2.80 -> 2.81 problem with calculateTimeOfImpact()

StefanEgo
Posts: 4
Joined: Thu Sep 27, 2012 7:35 pm

2.80 -> 2.81 problem with calculateTimeOfImpact()

Post by StefanEgo »

Hi,

I'm in the process of upgrading Bullet in our project from 2.80 to 2.81.

I'm stuck with how to handle btCompoundCollisionAlgorithm::calculateTimeOfImpact(). According to the source it states:

Code: Select all

btAssert(0);
//needs to be fixed, using btCollisionObjectWrapper and NOT modifying internal data structures
was that an oversight (i.e. should that code have been updated before release?
If not, is there a suggested way to handle the situation (i.e. change the interface and make it take a const btCollisionObjectWrapper* instead of the btCollisionObject* or something else)?

Regards,
Stefan
User avatar
Erwin Coumans
Site Admin
Posts: 4221
Joined: Sun Jun 26, 2005 6:43 pm
Location: California, USA

Re: 2.80 -> 2.81 problem with calculateTimeOfImpact()

Post by Erwin Coumans »

All of the bt*CollisionAlgorithm::calculateTimeOfImpact is obsolete and isn't (or shouldn't be) used by any of Bullet.

Are you sure you rely on that implementation?
Thanks,
Erwin
StefanEgo
Posts: 4
Joined: Thu Sep 27, 2012 7:35 pm

Re: 2.80 -> 2.81 problem with calculateTimeOfImpact()

Post by StefanEgo »

I double-checked and actually in our case that was logically unreachable code.

We weren't using that function in our code directly and the only usage in Bullet 2.81 is as far as I can tell in btCollisionDispatcher::defaultNearCallback(), when using a dispatcher set to contiunous dispatching (i.e. DISPATCH_CONTINUOUS).

Am I reading the fact then correctly, that this means that setting dispatchers to continuous dispatching isn't supported any longer, either and also is considered deprecated as of Bullet 2.81?