Petr Viktorin [Wed, 25 Sep 2024 21:30:40 +0000 (23:30 +0200)]
gh-121277: Allow `.. versionadded:: next` in docs (GH-121278)
Make `versionchanged:: next`` expand to current (unreleased) version.
When a new CPython release is cut, the release manager will replace
all such occurences of "next" with the just-released version.
(See the issue for release-tools and devguide PRs.)
Co-authored-by: Adam Turner <9087854+AA-Turner@users.noreply.github.com> Co-authored-by: Hugo van Kemenade <1324225+hugovk@users.noreply.github.com>
gh-119400: make_ssl_certs: update reference test data automatically, pass in expiration dates as parameters #119400 (GH-119401)
* Lib/test/certdata: do not hardcode reference cert data into tests
The script was simply printing the reference data and asking
users to update it by hand into the test suites. This can
be easily improved by writing the data into files and
having the test cases load the files.
* make_ssl_certs: make it possible to pass in expiration dates from command line
Note that in this commit, the defaults are same as they were,
so if nothing is specified the script works as before.
---------
Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com>
gh-119180: Avoid going through AST and eval() when possible in annotationlib (#124337)
Often, ForwardRefs represent a single simple name. In that case, we
can avoid going through the overhead of creating AST nodes and code
objects and calling eval(): we can simply look up the name directly
in the relevant namespaces.
Co-authored-by: Victor Stinner <vstinner@python.org>
Mat S [Wed, 25 Sep 2024 09:24:47 +0000 (10:24 +0100)]
gh-123445: calendar: Improve descriptions for day and month attributes (#123483)
Co-authored-by: Adam Turner <9087854+aa-turner@users.noreply.github.com> Co-authored-by: Eric V. Smith <ericvsmith@users.noreply.github.com> Co-authored-by: Hugo van Kemenade <1324225+hugovk@users.noreply.github.com>
typing.get_origin() does what we need here, without reaching into
typing internals. This shouldn't change any behavior (so I am going
to skip news), but it sets a good example for other users introspecting
typing objects.
Sam Gross [Tue, 24 Sep 2024 20:08:18 +0000 (13:08 -0700)]
gh-123923: Defer refcounting for `f_funcobj` in `_PyInterpreterFrame` (#124026)
Use a `_PyStackRef` and defer the reference to `f_funcobj` when
possible. This avoids some reference count contention in the common case
of executing the same code object from multiple threads concurrently in
the free-threaded build.
* Detect source file encoding.
* Use the "replace" error handler even for UTF-8 (default) encoding.
* Remove the BOM.
* Fix detection of too long lines if they contain NUL.
* Return the head rather than the tail for truncated long lines.
gh-72795: Make positional arguments with nargs='*' or REMAINDER non-required (GH-124306)
This allows to use positional argument with nargs='*' and without default
in mutually exclusive group and improves error message about required
arguments.
gh-63143: Fix parsing mutually exclusive arguments in argparse (GH-124307)
Arguments with the value identical to the default value (e.g. booleans,
small integers, empty or 1-character strings) are no longer considered
"not present".
Petr Viktorin [Tue, 24 Sep 2024 00:40:53 +0000 (02:40 +0200)]
gh-100980: ctypes: Test, document, and fix finalizing _fields_ (GH-124292)
- If setting `_fields_` fails, e.g. with AttributeError, don't set the attribute in `__dict__`
- Document the “finalization” behaviour
- Beef up tests: add `getattr`, test Union as well as Structure
- Put common functionality in a common function
Co-authored-by: Peter Bierma <zintensitydev@gmail.com>
Co-authored-by: Russell Keith-Magee <russell@keith-magee.com> Co-authored-by: T. Wouters <thomas@python.org> Co-authored-by: Adam Turner <9087854+AA-Turner@users.noreply.github.com>
GH-65961: Document the deprecation of `__package__` and `__cached__` (GH-124377)
The code changes for warning related to `__package__` landed in Python 3.12. `__cached__` doesn't have any changes as it isn't used but only set by the import system.
GH-87041: Fix incorrect indentation in argparse help (GH-124230)
In case of usage a long command along with max_help_position more than
the length of the command, the command's help was incorrectly started
on the new line.
Co-authored-by: Pavel Ditenbir <pavel.ditenbir@gmail.com>
gh-119180: Fix annotationlib.ForwardRef.evaluate with no globals (#124326)
We were sometimes passing None as the globals argument to eval(), which makes it
inherit the globals from the calling scope. Instead, ensure that globals is always
non-None. The test was passing accidentally because I passed "annotationlib" as a
module object; fix that. Also document the parameters to ForwardRef() and remove
two unused private ones.
Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com>
Replace the term `Immutable` with a `Hashable` in the `sequence` entry of the Glossary (#124350)
The term `Immutable` in the `sequence` entry of the glossary is used incorrectly, in fact dicts accepts hashable keys, which is not the same as immutable.
Sam James [Mon, 23 Sep 2024 07:03:30 +0000 (08:03 +0100)]
GH-113655: Lower the C recursion limit for HPPA, PPC64 and SPARC (#124264)
Lower the C recursion limit for HPPA, PPC64 and SPARC, as they use
relatively large stack frames that cause e.g. `test_descr` to hit
a stack overflow. According to quick testing, it seems that values
around 8000 are max for HPPA and PPC64 (ELFv1 ABI) and 7000 for SPARC64.
To keep things safe, let's use 5000 for PPC64 and 4000 for SPARC.
Terry Jan Reedy [Sun, 22 Sep 2024 19:51:09 +0000 (15:51 -0400)]
gh-112938: IDLE - Fix uninteruptable hang when Shell gets rapid continuous output. (#124310)
https://github.com/python/cpython/issues/88496 replaced text.update with text.update_idletasks in colorizer.py and outwin.py to fix test failures on macOS. While theoretically correct, the result was Shell freezing when receiving continuous short strings to print. Test: `while 1: 1`.
The guess is that there is no idle time in which to do the screen update. Reverting the change in one of the files,
outwin, fixes the issue. Colorizer runs ever 1/20 second and seems to work fine.
When running test-outwin on macOS, alias 'update'
to 'update_idletasks on the text used for testing.
Add a helper function that checks whether the test suite is running
inside a systemd-nspawn container, and skip the few tests failing
with `--suppress-sync=true` in that case. The tests are failing because
`--suppress-sync=true` stubs out `fsync()`, `fdatasync()` and `msync()`
calls, and therefore they always return success without checking for
invalid arguments.
Call `os.open(__file__, os.O_RDONLY | os.O_SYNC)` and check the errno to
detect whether `--suppress-sync=true` is actually used, and skip
the tests only in that scenario.
Debian (and derivatives) provide a /usr/bin/pager binary, managed by the
alternatives system, that always points to an available pager utility.
Allow _pyrepl to use it, to follow system policy.
This is a very trivial change, from a patch that Debian has been
carrying since 2.7 era. Seems appropriate to upstream.
https://bugs.debian.org/799555