I test your patch, and get some problems below:
1. space.cpp : (ODEObject) ERROR: Space returned empty ODE handle
I restore to the orignal version, and it works ok, and no this error.
2. pure virtual method called, terminate called without an active exception
I do not know how this happen, and it rarely happen
3. segment fault : texture.cpp 41: glDeleteTextures(1, &mTexID);
you create a thread for 'rendercontrol', though its 'endcycle' is executed
in the main thread, but there is something wrong of the 'Destruction'.
as you have memtioned, but how to solve it, could you show an example?
5. simcontrolnode simstep
each simulation control node hold a certain simstep, and they are different,
for example, 'rendercontrol' is 0.04, however, this is easy to fix.
And finally, with the usage of 'boost::barrier' the code is more
simpler and easy to be understood, and the quit problem is caused by
*deadlock*, I think. Thank you :-)
Ben from Apollo3D
已 经 超 过 100 万 台 电 脑 安 装 了 网 易 有 道 免 费 英 汉 词 典 ， 点 击 此 处 可 以 快 速 下 载