[Bf-committers] A subject -- for (no subject)

joeedh joeeagar at prodigy.net
Mon Feb 7 01:29:16 CET 2005


Alexander Ewering wrote:

>
> I know that my posts are mostly ignored by now, but I have a lot of time,
> so why not try :)
>
> Ton, do you still remember how the "collision jitter" problem was solved
> in 2.25's game engine, and why it isn't solved after open-sourcing?
>
> Have the following situation: A completely even plane, and a Sphere
> which bounces down on it.
>
> After a few bounces (say, Restitution = 0.5), the sphere should sit on
> the plane and not move at all anymore.
>
> 2.25's engine did this correctly. When the sphere sat on the floor, any
> attached collision sensors did not trigger anymore, and it didn't visibly
> move.
>
> In 2.36 (after I fixed resitution), a sphere *seems* to sit on the floor
> as well after a few bounces, but it's not the case. It constantly
> "jitters" around, and collision sensors trigger all the time.
>
> My question is: How has this problem been solved in 2.25, and why has the
> code changed so substantially that 2.36 suddenly reveals this problem?

One of the problems the gameengine encountered after it's comeback was 
the fact that Gino van den Bergen released SOLID 3.5, instead of 3, and 
the API had changed.  In face I believe that that was one of the first 
things Kester did, was to change the code so as to be compatible with 
SOLID 3.5.

Unfortunately, I suspect this also may have caused some bugs.

Kester: am I right about that?

joeedh


More information about the Bf-committers mailing list