Dear Nikola and others, With respect to your request something else just came to my mind. We currently have two sets of patterns: - Serbo-Croatian in Cyrillic script (last modified by Dejan Muhamedagić in 2008) - Serbian in Cyrillic script (last modified by Aleksandar Jelenak in 2003; based on patterns by Dejan)
Serbo-Croatian patterns (older) are supposed to cover Serbian pretty well (at least that is what Dejan says). I'm not sure what the reason for further modifications by Aleksandar was and to what extent he modified the patterns and Dejan didn't do much checking either. (I hope that Aleksandar didn't modify them manually over existing patgen-generated patterns.) So now we have: - hyph-sh-latn.tex -> serbian - hyph-sh-cyrl.tex (equal as sh-latn, but not used anywhere) - hyph-sr-cyrl.tex -> serbianc I have now merged serbian and serbianc for XeTeX and I will soon commit a version that merges them for LuaTeX. However now a semi-serious problem arises. If you try to typeset some text in both scripts and if we use hyph-sh-latn + hyph-sr-cyrl then you will end up in different hyphenation. This might have been another reason why the serbian hyphenation patterns were disabled. I simply have no idea which ones should be included and why one set of patterns would be better than the other. I would prefer if we would either create hyph-sr-latn (transliteration from hyph-sr-cyrl) or include hyph-sh-latn+hyph-sh-cyrl. (We can prepare a longer document with hyphenation points colored, so that one could compare differences between hyph-sr-cyrl and hyph-sh-cyrl, but I'm not qualified to judge about differences and which ones are right or wrong.) Mojca 2011/6/1 Nikola Lečić wrote: > Hello, > > Regarding these old posts: > > http://tug.org/pipermail/tex-hyphen/2010-July/000681.html > http://tug.org/pipermail/tex-hyphen/2010-July/000682.html > http://tug.org/pipermail/tex-hyphen/2010-July/000683.html > etc. > > After a fresh tl2011-pretest install, Serbian Cyrillic patterns are > still not loaded by default and the manual intervention similar to what > is described in the first of those mails is still necessary. > > Can someone take a look before tl2011 release? > > Thanks, > -- > Nikola Lečić = Никола Лечић > fingerprint : FEF3 66AF C90E EDC3 D878 7CDC 956D F4AB A377 1C9B > ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ >
