#4 Tsearch2 ispell parser unicode bug

[Anonymous]

I tried postgresql-8.0beta and rc (up to rc3) and all
versions contains bug in tsearch2. (probably in ispell
code part)
I can set up tsearch without problems for unicode
fulltext (for czech language), but when i tried the
test command like this (and some similars).

SELECT lexize('cz_ispell','jablkům');

postgresql fails with SIG11.

I tried to backport the ispell part and now it works
without problems.

(my backport patch is included)

[Nobody/Anonymous]