Am So., 28. Juni 2026 um 19:28 Uhr schrieb Dan Eble
<[email protected]>:
>
> On 2026-06-28 08:33, Thomas Morley wrote:
> >
> > To fix #6946 it would be sufficient to create only those Glissando
> > grobs really needed, then there would be no need to delete some of
> > them afterwards.
> > There seem to be some conditions for that purpose in the original
> > cc-engraver, but I don't understand them sufficiently.
>
> Harm,
>
> Issue #6946 seems to be a recurrence of issue #1639, which was fixed in
> b8dfb88ca356cdde12cccb4cf57ecc64397c137b. That change defers grob
> suicide until other engravers have had the opportunity to acknowledge
> the end of the spanner.
>
> Probably, the best response would be to update the Scheme implementation
> of Glissando_engraver to defer grob suicide similarly. Since the issue
> is a regression, the change would ideally be kept minimal so that it
> could easily be back-ported to 2.26.
For the initial codes-example of this thread that would probably mean:
#(define-public My_glissando_engraver
(lambda (context)
(let ((glissando-event #f)
(start-glissando-line #f)
(stop-glissando-line #f)
(to-kill '())
(gliss-grob #f))
(make-engraver
(listeners
((glissando-event this-engraver event)
(set! glissando-event event)
(set! start-glissando-line #t)))
(acknowledgers
((note-column-interface this-engraver grob source-engraver)
(when stop-glissando-line
(ly:engraver-announce-end-grob this-engraver gliss-grob
(ly:context-property context 'currentMusicalColumn))
(set! to-kill (cons gliss-grob to-kill))
(set! stop-glissando-line #f)
(set! gliss-grob #f))
(when start-glissando-line
(set! gliss-grob
(ly:engraver-make-grob
this-engraver 'Glissando glissando-event))
(set! start-glissando-line #f)
(set! glissando-event #f)
(set! stop-glissando-line #t))))
((stop-translation-timestep this-engraver)
(for-each ly:grob-suicide! to-kill))
;((finalize this-engraver)
; (for-each ly:grob-suicide! to-kill))
))))
Doing it in `stop-translation-timestep` seems to work, though Mike's
fix for #1639 does it in `finalize`.
Do you have an advice on it?
Well, I could try it out by running regtests, but maybe you have some
insight already.
> Secondarily, it might be helpful to make Spanner_break_forbid_engraver
> more robust in the presence of dead grobs. I'll investigate this part.
That would be great.
> --
> Dan
>
Many thanks,
Harm
