Xinsong Lin

Yes, it does work in OMPL. The SO2 object correctly represents the joint rotation. If we were to support this in constrained planning, then it is indeed possible to have...

curiously, in sapien 3, the base in two_stage_motion doesn't move anymore even though when switching to sapien 2 it moves. Checked the trajectory and everything looked right. Maybe can investigate...

> curiously, in sapien 3, the base in two_stage_motion doesn't move anymore even though when switching to sapien 2 it moves. Checked the trajectory and everything looked right. Maybe can...

Was checking out the segfault seems sapien renderer related. When running just a single test inside the suite, get `RuntimeError: failed to find a rendering device.` But when running more...

Hi! If you try newer versions of the library, we have a matrix in which you can set [allowed collision](https://motion-planning-lib.readthedocs.io/latest/reference/collision_detection/index.html#mplib.collision_detection.AllowedCollisionMatrix).

hi, unfortunately, we do not currently compile a wheel for macos. It is theoretically possible though, since each component can be compiled individually on mac. I can try it later...

sorry unfortunately yes as of right now

What version are you guys using? There might have been a convex bug before iirc. Can you please try the nightly wheels [here](https://github.com/haosulab/MPlib/releases/tag/nightly)?

convex hull should work if your geometry is more or less convex to begin with. at worst your collision detection will be a little preemptive

Please try the newer versions like 0.2.1 or the [nightly release](https://github.com/haosulab/MPlib/releases/tag/nightly)!