Cheryl Sabella [Wed, 31 Jan 2018 21:37:51 +0000 (16:37 -0500)]
bpo-32735: Fix typo in f-strings datetime format specifier example (GH-5464)
The f-string example for using datetime format specifier does not match the given output.
Changed the format from %b to %B so it matches the output of "January".
Ned Deily [Tue, 30 Jan 2018 12:42:14 +0000 (07:42 -0500)]
bpo-32726: macOS installer and framework enhancements and changes for 3.7.0 (GH-5448)
This issue covers various changes for the macOS installers provided via python.org for 3.7.0.
- Provide a provisional new installer variant for macOS 10.9 and later systems with 64-bit (x86_64) architecture only. Apple has made it known that future versions of macOS will only fully support 64-bit executables and some other third-party software suppliers have chosen 10.9 as their oldest supported system.
- Support **Tcl/Tk 8.6** with the 10.9 installer variant.
- Upgrade **OpenSSL** to 1.1.0g and **SQLite** to 3.22.0.
- The compiler name used for the interpreter build and for modules built with **Distutils / pip** is now _gcc_ rather than _gcc-4.2_. And extension module builds will no longer try to force use of an old SDK if present.
Ned Deily [Tue, 30 Jan 2018 10:14:09 +0000 (05:14 -0500)]
bpo-28440: Don't add /Library/Python/3.x/site-packages to sys.path (#5445)
No longer add /Library/Python/3.x/site-packages, the Apple-supplied
system Python site-packages directory, to sys.path for macOS framework
builds in case Apple ships a version of Python 3. A similar change
was made earlier to Python 2.7 where it was found that the coupling
between the system Python and a user-installed framework Python often
caused confusion or pip install failures.
Eric Appelt [Tue, 30 Jan 2018 07:13:03 +0000 (01:13 -0600)]
closes bpo-30117: fix lib2to3 ParserIdempotency test (GH-1242)
Fix two (in my opinion) spurious failure conditions in the lib2to3.tests.test_parser.TestParserIdempotency test_parser test.
Use the same encoding found in the initial file to write a temp file for a diff. This retains the BOM if the encoding was initially utf-8-sig.
If the file cannot be parsed using the normal grammar, try again with no print statement which should succeed for valid files using future print_function
For case (1), the driver was correctly handling a BOM in a utf-8 file, but then the test was not writing a comparison file using 'utf-8-sig' to diff against, so the BOM got removed. I don't think that is the fault of the parser, and lib2to3 will retain the BOM.
For case (2), lib2to3 pre-detects the use of from __future__ import print_function or allows the user to force this interpretation with a -p flag, and then selects a different grammar with the print statement removed. That makes the test cases unfair to this test as the driver itself doesn't know which grammar to use. As a minimal fix, the test will try using a grammar with the print statement, and if that fails fall back on a grammar without it. A more thorough handling of the idempotency test would to be to parse all files using both grammars and ignore if one of the two failed but otherwise check both. I didn't think this was necessary but can change.
Gregory P. Smith [Tue, 30 Jan 2018 05:27:39 +0000 (21:27 -0800)]
bpo-25942: make subprocess more graceful on ^C (GH-5026)
Do not allow receiving a SIGINT to cause the subprocess module to trigger an
immediate SIGKILL of the child process. SIGINT is normally sent to all child
processes by the OS at the same time already as was the established normal
behavior in 2.7 and 3.2. This behavior change was introduced during the fix to https://bugs.python.org/issue12494 and is generally surprising to command line
tool users who expect other tools launched in child processes to get their own
SIGINT and do their own cleanup.
In Python 3.3-3.6 subprocess.call and subprocess.run would immediately
SIGKILL the child process upon receiving a SIGINT (which raises a
KeyboardInterrupt). We now give the child a small amount of time to
exit gracefully before resorting to a SIGKILL.
This is also the case for subprocess.Popen.__exit__ which would
previously block indefinitely waiting for the child to die. This was
hidden from many users by virtue of subprocess.call and subprocess.run
sending the signal immediately.
Behavior change: subprocess.Popen.__exit__ will not block indefinitely
when the exiting exception is a KeyboardInterrupt. This is done for
user friendliness as people expect their ^C to actually happen. This
could cause occasional orphaned Popen objects when not using `call` or
`run` with a child process that hasn't exited.
Refactoring involved: The Popen.wait method deals with the
KeyboardInterrupt second chance, existing platform specific internals
have been renamed to _wait().
Also fixes comment typos.
Christian Heimes [Mon, 29 Jan 2018 21:37:58 +0000 (22:37 +0100)]
bpo-28134: Auto-detect socket values from file descriptor (#1349)
Fix socket(fileno=fd) by auto-detecting the socket's family, type,
and proto from the file descriptor. The auto-detection can be overruled
by passing in family, type, and proto explicitly.
Without the fix, all socket except for TCP/IP over IPv4 are basically broken:
Pablo Galindo [Mon, 29 Jan 2018 20:34:42 +0000 (20:34 +0000)]
bpo-20104: Fix leaks and errors in new os.posix_spawn (GH-5418)
* Fix memory leaks and error handling in posix spawn
* Improve error handling when destroying the file_actions object
* Py_DECREF the result of PySequence_Fast on error
* Handle uninitialized pid
* Use OSError if file actions fails to initialize
* Move _file_actions to outer scope to avoid undefined behaviour
* Remove HAVE_POSIX_SPAWN define in Modules/posixmodule.c
* Unshadow exception and clean error message
rename _imp initialization function to follow conventions (#5432)
When the C imp module became _imp in 6f44d66bc491bad5b8d897a68da68e009e27829d, the initialization function should have been renamed from PyInit_imp to PyInit__imp.
