Hi,
> The bigger problem is that hyph-sr-latn and hyph-sr-cyrl differ (and
> both make errors). Take a look at this concrete example.
Thanks for the examples. I'd like to point out that while most of the
incorrect ones are indeed flat out wrong, some of them are border cases:
for example
> --------------------------------------
> српско incorrect
> srp-sko correct
> --------------------------------------
is only incorrect in that it doesn't find all the breakpoints in the
Cyrillic version; but it doesn't insert wrong ones (all the ones it
finds are right -- OK, it's 0 in that case ;-)
This one:
> --------------------------------------
> на-сто-ја-тељ correct
> na-sto-ja-te-lj incorrect
> --------------------------------------
may be a different issue: it might be that the Latin patterns have been
generated from patgen with a value of \righthyphenmin greater than 2
(strictly); in that case setting \righthyphenmin to 1 or 2 would simply
produce wrong results.
Other than that, it strikes me that many of the incorrect breakpoints
involve the liquid consonants 'l' and 'r'.
Arthur
