Re: language.dat: Should we just enable all the patterns?
Hilmar Preusse <hille42@web.de> schrieb:
> On 01.03.04 Frank Küster (frank@kuesterei.ch) wrote:
>
> Hai,
>
>> What about kicking out all my nice code for handling language.dat,
>> and all debconf related language.dat stuff, and simply enabling
>> _all_ patterns we ship?
>>
> Just from the changelog of latest beta release:
>
> Changes in the texmf area:
> <snip>
> * enable all available languages in language.dat
>
> So, what now?
I am not totally without responsibility for this...
I would suggest to follow upstream here, and do the same in our
tetex-2.0.2 packages. Walter Schmidt has convinced me that this will not
cause big problems: The only parameters affected during normal operation
are the "number of strings" and "multiletter control sequences", and
they are
- not critical under normal circumstances
- only affect relatively minor, that is: If this change causes an
overflow, the next added package that creates control sequences will
do the same.
During format generation, there may be overflows in pattern_space and
some other registers. For tetex itself, there's no problem. For other
packages depending on us, problems can occur. However, I think it might
actually be a good thing to trigger this by a change in the package:
This way, the maintainer of the respective package will notice that
there is a problem and can fix it early. Otherwise, the bug reports will
only drop in occasionally once a user enables too many patterns. (And
then it's perhaps tetex-*s postinst that fails, if that is upgraded
first).
One example is the bug I submitted against jtex-bin, #257825. With the
fix I suggested, and all patterns enabled in
/etc/texmf/jtex/language.dat, except dutch and magyar, it indeed fails
with norwegian and a TeX capacity exceeded error. But probably with
jlatex one wouldn't want to use so many language patterns. And I'm sure
it can be fixed, I'm only to dumb to see it easily (also because jtex is
based on tetex-1.0.7). Anyway, the easy fix is to activate only a small
subset in /etc/texmf/jtex/language.dat.
Regards, Frank
--
Frank Küster, Biozentrum der Univ. Basel
Abt. Biophysikalische Chemie
Reply to: