[XeTeX] Sanskrit hyphenation
jonathan_kew at sil.org
Thu Jul 29 12:57:30 CEST 2004
On 29 Jul 2004, at 11:40 am, Yves Codet wrote:
> I tried and wrote a hyphenation file for Sanskrit after Jonathan's
> Hindi hyphenation file. It looks good but it doesn't do anything. I
> followed Jonathan's instructions though and log files seem to show
> that patterns were loaded. If somebody could check the attached
> archive and find what went wrong I would be very thankful. The input
> text file is "mbh.tex".
It looks to me as though your input file is lacking a command to tell
TeX that the current language is Sanskrit, so the default (English)
patterns are active. Try adding \language=\sanskrit at the start of
In addition, when I actually try this, no hyphenations occur in this
sample; but if you use a smaller \hsize than the default, making it
harder to find good line-breaks, then they do start to appear.
> There's something else I don't understand. When I compiled a format,
> some letters were declared non letters. I had to comment them out.
> They can be seen in "sanhyph.tex".
Ah, looks like those are the vocalic rr and ll letters and vowel signs;
checking the Unicode standard, I see that they're at U+0960..0963. So
\loop \lccode\n=\n \ifnum\n<"095F \advance\n by 1 \repeat
in the patterns file doesn't set \lccode for these, because it stops at
095F. Just change the ending value for that loop to 0963, and they'll
be included as acceptable letters for hyphenation patterns.
More information about the XeTeX