- return the full size of the sockaddr_un structure, without which
  bind() fails with EINVAL;
- set test_socketserver to use a socket name that meets the form
  required by the underlying implementation;
- don't bother exercising the forking AF_UNIX tests on EMX - its
  fork() can't handle the stress.

M IOBinding.py
M NEWS.txt
M configDialog.py

- If nulls somehow got into the strings in recent-files.lst
  EditorWindow.update_recent_files_list() was failing.  Python Bug 931336.

did not stand-up to additional timings.

globaltrace_countfuncs() into file_module_function_of().

In that function use Michael Hudson's suggestion of gc.get_referrers() to
back up from the code object to a function, then to a class's dict and
finally to a class object if one exists.

relationships at program exit.  Output is a bit prettier than that for
--listfuncs but won't parse as easily using downstream postprocessing tools.

pre-increment forms to post-increment forms.  Post-incrementing
also eliminates the need for negative array indices for oparg fetches.

* In exception handling code, check for class based exceptions before
  the older string based exceptions.

BINARY_SUBSCR:
    * invert test for normal case fall through
    * eliminate err handling code by jumping to slow_case

LOAD_LOCALS:
    * invert test for normal case fall through
    * continue instead of break for the non-error case

STORE_NAME and DELETE_NAME:
    * invert test for normal case fall through

LOAD_NAME:
    * continue instead of break for the non-error case

DELETE_FAST:
    * invert test for normal case fall through

LOAD_DEREF:
    * invert test for normal case fall through
    * continue instead of break for the non-error case

Restores the self-documenting enum declaration.

This allows multiple flags to be tested in a single compare
which eliminates unnecessary compares and saves a few bytes.

been used. (Suggested by Martin v. Loewis)

tests of "why" against WHY_YIELD became useless.  This patch removes them,
but assert()s that why != WHY_YIELD everywhere such a test was removed.
The test suite ran fine under a debug build (i.e., the asserts never
triggered).

attribute name lookup.

the socket name must start with /socket32/.  Unlike Unix systems, this
file never exists in the file system.

with major C compilers (VACPP, EMX+gcc and [Open]Watcom).

Also tidy up the export of spawn*() symbols in the os module to match what
is found/implemented.

close() calls would attempt to free() the buffer already free()ed on
the first close().     [bug introduced with patch #788249]

Making sure that the buffer is free()ed in file object deallocation is
a belt-n-braces bit of insurance against a memory leak.

Also, move pystack comment down next to it

as directly setting the .prog attribute (which should be supported based on
the class docstring).
Closes SF bug #850964.

Revise the long comment that explained details of the magic number
in gory detail.

[Bugfix candidate] Escape traceback type and value.  There are probably additional cases where cgitb.py doesn't escape as paranoidly as it should (e.g. attribute names)

Backported to 2.3.

inherit those bits, causing the test_mkdtemp.test_mode() test to fail.
Remove those before comparing the actual mode to the expected mode.

PDF and PostScript versions of the docs (the CSS already does this for HTML)

Closes SF bug #837228; backporting for Python 2.3.4.