[vtkusers] vtkIntersectionPolyDataFilter::TriangleTriangleIntersection corrupts the stack?

[James Johnston]

Cory,

Sorry for the delay - been busy lately!

Your patch is identical to the workaround I implemented to move on in my
project.  Given that: (1) the rest of the code seemed to be written with the
expectation that up to 2 elements would be stored in the arrays, (2) my
workaround was working for my particular application but I wasn't sure about
other applications, (3) and that I wasn't trying to understand how the
function worked, then I wasn't sure about pushing that change myself since I
didn't know if more needed to be changed to the function to ensure
reasonable behavior in existing applications.

The patch does show that the code itself is at least guaranteed not to
exhibit undefined behavior by killing the stack - it's easy to see this from
a quick reading.  It works for me & my application, anyway...

Best regards,

James Johnston

> -----Original Message-----
> From: Cory Quammen [mailto:cquammen at cs.unc.edu]
> Sent: Monday, October 08, 2012 20:40
> To: James Johnston
> Cc: vtkusers at vtk.org
> Subject: Re: [vtkusers]
> vtkIntersectionPolyDataFilter::TriangleTriangleIntersection corrupts the
stack?
> 
> James,
> 
> This should be fixed in the master branch of VTK. Could you give it a try?
> 
> Thanks,
> Cory