MantisBT - ParaView
View Issue Details
IDProjectCategoryView StatusDate SubmittedLast Update
0013316ParaView(No Category)public2012-07-20 05:242012-10-29 17:04
ReporterIan Curington 
Assigned ToKyle Lutz 
PriorityhighSeverityminorReproducibilityhave not tried
StatusclosedResolutionfixed 
PlatformOSOS Version
Product Version3.14.1 
Target VersionFixed in Version3.98.0 
ProjectTBD
Topic Name
Typecrash
Summary0013316: pvpython crash on exit when including programmable source
Descriptionpvpython script demonstrates crash on completion due to memory free bug, particularly when script also includes use of a python programmable filter.

How to reproduce test case:
1. Start ParaView
2. Sources->Programmable Source
3. In script area pase: pdo = self.GetPolyDataOutput()
4. Press apply button
5. Press close button in "Output Messages" dialog if it appears
6. File->Save State
7. Save state with name pvpython_crash.pvsm
8. exit

crash happens on:
pvpython pvpython_crash.py
but not on
python pvpython_crash.py.

Crash place is after all lines in pvpython_crash.py are executed.

If comment line #12 "view.WriteImage("plot.png", "vtkPNGWriter", 1)" in pvpython_crash.py, then crash will gone.

crash place is:
705 if (!PyString_CheckExact(key) ||
706 (hash = ((PyStringObject *) key)->ob_shash) == -1)

Issue at http://paraview.org/Bug/print_bug_page.php?bug_id=10506 [^] 0010506 can be related.

pvpython_crash_gdb_backtrace.txt contains gdb backtrace at crash place
pvpython_crash_valgrind.txt containes valgrind output near crash place


Crash happens on more complex Programmable Filters too:
 for intance on http://www.paraview.org/Wiki/Python_Programmable_Filter#Generating_Data_.28Programmable_Source.29 [^]
__
environment:

CentOS release 5.8 (Final) with latest updates, x64
Python-2.7.3, built with "./configure --enable-shared && make"
ParaView-3-14.1, built with "BUILD_SHARED_LIBS=ON, PARAVIEW_ENABLE_PYTHON=ON, other options are default"

SEE: attached pvpython_crash_case_29june2012.zip for test case files.

Steps To Reproduce
Additional Information
TagsNo tags attached.
Relationships
duplicate of 0013222closed Kyle Lutz Python Programmable Filter Test Fails 
Attached Fileszip pvpython_crash_case_29june2012.zip (3,937) 2012-07-20 05:24
https://www.vtk.org/Bug/file/9234/pvpython_crash_case_29june2012.zip
Issue History
Date ModifiedUsernameFieldChange
2012-07-20 05:24Ian CuringtonNew Issue
2012-07-20 05:24Ian CuringtonFile Added: pvpython_crash_case_29june2012.zip
2012-07-24 14:15Utkarsh AyachitRelationship addedrelated to 0013222
2012-07-24 14:16Utkarsh AyachitNote Added: 0028909
2012-07-24 14:16Utkarsh AyachitRelationship replacedduplicate of 0013222
2012-07-24 14:16Utkarsh AyachitStatusbacklog => gatekeeper review
2012-07-24 14:16Utkarsh AyachitFixed in Version => git-master
2012-07-24 14:16Utkarsh AyachitResolutionopen => fixed
2012-07-24 14:16Utkarsh AyachitAssigned To => Kyle Lutz
2012-07-24 14:16Utkarsh AyachitStatusgatekeeper review => customer review
2012-07-25 14:09Alan ScottNote Added: 0028926
2012-07-25 14:09Alan ScottStatuscustomer review => closed
2012-10-29 17:04Utkarsh AyachitFixed in Versiongit-master => 3.98.0

Notes
(0028909)
Utkarsh Ayachit   
2012-07-24 14:16   
Cannot reproduce. Maybe related to fixed bug 0013222
(0028926)
Alan Scott   
2012-07-25 14:09   
I also tried to reproduce this crash, and failed. Linux, master. I used the "How to reproduce test case" instructions, creating a .pvsm file. I then read this file in as a state file to pvpython. Please reopen and explain what we are doing wrong if the bug still exists.