Crash of normalInelasticity scripts (and others ?)

Crash in normalinelasticity test caused by ForceContainer::reset, namely the lines :

memset(&_force [0], 0,sizeof(Vector3r)*size);
memset(&_torque[0], 0,sizeof(Vector3r)*size);

Hi,
could you, please, provide a script, which leads to a crash?
Thanks

Anton

On Thu, Apr 28, 2011 at 11:11 AM, jduriez <email address hidden> wrote:
> Crash in normalinelasticity test caused by ForceContainer::reset, namely
> the lines :
>
> memset(&_force [0], 0,sizeof(Vector3r)*size);
> memset(&_torque[0], 0,sizeof(Vector3r)*size);
>
> --
> You received this bug notification because you are subscribed to Yade.
> https://bugs.launchpad.net/bugs/739429
>
> Title:
>  SimpleShear Preprocessor to update
>
> Status in Yet Another Dynamic Engine:
>  New
>
> Bug description:
>  Changes to commit about this Preprocessor. And (new ?) problems with
>  normalinelasticity test, leading to crash, at O. run()
>

scripts/normalInelasticity-test.py which is in trunk. The crash is caused by line 65 : O.run(2,True)

As said before, in the iteration, the problem appears in ForceContainer::reset ? I'm trying to get more info.
Launching in debug did not help me : no output specially related to "my" files.

Thanks

A piece of the gdb infos at the crash. Points out ForceContainer::reset, but I do not see really what to do. Trying other scripts..

Thread 2 (Thread 0x7f0bdcf01700 (LWP 31405)):
#0 0x00007f0bf99693ad in waitpid () from /lib/libc.so.6
#1 0x00007f0bf9907299 in ?? () from /lib/libc.so.6
#2 0x00007f0bf99075d0 in system () from /lib/libc.so.6
#3 0x00007f0bf8a5b0da in crashHandler (sig=11) at core/main/pyboot.cpp:45
#4 <signal handler called>
#5 0x00007f0bedc10c52 in Eigen::MatrixBase<Eigen::Matrix<double, 3, 1, 2, 3, 1> >::operator[] (this=0x0, index=2) at /usr/include/eigen2/Eigen/src/Core/Coeffs.h:134
#6 0x00007f0bedc2a6e1 in Eigen::MatrixBase<Eigen::Matrix<double, 3, 1, 2, 3, 1> >::z (this=0x0) at /usr/include/eigen2/Eigen/src/Core/Coeffs.h:219
#7 0x00007f0bedc1352c in operator<< <double> (os=..., v=...) at /home/3S-LAB/jduriez/build-cluster-cluster/dbg/include/yade/lib/base/Math.hpp:98
#8 0x00007f0bee674a67 in ForceContainer::reset (this=0x25dbb28, iter=0) at /home/3S-LAB/jduriez/build-cluster-cluster/dbg/include/yade/core/ForceContainer.hpp:166
#9 0x00007f0bee667263 in ForceResetter::action (this=0x38faf10) at /home/3S-LAB/jduriez/yadeTrunk/pkg/common/ForceResetter.cpp:8
#10 0x00007f0bf542aac3 in Scene::moveToNextTimeStep (this=0x25dbb20) at /home/3S-LAB/jduriez/yadeTrunk/core/Scene.cpp:97
#11 0x00007f0bf5431a8b in SimulationFlow::singleAction (this=0x25ad630) at /home/3S-LAB/jduriez/yadeTrunk/core/SimulationFlow.cpp:20

In fact scripts/test/checks/checkWeight.py crashes also with the same manner... (with r2812)
Something on my computer ?

Sorry, difficult to say for the moment.
check script performs at each commit and it does not give a crash...

Anton

On Fri, Apr 29, 2011 at 3:03 PM, jduriez <email address hidden> wrote:
> In fact scripts/test/checks/checkWeight.py crashes also with the same manner... (with r2812)
> Something on my computer ?
>
> ** Summary changed:
>
> - SimpleShear Preprocessor to update
> + Crash of normalInelasticity scripts (and others ?) - SimpleShear Preprocessor to update
>
> --
> You received this bug notification because you are subscribed to Yade.
> https://bugs.launchpad.net/bugs/739429
>
> Title:
>  Crash of normalInelasticity scripts (and others ?) - SimpleShear
>  Preprocessor to update
>
> Status in Yet Another Dynamic Engine:
>  New
>
> Bug description:
>  Changes to commit about this Preprocessor. And (new ?) problems with
>  normalinelasticity test, leading to crash, at O. run()
>

Yes, I will try with other scripts with the latest revision to be sure but I'm afraid that on my Pc at least it goes wrong

After update and a new compilation check tests are ok (?), sorry.

Still crash for normalInelasticityTest when accessing _force[0] unfortunately.

I have got the crash with that script as well.

Anton

On Fri, Apr 29, 2011 at 5:48 PM, jduriez <email address hidden> wrote:
> After update and a new compilation check tests are ok (?), sorry.
>
> Still crash for normalInelasticityTest when accessing _force[0]
> unfortunately.
>
> --
> You received this bug notification because you are subscribed to Yade.
> https://bugs.launchpad.net/bugs/739429
>
> Title:
>  Crash of normalInelasticity scripts (and others ?) - SimpleShear
>  Preprocessor to update
>
> Status in Yet Another Dynamic Engine:
>  New
>
> Bug description:
>  Changes to commit about this Preprocessor. And (new ?) problems with
>  normalinelasticity test, leading to crash, at O. run()
>

Crash always in ForceContainer::reset (also with FrictMat) but crash occurs only when interactions exist.

Ok, reason is absence of NewtonIntegrator.

See http://<email address hidden>/msg03085.html

Please, check r2849

r2849 allows indeed to not use NewtonIntegrator. Thank you !
(changing the name of the bug since most was for these crashs, and also since SimpleShear Preprocessor is not yet modified !)