Lots of changes to get the paragraph marking to work, and not go into
an infinite recursion.
Start to rationalize markup of method signatures; not complete.
find_all_elements(): similar to getElementsByTagName(), but operates
on the not-quite-legal fragments we deal with here.
Simplifies several individual transforms.
fixup_sectionauthors(): Convert \sectionauthor items to <author> just
after the <section>'s <title>.
general identifiers for which closing tags will be omitted
when SGML is generated. This can be used to tell the markup
generator to drop stuff like </para>. Note that it needs to
be possible for the closing tag to *always* be omitted for it
to be included in "autoclose".
main(): Added command-line option "-a" / "--autoclose" to set the
list of general identifiers passed to the convert() function
as the "autoclose" parameter. The list may only be specified
once (not additive) and GIs should be comma-separated. The
default list includes only "para".
any subdirectory; don't continue with remaining subdirs.
Added "api", "ext", "lib", "mac", "ref", and "tut" targets to only do
submakes in those directories. This is just a lot easier to use than
to cd into the subdir and use make.rules directly.
currently generated by the LaTeX and LaTeX2HTML processes is generated
here as well, making it more flexible in the SGML version.
Reduce the <args> element so that <optional> goes away; just use
square brackets to indicate what's optional. This makes it easier to
read than the LaTeX, and the processor can do any checking it needs to
in order to make sure it's legit. Possible shortcoming: DSSSL
processors may need more explicit markup. Can probably hack around it
for this case, but we'll see.
used to deal with the table headings and entries.
An additional flag in the element table is used to indicate elements
which have no "general" content, but which do have subelement
content. These must be flagged distinctly from empty elements.
Currently used for \lineii, \lineiii, and \lineiv.
<moduleinfo> element based on various meta information, and
strip some cruftiness. This is more usable for information
extraction, and organizes the information more clearly.
cleanup_synopses(): Rewrite to use create_module_info(), so this
will work with multi-rooted "documents".