jcarius-nv

Hard to tell from your screenshotted backtrace... apparently it's happening [here](https://github.com/NVIDIAGameWorks/PhysX/blob/acbd071531ca4f3e2a82525b92f60824178c39fa/PhysX_3.4/Source/GeomUtils/src/contact/GuContactConvexMesh.cpp#L341). That rotate method eventually calls `PxMat33::transform(const PxVec3& other)`, but I don't see where this can crash. Maybe your stack...

Hi @TheDonsky , I looked at the video in your original post, indeed something is misbehaving there. I tried to reproduce but in my tests the flags are actually set...

ok, now I understand it and I can also reproduce it. The issue is that you're setting the flag **between** `simulate()` and `fetchResults()`, and I was just testing the lockflag...

I can confirm that there's a missing flush operation in the double-buffering scheme such that the RigidDynamicLockFlags set in the buffer never make it to the simulation. We've fixed it...

Hm, I don't know how I can help you further without an actual scene that can reproduce your issue. I tried reproducing once again in the SnippetHelloWorld by adding ```...

[`PxRigidDynamic::setKinematicTarget()`](https://github.com/NVIDIAGameWorks/PhysX/blob/93c6dd21b545605185f2febc8eeacebe49a99479/physx/include/PxRigidDynamic.h#L117) should do the job.

`PxRigidBody::setForceAndTorque` allows for different modes, e.g., applying force, impulse (momentum), and velocity change. If you don't want a shape to partake in collisions you can set the `eSIMULATION_SHAPE` of the...

Unfortunately there is no replacement. It just has to be done by the user through addForce & scene queries or triggers etc.

Thanks for reporting. While I don't see your exact error, compilation with gcc11 indeed fails. We've narrowed down the issue and it will be fixed in the next version. As...

This is surprising because we're testing the release in a clean containerized Ubuntu environment the respective default clang and gcc versions, i.e., Ubuntu 20.04 LTS with gcc 9 or clang...