Tempo / tuplet bracket bug

[Trevor Bača]

Hi,

The following snippet exhibits a bug in the form of contention between
metronome mark and tuplet instantiation:


%%% BEGIN TEMPO INSTANTIATION BUG %%%

\version "2.13.9"

\new Staff {
   \time 2/8
   \times 2/3 {
      \tempo 8=52
      c'8 (
      c'8
      c'8
    }
   \time 2/8
      c'8
      c'8 )
}

%%% END TEMPO INSTANTIATION BUG %%


The interpreter gives the following warnings:

GNU LilyPond 2.13.9
Processing `0329.ly'
Parsing...
Interpreting music...
programming error: stopped tuplet bracket has left nor right bound.
continuing, cross fingers
0329.ly:10:10: warning: unterminated slur
      c'8
          (
0329.ly:16:10: warning: cannot end slur
      c'8
          )
Preprocessing graphical objects...
Solving 1 page-breaking chunks...[1: 1 pages]
Drawing systems...

And the output is attached here as tuplet-bracket-bound-bug.png.


The problem apparently has to do with the lexical spot at which the tempo
mark instantiates. Moving the tempo mark prior to the tuplet \times command
provides a workaround:

%%% BEGIN WORKAROUND %%%

\version "2.13.9"

\new Staff {
   \time 2/8
   \tempo 8=52
   \times 2/3 {
      c'8 (
      c'8
      c'8
    }
   \time 2/8
      c'8
      c'8 )
}

%%% END WORKAROUND %%%


Note that what makes this one particularly slippery is that the output from
the buggy example (shown in the png attached here) is actually syntactically
wrong: the measure of 2/8 prints with three eighth notes (because the tuplet
bracket goes missing). The rhythmic inaccuracy is easy enough to spot in
this pared-down example; but in a situation of considerably more rhythmic
complexity this can be lethal. (In the realworld score that I am working on
that surfaced the bug it was, in fact, only the very visible absence of an
extended slur that made this one possible to find.)


Trevor.


-- 
Trevor Bača
address@hidden

tuplet-bracket-bound-bug.png
Description: PNG image