Crash solely occur when using Annotations:
0) start slicer (welcome to slicer being the default module)
1) load a volume
2) open annotations module
3) create a fiducial on 2D view
4) quit slicer

Memory leaks occur when:
0) start slicer (welcome to slicer being the default module)
1) create a fiducial on 2D view
2) quit slicer

Now turned off on release builds - still an issue on debug builds.

Is this still an issue ? Thanks Jc

No crashes or leaks in debug mode using those steps right now.

Marking no change required since it was fixed through an unknown revision.

I still see crash on exit most of the time.

Could somebody confirm that the workflow described by Julien is still causing a crash ?

Confirmed for the mac nightly 5-8

I fixed a crash on exit yesterday, can you confirm it's still crashing on exit with the 5-9 nightly? I tested the steps that Julien included here, Ron, if you see a crash or leak with a different set of steps, please include them in a note.

This bug is a bit of a moving target, since we get crashes on exit due to lots of different changes in the unstable code base. I agree that they have to be cleared out, as well as memory leaks, but I'd prefer to have one mantis issue per specific crash or memory leak, with each report including the list of leaking classes or the crash back trace (ideally).

As soon as the crash related to the workflow that Julien described is fixed, I guess the issue should be closed.

The workflow described should also be turned into either a python unittest (or regular Cpp test) reproducing it. That way, the dashboard will turn red if there is a regression.

The following backtrace reported by Ron May 10, 2012, can be obtained following the workflow described at the top of this issue.

Process: Slicer [4111]
Path: /Users/USER/Downloads/Slicer-4.1.0-2012-05-08-macosx-amd64.app/Contents/MacOS/Slicer
Identifier: ???
Version: ??? (???)
Code Type: X86-64 (Native)
Parent Process: launchd [134]

Date/Time: 2012-05-10 08:12:02.156 +0800
OS Version: Mac OS X 10.7.3 (11D50)
Report Version: 9

Interval Since Last Report: 3544678 sec
Crashes Since Last Report: 1222
Per-App Interval Since Last Report: 709666 sec
Per-App Crashes Since Last Report: 111
Anonymous UUID: F3CF6168-B63E-40A4-B96C-BB98DB23AB01

Crashed Thread: 0 Dispatch queue: com.apple.main-thread

Exception Type: EXC_BAD_ACCESS (SIGSEGV)
Exception Codes: KERN_INVALID_ADDRESS at 0x0000000000000000

VM Regions Near 0:
-->
__TEXT 0000000100000000-0000000100025000 [ 148K] r-x/rwx SM=COW /Users/USER/Downloads/Slicer-4.1.0-2012-05-08-macosx-amd64.app/Contents/MacOS/Slicer

Application Specific Information:
objc[4111]: garbage collection is OFF

Thread 0 Crashed:: Dispatch queue: com.apple.main-thread
0 libMRMLCore.dylib 0x0000000105d4358f vtkMRMLSceneViewNode::~vtkMRMLSceneViewNode() + 31
1 libMRMLCore.dylib 0x0000000105d36bfa vtkMRMLScene::~vtkMRMLScene() + 346
2 libMRMLLogic.dylib 0x00000001031b4066 vtkMRMLAbstractLogic::SetMRMLScene(vtkMRMLScene*) + 54
3 libMRMLLogic.dylib 0x00000001031b4a14 vtkMRMLAbstractLogic::~vtkMRMLAbstractLogic() + 36
4 libSlicerBaseLogic.dylib 0x000000010308f910 vtkSlicerApplicationLogic::~vtkSlicerApplicationLogic() + 1280
5 libMRMLDisplayableManager.dylib 0x000000010315aa8b vtkMRMLDisplayableManagerFactory::~vtkMRMLDisplayableManagerFactory() + 59
6 libMRMLDisplayableManager.dylib 0x000000010315b74b vtkMRMLThreeDViewDisplayableManagerFactory::~vtkMRMLThreeDViewDisplayableManagerFactory() + 27
7 libMRMLDisplayableManager.dylib 0x000000010316262d tcf_1 + 45
8 libsystem_c.dylib 0x00007fff8a73a7c8 cxa_finalize + 274
9 libsystem_c.dylib 0x00007fff8a73a652 exit + 18
10 0x000000010001ef0f start + 59

