The bitshift and xor op for 32bit builds has changed the order of hash values.

Issue #19183: test_gdb's test_dict was failing on some machines as the order or dict keys has changed again.

Python now uses SipHash24 on all major platforms.

Now './configure --silent && make -s' doesn't print any message to stdout or stderr.

"hf" type is INT_PTR, it is used to store an int in _msi.c.

On AIX, the C locale encoding uses the ISO-8859-1 encoding, not ASCII.

Some compilers (ex: Visual Studio) decode -2147483648 as a unsigned integer
instead of an signed integer.

INT32_MIN and INT32_MAX constants are unknown on Windows.

compiler warnings on Windows 64-bit. Use Py_SAFE_DOWNCAST() where the final
downcast is needed.

The bytecode doesn't support integer parameters larger than 32-bit yet.

csv was handling non-string keys fine except for the error message
generated when a non-string key was not in 'fieldnames'.

Fix by Tomas Grahn, full patch-with-test by Vajrasky Kok (tweaked slightly).

Also, add a link to the Babel package.

The zlib and hex codecs throw custom exception types with
weakref support if the input type is valid, but the data
fails validation. Make sure the exception chaining in the
codec infrastructure can wrap those as well.

The utf-16* and utf-32* encoders no longer allow surrogate code points
(U+D800-U+DFFF) to be encoded.
The utf-32* decoders no longer decode byte sequences that correspond to
surrogate code points.
The surrogatepass error handler now works with the utf-16* and utf-32* codecs.

Based on patches by Victor Stinner and Kang-Hao (Kenny) Lu.

This will be merged into default when PEP451 is merged in.  See
changeset 5d38989191bb in features/pep-451

to avoid reporting success on empty tests.

exists in 3.4 but apparently wasn't done for 3.3).