View Issue Details
ID | Project | Category | View Status | Date Submitted | Last Update |
---|---|---|---|---|---|
0001537 | Slicer4 | Core: Base Code | public | 2011-11-10 14:24 | 2013-02-12 09:40 |
Reporter | pohl | Assigned To | jcfr | ||
Priority | high | Severity | crash | Reproducibility | always |
Status | closed | Resolution | fixed | ||
Platform | Linux | OS | Ubuntu | OS Version | 11.10 |
Product Version | |||||
Target Version | Slicer 4.3.0 | Fixed in Version | Slicer 4.3.0 | ||
Summary | 0001537: Exit from tcl adapter causes seg fault when exiting program | ||||
Description | JC helped me create the attached test, which exits with a seg fault. JC wanted to integrate the code as base function in Slicer (minus the EMSegmenter specific part), debug it, and include the test in to the base | ||||
Tags | No tags attached. | ||||
Thanks for the remainder. Will be working on that issue on Monday. |
|
can you then update Tcl_Interp* vtkSlicerCommonInterface::Startup accordingly ? |
|
I started to refactor the code so that the overall process to execute python from a standalone executable like EMSegmentCommandLine is simplified. On the other hand, I couldn;t reproduce the error. Would it be possible to provide me with a stack trace ? |
|
I am trying to run it but all my tests now simply hang : [14] [11:14am] [sbia-pc82 /software/Slicer4/Slicer4-SuperBuild/Slicer-build/QTModules] >ctest -VV -R vtkSlicerCommonInterfaceStartup 66: Test command: /software/Slicer4/Slicer4-SuperBuild/Slicer-build/Slicer "--launcher-no-splash" "--launch" "/software/Slicer4/Slicer4-SuperBuild/Slicer-build/bin/vtkSlicerCommonInterfaceStartupTest" nothing happens afterwards |
|
Could you either provide the version of "vtkSlicerCommonInterfaceStartup" you are using or commit the test into EMSegment repository ? Adding the test into EMSegment/Testing and updating EMSegment/CMakeLists.txt, after re-configuring and re-compiling, the following tests pass: jchris@karakoram:~/Projects/Slicer4-Superbuild-Debug/Slicer-build $ ctest -R vtkSlicerCommonInterface 100% tests passed, 0 tests failed out of 2 Total Test time (real) = 3.23 sec |
|
Following 16910, vtkSlicerCommonInterfaceStartupTest can be simplified. The two "HACK" where mainContext and updatePythonOsEnviron are called can be removed. See also https://github.com/Slicer/Slicer/commit/72231b6e9880b3ad1ffc93d415e937bcaf192ed6 |
|
Now I get to ctest -VV vtkSlicerCommonInterfaceStartupTest 1: Test command: /software/Slicer4/Slicer4-SuperBuild/Slicer-build/Slicer "--launcher-no-splash" "--launch" "/software/Slicer4/Slicer4-SuperBuild/Slicer-build/bin/ColorsCxxTests" "qSlicerColorsModuleWidgetTest1" a window shortly pops up and nothing happens afterwards |
|
How is progress with this? |
|
2011-12-08 06:10
|
vtkSlicerCommonInterfaceStartupTest.cxx (2,625 bytes) |
A test named qSlicerApplicationTpyclEMSegmentIntegrationTest has been added. See https://github.com/Slicer/Slicer/commit/bb82ef199e72665a4955dc330c39de9dbb3fec20 The test passes on Ubuntu 10.04 / gcc (Ubuntu 4.4.3-4ubuntu5) |
|
Linking issue on windows should now be fixed. |
|
I still have the same issue - how should I debug it ? Linux sbia-pc82 3.0.0-14-generic 0000023-Ubuntu SMP Mon Nov 21 20:28:43 UTC 2011 x86_64 x86_64 x86_64 GNU/Linux [5] [9:10pm] [sbia-pc82 /data/software/Slicer4-2011-12-15-18938/Slicer4-SuperBuild/Slicer-build] >ctest -VV -R qSlicerApplicationTpyclEMSegmentIntegrationTest 466: Test command: /data/software/Slicer4-2011-12-15-18938/Slicer4-SuperBuild/Slicer-build/Slicer "--launcher-no-splash" "--launch" "/data/software/Slicer4-2011-12-15-18938/Slicer4-SuperBuild/Slicer-build/bin/SlicerQTCxxTests" "qSlicerApplicationTpyclEMSegmentIntegrationTest" "/data/software/Slicer4-2011-12-15-18938/Slicer4-SuperBuild/Slicer-build/lib/Slicer-4.0/cli-modules/EMSegmentCommandLine" 0% tests passed, 1 tests failed out of 1 Total Test time (real) = 2063.37 sec The following tests FAILED: |
|
Using the debugger could be useful. We need to find out why the test is timing out |
