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

Re: tex-common, languages, and finding language.dat



Norbert Preining <preining@logic.at> wrote:

> Hi Frank, hi maintainers!
>
> I have a question concerning finding language.dat.
>
> I want to put a language.dat into TEXMFSYSVAR = /var/lib/texmf/
> (into tex/generic/config) instead of linking it from /etc/texmf
> into TEXMFMAIN.
>
> Now the strange thing is that with
> 	$ kpsewhich language.dat
> 	/var/lib/texmf/tex/generic/config/language.dat
> 	$ TEXMFVAR=/var/lib/texmf TEXMFCONFIG=/usr/share/texmf kpsewhich language.dat
> 	/usr/share/texmf/tex/generic/config/language.dat
>
> So fmtutil-sys cannot work as it takes the wrong language.dat.

If we have language.dat in TEXMFVAR, why not simply ship nothing
elsewhere?  teTeX's original is in /usr/share/tetex-base/ anyway.

> Thomas answered me on this question:
> ---------------- 8< ---------------------
> I think this can be explained if you look at your definition of the
> order of TEXMF trees given in the TEXMF variable.
>
> In teTeX's default setup, there is no overlap among the directories
> mentioned in $TEXMF, i.e. all of TEXMFDIST, TEXMFHOME, TEXMFMAIN,
> TEXMFCONFIG, TEXMFVAR, TEXMFSYSVAR and TEXMFSYSCONFIG are different
> directories.
>
> If you do something else, you should know what you are doing and also
> know that I don't recommend to do this...
> ---------------- 8< ------------------------
>
> Well this is now my question: (1) In fact I don't see why, and (2) can
> we fix this in tex-common/texmf.cnf?

As for (1), I don't see it immediately, either.

I think simply dropping /usr/share/texmf/tex/generic/config/language.dat
is the way to go.  However, we should really take time somewhen and
again collect the arguments for setting TEXMFSYSCONFIG to the same as
TEXMFMAIN.  It is the easiest way to do it, but in one of the last mails
I exchanged with people on that subject some new arguments against it
were raised.  I think it was in de.comp.text.tex, and it might be that
my discussion partner wasn't TE this time, but Markus Kohm.

Regards, Frank
-- 
Frank Küster
Inst. f. Biochemie der Univ. Zürich
Debian Developer



Reply to: