Re: cross-staff bug?

[Mike Solomon]

On Jul 25, 2014, at 6:19 PM, David Nalesnik <address@hidden> wrote:

> Hi,
> 
> Attempting to put dynamics above cross-staff notes using the code below leads 
> to some awful collisions.  I've recast my example in three ways, to no avail.
> 
> In the first example, besides the collisions, the final stem is reversed and 
> I get a number of programming errors:  "Grob direction requested while 
> calculation in progress."  (See attached for image.)
> 
> The second still has collisions and throws the programming error "My 
> pure_y_common is a VerticalAlignment, which might contain several staves."
> 
> The third throws no errors, but the collisions are still there.
> 
> I could of course use overrides to move objects out of the way, but I'd much 
> rather learn that I'm doing something wrong, recast the example, and let 
> LilyPond handle placement for me.
> 
> (To the suggestion that I simply put the dynamics below rather than above: 
> this won't always work in the piece I'm engraving.)
> 
> Thanks,
> David
> 
> %%%%%%%%%%%%%%%%%%
> \version "2.19.10"
> 
> up = \change Staff = "up"
> dn = \change Staff = "down"
> 
> \new PianoStaff <<
>   \new Staff = "up" {
>   s8*7^\p^\< s8^\f
>   }
>   \new Staff = "down" {
>     e'8[( \up b' \dn g' \up d'' \dn b' \up f'' \dn d'' \up 
> a''\marcato\fermata])
>   }
> >>
> 
> \new PianoStaff <<
>   \new Staff = "up" {
>     s1
>   }
>   \new Staff = "down" {
>     e'8[(^\p^\< \up b' \dn g' \up d'' \dn b' \up f'' \dn d'' \up 
> a''\marcato\fermata^\f])
>   }
> >>
> 
> \new PianoStaff <<
>   \new Dynamics {
>     s8*7^\p^\< s8\f
>   }
>   \new Staff = "up" {
>     s1
>   }
>   \new Staff = "down" {
>     e'8[( \up b' \dn g' \up d'' \dn b' \up f'' \dn d'' \up 
> a''\marcato\fermata])
>   }
> >>
> %%%%%%%%%%%%%%%%%%%%%%%%
> 

Unfortunately, I don’t think that there is anything you can do.

Cross-staff objects in LilyPond are a major area in need of improvement.  
Several cases like this were fixed by a patch a couple years ago, but the patch 
was reverted because it caused more problems than it fixed.

I don’t think there is a sane workaround yet for this sorta thing (meaning 
something that doesn’t require extra-y-offset), although I’d be interested to 
see if people have found workarounds.

Cheers,
MS