search

humanoid-path-planner / hpp-rbprm

"Implementation of RB-PRM planner using hpp."
5 stars 7 forks source link

fix segfault when environment is not BV_OBBRSS type with triangles #57

Closed pFernbach closed 4 years ago

pFernbach commented 4 years ago

In the shooter, there was a failing assert (in Debug) or a segfault (in Release) when the environment was not composed of fcl::BV_OBBRSS nodes, with modelType==BVH_MODEL_TRIANGLES.

This temporary fix avoid the failling assert/segfault but is not sufficient : the contact normal is not correctly computed yet in this cases (it is set to +z)

jmirabel commented 4 years ago

After discussion with @pFernbach: when https://github.com/humanoid-path-planner/hpp-fcl/issues/92 is solved, there won't be a need to put the BV type specifically to get the triangles. This specification is hazardous because it won't work with OBB.

stonneau commented 4 years ago

Ok, so should we merge the request or not ?

pFernbach commented 4 years ago

I think yes, I opened an issue to remember that we still need to come with a better fix, but I won't have the time to take care of this soon.

stonneau commented 4 years ago

Well actually why not raise an error ? I mean if it won't work it won't work no ?

pFernbach commented 4 years ago

Teguh need to be able to use the shooter with primitives type. For this use case (and most of the others) it's not an issue if you don't have the correct normal at this step of the algorithm.