such that rendering programs DTRT for both œôòøfooœôòù and œôòøbarœôòù.  

I don't see the problem, in principle.  The information about what font
should be used for what index entry is completely determined (and
deterministic) by other known information.

As Patrice says, though, it could be inconvenient to trace back the
@syn*indexes and @def*indexes.  I have no objection to adding the info
explicitly in the XML (or whatever) output.

k


Reply via email to