Floody water

[xoft]

Nice ideas, but I think I'll leave those for later.

[ThuGie]

Looking really promising .
Actually amazed at the progress of the development.

Multi-version support! is already something i never expected.

[xoft]

I guess I'll tear the branch down and restart later - there's quite a lot of cleanup I need to do first.

[xoft]

Branch torn down; I've decided to develop this in the trunk.
Also, he first step has been taken - I've cleaned up the simulators, and moved the original simulator code into a descendant class. Now the server admins will be able to choose fluid simulators for each world and each fluid separately using the world.ini.

[l0udPL]

I dunno but is there any reason to keep the 'classic' simulator even it does not work properly? Of course if I remember right, that it has bugs so anyone with the bucket can crash the server?

[xoft]

I'll keep it until the other simulators are deemed more stable and more useful.
Still, I expect the final version to have at least two fluid simulators - vanilla and floody. Not to mention that a world needs to know whether its lava spreads 3 blocks (overworld) or 7 blocks (nether). So the world.ini configuration will stay as well.

I'm not aware of any crashes due to fluid simulation, do you have any more information? Best case scenario, a fixed sequence that will always lead to a crash. Or at least a crashdump.

[xoft]

First iteration: forgot to spread downwards.

---

Second iteration: Floody really is floody


The spreading works okay now, but drying up doesn't, because the block updates aren't propagated correctly, so drying up at one block doesn't trigger a re-calc at the neighboring blocks

---

The code has been commited as rev 963, you can play with it
It gets enabled by setting [Physics]:WaterSimulator to Floody in world.ini. [Physics]:LavaSimulator for lava.

[FakeTruth]

Lookin good, it's definitely floody XD

[xoft]

Rev 964 implements proper drying up, too. But I've run into issues that I'm not yet sure how to solve: multithreading concurrency.
The simulators, blockhandlers and all stuff expects that the world doesn't change while they're executing. Unfortunately it seems the world is changing. If you try to place blocks in water's path, sooner or later you're gonna fool the simulator and it will start glitching.

[NiLSPACE]

the good news is its much better than the old water. and if flowers break they don't turn into dirt blocks. the bad news is, if water hits lava there is no stone or cobblestone or obsydian.