Two crashes with corrupted file
Attempts to fix frandog's file leads to two different crashes in 2.0.3 32-bit. Sorry this isn't a minimal example, but minimizing it causes a crash. ;) Just load the file and ignore the corruption.
Drag the 2/4 time signature to the second measure. The crash may take a while.
Select measure 22, staff 1. Press ^
The crashes occur on either the Ubuntu PPA or the AppImage 2.0.3 on either Ubuntu 14.04 or Fedora 16.
Based on stack trace, they appear to be two different crashes. I have a screen shot of a partial stack trace for Crash 1, if that helps. I can post these to the issue tracker if you want. Uncertain whether they have been dealt with.
|American Patrol.mscz||92.71 KB|
Crash 1 and crash 2:
1) File -> Parts -> Delete (all the parts) -> Ok
2) Drag the 4/4 time signature into the second measure
3) Drag the 2/4 time signature into the first measure
4) Select and Cut the 4/4 time signature of the second measure
5) File -> Parts -> New all -> Ok
So: American Patr.mscz
EDIT: Have you an idea of how the score got in this state? What operations eg with the time signatures and the parts, and in which order (steps)?
In reply to Crash 1 and crash 2: 1) File by cadiz1
1. Delete all parts.
2. Delete measure 22. OK
3. Reopen original file and delete first four parts.
4. Delete measure 22. Crash.
5. Reopen original file and delete second four parts.
6. Delete measure 22. Crash.
That narrows it down a little more.
The file is from here: https://musescore.org/en/node/117306 . I have no idea how it got that way. The objective here is to prevent the crashes, not to fix the file--although how it got that way is another possible issue.
In reply to 1. Delete all parts. 2. by RexC
"The objective here is to prevent the crashes, not to fix the file"
Indeed. You preach a convinced. And you do not believe say the words as well. I have done so many times and for a few years to replicate from scratch many dozens of crashes. This is the hardest and the longest. But the fact to fix this file give me already some pointers (it's the first step). I hope to succeed again, but when the indications of the user are minimal, incomplete (sometimes erroneous or inaccurate, unintentionally of course), the task is not easy.
Reported as issues #117696 and #117701.