The numbers from 0-9 and different symbols like %,@,# etc are nowhere in the dictionary, but they are still shown correct whenever we write that. How does this happen?
Word breaking filter of the command line Hunspell checks only words (letter sequences) for spell checking. Spell checking numbers and other characters you need to add or extend WORDCHARS option in the dictionary affix file.
I think my problem is related to this one and I was wondering if nemethl would be able to help me here. I'm writing a mozilla add-on for a spell check of a minority language. I have my wordlist ready. Everything works, except one character that is a regular letter in Sinama (sml) language of the Philippines is not spell checked by Hunspell. Therefore when I type hˈllaku, the utf-8 character U02C8 is not checked and instead h is checked and llaku is checked. llaku then suggests the correct word in spell checker, but the problem repeats itself since the word was actually repeated.
I fear that I will have to create a locale for the sinama language, but I don't have the technical know how and there is little published work verifying the orthography that the linguists are currently using. Is there any quick fix that can be added to the .aff file. I tried putting U02C8 in the WORDCHARS title, but it didn't change the problem :-(
Adding a char to wordchars should work. Only problem left is the issue whether the calling application knows about this char.
You had better test from the command line, using input text file contaning the test words.