Message-ID:  <199910312013.VAA00827@istrati.zdv.uni-mainz.de>
Reply-To: Mailing list for the LaTeX3 project
              <LATEX-L@URZ.UNI-HEIDELBERG.DE>
Date:         Sun, 31 Oct 1999 21:13:27 +0100
From: Frank Mittelbach <frank.mittelbach@LATEX-PROJECT.ORG>
Sender: Mailing list for the LaTeX3 project
              <LATEX-L@URZ.UNI-HEIDELBERG.DE>
To: Multiple recipients of list LATEX-L
              <LATEX-L@URZ.UNI-HEIDELBERG.DE>
Subject:      extending xparse --- any takers?
Status: R

we had a bit of discussion some time ago on how to extend xparse

here is one suggestion which i think might be of interest:

suppose we have a parsing class "x" in the second argument of
\DeclareUserCommand which would take a list of tokens (for example comma
spearated) and parses the input to see if any of those tokens follows. If so
it would return \BooleanTrue otherwise \BooleanFalse. eg

\DeclareUserCommand \footnotemark { o x{\footnotemark,\footnote} }

would scan for an optional argument and returns it as #1 (or \NoValue if not
found) and then scans for either \footnotemark or \footnote and returns
\BooleanTrue or False respectively

if this is a sensible extension (not sure) one has to be clear about a couple
of points:

 a) is the parsing happening only on top-level of the source document or does
 one parse using expansion? That is, if the user says
 \newcommand\fn[1]{\footnote{#1}} would \footnotemark\fn{foo} see the hidden
 \footnote or not? --- i think it probably should find it

 b) what do we do about spaces: is the first  in \footnote{foo} \footnote{bar}
 seeing the second one despite the fact that there is a space inbetween and if
 so what happens to this space? --- again i think it should probably see the
 second footnote (or perhaps there should be a variation class "X" that can
 look past such spaces) but it would need to communicate that such a space is
 there (somehow)

any suggestions/comments on something like this? --- it certainly would need
some further thought to become useful (if it is useful at all)

anybody out there who needs a nightcap puzzle?  would be interesting to
experiment with such an implementation.

frank