Hi, I've done some experimentation with speechd-up and espeak, and found the UTF-8 reading to work when hearing characters automatically. I used the example of various accents on the letter a, typed into Vim. Both A acute and A grave are read properly by espeak, but when reviewed with speakup's review keys they appear as Greek letters. I suspect this is a consequence of the way character pronounciations are handled currently. /sys/modules/speakup/parameters/characters isn't UTF-8 aware. I tried catting /dev/null to it, but that just reset it to defaults, as did /dev/zero. Realistically speaking I don't need UTF-8 characters myself, but it is still a little annoying to hear them misidentified. Hope this helps somewhat, Zack.
