Some semi-random thoughts on terrain generation

[FakeTruth]

Maybe you could even try to use the wiggly lines for terrain, if you can get it to generate 3d wiggly planes

[NiLSPACE]

Maybe you could even try to use the wiggly lines for terrain, if you can get it to generate 3d wiggly planes

That would be an interesting heightmap.


There might be something wrong with the FluidSprings generator. I don't have a large stacktrace, so here is a small one:

> MCServer.exe!cProbabDistrib::MapValue(int a_OrigValue) Line 137 C++
MCServer.exe!cFinishGenFluidSprings::GenFinish(cChunkDesc & a_ChunkDesc) Line 746 C++
MCServer.exe!cComposableGenerator::DoGenerate(int a_ChunkX, int a_ChunkZ, cChunkDesc & a_ChunkDesc) Line 209 C++
MCServer.exe!cChunkGenerator::DoGenerate(int a_ChunkX, int a_ChunkY, int a_ChunkZ) Line 279 C++
MCServer.exe!cChunkGenerator::Execute() Line 264 C++
MCServer.exe!cIsThread::thrExecute(void * a_Param) Line 78 C++

EDIT:
Weird. I can't reproduce this in debug mode.

[xoft]

How do you like them ravines?

[NiLSPACE]

Love it

[NiLSPACE]

I have a proposal for the cFinishGenSingleBiomeSingleTopBlock finisher. Rename it to cFinishGenSingleTopBlock and make it accept a list/vector for the biomes it can place in and a list/vector for the allowed blocks below.

[xoft]

Sounds reasonable.

[NiLSPACE]

Then I'm going to work on that now

[NiLSPACE]

Done: https://github.com/mc-server/MCServer/pull/1254

[NiLSPACE]

I also started working on acacia trees. They might be a little too random though:

[xoft]

A bit too random, as you said.

I think you should start with a variable-length vertical trunk, then pick whether to make a 1-top or 2-top tree, then for the first top pick one of the 8 directions, and for the optional second top pick a direction that is opposite or almost opposite to the first one. Keep going in that direction only, for 3 or 4 blocks diagonally (1 step in the direction with 1 step in Y++), then place the top.