]> granicus.if.org Git - python/log
python
5 years agoUpdate concurrent.futures.rst (GH-14061)
Géry Ogam [Fri, 14 Jun 2019 14:39:43 +0000 (16:39 +0200)]
Update concurrent.futures.rst (GH-14061)

This PR adds missing details in the [`concurrent.futures`](https://docs.python.org/3/library/concurrent.futures.html) documentation:

* the mention that `Future.cancel` also returns `False` if the call finished running;
* the mention of the states for `Future` that did not complete: pending or running.

5 years agobpo-37278: Fix test_asyncio ProactorLoopCtrlC (GH-14074)
Victor Stinner [Fri, 14 Jun 2019 11:02:51 +0000 (13:02 +0200)]
bpo-37278: Fix test_asyncio ProactorLoopCtrlC (GH-14074)

Join the thread to prevent leaking a running thread and leaking a
reference.

Cleanup also the test:

* asyncioWindowsProactorEventLoopPolicy became the default policy,
  there is no need to set it manually.
* Only start the thread once the loop is running.
* Use a shorter sleep in the thread (100 ms rather than 1 sec).
* Use close_loop(loop) rather than loop.close().
* Use longer variable names.

5 years agobpo-37249: add declaration of _PyObject_GetMethod (GH-14015)
Jeroen Demeyer [Fri, 14 Jun 2019 10:37:15 +0000 (12:37 +0200)]
bpo-37249: add declaration of _PyObject_GetMethod (GH-14015)

5 years agobpo-37269: Correctly optimise conditionals with constant booleans (GH-14071)
Pablo Galindo [Fri, 14 Jun 2019 05:54:53 +0000 (06:54 +0100)]
bpo-37269: Correctly optimise conditionals with constant booleans (GH-14071)

Fix a regression introduced by af8646c8054d0f4180a2013383039b6a472f9698 that was causing code of the form:

if True and False:
   do_something()

to be optimized incorrectly, eliminating the block.

5 years agobpo-37077: Add native thread ID (TID) for AIX (GH-13624)
Michael Felt [Thu, 13 Jun 2019 22:34:46 +0000 (00:34 +0200)]
bpo-37077: Add native thread ID (TID) for AIX (GH-13624)

This is the followup  for issue36084

https://bugs.python.org/issue37077

5 years agobpo-36710: Pass explicitly tstate in sysmodule.c (GH-14060)
Victor Stinner [Thu, 13 Jun 2019 20:41:23 +0000 (22:41 +0200)]
bpo-36710: Pass explicitly tstate in sysmodule.c (GH-14060)

* Replace global var Py_VerboseFlag with interp->config.verbose.
* Add _PyErr_NoMemory(tstate) function.
* Add tstate parameter to _PyEval_SetCoroutineOriginTrackingDepth()
  and move the function to the internal API.
* Replace _PySys_InitMain(runtime, interp)
  with _PySys_InitMain(runtime, tstate).

5 years agobpo-37213: Handle negative line deltas correctly in the peephole optimizer (GH-13969)
Pablo Galindo [Thu, 13 Jun 2019 18:16:22 +0000 (19:16 +0100)]
bpo-37213: Handle negative line deltas correctly in the peephole optimizer (GH-13969)

The peephole optimizer was not optimizing correctly bytecode after negative deltas were introduced. This is due to the fact that some special values (255) were being searched for in both instruction pointer delta and line number deltas.

5 years agobpo-6689: os.path.commonpath raises ValueError for different drives isn't documented...
Makdon [Thu, 13 Jun 2019 13:59:49 +0000 (21:59 +0800)]
bpo-6689: os.path.commonpath raises ValueError for different drives isn't documented (GH-14045)

It would raise ValueError("Paths don't have the same drive") if the paths on different drivers, which is not documented.

os.path.commonpath raises ValueError when the *paths* are in different drivers, but it is not documented.
Update the document according @Windsooon 's suggestion.
It actually raise ValueError according line 355 of [test of path](https://github.com/python/cpython/blob/master/Lib/test/test_ntpath.py)

https://bugs.python.org/issue6689

5 years agobpo-37261: Fix support.catch_unraisable_exception() (GH-14052)
Victor Stinner [Thu, 13 Jun 2019 12:44:54 +0000 (14:44 +0200)]
bpo-37261: Fix support.catch_unraisable_exception() (GH-14052)

The __exit__() method of test.support.catch_unraisable_exception
context manager now ignores unraisable exception raised when clearing
self.unraisable attribute.

5 years agobpo-37210: Fix pure Python pickle when _pickle is unavailable (GH-14016)
Victor Stinner [Thu, 13 Jun 2019 11:58:51 +0000 (13:58 +0200)]
bpo-37210: Fix pure Python pickle when _pickle is unavailable (GH-14016)

Allow pure Python implementation of pickle to work
even when the C _pickle module is unavailable.

