On Tue, Dec 31, 2024 at 11:02:17AM -0800, H. S. Teoh via LilyPond user
discussion wrote:
[...]
> I'm testing out the latest git HEAD version of lilypond with an
> orchestral score, and I'm getting this error:
>
> -------------
> Parsing...ERROR: In procedure %resolve-variable:
> Unbound variable: ly:duration-length
> ice-9/eval.scm:159:9: Throw to key `numerical-overflow' with args `("divide"
> "Numerical overflow" #f #f)'.
> -------------
[...]
It's been more than a year, but since then, I've figured out that there
are two separate problems here. The issue with ly:duration-length can
be fixed by substituting it with ly:duration->moment, but it has nothing
to do with the numerical-overflow error.
Turns out that the numerical-overflow error appears to be a bug /
incompatibility of \articulate with trill spans. I've managed to reduce
the failing case from my large complex score to an almost-minimal case
that reproduces the problem:
```
\version "2.27.2"
\include "articulate.ly"
\score {
\articulate {
d='''1\startTrillSpan
s2... s16\stopTrillSpan
}
}
```
On latest git HEAD, lilypond aborts with this error:
```
Processing `test.ly'
Parsing...ice-9/eval.scm:159:9: Throw to key `numerical-overflow' with args
`("divide" "Numerical overflow" #f #f)'.
```
I'm not exactly sure what's going on in articulate.ly that's triggering
this error, but removing \startTrillSpan and \stopTrillSpan makes the
problem go away.
Perhaps it's a bug in articulate.ly? Even if we're not supposed to use
trill spans with \articulate, it should at least give a more helpful
error message (e.g., point out which line of the input is the offending
one).
How do I go about filing a bug for this?
T
--
War doesn't prove who's right, just who's left. -- BSD Games' Fortune
