MantisBT - ParaView
View Issue Details
IDProjectCategoryView StatusDate SubmittedLast Update
0012418ParaView(No Category)public2011-07-28 12:402011-08-25 17:11
ReporterKyle Lutz 
Assigned ToKyle Lutz 
PrioritynormalSeverityminorReproducibilityhave not tried
StatusclosedResolutionfixed 
PlatformOSOS Version
Product Version 
Target VersionFixed in Version3.12 
ProjectTBD
Topic Name12418-fix-quadric-clustering-segfault
Typecrash
Summary0012418: Segfault in vtkQuadricClustering::AddEdges()
DescriptionOn Wed, Jun 29, 2011 at 11:21 AM, Dennis McWherter <dmcwhe2@illinois.edu> wrote:
> Hello,
>
> I have generated a Legacy VTK file using VTK's Python wrapper. This file
> crashes ParaView 3.10.1 (Windows 7) when attempting to rotate it in
> ParaView's viewport, however, it seems that I can appropriately apply
> filters and such. However, this data did work properly on ParaView 3.8.1.
>
> If anyone can help in inspecting the file for potential bugs, it can be
> found here (~55MB):
> https://netfiles.uiuc.edu/xythoswfs/webview/_xy-22178502_2 [^]
>
> ParaView crashes in vtkQuadricClustering::AddEdges at this line:
>
> points->GetPoint(ptIds[0], pt0);
>
> ptIds is NULL.
>
> The data itself is a PolyData object with PolyLines and arrays. I presume
> there may be a bad/missing data structure within the data, can any ParaView
> expert shed some light on the issue?
>
> Regards,
> Dennis McWherter
Steps To Reproduce
Additional Information
TagsNo tags attached.
Relationships
Attached Files
Issue History
Date ModifiedUsernameFieldChange
2011-07-28 12:40Kyle LutzNew Issue
2011-07-28 12:51Kyle LutzAssigned To => Kyle Lutz
2011-07-28 12:51Kyle LutzStatusbacklog => todo
2011-08-03 13:17Kyle LutzNote Added: 0027214
2011-08-03 14:21Kyle LutzStatustodo => active development
2011-08-03 14:35Kyle LutzTopic Name => 12418-fix-quadric-clustering-segfault
2011-08-03 14:35Kyle LutzNote Added: 0027219
2011-08-03 14:35Kyle LutzStatusactive development => gatekeeper review
2011-08-03 14:35Kyle LutzFixed in Version => 3.12
2011-08-03 14:35Kyle LutzResolutionopen => fixed
2011-08-04 18:08Utkarsh AyachitNote Added: 0027232
2011-08-05 11:32Utkarsh AyachitNote Added: 0027270
2011-08-12 14:23Utkarsh AyachitStatusgatekeeper review => customer review
2011-08-12 14:23Utkarsh AyachitNote Added: 0027356
2011-08-25 17:11Kyle LutzStatuscustomer review => closed

Notes
(0027214)
Kyle Lutz   
2011-08-03 13:17   
Bisected:

commit b0391d81669b40f31edd4cf447c413f9d75d290a
Author: François Bertel <francois.bertel@kitware.com>
Date: Wed Jul 14 10:31:30 2010 -0400

    BUG:Fixed vtkCellArray::GetNextCell when traversal is done.
    
    npts and pts were not initialized to 0. vtkVRMLImporter is
    using GetNextCell directly without checking for the return value
    and then has a loop involving npts. This also fixes a warning about
    the fact that npts may be used uninitialized in vtkVRMLImporter.

diff --git a/Filtering/vtkCellArray.h b/Filtering/vtkCellArray.h
index 7593639..283aadf 100644
--- a/Filtering/vtkCellArray.h
+++ b/Filtering/vtkCellArray.h
@@ -313,6 +313,8 @@ inline int vtkCellArray::GetNextCell(vtkIdType& npts, vtkIdType* &pts)
     this->TraversalLocation += npts;
     return 1;
     }
+ npts=0;
+ pts=0;
   return 0;
 }
(0027219)
Kyle Lutz   
2011-08-03 14:35   
Merge topic '12418-fix-quadric-clustering-segfault-next' into next

c750580 Merge topic '11580_fix_extract_surface_subdivision-next' into 12418-fix-quadric-clustering-segfault
8954727 Update VTK to bring in '12418-fix-quadric-clustering-segfault'


Merge topic '12418-fix-quadric-clustering-segfault' into pv-next

58414a3 Check for a non-empty cell before using its points
(0027232)
Utkarsh Ayachit   
2011-08-04 18:08   
Developer note: Issue slated for gatekeeper review on Aug 7, 2 PM EDT.
(0027270)
Utkarsh Ayachit   
2011-08-05 11:32   
verified.
(0027356)
Utkarsh Ayachit   
2011-08-12 14:23   
Merged into git-master