Re: CVS yesterday not printing longas

[Han-Wen Nienhuys]

address@hidden writes:
> 
> This is yet another good reason for introducing a 16-bit font character
> lookup table, as I recently proposed.  But this is probably unlikely to be
> introduced before 1.7, since it requires major changes to many parts of
> lily.

I don't think this solves the real problem. You only need 16-bit fonts
if you use character sets that require all those 16 bits at the same
time.  Unless you absolutely need to have Chinese characters as
noteheads, this solution is overweight and overengineered.
 
The real problem is that you have to have some silly kind of logic
that decides that for a particular style (eg. "baroque") you have to
combine different basic styles of noteheads (modern, mensural), and
that the current form of this logic is too closely tied to the old
layout of the font.

The real solution is to rewrite the interface to find-notehead-symbol,
such that it also looks up the font-metric to be used. Or perhaps an
interface that returns the molecule directly, i.e.

          (define (find-note-head-molecule grob)
                  "return molecule for GROB note head"
                  (...do funky stuff with looking up fonts ...)
          )

This is easy to do; left to the reader as an excercise. If you submit
quickly enough, it might even make 1.5.60.

-- 
Han-Wen Nienhuys | address@hidden | http://www.cs.uu.nl/~hanwen/