[vtk-developers] backtrace of segfault in Charts-TestMultipleScalarsToColors on bubbles OSMesa build

[Marcus D. Hanwell]

On Wed, Oct 27, 2010 at 1:41 PM, Kevin H. Hobbs <hobbsk at ohiou.edu> wrote:
> On 10/20/2010 09:46 AM, Marcus D. Hanwell wrote:
>>
>> That looks like an OSMesa issue, querying for supported OpenGL
>> extensions shouldn't cause a segfault. I will add this to my list, and
>> see if I can possibly trace into OSMesa. It would be nice to get this
>> working, but at first glance this is just due to us checking for
>> OpenGL 1.2.
>>
>
> Brian Paul from Mesa said,
>
> "This reminds me of someone else who reported a bug in which calling
> glGetString() without a current context caused a crash, and I think it
> was x86-64 only.  Maybe that's the issue."
>
> The thread in the Mesa developers list starts here :
>
> http://lists.freedesktop.org/archives/mesa-dev/2010-October/003679.html
>
Thanks Kevin. I don't know of anywhere else where this is failing, but
thinking about it I don't think we should be calling
ExtensionSupported until we have an OpenGL context. If you look in the
TestGLSL.cxx test (in the same directory) I am very careful to test
for OpenGL features after we get a valid context.

I should probably fix this up, and I am not sure if the Mesa folks
will want to fix up the segfault?

Marcus