[vtk-developers] Dashboard failures on Londinium

[Berk Geveci]

As another data point, this happens with GCC 4.8 and 4.6.3 but does not
happen with clang on the same machine. Neither does it happen on Utkarsh's
machine which happen to have GCC 4.7.2. Another interesting point is that
the loader seems to binding symbols somewhat differently between Utkarsh's
machine and the other ones where the example crashes. If we use LD_DEBUG on
Utkarsh's machine, we get the following for pthread_create resolution:

    10888: binding file /usr/lib/nvidia-current/libGL.so.1 [0] to
/lib/x86_64-linux-gnu/libpthread.so.0 [0]: normal symbol
`pthread_create' [GLIBC_2.2.5]
     10888: binding file /lib/x86_64-linux-gnu/libpthread.so.0 [0] to
/lib/x86_64-linux-gnu/libpthread.so.0 [0]: normal symbol
`pthread_create' [GLIBC_2.2.5]
     10888: binding file /lib/x86_64-linux-gnu/libpthread.so.0 [0] to
/lib/x86_64-linux-gnu/libpthread.so.0 [0]: normal symbol
`pthread_create' [GLIBC_2.2.5]
     10888: binding file
/home/utkarsh/Kitware/ParaView3/ParaViewBin/lib/libvtkCommonCore-pv3.98.so.1
[0] to /lib/x86_64-linux-gnu/libpthread.so.0 [0]: normal symbol
`pthread_create' [GLIBC_2.2.5]

Where on Londinium, the last line:

     10888: binding file
/home/utkarsh/Kitware/ParaView3/ParaViewBin/lib/libvtkCommonCore-pv3.98.so.1
[0] to /lib/x86_64-linux-gnu/libpthread.so.0 [0]: normal symbol
`pthread_create' [GLIBC_2.2.5]

is missing. So it is not showing how the loader is resolving pthread_create
for libvtkCommonCore which happens to contain vtkMultiThreader and the last
VTK-level entry point of the stack at crash time.

I am still stumped and would welcome suggestions dearly.

-berk



On Wed, Apr 24, 2013 at 11:13 AM, Berk Geveci <berk.geveci at kitware.com>wrote:

> I'll check other compilers but this is reproducible even with a debug
> build.
>
>
> On Wed, Apr 24, 2013 at 10:48 AM, Sean McBride <sean at rogue-research.com>wrote:
>
>> On Wed, 24 Apr 2013 10:13:17 -0400, Berk Geveci said:
>>
>> >I would be grateful if anyone had any ideas.
>>
>> I notice it uses the newest gcc (4.8) and is building in release.
>>  Compiler optimizers are getting more and more devious these days... does
>> the bug repro at lower optimization levels?  Does it repro if you build
>> with another compiler, say clang?
>>
>> Cheers,
>>
>> --
>> ____________________________________________________________
>> Sean McBride, B. Eng                 sean at rogue-research.com
>> Rogue Research                        www.rogue-research.com
>> Mac Software Developer              Montréal, Québec, Canada
>>
>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/vtk-developers/attachments/20130501/3714ac98/attachment.html>