bugDenemo - Bugs: bug #36937, 0.9.5 rc2 Windows crashes

 
 

bug #36937: 0.9.5 rc2 Windows crashes

Submitted by:  John Beattie <jabeattie>
Submitted on:  Tue 24 Jul 2012 01:05:34 PM UTC  
 
Category: NoneSeverity: 3 - Normal
Item Group: NoneStatus: Fixed
Privacy: PublicAssigned to: None
Open/Closed: Closed

Add a New Comment (Rich MarkupRich Markup):
   

You are not logged in

Please log in, so followups can be emailed to you.

 

(Jump to the original submission Jump to the original submission)

Sat 28 Jul 2012 07:32:30 PM UTC, comment #6:

Next crash after 9 & quarter hours - no crash recovery so may be same as last one.
This time I did have stderr and it suggested it was same as 36968 so information logged on that bug.

I think we need to wait for rc3 on this one to see what effect the changes have.

John Beattie <jabeattie>
Sat 28 Jul 2012 08:36:09 AM UTC, comment #5:

Looks like it lasted over 8 hours then stopped.
No crash recovery file created this time so just going on time of last autosave. Of course I forgot to turn on the logging so can't be sure it is same thing. Will set it off again.

John Beattie <jabeattie>
Fri 27 Jul 2012 08:41:40 AM UTC, comment #4:

Highlight the cursor was ticked.
I will turn it off and leave it running during the day.

John Beattie <jabeattie>
Fri 27 Jul 2012 07:55:34 AM UTC, comment #3:

This looks like some bug drawing the display. Do you have the Denemo cursor in "highlight" mode (it is the default, see view->cursor highlight). I ask because this causes a re-draw for every blink of the cursor, and so could eventually cause a crash if there were some lurking bug on draw.
I could imagine some threading bug in the gtk library under windows giving rise to such a thing - Denemo's own threads are the playback ones, but I would imagine as they don't do anything much in a quiescent program they wouldn't be responsible.
It could be worthwhile testing if it occurs with the blinking of the cursor turned off.

Richard Shann <rshann>
Project Administrator
Thu 26 Jul 2012 05:19:23 PM UTC, comment #2:

How had two failures using denemo-console in both cases the crashrecovery file has same time as autosave file. I wasn't present when either crashed as I just left them running (>6 hours).
First time I just ran denemo-console and console died along with main screen.
Second time I ran denemo-console from a dos prompt and redirected Stderr to a file.
The dos screen shows lots of Autosaving and then No of displays : 1 and that's the end.
Stderr shows the following:

(denemo-console.exe:47396): Gdk-WARNING **: /home/jjbenham/src/gub/target/mingw/src/gtk+-2.22.1/gdk/win32/gdkpixmap-win32.c:302: CreateDIBSection failed: Not enough storage is available to process this command.

(denemo-console.exe:47396): Gdk-WARNING **: /home/jjbenham/src/gub/target/mingw/src/gtk+-2.22.1/gdk/win32/gdkpixmap-win32.c:114: DeleteObject failed: The operation completed successfully.

(denemo-console.exe:47396): GLib-GObject-CRITICAL **: g_object_ref: assertion `G_IS_OBJECT (object)' failed

(denemo-console.exe:47396): Gdk-CRITICAL **: gdk_drawable_get_size: assertion `GDK_IS_DRAWABLE (drawable)' failed

No obvoius signs of my allocated memory increasing so don't suspect memory leak perhaps something else just happens to be using a lot at the wrong time.

NB the stderr contains load of warning at start up. These are not related to this bug but just in case anybody is interested this is what I got.

(denemo-console.exe:47396): GdkPixbuf-WARNING **: Cannot open pixbuf loader module file 'C:\Program Files\Denemo\usr\lib\gdk-pixbuf-2.0\2.10.0\loaders.cache': No such file or directory

    • (denemo-console.exe:47396): WARNING **: couldn't open output stream
    • (denemo-console.exe:47396): WARNING **: initializing audio backend 'portaudio' failed, falling back to dummy

(denemo-console.exe:47396): Gtk-WARNING **: Error loading icon from file 'C:\Program Files\Denemo\usr\share\denemo\icons\denemo.png':
Failed to open file 'C:\Program Files\Denemo\usr\share\denemo\icons\denemo.png': No such file or directory

(denemo-console.exe:47396): GdkPixbuf-WARNING **: Cannot open pixbuf loader module file 'C:\Program Files\Denemo\usr\lib\gdk-pixbuf-2.0\2.10.0\loaders.cache': No such file or directory

This is followed by load of warning when loading scripts one example is:

    • (denemo-console.exe:47396): WARNING **: Cannot place LineBreak after Anacrusis as requested,

because I haven't seen Anacrusis yet
To fix this delete the LineBreak command save the command set,
Quit and restart Denemo
Then re-install LineBreak by right clicking on the Anacrusis item
and choosing More Commands
And finally saving command set again

John Beattie <jabeattie>
Tue 24 Jul 2012 05:08:35 PM UTC, comment #1:

I have left Denemo with an altered score for an hour or two using rc2 and it stayed up, I could see that it was not gobbling up memory either.

There is an alternative binary called denemo-console.exe which does not hide the console, though I am not sure what might happen then if Denemo crashed.

The "restore" code is of dubious quality - in fact I don't ever recall it doing anything useful for me. It is (IIRC) separate from the autosave file, which is also something that I have never had occasion to use - I don't think I've ever noticed code that would load that file, so I suspect it is someone's unfinished project :)

Richard Shann <rshann>
Project Administrator
Tue 24 Jul 2012 01:05:34 PM UTC, original submission:

Here's today's impossible bug!

Using 0.9.5. rc2 on Windows but I've also seen it on rc1.

Denemo was left open while I worked on other things eg bugs.
After a while it crashed and got the windows message that it had
stopped working. This last time I had been looking at Score Layout before I started writing it up. When I restarted it wasn't recognised as a crash and so it was restored to last save not autosave so the last changes weren't restored (they weren't things I wanted saved anyway so no loss). I think that some of the previous times it happened it was recognised as a crash & restored so there might be 2 problems here- the crash & the failure to restore.

I have no idea what it was doing at the time of the crash (hence impossible bug). Perhaps is was doing an autosave or something and causing an assertion due to lack of resources or something. Is there a way to turn on any debugging that would tell us more?

John Beattie <jabeattie>

 

(Note: upload size limit is set to 16384 kB, after insertion of the required escape characters.)

Attach File(s):
   
   
Comment:
   

No files currently attached

 

Depends on the following items: None found

Items that depend on this one: None found

 

Carbon-Copy List
  • -unavailable- added by rshann (Posted a comment)
  • -unavailable- added by jabeattie (Submitted the item)
  •  

    Do you think this task is very important?
    If so, you can click here to add your encouragement to it.
    This task has 0 encouragements so far.

    Only logged-in users can vote.

     

    Please enter the title of George Orwell's famous dystopian book (it's a date):

     

     

    Follow 3 latest changes.

    Date Changed By Updated Field Previous Value => Replaced By
    Sun 20 Jan 2013 06:51:40 PM UTCrshannStatusNeed Info=>Fixed
      Open/ClosedOpen=>Closed
    Tue 24 Jul 2012 05:08:35 PM UTCrshannStatusNone=>Need Info

    Back to the top


    Powered by Savane 3.1-cleanup