repo stringclasses 1
value | instance_id stringlengths 20 22 | problem_statement stringlengths 126 60.8k | merge_commit stringlengths 40 40 | base_commit stringlengths 40 40 |
|---|---|---|---|---|
python/cpython | python__cpython-106211 | # Emscripten import trampoline is not needed
As far as I know the import isn't actually needed. It was originally added to work around typos in the CPython test suite, but those typos have since been fixed and I am unaware of any user code that requires the trampoline.
<!-- gh-linked-prs -->
### Linked PRs
* gh-10621... | e7bc8d16364bde54487eab349a29d58345e35f28 | cea9d4ea82abcb2c6f1d83a2fe819859da4bbda4 |
python/cpython | python__cpython-106201 | # Multiple unused imports in the source code
Found by https://github.com/python/core-workflow/issues/505
Affected:
```
Lib/poplib.py:454:5: F811 redefinition of unused 'sys' from line 19
Lib/test/test_capi/test_misc.py:1826:9: F811 redefinition of unused 'json' from line 9
Lib/test/test_capi/test_misc.py:1905... | d830c4a944bcdcc8fe729a60f438fc762965eec1 | 6c60684bf5d34fae27a2f6a142ff794b38cefe1b |
python/cpython | python__cpython-106198 | # `test_multiple_inheritance_buffer_last` is duplicated in `test_buffer`
Based on https://github.com/python/core-workflow/issues/505
Source: https://github.com/python/cpython/blob/bbf722dcd39c66418e45991dcf1cdf140c2ce20e/Lib/test/test_buffer.py#L4697-L4754
There are two different tests with the same name. One mu... | c283a0cff5603540f06d9017e484b3602cc62e7c | 18f51f91e24402a24a0daa53fcbc81a5a2e9af94 |
python/cpython | python__cpython-106196 | # `test_curses` has duplicated tests
From https://github.com/python/core-workflow/issues/505
```
Lib/test/test_curses.py:1371:5: F811 redefinition of unused 'test_move_left' from line 1362
```
Source: https://github.com/python/cpython/blob/bbf722dcd39c66418e45991dcf1cdf140c2ce20e/Lib/test/test_curses.py#L1362-... | 3fb7c608e5764559a718ce8cb81350d7a3df0356 | 4bde89462a95e5962e1467cfc1af5a6094c0c858 |
python/cpython | python__cpython-109139 | # `test_monitoring` has duplicated tests
Based on https://github.com/python/core-workflow/issues/505 by @hugovk
```
Lib/test/test_monitoring.py:973:5: F811 redefinition of unused 'test_line_then_instruction' from line 950
Lib/test/test_monitoring.py:978:5: F811 redefinition of unused 'test_instruction_then_line'... | ea530f2f9ae63e81c22a1818bec0a650ccf758d2 | 17d65547df55eaefe077c45242a7f2d175961dfd |
python/cpython | python__cpython-107016 | # email.parser header-only parsing records MultipartInvariantViolationDefect for valid multipart emails
# Bug report
A valid multipart email message, when parsed with `email.parser.HeaderParser(policy=email.policy.default)` will record a `email.errors.MultipartInvariantViolationDefect`.
If the parser isn't going... | c65592c4d6d7552fb6284442906a96a6874cb266 | 54632528eeba841e4a8cc95ecbd84c9aca8eef57 |
python/cpython | python__cpython-106187 | # `test_traceback` does not run one set of `CPythonTracebackErrorCaretTests`
Based on @hugovk work in https://github.com/python/core-workflow/issues/505 I found that we have two duplicated class `CPythonTracebackErrorCaretTests` here: https://github.com/python/cpython/blob/a3dd8cce58fe2b27eea4eed572d086dc8a7e1bb8/Lib/... | 7c89f1189229c5c67a3766e24ecf00cde658b7fd | 233b8782886939176982a90f563d552757cbf34e |
python/cpython | python__cpython-106183 | # `sys.getfilesystemencoding()` should return interned string
# Feature or enhancement
Intern `sys.getfilesystemencoding()` output.
# Pitch
fsencoding string may be cached and reused.
(After `sys._enablelegacywindowslegacyfsencoding()` is removed, we can cache fsencoding object.)
<!-- gh-linked-prs -->
##... | f1034ba7f67400e7ed7e299dc8b22521c4e43246 | 77ddc9a7b1b28c8b8aee6cc97e483185a56819a6 |
python/cpython | python__cpython-106164 | # C API: Check in PyTuple_SET_ITEM() and PyList_SET_ITEM()
The PyTuple_SET_ITEM() and PyList_SET_ITEM() functions don't check that the index is valid. It should be checked with an assertion.
<!-- gh-linked-prs -->
### Linked PRs
* gh-106164
* gh-111480
* gh-111618
* gh-111683
<!-- /gh-linked-prs -->
| 3f8483cad2f3b94600c3ecf3f0bb220bb1e61d7d | 161012fc25910a47423bae8012398bf519a88140 |
python/cpython | python__cpython-106181 | # test_array modifies `warnings.filter`
```
$ ./python -m test test_array --fail-env-changed -m test.test_array.ArrayReconstructorTest.test_error
0:00:00 load avg: 17.75 Run tests sequentially
0:00:00 load avg: 17.75 [1/1] test_array
Warning -- warnings.filters was modified by test_array
Warning -- Before: (140... | a3dd8cce58fe2b27eea4eed572d086dc8a7e1bb8 | 541a10f9ed193a79aeea5e244bca6f6485d0689f |
python/cpython | python__cpython-106167 | # Fix `test_gzip` failure under WASI
https://github.com/python/cpython/commit/1858db7cbdbf41aa600c954c15224307bf81a258 introduced a dependency on `zlib` in `test_gzip` (see https://buildbot.python.org/all/#/builders/1046/builds/2315 for the first failure), but `zlib` is not guaranteed to exist.
/cc @Yhg1s
<!-- gh... | 161012fc25910a47423bae8012398bf519a88140 | 84caa3324aaefb900895de2f946607cfdbe1be70 |
python/cpython | python__cpython-106161 | # [3.12] cProfile counts 0 primitive calls for builtins.exec in certain scenario
# Bug report
In 3.12, cProfile can count 0 primitive calls and miscalculate cumulative time in the following scenario.
