pipeworks/autoplace_tubes.lua

154 lines
4.8 KiB
Lua
Raw Normal View History

-- autorouting for pneumatic tubes
function pipeworks.scan_for_tube_objects(pos)
if pos == nil then return end
tube_autoroute({ x=pos.x-1, y=pos.y , z=pos.z })
tube_autoroute({ x=pos.x+1, y=pos.y , z=pos.z })
tube_autoroute({ x=pos.x , y=pos.y-1, z=pos.z })
tube_autoroute({ x=pos.x , y=pos.y+1, z=pos.z })
tube_autoroute({ x=pos.x , y=pos.y , z=pos.z-1 })
tube_autoroute({ x=pos.x , y=pos.y , z=pos.z+1 })
tube_autoroute(pos)
end
function in_table(table,element)
for _,el in ipairs(table) do
if el==element then return true end
end
return false
end
function is_tube(nodename)
return in_table(tubenodes,nodename)
end
Multiple updates: 1) Refactor autoplace, There was a lot of redundant code and like a dozen unneccessary string scans for every node next to every tube placed! I put it all into indexed tables and loops instead of bizarre and unexplainable variable names and copy and pasted code. There was also no support for notifying a chest when an item has been taken from it by a filter, so I added something for that. I also thought it prudent to fall back on the allow_metadata_inventory_take function should a special can_remove not exist. In fact if can_insert doesn't exist, it calls allow_metadata_inventory_put instead. I also added a thing for allowing pipes to attach to nodes of other modules, without having to hard code type all those node names into autoplace.lua. Basically node.tube.collects(i,param2) and i is the direction from the pipe and param2 is the param2 of what it's pointing at. I also abstracted the inscrutable correlation between i and param2 by trial and error (and the paramwand mod) into understandable functions. There was no pipeworks namespace so I created it, and put these functions into pipeworks.collects (as distinguished from a node.tube.collects function, which uses those functions) And now it's too late to cart my old clothes to the thrift store, dangit. 2) My "node.tube.collects" idea might be redundant with the node.tube.connect_sides thing, though possibly more versatile so I'll leave it in. 3) I was using node.tube.connects and fancy functions for checking if it's the sides or top or whatnot, and this connect_side thing came in. This should make both my way and the way using connect_side work. Also removed some debugging cruft
2013-10-15 05:45:07 +02:00
if pipeworks == nil then
pipeworks = {}
end
--a function for determining which side of the node we are on
Multiple updates: 1) Refactor autoplace, There was a lot of redundant code and like a dozen unneccessary string scans for every node next to every tube placed! I put it all into indexed tables and loops instead of bizarre and unexplainable variable names and copy and pasted code. There was also no support for notifying a chest when an item has been taken from it by a filter, so I added something for that. I also thought it prudent to fall back on the allow_metadata_inventory_take function should a special can_remove not exist. In fact if can_insert doesn't exist, it calls allow_metadata_inventory_put instead. I also added a thing for allowing pipes to attach to nodes of other modules, without having to hard code type all those node names into autoplace.lua. Basically node.tube.collects(i,param2) and i is the direction from the pipe and param2 is the param2 of what it's pointing at. I also abstracted the inscrutable correlation between i and param2 by trial and error (and the paramwand mod) into understandable functions. There was no pipeworks namespace so I created it, and put these functions into pipeworks.collects (as distinguished from a node.tube.collects function, which uses those functions) And now it's too late to cart my old clothes to the thrift store, dangit. 2) My "node.tube.collects" idea might be redundant with the node.tube.connect_sides thing, though possibly more versatile so I'll leave it in. 3) I was using node.tube.connects and fancy functions for checking if it's the sides or top or whatnot, and this connect_side thing came in. This should make both my way and the way using connect_side work. Also removed some debugging cruft
2013-10-15 05:45:07 +02:00
local function nodeside(node, tubedir)
if not tubedir or
not node or
not (type(node.param2) == "number" and node.param2 > 0 and node.param2 < 23) then
return "back"
end
Multiple updates: 1) Refactor autoplace, There was a lot of redundant code and like a dozen unneccessary string scans for every node next to every tube placed! I put it all into indexed tables and loops instead of bizarre and unexplainable variable names and copy and pasted code. There was also no support for notifying a chest when an item has been taken from it by a filter, so I added something for that. I also thought it prudent to fall back on the allow_metadata_inventory_take function should a special can_remove not exist. In fact if can_insert doesn't exist, it calls allow_metadata_inventory_put instead. I also added a thing for allowing pipes to attach to nodes of other modules, without having to hard code type all those node names into autoplace.lua. Basically node.tube.collects(i,param2) and i is the direction from the pipe and param2 is the param2 of what it's pointing at. I also abstracted the inscrutable correlation between i and param2 by trial and error (and the paramwand mod) into understandable functions. There was no pipeworks namespace so I created it, and put these functions into pipeworks.collects (as distinguished from a node.tube.collects function, which uses those functions) And now it's too late to cart my old clothes to the thrift store, dangit. 2) My "node.tube.collects" idea might be redundant with the node.tube.connect_sides thing, though possibly more versatile so I'll leave it in. 3) I was using node.tube.connects and fancy functions for checking if it's the sides or top or whatnot, and this connect_side thing came in. This should make both my way and the way using connect_side work. Also removed some debugging cruft
2013-10-15 05:45:07 +02:00
--get a vector pointing back
local backdir = minetest.facedir_to_dir(node.param2)
