix` decorators for tests that
require a specific a specific path flavour. These aren't much used yet, but
will be later.
- consistently use correct parameter markup
- consistently use submodule name as database name
- improve accuracy of the dbm.dumb.open() spec
- remove dumbdbm class refs and replace them with generic "database object"
- use parameter list for dbm.dumb.open()
- add abbreviation directives for NDBM and GDBM
- consistently spell NDBM as NDBM
- silence broken ndbm class refs
- improve accuracy of dbm.ndbm.open() spec
- use replacement text for NDBM/GDBM file format incompatibility note
- add refs to other parts of the docs (dict, bytes, etc.)
- clarify whichdb() return value by using list markup
- silence refs to example or generic submodule methods (keys, get, etc.)
In 49f90ba we added support for the recursive wildcard `**` in
`pathlib.PurePath.match()`. This should allow arbitrary prefix and suffix
matching, like `p.match('foo/**')` or `p.match('**/foo')`, but there's a
problem: for relative patterns only, `match()` implicitly inserts a `**`
token on the left hand side, causing all patterns to match from the right.
As a result, it's impossible to match relative patterns from the left:
`PurePath('foo/bar').match('bar/**')` is true!
This commit reverts the changes to `match()`, and instead adds a new
`full_match()` method that:
- Allows empty patterns
- Supports the recursive wildcard `**`
- Matches the *entire* path when given a relative pattern
Nikita Sobolev [Thu, 25 Jan 2024 19:46:32 +0000 (22:46 +0300)]
gh-114315: Make `threading.Lock` a real class, not a factory function (#114479)
`threading.Lock` is now the underlying class and is constructable rather than the old
factory function. This allows for type annotations to refer to it which had no non-ugly
way to be expressed prior to this.
---------
Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
Sam Gross [Thu, 25 Jan 2024 18:27:36 +0000 (13:27 -0500)]
gh-112529: Implement GC for free-threaded builds (#114262)
* gh-112529: Implement GC for free-threaded builds
This implements a mark and sweep GC for the free-threaded builds of
CPython. The implementation relies on mimalloc to find GC tracked
objects (i.e., "containers").
Dino Viehland [Thu, 25 Jan 2024 17:34:03 +0000 (09:34 -0800)]
gh-112075: Add try-incref functions from nogil branch for use in dict thread safety (#114512)
* Bring in a subset of biased reference counting:
https://github.com/colesbury/nogil/commit/b6b12a9a94e
The NoGIL branch has functions for attempting to do an incref on an object which may or may not be in flight. This just brings those functions over so that they will be usable from in the dict implementation to get items w/o holding a lock.
There's a handful of small simple modifications:
Adding inline to the force inline functions to avoid a warning, and switching from _Py_ALWAYS_INLINE to Py_ALWAYS_INLINE as that's available
Remove _Py_REF_LOCAL_SHIFT as it doesn't exist yet (and is currently 0 in the 3.12 nogil branch anyway)
ob_ref_shared is currently Py_ssize_t and not uint32_t, so use that
_PY_LIKELY doesn't exist, so drop it
_Py_ThreadLocal becomes _Py_IsOwnedByCurrentThread
Add '_PyInterpreterState_GET()' to _Py_IncRefTotal calls.
Sam Gross [Tue, 23 Jan 2024 20:14:46 +0000 (15:14 -0500)]
gh-101438: Avoid reference cycle in ElementTree.iterparse. (GH-114269)
The iterator returned by ElementTree.iterparse() may hold on to a file
descriptor. The reference cycle prevented prompt clean-up of the file
descriptor if the returned iterator was not exhausted.
mpage [Tue, 23 Jan 2024 19:25:41 +0000 (11:25 -0800)]
gh-113884: Make queue.SimpleQueue thread-safe when the GIL is disabled (#114161)
* use the ParkingLot API to manage waiting threads
* use Argument Clinic's critical section directive to protect queue methods
* remove unnecessary overflow check
Co-authored-by: Erlend E. Aasland <erlend.aasland@protonmail.com>
The `--disable-gil` builds occasionally need to pause all but one thread. Some
examples include:
* Cyclic garbage collection, where this is often called a "stop the world event"
* Before calling `fork()`, to ensure a consistent state for internal data structures
* During interpreter shutdown, to ensure that daemon threads aren't accessing Python objects
This adds the following functions to implement global and per-interpreter pauses:
* `_PyEval_StopTheWorldAll()` and `_PyEval_StartTheWorldAll()` (for the global runtime)
* `_PyEval_StopTheWorld()` and `_PyEval_StartTheWorld()` (per-interpreter)
(The function names may change.)
These functions are no-ops outside of the `--disable-gil` build.
Sam Gross [Tue, 23 Jan 2024 18:05:15 +0000 (13:05 -0500)]
gh-112984: Fix link error on free-threaded Windows build (GH-114455)
The test_peg_generator test tried to link the python313_d.lib library,
which failed because the library is now named python313t_d.lib. The
underlying problem is that the "compiler" attribute was not set when
we call get_libraries() from distutils.
gh-113317: Don't use global clinic instance in bad_argument() (#114330)
Make it possible for a converter to have multiple includes, by collecting
them in a list on the converter instance. This implies converter includes
are added during template generation, so we have to add them to the
clinic instance at the end of the template generation instead of in the
beginning.
Jim Porter [Tue, 23 Jan 2024 08:53:04 +0000 (00:53 -0800)]
gh-89427: Provide the original prompt value for VIRTUAL_ENV_PROMPT (GH-106726)
This improves the implementation in gh-106643.
Previously, venv passed "(<prompt>) " to the activation scripts, but we want
to provide the original value so that users can inspect it in the
$VIRTUAL_ENV_PROMPT env var.
Note: Lib/venv/scripts/common/Activate.ps1 surrounded the prompt value with
parens a second time, so no change was necessary in that file.
Gregory P. Smith [Mon, 22 Jan 2024 23:55:12 +0000 (15:55 -0800)]
Add me to codeowners for hashlib & multiprocessing (#114454)
I already effectively own these. (multiprocessing reluctantly, but I've
spent enough time in the code of late, it is important, and and
championing some changes, so I may as well be looped in there).
This adds support for visiting abandoned pages in mimalloc and improves
the performance of the page visiting code. Abandoned pages contain
memory blocks from threads that have exited. At some point, they may be
later reclaimed by other threads. We still need to visit those pages in
the free-threaded GC because they contain live objects.
This also reduces the overhead of visiting mimalloc pages:
* Special cases for full, empty, and pages containing only a single
block.
* Fix free_map to use one bit instead of one byte per block.
* Use fast integer division by a constant algorithm when computing
block offset from block size and index.
Serhiy Storchaka [Mon, 22 Jan 2024 14:14:42 +0000 (16:14 +0200)]
gh-102512: Turn _DummyThread into _MainThread after os.fork() called from a foreign thread (GH-113261)
Always set a _MainThread as a main thread after os.fork() is called from
a thread started not by the threading module.
A new _MainThread was already set as a new main thread after fork if
threading.current_thread() was not called for a foreign thread before fork.
Now, if it was called before fork, the implicitly created _DummyThread will
be turned into _MainThread after fork.
It fixes, in particularly, an incompatibility of _DummyThread with
the threading shutdown logic which relies on the main thread
having tstate_lock.
Co-authored-by: Marek Marczykowski-Górecki <marmarek@invisiblethingslab.com>