For this layout:
```
script.py
project/
typing.py
```
And the file contents:
script.py
```python
i... | cea9d4ea82abcb2c6f1d83a2fe819859da4bbda4 | 7b2d94d87513967b357c658c6e7e1b8c8d02487d |
python/cpython | python__cpython-106150 | # tidy up the division of work between compile.c and flowgraph.c
Following the split of the compiler into 3 parts in #87092, the resolution of jump targets is still in the optimizer, but it should move to the assembler.
There are also a few operations in compile.c that could move into the optimizer.
<!-- gh-link... | 529088100952b31797a29ef7e0f5716613b32d66 | eaa1eae55ea66d74c5303924320185dac74d4eb1 |
python/cpython | python__cpython-106224 | # AST nodes for PEP 695 type param syntax do not require `end_lineno` nor `end_col_offset`
# Bug report
Unlike AST nodes for 3.10 match syntax, AST nodes for 3.12 type param syntax (PEP 695) do not require `end_lineno` nor `end_col_offset`.
For 3.10 match syntax, this question was discussed [here](https://github.c... | 46c1097868745eeb47abbc8af8c34e8fcb80ff1d | 904aef994262383ae916545908f0578c2d53cf31 |
python/cpython | python__cpython-106143 | # Reorder some fields to facilitate out-of-process inspection
Some of the relevant fields in the interpreter state and the frame state in 3.12 are **very** challenging to fetch from out of process tools because they are in offsets that depend on compilation or platform variables that are different in different platfor... | 2d5a1c281161d037148ffb5983decc6d31c2557d | bb578a0c304dffe43bb28b36b2b1c9153c78b659 |
python/cpython | python__cpython-106271 | # Add more cases to `test_patma`
While anwsering https://github.com/python/cpython/issues/106133 I've noticed that `test_patma` does not have tests for some corner-cases:
1. Case `case [x] | x: ...` is mentined in https://peps.python.org/pep-0634/#capture-patterns but is never tested, the closest we have to it is `ca... | 904aef994262383ae916545908f0578c2d53cf31 | 2062e115017d8c33e74ba14adef2a255c344f747 |
python/cpython | python__cpython-106124 | # Modules/_sha3 is created by configure, but no longer an actual source directory
# Bug report
I noticed that `make distclean` didn't compare correctly with a fresh clone (after accounting for files I expected it wouldn't be expected to clean up). It left an empty `Modules/_sha3` directory which used to hold files,... | 0345b0c2bbf251a0f475cf53e0fb04c79a220e52 | 51fc72511733353de15bc633a3d7b6da366842e4 |
python/cpython | python__cpython-106120 | # 3.12.0b3 build issue: 'O_CLOEXEC' undeclared in sysmodule.c
I am building Python 3.12.0b3 (3.11.X builds fine) on CentOS 5 using glibc-2.5 and gcc-4.8.5 and I encounter this build error:
gcc-4.8 -std=gnu11 -pthread -c -fno-strict-overflow -Wsign-compare -DNDEBUG -g -O3 -Wall -std=c11 -Wextra -Wno-unused-parameter... | 6c60684bf5d34fae27a2f6a142ff794b38cefe1b | bbf722dcd39c66418e45991dcf1cdf140c2ce20e |
python/cpython | python__cpython-106112 | # The zipapp module documentation includes out of date notes on creating a Windows executable
# Documentation
As noted in #72434, the zipapp documentation contains a section covering "How to create a Windows executable". This section was never entirely complete, and is now inaccurate as the `distutils` library is n... | 5d4dbf0e309255e5bce9e31d805a8f950ebf9161 | 1a2bc94fc2bbdf5f810b441ebbbd8fec95a3207c |
python/cpython | python__cpython-106109 | # Docs: dataclasses: wrong error documented
# Documentation
[Data Classes - Mutable default values](https://docs.python.org/3/library/dataclasses.html#mutable-default-values) says in the second blockquote
> x: list = [] # This code raises **ValueError**
but in the explanantion (two paragraphs down)... | 512f299e557f4ab60768d36cee9968bd92116367 | 219effa876785408a87bd6acb37c07ee0d25f3f9 |
python/cpython | python__cpython-106875 | # Segmentation fault in 3.11.4, 3.12.0b3; _PyInterpreterFrame ownership issue
# Crash report
This test case (minimized from a crash in the Zulip test suite) causes a segmentation fault in Python 3.11.4 and 3.12.0b3.
https://github.com/andersk/python-segfault
```console
$ git clone https://github.com/andersk/... | 557b05c7a5334de5da3dc94c108c0121f10b9191 | 052a0d1106fa3ee0c955a3b7ba48e82c49424e20 |
python/cpython | python__cpython-106085 | # C API: Remove private functions from abstract.h
Over the years, we accumulated many private functions as part of the public C API in abstract.h header file. I propose to remove them: move them to the internal C API.
<!-- gh-linked-prs -->
### Linked PRs
* gh-106085
* gh-106088
* gh-106103
* gh-106106
* gh-106159
<!... | 00e75a33728cdad7c10088acc36bc55b2f4a0efe | 93a970ffbce58657cc99305be69e460a11371730 |
python/cpython | python__cpython-106079 | # Isolate the `_decimal` extension module
This issue is used to track the PRs split from [gh-103092](https://github.com/python/cpython/pull/103381). Isolate the `_decimal` extension module by:
* Establish a global module state and convert static types to heap types
* Move other global static variables to the glob... | fb0d9b9ac1ec3ea13fae8b8ef6a4f0a5a80482b3 | 0e24499129f3917b199a6d46fa33eeedd2c447fc |
python/cpython | python__cpython-106090 | # `asyncio.__init__` does not include `asyncio.taskgroups` in its `__all__`
I'm not sure if I should file this as a bug or a feature...I just noticed it in [a discussion](https://discuss.python.org/t/add-the-export-keyword-to-python/28444/13), and thought I'd raise the issue. It's possible this is intended behavior.
... | a12e8ffb49e05a1d1874389318911ce9685db232 | 3eeb8c89063d5ac22c0b1d26e4ae2fd12c149650 |
python/cpython | python__cpython-106108 | # Memory leak in AST parsing (OSS-Fuzz #60074)
# Bug report
Reported by OSS-Fuzz (issue [60074](https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=60074)).