Fix test_pickle when _pickle is missing: declare PyPicklerHookTests
outside "if has_c_implementation:" block.

5 years agotbpo-36402: Fix threading.Thread._stop() (GH-14047)
Victor Stinner [Thu, 13 Jun 2019 10:06:24 +0000 (12:06 +0200)]
tbpo-36402: Fix threading.Thread._stop() (GH-14047)

Remove the _tstate_lock from _shutdown_locks, don't remove None.

5 years agobpo-37231: optimize calls of special methods (GH-13973)
Jeroen Demeyer [Thu, 13 Jun 2019 09:26:44 +0000 (11:26 +0200)]
bpo-37231: optimize calls of special methods (GH-13973)

5 years agobpo-37253: Remove PyAST_obj2mod_ex() function (GH-14020)
Victor Stinner [Thu, 13 Jun 2019 07:18:45 +0000 (09:18 +0200)]
bpo-37253: Remove PyAST_obj2mod_ex() function (GH-14020)

PyAST_obj2mod_ex() is similar to PyAST_obj2mod() with an additional
'feature_version' parameter which is unused.

5 years agobpo-35070: test_getgrouplist may fail on macOS if too many groups (GH-13071)
Jeffrey Kintscher [Thu, 13 Jun 2019 07:01:29 +0000 (00:01 -0700)]
bpo-35070: test_getgrouplist may fail on macOS if too many groups (GH-13071)

5 years agobpo-37216: update version to 3.9 in mac using document (GH-13966)
Makdon [Thu, 13 Jun 2019 05:04:13 +0000 (13:04 +0800)]
bpo-37216: update version to 3.9 in mac using document (GH-13966)

5 years agobpo-37257: obmalloc: stop simple arena thrashing (#14039)
Tim Peters [Thu, 13 Jun 2019 03:41:03 +0000 (22:41 -0500)]
bpo-37257:  obmalloc:  stop simple arena thrashing (#14039)

GH-14039:  allow (no more than) one wholly empty arena on the usable_arenas list.

This prevents thrashing in some easily-provoked simple cases that could end up creating and destroying an arena on each loop iteration in client code.   Intuitively, if the only arena on the list becomes empty, it makes scant sense to give it back to the system unless we know we'll never need another free pool again before another arena frees a pool.  If the latter obtains, then - yes - this will "waste" an arena.

5 years agoAdd 3.9 whatsnew file (GH-14040)
Ned Deily [Thu, 13 Jun 2019 03:31:45 +0000 (23:31 -0400)]
Add 3.9 whatsnew file (GH-14040)

5 years agobpo-37253: Fix typo in PyCompilerFlags doc (GH-14036)
Victor Stinner [Thu, 13 Jun 2019 00:17:14 +0000 (02:17 +0200)]
bpo-37253: Fix typo in PyCompilerFlags doc (GH-14036)

Remove ";" to fix Sphinx formatting.

5 years agobpo-37253: Add _PyCompilerFlags_INIT macro (GH-14018)
Victor Stinner [Thu, 13 Jun 2019 00:16:41 +0000 (02:16 +0200)]
bpo-37253: Add _PyCompilerFlags_INIT macro (GH-14018)

Add a new _PyCompilerFlags_INIT macro to initialize PyCompilerFlags
variables, rather than initializing cf_flags and cf_feature_version
explicitly in each variable.

5 years agobpo-37253: Document PyCompilerFlags.cf_feature_version (GH-14019)
Victor Stinner [Thu, 13 Jun 2019 00:01:29 +0000 (02:01 +0200)]
bpo-37253: Document PyCompilerFlags.cf_feature_version (GH-14019)

* Update PyCompilerFlags structure documentation.
* Document the new cf_feature_version field in the Changes in the C
  API section of the What's New in Python 3.8 doc.

5 years agobpo-36402: Fix threading._shutdown() race condition (GH-13948)
Victor Stinner [Wed, 12 Jun 2019 23:30:17 +0000 (01:30 +0200)]
bpo-36402: Fix threading._shutdown() race condition (GH-13948)

Fix a race condition at Python shutdown when waiting for threads.
Wait until the Python thread state of all non-daemon threads get
deleted (join all non-daemon threads), rather than just wait until
Python threads complete.

* Add threading._shutdown_locks: set of Thread._tstate_lock locks
  of non-daemon threads used by _shutdown() to wait until all Python
  thread states get deleted. See Thread._set_tstate_lock().
* Add also threading._shutdown_locks_lock to protect access to
  threading._shutdown_locks.
* Add test_finalization_shutdown() test.

5 years agobpo-36779: time.tzname returns empty string on Windows if default cod… (GH-13073)
Paul Monson [Wed, 12 Jun 2019 23:13:27 +0000 (16:13 -0700)]
bpo-36779: time.tzname returns empty string on Windows if default cod… (GH-13073)

Calling setlocale(LC_CTYPE, "") on a system where GetACP() returns CP_UTF8 results in empty strings in _tzname[].

