properties:  the docstring (if any) is displayed, and the getter, setter
  and deleter (if any) functions are named.  All that is shown indented
  after the property name.

+ Text-mode pydoc class display now draws a horizontal line between
  class attribute groups (similar to GUI mode -- while visually more
  intrusive in text mode, it's still an improvement).

- property() now takes 4 keyword arguments:  fget, fset, fdel, doc.
  Note that the real purpose of the 'f' prefix is to make fdel fit in
  ('del' is a keyword, so can't used as a keyword argument name).

- These map to visible readonly attributes 'fget', 'fset', 'fdel',
  and '__doc__' in the property object.

- fget/fset/fdel weren't discoverable from Python before.

- __doc__ is new, and allows to associate a docstring with a property.

declarations and weird markup that we used to accept & ignore that recent
versions raised an exception for; the original behavior has been restored
and augmented (the user can decide what to do if they care; the default is
to ignore it as done in early versions).

Use a new internal method, error(), consistently to raise parse errors;
the new base class also uses this.
Adjust the parse_comment() method to return the new offset into the buffer
instead of the number of characters scanned; this was the only helper
method that did it this way, so we have better consistency now.  Required
to share the new base class.
This fixes SF bug #448482 and #453706.

Use a new internal method, error(), consistently to raise parse errors;
the new base class also uses this.

and HTMLParser modules (and indirectly for the htmllib.HTMLParser class).

This has all the support for scanning over DOCTYPE declarations; it warrants
having a base class since this is a fair amount of tedious code (since it's
fairly strict), and should be in a separate module to avoid compiling many
REs that are not used (which would happen if this were placed in either then
sgmllib or HTMLParser module).

popped frame-block. What an embarrassing bug! Especially for Jeremy, since
he accepted the patch :-)

This fixes SF bugs #463359 and #462937, and possibly other, *very* obscure
bugs with very deeply nested loops that continue the loop and then break out
of it or raise an exception.

to one with a static metatype raised an obscure error.

and the information provided to the profiler.  This stuff is a mess!

function returns NotImplemented when comparing objects whose
tp_richcompare slot is not itself.

input to .write() too.

this week (apparently me and Bob Heeter at more or less the same time).

arguments are subclasses of str, as long as they don't override rich
comparison.

looks like <X object at ...>, sometimes it says <X instance at ...>.
Make this uniformly say <X object at ...>.

introduced in Python 2.2.
Add documentation for the slice object interface (not complete).
Added version annotations for several of the Python 2.2 APIs already
documented.

Fix two errors

+ Minor code cleanup, generalization and simplification.

+ "Do something" to make the attribute aggregation more apparent:
    - In text mode, stick a "* " at the front of subgroup header lines.
    - In GUI mode, display a horizontal rule between subgroups.
   For GUI mode, this is a huge improvement, at least under IE.

mode (identify the source class for class attrs; segregate attrs according
to source class, and whether class method, static method, property, plain
method, or data; display data attrs; display docstrings for data attrs
when possible).

Alas, this is mondo ugly, and I'm no HTML guy.  Part of the problem is
that pydoc's GUI mode has always been ugly under IE, largely because
<small> under IE renders docstrings unreadably small (while sometimes
non-docstring text is painfully large).  Another part is that these
segregated listings of attrs would *probably* look much better as bulleted
lists.  Alas, when I tried that, the bullets all ended up on lines by
themselves, before the method names; this is apparently because pydoc
(ab?)uses definition lists for format effects, and at least under IE
if a definition list is the first chunk of a list item, it gets rendered
on a line after the <li> bullet.

An HTML wizard would certainly be welcomed here.

This almost entirely replaces how pydoc pumps out class docs, but only
in text mode (like help(whatever) from a Python shell), not in GUI mode.

A class C's attrs are now grouped by the class in which they're defined,
attrs defined by C first, then inherited attrs grouped by alphabetic order
of the defining classes' names.

Within each of those groups, the attrs are subgrouped according to whether
they're plain methods, class methods, static methods, properties, or data.
Note that pydoc never dumped class data attrs before.  If a class data
attr is implemented via a data descriptor, the data docstring (if any)
is also displayed (e.g., file.softspace).

Within a subgroup, the attrs are listed alphabetically.

This is a friggin' mess, and there are bound to be glitches.  Please
beat on it and complain!  Here are three glitches:

1. __new__ gets classifed as 'data', for some reason.  This will
   have to get fixed in inspect.py, but since the latter is already
   looking for any clue that something is a method, pydoc will
   almost certainly not know what to do with it when its classification
   changes.

2. properties are special-cased to death.  Unlike any other kind of
   function or method, they don't have a __name__ attr, so none of
   pydoc's usual code can deal with them.  Worse, the getter and
   setter and del'er methods associated with a property don't appear
   to be discoverable from Python, so there's really nothing I can
   think of to do here beyond just listing their names.

   Note that a property can't be given a docstring, either (or at least
   I've been unable to sneak one in) -- perhaps the property()
   constructor could take an optional doc argument?

3. In a nested-scopes world, pydoc still doesn't know anything about
   nesting, so e.g. classes nested in functions are effectively invisible.

Create & populate the new Lib/test/data directory.

suite.  Note that other tests can put input data in this directory.

<http://sf.net/projects/mimelib>.  There /are/ API differences between
mimelib and email, but most of the implementations are shared (except
where cool Py2.2 stuff like generators are used).

Added signatures for some new PyType_*() functions.

point out, pydoc doesn't tell you where class attributes were defined,
gets several new 2.2 features wrong, and isn't aware of some new features
checked in on Thursday <wink>.  pydoc is hampered in part because
inspect.py has the same limitations.  Alas, I can't think of a way to
fix this within the current architecture of inspect/pydoc:  it's simply
not possible in 2.2 to figure out everything needed just from examining
the object you get back from class.attr.  You also need the class
context, and the method resolution order, and tests against various things
that simply didn't exist before.  OTOH, knowledge of how to do that is
getting quite complex, so doesn't belong in pydoc.

classify_class_attrs takes a different approach, analyzing all
the class attrs "at once", and returning the most interesting stuff for
each, all in one gulp.  pydoc needs to be reworked to use this for
classes (instead of the current "filter dir(class) umpteen times against
assorted predicates" approach).

restores the 2.1 ability of Tools/scripts/ndiff.py to start producing
output before the entire comparison is complete.

easy for 2.2 new-style classes, but trickier for classic classes, and
different approaches are needed "depending".  The function will allow
later code to treat all flavors of classes uniformly.

somewhere inside a line, use ndiff so that intraline difference marking
can point out what changed within a line.  I don't remember diff-style
abbreviations either (haven't used it since '94, except to produce
patches), so say the rest in English too.

cStringIO's can participate in the iterator protocol.

Fill the Itype.tp_iter slot with I_getiter()