(suggested by Jack Jansen)

externally visible difference is that the factory defaults are
now in a plist file in the bundle, in stead of being hard-coded in
the application.

whether they crash or not is not a matter of using a framework or not,
but whether the code is run from an app bundle or not. And that it
_can_ crash (instead of tracing back) is Apple's bug anyway, not ours.

Include all First/Last blocks.

guarantee abs(x%y) < abs(y) in all cases when a float is involved.
math.fmod() should, though, so noted that too.

Bugfix candidate.  Someone should check the LaTeX here first, though.

file directly (instead of sucking it all into a list of lines first).

and only if the test input file doesn't exist.

data it needs doesn't exist in the project.  If it's a huge file,
maybe the test should be changed to be one of the -u thingies.

XXX If NormalizationTest.txt is required to run this test, why isn't it
checked into the project?

dereference rather than the caller.

both have a nb_remainer slot.

whether this is a correct thing to do:

+ There are linker warnings (see PCbuild\readme.txt).
+ test_bsddb passes, in both release and debug builds now.
+ test_bsddb3 has several failures, but it did before too.

Also made pythoncore a dependency of the _bsddb project, updated
build instructions, added database conversion XXX to NEWS, and fiddled
the Windows installer accordingly.

Verify that the lookup for an existing name succeeds.

arguments.

dict() constructor. Example:
  >>> dict(a=1, b=2)
  {'a': 1, 'b': 2}
  >>>

+ News blurb, but as much XXX as news.
+ Updated installer (install the new bsddb package, and the Berkeley DLL;
  still don't know how to fold that into _bsddb.pyd).
+ Fleshed out build instructions.
+ Debug Python still blows up.