This causes time.tzname to be an empty string.
I have reported the bug to the UCRT team and will follow up, but it will take some time get a fix into production.

In the meantime one possible workaround is to temporarily change the locale by calling setlocale(LC_CTYPE, "C") before calling _tzset and restore the current locale after if the GetACP() == CP_UTF8 or CP_UTF7

@zooba

https://bugs.python.org/issue36779

5 years agobpo-37069: regrtest uses sys.unraisablehook (GH-13759)
Victor Stinner [Wed, 12 Jun 2019 23:09:04 +0000 (01:09 +0200)]
bpo-37069: regrtest uses sys.unraisablehook (GH-13759)

regrtest now uses sys.unraisablehook() to mark a test as "environment
altered" (ENV_CHANGED) if it emits an "unraisable exception".
Moreover, regrtest logs a warning in this case.

Use "python3 -m test --fail-env-changed" to catch unraisable
exceptions in tests.

5 years agobpo-37223, test_io: silence last 'Exception ignored in:' (GH-14029)
Victor Stinner [Wed, 12 Jun 2019 21:57:11 +0000 (23:57 +0200)]
bpo-37223, test_io: silence last 'Exception ignored in:' (GH-14029)

Use catch_unraisable_exception() to ignore 'Exception ignored in:'
error when the internal BufferedWriter of the BufferedRWPair is
destroyed. The C implementation doesn't give access to the
internal BufferedWriter, so just ignore the warning instead.

5 years agoMake asyncio stream sendfile fail on error (was hang) (GH-14025)
Andrew Svetlov [Wed, 12 Jun 2019 18:50:23 +0000 (21:50 +0300)]
Make asyncio stream sendfile fail on error (was hang) (GH-14025)

5 years agobpo-37236: pragma optimize off for _Py_c_quot on Windows arm64 (GH-13983)
Paul Monson [Wed, 12 Jun 2019 18:08:40 +0000 (11:08 -0700)]
bpo-37236: pragma optimize off for _Py_c_quot on Windows arm64 (GH-13983)

5 years agobpo-37201: fix test_distutils failures for Windows ARM64 (GH-13902)
Paul Monson [Wed, 12 Jun 2019 17:16:49 +0000 (10:16 -0700)]
bpo-37201: fix test_distutils failures for Windows ARM64 (GH-13902)

5 years agobpo-37160: Thread native ID NetBSD support (GH-13835)
David Carlier [Wed, 12 Jun 2019 15:37:56 +0000 (15:37 +0000)]
bpo-37160: Thread native ID NetBSD support (GH-13835)

5 years agobpo-35545: Skip `test_asyncio.test_create_connection_ipv6_scope` on AIX (GH-14011)
Michael Felt [Wed, 12 Jun 2019 12:00:56 +0000 (14:00 +0200)]
bpo-35545: Skip `test_asyncio.test_create_connection_ipv6_scope` on AIX (GH-14011)

because "getaddrinfo()" behaves different on AIX

https://bugs.python.org/issue35545

5 years agobpo-29505: Fuzz json module, enforce size limit on int(x) fuzz (GH-13991)
Ammar Askar [Wed, 12 Jun 2019 04:30:35 +0000 (21:30 -0700)]
bpo-29505: Fuzz json module, enforce size limit on int(x) fuzz (GH-13991)

* bpo-29505: Enable fuzz testing of the json module, enforce size limit on int(x) fuzz and json input size to avoid timeouts.

Contributed by by Ammar Askar for Google.

5 years agobpo-32625: Updated documentation for EXTENDED_ARG. (GH-13985)
Yao Zuo [Wed, 12 Jun 2019 03:46:09 +0000 (20:46 -0700)]
bpo-32625: Updated documentation for EXTENDED_ARG. (GH-13985)

Python 3.6 changed the size of bytecode instruction, while the documentation for `EXTENDED_ARG` was not updated accordingly.

5 years agobpo-26219: Fix compiler warning in _PyCode_InitOpcache() (GH-13997)
Victor Stinner [Wed, 12 Jun 2019 02:41:16 +0000 (04:41 +0200)]
bpo-26219: Fix compiler warning in _PyCode_InitOpcache() (GH-13997)

Fix MSVC warning:

    objects\codeobject.c(285): warning C4244: '=':
    conversion from 'Py_ssize_t' to 'unsigned char',
    possible loss of data

5 years agobpo-36918: Fix "Exception ignored in" in test_urllib (GH-13996)
Victor Stinner [Wed, 12 Jun 2019 02:07:38 +0000 (04:07 +0200)]
bpo-36918: Fix "Exception ignored in" in test_urllib (GH-13996)

Mock the HTTPConnection.close() method in a few unit tests to avoid
logging "Exception ignored in: ..." messages.

