Re: make-non-relative-music & partcombine

[David Kastrup]

Dan Eble <address@hidden> writes:

> I don’t see any differences in the regression tests when I remove 
> make-non-relative-music from this code.  If it’s really necessary, can anyone 
> suggest or provide a new regression test that covers it?  I’d be happy to see 
> it through the review process.  Thanks.
>
> (define-public (make-part-combine-music music-list direction chord-range)
>   (let* ((m (make-music 'PartCombineMusic))
>          (m1 (make-non-relative-music (context-spec-music (first music-list) 
> 'Voice "one")))
>          (m2  (make-non-relative-music  (context-spec-music (second 
> music-list) 'Voice "two")))
>          (listener (ly:parser-lookup 'partCombineListener))
>          (evs2 (recording-group-emulate m2 listener))
>          (evs1 (recording-group-emulate m1 listener)))
>
>     (set! (ly:music-property m 'elements) (list m1 m2))

It would obviously be relevant for

\relative {
  \partcombine { ... } { ... }
}

where the ... itself does not have a \relative inside.  We might not
have something like that in the regtests but I could well imagine that
stuff gets erratic in that case.

-- 
David Kastrup