Greetings,

I've been working with a slightly modified Hunspell 1.1.12 with the Ubuntu 7.10 
default
version of OpenOffice 2.3.  My adjustments to libhunspell are certainly trivial.

In the Hunspell::spell method I've added a line to print the "word" parameter 
to stderr.
What I've observed is that the printed "word" may be the expected UTF-8 
characters, but often
will be a sequence of '?' symbols, one per unicode symbol (NOT per UTF-8 byte). 
I've printed
the character addresses here to verify that these are literal 0x3f question 
marks and not
a terminal display mapping.  Naturally this causes dictionary lookups to fail.

What triggers OO.o Writer to pass the sequence of '?' to Hunspell in place of 
letter
symbols?  Is it a side effect of options settings? I'd like to keep this off 
but haven't
found a combination of locale and OO.o options to consistantly keep the letter 
from
being mapped into questions marks.

thanks!

-Daniel

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to