Build cpython with:
```sh
CC=clang CFLAGS="-fsanitize=address" LDFLAGS="-fsanitize=address" ./configure --prefix=<prefix>
ASAN_OPTIONS=detect_leaks=0... | 24fb627ea7a4d57cf479b7516bafdb6c253a1645 | e1d45b8ed43e1590862319fec33539f8adbc0849 |
python/cpython | python__cpython-106515 | # Possessive quantifier matches where an ordinary quantifier doesn't
```python
>>> import re
>>> re.fullmatch('(?:ab?c)*', 'a')
>>> re.fullmatch('(?:ab?c)*+', 'a')
<re.Match object; span=(0, 1), match='a'>
>>>
```
I'm not sure that I completely understand the behavior of the possessive quantifiers, but I thin... | 7b6e34e5baeb4162815ffa4d943b09a58e3f6580 | 73507382ac184a72b59ebb0c2f85e8b1d2dfa58e |
python/cpython | python__cpython-106082 | # Improve error message from `os.fspath` if `__fspath__` is set to `None`
# Feature or enhancement
A common Python idiom is to set a magic method to `None` in a subclass if you want to disable the behaviour that the magic method enables. If you do so, a nice error message will be given if a user tries to "call" the... | 93a970ffbce58657cc99305be69e460a11371730 | 8c24a837371439b8e922ff47275085b581f510c5 |
python/cpython | python__cpython-106034 | # PyDict_GetItem and PyObject_HasAttr should not be used
These functions are broken by design, because they discard any exceptions raised inside, including MemoryError and KeyboardInterrupt. There were several rounds of getting rid of them in past (for example 567eba1852ed89e5cf93dbce33f7e2ca73e8f05d, #11112, #75753),... | 1d33d5378058671bfabb6f4d4b5bfd4726973ff9 | 41ad4dfc04c201728ce9fa12b1a96922dd15a368 |
python/cpython | python__cpython-106031 | # Miscellaneous fixes in Python/suggestions.c
* PyDict_GetItem() and PyObject_HasAttr() suppress arbitrary errors and
should not be used.
* PyUnicode_CompareWithASCIIString() only works if the second argument
is ASCII string.
* Refleak in get_suggestions_for_name_error.
* Use of borrowed pointer after possibl... | c8c162ef5294cddb7ac75fe93ab918e5661c68ee | 9499b0f138cc53b9a2590350d0b545d2f69ee126 |
python/cpython | python__cpython-106257 | # C API: What's the status of the half-private FASTCALL calling convention?
Hi,
In 2017, I added a new experimental FASTCALL calling convention. See my articles about it: [The start of the FASTCALL project](https://vstinner.github.io/start-fastcall-project.html) and [FASTCALL microbenchmarks](https://vstinner.githu... | 8c5f74fc89e35827c52753fe620b32207d537319 | e7bc8d16364bde54487eab349a29d58345e35f28 |
python/cpython | python__cpython-106017 | # Crash in test_import: Assertion error about monitoring version.
This
`./python -m test -j1 -R 3:3 test_import -v -m test_concurrency -m test___cached___legacy_pyc -m test_package___cached___from_pyc`
Crashes with `Assertion `code->_co_instrumentation_version == tstate->interp->monitoring_version' failed.`
<!-... | 9339d70ac2d45743507e320e81e68acf77e366af | a72683ba8e0337650cc490dbe593a5e46aba60cb |
python/cpython | python__cpython-106003 | # Make implicit boolean conversions explicit
...as discussed in https://github.com/faster-cpython/ideas/issues/568.
By adding a dedicated instruction for converting values to bool, we can easily specialize the conditions of all remaining branches in the bytecode while keeping the branches themselves as "dumb" and s... | 7b2d94d87513967b357c658c6e7e1b8c8d02487d | 6e9f83d9aee34192de5d0ef7285be23514911ccd |
python/cpython | python__cpython-106005 | # C API: Add PyDict_GetItemRef() function
The PyDict C API has a bad history. PyDict_GetItem() ignores all exception: error on hash(), error on "key == key2", KeyboardInterrupt, etc. PyDict_GetItemWithError() was added to fix this design. Moreover, Python 3.9 and older allowed to call PyDict_GetItem() with the GIL rel... | 41ca16455188db806bfc7037058e8ecff2755e6c | 0ba07b2108d4763273f3fb85544dde34c5acd40a |
python/cpython | python__cpython-105995 | # asyncio.EventLoop.start_tls() returns None
# Bug report
The documentation says the function should return a transport:
https://docs.python.org/3/library/asyncio-eventloop.html#asyncio.loop.start_tls
But, in some cases it actually returns None. It looks like this may happen when it receives a closing transport... | 6b52a581c151914e59c8c367a03bc7309713a73b | a3dd8cce58fe2b27eea4eed572d086dc8a7e1bb8 |
python/cpython | python__cpython-105989 | # Crash in `_asyncio._swap_current_task` due to improper reference counting
First appeared in a474e04388c2ef6aca75c26cb70a1b6200235feb, where given function was introduced.
Repro:
```python
import _asyncio
class DummyLoop:
pass
class DummyTask:
pass
l = DummyLoop()
_asyncio._swap_current_task(... | d2cbb6e918d9ea39f0dd44acb53270f2dac07454 | 4849a80dd1cbbc5010e8749ba60eb91a541ae4e7 |
python/cpython | python__cpython-105980 | # Assertion failure in `_imp.get_frozen_object` if data object contains bad marshal data
Repro:
```python
import _imp
>>> _imp.get_frozen_object('x', b"6\'\xd5Cu\x12")
```
Output:
```
python: Objects/object.c:541: PyObject_Repr: Assertion `!_PyErr_Occurred(tstate)' failed.
