Re: \partcombine causes wrong tuplet bracket when tuplet begins with rest

[Dan Eble]

James Lowe <pkx <at> gnu.org> writes:
> On 15/05/15 22:38, H. S. Teoh wrote:
> > % The triplet in the first argument to \partcombine is wrongly
> > % formatted in the output; the tuplet bracket only spans the two quarter
> > % rests, whereas it should span both rests and the single note.
> > %
> > % - Note: this is not the same as issue 913; padding the first voice
> > %   to have the same length as the second does not change the wrong
> > %   tuple bracket.
> > % - For some reason, deleting the r4 in the second voice makes the
> > %   problem disappear, for no apparent reason.
> > % - Starting the tuplet with a non-rest note also makes this problem go
> > %   away.
> > % - This problem does not occur if \partcombine is not used.
> > 
> > \version "2.19.19"
> > \score {
> >     \partcombine
> >         { \tuplet 3/2 { r r c'4 } }
> >         { R1 r4 }
> > }

> Is this just another symptom of
> https://code.google.com/p/lilypond/issues/detail?id=913
> ?

Not the same as 913.  In this case, it is probably that the part
combiner does not recognize at the time it is assigning notes
to voices that a bracket is desired in the output.

You might be able to use invisible slurs (or phrasing slurs) or
the manual override commands (\partcombineApart etc.) to
work around this.

There are similar problems with beams.  The notes are assigned to
voices before beaming, yielding ugly results.  I have resorted to
beaming entire songs manually in extreme cases.