Mailman 3 June 2021 - Docs-bugs

[issue44388] venv API Docs for EnvBuilder.ensure_directories incorrectly describe behavior
by Matthew Clapp 28 Jul '21

28 Jul '21

New submission from Matthew Clapp <itsayellow+dev(a)gmail.com>: The docs for the venv module, EnvBuilder class, ensure_directories method, describe behavior that doesn't match what its actual behavior is, (and what the code is). I propose to update the documentation of the API to match the actual behavior. https://docs.python.org/3.9/library/venv.html#venv.EnvBuilder.ensure_direct… The last sentence for ensure_directories(env_dir) reads: The directories are allowed to exist already, as long as either clear or upgrade were specified to allow operating on an existing environment directory. After testing and looking at the code (including past commits back over 6 years), this is not true. ensure_directories completely disregards the `upgrade` attribute. Also it allows directories to exist already unconditionally (always operating without error), whether the `clear` attribute is set or not. In addition, if `clear` is not set, it doesn't make any changes to the directories, but helpfully still returns the context of venv paths. Ref: https://github.com/python/cpython/blob/3ce35bfbbe29664942f9a8c50c177a4575a3… ---------- assignee: docs@python components: Documentation messages: 395603 nosy: docs@python, itsayellow priority: normal severity: normal status: open title: venv API Docs for EnvBuilder.ensure_directories incorrectly describe behavior versions: Python 3.10, Python 3.11, Python 3.6, Python 3.7, Python 3.8, Python 3.9 _______________________________________ Python tracker <report(a)bugs.python.org> <https://bugs.python.org/issue44388> _______________________________________

4 5

[issue44399] log rotator cookbook example might waste disk space
by mori-b 28 Jul '21

28 Jul '21

New submission from mori-b <moribirom(a)gmail.com>: In https://docs.python.org/3/howto/logging-cookbook.html#using-a-rotator-and-n…, the log rotator example deletes the original log file after compressing it. However, running on Linux the command "lsof +S1" shows that the deleted original log file might still hold the same disk space, and keep growing. Replacing the command "os.remove(source)" with "os.truncate(source,0)" seems to solve the issue by freeing the original log file disk space at each rotation. ---------- assignee: docs@python components: Documentation messages: 395658 nosy: docs@python, mori-b priority: normal severity: normal status: open title: log rotator cookbook example might waste disk space type: resource usage _______________________________________ Python tracker <report(a)bugs.python.org> <https://bugs.python.org/issue44399> _______________________________________

4 16

[issue44473] logging.handlers.QueueHandler acts unexpected
by Kai Müller 28 Jul '21

28 Jul '21

New submission from Kai Müller <kai.jmueller(a)gmail.com>: According to the docstring of logging.handlers.QueueHandler "The base implementation formats the record to merge the message and arguments, and removes unpickleable items from the record in-place." But, if a just a log message is used w/o any arguments, the arguments are still set to None, which IMHO unexpected. Especially, according to the typeshed project, the "args" is ALWAYS either a dict or a tuple. But in this case, they are set to None which surprised my a lot. Would it be possible to improve the docstring to state, that args, exc_info, exc_text are set to None and that msg is overwritten. If you miss this tiny but very important detail, additional handlers can act very wrong. In addition, it seems to be that the type information needs to be improved in typeshed or are there any other plans on your side? ---------- assignee: docs@python components: Documentation messages: 396243 nosy: docs@python, kai.jmueller priority: normal severity: normal status: open title: logging.handlers.QueueHandler acts unexpected type: enhancement versions: Python 3.10, Python 3.11, Python 3.6, Python 3.7, Python 3.8, Python 3.9 _______________________________________ Python tracker <report(a)bugs.python.org> <https://bugs.python.org/issue44473> _______________________________________

5 6

[issue44514] configparser.rst & bz2.rst leave temp files after 'make doctest'
by Kevin Follstad 13 Jul '21

13 Jul '21

New submission from Kevin Follstad <kfollstad(a)gmail.com>: Both Docs/library/configparser.rst and Docs/library/bz2.rst create untracked temp files on the filesystem when 'make doctest' is run because testcleanup directives are absent in these files. ---------- assignee: docs@python components: Documentation messages: 396541 nosy: docs@python, kfollstad priority: normal severity: normal status: open title: configparser.rst & bz2.rst leave temp files after 'make doctest' type: behavior versions: Python 3.10, Python 3.11, Python 3.8, Python 3.9 _______________________________________ Python tracker <report(a)bugs.python.org> <https://bugs.python.org/issue44514> _______________________________________

