By accident the size of the empty dict keys object matched the
size of values array.
(cherry picked from commit 39a156c5)

(cherry picked from commit cb04f751)

(cherry picked from commit 4135c893)

bpo-31764: Prevent a crash in sqlite3.Cursor.close() in case the Cursor object is uninitialized (GH-3958) (#4303)

(cherry picked from commit edb13ae4)

bpo-31770: Prevent a crash and refleaks when calling sqlite3.Cursor.__init__() more than once (GH-3968) (#4301)

(cherry picked from commit e56ab746)

Manual partial backport of patch that fixed nearly 50 other files.

(cherry picked from commit aed08562)

(cherry picked from commit 0d2a9088)

(cherry picked from commit aafece7a)

(cherry picked from commit 82cd3ced)

(cherry picked from commit b838cc3f)

* Using ifconfig on NetBSD and OpenBSD.
* Using arp on Linux, FreeBSD, NetBSD and OpenBSD.

Based on patch by Takayuki Shimizukawa.
(cherry picked from commit ee1a9a2b)

(cherry picked from commit a7723d8b)

All Blake2 params have to be encoded in little-endian byte order. For
the two multi-byte integer params, leaf_length and node_offset, that
means that assigning a native-endian integer to them appears to work on
little-endian platforms, but gives the wrong result on big-endian. The
current libb2 API doesn't make that very clear, and @sneves is working
on new API functions in the GH issue above. In the meantime, we can work
around the problem by explicitly assigning little-endian values to the
parameter block.

See https://github.com/BLAKE2/libb2/issues/12.
(cherry picked from commit dcfb0e3c)

Bug report and patch by Jeroen Demeyer.
(cherry picked from commit f6f90ff0)

[3.6] bpo-31310: multiprocessing's semaphore tracker should be launched again if crashed (GH-3247) (#4254)

* bpo-31310: multiprocessing's semaphore tracker should be launched again if crashed

* Avoid mucking with process state in test.
Add a warning if the semaphore process died, as semaphores may then be leaked.

* Add NEWS entry
(cherry picked from commit cbe1756e)

When a single .c file contains several functions and/or methods with
the same name, a safety _METHODDEF GH-define statement is generated
only for one of them.

This fixes the bug by using the full name of the function to avoid
duplicates rather than just the name..
(cherry picked from commit 4f57409a)

[3.6] bpo-31308: If multiprocessing's forkserver dies, launch it again when necessary (GH-3246) (#4252)

* bpo-31308: If multiprocessing's forkserver dies, launch it again when necessary.

* Fix test on Windows

* Add NEWS entry

* Adopt a different approach: ignore SIGINT and SIGTERM, as in semaphore tracker.

* Fix comment

* Make sure the test doesn't muck with process state

* Also test previously-started processes

* Update 2017-08-30-17-59-36.bpo-31308.KbexyC.rst

* Avoid masking SIGTERM in forkserver.  It's not necessary and causes a race condition in test_many_processes..
(cherry picked from commit fc6b348b)

netrc file format doesn't support quotes and escapes.

See https://linux.die.net/man/5/netrc
(cherry picked from commit b24cd055)

(cherry picked from commit ac03c03b)

(cherry picked from commit 0d2c645d)

(cherry picked from commit 4f469c09)

Skip a test for unget_wch()/get_wch() on OpenBSD since they are broken
in ncurses 5.7.
(cherry picked from commit 7e68790f)

(cherry picked from commit 894ebd06)

bpo-31910: test_create_connection() now catchs also EADDRNOTAVAIL to
fix the test on Travis CI.
(cherry picked from commit 280c22a8)

Replace occurence of nested comments in blake2 reference implementation
with preprocessor directive for disabling unused code.

`blake2s-load-xop.h` is conditionally pulled in only on chips with XOP
support, among others the AMD Bulldozer. The malformed comments in the
source file breaks the build of `hashlib`'s `_blake2` on GCC 6.3.0.

Official reference code on github uses `GH-if` so this change should be
uncontroversial.
(cherry picked from commit 388cd85e)

Modify the code to use ncurses is_pad() instead of checking WINDOW
_flags field.  If your platform does not provide the is_pad(), the
existing way that checks the field will be enabled.

Note: This change does not drop support for platforms where do not
have both WINDOW _flags field and is_pad().
(cherry picked from commit 8bc7d635)

* Fix a compilation error on FreeBSD.
* Fix the data attribute size on Mac OS X.
(cherry picked from commit 2298fad5)

Document the following functions:

* PyObject_Malloc()
* PyObject_Calloc()
* PyObject_Realloc()
* PyObject_Free()

Fix also PyMem_RawFree() documentation.
(cherry picked from commit ec2cbdd1)

bpo-31897: Convert unexpected errors when read bogus binary plists into InvalidFileException. (GH-4171) (#4192)

(cherry picked from commit db91e0fe)

Removed a code that incorrectly detected in-place resizing in realloc()
 and wrote to freed memory.
(cherry picked from commit b484d560)

* Fixed the layout of the kqueue_event structure on OpenBSD and NetBSD.
* Fixed the comparison of the kqueue_event objects.
(cherry picked from commit b9052a0f)

(cherry picked from commit baac01e6)

test_curses now saves/restores signals. On FreeBSD, the curses module
sets handlers of some signals, but don't restore old handlers when
the module is deinitialized.

(cherry picked from commit 19f68301)

Update the kwarg in the documentation of `multiprocessing.pool.Pool.starmap_async`, from `error_back` to `error_callback` to match the source code.
(cherry picked from commit 11225753)

separators that are not bytes-like objects..
(cherry picked from commit a2314283)

(cherry picked from commit 006617ff)

(cherry picked from commit 5a4bbcd4)