Thread 1:: Dispatch queue: com.apple.libdispatch-manager
0 libsystem_kernel.dylib 0x00007fff9292c7e6 kevent + 10
1 libdispatch.dylib 0x00007fff8a52c5be _dispatch_mgr_invoke + 923
2 libdispatch.dylib 0x00007fff8a52b14e _dispatch_mgr_thread + 54

Thread 2:: com.apple.CFSocket.private
0 libsystem_kernel.dylib 0x00007fff9292bdf2 select + 10
1 com.apple.CoreFoundation 0x00007fff941bbcdb CFSocketManager + 1355
2 libsystem_c.dylib 0x00007fff8a7478bf _pthread_start + 335
3 libsystem_c.dylib 0x00007fff8a74ab75 thread_start + 13

Thread 3:
0 libsystem_kernel.dylib 0x00007fff9292bbca __psynch_cvwait + 10
1 libsystem_c.dylib 0x00007fff8a74b274 _pthread_cond_wait + 840
2 QtWebKit 0x0000000101fd3048 WTF::TCMalloc_PageHeap::scavengerThread() + 520
3 QtWebKit 0x0000000101fd30e9 WTF::TCMalloc_PageHeap::runScavengerThread(void*) + 9
4 libsystem_c.dylib 0x00007fff8a7478bf _pthread_start + 335
5 libsystem_c.dylib 0x00007fff8a74ab75 thread_start + 13

Thread 4:
0 libsystem_kernel.dylib 0x00007fff9292c192 __workq_kernreturn + 10
1 libsystem_c.dylib 0x00007fff8a749594 _pthread_wqthread + 758
2 libsystem_c.dylib 0x00007fff8a74ab85 start_wqthread + 13

Thread 5:
0 libsystem_kernel.dylib 0x00007fff9292c192 __workq_kernreturn + 10
1 libsystem_c.dylib 0x00007fff8a749594 _pthread_wqthread + 758
2 libsystem_c.dylib 0x00007fff8a74ab85 start_wqthread + 13

Thread 6:
0 libsystem_kernel.dylib 0x00007fff9292c192 __workq_kernreturn + 10
1 libsystem_c.dylib 0x00007fff8a749594 _pthread_wqthread + 758
2 libsystem_c.dylib 0x00007fff8a74ab85 start_wqthread + 13

Thread 7:

Thread 0 crashed with X86 Thread State (64-bit):
rax: 0x0000000105e61650 rbx: 0x000000011866f3c0 rcx: 0x000000000000005e rdx: 0x0000000000000000
rdi: 0x0000000000000000 rsi: 0x0000000000000000 rbp: 0x00007fff5fbff910 rsp: 0x00007fff5fbff8f0
r8: 0x000000000000001f r9: 0x000000011865b260 r10: 0x0000000020cda8d8 r11: 0x0000000105d43570
r12: 0x000000010bc622b8 r13: 0x000000010bc415d0 r14: 0x000000010bc62190 r15: 0x000000010bc62220
rip: 0x0000000105d4358f rfl: 0x0000000000010206 cr2: 0x0000000000000000
Logical CPU: 4

Jc: Looking at the backtrace, the crash seems to occur within the destrcutor of ae vtkMRMLSceneViewNode.

Since the problem still occur in today nightly and nicole commit [1] has been integrated before the nightly started (30 hrs ago) [2], I believe the problem is not fixed.

[1] http://viewvc.slicer.org/viewvc.cgi/Slicer4?view=revision&revision=20043
[2] http://slicer.cdash.org/viewUpdate.php?buildid=18582

Couldn't reproduce crash/leak on r20102

I still see crashes on exit. Just less frequently. I will report as a new issue, once I have something.