Mikhael Goikhman wrote: > Ok, I did some changes. Including removing alloca (not that > this helps with speed and core dumps).
Thanks Olivier/Mikhael for all your help (it would have taken me months to debug that myself). I just fixed the combinational character problem (that was noted by myself in the code) and reformatted the code a bit for better readability and for it to be no more than 80 characters wide (so I'm including the entire file so that you can simply substitute what's in CVS with what's noted (FBidiJoin.c) - I've been running with it for awhile now). I'm also including the arabic .po files along with the minor patch to include 'ar' as a "lang". Arabic (& derivatives) are now 99% complete (I will mail later about the remaining 1% with regards to harakat/tanween (eg. U+064E) characters). BTW: Olivier, did those UCS4<->UTF8 functions look ok ? Much thanks. - Nadim (CC for faster replies) __________________________________________________ Do you Yahoo!? Yahoo! Mail Plus - Powerful. Affordable. Sign up now. http://mailplus.yahoo.com
arabic_fixes.tar.gz
Description: arabic_fixes.tar.gz
