I noticed that some translations files are encoded in utf-8 and some are encoded in iso or windows charsets.
As a result, search phrases and keywords are not correctly displayed even with utfdecode plugin.
This was a problem for me because I have both accented french and cyrillic russian requests.
Why not using utf-8 for every languages?
I did the conversion (using distribution 6.7) to utf-8 for fr,en,ru,bg,de & ua and that works great.
Would you be interested to include these files in the next release?