5 years agobpo-35766: compile(): rename feature_version parameter (GH-13994)
Victor Stinner [Wed, 12 Jun 2019 00:52:16 +0000 (02:52 +0200)]
bpo-35766: compile(): rename feature_version parameter (GH-13994)

Rename compile() feature_version parameter to _feature_version and
convert it to a keyword-only parameter.

Update also test_type_comments to pass feature_version as a tuple.

5 years agobpo-35766: Change format for feature_version to (major, minor) (GH-13992)
Guido van Rossum [Wed, 12 Jun 2019 00:23:12 +0000 (17:23 -0700)]
bpo-35766: Change format for feature_version to (major, minor) (GH-13992)

(A single int is still allowed, but undocumented.)

https://bugs.python.org/issue35766

5 years agobpo-37238: Enable building for Windows using Visual Studio 2019 (GH-13988)
Paul Monson [Tue, 11 Jun 2019 22:03:17 +0000 (15:03 -0700)]
bpo-37238: Enable building for Windows using Visual Studio 2019 (GH-13988)

5 years agobpo-35766: What's new in the ast and typing modules (#13984)
Guido van Rossum [Tue, 11 Jun 2019 20:42:35 +0000 (13:42 -0700)]
bpo-35766: What's new in the ast and typing modules (#13984)

5 years agoFix test_posix if RWF_HIPRI is defined but not preadv2. (GH-13980)
Benjamin Peterson [Tue, 11 Jun 2019 17:15:31 +0000 (10:15 -0700)]
Fix test_posix if RWF_HIPRI is defined but not preadv2. (GH-13980)

If preadv2 is not available, preadv will raise NotImplementedError.

5 years agocloses bpo-33758: Skip test_get_type_hints_modules_forwardref. (GH-13977)
Benjamin Peterson [Tue, 11 Jun 2019 16:18:31 +0000 (09:18 -0700)]
closes bpo-33758: Skip test_get_type_hints_modules_forwardref. (GH-13977)

This test "works" if things are run in the right order, so it's better to use @skip than @expectedFailure here.

5 years agobpo-36607: Eliminate RuntimeError raised by asyncio.all_tasks() (GH-13971)
Andrew Svetlov [Tue, 11 Jun 2019 15:27:30 +0000 (18:27 +0300)]
bpo-36607: Eliminate RuntimeError raised by asyncio.all_tasks() (GH-13971)

If internal tasks weak set is changed by another thread during iteration.

https://bugs.python.org/issue36607

5 years agobpo-37219: Remove erroneous optimization for differencing an empty set (GH-13965)
Raymond Hettinger [Tue, 11 Jun 2019 08:15:24 +0000 (01:15 -0700)]
bpo-37219: Remove erroneous optimization for differencing an empty set (GH-13965)

5 years agocloses bpo-35184: Fix XML_POOR_ENTROPY option that breaks makesetup parsing of pyexpa...
aaronpaulhurst [Tue, 11 Jun 2019 01:54:24 +0000 (18:54 -0700)]
closes bpo-35184: Fix XML_POOR_ENTROPY option that breaks makesetup parsing of pyexpat line in Setup. (GH-13064)

When the line is uncommented, the equals character causes it to be incorrectly interpreted
as a macro definition by makesetup.  This results in invalid Makefile output.

The expat code only requires XML_POOR_ENTROPY to be defined; the value is unnecessary.

5 years agobpo-37223: test_io: silence destructor errors (GH-13954)
Victor Stinner [Tue, 11 Jun 2019 01:10:59 +0000 (03:10 +0200)]
bpo-37223: test_io: silence destructor errors (GH-13954)

Implement also MockNonBlockWriterIO.seek() method.

5 years agobpo-18748: Fix _pyio.IOBase destructor (closed case) (GH-13952)
Victor Stinner [Tue, 11 Jun 2019 00:49:06 +0000 (02:49 +0200)]
bpo-18748: Fix _pyio.IOBase destructor (closed case) (GH-13952)

_pyio.IOBase destructor now does nothing if getting the closed
attribute fails to better mimick _io.IOBase finalizer.

5 years agobpo-37215: Fix dtrace issue introduce by bpo-36842 (GH-13940)
Christian Heimes [Mon, 10 Jun 2019 15:19:48 +0000 (17:19 +0200)]
bpo-37215: Fix dtrace issue introduce by bpo-36842 (GH-13940)

Signed-off-by: Christian Heimes <christian@python.org>
https://bugs.python.org/issue37215

5 years agoDo not use explicit inheritance from object in the documentation. (GH-13936)
Serhiy Storchaka [Mon, 10 Jun 2019 10:35:52 +0000 (13:35 +0300)]
Do not use explicit inheritance from object in the documentation. (GH-13936)

