Re: Font Problems

[Owen Taylor <otaylor redhat com>]

Eric Mader <mader jtcsv com> writes:

> At 02:17 PM 8/20/2002, I wrote:
> > * Arabic text doesn't seem to be contextually forming. I'm trying to
> > debug this, but don't have any information yet, other than it does
> > seem to call arabic-xft.c to process the Arabic text. (Also, while
> > debuggin this, it looks like arabic-xft.c is getting called for
> > *each word* in the Arabic text. I guess that it's finding the space
> > character in another font. This is probably a performance hit...)
> > I'm seeing some error messages about invalid GPOS and GSUB tables,
> > so that may be some of the problem... did we perhaps loose some of
> > the bullet-proofing when things got updated to ft2?
> 
> I tracked down the message about invalid GSUB/GPOS tables. It was
> comming from a beta version of Raavi, the Windows XP Gurmukhi
> font. This font has an empty LangSys for the Devanagari script. I
> patched Load_ScriptList in xftopen.c to just ignore the bad script
> rather than causing the whole font to be ignored. Should I submit this
> patch to CVS?
> 
> (The production version of Raavi doesn't have this problem, but it's
> still probably worth making any other scripts in the font work, even
> if one of them is bad...)

How is an "empty LangSys" represented in a font? I'm not opposed
to adding in some sort of tolerance if it's pretty clean, but if
there is a LangSys pointing off the end of the font or something,
then I think we should definitely refuse to load the font.

In general, I think we should be strict about validation of fonts
until we actually have reports about production fonts that are
causing problems ... otherwise, we are just encouraging (in a small
way) such broken fonts.

Regards,
                                        Owen