Crash when deleting soundfont during playback

• Feb 17, 2015 - 21:59
Reported version
P2 - Medium
S1 - Blocker

1. Open score.
2. 'Play'.
3. 'View'>'Synthesiser'.
4. 'Delete'.

Result: Crash.


Not always reproducible.
See attached log.

Using MuseScore 2.0 Nightly Build 13eccf9 - Mac 10.7.5.


Status (old) active needs info

I can't reproduce, and the crash log looks totally unrelated. Could be a matter of memory being corrupted and only showing effects later. Meaning, the deletion of the soundfont might not actually be *cause* but only the trigger. I'd need precise steps to reproduce in order to know for sure.

Reported version 2.1

Well, the stable version now is 2.1, this issue stems from a time before 2.0 and apparently never got fixed?
I can't reproduce it though, not in 2.1 on Windows 7

  1. Open attached score (produced in 2.2.1).
  2. Enable 'Loop'.
  3. 'View'>'Synthesiser'.
  4. 'Play'.
  5. Add 'MuseScore_General.sf3' on beat 4.

Result: Crash (see attached log).


Also reproducible if deleting, or moving.
I've had similar crashing threads in other instances without using loop, hence why I've not confined the cause to that.

Using MuseScore 2.3 RC 16bd6b6 - Mac 10.11.6.

I believe it affects 3.0-dev too.

Marc said this on IRC (I think about the one above): 'it's deep inside synth code, not something trivial from what I can see'

There is another crash sometimes (duplicate the same bar to about 16) when moving multiple soundfonts around (three MuseScore_General.sf3 and Arachno SoundFont - Version 1.0.sf2). I think it shared the same crash thread as another when I load a score, perhaps mess with soundfonts(?), open another score (which stops playback) and, in Synthesiser, 'Load from Score'.

Potentially useful: When I hit Play immediately upon load of application and score, the icon reacts but it doesn't actually do anything else until I press it again. Maybe the synthesiser is loading?

About my comment on 'Load from Score', this could be related. That one might be to do with startup only (different crash thread), but maybe I'm wrong.

Severity S2 - Critical S1 - Blocker
Priority P2 - Medium
Regression No
Reproducibility Always
Workaround No