On 2025-12-08 00:41, Werner LEMBERG wrote:
> Is this an intended change or do I miss something? If it is the
former, the 'Changes' document should be updated, together with the
`compound-time-signature.ly` snippet.
The change involves an error in 2.24 that has been fixed.
https://gitlab.com/lilypond/lilypond/-/issues/6789
We don't usually list fixed bugs in the Changes doc, right? I could add
an item anyway if you still think it would be helpful.
Aside from that, one could ask whether the default time signature
settings for 2/8 should allow whole-measure beams. I am aware of issue
6828, but I don't have Behind Bars at hand or time to investigate today.
Overriding the default settings for 2/8 changes the output of 2.25:
```
\relative c' {
\overrideTimeSignatureSettings 2/8 #1/8 2 #'()
\compoundMeter #'((2 8) (3 8))
c8 d e fis gis
c8 fis, gis e d
c8 d e4 gis8
}
```
This could also be considered for the case you have presented:
```
\relative c' {
\compoundMeter #'((2 3 8)) % <-- different
c8 d e fis gis
c8 fis, gis e d
c8 d e4 gis8
}
```
--
Dan
