clean_key(): Override the standard LaTeX2HTML clean_key() to remove a
leading HTML tag, if present. This broke the indexes for the library
reference (at least) since many of the strings began with <code> or
something similar.
functions and constants together).
Make explicit datadesc sections for each of the constants which might appear,
and have a description of each. (Descriptions are based on the Linux
documentation and sources and the Solaris man pages.)
Hopefully Jeremy won't mind, because I didn't ask. ;-)
Use \file{} for file names.
Prefer \code{blat} and \emph{blat} to {\tt blat} and {\em blat}; this matches
current style in the Library Reference a bit better.
Made the example startup banner current. The version number should be
bumped before the next release.
Remove spurious underscore following book title. Added specific reference to
a Win32 networking book.
Changed \indexsubitem from (in module SocketServer) to (SocketServer protocol),
since it's talking about a protocol supported by a collection of classes.
Removed the large comment remaining from the template documentation
section; the template tells us to remove these once they're not needed.
Remove some trailing whitespace from VM instruction pseudo-code.
includes the string in the returned value instead of the dummy
filler character.
add_idx(): Override the latex2html function of the same name; this gets
the anchor properly embedded in the <dt> element, so the index
works in Grail too.
to use the opcodedesc environment.
Changed a \code{} to a \file{} near the start where a file is referenced.
Fixed a typo: "on" --> "one" in ROT_THREE description.
Wherever opcodes were referenced by name, use \code{OPCODE_NAME}; usage was
inconsistent. Ideally, another macro would be defined since these don't
represent code a programmer would type, but that's minor even for me. It'll
probably get converted automatically in the SGML conversion project. Whether
that matters or not isn't relevant at this point.
the libdis.tex file I'm about to check in.
I'm not sure this is really an optimal solution yet, but it may be the best
alternative. It avoids describing the instructions as either data items or
functions.
This change was discussed with Guido. (Guido: Take a look at the LaTeX
output for this; if this is reasonable I'll go ahead and update the perl
code in myformat.perl to match.)
messed up (but not checked in) version in my work files and didn't
notice before releasing 1.5 -- at least the distributed latex file
doesn't have the bogus warning about changed __init__ semantics.
Fred Drake
Guido van Rossum
Barry Warsaw