Bugs (Applies to 2.42 preview27)

[goldentaiji]

Had some time to test the newest build and here's some recurring bugs, lol, updated on top.

Yeah, I just took a look at the problems I had previously, and it's down to 2 bugs. They're kinda necessary. (2) definitely. (1) important for now until BOgre starts working.

Jason Lin

[Icoxo]

At which preview do you expect to put resitution in, it is important to simulate different materials because if you dropped a box in Blender now using physics using the box bounds it bounces a little, imagine if that was supposed to be lead

[Erwin Coumans]

At which preview do you expect to put resitution in, it is important to simulate different materials because if you dropped a box in Blender now using physics using the box bounds it bounces a little, imagine if that was supposed to be lead

Icoxo, restitution is quite a bit of work, to get in the 2.42 release.
But you can play with this special Icoxo build: blender-2.42-win32-preview-icoxo.zip !

http://www.continuousphysics.com/ftp/pu ... -icoxo.zip


Thanks,
Erwin

[Icoxo]

The restitution is very nice, thanks. I could use it for bouncing balls off of things and simulations, I don't see why this current version can't get in to 2.42.

Sorry if I annoyed you Erwin, its just that restitution is standard in all physics engines and was one of the few remaining things in sumo still not ported over.

I look foward to character support and compounds in the coming months.

[goldentaiji]

Ah darn I found a bug, lol. suspendDynamics() and restoreDynamics() does not work in bullet.

Jason Lin

[pildanovak]

I think i found out "what's the bug" in the car simulation overnight - in fact,approaching any slope makes a wild collision with the car flying around. -
1.st, the car should stay more in the direction where it moves, there should be an way how to damp the bounces( since cars are not rigid, they deform normally)

2.nd the collision of wheels fails often, which could be the reason - the chassis hits the ground instead of wheels and bounces.
- how do the wheels collide?(standard cylinders?)
- maybe the collision for wheels could be measured more exactly, but how?
- my model has exactly the same scale as the car in the demofiles, also weights, etc..

I would also like to know, if there are functions to:
get the Speed of vehicle(easy to script, but if there's an internal one...)
now if the vehicle is in the air - to measure jumps, etc. - I also noticed that when in the air, the wheels don't rotate, which is not much realistic, and also the reaction of the car is different if you fall on the ground with wheels rotating or not.

[pildanovak]

hello I found 1 more bug with 2 different aspects - the files showing the problem are here:
http://plant.ffa.vutbr.cz/~novak/dwnflz/bugs/

in fact, physics objects added as children of an static object(imagine adding a piece of level with some objects) are added good only if the static(parent) object is added on it's location on hidden layers. when on different place 2 errors can occur - dynamic objects appear on their original location, objects appear quite correctly, but don't collide.
More bug reports are to come !

[meialua]

Found a bug with message sensor: when put into pulse mode it just keeps triggering, without actually receiveing any messages. Works as expected in 2.41, but changes when loaded into 2.42 (preview 24)

EDIT:
Also, Added Objects do not have unique names, so when you add 2 objects from the same original, they are called the same thing, which can cause a crash quite easily. eg: you have an empty that adds cubes every frame, in 2.41 they would be called Cube, Cube.001, Cube.002 etc.. in 2.42, they are all called Cube (a problem when you need to delete inidividual added objects, causes a crash).

EDIT (To the above edit):
Have found that the above statement is not quite true, and the same naming thing happens in 2.41, however, it would be very useful if added objects could have unique names!