On 26/07/2010 14:43, Arthur Reutenauer wrote:
The main problem is with Babel. If we do that then Babel will load T2A
and T1 patterns at the same time and will most probably produce
disastrous results.
Right, I hadn't thought of that. Then that's probably no solution as
far as Babel is concerned; and in Polyglossia there already is a
solution.
else we need to split language.dat (one copy for babel
and one for Polyglossia).
Obviously, that's out of the question.
Indeed. It would be technically possible to have a hyphen.cfg specifically tailored for
xelatex, and thus use another "language.dat", but this is not a road we should follow.
At the moment one has option [script=Latin] and [script=Cyrillic] in
Polyglossia, maybe one could also have an option [script=All] or
script=Both or something similar
That might be useful in case one needs to use both scripts in the same
document, but that's a different issue, and that still doesn't solve the
problem we have in Babel.
Also keep in mind that Script=All or Both makes no sense, as the option determines not
only which hyphenation patterns to load, but also the definitions of captions and date
formatting...
I'd prefer to have:
serbian loadhyph-sr-cyrl.tex
=serbiancyr
serbianlat loadhyph-sr-latn.tex
Wdyt?
