Re: [Lilypond-auto] Issue 2468 in lilypond: augmentation dot of a moved rest is placed on a line

[lilypond]

Comment #8 on issue 2468 by address@hidden: augmentation dot of a  
moved rest is placed on a line
http://code.google.com/p/lilypond/issues/detail?id=2468

> X and Y calculations should never be inter-dependent
That would be convenient, but sometimes X and Y positions inter-depend.   
Issue 1088 shows how a dot on a rest can influence horizontal note-spacing,  
depending on how high the rest needs to be. (You can fool ver 2.14, though,  
with {c'16[ r16. cis']})

In the patch for issue 2180, it looked like the plan was to let the code in  
dot-column.cc run before line-breaking.  It looked like 'pure' estimates of  
Y-positions were good enough to complete the dot configuration.  A rest  
might later move to avoid a beam, but such motions never require  
re-configuring the dots, because rests are vertically far enough from the  
clusters of dots on chords.  The dot can always simply follow her rest.

Minor variations on the bug report show the dot at strange fractions of a  
staff-space, so it looks like the dot is not simply following the rest.   
The rest always moves by whole-staff-spaces, but the dot does not.

> the only way to sort this out is to revamp Dots' vertical shifting  
> mechanism
Not the /only/ way; we can rewind the patches for issues 2180, 1986, and  
1952.