Crash when activating collisions after cell change
Crash when activating collisions after cell change
I noticed that every time you change cell and the try to activate collisions, you get a crash.
Re: Crash when activating collisions after cell change
Confirmed. Another regression, probably from the mwrender refactoring
Re: Crash when activating collisions after cell change
It is npc and creature physics that is causing this, if you comment out insertObject() for them it works. Don't know how to fix this yet.
Re: Crash when activating collisions after cell change
Got this sort of fixed with my latest commit. You can switch cells and then turn on collision and it will work. If you already have collision turned on though and try to change cells it crashes.
Re: Crash when activating collisions after cell change
Sorry, no. Please revert your last commit. You are potentially breaking the entire world model and definitely parts of the game mechancis. Anyway, working around the problem like this won't do us any good. We need to analyse what exactly the problem is and then properly fix it.
Re: Crash when activating collisions after cell change
Well I just made another commit and it works correctly now, even if collision is already on.
Re: Crash when activating collisions after cell change
Revert please. This may make the physics work, but it breaks other stuff. It is also the wrong way to address the problem.
Re: Crash when activating collisions after cell change
Thanks. Sorry for shooting you down like this, but this particular change would have caused countless problems further down the line.
Re: Crash when activating collisions after cell change
I have been reviewing the code in World, Scene, the class hierarchy and the mwrender subsystem for about an hour now without being able to find a flaw in the logic or the implementation. Well, the enable/disable implementation is a bit odd (see tracker), but that does not touch physics at all.
That could mean that the bug is actually in the physics implementation. Or it could mean that I haven' looked hard enough.
I will continue searching for the problem over the next couple of days, but it would be helpful if in parallel an investigation could be done inside the physics system.
That could mean that the bug is actually in the physics implementation. Or it could mean that I haven' looked hard enough.
I will continue searching for the problem over the next couple of days, but it would be helpful if in parallel an investigation could be done inside the physics system.