Comment #23 on issue 1110 by [email protected]: Wrong octave of repetition chord
with \relative and #{ #} syntax
http://code.google.com/p/lilypond/issues/detail?id=1110
Well, a bounty does not make all that much sense before one has decided on
a strategy. One strategy is to consider this a \relative-only problem and
let the behavior of q be to assign the pitches of the last chord in the
parser, and keep enough information around when doing \relative to deliver
an error message when the connection between last chord and repetition
looks fishy.
The second strategy would be to delegate the pitch assignment to the stage
of iteration. In that case, you could juggle music expressions around, and
whatever chord ends up before a q will get to be its reference.
That's trickier since you can't just keep a last-chord reference per Voice,
as that would make chords in simultaneous music decidedly awkward. Unless
one decides that temporal order is good enough since, frankly, no sane
person would want to do
<< {<a c> q} {<c e> q} >> anyway: the simultaneous chords would all have to
share one stem. But it might be relevant for something like
<< {<a c> s q} {s <c e> s} >>
where people might be surprised if q produces <c e> instead of <a c>.
_______________________________________________
bug-lilypond mailing list
[email protected]
https://lists.gnu.org/mailman/listinfo/bug-lilypond
