5

I have a perfectly working setup for hunspell with emacs (flyspell) using 

("german”  "[[:alpha:]]" "[^[:alpha:]]" "[']" t ("-d" "de_DE_frami" ) nil utf-8)

as dictionary definition for ispell-dictionary-alist.

German Umlaute are no problem as long as the main dict “de_DE_frami” contains them. However, as soon as I put a word with german Umlaute like “Bestürzenste” in the personal dictionary of hunspell, it is not recognized as being correct whereas all other added words that contain no german Umlaut are well recognized as being correct.

On top of this, if I check the personal dictionary of hunspell all words appear there nicely in proper utf8 encoding. What’s even more is that hunspell, if executed on the command line with the only option "-d de_DE_frami" does recognize words with Umlaute and e.g. “Bestürzenste” is reported as being correct.

So I guess the trouble is caused by ispell or flyspell. Might this be a bug concerning only lookups in the personal dictionary. Did anybody have the some problem? Is there a workaround that I wasn’t able to find aber hours of searching the net?

Thanks for helping out here!

Btw. I also tried with the encoding iso-8859-1 instead of ut8 in the dict definition above --> not change.

emelin
  • 101
  • 5

1 Answers1

2

I had a similiar problem using hunspell on the commandline to generate a list of misspelled words in a utf-8 encoded text file: words with Umlaute in the main dict de_DE were recognized as correct, but not those in my personal dict.

The solution to the problem was the observation, that the main dict was latin1 encoded, the personal dict utf-8 encoded; so I recoded my personal dict to latin1 and everything was fine.

Hope, this helps.

Klaus Horn
  • 21
  • 2
  • Thank you for this comment. Finally, I had to patch hunspell to make it work with utf-8 because of this (more or less) well known bug: http://debbugs.gnu.org/cgi/bugreport.cgi?bug=7781#31 – emelin Jun 29 '14 at 09:56