Real-Time Physics Simulation Forum

I too was distracted by localShapeInfo in btCollisionWorld, and it turns out that it does contain what we need. I've modded the ClosestRayResultCallback to store the local ID of a a hit to a compound shape. In the struct, add int m_childShapeID; In addSingleResult(), add if (rayResult.m_loclShapeInf...

I will also need that much detail sooner or later. Plan would be to have a unique btCollisionShape for all child shapes, then somehow cast the ray/rigid-collsion result to btCollisionShape and extracting the userData/userIndex. It might involve modding the library's source or perhaps there is anothe...

svedach wrote:I think in many cases create a copy of the world after the simulation will be slow operation. We need to make to the existing world raytest multithreaded.

Yes it would be. My suggestion was to do so just once at runtime, but that is game specific based on a static world (not dynamic actors).

It kinda looks expected, but I'm just seeing solid shapes which will effectively weigh the same as the base. Perhaps tune the mass of the base, or even constrain it to the ground. Also has me thinking about playing with gyro effect!

Might not be what you're looking for, but as the scene is static then you could have a duplicate dynamics world running on a seperate thread with a raytest queue. I found that this previously produced results quickly enough for my needs (camera, particles).

I'm not sure what you would like to do differently?

We find the min and max heights, then position the shape at the center location between the two extremes.
Just offset your heightfield data to start with (when generating or importing) if you want everything non-zero origin.

I notice in btVector3.cpp (Ln 60) that maxIndex defaults to -1L, you could try defaulting it to 0 to avoid the case of -1 ever being returned. Apart from that, good luck reproducing the crash! Hey your game looks awesome, adding to wishlist

1 unit would equate to 1m * local scaling. So the heightfield will be the size of your width/length * local scaling. The 'square' that you mention is incorrect and would actually cover 2x2 1m scaled grid slots made from 9 heights, or 1x1 2m scaled from 4 heights. x-x-x | | | x-x-x | | | x-x-x or x--...

ugh! so 'btIndexedMesh' has allowed for short indices. now I'm back to versions of earlier issues. Any suggestions are very welcome, please! SOLVED I was still providing normals in a non-btVector3 format after sorting the same problem with vertices. ...now to incorporate the changes in to serializat...

I've been trying to interpolate mesh normals from a vehicle's raycast using (forum member) Kester 's code. There are two problems at the moment: 1) btAssert(indicestype == PHY_SHORT) kicks in, and 2) manually setting a "normal offset" to external vertex normals doesn't seem to be correct. ...

Oh, well this should solve :lol: things for you. If you browse the constraint demos (at least) one of them has optional MLCP solvers (all references remarked out by default). A quick search online revealed that the forklift demo also references some of them. Selecting a desired solver is normally at...

Had a strange experience earlier where the constraints were like a sponge again. It took a while to understand how, and I'm still investigating, but deduced that the vehicles were being spawned at a higher location (110m instead of 10m) which somehow made the constraints loose...that 0.01 constraint...

Are you looking to create from a heightfield or meshes? I've always used btTriangleMesh/btBvhTriangleMeshShape with the intention of having caves. These shapes are serialized to disk - which takes a few moments - or quickly loaded back next time (if exists). So for a large area it is split up in to ...

Hi, I'm attaching vehicle chassis rigids to a single static rigid (for parking anchors), with a prismatic constraint constructed from a btGeneric6DofSpringConstraint. Linear limits (setLinearLowerLimit...) for the slide are set on the Y axis, to allow for gradients/cambers of the road. The constrain...

I will speculate (to get my account started here!), the floaters are sleeping rigids (which have prematurely detected non movements). Either lower the angular/linear amounts for them to sleep, or set the flag to always keep them awake.