Yury Selivanov [Mon, 29 Jan 2018 19:31:47 +0000 (14:31 -0500)]
bpo-32703: Fix coroutine resource warning in case where there's an error (GH-5410)
The commit removes one unnecessary "if" clause in genobject.c. That "if" clause was masking un-awaited coroutines warnings just to make writing unittests more convenient.
Christian Heimes [Mon, 29 Jan 2018 13:25:13 +0000 (14:25 +0100)]
bpo-28414: ssl module idna test (#5395)
Add test cases for IDNA 2003 and 2008 host names. IDNA 2003
internationalized host names are working since bpo-31399 has landed. IDNA
2008 deviations are still broken and will be fixed in another patch.
Signed-off-by: Christian Heimes <christian@python.org>
Christian Heimes [Mon, 29 Jan 2018 13:10:18 +0000 (14:10 +0100)]
bpo-31429: Define TLS cipher suite on build time (#3532)
Until now Python used a hard coded white list of default TLS cipher
suites. The old approach has multiple downsides. OpenSSL's default
selection was completely overruled. Python did neither benefit from new
cipher suites (ChaCha20, TLS 1.3 suites) nor blacklisted cipher suites.
For example we used to re-enable 3DES.
Python now defaults to OpenSSL DEFAULT cipher suite selection and black
lists all unwanted ciphers. Downstream vendors can override the default
cipher list with --with-ssl-default-suites.
Signed-off-by: Christian Heimes <christian@python.org>
socketserver.ForkingMixIn.server_close() and
socketserver.ThreadingMixIn.server_close() now waits until all child
processes and non-daemonic threads complete.
Martin Teichmann [Sun, 28 Jan 2018 04:17:46 +0000 (05:17 +0100)]
bpo-30306: release arguments of contextmanager (GH-1500)
The arguments to a generator function which is declared as a
contextmanager are stored inside the context manager, and
thus are kept alive, even when it is used as a regular context
manager, and not as a function decorator (where it needs
the original arguments to recreate the generator on each
call).
This is a possible unnecessary memory leak, so this changes
contextmanager.__enter__ to release the saved arguments,
as that method being called means that particular CM instance
isn't going to need to recreate the underlying generator.
Eric V. Smith [Sun, 28 Jan 2018 00:07:40 +0000 (19:07 -0500)]
bpo-32513: Make it easier to override dunders in dataclasses. (GH-5366)
Class authors no longer need to specify repr=False if they want to provide a custom __repr__ for dataclasses. The same thing applies for the other dunder methods that the dataclass decorator adds. If dataclass finds that a dunder methods is defined in the class, it will not overwrite it.
Ammar Askar [Sat, 27 Jan 2018 20:53:56 +0000 (15:53 -0500)]
Fix changed file detection on Travis (GH-3129)
Travis when merging changes from a pull request onto the target branch
does not perform a rebase, instead it does a simple merge which causes
the PR commits to retain their commit dates. This means that the commit
log can potentially look like:
PR merge <-- HEAD
normal master commit <- master
more commits from normal workflow
PR commit 1
another master commit
PR commit 2
Performing a git diff from PR commit 2 to master will accidentally
include files that should not be there.
Christian Heimes [Sat, 27 Jan 2018 14:51:38 +0000 (15:51 +0100)]
bpo-31399: Let OpenSSL verify hostname and IP address (#3462)
bpo-31399: Let OpenSSL verify hostname and IP
The ssl module now uses OpenSSL's X509_VERIFY_PARAM_set1_host() and
X509_VERIFY_PARAM_set1_ip() API to verify hostname and IP addresses.
* Remove match_hostname calls
* Check for libssl with set1_host, libssl must provide X509_VERIFY_PARAM_set1_host()
* Add documentation for OpenSSL 1.0.2 requirement
* Don't support OpenSSL special mode with a leading dot, e.g. ".example.org" matches "www.example.org". It's not standard conform.
* Add hostname_checks_common_name
Signed-off-by: Christian Heimes <christian@python.org>
Christian Heimes [Sat, 27 Jan 2018 08:53:43 +0000 (09:53 +0100)]
bpo-32433: Optimized HMAC digest (#5023)
The hmac module now has hmac.digest(), which provides an optimized HMAC
digest for short messages. hmac.digest() is up to three times faster
than hmac.HMAC().digest().
Signed-off-by: Christian Heimes <christian@python.org>
bpo-32591: fix abort in _PyErr_WarnUnawaitedCoroutine during shutdown (#5337)
When an unawaited coroutine is collected very late in shutdown --
like, during the final GC at the end of PyImport_Cleanup -- then it
was triggering an interpreter abort, because we'd try to look up the
"warnings" module and not only was it missing (we were prepared for
that), but the entire module system was missing (which we were not
prepared for).
I've tried to fix this at the source, by making the utility function
get_warnings_attr robust against this in general. Note that it already
has the convention that it can return NULL without setting an error,
which is how it signals that the attribute it was asked to fetch is
missing, and that all callers already check for NULL returns.
There's a similar check for being late in shutdown at the top of
warn_explicit, which might be unnecessary after this fix, but I'm not
sure so I'm going to leave it.
Guido van Rossum [Fri, 26 Jan 2018 16:20:18 +0000 (08:20 -0800)]
String annotations [PEP 563] (#4390)
* Document `from __future__ import annotations`
* Provide plumbing and tests for `from __future__ import annotations`
* Implement unparsing the AST back to string form
This is required for PEP 563 and as such only implements a part of the
unparsing process that covers expressions.