Crash on part generation in the presence of tremolos
Reported version
3.0
Priority
P0 - Critical
Type
Functional
Frequency
Once
Severity
S1 - Blocker
Reproducibility
Always
Status
closed
Regression
No
Workaround
No
Project
a8a8ee0
I don't see the pattern in the scores affected - I can generate parts for the Nutcracker Overture, which has no notes for Contrabass and Violoncello and is large. I can for a small wind section for Megalovania. I can't for either movement of Dvorak, which is an issue as the Dvorak parts were the entire point of that score.
Dvorak crashes if I so much as try to generate a single Flute I part. Maybe voices have something to do here? I put in cues in other voices on both Dvorak Scores, but not for either the Nutcracker Overture or Megalovania.
Edit: definitely voices, I could generate a part from the one staff that didn't have cues.
Attachment | Size |
---|---|
Dvorak_Symphony_No._8_Mvt1.mscz | 217.4 KB |
Comments
Confirmed, although there must be more to it than that since other scores with multiple voices can have parts generated successfully.
Alright, maybe it isn't voices, or the cue notes? I managed to generate parts on Contrabasso and Violoncello in Mvt1, so I tried some more things to isolate, short of actually building MS3, which is a thing I should do.
I found that when I tried to fake generate a part by making staves invisible, that also crashed the program. I did manage to find out that I could have both Violin II and Viola as the only two visible staves without crashing, though simply moving around the score crashed once, and I got this fun artifact
So maybe it's another layout thing? There are no mmrests, and Vl. II and Viola both crash if they are the only staves visible, but don't if they have even one other staff onscreen. To save clicks, attaching score with everything invisible except those two staves.
In reply to Alright, maybe it isn't… by Laurelin
Zooming in all the way on the artifact - the measure numbers 264 and 272 are displayed above that top stave, and yes, the measure overlay matches those numbers. The bottom system collision looks to be measures 268 and 277.
Does the crash still happen? I cannot reproduce it with the latest master.
Still happens. Use the attached score. It's not voices, my previous post. No idea what it actually is, maybe someone can run debug, see what happens when it crashes?
In reply to (No subject) by Jojo-Schmitz
OS: Windows 7 SP 1 (6.1), Arch.: x86_64, MuseScore version (64-bit): 3.0.0.4594, revision: 9769dff
Tremolo problem. In the attached score above, delete all tremolos avoids the issue.
From scratch:
1) Default score
2) Enter two notes -> add tremolo
3) Generate parts -> New all -> Ok
----> Crash
See:
Brilliant, thanks for the investigation! Very liekly related to #280012: Crash opening vtest/tremolo-1.mscx
Appears on December 12
There is a few commits between these two nightlies/commits, eg this one (the most likely here?): https://github.com/musescore/MuseScore/commit/fedb04b66e6f1825b45c91b75…
Does this extend to other staves in the piece? Because if I try generating a flute part, it crashes there too, and the flutes don't have tremolos.
Might be more than one issue.
But yes, seriously awesome find.
In reply to Does this extend to other… by Laurelin
Flute tremolos between notes start in measure 273 of Movement 3.
The debugger doesn't lie :-). It's the same crash, happening during tremolo layout. Flute has them around letter "N".
Lol, right, they have that transition that Vl. II had earlier. I think. Too many scores. You'd think I'd remember something as odd as flute tremolos, though. Thank you!
See https://github.com/musescore/MuseScore/pull/4444.
Fixed in branch master, commit 483740b42e
fix #280095 Crash on part generation in the presence of tremolos
Fixed in branch master, commit 62f5d132bf
fix #280095, fix #278964: rework duration types hanling with tremolo
Automatically closed -- issue fixed for 2 weeks with no activity.