Re: language.dat: Should we just enable all the patterns?
Atsuhito Kohda <kohda@pm.tokushima-u.ac.jp> wrote:
> I see. I'm neutral at the moment but would like
> to comment a bit from my quick look.
>
> - it seemed english should always be defined but
> it could be an alias for american or british.
> I don't know how they differ but IMHO it could be
> a problem.
They do differ AFAIK, but also now it's the users choice whether he
wants to change the alias. Currently (from upstream tetex) english is
american if UKenglish is not enabled, but gets UKenglish as soon as this
is enabled. No problem to have both, as long as the alias is defined.
> - there are some languages which provide multiple
> hyphenation patterns like norsk, ukrainian.
> Especially, ukrainian seems to use only one of
> candidates. I might be wrong but with all candidates
> enabled "texconfig init" showed an error message
> `tex -ini -jobname=latex -progname=latex latex.ini' possibly failed.
> though latex.fmt seemed generated. ^^^^^^^^
Yes, same with norsk. We have to make a decision (or a good guess), and
users who disagree would have to edit language.dat again, as in before
debconf took over. But I think for someone who cares about the details
of hyphenation, this shouldn't be really hard. (And we would of course
document it in README.Debian).
> - jlatex (of jtex-bin) caused error with dutch and
> magyar at first trial, and further, after commenting
> out these in language.dat, it failed to generate jlatex.fmt
> with an error;
>
> (/usr/share/texmf/tex/generic/hyphen/nohyphb.tex Norsk hyphenation patterns
> ! TeX capacity exceeded, sorry [pattern memory ops=1501].
> l.1422 .u2r2i
>
> If you really absolutely need more capacity,
> you can ask a wizard to enlarge me.
TeX capacity is the only thing I am worried about. And Walter Schmidt
said he really doesn't know about this, but has not encountered
problems.
Strangely, I cannot reproduce this - I get a different error. The first
comes when dutch hyphenation patterns are processed, so I commented them
out. Then a very similar error comes with the hungarian ones:
\l@magyar=\language14
(/usr/share/texmf/tex/generic/hyphen/huhyph.tex Hungarian hyphenation patterns
! Improper alphabetic or japanese constant.
<to be read again>
\á
l.26 catcode`\á=11\uccode`\á
=`\^^c1\lccode`\^^e1=`\^^e1^@
A one-character control sequence belongs after a ` mark.
So I'm essentially inserting \0 here.
(with dutch it's \ä).
Anyway, I will try to find somebody who can comment on the memory and
registers needed.
>> Unfortunately, I wasn't able
>> to read the documentation - xdvi segfaults after some font errors, and
>> dvips doesn't generate output. Brr.
>
> Ah, this is completely different issue. You need xdvik-ja
> and dvipsk-ja to read the documentation but if you installed
> these it would be difficult for you to find out problems
> of tetex-bin (i.e. xdvi or dvips) because xdvi and dvips
> would call binaries of xdvik-ja and dvipsk-ja but not of
> tetex-bin.
Hm, yes. Of course I didn't want to have them installed permanently, but
I tried to. In fact I had some binaries lying around from a tetex-3.0
test install in /usr/local/bin, and this influenced it. Now with only
empty directories under /usr/local and xdvik-ja and dvips-ja installed,
I get a ps file - but this contains only strange multibyte-looking
stuff, not japanese or whatever fonts (and I'd like to have _some_
documentation for a package available in english, anyway).
xdvi-ja only blinks and says:
Unexpected error in "read_ZEIT_char()"
But you really needn't bother about that. I have reported this as bug
#235626, and it's not urgent, I think.
Regards, Frank
--
Frank Küster, Biozentrum der Univ. Basel
Abt. Biophysikalische Chemie
Reply to: