VPhysics-Jolt
VPhysics-Jolt copied to clipboard
Joshua-Ashton
Collision issue cause problems with many addons, including LFS, Simfphys, ect.
Both props are set to no-collide with each other, but still continue to interact with each other
the higher the speed the more times they interact with each other, this problem affect not only addons, but anything that uses no-collide.
this problem is also most likely to cause this behavior:
https://steamcommunity.com/sharedfiles/filedetails/?id=2857489430
Main problem: https://github.com/Joshua-Ashton/VPhysics-Jolt/issues/86 https://github.com/Joshua-Ashton/VPhysics-Jolt/issues/28
LFS, Simfphys and other: https://github.com/Joshua-Ashton/VPhysics-Jolt/issues/87 https://github.com/Joshua-Ashton/VPhysics-Jolt/issues/75 https://github.com/Joshua-Ashton/VPhysics-Jolt/issues/55 https://github.com/Joshua-Ashton/VPhysics-Jolt/issues/50 https://github.com/Joshua-Ashton/VPhysics-Jolt/issues/48 https://github.com/Joshua-Ashton/VPhysics-Jolt/issues/30
Simfphys/LFS author has created a successor that likely solves this issue https://steamcommunity.com/sharedfiles/filedetails/?id=2912816023
I've tried to read the code for half a day, and I came to some conclusions. I'm not familiar with both VPhysics-Jolt and Jolt Physics, so they might be wrong.
The GMod's no collide constraint is made with logic_collision_pair, what it does is add an object hash pair into IPhysicsObjectHashPair (JoltPhysicsObjectHashPair). The VPhysics-Jolt itself doesn't use it at all, it only provides it.
Inside GMod, I can only guess it uses this inside its solver's ShouldCollide. The solver does get called, but there's a problem. Its result doesn't seem to be passed to Jolt Physics. It's being ignored.
As a result, Jolt Physics doesn't know about this, causing the issue.
About the reason for not using the result, I don't know exactly. I can't fix and test it since I don't know much about these things.
I have high hopes for this https://github.com/miguelmazetto/gmod_better_vphysics
Should be fixed now, previously ShouldCollide was not being respected when linear cast was used.
Also I do not know why you would have high hopes for the exact same physics engine.
The flaws with IVP are pretty fundamental with it's solver and will not be solved by some dude importing a vector math library and trying to optimize things that are not the problem.
Also I do not know why you would have high hopes for the exact same physics engine.
The flaws with IVP are pretty fundamental with it's solver and will not be solved by some dude importing a vector math library and trying to optimize things that are not the problem.
in such old and ancient engine that based on hundreds hacks and patches just to make it work somehow, it's all about people and ideas, the more people working on something the better for the community. At least He claiming that it will be "original physics engine with performance improvements" with is still milion% better that the original gmod physics if true. But dont get me wrong, you did an amazing job. I still believe one day we will have almost perfect vphysics replica that will allow server owners to unlock true power of modern CPU for server with high amount of players to enjoy without all existing today problems.