[XeTeX] segmentation fault with a given font

Pablo Rodríguez oinos at web.de
Fri Jan 26 22:11:39 CET 2007


Jonathan Kew wrote:
> On 16 Jan 2007, at 5:18 pm, Pablo Rodríguez wrote:
>>
>> I get a segmentation fault. I know that the problem is in the font
>> naming. ....
> 
> Naming looks OK....
> 
>> And the backtrace is the following:
> 
> This looks like a problem with a kerning table in the font. I can't  
> tell from the report here whether there's something wrong with the  
> font itself, or with the ICU library functions that interpret it. If  
> you can run MS Font Validator on the font, it would be interesting to  
> see what it reports.

Sorry for not having answered before.

I asked about naming because I had installed a previous version and it
worked fine and the only thing I did for the new release was changing
the TTF name. The previous version is affected by the same kind of
problem. Since I have the Fontforge source, I investigated further.

Checking all tables with spot tool from the Adobe Font Development Kit
for OpenType, all is well, excepting the GPOS table, that makes spot crash.

> I'm not familiar with this font; is it a freely available font that I  
> could take a look at, or a commercial product?

It is a font of mine that mixes glyphs released under different licenses
(GPL & OFL), whose copyleft provisions render them incompatible with
each other. So I'm not allowed to distribute the font until one of them
is dual-licensed (or a newer OFL version allows this).

I realized that I get the same result removing all characters in the
Greek and Greek extended panes from the actual Junicode font. You cand
download the FontForge source at http://ousia.f2o.org/juni.sfd.bz2 and
the OpenType font at http://ousia.f2o.org/juni.sfd.bz2.

I'm afraid it might be a bug in FontForge.

Thanks for your help and sorry for the delayed answer,


Pablo


More information about the XeTeX mailing list