5 years agobpo-36785: PEP 574 What's New entry (#13931)
Antoine Pitrou [Sun, 9 Jun 2019 12:47:15 +0000 (14:47 +0200)]
bpo-36785: PEP 574 What's New entry (#13931)

5 years agoAdd some placeholder notes for major 3.8 features (GH-13927)
Nick Coghlan [Sun, 9 Jun 2019 09:07:42 +0000 (19:07 +1000)]
Add some placeholder notes for major 3.8 features (GH-13927)

5 years agobpo-11122: fix hardcoded path checking for rpmbuild in bdist_rpm.py (GH-10594)
Marcin Niemira [Sat, 8 Jun 2019 21:05:06 +0000 (07:05 +1000)]
bpo-11122: fix hardcoded path checking for rpmbuild in bdist_rpm.py (GH-10594)

5 years agobpo-37178: Allow a one argument form of math.perm() (GH-13905)
Raymond Hettinger [Sat, 8 Jun 2019 15:58:11 +0000 (08:58 -0700)]
bpo-37178: Allow a one argument form of math.perm() (GH-13905)

5 years agobpo-34886: Fix subprocess.run handling of exclusive arguments (GH-11727)
Rémi Lapeyre [Sat, 8 Jun 2019 14:56:24 +0000 (16:56 +0200)]
bpo-34886: Fix subprocess.run handling of exclusive arguments (GH-11727)

Fix an unintended ValueError from :func:`subprocess.run` when checking for
conflicting `input` and `stdin` or `capture_output` and `stdout` or `stderr` args
when they were explicitly provided but with `None` values within a passed in
`**kwargs` dict rather than as passed directly by name.

5 years agobpo-29505: Fix interpreter in fuzzing targets to be relocatable (GH-13907)
Ammar Askar [Sat, 8 Jun 2019 14:43:16 +0000 (07:43 -0700)]
bpo-29505: Fix interpreter in fuzzing targets to be relocatable (GH-13907)

5 years agobpo-37173: Show passed class in inspect.getfile error (GH-13861)
Philipp A [Sat, 8 Jun 2019 12:05:46 +0000 (14:05 +0200)]
bpo-37173: Show passed class in inspect.getfile error (GH-13861)

Currently, inspect.getfile(str) will report nonsense:

```pytb
>>> inspect.getfile(str)
TypeError: <module 'builtins' (built-in)> is a built-in class
```

This fixes that

https://bugs.python.org/issue37173

5 years agocross port importlib-metadata PR #76 (#13903)
Anthony Sottile [Fri, 7 Jun 2019 21:23:39 +0000 (14:23 -0700)]
cross port importlib-metadata PR #76 (#13903)

https://gitlab.com/python-devs/importlib_metadata/merge_requests/76

5 years agobpo-37150: Throw ValueError if FileType class object was passed in add_argument ...
zygocephalus [Fri, 7 Jun 2019 20:08:36 +0000 (23:08 +0300)]
bpo-37150: Throw ValueError if FileType class object was passed in add_argument (GH-13805)

There is a possibility that someone (like me) accidentally will omit parentheses with `FileType` arguments after `FileType`, and parser will contain wrong file until someone will try to use it.

Example:
```python
parser = argparse.ArgumentParser()
parser.add_argument('-x', type=argparse.FileType)
```

https://bugs.python.org/issue37150

5 years agobpo-37138: fix undefined behaviour with memcpy() on NULL array (GH-13867)
Jeroen Demeyer [Fri, 7 Jun 2019 18:01:53 +0000 (20:01 +0200)]
bpo-37138: fix undefined behaviour with memcpy() on NULL array (GH-13867)

5 years agobpo-37181: Fix test_regrtest failures on Windows arm64 (GH-13872)
Paul Monson [Fri, 7 Jun 2019 17:58:41 +0000 (10:58 -0700)]
bpo-37181: Fix test_regrtest failures on Windows arm64 (GH-13872)

5 years agobpo-37191: Move TestPEP590 from test_capi to test_call (GH-13892)
Victor Stinner [Fri, 7 Jun 2019 15:51:28 +0000 (17:51 +0200)]
bpo-37191: Move TestPEP590 from test_capi to test_call (GH-13892)

5 years agobpo-37169: Rewrite _PyObject_IsFreed() unit tests (GH-13888)
Victor Stinner [Fri, 7 Jun 2019 14:22:21 +0000 (16:22 +0200)]
bpo-37169: Rewrite _PyObject_IsFreed() unit tests (GH-13888)

Replace two Python function calls with a single one to ensure that no
memory allocation is done between the invalid object is created and
when _PyObject_IsFreed() is called.

