Joona Aalto comments

Results 115 comments of


                                            Joona Aalto

Store square roots for world-space inverse angular inertia

#793 implemented the inertia tensor inversion fix. I think this issue is *mostly* redundant and it's likely not worth going the sqrt route. Still not entirely sure why exactly Rapier...

Rigidbodies squash and bounce off each other when restitution is zero

This is "expected" behavior, in that it's a known limitation. It is briefly mentioned in the docs under the Accuracy section [here](https://docs.rs/avian2d/0.1.2/avian2d/dynamics/index.html#accuracy). The issue here is the large mass ratio...

Issue: Incorrect inelastic collision

Is this in version 0.1? I suspect this was fixed by #542 on the main branch. You can also see the second video there for a scenario that is quite...

Issue: Incorrect inelastic collision

The main branch already uses Bevy 0.15-rc, but not Bevy's main branch. My new PR for interpolation and extrapolation support (#566) *does* depend on Bevy's main branch though, since it...

Rework mass properties

I made a write-up laying out a lot of the current problems, along with a proposal for overhauling mass properties: #499

Make collision reporting opt-in

I don't have any up-to-date branch with collision events based on observers or one-shot systems, but on the main branch, `CollisionStarted` and `CollisionEnded` events are now opt-in via the `CollisionEventsEnabled`...

Negative scales coming from an ancestor are ignored by colliders

#472 fixes the second case with the building's collider not being scaled. The mirroring bug for negative scales is a separate issue.

Panic when simultaneously querying for ```LinearVelocity``` and ```Forces```

The `ForcesItem` retrieved when querying `Forces` has [`linear_velocity`](https://docs.rs/avian3d/latest/avian3d/dynamics/rigid_body/forces/trait.RigidBodyForces.html#method.linear_velocity) and [`linear_velocity_mut`](https://docs.rs/avian3d/latest/avian3d/dynamics/rigid_body/forces/trait.RigidBodyForces.html#method.linear_velocity_mut) accessors for this purpose; you can use those when you get query conflicts. > Also, probably you want to calculate...

Sometimes `shape_cast` does not return all hit entities

For me, in 0.4, this test fails immediately with a `max_distance` of 0.0 but runs for some time with anything greater (e.g. even 0.00001). I'm guessing that's something on Parry's...

assertion failed: b.min.cmple(b.max).all() in `ColliderAabb::grow`

This commonly means that the position or rotation is somehow NaN or infinite, causing the computed world-space AABB to be invalid. I'd maybe try to add some logs to see...