Aborted (core dumped)
```
I'll subm... | cd5280367a3a7065d13b8f7234474f7a2e9a18fd | 46a3190fcf8580f322047395408cd60feba67041 |
python/cpython | python__cpython-105976 | # Behaviour change in py312 for protocols with non-callable members and custom `__subclasshook__` methods
# Bug report
On Python 3.11:
```pycon
>>> from typing import *
>>> @runtime_checkable
... class Foo(Protocol):
... x = 1
... @classmethod
... def __subclasshook__(cls, other):
... ... | 9499b0f138cc53b9a2590350d0b545d2f69ee126 | 968435ddb1c1af9333befb26f7970cded8a5c710 |
python/cpython | python__cpython-112615 | # crc32 function outputs wrong result for large data on the macOS arm64 platform
# Bug report
The functions zlib.crc32 and binascii.crc32 share the problematic behavior. When computing the CRC for data >= 2GB macOS arm64 binaries result in different values than all other platforms such as macOS x64, Windows x64, Li... | 4eddb4c9d9452482c9af7fa9eec223d12b5a9f33 | a1551b48eebb4a68fda031b5ee9e5cbde8d924dd |
python/cpython | python__cpython-105939 | # DeprecationWarning when escaping curly braces in an f-string
# Bug report
A `DeprecationWarning` is emitted when escaping a curly brace in an f-string. The correct warning is a `SyntaxWarning` as of Python 3.12.
```python3
cpython on test-fstring-syntaxwarnings [$] via C v14.0.3-clang via 🐍 pyenv 3.11.3 to... | 6586cee27f32f0354fe4e77c7b8c6e399329b5e2 | 155577de1b6a7f4404b2bf90bcc1a588201550da |
python/cpython | python__cpython-108671 | # Surprising behaviour with compileall -s STRIPDIR parameter
The '-s' stripdir parameter to compileall has very unusual semantics that could lead to unexpected and potentially confusing outcomes.
For example, when run like this: `python -m compileall -s/path/to/build/dst -p /lib /path/to/build/src/file.py`
The s... | 3726cb0f146cb229a5e9db8d41c713b023dcd474 | 52e902ccf0178d7a3f26de4bba7922b01c0d4d3c |
python/cpython | python__cpython-105928 | # C API: Add PyWeakref_GetRef() function
PyWeakref_GET_OBJECT() and PyWeakref_GetObject() return a borrowed reference to the object, or a borrowed reference to None if the object has been finalized. This API is error-prone and not easy to use.
This API was discussed in 2016: https://mail.python.org/archives/list/py... | 7f97c8e367869e2aebe9f28bc5f8d4ce36448878 | a5c2ad0c3d23d2b1e61ab8e0d7ee64f7e1288547 |
python/cpython | python__cpython-105923 | # C API: Add PyImport_AddModuleRef() function
The C API [PyImport_AddModule()](https://docs.python.org/dev/c-api/import.html#c.PyImport_AddModule) returns a borrowed reference using a special dance added by commit https://github.com/python/cpython/commit/4db8988420e0a122d617df741381b0c385af032c of issue #86160:
```... | 03f1a132eeb34c738812161947ef171b21d58c25 | 7f97c8e367869e2aebe9f28bc5f8d4ce36448878 |
python/cpython | python__cpython-105940 | # SyntaxWarnings in `test_fstring`
# Bug report
Running `python -m test test_fstring` on `main` results in several `SyntaxWarnings` being emitted.
If I just run `python -m test test_fstring` locally, I get this:
```pytb
C:\Users\alexw\coding\cpython>python -m test test_fstring
Running Release|x64 interprete... | 4b431d2e90bf5760a57aa40af2dd78e7bbf0b1ae | 6586cee27f32f0354fe4e77c7b8c6e399329b5e2 |
python/cpython | python__cpython-112871 | # Crash in urllib/request.py proxy_bypass_macosx_sysconf
<!--
Use this template for hard crashes of the interpreter, segmentation faults, failed C-level assertions, and similar.
Do not submit this form if you encounter an exception being unexpectedly raised from a Python function.
Most of the time, these shou... | 22511f77c2818a138a252e6ddae89725d082f8b0 | a723a13bf135306cdc5999a959596bfb487e8f4f |
python/cpython | python__cpython-105909 | # syntactical `__future__` import `barry_as_FLUFL` does not work in the REPL anymore
<!--
If you're new to Python and you're not sure whether what you're experiencing is a bug, the CPython issue tracker is not
the right place to seek help. Consider the following options instead:
- reading the Python tutoria... | 28187a9c4f95affe50fd37e0db0db177e2b9c2e9 | 1858db7cbdbf41aa600c954c15224307bf81a258 |
python/cpython | python__cpython-114035 | # Documentation for pathlib.Path.symlink_to is incorrectly constrained
# Documentation
The docs for `symlink_to` state:
> _target_is_directory_ must be true (default `False`) if the link’s target is a directory.
That's not precisely correct. I experimented and `symlink_to` seems to follow the same behavior a... | b822b85ac11e73bbe4417bf03ee770ab116bb42d | 32c227470aa6f72950b76206ffc529c258b4b8fa |
python/cpython | python__cpython-105885 | # Allow eval and exec to take keyword arguments
# Feature or enhancement
Allow for `globals` and `locals` to be passed in as keyword arguments to `exec` and `eval`.
# Pitch
`exec` and `eval` both can take `globals` and `locals` as positional arguments.
The built-in functions documentation shows the argument... | 2770d5caca42d48102f8e18210132a964c34af7c | 72867c962cc59c6d56805f86530696bea6beb039 |
python/cpython | python__cpython-105876 | # Bump required SQLite version to 3.15.2
(Also [posted on Discord](https://discuss.python.org/t/bump-sqlite-minimum-version-requirement/27999?u=erlendaasland).)
In January 2021 (mid 3.10 development), Sergey Fodoseev and I raised the compile time (and runtime) SQLite version requirements for the sqlite3 standard li... | 6849acb3feacda63ee43f1dc9be28fac1075ca7d | bc07c8f096791d678ca5c1e3486cb9648f7a027b |
python/cpython | python__cpython-105874 | # `_xxsubinterpreters`: the name of shared exception type is `str(exc_type)`, not `exc_type.__name__`
Example:
```python
>> import _xxsubinterpreters
>> i = _xxsubinterpreters.create()
>> _xxsubinterpreters.run_string(i, "(")
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
_xxsubinterpr... | 69a39bd9ad52241ca0e9a1926b4536c73017d067 | 2ef1dc37f02b08536b677dd23ec51541a60effd7 |
python/cpython | python__cpython-105870 | # dataclasses: implicitly defined __dict__ and __weakref__ slots of inherited classes are ignored.
