so that the current directory does not get added to sys.path.
Also fixes #1526785.

- update header checks, using autoconf
- provide dummies for getenv, environ, and GetVersion
- adjust MSC_VER check in socketmodule.c

mixing signed and unsigned types in comparison.
Relatedly, `dataSize` is declared as DWORD, not as
int, so change relevant cast from (int) to (DWORD).

This was probably fixed in rev 1.32 of getpath.c, but there are so
many paths thru the code that invoke joinpath() it's not at all
obvious that it *is* fixed.  It doesn't help confidence that a crucial
precondition for calling joinpath() was neither documented nor verified.
It is now, and joinpath() will barf with a fatal error now rather than
overrun the buffer, if the precondition isn't met.

Note that this patch only changes the Windows flavor.  I attached another
patch to the bug report for the POSIX flavor (which I can't test
conveniently).

Will backport to release23-maint.

Note this code is not used by the core on Win32, but in a block used only
by Windows CE.

- new import hooks in import.c, exposed in the sys module
- new module called 'zipimport'
- various changes to allow bootstrapping from zip files

I hope I didn't break the Windows build (or anything else for that
matter), but then again, it's been sitting on sf long enough...

Regarding the latest discussions on python-dev: zipimport sets
pkg.__path__ as specified in PEP 273, and likewise, sys.path item such as
/path/to/Archive.zip/subdir/ are supported again.

Rename all occurrences of MS_WIN32 to MS_WINDOWS.

If after calculating sys.path we do not have sys.prefix set, we loop over all path entries checking if one can point to our home directory.

and PYTHONPATH).

Checkin updated version of patch #103933 .  As Thomas says, fixes the bugs #131064, #129584, #127722. See the discussion in bug #131064

Prevent possible buffer overflow exploits under Windows.  As per (the very quick) patch Patch #101801.

If there was a NULL registry key, Python could barf.

Also wraps some surrounding lines to 80 chars.

This should match the situation in the 1.6b1 tree.

they include prototypes.

char**) and return an int even on PC platforms. If not, please fix
PC/utils/makesrc.c ;-P

Python on UNIX now trusts PYTHONHOME unconditionally

Modules/getpath.c:

Landmark changed to os.py.

Setting PYTHONHOME now unconditionally sets sys.prefix
(and sys.exec_prefix). No further checks are done whether the
standard lib can be found in that location or not. This is in
sync with the PC subdir getpath implementations.

PC/getpathp.c:

Landmark changed to os.py.

PC/os2vacpp/getpathp.c:

Landmark changed to os.py.

Note: BAW's checkin on exceptions.c eliminates earlier concerns about
a bogus PYTHONHOME value leading to a core dump.  Instead it causes a
useless sys.path and prevents imports.

light of three different situations: (1) running from build; (2)
running from installed; (3) running without being able to find an
installation (e.g. as a COM object).  The system paths in the
repository are only used for (3); the path deduced from the
installation location are used otherwise.  PYTHONHOME overrides in all
cases.

Read the comments for more details.

running out of the build directory.  This means that it will no longer
try to use an older version of the library when an older version has
been installed.

The registry always comes first and the default is always appended.

Make an explicit test for whether the prefix is in fact the
source directory, and then don't use the registry.

when it gets the path from the registry, it no longer appends the
default path to the end (which would mostly be a duplication).

- MS_DLL_ID is now set to "1.5" instead of "1.5.x" so Mark Hammond's
  extensions won't have to be changed.
- FILEVERSION and PRODUCTVERSION are set to 1,5,2,1.
  (The last number could be the build# or the alpha release# or so.)

(1) Use Py_GetPythonHome() instead of getenv("PYTHONHOME");

(2) Mark Hammond's patch to search for .pyc/.pyo landmark as well.

that was only causing confusing).  Add free(userpath) and
free(machinepath) statements to prevent some leaks.

again (Mark Hammond is the cause of all this).

is now obsolete), and changed the default path calculations.
$PYTHONPATH is now added as a prefix (like it's always been on Unix);
$PYTHONHOME takes precedence over the program pathname; and only one
landmark is needed.

lives to the end of the path.

(Still to do: add $PYTHONPATH to the front instead of using it as is;
add the Win32 registry paths as in getpath_nt.c (which can then
retire).)