On Fri, 7 Sep 2001, Markus Kuhn wrote:
> I assume the only purpose of any "visual order" produced by fibidi is to
> make text readable under terminal emulators without bidi support,
> therefore it will obviously have to follow for combining characters the
> normal Unicode postscript convention (combining character follows base
> character). So what you describe sounds like a bug in fribidi to me.
That is the point of L3 part of the Unicode Bidi Algorithm, and is really
application-dependent. Both of the methods (NSMs after base; NSMs before
base) have their pros and cons. This should be an option in invoking the
reordering method. I have discussed this with Behdad, current maintainer
of fribidi, and he told me that he is still thinking about the API for it.
If anyone has some idea, please contact the fribidi mailing list at:
http://lists.sourceforge.net/lists/listinfo/fribidi-discuss
or contact him directly at the address CC-ed in this mail.
roozbeh
-
Linux-UTF8: i18n of Linux on all levels
Archive: http://mail.nl.linux.org/linux-utf8/