# Bug report
The __weakref__ slot is redefined when inherited from a class which didn't specify slots at all (and thus has a __dict__ and __weakref__ slots).
```
from dataclasses import dataclass
class A:
... | a22d05f04c074dbb4f71e7837f54c0bb693db75d | 1d82a41235ac5619d36ac7e289fcbb686c1d9350 |
python/cpython | python__cpython-105880 | # Improve the constructors of AST nodes
Currently, the constructors for AST nodes accept arbitrary keyword arguments and don't enforce any value:
```pycon
>>> node=ast.FunctionDef(what="is this")
>>> node.what
'is this'
>>> node.name
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
At... | d53560deb2c9ae12147201003fe63b266654ee21 | e72576c48b8be1e4f22c2f387f9769efa073c5be |
python/cpython | python__cpython-107986 | # asyncio subprocess stdin/out/err can be filehandles, but this is undocumented
# Documentation
I was trying to write some asyncio subprocess code to pipe from one process to another, struggled until I found examples with os.pipe(). But going back to the docs I was confused because passing a filehandle isn't mentio... | 13966da71b693b1fae1a8ef66e34e2f0a90ec6c0 | 39de79b345f925ce3bbb79b33534872fe0c90877 |
python/cpython | python__cpython-105851 | # Clarify definition of "minor" vs "major" release
# Documentation
I noticed this in a discussion but it was off-topic there. Either I'm misreading things, but it seems like the definition of major and minor release is inconsistent in the documentation. See [this page](https://discuss.python.org/t/a-fast-free-threa... | 0bffe1acd78069ea21f6b1347bec9cc9747342cb | 0d0963737a0f4b7cadedfae7e8fd33ed18269289 |
python/cpython | python__cpython-105847 | # Assertion failure when specializing functions with too many `__defaults__`
When specializing some Python calls, we assert that we don't have more defaults than we have arguments. This isn't always true, though:
```py
>>> def f():
... pass
...
>>> f.__defaults__ = (None,)
>>> for _ in range(2):
... ... | 2beab5bdef5fa2a00a59371e6137f769586b7404 | b356a4749acb3e6f8c50e8abeb7b2d2b267738d7 |
python/cpython | python__cpython-105835 | # Untested code in `typing.py`
# Bug report
If you apply this diff to `typing.py`, all tests continue to pass:
```diff
diff --git a/Lib/typing.py b/Lib/typing.py
index 1dd9398344..98e19644a2 100644
--- a/Lib/typing.py
+++ b/Lib/typing.py
@@ -1931,6 +1931,7 @@ def _proto_hook(other):
if... | 70c075c194d3739ae10ce76265f05fa82ed46487 | 101d5ec7d7fe122fa81a377c8ab8b562d1add9ee |
python/cpython | python__cpython-105833 | # F-string debug mode does not print correctly when in last line of file
Where there's an f-string that has a debug expression in the last line of a file, the debug expression buffer is one character too short. For example:
```python3
```python3
cpython on main via C v14.0.3-clang via 🐍 pyenv 3.11.3 took 15s
... | 3af2dc7588614c65e9d1178ad9b4a11a19c14dde | d382ad49157b3802fc5619f68d96810def517869 |
python/cpython | python__cpython-108513 | # `concurrent.futures.ProcessPoolExecutor` pool deadlocks when submitting many tasks
# Bug report
Submitting many tasks to a `concurrent.futures.ProcessPoolExecutor` pool
deadlocks with all three start methods.
When running the same example with `multiprocessing.pool.Pool` we have NOT been
able to cause a dead... | 405b06375a8a4cdb08ff53afade09a8b66ec23d5 | e94a2232eac07eb526ec93ef01699513cf9b0fa3 |
python/cpython | python__cpython-105822 | # `test___all__.AllTest.test_all` is failing on every PR
# Bug report
Following https://github.com/python/cpython/pull/105801, `test___all__.AllTest.test_all` is failing on every non-docs PR on the "Hypothesis tests on Ubuntu" CI job.
The traceback is:
```pytb
================================================... | 09ce8c3b48f940eb8865330f029b8069854c3106 | c5111aec2bac464b6643e21fe0844c9b8c4c661a |
python/cpython | python__cpython-105828 | # `patchcheck.py` raises `SystemError` on `main`
# Bug report
Following #105801, running `python Tools/patchcheck/patchcheck.py` now results in `SystemError` being raised.
# To reproduce
1. Make any change to a `.py` file, e.g.:
```diff
--- a/Lib/test/test_fstring.py
+++ b/Lib/test/test_fstring.py
@@ -1... | 3af2dc7588614c65e9d1178ad9b4a11a19c14dde | d382ad49157b3802fc5619f68d96810def517869 |
python/cpython | python__cpython-105910 | # GzipFile.flush doesn't flush compressor in 3.12 beta
<!--
If you're new to Python and you're not sure whether what you're experiencing is a bug, the CPython issue tracker is not
the right place to seek help. Consider the following options instead:
- reading the Python tutorial: https://docs.python.org/3/t... | 1858db7cbdbf41aa600c954c15224307bf81a258 | 7a56a4148c521969d64164d2776641f19e3ca9e8 |
python/cpython | python__cpython-105801 | # f-strings do not show warnings about invalid escapes
In 3.12 and 3.13, f-strings are not warning about invalid escapes that get warnings in real strings. This seems like a bug.
```
>>> f'\?'
'\\?'
>>> len(f'\?')
2
>>> '\?'
<stdin>:1: SyntaxWarning: invalid escape sequence '\?'
'\\?'
_Originally posted by... | 12b6d844d8819955508bd86db106f17516be3f77 | 698a0da7d440856a90b45964e9082b5a55387b80 |
python/cpython | python__cpython-105794 | # Support for not following symlinks in pathlib.Path.is_dir()
# Feature or enhancement
Add a *follow_symlinks* argument to [`pathlib.Path.is_dir()`](https://docs.python.org/3/library/pathlib.html#pathlib.Path.is_dir), defaulting to `True`
# Pitch
Pathlib's `walk()` and `glob()` implementations are built upon ... | 219effa876785408a87bd6acb37c07ee0d25f3f9 | 5d4dbf0e309255e5bce9e31d805a8f950ebf9161 |
python/cpython | python__cpython-108343 | # Python 3.11.4: Running "make test" fails under gcc 4.8.5 because both -std=gnu11 and -std=c++11 are specified when compiling _testcpp11ext
In Python 3.11.4, running "make test" fails under gcc 4.8.5 because both -std=gnu11 and -std=c++11 are specified when compiling _testcpp11ext.
