Re: Font Problems

[Eric Mader <mader jtcsv com>]

At 03:37 PM 8/23/2002, Keith Packard wrote:

> Around 17 o'clock on Aug 23, Owen Taylor wrote:
>
> > However, it doesn't do us any good for Xft, because fontconfig
> > pre-prunes the list of candidate fonts based on the coverage
> > information they report.
>
> Hmm.  We need to move fonts with GSUB support ahead of fonts without GSUB,
> but only for arabic shaping.  One possibility is to have a property
> indicating GSUB support; apps supporting shaping would request GSUB and
> prefer fonts with GSUB bits.

We probably need to check for GPOS too; I *think* it makes sense for a font 
to have GPOS but not GSUB. There are other OT layout tables too, but I 
don't think they make any sense w/o GSUB or GPOS present...

Also, my Indic shaper uses OpenType too, so it's not just Arabic shaping 
that matters. In principle, there could be OT-based shapers for any script. 
(MS has already defined specs for a whole bunch of languages...)

There could be, in principle at least, shapers for AAT / QuickDraw GX fonts 
which also contain TrueType tables to describe the layout process. (Not to 
mention SIL's Graphite system)

Ideally, I think we need to associate a layout technology with each font, 
and keep one font covering each script for each technology. If the shapers 
also say which technologies they can deal with we can be sure that we're 
always giving the font to a shaper that can deal with it.

The current situation seems to be that the order of things in font.conf has 
to match the order in which Pango stores the shapers, which as I have 
already demonstrated, is too fragile :-)

> Keith Packard        XFree86 Core Team        HP Cambridge Research Lab

Regards,
Eric