5 years agobpo-37151: simplify classmethoddescr_call (GH-13340)
Jeroen Demeyer [Fri, 7 Jun 2019 10:20:24 +0000 (12:20 +0200)]
bpo-37151: simplify classmethoddescr_call (GH-13340)

5 years agoUse assertEqual(). (#13883)
Stefan Krah [Fri, 7 Jun 2019 09:18:34 +0000 (11:18 +0200)]
Use assertEqual(). (#13883)

5 years agoIDLE: Standardize naming convention for DummyEditwin in tests (GH-13876)
Cheryl Sabella [Fri, 7 Jun 2019 09:08:20 +0000 (05:08 -0400)]
IDLE: Standardize naming convention for DummyEditwin in tests (GH-13876)

* Change from Dummy_Editwin to DummyEditwin to match other tests.

5 years agobpo-37188: Fix a divide-by-zero in arrays of size-0 objects (#13881)
Eric Wieser [Fri, 7 Jun 2019 08:13:26 +0000 (01:13 -0700)]
bpo-37188: Fix a divide-by-zero in arrays of size-0 objects (#13881)

5 years agobpo-37177: make IDLE's search dialogs transient (GH-13869)
Tal Einat [Fri, 7 Jun 2019 05:54:40 +0000 (08:54 +0300)]
bpo-37177: make IDLE's search dialogs transient (GH-13869)

This avoids the search dialogs being hidden behind the editor window.

5 years agobpo-37134: Add PEP570 notation to the signature of byte{array}.translate (GH-13874)
Pablo Galindo [Thu, 6 Jun 2019 23:38:41 +0000 (00:38 +0100)]
bpo-37134: Add PEP570 notation to the signature of byte{array}.translate (GH-13874)

5 years agobpo-37170: Fix the cast on error in PyLong_AsUnsignedLongLongMask() (GH-13860)
Zackery Spytz [Thu, 6 Jun 2019 20:39:23 +0000 (14:39 -0600)]
bpo-37170: Fix the cast on error in PyLong_AsUnsignedLongLongMask() (GH-13860)

5 years agobpo-36520: Email header folded incorrectly (#13608)
websurfer5 [Thu, 6 Jun 2019 19:53:27 +0000 (12:53 -0700)]
bpo-36520: Email header folded incorrectly (#13608)

* bpo-36520: reset the encoded word offset when starting a new
line during an email header folding operation

* 📜🤖 Added by blurb_it.

* bpo-36520: add an additional test case, and provide descriptive
comments for the test_folding_with_utf8_encoding_* tests

* bpo-36520: fix whitespace issue

* bpo-36520: changes per reviewer request -- remove extraneous
backslashes; add whitespace between terminating quotes and
line-continuation backslashes; use "bpo-" instead of
"issue #" in comments

5 years agobpo-37156: Fix libssl DLL tag in MSI sources (GH-13866)
Steve Dower [Thu, 6 Jun 2019 16:06:51 +0000 (09:06 -0700)]
bpo-37156: Fix libssl DLL tag in MSI sources (GH-13866)

5 years agobpo-36763, _testembed: enable assert() in release mode (GH-13857)
Victor Stinner [Thu, 6 Jun 2019 12:25:18 +0000 (14:25 +0200)]
bpo-36763, _testembed: enable assert() in release mode (GH-13857)

5 years agobpo-2661: Make mapping tests better usable for custom mapping classes. (GH-11157)
Walter Dörwald [Thu, 6 Jun 2019 10:13:08 +0000 (12:13 +0200)]
bpo-2661: Make mapping tests better usable for custom mapping classes. (GH-11157)

In test_fromkeys() the derived test class now supports all arguments in its
constructor so that the class to be tested can use its own constructor in its
fromkeys() implementation.

In test_mutatingiteration() the test fails as soon as iterating over a
dictionary with one entry and adding new entries in the loop iterates more
than once (to avoid endless loops in faulty implementations).

https://bugs.python.org/issue2661

5 years agobpo-35551: remove mac_centeuro encoding (GH-13856)
Inada Naoki [Thu, 6 Jun 2019 05:38:52 +0000 (14:38 +0900)]
bpo-35551: remove mac_centeuro encoding (GH-13856)

It is alias to mac_latin2 now.

5 years agoAdd importlib.metadata to what's new. (#13855)
Barry Warsaw [Thu, 6 Jun 2019 02:40:19 +0000 (19:40 -0700)]
Add importlib.metadata to what's new. (#13855)

5 years agobpo-37134: Use PEP570 syntax for sum() (GH-13851)
Pablo Galindo [Wed, 5 Jun 2019 23:11:46 +0000 (00:11 +0100)]
bpo-37134: Use PEP570 syntax for sum() (GH-13851)

5 years agobpo-37165: Convert _count_elements to the argument clinic (GH-13848)
Raymond Hettinger [Wed, 5 Jun 2019 23:05:25 +0000 (16:05 -0700)]
bpo-37165: Convert _count_elements to the argument clinic (GH-13848)

5 years agobpo-35551: encodings update (GH-11446)
Ashwin Ramaswami [Wed, 5 Jun 2019 22:18:07 +0000 (15:18 -0700)]
bpo-35551: encodings update (GH-11446)

