The czech ispell hash we provide for AbiWord is not working, it contains only chartypes nrof and tex but not latin2. Consequently, when the hash is loaded the spellchecher's iconv is set to the default latin1, and any words containing non-latin1 characters are incorrectly processed. As I have no experience with building ispell hashes, I have no idea how to fix this, but I supsect that the same problem might apply to all of our non-latin1 hashes.
Tomas
