[tex-hyphen] luatex-hyphen & TL

Manuel Pégourié-Gonnard mpg at elzevir.fr
Sun May 23 23:35:37 CEST 2010

Le 23/05/2010 20:53, Mojca Miklavec a écrit :
> My idea in your nomenclature:
> - group 1: for pdftex and related (language.dat)
> - group 2: for XeTeX (language.dat in different xetex-specific tree)
> - group 3: for LuaTeX (using language.dat.lua)
> - The set of languages is exactly equal for group 2 and 3, it's only
> that they are loaded in a different way.  (Though there might be a
> slight catch since luatex should be able to use "more advanced
> patterns", so it may happen that the files won't be equal any more in
> future; but that doesn't depend on me; an example is Hungarian that
> uses different patterns for OpenOffice and TeX, generated by the same
> author)
> - Group 1 is not a subgroup of 2/3. The majority of languages is the
> same, but some are only usable in group 2/3 and some are only usable
> in group 1.
> - Both 8-bit and XeTeX patterns would still be loaded using the same
> loadhyph-xx.tex.
Currently, the symetric difference between group 1 and group 2, unless I missed
something, would only be that ibycus and mongolian-mlc are in G1 minus G2, and
G2 minus G1 would be empty. I don't see much to be gained by disabling those two
language in XeTeX based format (a few tens on seconds at format generation
time). So, IMO, it's not worth it in the current state of things. (Of course it
may change later, and we may want to change the system when it is needed.)


More information about the tex-hyphen mailing list