Re: LILF table in emmentaler

[Tobias Hoffmann]

On 28/11/13 12:48, Werner LEMBERG wrote:
> > Now, for recent versions of emmentaler (same font file), LILC contains:
> >    (subfont . "feta20")
> >    (subfont . "feta-alphabet20")
> >    (subfont . "feta-flags20")
> >    (subfont . "feta-noteheads20")
> >    (subfont . "parmesan20")
> >    (subfont . "parmesan-noteheads20")
> >
> > [...]
> >
> > but the subfonts array has not been updated in
> > these commits.
> Please test the attached patch.

I can't easily run a full test right now, because installing fontforge 
would require me to update fontconfig to 2.11, which currently breaks 
xpdf (debian testing) ...

Nevertheless I tried your patch as far as possible and it produces the 
following emmentaler-20.subfonts file (fontforge compiles this as-is 
into the LILF table), which LGTM:
   feta20 feta-noteheads20 feta-flags20 parmesan20 parmesan-noteheads20 
feta-alphabet20

But I noticed that the generated dependency files also have some 
problems, for example take emmentaler-20.dep:
   emmentaler-20.otf: $(outdir)/feta20.pfa \
    $(outdir)/parmesan20.pfa  \
    $(outdir)/feta-alphabet20.pfa feta20.otf-table \
    $(outdir)/feta-alphabet20.pfa feta20.otf-gtable

Notice that feta-alphabet20.pfa appears twice, and some of the other 
subfonts do not...
Even worse: The actually generated and merged Type 1 font are  *.pfb , 
not  *.pfa !

I don't really understand the bigger picture of lilyponds build process, 
but from what I see, I'd suggest  the subfonts-array to be defined 
somewhere at the top of gen-emmentaler-script.py,
and then to generate not only the .subfonts - files, but also the 
dependencies and the MergeFonts directives from it. I'm not sure about 
the MergeFonts("feta-alphabet...") though, I seems to be deliberately 
after the SetUnicodeValue-loop...

  Tobias