the process of writing docs for the other "missing" debug support
functions.

on option flags; added a versionadded decoration.

"soon", after I repair the LaTeX I somehow damaged.

unittest support function, from the public interface.  If they're not
documented, they shouldn't be public.

example.  Squash repeated argument descriptions.  Minor rewordings.

and the LaTeX docs are in increasingly good shape.

squashed massive duplication of common argument descriptions.

get rid of the itemize structure.

ELLIPSIS can be used to deal with examples that embed object addresses.

part of my previous last changes.

reStructuredText format.  Remove words describing the return value of
testmod() and testfile() in the intro sections, since it's never
useful in such simple cases.

- Add more ulink's to modules (this should probably become a standard macro).

buffers, compute the new buffer size based on the old buffer size.
Fixes SF bug #1034242.

'base').  Fixes SF bug #1033720.  Move offset sanity checking to
buffer_from_memory().

its documentation.

* Documented that the compiled re methods are supposed to be more full
  featured than their simpilified function counterparts.

* Documented the existing start and stop position arguments for the
  findall() and finditer() methods of compiled regular expression objects.

* Added an optional flags argument to the re.findall() and re.finditer()
  functions.  This aligns their API with that for re.search() and
  re.match().

platforms where that macro works, NaN compared to an int or long works
the same as NaN compared to a finite float.

today.  pyconfig.h can override it if not, and can also override
Py_IS_INFINITY now.  Py_IS_NAN and Py_IS_INFINITY are overridden now
for Microsoft compilers, using efficient MS-specific spellings.

When an integer is compared to a float now, the int isn't coerced to float.
This avoids spurious overflow exceptions and insane results.  This should
compute correct results, without raising spurious exceptions, in all cases
now -- although I expect that what happens when an int/long is compared to
a NaN is still a platform accident.

Note that we had potential problems here even with "short" ints, on boxes
where sizeof(long)==8.  There's #ifdef'ed code here to handle that, but
I can't test it as intended.  I tested it by changing the #ifdef to
trigger on my 32-bit box instead.

I suppose this is a bugfix candidate, but I won't backport it.  It's
long-winded (for speed) and messy (because the problem is messy).  Note
that this also depends on a previous 2.4 patch that introduced
_Py_SwappedOp[] as an extern.

Makes it easier on the eyes and a bit more snappy.