Notes |
|
(0015132)
|
Utkarsh Ayachit
|
2009-02-18 09:08
|
|
Another test:
* 2 servers
* simply create sphere and apply Clean. The server segfaults. |
|
|
(0016231)
|
Utkarsh Ayachit
|
2009-04-28 17:08
|
|
Berk,
Does this patch make sense:
Index: VTK/Graphics/vtkCleanPolyData.cxx
===================================================================
RCS file: /cvsroot/ParaView3/ParaView3/VTK/Graphics/vtkCleanPolyData.cxx,v
retrieving revision 1.79
diff -u -3 -p -r1.79 vtkCleanPolyData.cxx
--- VTK/Graphics/vtkCleanPolyData.cxx 27 Apr 2009 17:25:48 -0000 1.79
+++ VTK/Graphics/vtkCleanPolyData.cxx 28 Apr 2009 21:05:24 -0000
@@ -116,7 +116,7 @@ int vtkCleanPolyData::RequestUpdateExten
}
else
{
- inInfo->Set(vtkStreamingDemandDrivenPipeline::UPDATE_PIECE_NUMBER(), -1);
+ inInfo->Set(vtkStreamingDemandDrivenPipeline::UPDATE_PIECE_NUMBER(), 0);
inInfo->Set(vtkStreamingDemandDrivenPipeline::UPDATE_NUMBER_OF_PIECES(),
0);
}
The problem is that the original code is setting UPDATE_PIECE_NUMBER to -1 which is < 0 and hence filters still try to execute and then seg fault with divide by 0 or such when they try to produce the -1th piece of 0 pieces. |
|
|
(0016462)
|
Utkarsh Ayachit
|
2009-05-14 21:27
|
|
Reminder sent to: Berk Geveci Berk,
Does the patch look good. I guess my question is what's the recommend way of telling the input that I don't want any piece at all. |
|
|
(0016475)
|
Utkarsh Ayachit
|
2009-05-15 09:49
|
|
/cvsroot/ParaView3/ParaView3/VTK/Graphics/vtkCleanPolyData.cxx,v <-- VTK/Graphics/vtkCleanPolyData.cxx
new revision: 1.80; previous revision: 1.79 |
|
|
(0016579)
|
Alan Scott
|
2009-05-26 21:19
|
|
|