5 years agobpo-37134: Add PEP570 notation to the documentation (GH-13743)
Pablo Galindo [Wed, 5 Jun 2019 20:24:28 +0000 (21:24 +0100)]
bpo-37134: Add PEP570 notation to the documentation (GH-13743)

5 years agobpo-33725, multiprocessing doc: rephase warning against fork on macOS (GH-13841)
Victor Stinner [Wed, 5 Jun 2019 19:59:33 +0000 (21:59 +0200)]
bpo-33725, multiprocessing doc: rephase warning against fork on macOS (GH-13841)

Co-Authored-By: Barry Warsaw <barry@python.org>
5 years agobpo-21315: Fix parsing of encoded words with missing leading ws. (#13425)
Abhilash Raj [Wed, 5 Jun 2019 16:56:33 +0000 (12:56 -0400)]
bpo-21315: Fix parsing of encoded words with missing leading ws. (#13425)

* bpo-21315: Fix parsing of encoded words with missing leading ws.

Because of missing leading whitespace, encoded word would get parsed as
unstructured token. This patch fixes that by looking for encoded words when
splitting tokens with whitespace.

Missing trailing whitespace around encoded word now register a defect
instead.

Original patch suggestion by David R. Murray on bpo-21315.

5 years ago[3.9] bpo-37116: Use PEP 570 syntax for positional-only parameters. (GH-12620)
Serhiy Storchaka [Wed, 5 Jun 2019 15:22:31 +0000 (18:22 +0300)]
[3.9] bpo-37116: Use PEP 570 syntax for positional-only parameters. (GH-12620)

Turn deprecation warnings added in 3.8 into TypeError.

5 years agobpo-37158: Simplify and speed-up statistics.fmean() (GH-13832)
Raymond Hettinger [Wed, 5 Jun 2019 14:39:38 +0000 (07:39 -0700)]
bpo-37158: Simplify and speed-up statistics.fmean() (GH-13832)

5 years agobpo-26836: Document os.memfd_create() name parameter (GH-13838)
Victor Stinner [Wed, 5 Jun 2019 10:24:52 +0000 (12:24 +0200)]
bpo-26836: Document os.memfd_create() name parameter (GH-13838)

https://bugs.python.org/issue26836

5 years agobpo-34767: Do not always create a collections.deque() in asyncio.Lock() (GH-13834)
Zackery Spytz [Wed, 5 Jun 2019 09:33:27 +0000 (03:33 -0600)]
bpo-34767: Do not always create a collections.deque() in asyncio.Lock() (GH-13834)

https://bugs.python.org/issue34767

5 years agoFix documentation (GH-13721)
Benjamin Yeh [Wed, 5 Jun 2019 09:08:04 +0000 (02:08 -0700)]
Fix documentation (GH-13721)

Based on the source code https://github.com/python/cpython/blob/4a686504eb2bbf69adf78077458508a7ba131667/Lib/multiprocessing/pool.py#L755 AsyncResult.successful() raises a ValueError, not an AssertionError.

5 years agobpo-36373: Deprecate explicit loop parameter in all public asyncio APIs [streams...
Emmanuel Arias [Wed, 5 Jun 2019 05:45:53 +0000 (02:45 -0300)]
bpo-36373:  Deprecate explicit loop parameter in all public asyncio APIs [streams] (GH-13671)

This PR deprecate explicit loop parameters in all public asyncio APIs

This issues is split to be easier to review.

Second step: streams.py

https://bugs.python.org/issue36373

5 years ago bpo-35763: Make IDLE calltip note about '/' less obtrusive (GH-13791)
Terry Jan Reedy [Wed, 5 Jun 2019 01:55:37 +0000 (21:55 -0400)]
 bpo-35763: Make IDLE calltip note about '/' less obtrusive (GH-13791)

Add it to the end of the first line if there is room.  Tests were reworked.

5 years agoDoc: Python 3.9 in sidebar and version switcher. (GH-13824)
Julien Palard [Tue, 4 Jun 2019 23:15:33 +0000 (01:15 +0200)]
Doc: Python 3.9 in sidebar and version switcher. (GH-13824)

5 years agobpo-34282: Remove deprecated enum _convert method (GH-13823)
Zachary Ware [Tue, 4 Jun 2019 21:03:10 +0000 (16:03 -0500)]
bpo-34282: Remove deprecated enum _convert method (GH-13823)

5 years agoBump version in AppVeyor config (#13822)
Zachary Ware [Tue, 4 Jun 2019 20:57:15 +0000 (15:57 -0500)]
Bump version in AppVeyor config (#13822)

5 years agoBump to 3.9.0a0
Łukasz Langa [Tue, 4 Jun 2019 20:12:32 +0000 (22:12 +0200)]
Bump to 3.9.0a0

