Re: Hamza U+0621

["Mete Kural" <metek at touchtonecorp dot com>]

Hello Meor,

From: Meor Ridzuan Meor Yahaya <meor dot ridzuan at gmail dot com
>Tom,
>I'm not sure it's OpenType (the spec) fault, or it's implementation
>(uniscribe, pango) fault, or it might just be plain VOLT fault. I
>think  the problem lies between VOLT (the software the generate
>opentype table) to Uniscribe/OTLS  implementation. OpenType , the spec
>is not the problem. Anyway, I've to admit, last time I did not test it
>on Linux when the problems occurs, so not sure how it behaves. I only
>test it after it works under windows (which still have some problems
>under Linux).

Meor, I think the first thing I would suggest you to do would be to test the U+0621 with both the first case (al-aakhirah) and the second case (ya'aadamu) on Linux with Pango. If it works with Pango but it doesn't work with Uniscribe then we know that VOLT does its job correctly. If it doesn't work in Pango either then VOLT would be suspect.

>I might be able to eliminate all of the problem by creating the
>OpenType table (GSUB mostly) by hand and recompile it using TTX, but
>that will take a lot of my time (unless I can work full time on it, I
>might do it).  I still think that somehow VOLT produces a buggy GSUB
>table, but difficult to prove it.

Maybe trying to create the GSUB table in Visual TrueType might be another alternative..

Regards,
Mete

--
Mete Kural
Touchtone Corporation
714-755-2810
--