Run the following:
```shell
... | 9173b2bbe13aeccc075b571da05c653a2a91de1b | 3a1ac87f8f89d3206b46a0df4908afae629d669d |
python/cpython | python__cpython-106059 | # Remove `LOAD_CLOSURE`
`LOAD_CLOSURE` is identical to `LOAD_FAST_CHECK` in every way except its name and number.
The justification for its existence is that "We keep LOAD_CLOSURE so that the bytecode stays more readable.".
Which is insufficient justification to keep it given that it:
* Uses an instruction, a lim... | 8bff940ad69ce176dcd2b8e91d0b30ddd09945f1 | 3c70d467c148875f2ce17bacab8909ecc3e9fc1d |
python/cpython | python__cpython-106093 | # normalize() method of Decimal class does not always preserve value
I have encountered unexpected behavior while using the ```normalize()``` method of the ```Decimal``` class. I performed the following test using Python 3.10.10 and 3.11.4.
```python
from decimal import Decimal
v1 = Decimal("0.999999999999999999... | a8210b6df1ed2793c484b3380182ba56c4254a4e | 0345b0c2bbf251a0f475cf53e0fb04c79a220e52 |
python/cpython | python__cpython-107519 | # "mem" and "object" Allocators are No Longer Protected by the GIL
Once we moved to per-interpreter GIL, the promises in [the docs](https://docs.python.org/3.12/c-api/memory.html#allocator-domains) no longer hold:
```
...where the allocation must be performed with the GIL held.
```
It's still fine for pymalloc... | db361a340af3970c279908c8746a6b9ed45f47b8 | fb344e99aa0da5bef9318684ade69978585fe060 |
python/cpython | python__cpython-105758 | # Rework test_ctypes
The ctypes module was first maintained outside Python, then moved into Python stdlib. Its test suite wasn't cleaned recently. I create this issue to track work on this topic.
<!-- gh-linked-prs -->
### Linked PRs
* gh-105758
* gh-105762
* gh-105768
* gh-105797
* gh-105798
* gh-105803
* gh-105814
... | ac7b551bde7a362bc77d2cb84accf755ac30eb09 | b87d2882754a7c273e2695c33384383eba380d7d |
python/cpython | python__cpython-105746 | # `webbrowsers.Konqueror` is broken
`webbrowsers.Konqueror` is broken after https://github.com/python/cpython/pull/102872#issuecomment-1589958849
It looks like `def open` was not removed from `Grail` and was moved to `Konqueror`.
Originially discovered by @vstinner in https://github.com/python/cpython/pull/10287... | e5d45b7444733861153d6e8959c34323fd361322 | 67f69dba0a2adc68c631bad5d970bdd22fc05d91 |
python/cpython | python__cpython-105734 | # ctypes: Deprecate SetPointerType() and ARRAY() functions
The ctypes module has two undocumented functions: SetPointerType() and ARRAY(). These functions are marked as ``XXX Deprecated``, but only in comments.
I propose to deprecate ``ctypes.SetPointerType()`` and ``ctypes.ARRAY()`` by emitting a DeprecationWarnin... | 2211454fe210637ed7fabda12690dac6cc9a8149 | b97e14a806477af4225777d215ac38c0d9b845f0 |
python/cpython | python__cpython-106035 | # More callables should be usable as Exception Group predicates
# Bug report
If a bound method is used as the condition for BaseExceptionGroup.split the following exception is raised:
> TypeError: expected a function, exception type or tuple of exception types
Consider the following example:
```python
class H... | d8ca5a11bc55e2a69cab4f8795d0a5aa6932a41b | 1d33d5378058671bfabb6f4d4b5bfd4726973ff9 |
python/cpython | python__cpython-106771 | # `contextlib.ContextManager` doesn't contain `__slots__ = ()`
Python's ABCs should have `__slots__` attributes to not interfere with slotting in derived classes (especially when the class has no members). However, it is not the case for `contextlib.ContextManager` (as for other `contextlib` ABCs):
```python
from ... | 55408f86d78259f18c56c5e1ea51e0f8dcdbeb67 | cc25ca16ee406db936dfbd2337cbd14b12ccc4b7 |
python/cpython | python__cpython-105728 | # 3.12+: segfault tokenizing multiline fstring placeholder
# Bug report
```python
f'{
hello
}:{world}'
```
output:
```console
$ python3.12 -m tokenize t.py
free(): invalid pointer
Aborted (core dumped)
```
here's a backtrace:
```
(gdb) bt
#0 __pthread_kill_implementation (no_tid=0, s... | abfbab6415fb029e7dca19ecc8d29a13da37bf71 | d0f1afd9425e28409fbf535bb7d43472bfcffcef |
python/cpython | python__cpython-109921 | # _xxsubinterpreters.run_string() doesn't allow subthreads to be running still
Currently `_xxsubinterpreters.run_string()` fails if the interpreter has other threads running, e.g. subthreads created by a previous `run_string()` call. This should be fixed to allow other threads to be running.
<!-- gh-linked-prs -->
#... | 5a76d1be8ef371b75ca65166726923c249b5f615 | bbee57fa8c318cb26d6c8651254927a1972c9738 |
python/cpython | python__cpython-105723 | # 3.12+: tokenize of mixed tabs and spaces now produces an error
I read the parts in [here](https://docs.python.org/3.12/whatsnew/3.12.html#changes-in-the-python-api) and it does mention that some `ERRORTOKEN`s will be converted to exceptions but it doesn't seem to cover this case.
this breaks the `E101` check in `... | ed8217b493e19cea0f3f539e55b592c09ceb9323 | c3d2d64b4c53203719466b32df60ea76f7312891 |
python/cpython | python__cpython-105740 | # Crash During Subinterpreter Finalization
There's an isolation leak somewhere. It may be just in the _xxsubinterpreters module, but I suspect it's not.
See https://github.com/python/cpython/pull/99114#issuecomment-1520952650.
