Abm check all neighbours

[markthesmeagol]

I am writing a 3d game of life thing, since the one that existed previously has no link, and uses 0.4. I am currently adding the actual rule thing (using an abm). However, for that I would need to check all the neighbours of a given node. Is there a fast way to do this, or do I need to manually check all 26 neighbours?

[Krock]

Within the ABM registration you can specify neighbour nodes. One of them is required in order to trigger the "action" function. It your code should only run for - say - three similar neighbours, then you'd need to guard the code within the "action" function using a minetest.find_nodes_in_area check.

Such code could look as follows: (untested, quote to preserve tabulators)

Code: Select all

action = function(pos, node)
	local positions, nodes_count = minetest.find_nodes_in_area(
		vector.add(pos, -1), vector.add(pos, 1), {"default:mese"})

	if (nodes_count["default:mese"] or 0) < 2 then
		return
	end

	-- Code to run if there are >= 3 mese nodes
end

[markthesmeagol]

Within the ABM registration you can specify neighbour nodes. One of them is required in order to trigger the "action" function. It your code should only run for - say - three similar neighbours, then you'd need to guard the code within the "action" function using a minetest.find_nodes_in_area check.

Such code could look as follows: (untested, quote to preserve tabulators)

Code: Select all

action = function(pos, node)
	local positions, nodes_count = minetest.find_nodes_in_area(
		vector.add(pos, -1), vector.add(pos, 1), {"default:mese"})

	if (nodes_count["default:mese"] or 0) < 2 then
		return
	end

	-- Code to run if there are >= 3 mese nodes
end

Thanks! That is actually a lot easier to understand then the api, or the book! I'mma go implement that now...

[rubenwardy]

Thanks! That is actually a lot easier to understand then the api, or the book! I'mma go implement that now...

Because he wrote the exact code you needed to use. The API is a list of functions, and the book can't contain every single possible snippet of code

[markthesmeagol]

Thanks! That is actually a lot easier to understand then the api, or the book! I'mma go implement that now...

Because he wrote the exact code you needed to use. The API is a list of functions, and the book can't contain every single possible snippet of code

Actually, the api said something about a second return value, which doesn't make very much sense to my c/c++ background, and the book just sticks a # or something at the end, without explaining anything.

[rubenwardy]

Actually, the api said something about a second return value, which doesn't make very much sense to my c/c++ background

Higher-level languages add convenience features. In C, register 0 is used to return a single value. In Lua, the equivalent of register 0 indicates how many values should be returned. The caller of a function unpacks the return values like so:

Code: Select all

local one, two = do_thing()
print(one)
print(two)

and the book just sticks a # or something at the end, without explaining anything.

The # is an anchor link which allows you to link to specific section like so: https://rubenwardy.com/minetest_modding ... -modifiers

This section of the book is fairly old, and so doesn't "tell a story" as well as other sections. Perhaps it should introduce a simple ABM before introducing chances and more complex ABMs

I'm trying to work out exactly what it is that's unclear, so I can improve it

Edit: are you referring to the Basic Map Operations section? This makes what you said make sense, as the book doesn't explain it very well - it doesn't mention the second return value, and just uses the # operator to count all values in the table.