Re: Issue 430 in lilypond: Cross-staff beam craziness (when down-markup combines with down-articulation)

[Han-Wen Nienhuys]

2007/8/31, Joe Neeman <address@hidden>:
> On Friday 31 August 2007 13:23, address@hidden wrote:
> > Issue 430: Cross-staff beam craziness (when down-markup combines with
> > down-articulation) http://code.google.com/p/lilypond/issues/detail?id=430
> >
> > Comment #2 by hanwenn:
> > Hi Joe,
> >
> > can you review my patch for this fix?
>
> My preferred way for dealing with these cyclic dependency issues is to mark
> something cross-staff. If a grob has the cross-staff property set to true, it
> gets ignored for all pure-height calculations, so it's a good way to break
> this sort of cycle (plus, it's encapsulated in its own grob property).

Yes, but wouldn't this result in staves colliding if the script is
really large?  The advantage of doing it with Stem direction is that
you break the cycle at the exact point where unnecessary information
is requested.  Of course, it would be best if we could have that idea
encapsulated in a property by itself.  One extreme idea could be to
calculate up and down extents separately, perhaps with a

  Y-extent -> (up-extent  , down-extent)

dependency, which we would have just for Stem for now.

-- 
Han-Wen Nienhuys - address@hidden - http://www.xs4all.nl/~hanwen