On Feb 14, 2012, at 10:10 AM, Joe Neeman wrote:
>
>
> Actually, I was referring to a comment that has since disappeared:
>
> - because this loop results in the looking up of extents, it may
> trigger
> - the calling of this function a second time before the first one
> finishes
> - to execute. thus, the outside-staff-priority may be set to false.
> that
> - is why this check is necessary.
>
Ah, OK. Yeah, the current comment is a rewrite of the older one.
I can tell you very concretely where in the code this creeps up. In
metronome-text.ly, when you run the regtests (but not when you compile the file
normally), there is an outside staff priority that is set to false somewhere
between the sorting using staff_priority_less in skyline_spacing (line 819 of
my patch) and the work on line 715. Barring truly strange behavior, the only
two lines that seem like they could possibly cause outside-staff-priority to be
set to false between these two points are lines 694 and 695 (the relative
coordinate lookups). If you comment this bit around 715 out and then run the
regtests (meaning comment out 716-724), you should see it show up in the
regtests.
I can try to do some valgrinding to track down the how/where/whys of this. My
guess is that it may have something to do with metronome marks' after line
breaking, but I'm really not sure. It takes me a long time because it's
behavior that only shows up in the regtests.
Cheers,
MS
_______________________________________________
lilypond-devel mailing list
[email protected]
https://lists.gnu.org/mailman/listinfo/lilypond-devel