|
As of Dec 20 2011, the test "qSlicerApplicationTpyclEMSegmentIntegrationTest" passes on the following platform:
|
|
I followed the instructions on that web site but it did not help a lot - instead I executed the following : [12] [9:34pm] [sbia-pc82 /software/Slicer4/Slicer4-SuperBuild/Slicer-build] >gdb /software/Slicer4/Slicer4-SuperBuild/Slicer-build/bin/vtkSlicerCommonInterfaceStartupTest Program received signal SIGABRT, Aborted. (gdb) backtrace Any help ? |
|
http://forum.dcmtk.org/viewtopic.php?t=1119 Somewhere in your test, you might want to add "OFCondition cond = EC_Normal;" |
|
After preventing the DICOM related modules from loading, it resumes the issue is related to FD message handler. The following backtrace has been obtained by attaching gdb to the SlicerQTCxxTests. (gdb) backtrace |
|
When executing make SlicerQTCxxTests && ctest -VV -R EMSegmentInt with then the test passes. if we include then the test crashes 481: pure virtual method called with the following backtrace: Program received signal SIGABRT, Aborted. which is due to dcmtk |
|
To pass it disable DICOM (temporary solution)
then the above test passes |
|
if comment out //this->ErrorLogModel->registerMsgHandler(new ctkErrorLogFDMessageHandler); in Slicer4/Base/QTCore/qSlicerCoreApplication.cxx then the entire test passes |
|
If I execute the test in release mode same bug comes up test 481 481: Test command: /data/software/Slicer4-2011-12-15-18938/Slicer4-SuperBuild/Slicer-build/Slicer "--launcher-no-splash" "--launch" "/data/software/Slicer4-2011-12-15-18938/Slicer4-SuperBuild/Slicer-build/bin/SlicerQTCxxTests" "qSlicerApplicationTpyclEMSegmentIntegrationTest" "/data/software/Slicer4-2011-12-15-18938/Slicer4-SuperBuild/Slicer-build/lib/Slicer-4.0/cli-modules/EMSegmentCommandLine" now it hangs |
|
The OFCondition crash at exit is due to: I tried to reproduce the error in CTK only by running:
|
|
I do not understand the entry completely - it does not seem to replicate the bug - correct ? |
|
In CTK, ctkSimplePythonShell reproduces the OFCondition crash at exit (pure virtual method called). |
|
Ok, I can now reproduce in CTK without the need of using python. |
|
From Julien: With what test do you have the crash ? When I fix the DICOM error, I don't have the msghandler problem: then all the tests but 2 pass. |
|
so I checked out Slicer edited Slicer-Superbuild/CTK/Libs/Scripting/Python/Core/Python/ctk/init.py.in I changed try: to build CTK and SLICER and executed ctest -VV -R qSlicerApplicationTpyclEMSegmentIntegrationTest The test still gets stuck - nothing has changed. Is that also true for you ? If yours is passing do you have the 32 or 64 bit version of ubuntu installed ? Linux sbia-pc82 3.0.0-15-generic 0000026-Ubuntu SMP Fri Jan 20 17:23:00 UTC 2012 x86_64 x86_64 x86_64 GNU/Linux |
|
Could you make sure the file "CTK-build/CTK-build/bin/Python/ctk/init.py" still contains the modification after rebuilding ? If not, I would suggest to directly update the file "CTK-build/CTK-build/bin/Python/ctk/init.py" |
|
it does |
|
2012-02-22 13:53
|
|
The problem with ctkErrorLogFDMessageHandler seems to be resolved thanks to JC . We now still have the issue with DCMTK. So If we run Slicer without DCMTK by Additional notes from JC: Each time you will update EMSegment code, make sure to either just rebuild the corresponding target in Slicer or remove the DICOM2FullBrainTractography.py and DICOM.py again. |
|
Reminder sent to: finetjul, pieper Steve, Julien, I changed the target version to 4.2.0 (Summer AHM), does that sounds reasonable. It's now too long we are keeping that issue open. I also created issue 0002006 allowing to track the update of DCMTK version. |
|
Now when I 1) removing DICOMWidgets' and 'DICOMCore' from CTK-build/CTK-build/bin/Python/ctk/init.py ctest qSlicerApplicationTpyclEMSegmentIntegrationTest fails |
|
