Repair so that it's back to how it was in 2.4.3.

Needs to be forward-ported to 2.5 branch and trunk.

_bsddb on Windows (Martin recently repaired that -- thanks!).

fixing an unlikely crash bug in dict resizing, SF
bug 1456209.

The rest of rev 46589 changes whether Python suppresses
exceptions during some dict-related comparisons.  While I
think that's a good idea, it does change visible behavior at
times, and there was already some complaining about that on
the trunk.  Not a good idea for backporting.  The part of
46589 checked in here can at worst stop segfaults, and I doubt
anyone will gripe about that ;-)

- reset errno before calling confstr - use confstr() doc to simplify
  checks afterwards
- Correct implementation and documentation of os.confstr.  Add a simple
  test case.  I've yet to figure out how to provoke a None return I can test.
- Address issues brought up by MvL on python-checkins.
  I tested this with valgrind on amd64.

  The man pages I found for diff architectures are inconsistent on this.
  I'm not entirely sure this change is correct for all architectures
  either.

  Perhaps we should just over-allocate and not worry about it?

The change to return None instead of "" in case of unconfigured
values has not been backported.

Modules".