Re: Unexpected font rendering with pango

[Keith Packard <keithp keithp com>]

Around 14 o'clock on Jun 27, Eric Mader wrote:

>  It might also make sense to take the language into account too; possibly
> with a notion of what characters are used to write a given language in a
> given script. I haven't thought about that enough to say for sure.

That's precisely what fontconfig does.  It calculates language coverage 
based on Unicode coverage of language orthography.  It's a relatively 
straightforward technique aside from the collection of language 
orthographies; fontconfig is still missing orthographies for 17 of the 139 
languages in ISO 639-1.

> Another point: for complex text like Arabic and Hindi, you really, 
> really want to try and keep it all in the same font, because that's the 
> only way to get the correct contextual behavior.

Arabic is not horribly problematic as most fonts provide coverage for most 
languages.  One issue for Arabic is that newer fonts are less likely to 
include the presentation forms and are starting to expect shapers to 
perform compositing.  That may require some additional information for 
font matching.

I've seen significantly more trouble with languages presented in Latin or 
Cyrillic scripts where the lack of a language tag often results in uncommon
glyphs being presented in an unexpected font.

If an application wants to ensure that a complete document is presented in
a single font, it can pass the set of Unicode values from the document to
the font selection mechanism and request a single font covering those
values.  Language tags simply provide a convenient shorthand for this
mechanism.

-keith