Reproducers:
* https://gist.github.com/tonybaloney/262986212e1061b97908657a53a605d... | b87d2882754a7c273e2695c33384383eba380d7d | fcf0647cf2a78de2c2b603af2709d58c8567df67 |
python/cpython | python__cpython-105688 | # Remove deprecated `re.template`, `re.T`, `re.TEMPLATE`
# Feature or enhancement
It is time to remove deprecated stuff from `re` module.
The 3.11 says:
> The :func:`re.template` function and the corresponding :const:`re.TEMPLATE`
and :const:`re.T` flags are deprecated, as they were undocumented and
lack... | 67f69dba0a2adc68c631bad5d970bdd22fc05d91 | fb655e0c4581ca4bed80db0a083884b29fe142d2 |
python/cpython | python__cpython-105685 | # Get rid of _set_task_name in asyncio
It is already deprecated to have a task which doesn't not supports the name, its time to get rid of this old workaround.
<!-- gh-linked-prs -->
### Linked PRs
* gh-105685
<!-- /gh-linked-prs -->
| 840d02f3f0cd341207db6d918ce7f0987be9952e | 829ac13b69a2b53153e1b40670e6ef82f05130c1 |
python/cpython | python__cpython-105680 | # Remove irregular stack effects
As we move towards generating more components from the instruction definition file (bytecodes.c), it helps if the instructions have regular formats and stack effects.
Currently the stack effects can be simple, variable, conditional or complex.
* Simple: Does not depend on the operand... | 09ffa69e2e84950751739ab500f820725e00a3dd | 217589d4f3246d67c6ef0eb0be2b1c33987cf260 |
python/cpython | python__cpython-105674 | # New warning: `‘value’ may be used uninitialized in this function [-Wmaybe-uninitialized]`
<img width="981" alt="Снимок экрана 2023-06-12 в 09 10 49" src="https://github.com/python/cpython/assets/4660275/84a2242a-d8d3-4777-9886-ae64c67d79f2">
I am working on a fix already.
<!-- gh-linked-prs -->
### Linked PRs
* ... | a8d69fe92c65d636fc454cfb1825c357eb2e6325 | 58f5227d7cdff803609a0bda6882997b3a5ec4bf |
python/cpython | python__cpython-109384 | # 3.12 tracing regression: a conditional in a finally block will revisit the condition before exiting the block.
Python 3.12 introduced a change in tracing behavior. Now a conditional in a finally block will revisit the condition before exiting the block.
Here is test.py:
```python
import sys
print(sys.version)... | 4a54074a0f5579d417445ec28427cd0ed5aa01f4 | 1ce9ea0453f7dc69dd41684f3bc9310259513de8 |
python/cpython | python__cpython-105628 | # Change the default return value of `HTTPConnection.get_proxy_response_headers` from `{}` to `None`
See my reasoning here: https://github.com/python/cpython/pull/104248#issuecomment-1585579473
<!-- gh-linked-prs -->
### Linked PRs
* gh-105628
* gh-106738
<!-- /gh-linked-prs -->
| 490295d651d04ec3b3eff2a2cda7501191bad78a | 025995feadaeebeef5d808f2564f0fd65b704ea5 |
python/cpython | python__cpython-105887 | # Poor performance on logging.RotatingFileHandler due to fix to #89564
The fix to #89564 produced a significant performance degradation when logs are on an NFS filesystem. The fix for #89564 was for a bug found with the TimedRotatingFileHandler but an fix was added to the sized base file rotator too.
Here is the... | e9b4ec614b66d11623b80471409c16a109f888d5 | 0890ad7c024ccf29614849b6ffadcb92c0e91ce7 |
python/cpython | python__cpython-105620 | # Rename "own_gil" Field of PyInterpreterConfig to "gil"
We've added the `PyInterpreterConfig` struct in 3.12. For PEP 684 (per-interpreter GIL), likewise 3.12, we added a new int-boolean field, "own_gil", to indicate if `PyInterpreterState.ceval.gil` should point to the interpreter's own GIL.
Having finally just ... | b97e14a806477af4225777d215ac38c0d9b845f0 | abfbab6415fb029e7dca19ecc8d29a13da37bf71 |
python/cpython | python__cpython-105589 | # Some object-to-AST conversions are missing error checks
The generated code in `Python-ast.c` is missing error checks following the construction of C-level `alias`, `arg`, `comprehension`, `keyword`, `match_item`, and `withitem` nodes from their Python object counterparts. This means it's possible to crash the interp... | a4056c8f9c2d9970d39e3cb6bffb255cd4b8a42c | 3af2dc7588614c65e9d1178ad9b4a11a19c14dde |
python/cpython | python__cpython-105638 | # 3.12.0b2: "Assertion failed: immortal object has less refcnt than expected _Py_IMMORTAL_REFCNT"
# Crash report
When I run [my project](https://qutebrowser.org/) with Python 3.12, sometimes on exit I get the error below.
~~Given the circumstances, it seems almost impossible to construct a minimal example with t... | 6199fe3b3236748033a7ce2559aeddb5a91bbbd9 | dab5a3ebe8063e7ad39c9ba04421d8fe11927d43 |
python/cpython | python__cpython-105780 | # Deprecate unusual ways of creating empty TypedDicts
# Feature or enhancement
I propose that we deprecate the following two ways of creating empty `TypedDict`s:
```py
from typing import TypedDict
T = TypedDict("T")
T2 = TypedDict("T2", None)
```
# Pitch
Users currently have four distinct options if ... | 7b1f0f204a785485de1daf9d26828a81953537e4 | d32e8d6070057eb7ad0eb2f9d9f1efab38b2cff4 |
python/cpython | python__cpython-105609 | # Deprecate unusual ways of creating `typing.NamedTuple` classes
# Feature or enhancement
I propose that we deprecate in Python 3.13 the following unusual ways of constructing a `typing.NamedTuple`:
```py
from typing import NamedTuple
Foo = NamedTuple("Foo", x=int, y=int) # NamedTuple with "x" and "y" field... | ad56340b665c5d8ac1f318964f71697bba41acb7 | fc8037d84c5f886849a05ec993dd0f79a356d372 |
python/cpython | python__cpython-105565 | # Artificial newlines show in the "line" attribute of tokens in the tokenizer module
Example:
```
from tokenize import generate_tokens
from io import StringIO
import pprint
pprint.pprint(list(generate_tokens(StringIO('a').readline)))
```
prints:
```
[TokenInfo(type=1 (NAME), string='a', start=(1, 0), en... | d7f46bcd989580340675bf0a9fdbfa1505a37e81 | 1dd267af642ed6df05a1c106e9dafb8252d826e6 |
python/cpython | python__cpython-105946 | # Recommend DateType in the datetime documentation
# Documentation
At the PyCon US 2023 sprints, I spent some time with @glyph, @hauntsaninja, @AlexWaygood and others trying to put together a plan to make the type stubs for `datetime` distinguish between naïve and aware datetimes (even if the runtime types don't re... | 8f9ea43ee805f98391f857397daac9df7ffa71cd | 6f97eeec222f81bd7ae836c149872a40b079e2a6 |
python/cpython | python__cpython-105558 | # test_sqlite3.test_userfunctions redefines test_func_return_too_large_int() method
Lib/test/test_sqlite3/test_userfunctions.py defines to methods called test_func_return_too_large_int() in the same class, so the first one is never called.
