Crash when deleting a horizontal frame if corrupt part
Reported version
3.0
Priority
P1 - High
Type
Functional
Frequency
Few
Severity
S2 - Critical
Reproducibility
Always
Status
duplicate
Regression
Yes
Workaround
Yes
Project
Open the MCSZ file attached.
Delete horizontal frame between section C and the coda. This occurs either by clicking and pressing the delete key, or by right clicking and selecting Delete.
| Attachment | Size |
|---|---|
| Iko_Iko.mscz | 38.21 KB |
Fix version
3.1.0
Comments
Confirmed for this score, but unable to reproduce in others. Not obvious what the trigger is, debugger didn't catch the crash either.
In fact, the main problem is not the removal of the frame, but corruption - image below - in parts (last measure).
![parts.jpg]( "parts.jpg")
By saving a part, for example, Iko_Tenor_Saxophone.mscz , it crashes at the opening.
Delete the parts, and no more issue, the frame is cleared as expected.
So the question is rather how did corruption get to that place?
I leave the title almost as such, but it will evolve when the origin of corruption is understood (the real subject)
Or/and it also could be related to this issue: #283545: Crash on undoing removal horizontal frame
The fact the file is corrupted does not surprise me. I was encountering a lot of the widely reported issues, like crashing at startup, extraneous key signatures, parts reverting to concert key. So, I was reinstalling prior versions to try to get around these problems. In 3.0.4, I created a new file, loaded my style, then cut and pasted the bars from the corrupted file into a new file, and now it's all fine. So, my short term problem is fixed.
Same issue here. What is the workaround?
We don't fully understand the issue, but since it seems to be related to a corrupt score, the workaround is presumably to fix the corruption. Feel free to attach your score here so we can assist, and maybe it will also help us understand the root cause of the issue.
There is no corruption being reported. And no crash in lates build from the master branch, nor with 3.1 RC, so I guess we can mark this fixed.
Marked fixed. ;-)
oops
Automatically closed -- issue fixed for 2 weeks with no activity.
I reopen and mark this issue as a duplicate of #290559: Appending measures makes saving crash if corrupt part
But reversing it could have been just as true, if not more so, since it is earlier. But let's do it this way.
And so, likely another duplicate/another scenario of the root cause: #290096: Remove select range in a score with parts results in corruption