Julian Gilbey <[EMAIL PROTECTED]> wrote:
> Sorry for dropping into the discussion so late. Is it reasonable to
> treat any format which has an invocation line beginning with "&..." in
> a similar way? Would that help to solve the problem more generally?
Yes, yes, very well. If you do the coding ;-). It's not a trivial
change, though, because up to 1.7, update-fontlang just cat'ed the
files, now it at least records that it encountered
10texlive-latex-base.cnf and checks whether it is working on
40{jadetex,xmltex}.cnf. Looking at file *contents* instead would be
much harder. And in order to solve the problem generally, we would even
have to look forward, since order doesn't matter in fmtutil.cnf: A
02local.cnf which loads "&latex" is fine only if latex is defined
later.
After considering this (and whether this would be faster in bash or in a
Perl rewrite) we decided to just special-case the jadetex and xmltex
packages...
Regards, Frank
--
Frank Küster
Single Molecule Spectroscopy, Protein Folding @ Inst. f. Biochemie, Univ. Zürich
Debian Developer (teTeX/TeXLive)