cc @erlend-aasland
<!-- gh-linked-prs -->
### Linked PRs
* gh-105558
* gh-... | b8fa7bda4f286503447dc12327b789bbfc836458 | 9bf8d825a66ea2a76169b917c12c237a6af2ed75 |
python/cpython | python__cpython-105555 | # `generate_tokens` starts to give `SyntaxError`
<!--
If you're new to Python and you're not sure whether what you're experiencing is a bug, the CPython issue tracker is not
the right place to seek help. Consider the following options instead:
- reading the Python tutorial: https://docs.python.org/3/tutoria... | b047fa5e56ba725362c64ca3d6fccbdcf51d0cab | 00b599ab5a76023fa0083d7cc5d3c569342a5191 |
python/cpython | python__cpython-105546 | # Remove deprecated `MacOSXOSAScript._name` attribute
# Feature or enhancement
It was deprecated in 3.11: https://github.com/python/cpython/blob/3.11/Lib/webbrowser.py#L676-L688
It is time to remove it.
My PR from 2 years ago: https://github.com/python/cpython/pull/30241
CC @corona10 as the original PR reviewe... | 947ec7ab02e7673956eb7f235c330a49f11e157a | 3e525d22128cf040b3fd164f52cc6ae20ca58455 |
python/cpython | python__cpython-105707 | # cases_generator tests aren't run by CI, and are broken
There are some tests for the cases generator in Tools/cases_generator/test_generator.py, but several things are wrong with these:
- [x] They are currently broken
- [x] They depend on pytest
- [x] They aren't run in CI
- [ ] They are incomplete
Let's fix... | 8da9d1b16319f4a6bd78435016ef1f4bef6e2b41 | ca3cc4b95d66f7527ebe0ba4cdb1907082d9bfc8 |
python/cpython | python__cpython-108015 | # Emit resource warning if sqlite3 fails to close the database
_See https://github.com/python/cpython/issues/103837#issuecomment-1541808307 and https://github.com/python/cpython/issues/103837#issuecomment-1543936079_:
EAA:
> I still think we should consider changing the underlying API so we can emit a resource war... | 1a1bfc28912a39b500c578e9f10a8a222638d411 | 86617518c4ac824e2b6dc20691ba5a08df04f285 |
python/cpython | python__cpython-107119 | # Document that enums with unhashable members are created non-performantly
# Documentation
#28907 fixes enum creation taking quadratic time relative to the number of members, but _only_ for members whose values are hashable. If not fixed, ideally this would be documented somewhere as it could potentially be a big p... | 601f3a7b3391e9d219a8ec44a6c56d00ce584d2a | 735fc2cbbcf875c359021b5b2af7f4c29f4cf66d |
python/cpython | python__cpython-105523 | # Modernize `test_enum` by removing unused exception handling
Right now there are places in `test_enum` that for some reason contain code like https://github.com/python/cpython/blob/4ff5690e591b7d11cf11e34bf61004e2ea58ab3c/Lib/test/test_enum.py#L53-L76
For some reason we try to catch exceptions that won't ever happ... | 199438b7cc2ef669b8d005d38797477a18b610cb | fce93c80ae2d792b8ca443b044e28abbf28bb89a |
python/cpython | python__cpython-105510 | # Simplify the implementation of `typing.Annotated`
# Feature or enhancement
`typing.Annotated` is currently implemented as a class, but doesn't need to be. All other objects like it in the `typing` module are implemented as instances of `typing._SpecialForm`, and we can do the same here. This simplifies the code a... | a1cbace91b624681dd7bb8eb82ba187bda55d785 | 8f9ea43ee805f98391f857397daac9df7ffa71cd |
python/cpython | python__cpython-105511 | # Merge `typing.Union` and `types.UnionType`
Currently, unions created through `typing.Union[A, B]` and through the PEP-604 syntax `A | B` are at runtime instances of completely different types, and they differ in exactly what elements they accept. This is confusing and makes it harder for users to detect unions at ru... | 0f511d8b44dd9993474402411af8c83f4964bc95 | 9127b4602e4e0e110eab7f2f6a8ac54fe66b0a72 |
python/cpython | python__cpython-105542 | # 3.11.4: ValueError when inverting enum.Flag member with mask member
With code such as:
```python
import enum
class Flag(enum.Flag):
A = 0x01
B = 0x02
Mask = 0xff
print(~Flag.A)
```
Python 3.10.11 prints `Flag.B`, and so does Python 3.11.3. However, with Python 3.11.4, this happens instea... | 59f009e5898a006cdc8f5249be589de6edfe5cd0 | 8e755923c97d689ba7c7fe8deb50c1b169263264 |
python/cpython | python__cpython-105488 | # `types.GenericAlias` has extra `__copy__` and `__deepcopy__` in `__dir__`
Repro:
```python
>>> type A[X] = list[X]
>>> dir(A[int])
[..., '__copy__', '__deepcopy__', ...]
```
We can access all other attributes, but not `__copy__` and `__deepcopy__`:
```pytb
>>> A[int].__copy__
Traceback (most recent ca... | eb7d6e7ad844955f9af88707d296e003c7ce4394 | e212618bafaa4f775502e3442de0affb80205b5e |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.