[tex-hyphen] tl2011-pretest: loadhyph-sr-cyrl.tex still not activated by default?

[Nikola Lečić]

Hi Arthur,

On Wed, 1 Jun 2011 15:02:26 +0100
Arthur Reutenauer <arthur.reutenauer at normalesup.org> wrote:
 
> 	Hi,
> 
> > I understood that conclusion was reached to load both patterns by
> > default in Unicode-aware engines. Anyway, IMHO loading both is the
> > most convenient for end users. These patterns are disjunct. I've
> > been using them in that way for many years.
> > 
> > (I think that Karl and Arthur made a decision on how to do this.)
> 
>   There has been no real decision at the time; Karl has only reacted
> early in the discussion you mentioned to say that having TeX Live's
> build scripts generate "serbian loadhyph-sr-cyrl.tex
> loadhyph-sr-latn.tex" in language.dat would be require a lot of
> changes in the infrastructure, and I replied that it would be no
> problem for Mojca and me to generate a loadhyph-sr.tex that contains
> both pattern sets (so that language.dat would only need to contain
> the line "serbian loadhyph-sr.tex").  But we haven't decided for any
> particular behaviour at the time.

I see. I apologise for misunderstanding.

> > I know. I'd slightly prefer to have Cyrillic as default; I just
> > don't know what kind of arguments apply here. If babel is the most
> > relevant factor for pdftex (and for XeTeX/LuaTeX it isn't), let's
> > decide according to the state of Serbian language in babel.
> 
>   In general, I would like to say that it's better to follow Babel
> (since there is no clear-cut consensus for either option); however
> Babel is completely broken in that particular case, as Mojca remarked
> in last year's discussion: back when we started the hyph-utf8 project
> in 2008, it loaded Cyrillic patterns while setting labels in Latin.
> This is why we switched the default patterns to Latin, in order to
> preserve consistency.  The problem was that we could change the
> behaviour of the hyphenation patterns since we were maintaining them,
> but not the settings in Babel (and talking to Johannes Braams was no
> use either; I tried).  Not a very good reason, but there was no
> better solution.

That's why I asked what arguments apply here. Thanks for the
clarification. Knowing this of course it makes sense to left it set to
Latin.

> > Yes, it's possible to use babel with Serbian Cyrillic, at least with
> > XeTeX. This is excerpt from one old file of mine (2007), don't know
> > it it's relevant nowadays:
> > 
> > \usepackage{fontspec}
> > \usepackage[serbian,greek,english]{babel}
> > \def\cyrillicencoding{EU1}
> > \let\greektext\relax
> > 
> > along with adding 'serbian xu-srhyphc.tex' to language.dat.
> 
>   Hum, that's good to know

That's an old recipe by Yves Codet and Jonathan Kew, from an era before
polyglossia.

> but not really relevant here, as
> Polyglossia does the job just as well.  I really think Mojca meant
> "Babel *and* pdfTeX".

Indeed. However, I'm afraid I'm not competent to explain anything
Babel+pdfTeX. I've never used that combination.

> > To sum up: I strongly suggest loading both patterns for
> > XeTeX/LuaTeX.
> 
>   I see no problem with that.  And I would advise to not change the
> default in pdfTeX, because that would make the life of the
> (presumably) few users who try to typeset Serbian with Babel pdfTeX
> even more difficult.  Mojca?

Loading both for (Xe|Lua)(La)TeX and leaving Latin for pdfTeX sounds
optimal to me.

Thanks,
-- 
Nikola Lečić = Никола Лечић
fingerprint : FEF3 66AF C90E EDC3 D878  7CDC 956D F4AB A377 1C9B
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 314 bytes
Desc: not available
URL: <http://tug.org/pipermail/tex-hyphen/attachments/20110601/5ddf6970/attachment.bin>