from Tim Hochberg.  Also mucho fiddling to change the way doctest
determines whether a thing is a function, module or class.  Under 2.2,
this really requires the functions in inspect.py (e.g., types.ClassType
is close to meaningless now, if not outright misleading).

public interface, so we can simplify the documentation.

try to explain the complex general scheme we actually use now, I decided
to spell out only what equality means (which is easy to explain and
intuitive), leaving the other outcomes unspecified beyond consistency.

The patch repaired internal gcc compiler errors on BeOS.
This checkin repairs them in a simpler way, by explicitly casting the
platform INFINITY to double.

added local escape method (made the dumps method some 50-80% faster)
minor tweaks to the unmarshalling code

Patch from Steve Scott to add SIGBREAK support (unique to Windows).

top level.

- SLOT1BINFULL() macro: changed this to check for __rop__ overriding
  __op__, like binary_op1() in abstract.c -- the latter only calls the
  slot function once if both types use the same slot function, so the
  slot function must make both calls -- which it already did for the
  __op__, __rop__ order, but not yet for the __rop__, __op__ order
  when B.__class__ is a subclass of A.__class__.

- slot_sq_contains(), slot_nb_nonzero(): use lookup_maybe() rather
  than lookup_method() which sets an exception which we then clear.

- slot_nb_coerce(): don't give up when left argument's __coerce__
returns NotImplemented, but give the right argument a chance.

(to see whether __rop__ should go before __op__) only when slotv is
set.  This saves a test+branch when only slotw is set.

"boolean" --> "Boolean" (per the style guide).

as container objects rather than as mapping objects (in the index entries).
Change the section heading and intro sentence to be a little more general,
since that's how things have actually evolved.

  I modified nntplib so the body method can accept an
  optional second parameter pointing to a filehandle or
  filename (string). This way, really long body
  articles can be stored to disk instead of kept in
  memory. The way I made the modification should make
  it easy to extend this functionality to other extended
  return methods.

Decode and encode underscores for header style encoding. Fixes bug #463996.

it acts more like assigning to keys in locals(), i.e. modifications to
function locals aren't reflected in the locals when execfile() returns.

Replaced 3 instances of "iter() of non-sequence" with
"iteration over non-sequence".
Restored "unpack non-sequence" for stuff like "a, b = 1".

Generalize PyLong_AsLongLong to accept int arguments too.  The real point
is so that PyArg_ParseTuple's 'L' code does too.  That code was
undocumented (AFAICT), so documented it.

recent user feedback: you must end the input with \n and you must use
\n, not \r\n to represent line endings.

whitespace characters to a single space.
Small changes elsewhere, mostly to clean up the code a little.

This is probably a little bit faster, but mostly is just cleaner code.
The old-style support is still used for Python versions < 2.2 so this
source file can be shared with PyXML.

    every time this gets called; move it out as a global helper function.
    Simplify the call to the _dispatch() method of the registered instance.