Re: LilyPond 2.1.30 released

[Werner LEMBERG]

> > How will you handle arbitrary text strings too complex to be
> > handled by lilypond itself (exotic scripts, mathematics, etc.)?
> 
> We'll leave the TeX output intact, but users will have to write
> their own glue/titling code (or use lilypond-book).
> 
> In a far future, I hope that we can link Lily to Pango or Omega so
> that we can "outsource" text typesetting.  Unfortunately, Pango
> doesn't do print.  I'm not sure where Omega is standing, last time I
> looked, development wasn't moving much.

You want to send LaTeX code to a library and getting back small EPS
files?  This is non-trivial due to the embedded fonts.  IMHO the
current dvips is not powerful enough to do that; the created eps files
are far too big because it still uses pfb without converting them to
cff.  Using PDF (either pdflatex or latex+dvipdfmx) might be a
possible alternative.

A first step could be this: Emit all text strings as a large LaTeX
document where the entries are put into \hboxes.  Then run LaTeX on
this file, writing the box dimensions into the aux file.

This would make two passes with LilyPond:

  lilypond --generate-latex-data foo.ly > foo.tex
  latex foo.tex
  lilypond --use-latex-data foo.ly

The --generate-latex-data option makes lilypond exit after parsing the
input file (and emitting the latex data).
  
I would like to see this anyway to have a chance to reliably select
non-cmr fonts for text strings in LilyPond; if you code the lilypond
part I'll provide the LaTeX part.


    Werner