[vtk-developers] RE: Help with vtkLineRepresentationTest1 test
David Cole
dlrdave at aol.com
Wed Mar 20 17:15:25 EDT 2013
There’s already this code:
if (size == NULL)
{
return;
}
Why not just add another similar block?
if (size[0] == 0 && size[1] == 0)
{
return;
}
I bet all of those tests that assert go through that same code again later after the window really does have a size, and the correct result is used way after this “failure” is presently occurring...
D
From: Sean McBride
Sent: March 20, 2013 5:04 PM
To: vtk-developers at vtk.org
Subject: [vtk-developers] Help with vtkLineRepresentationTest1 test
Hi all,
The test 'vtkInteractionWidgetsCxx-vtkLineRepresentationTest1' divides by 0, thus creating a (double) NaN, which it then later tries to convert to float, which is undefined behaviour[1].
The divide-by-zero happens in vtkViewport::DisplayToView() because size[0] and size[1] are zero, which I'm thinking they probably shouldn't be. Backtrace:
#0 vtkViewport::DisplayToView (this=0x10ff51fc0) at VTK/Rendering/Core/vtkViewport.cxx:209
#1 vtkViewport::DisplayToWorld (this=0x10ff51fc0) at vtkViewport.h:169
#2 vtkInteractorObserver::ComputeDisplayToWorld (ren=0x10ff51fc0, x=-2.5, y=-2.5, z=0.99000990000000011, worldPt=0x7fff5fbfba40) at VTK/Rendering/Core/vtkInteractorObserver.cxx:268
#3 vtkWidgetRepresentation::SizeHandlesInPixels (this=0x10ff21900, factor=1.3500000000000001, pos=0x10ff37a98) at VTK/Interaction/Widgets/vtkWidgetRepresentation.cxx:194
#4 vtkLineRepresentation::SizeHandles (this=0x10ff21900) at VTK/Interaction/Widgets/vtkLineRepresentation.cxx:673
#5 vtkLineRepresentation::BuildRepresentation (this=0x10ff21900) at VTK/Interaction/Widgets/vtkLineRepresentation.cxx:738
#6 vtkLineRepresentationTest1 () at VTK/Interaction/Widgets/Testing/Cxx/vtkLineRepresentationTest1.cxx:21
#7 main (ac=7, av=0x7fff5fbff7b0) at VTK-big-bin/Interaction/Widgets/Testing/Cxx/CxxTests.cxx:587
You can easily reproduce this with any compiler if you add:
assert(sizex > 0);
assert(sizey > 0);
to vtkViewport::DisplayToView(). In fact doing so results in 6 different tests asserting.
I don't know how to address this... any thoughts?
[1] or at least sorta undefined:
<http://lists.cs.uiuc.edu/pipermail/cfe-dev/2013-March/028596.html>
Thanks,
--
____________________________________________________________
Sean McBride, B. Eng sean at rogue-research.com
Rogue Research www.rogue-research.com
Mac Software Developer Montréal, Québec, Canada
_______________________________________________
Powered by www.kitware.com
Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html
Follow this link to subscribe/unsubscribe:
http://www.vtk.org/mailman/listinfo/vtk-developers
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/vtk-developers/attachments/20130320/f079624b/attachment.html>
More information about the vtk-developers
mailing list