[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: tex-common: Please provide dh_installtex functions for packages that need only mktexlsr



Norbert Preining <preining@logic.at> wrote:

> On Die, 24 Okt 2006, Florent Rougon wrote:
>> I fear that, in addition to the problems caused by the partial mktexlsr
>> calls that I pointed out in my last mail (partial in that not all TEXMF
>> trees are scanned and updated), update-updmap and its friends won't be
>> called if nothing is found in the TEXMF trees. Hmmm? Because you moved
>> these calls to post{inst,rm}-texlsr and now, you're saying these
>> snippets are only included "IF we have some tree present".
>> 
>> (maybe in every real situation where update-updmap and friends are
>> needed, there is something in the TEXMF trees, but still...)
>
> Hmm, you are right, at least partly. But it is true that in most cases
> (in fact I cannot imagine the case where it is not happening) IF we have
> to call one of the update-* scripts, then also something HAS to be
> installed into /usr/share/texmf.

If I write a program (and package it) that generates code which

- needs much larger settings of TeX's memory variables

- accesses new fonts outside the TEXMF trees with clever map lines,
  either in the document (new pdftex) or in updmap's output files

it might be necessary to add files in texmf.d or updmap.d, but not in
TEXMFMAIN. 

> Another option would be to call something IF a file was installed into
> /var/lib/tex-common/* where the list files are stored.
>
> But now, what do all of you think? Should we follow this path or not? I
> don't mind throwing out the whole mktexlsr-magic code, and leave only
> the separation between mktexlsr and the rest-magic in there, so that an
> empty dh_installtex just calls 
> 	mktexlsr
> in the right way.
>
> Frank, it was your bug report, what is your opinion?

I think that it might be a good approach not to put too much magic into
the script, but leave more to the maintainer: They have a better chance
of knowing what is going on.  This would mean to just separate out the
mktexlsr stuff, let it always scan TEMXFSYSCONFIG, TEXMFSYSVAR and
TEXMFMAIN, plus TEXMFDIST (and TEXMFLOCAL?) upon request (when a basic
TeX package is installed or updated), and do the rest only if requested
by a commandline option, without additional magic (but of course doing
the right thing wrt order and mktexlsr calling).

Regards, Frank
-- 
Dr. Frank Küster
Single Molecule Spectroscopy, Protein Folding @ Inst. f. Biochemie, Univ. Zürich
Debian Developer (teTeX/TeXLive)



Reply to: