Back-porting to release22-maint.
This closes SF patch #496705.

    and returns None.  This allows any object that supports the fileno()
    method to be passed as a file descriptor, not just an integer.

posix_fchdir():  New exposed function: implements posix.fchdir().  This
    closes SF feature #536796.

posix_fsync(), posix_fdatasync():  Convert to use posix_fildes() instead
    of posix_int().  This also changes them from METH_VARARGS to METH_O
    functions.

setup_confname_table():  Remove unused variable.  Change to take a module
    rather than a dict to save the resulting table into.

setup_confname_tables():  Change to take a module instead of a dict to
    pass to setup_confname_table().

Thanks to Thomas Heller for the sharp eye.

This is part of SF feature #536796.

the Extending & Embedding manual use it.

instance, otherwise make a copy.

This moves styling to the stylesheet; the use of <dl> structures to control
style sometimes produced improper indentation of subsequent text in many
browsers when the text was already part of the <dl> structure (as in a
function or class description).

isn't constant, so why bother.

Folded long lines.

Whitespace normalization.

Add a method zfill to str, unicode and UserString and change
Lib/string.py accordingly.

This activates the zfill version in unicodeobject.c that was
commented out and implements the same in stringobject.c. It also
adds the test for unicode support in Lib/string.py back in and
uses repr() instead() of str() (as it was before Lib/string.py 1.62)

is the same in all three cases (mostly because // calls divmod :-).

Complex numbers implement divmod() and //, neither of which makes one
lick of sense.  Unfortunately this is documented, so I'm adding a
deprecation warning now, so we can delete this silliness, oh, around
2005 or so.

Bugfix candidate (At least for 2.2.2, I think.)

A stupid bug in object_set_class(): didn't check for value==NULL
before checking its type.

Bugfix candidate.

#543674.

Bugfix candidate.

In DatagramRequestHandler.setup(), the wfile initialization should be
StringIO.StringIO(), not StringIO.StringIO(slf.packet).

Bugfix candidate (all the way back to Python 1.5.2 :-).

When os.stat() for a file raises OSError, turn it into IOError per
documentation.

Bugfix candidate.

The problem was that an exception can occur in the text.get() call or
in the write() call, when the text buffer contains non-ASCII
characters.  This causes the previous contents of the file to be lost.

The provisional fix is to call str(self.text.get(...)) *before*
opening the file, so that if the exception occurs, we never open the
file.

Two orthogonal better solutions have to wait for policy decisions:

1. We could try to encode the data as Latin-1 or as UTF-8; but that
   would require IDLE to grow a notion of file encoding which requires
   more thought.

2. We could make backups before overwriting a file.  This requires
   more thought because it needs to be fast and cross-platform and
   configurable.

complex_subtype_from_string():  this stopped parsing at the first 0
byte, as if that were the end of the input string.

Bugfix candidate.

Highlights: import and friends will understand any of \r, \n and \r\n
as end of line. Python file input will do the same if you use mode 'U'.
Everything can be disabled by configuring with --without-universal-newlines.

See PEP278 for details.

Added code to call this when PYMALLOC_DEBUG is enabled, and envar
PYTHONMALLOCSTATS is set, whenever a new arena is obtained and once
late in the Python shutdown process.