Re: pango_ot_ruleset_shape and GPOS

[Eric Mader <mader jtcsv com>]

At 08:21 AM 2/1/2002, Owen Taylor wrote:

> Eric Mader <mader jtcsv com> writes:
>
> > [...]
> >
> > I looked at pango_ot_ruleset_shape, which seems to be the logical
> > place to call TTO_GPOS_Apply_String, and it's not called there. Is
> > this just a case of "you're the first person who needs it, so you get
> > to write it"
>
> The patch in:
>
>  http://bugzilla.gnome.org/show_bug.cgi?id=69329
>
> Includes this. I haven't had a chance to review it yet, unfortunately.
>
> Regards,
>                                         Owen

I installed and tested this patch. The patches it makes to arabic-xft.c 
don't compile (it calls pango_ot_ruleset_shape_2 instead of 
pango_ot_ruleset_shape). Also, since I'm using an Arabic font with lots of 
ligatures, I hit the num_glyphs == num_char assertion. I was able to 
proceed by commenting out the assertion.

(I think this is actually a symptom of a more general need with OT 
processing; a way to relate the input character string to the resultant 
glyph string. I need something like this in my Indic engine to do post-GSUB 
matra fixups for Tamil and Malayalam... I'll have to think about this some 
more.)

After I adjusted my Indic code to set both GSUB and GPOS features in the 
ruleset, I got "Error loading GPOS table 4128" 
(TTO_Err_Invalid_GPOS_SubTable_Format) I'm going to start digging in the 
code to figure out what's happening here. Just reporting in, 'cause I don't 
know how long it's going to take to find this...

Eric Mader