3 8

[issue44538] ast.Slice 3.9.6 documentation bug
by Tim 30 Jun '21

30 Jun '21

New submission from Tim <tim.baer(a)c3.ai>: Based off the ast 3.9.6 documentation (https://docs.python.org/3/library/ast.html), we would expect `Slice` to inherit from `expr`. However, looking at `ast.Slice.__mro__` produces the following output: `(<class '_ast.Slice'>, <class '_ast.slice'>, <class '_ast.AST'>, <class 'object'>)`. It appears that instead of inheriting from `expr`, `Slice` inherits from `slice` which appears to be a deprecated type. ---------- assignee: docs@python components: Documentation messages: 396765 nosy: c3-timjbaer, docs@python priority: normal severity: normal status: open title: ast.Slice 3.9.6 documentation bug type: behavior versions: Python 3.9 _______________________________________ Python tracker <report(a)bugs.python.org> <https://bugs.python.org/issue44538> _______________________________________

2 2

[issue34798] pprint ignores the compact parameter for dicts
by Irit Katriel 30 Jun '21

30 Jun '21

Irit Katriel <iritkatriel(a)gmail.com> added the comment: I made PR26967 to break up the paragraph about the constructor params so that it's easier to read, and also added emphasis around the "sequence" point. ---------- versions: +Python 3.10, Python 3.11, Python 3.9 -Python 3.8 _______________________________________ Python tracker <report(a)bugs.python.org> <https://bugs.python.org/issue34798> _______________________________________

1 0

[issue34798] pprint ignores the compact parameter for dicts
by Irit Katriel 30 Jun '21

30 Jun '21

Change by Irit Katriel <iritkatriel(a)gmail.com>: ---------- keywords: +patch pull_requests: +25531 stage: -> patch review pull_request: https://github.com/python/cpython/pull/26967 _______________________________________ Python tracker <report(a)bugs.python.org> <https://bugs.python.org/issue34798> _______________________________________

1 0

[issue44537] Document PGO in devguide
by Guido van Rossum 30 Jun '21

30 Jun '21

New submission from Guido van Rossum <guido(a)python.org>: The only docs for PGO seem to be in the toplevel README.rst. Sinec that is what's shown to all visitors to the GitHub repo, perhaps this information is not so useful there, and instead a section could be added to the devguide? (Ditto for LTO.) Note that for Windows this info is in PCbuild/readme.txt. It's perhaps okay to keep it there, and just mention in the devguide that the Windows info in there. FWIW, maybe we should get rid of the tradition that the toplevel README.rst file's heading gives the version? That tradition dates from the times when source distributions were done as tarballs... It would save one more place to update (there's another mention of the version later in README.rst which still points to 3.10). ---------- assignee: docs@python components: Documentation messages: 396764 nosy: docs@python, gvanrossum priority: normal severity: normal status: open title: Document PGO in devguide versions: Python 3.11 _______________________________________ Python tracker <report(a)bugs.python.org> <https://bugs.python.org/issue44537> _______________________________________

1 1

[issue44523] weakref.proxy documentation might be outdated
by Alexander Neumann 30 Jun '21

30 Jun '21

New submission from Alexander Neumann <aleneum(a)gmail.com>: The documentation currently states: > Proxy objects are not hashable regardless of the referent; this avoids a number of problems related to their fundamentally mutable nature, and prevent their use as dictionary keys. callback is the same as the parameter of the same name to the ref() function. However, it seems with commit 96074de573f82fc66a2bd73c36905141a3f1d5c1 (https://github.com/python/cpython/commit/96074de573f82fc66a2bd73c36905141a3…) hash/reversed pass throughs have been introduced that enable weakref.proxy to be used as a dictionary key. At least the following code fails with Python 3.8 but works with Python 3.9: ``` import weakref class Model: pass m = Model() proxy = weakref.proxy(m) ref = weakref.ref(m) print(hash(m)) print(hash(ref)) print(hash(proxy)) ``` ---------- assignee: docs@python components: Documentation messages: 396637 nosy: aleneum, docs@python priority: normal severity: normal status: open title: weakref.proxy documentation might be outdated type: behavior versions: Python 3.10, Python 3.11, Python 3.9 _______________________________________ Python tracker <report(a)bugs.python.org> <https://bugs.python.org/issue44523> _______________________________________

6 17

[issue38062] Clarify that atexit.unregister matches by equality, not identity
by Irit Katriel 29 Jun '21

29 Jun '21

Change by Irit Katriel <iritkatriel(a)gmail.com>: ---------- resolution: -> fixed stage: patch review -> resolved status: open -> closed _______________________________________ Python tracker <report(a)bugs.python.org> <https://bugs.python.org/issue38062> _______________________________________

1 0