[XeTeX] Character Variants

[Jonathan Kew]

On 27 Mar 2011, at 21:42, mskala at ansuz.sooke.bc.ca wrote:

> On Sun, 27 Mar 2011, Paul Isambert wrote:
>> Both, actually. The font by itself does nothing, it simply indicates what the
>> rendering engine should do. I suppose (because it's seems the simplest way
>> here) that this is implemented as a ligature.
> 
> By that logic it could be said that no software ever does anything, it
> only indicates what the computer hardware should do. What I meant was,
> does the handling of variation selectors consist entirely of a feature
> table with "sub xxx yyy by zzz" entries, so that it could be turned on in
> XeTeX the same way we turn on ligatures?  That's a simple yes/no question.
> As you say it would make sense for the answer to be "yes", but I haven't
> been able to find an authoritative answer on the Net.

Variation sequences _could_ be implemented using a ligature mechanism, but this is not in fact how it's normally done. There is a special 'cmap' subtable (format 14) used to map sequences of <base character, variation selector> to glyphs, without the requirement for the variation selector code by itself to map to a glyph ID at all.

I don't think the current xetex code supports this, although it should be a pretty minor update if someone wants to work on a patch....

JK