World.notifyBlockUpdate inaccurately mimics (bad) vanilla behaviour about mcmultipart HOT 10 CLOSED

Would this be the correct way to handle these methods? 2xsaiko@8c2f076

from mcmultipart.

No. markAndNotifyBlock needs to call all of them, and notifyBlockUpdate only the ones traditionally handled by IWorldEventListeners. getActualWorld() will not backtrack!

from mcmultipart.

Huh? World's markAndNotifyBlock works the same way with it calling notifyBlockUpdate, I copied that if block straight from World. Other than that, vanilla World.markAndNotifyBlock specifically only handles flags 1 and 16 itself. I did notice that the condition for flag 16 was inverted though, that needs to be fixed

Only IWorldEventListener I could find that did anything with the flags was RenderGlobal, and that only handles flag 8, which isn't handled at all in MCMP's code. Should I handle every flag in markAndNotifyBlock then?

from mcmultipart.

Huh? World's markAndNotifyBlock works the same way with it calling notifyBlockUpdate, I copied that if block straight from World. Other than that, vanilla World.markAndNotifyBlock specifically only handles flags 1 and 16 itself. I did notice that the condition for flag 16 was inverted though, that needs to be fixed

Yes, but the ActualWorld's markAndNotifyBlock will call the ActualWorld's notifyBlockUpdate, which won't then handle the multipart stuff.

from mcmultipart.

But I'm also calling the wrapped world's notifyBlockUpdate under the same conditions that the actual world's notifyBlockUpdate will get called in markAndNotifyBlock, so the multipart stuff is already handled. Right? I feel like I'm missing something here, haha

from mcmultipart.

notifyBlockUpdate(pos, iblockstate, newState, flags);

okay, i'm stupid

HOWEVER! This will cause getActualWorld().notifyBlockUpdate() to be called twice. Not as bad, but could still impact performance a little

from mcmultipart.

HOWEVER! This will cause getActualWorld().notifyBlockUpdate() to be called twice. Not as bad, but could still impact performance a little

do not forget

from mcmultipart.

yeah, the last commit should fix that.

from mcmultipart.

Aha! Seems like it. Thanks!

from mcmultipart.

I'll push a release before going to bed :)

from mcmultipart.