-
Notifications
You must be signed in to change notification settings - Fork 378
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Assert in PxcBruteForceOverlapBackface #302
Comments
Hi @korokuchum. Thanks for the report. |
This assert means that some precomputed values are incorrect for some of the convexes, potentially leading to missed collisions. There isn't much we can do without either a repro, or at least the vertices used in that example (so that we can cook the convex on our side and check what happens). This would be unrelated to aggregates. On the other hand, this is the "legacy" contact generation codepath which was known to have defects for convex-vs-convex collisions anyway. So even if we fix that assert, this specific codepath is ultimately unreliable - and it was the main reason for the introduction of the PCM codepath. I would therefore suggest switching to PCM (PxSceneFlag::eENABLE_PCM) here. |
With PCM enabled I have different assert. Call Stack
|
Here's the problematic mesh: https://drive.google.com/file/d/1CEb4LRIzhDEll5m37AKbLvewLbEML3Lb/view?usp=sharing Mesh description flags: physx::PxConvexFlag::eCOMPUTE_CONVEX | physx::PxConvexFlag::eSHIFT_VERTICES Cooking parameters:
|
Any updates on this issue? |
Library and Version
PhysX v5.4.0
Operating System
Windows 11
Steps to Trigger Behavior
Couldn't find any specific steps.
Here's how I got this issue:
Scene setup
Broad Phase Type: Any
Solver Type: Any
Friction Type: Any
Expected Behavior
Simulation runs without asserts
Actual Behavior
Assert during simulation
Call stack
The text was updated successfully, but these errors were encountered: