First, the RNG in whrandom.py sucks if you let it seed itself from the time.
The problem is the line:
			t = int((t&0xffffff) | (t>>24))
Since it ORs the two parts together, the resulting value has mostly
ON bits.  Change | to ^, and you don't lose any randomness.

with the other example of a structure description (struct _frozen).

Added index entries for the modules referenced in the document.

Use \emph{} when referring to title of any of the Python manuals, like is
done in the other manuals.

In some places, use \emph{} or \dfn{} instead of ``...''.

Use \samp{} instead of \code{} when marking identifier prefixes.

Use logical markup wherever it made sense.

Fixed a bunch of typos.

In several places, use "---" instead of "--" to get the emdash.

Start sentences with capital letters and end them with periods, as needed.

"it's" --> "its" in many places:  "it's" it *always* a contraction of "it is",
and "its" should always be used for the possessive.

"don't" --> "doesn't"

"should discards any" --> "should discard any"

In C function descriptions, use \var{} to mark parameters in the running
text instead of \code{}.  This matches usage in the other manauls better,
and is more consistent with the formatting of the function signatures.
Lots of little changes to implement this.

Lots of fixups needed due to appearant heavy cut-&-paste in the orignal
document.

Mention that the exception objects may be either classes or strings,
depending on the use of -X; they were unequivocally stated to be strings
in the section "Standard Exceptions".

"mkvalue()" --> "Py_BuildValue()"

Description of PyNumber_Power() indicates that the third value is option,
but not how to indicate that it was omitted.  Clarified.

Explain the behavior of PyString_FromStringAndSize() if the buffer is NULL.

Explain the Py_complex structure a little (tell what it's used for) and use
a {verbatim} environment for the structure definition itself.

Fix explanation of PyFile_SoftSpace().

Update the example version string to 1.5.

Combined the sections on defining new object types.

the mrsw lock and without 'condition' allocating a lock by default

(BTW, the Mac version was by Jack Jansen.)

place them elsewhere so they get the right capitalization.

"EOF" --> "\EOF{}"

removes extra vertical space from the list of names, and makes the display
more similar to that used in the socket module, where several constants share
a description.

\code{} at each embedded space, so that the words will wrap.  This keeps
it from being too ugly.

\code{} at each embedded space, so that the words will wrap.  This keeps it
from running off the side of the page, and is only slightly weird.

Explain what happens when a negative shift count is used (what exception).

Mark the title "Python Reference Manual" as \emph{}, for consistency.

"info" --> "information"

Tell more about the data attributes of file objects, using the {datadesc}
environment.

When refering the user to the language reference for information about
internal types, tell what internal types to expect information on.

This is an option for OS-es with case-insensitive but case-preserving
filesystems.  It is currently supported for Win32 and MacOS.  To
enable it, #define CHECK_IMPORT_CASE in your platform specific
config.h.  It is enabled by default on those systems where it is
supported.  On Win32, it can be disabled at runtime by setting the
environment variable PYTHONCASEOK (to any value).

When enabled, the feature checks that the case of the requested module
name matches that of the filename found in the filesystem, and raises
a NameError exception when they don't match.

&do_env_funcdescni:  New functions.  These support the non-indexing variety
	of the {datadesc} and {funcdesc} environments.

There's still some flakiness with the new indexsubitem support, but that's
low priority.

to indicate the interpreted letters from the format string.

in the running text.

For computed attribute and method names (where there's a \var{} part to
the name), use the non-indexing forms of \datadesc{} and \funcdesc{}.
This doesn't change the printed output, but removes 3 rejections from the
makeindex run and allows the LaTeX2HTML support to exclude these from the
index.

in the running text.

For computed method names (where there's a \var{} part to the name), use
the non-indexing form of \funcdesc{}.  This doesn't change the printed
output, but removes 3 rejections from the makeindex run and allows the
LaTeX2HTML support to exclude these from the index.

datadesc and funcdesc, but don't add entries to the index.

&do_cmd_setindexsubitem:  New function.  Set the indexsubitem value from
	\setindexsubitem{(...)}.

&do_env_opcodedesc:  By default, don't index byte codes.

$INDEX_OPCODES:  New flag.  If true, index the byte codes.  Default is off.

Normalize indentation to 4 spaces everywhere.

Minor nits.

Make all the indentations in {verbatim} environments have column 0 of the
listing in column 0 of the file.

Remove pagenumbering / pagestyle cruft.

Use more logical and less physical markup.

Make all the indentations in {verbatim} environments have column 0 of the
listing in column 0 of the file.

Remove pagenumbering / pagestyle cruft.

checkin of myformat.sty.

Change "\renewcommand{\indexsubitem}{(...)}" to "\setindexsubitem{(...)}"
everywhere.

Some other minor nits that I happened to come across.

Handle most (all?) of the page style / numbering magic here so the documents
don't have to do it individually.

Revise the \bcode / \ecode stuff so that the {verbatim} environment handles
it right directly.  \bcode / \ecode will be completely removed from all files
(to be checked in momentarily).

Have the {verbatim} environment get the samples indented a bit; this
appearantly had been attempted in the old code, but didn't work because
paragraphs weren't indented.

Make all headers, from chapters on down to subparagraphs, have sans-serif
titles.

\setindexsubitem{}:  New macro.  Replaces \renewcommand{\indexsubitem{}(...)}
	everywhere.  This allows LaTeX2HTML to be made to work correctly for
	this.  That was near impossible with the old mechanisms.

For all {*desc} environments, make the name of the described thing bold as
well as monospaced.

{opcodedesc} environment:  Don't index the byte code names; that doesn't seem
	terribly useful, and there are a lot of them.

\var{}:  More magic to make sure that the size is right even if embedded in
	\file{} or some other macro that uses the sans-serif font in running
	text.

\bfcode{}:  New macro.  Makes the font \code{} and bold.  (Was unreasonable
	using old LaTeX 2.09.)

\file{}:  Adjust the size of the sans-serif font a little.

\email{}, \url{},  Make these use the same font as \file{}, but not the
	surrounding single-quotes.

Update many comments.

Lots of minor nits and a little cleanliness.

main():  Include a comma between the module name and page number, to make
	 the generated index more like the "standard" generated index.