If the stack trace still mention OFCondition .. it means you failed to remove / rebuild without DCMTK. Could you check which stack trace you obtain ? See http://www.na-mic.org/Bug/view.php?id=1537#c3527 FYI - Few days ago, an option named Slicer_BUILD_DICOM_SUPPORT has been added. You could probably do a clean build in a different folder disabling this option. |
|
I confirm that the test does not fail anymore after performing a clean build with Slicer_BUILD_DICOM_SUPPORT disabled |
|
Either |
|
This should be fixed with the recent update made at the AHM2013 when working on the ITKv4 integration. See http://viewvc.slicer.org/viewvc.cgi/Slicer4?view=revision&revision=21602 |
|
Date Modified | Username | Field | Change |
---|---|---|---|
2011-11-10 14:24 | pohl | New Issue | |
2011-11-10 14:24 | pohl | Status | new => assigned |
2011-11-10 14:24 | pohl | Assigned To | => jcfr |
2011-11-10 14:24 | pohl | File Added: vtkSlicerCommonInterfaceStartupTest.cxx | |
2011-11-10 14:58 | jcfr | Note Added: 0003304 | |
2011-11-12 11:23 | pohl | Note Added: 0003311 | |
2011-11-17 16:42 | jcfr | Note Added: 0003328 | |
2011-11-24 06:15 | pohl | Note Added: 0003380 | |
2011-11-24 12:13 | jcfr | Note Added: 0003382 | |
2011-11-24 12:37 | jcfr | Note Added: 0003383 | |
2011-11-28 10:31 | pohl | Note Added: 0003404 | |
2011-12-08 05:28 | kikinis | Note Added: 0003429 | |
2011-12-08 06:10 | pohl | File Deleted: vtkSlicerCommonInterfaceStartupTest.cxx | |
2011-12-08 06:10 | pohl | File Added: vtkSlicerCommonInterfaceStartupTest.cxx | |
2011-12-13 15:41 | jcfr | Note Added: 0003444 | |
2011-12-14 05:57 | jcfr | Note Added: 0003446 | |
2011-12-15 18:33 | pohl | Note Added: 0003455 | |
2011-12-19 07:23 | jcfr | Note Added: 0003462 | |
2011-12-19 07:25 | jcfr | Note Edited: 0003462 | |
2011-12-20 06:04 | jcfr | Note Added: 0003468 | |
2012-01-04 16:40 | pohl | Note Added: 0003485 | |
2012-01-04 22:02 | finetjul | Note Added: 0003490 | |
2012-01-04 22:25 | finetjul | Note Edited: 0003490 | |
2012-01-11 11:17 | jcfr | Note Added: 0003496 | |
2012-01-12 05:56 | pohl | Note Added: 0003497 | |
2012-01-12 06:22 | pohl | Note Added: 0003498 | |
2012-01-12 06:54 | pohl | Note Added: 0003499 | |
2012-01-14 15:50 | pohl | Note Added: 0003500 | |
2012-01-19 06:15 | finetjul | Note Added: 0003505 | |
2012-01-19 13:31 | pohl | Note Added: 0003508 | |
2012-01-23 19:24 | finetjul | Note Added: 0003520 | |
2012-01-24 08:41 | finetjul | Note Added: 0003523 | |
2012-01-25 11:33 | jcfr | Note Added: 0003527 | |
2012-01-25 11:34 | jcfr | Assigned To | jcfr => pohl |
2012-01-25 11:34 | jcfr | Status | assigned => feedback |
2012-01-29 05:34 | pohl | Note Added: 0003570 | |
2012-01-30 06:33 | jcfr | Note Added: 0003571 | |
2012-01-30 18:17 | pohl | Note Added: 0003573 | |
2012-02-02 04:38 | finetjul | Status | feedback => assigned |
2012-02-22 13:53 | jcfr | File Added: Slicer-ctkErrorLogFDMessageHandler-hang.png | |
2012-02-23 10:07 | pohl | Note Added: 0003706 | |
2012-04-04 22:52 | pohl | Assigned To | pohl => finetjul |
2012-05-09 09:25 | jcfr | Target Version | Slicer 4.0.0 => Slicer 4.2.0 AHM Summer 2012 |
2012-05-09 09:29 | jcfr | Note Added: 0004237 | |
2012-07-24 11:07 | pohl | Note Added: 0005223 | |
2012-07-24 11:13 | jcfr | Note Added: 0005224 | |
2012-07-24 11:14 | jcfr | Note Edited: 0005224 | |
2012-07-24 13:24 | pohl | Note Added: 0005225 | |
2012-07-30 05:28 | pieper | Relationship added | related to 0002204 |
2012-08-14 12:55 | finetjul | Assigned To | finetjul => jcfr |
2012-08-14 12:56 | finetjul | Note Added: 0005550 | |
2012-08-20 11:37 | jcfr | Target Version | Slicer 4.2.0 - Feature freeze Sept 1st 2012 => Slicer 4.2.5 |
2012-08-21 09:39 | jcfr | Target Version | Slicer 4.2.5 => Slicer 4.3.0 |
2013-01-17 13:41 | jcfr | Note Added: 0007703 | |
2013-01-17 13:41 | jcfr | Status | assigned => resolved |
2013-01-17 13:41 | jcfr | Fixed in Version | => Slicer 4.3.0 |
2013-01-17 13:41 | jcfr | Resolution | open => fixed |
2013-01-17 14:38 | pohl | Status | resolved => closed |
2013-01-17 15:38 | jcfr | Relationship added | related to 0002459 |