[XeTeX] Help with XeTeX/ConTeXT and CTL

Jonathan Kew jfkthame at googlemail.com
Sun Jul 1 20:38:41 CEST 2012

On 1/7/12 19:21, Shriramana Sharma wrote:
> On Tue, Jun 26, 2012 at 8:30 PM, Jonathan Kew <jfkthame at googlemail.com> wrote:
>> OTOH, the people who wrote Graphite are still there, and if the graphite
>> engine is crashing, they may be able to help pinpoint the problem and
>> suggest a fix that could be applied to the library xetex is using, and/or a
>> way to work around the problem in your font.
>> Do *any* graphite fonts work in the current TL xetex builds?
> Hello Jonathan, Martin, Sharon and others. Sorry for the delay. I've
> since then tested XeTeX's Graphite ability with TeXLive XeTeX 2009
> packages that came with Kubuntu Precise. (Latest PPA from
> https://launchpad.net/~texlive-backports/+archive/ppa/ have been
> updated and are now downloading so I'll test that later.)
> Please find:
> https://sites.google.com/site/jamadagni/files/temp/xetex-graphite-smp-bug-20120701.tar.gz
> (The TTFs are OFL and the GDLs GPL v2+.)
> The Graphite Control ODT and PDF give the expected rendering.
> As for the TEX files, I am happy to note that both the Latex/Context
> behaviours with Xetex are equal (my god how do you people consistently
> capitalize these names in emails!) ... and that Latin Graphite rules
> (Charis SIL) and Tamil Graphite rules (Krishna Tamil) work. However,
> Brahmi Graphite rules don't work at all! :-(
> I am also unable to turn on the Tamil features on -- am I doing
> something wrong in specifying the features?

Yes; you need a semicolon to separate multiple features, not an 
ampersand. So try

\font\krishnaalt="Krishna Tamil/GR:Classical=True;Pulli=Ring" at 12 pt

and see if that works better.

> I wonder if there is some problem with the Graphite handling SMP code
> points? While typesetting (using Ctrl+T in TeXWorks) I still get the
> segfault I reported. I hope you people can reproduce that too.

I haven't tried to reproduce this (I'm not at the right computer just 
now), but I can easily believe there could be a bug that affects SMP 
codepoints in xetex/graphite. I wouldn't be surprised if you're the 
first person to try that...

> I note that if I comment out the \adinatha line and leave the SMP UTF
> text in, I get a series of boxes but no segfault, so it is not a
> problem with SMP text being part of the UTF-8 file but it seems to be
> a problem with the Graphite rendering trying to handle SMP text
> (maybe?).

Yes, SMP characters in themselves are not a problem for xetex; other 
people have definitely used them successfully (with opentype fonts). But 
using them with graphite may well be completely untested. Welcome to the 
bleeding edge.... (sorry!)

If you use the Adinatha font *without* the /GR modifier, so that xetex 
will render via ICU/OpenType, does it successfully generate a PDF with 
the expected glyphs for the Brahmi characters (but without any of the 
shaping from the graphite rules, of course)?

> I hope you people can look into this and help me out! Thanks!
> BTW that talk of HarfBuzz (and thereby, Gr2) integration into XeTeX is
> very interesting. May whoever is upstairs grant the capable people
> time to get this done sometime soon! :-)

More information about the XeTeX mailing list