sites.
do_cmd_ignorePlatformAnnotation(),
do_cmd_platform(),
do_cmd_platformof(): New functions to support platform dependency
information.
process_all_platformofs(): New function to post-process \platformof macros
using information collected during the initial pass.
process_python_state(): New function. Call all post-processing functions
defined in this file to avoid having to have too much knowledge of
the internals for this stuff in l2hinit.perl.
out a little better.
Make sure the boxes for the last column of tables always get drawn;
these are often left empty of content when used as a "Notes" column.
They look a little better this way.
the subitem text is updated correctly when each module section is
processed.
Print [<modulename>] when we start on \declaremodule so we can more
easily track progress during the library reference.
data file used by buildindex.py.
next_argument_id(): Removed; no longer used.
next_argument(): Doesn't use next_argument_id(), just does the work.
gen_target(), gen_target_name(): Removed, no longer used. The only
place that used them now does the (trivial) work directly.
Lots of little changes so we never have to pass around $br_id values
all over the place. The only places they're used now new IDs are
generated as needed.
still don't support things like \let^^M=\something, where ^^M could actually be
any active character. Print a decent warning if we find one we can't handle.
synonyms for macros. This supports:
\let\newname = \oldname
\let\newname\oldname
This is useful sometimes, and can at least avoid real errors at other times.
value.
ArabictoRoman(): Acquired function from Robert Donohue
<donahue@cfassp48.harvard.edu>; might be useful someday, and I don't
want to figure it out again!
left to the buildindex.py script.
Change the field separator used in the index.dat file so utilities like grep
are useful on these files; they can't typically handle NUL bytes.
Protect index entries for environment variables from being handled as
case-insensitive by buildindex.py.
special markup for the first column of a table using {tableii} or {tableiii}.
In LaTeX, this requires using "textrm" as the formatting function, and this
is needed to remove the bogus "<textrm>" that gets generated otherwise.
There should probably be a more general mapping of specified markup to useful
markup, but that doesn't appear needed at the moment.
do_env_tableii(), do_env_tableiii(): If the markup for the first column is
\textrm, make it empty.
do_cmd_lineii(), do_cmd_lineiii(): If the markup for the first column is
empty, don't add markup around the text.
Implemented some new features of python.sty for methods & members.
Fixed some bugs.
Cleaned up some perl code and made some basic regular expressions more robust.
