computing 2.0**-1074 accurately.  Using ldexp(1.0, -1074) should be
safer.

always set a dlerror() message.

to convert floats to and from hexadecimal strings respectively.

release build) when the format string of a pointer to an incomplete
structure is created.

Issue 2235: __hash__ is once again inherited by default, but inheritance can be blocked explicitly so that collections.Hashable remains meaningful

Added '#' formatting to integers.  This adds the 0b, 0o, or 0x prefix for bin, oct, hex.  There's still one failing case, and I need to finish the docs.  I hope to finish those today.

The fact that this was failing and went unnoticed so long seems like a good argument for being able to enable and disble py3kwarnings through Python.

process rather than both parent and child.

Does anyone actually use fork1()?  It appears to be a Solaris thing
but if Python is built with pthreads on Solaris, fork1() and fork()
should be the same.

Changed statements of the form Py_DECREF(obj), obj = 0 to Py_CLEAR(obj).

in Py_CLEAR().

Fix issue 3221 by emitting a RuntimeWarning instead of raising SystemError when the parent module can't be found during an absolute import (likely due to non-PEP 361 aware code which sets a module level __package__ attribute)

Correct a couple of errors in the updated catch_warning documentation (the Py3k version was fixed before being checked in)

Make test.test_support.catch_warnings more robust as discussed on python-dev. Also add explicit tests for itto test_warnings.

should have returned True.

Fixes issue #3339. Thanks, Henk Punt for the report and Andrii v. Mishkovskiyi
for attempting a patch.

#3343: Py_DisplaySourceLine should be a private function. Rename it to _Py_DisplaySourceLine.

some applications relied on them.

Also remove duplicated lines.