FOSS gamedev and creative worlds

I haven't found this posted yet.

Liquids cannot flow over different liquids eastward (i.e. from the west, the last neighbour checked).

To demonstrate if you have no other liquids, disable lavacooling and use default water and lava.
Place one on top of the other, and it will only flow in three directions. This is true whether the lower liquid is source or flowing. Also true whether starting from source or flowing.



Also, flows are not updated adjacent to a different flow that is removed. I.e. a hole is left instead of being filled.

I would be interested to know whether this happens in newer versions as well.

I wonder if it is this bug: https://github.com/minetest/minetest/issues/3263

oh kinda interesting... i already seen that liquids had a weird behavior but i never noticed this!

Fixerol wrote:I wonder if it is this bug: https://github.com/minetest/minetest/issues/3263

Thanks for the link, I didn't know about that repo.

These do appear to have the same underlying causes.

For the default/river water examples, https://github.com/minetest/minetest/issues/3263
I don't have river water, but the wiki states that it can only flow 2 nodes anyway, so most of the pics are expected behaviour(?). Although, not falling next to default water does resemble my second bug.
The last one however, is exactly the eastward flow bug.

For the lava/water columns, https://github.com/minetest/minetest/issues/2399
We can speculate a mechanism like this --

The falling liquid alerts adjacent gaps on touchdown.
The southern gap looks north, finds lava, expects lava, accepts lava.
The western gap looks north, finds none, looks east, finds lava, expects lava, accepts lava.
The northern gap looks north, finds water, expects water, rejects lava.
The eastern gap looks north, finds none, looks east, finds water, expects water, rejects lava.

This scheme fits with my second bug if we imagine the depleted flow holding an 'expectation' of its original type.
A similar story works for dysfunctional lava generators.

I have yet to think of a single characterisation that covers both bugs, and maybe they are distinct.