5 years agoMerge tag 'v3.8.0b1'
Łukasz Langa [Tue, 4 Jun 2019 19:53:26 +0000 (21:53 +0200)]
Merge tag 'v3.8.0b1'

Python 3.8.0b1

5 years agobpo-35047: Update whatsnew/3.8 for better mock error message (GH-13746)
Petter Strandmark [Tue, 4 Jun 2019 19:34:49 +0000 (21:34 +0200)]
bpo-35047: Update whatsnew/3.8 for better mock error message (GH-13746)

5 years agobpo-30835: email: Fix AttributeError when parsing invalid CTE (GH-13598)
Abhilash Raj [Tue, 4 Jun 2019 18:00:47 +0000 (14:00 -0400)]
bpo-30835: email: Fix AttributeError when parsing invalid CTE (GH-13598)

* bpo-30835: email: Fix AttributeError when parsing invalid Content-Transfer-Encoding

Parsing an email containing a multipart Content-Type, along with a
Content-Transfer-Encoding containing an invalid (non-ASCII-decodable) byte
will fail. email.feedparser.FeedParser._parsegen() gets the header and
attempts to convert it to lowercase before comparing it with the accepted
encodings, but as the header contains an invalid byte, it's returned as a
Header object rather than a str.

Cast the Content-Transfer-Encoding header to a str to avoid this.

Found using the AFL fuzzer.

Reported-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Andrew Donnellan <andrew@donnellan.id.au>
* Add email and NEWS entry for the bugfix.

5 years agoPython 3.8.0b1 v3.8.0b1
Łukasz Langa [Tue, 4 Jun 2019 17:44:34 +0000 (19:44 +0200)]
Python 3.8.0b1

5 years agobpo-35805: Add parser for Message-ID email header. (GH-13397)
Abhilash Raj [Tue, 4 Jun 2019 17:41:34 +0000 (13:41 -0400)]
bpo-35805: Add parser for Message-ID email header. (GH-13397)

* bpo-35805: Add parser for Message-ID header.

This parser is based on the definition of Identification Fields from RFC 5322
Sec 3.6.4.

This should also prevent folding of Message-ID header using RFC 2047 encoded
words and hence fix bpo-35805.

* Prevent folding of non-ascii message-id headers.
* Add fold method to MsgID token to prevent folding.

5 years agobpo-37153: test_venv.test_mutiprocessing() calls pool.terminate() (GH-13816)
Victor Stinner [Tue, 4 Jun 2019 17:03:13 +0000 (19:03 +0200)]
bpo-37153: test_venv.test_mutiprocessing() calls pool.terminate() (GH-13816)

test_venv.test_mutiprocessing() now explicitly calls pool.terminate()
to wait until the pool completes.

5 years agobpo-36742: Corrects fix to handle decomposition in usernames (#13812)
Steve Dower [Tue, 4 Jun 2019 15:55:30 +0000 (08:55 -0700)]
bpo-36742: Corrects fix to handle decomposition in usernames (#13812)

5 years agoFix extraneous whitespace in QueueListener.prepare (GH-13803)
Boris Feld [Tue, 4 Jun 2019 15:20:18 +0000 (17:20 +0200)]
Fix extraneous whitespace in QueueListener.prepare (GH-13803)

5 years agobpo-30699: Improve example on datetime tzinfo instances (GH-4290)
Mario Corchero [Tue, 4 Jun 2019 15:18:11 +0000 (16:18 +0100)]
bpo-30699: Improve example on datetime tzinfo instances (GH-4290)

* Improve example on tzinfo instances

Move from GMTX to TZX when naming the classes, as GMT1 might be rather
confusing as seen in the reported issue.
In addition, move to UTC over GMT and improve the tzname implementation.

* Simplify datetime with tzinfo example

Move the example in the documentation to just use timezone.utc and a
user defined Kabul timezone rather than having two user defined
timezones with DST.

Kabul timezone is still interesting as it changes its offset but not
based on DST. This is more accurate as the previous example was missing
information about the fold attribute. Additionally, implementing the fold
attribute was rather complex and probably not relevant enough for the
section "datetime with tzinfo".

5 years agobpo-36778: Remove outdated comment from CodePageTest (GH-13807)
Victor Stinner [Tue, 4 Jun 2019 15:09:10 +0000 (17:09 +0200)]
bpo-36778: Remove outdated comment from CodePageTest (GH-13807)

CP65001Test has been removed.

5 years agobpo-26219: Fix compiler warning in _PyCode_InitOpcache() (GH-13809)
Victor Stinner [Tue, 4 Jun 2019 15:08:24 +0000 (17:08 +0200)]
bpo-26219: Fix compiler warning in _PyCode_InitOpcache() (GH-13809)

Fix this MSVC warning:

    objects\codeobject.c(264): warning C4244: '=':
    conversion from 'Py_ssize_t' to 'unsigned char', possible loss of data