PhysX (v2.7.3) & Irrlicht (v1.4)

A forum to store posts deemed exceptionally wise and useful
frostysnowman
Posts: 83
Joined: Fri Apr 11, 2008 3:06 am
Location: Beaverton, Oregon, US

Post by frostysnowman »

char wrote:the bug is caused when creating device with openGL try to change EDT_OPENGL into EDT_DIRECT3D9.
I do believe changing the video driver would make no effect on physX.. I wrote a physX application of my own framework and all the problems dissapeared! I think I know why the bug where it crashes on actor->addForce(forceVec) crashes. As noted in my earlier post, the sdk gave a warning upon this same crash, issuing the message:
NpScene::createActor: Writelock is still acquired. Procedure call skipped to avoid a deadlock!
I found some similiar PhysX sdk errors on the internet, and I've deduced that the problem has something to do with creating a actor in the wrong place, near the call of simulate(). I noticed that in the main loop it checks if the key has been pressed, and if so to fire a box or ball. But, what it should be really doing is shooting the box/ball right when the event hits the event reciever. That's my implementation in my framework that I've written, and that error has never showed up when I shoot boxs. Additionally, the flickering boxes are gone :D. If anyone would like the VC++ project for my physX framework project, I guess I can upload it. It's not done though.
char
Posts: 2
Joined: Tue Jul 01, 2008 6:47 am

Post by char »

well.. i am having the same bug(crashes on actor->addForce(forceVec) ) . when i change the driver type, it works fine. i dont believe it :shock:... but it works for me :)
PhilippT
Posts: 4
Joined: Thu Jan 25, 2007 10:09 pm

Would be cool to take a look to your framework

Post by PhilippT »

I would be glad to take a look on your framework. I think a stable implementation of the physX stuff would gain irrlicht for serious usage.

And i am not very experienced in c++ and 3d but could learn fast with stable code.

However i would be happy.
Post Reply