[XeTeX] xetexref updated

[Jonathan Kew]

On 26 Jun 2007, at 8:32 am, François Charette wrote:

> Jonathan Kew a écrit :
>> Yes, I've tried Nafees Nastaleeq as well, and seen that there are  
>> some problems. The only OpenType Nastaliq font I'm aware of that  
>> currently renders properly with ICU in XeTeX is the OpenType  
>> version of Monotype's Noori Nastaliq, which is not freely  
>> available... but looks beautiful (see attached).
> Yes I like that font! Do you know if Linotype Qalmi also renders  
> well with XeTeX?

No, sorry - I don't have that font and don't know how it renders. I'd  
be interested in finding out! :)


> [snip]
> OK, perhaps it was just luck, but the above *did* work (see  
> attachment).

Yes, it was "just luck"... the diacritics appeared at their default  
"overstriking" locations, rather than being individually positioned  
on the particular base characters.

>>
>>> If the insertion of font features at the level of font mappings  
>>> were possible (is this at all feasible, Jonathan?), this could be  
>>> implemented quite easily with TECkit...
>>
>> No, not at present. Font mappings operate strictly on the  
>> character sequence that is actually going to be rendered in a  
>> given font -- where a "font" is a specific instance within XeTeX  
>> of a font at a particular size and with a certain set of features.
>>
> I know it is not possible with TECkit at the moment. What I meant  
> more precisely was more along the lines "Is an hypothetical  
> extension of TECkit that would allow the insertion of font features  
> (I am not talking of TeX macros here!) something that might be  
> theoretically or practically feasible?" I assume not.

Not at present; the prior issue would be devising a processing model  
that allows the text to carry such attributes without interfering  
with the standard shaping behavior. Color alone is relatively  
straightforward (though it still raises the problem of tracking the  
association from source characters to rendered glyphs, which may be  
far from simple), but people will want to vary many other attributes  
as well, such as OpenType features -- and that raises all sorts of  
complications for the whole OpenType shaping process.

JK