Am Di., 13. Okt. 2026 um 00:33 Uhr schrieb David Kastrup <[email protected]>:
>
>
> Hi folks,
>
> people following development will have seen that I have a comprehensive
> approach to fixing the infamous issue #34 once and for all in the
> pipeline. I had worked with a set of regtests created by a contributor
> who I managed to annoy to a degree where he rescinded permission to use
> his tests.
>
> This is your chance: if you have small, self-contained examples of
> things going wrong (usually not exceeding a dozen lines or so), you can
> send them over.
>
> The unifying element will be: parallel music with grace notes in at
> least one track.
>
> Of particular interest are: "timing tracks" (straight sequences of time
> signature, key and clef changes, line breaks, special bars, repeat
> structures with skips in between) that run in parallel with music
> containing grace notes. Also other constructs that may have tripped you
> up.
>
> Chances are that there will not be a lot that doesn't work with the
> current code. But it needs a new set of tests to make it into LilyPond
> 2.27.4.
>
> Thanks!
>
> --
> David Kastrup
>
>
Below the (shortened) code where I was hit by #34 for the very first
time (it was for 2.12.3 back in 2011).
eins =
\new Voice
\relative c'' {
\set Staff.instrumentName = "Gitarre I"
\key g \major
\time 2/2
\repeat volta 2 { g'4 b8. g16 d4 d }
}
zwei =
\new Voice
\relative c'' {
\set Staff.instrumentName = "Gitarre II"
\key g \major
\time 2/2
<< { g2 fis } \\ { g,4 r r2 } >>
}
drei =
\new Voice
\relative c'' {
\set Staff.instrumentName = "Gitarre III"
\key g \major
\time 2/2
<< { \acciaccatura { fis8 } <g, b g'>4 } \\ { <g, b d>4 } >> r4 r2
}
\score {
\new StaffGroup
<<
\new Staff \eins
\new Staff \zwei
\new Staff \drei
>>
}
Nowadays I'd code it a bit differently: instrumentName in \with. Alas,
we have some examples in NR doing as above.
The documented workaround would work, it's a nuisance, though,
regarding all the temporary Voices.
BTW, which of your currently proposed MRs should I apply to test #34
myself, !3153 or more?
Cheers,
Harm