Re: "Pausing" Voice

[Thomas Morley]

Am Do., 9. Juli 2020 um 12:30 Uhr schrieb David Kastrup <dak@gnu.org>:
>
> Thomas Morley <thomasmorley65@gmail.com> writes:
>
> > Hi all,
> >
> > if temporary Voices occur, then the main Voice may not contain
> > rhythmic-events for the duration of those temporary Voices.
> > I'd like to identify where the main Voice "pauses", like in this test-code:
>
> Just a vague idea: does the busyGrobs context property maintained by the
> Grob_pq_engraver help in any manner?
>
> --
> David Kastrup

Hi David,

I tried:

\version "2.20.0"

#(define (print-busy-grob-info context)
  (let ((busy-grobs (ly:context-property context 'busyGrobs)))
    (if (pair? busy-grobs)
        (for-each
          (lambda (bg)
            (format #t "\nAt ~a we see ~a ending at ~a, pitch is ~a"
              (ly:context-current-moment context)
              (cdr bg)
              (car bg)
              (ly:prob-property (ly:grob-property (cdr bg) 'cause) 'pitch)))
          busy-grobs)
        (format #t "\nAt ~a no busy grobs "
          (ly:context-current-moment context)))))

Test_busy-grob_engraver =
#(lambda (context)
  (make-engraver
    ((process-music this-engraver)
      (print-busy-grob-info context))))

\new Voice
  \with {
    \consists \Test_busy-grob_engraver
  }
  {
    %% mom 0
    b1\startTextSpan\((
    %% mom 1
    \new Voice c'
    %% mom 2
    d'\stopTextSpan)\)
    %% mom 3
  }

==>
At #<Mom 0> no busy grobs
At #<Mom 1> we see #<Grob NoteHead > ending at #<Mom 1>, pitch is #<Pitch b >
At #<Mom 2> no busy grobs
At #<Mom 3> we see #<Grob NoteHead > ending at #<Mom 3>, pitch is #<Pitch d' >

Which is very nice at first sight.

Alas, in IR busyGrobs is described as:
"
busyGrobs (list)

A queue of (end-moment . grob) cons cells. This is for internal (C++)
use only. This property contains the grobs which are still busy (e.g.
note heads, spanners, etc.)."

In the above posted output I miss the spanners, though. To be clear,
in my use-case I'll _wish_ to disregard spanners. I'm only interested
in rhythmic-events, including skip-events.
Can I rely on the busyGrobs-description being wrong?

Furthermore, I think I need to do some settings in
stop-translation-timestep. There the output would be different,
'busyGrobs is never '(), according to my tests.
So I would need to read 'busyGrobs in process-music and carry over the
results to stop-translation-timestep to rely on it.
If busyGrobs may contain spanners or other not-rhythmic-events I'd
need to do some selections.

Thus the question is, whats cheaper:
(1) calculating, based on current-moment and 'length of the
rhythmic-event (as in my initial post)?
(2) reading busyGrobs, move the result, probably select for rhythmic
grobs? Would this work for skips as well?

Any advice?

Many thanks,
  Harm