On Wed, 17 Mar 2004, Shachar Shemesh wrote:Yes, well. I did not think my suggestion consisted of a dirty interface. I even managed to convince Dov of that during a lift I gave him after the Go-Linux event last year. Maybe I shouldn't have changed those flights and gone through Toronto after all - a face to face helps in convincing of that :-P
Of course, my entire contribution to fribidi thus far has been a single
bug report, and suggesting a development direction for the "new
interface" that nobody seems to like. I am not a copyright holder, and
thus not one who's opinion binds anyone.
I really like your suggestion about doing bidi on UTF-anything, and constantly think about it. I have not yet found a good solution that satisfies all my requirements:
* No source/object code duplication.
* No significant overhead for UTF-32.
* No dirty interface.
In any case, I started working on that, and then realized that it's far too much work unless we seperate the character types and the actual resolving. Once that is done, I'm fairly confident I can convince you how to do that satisfying all of the above requirements.
Shachar
-- Shachar Shemesh Lingnu Open Systems Consulting http://www.lingnu.com/
_______________________________________________ Developer mailing list [EMAIL PROTECTED] http://lists.arabeyes.org/mailman/listinfo/developer
