Re: splitted rests

[Han-Wen Nienhuys]

address@hidden writes:
> > Not quite. The multi-measure rest is a spanner, and getting the
> > rests in the right spot horizontally will be difficult, not to
> > mention the fact that they won't participate in rest collisions. I
> > think the easiest approach is to force a few line breaks in a few
> > strategic places and use { r2 r2. } at those spots, and disallow
> > other breaks halfway the 5/4.
> 
> My idea is that both branches of the ifelse construction are computed
> by lilypond simultaneously:
> 
>    < { r2 r2. } \\
>      { R4*5   }  >
> 
> and that at the time of deciding the line breaks the improper branch
> is simply dropped.  I think this approach would work for all staffs
> with a single voice.

You could even do this today using Scheme level hacking. Add to the
after-linebreaking callback of all three grobs, and make them commit
suicide in the relevant cases.  IIRC all the necessary Scheme
functions are already there.

-- 

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