Multiple updates: 1) Refactor autoplace, There was a lot of redundant code and like a dozen unneccessary string scans for every node next to every tube placed! I put it all into indexed tables and loops instead of bizarre and unexplainable variable names and copy and pasted code. There was also no support for notifying a chest when an item has been taken from it by a filter, so I added something for that. I also thought it prudent to fall back on the allow_metadata_inventory_take function should a special can_remove not exist. In fact if can_insert doesn't exist, it calls allow_metadata_inventory_put instead. I also added a thing for allowing pipes to attach to nodes of other modules, without having to hard code type all those node names into autoplace.lua. Basically node.tube.collects(i,param2) and i is the direction from the pipe and param2 is the param2 of what it's pointing at. I also abstracted the inscrutable correlation between i and param2 by trial and error (and the paramwand mod) into understandable functions. There was no pipeworks namespace so I created it, and put these functions into pipeworks.collects (as distinguished from a node.tube.collects function, which uses those functions) And now it's too late to cart my old clothes to the thrift store, dangit. 2) My "node.tube.collects" idea might be redundant with the node.tube.connect_sides thing, though possibly more versatile so I'll leave it in. 3) I was using node.tube.connects and fancy functions for checking if it's the sides or top or whatnot, and this connect_side thing came in. This should make both my way and the way using connect_side work. Also removed some debugging cruft
2013-10-15 05:45:07 +02:00
--check whether the vector is equivalent to the tube direction; if it is, the tube's on the backside
if backdir.x == tubedir.x and backdir.y == tubedir.y and backdir.z == tubedir.z then
return "back"
end
--check whether the vector is antiparallel with the tube direction; that indicates the front
if backdir.x == -tubedir.x and backdir.y == -tubedir.y and backdir.z == -tubedir.z then
return "front"
end
--facedir is defined in terms of the top-bottom axis of the node; we'll take advantage of that
local topdir = ({[0]={x=0, y=1, z=0},
{x=0, y=0, z=1},
{x=0, y=0, z=-1},
{x=1, y=0, z=0},
{x=-1, y=0, z=0},
{x=0, y=-1, z=0}})[math.floor(node.param2/4)]
--is this the top?
if topdir.x == tubedir.x and topdir.y == tubedir.y and topdir.z == tubedir.z then
return "top"
end
--or the bottom?
if topdir.x == -tubedir.x and topdir.y == -tubedir.y and topdir.z == -tubedir.z then
return "bottom"
end
Multiple updates: 1) Refactor autoplace, There was a lot of redundant code and like a dozen unneccessary string scans for every node next to every tube placed! I put it all into indexed tables and loops instead of bizarre and unexplainable variable names and copy and pasted code. There was also no support for notifying a chest when an item has been taken from it by a filter, so I added something for that. I also thought it prudent to fall back on the allow_metadata_inventory_take function should a special can_remove not exist. In fact if can_insert doesn't exist, it calls allow_metadata_inventory_put instead. I also added a thing for allowing pipes to attach to nodes of other modules, without having to hard code type all those node names into autoplace.lua. Basically node.tube.collects(i,param2) and i is the direction from the pipe and param2 is the param2 of what it's pointing at. I also abstracted the inscrutable correlation between i and param2 by trial and error (and the paramwand mod) into understandable functions. There was no pipeworks namespace so I created it, and put these functions into pipeworks.collects (as distinguished from a node.tube.collects function, which uses those functions) And now it's too late to cart my old clothes to the thrift store, dangit. 2) My "node.tube.collects" idea might be redundant with the node.tube.connect_sides thing, though possibly more versatile so I'll leave it in. 3) I was using node.tube.connects and fancy functions for checking if it's the sides or top or whatnot, and this connect_side thing came in. This should make both my way and the way using connect_side work. Also removed some debugging cruft
2013-10-15 05:45:07 +02:00
--we shall apply some maths to obtain the right-facing vector
local rightdir = {x=topdir.y*backdir.z - backdir.y*topdir.z,
y=topdir.z*backdir.x - backdir.z*topdir.x,
z=topdir.x*backdir.y - backdir.x*topdir.y}
--is this the right side?
if rightdir.x == tubedir.x and rightdir.y == tubedir.y and rightdir.z == tubedir.z then
return "right"
end
--or the left?
if rightdir.x == -tubedir.x and rightdir.y == -tubedir.y and rightdir.z == -tubedir.z then
return "left"
end
--we should be done by now; initiate panic mode
minetest.log("error", "nodeside has been confused by its parameters; see pipeworks autoplace_tubes.lua, line 78")
Multiple updates: 1) Refactor autoplace, There was a lot of redundant code and like a dozen unneccessary string scans for every node next to every tube placed! I put it all into indexed tables and loops instead of bizarre and unexplainable variable names and copy and pasted code. There was also no support for notifying a chest when an item has been taken from it by a filter, so I added something for that. I also thought it prudent to fall back on the allow_metadata_inventory_take function should a special can_remove not exist. In fact if can_insert doesn't exist, it calls allow_metadata_inventory_put instead. I also added a thing for allowing pipes to attach to nodes of other modules, without having to hard code type all those node names into autoplace.lua. Basically node.tube.collects(i,param2) and i is the direction from the pipe and param2 is the param2 of what it's pointing at. I also abstracted the inscrutable correlation between i and param2 by trial and error (and the paramwand mod) into understandable functions. There was no pipeworks namespace so I created it, and put these functions into pipeworks.collects (as distinguished from a node.tube.collects function, which uses those functions) And now it's too late to cart my old clothes to the thrift store, dangit. 2) My "node.tube.collects" idea might be redundant with the node.tube.connect_sides thing, though possibly more versatile so I'll leave it in. 3) I was using node.tube.connects and fancy functions for checking if it's the sides or top or whatnot, and this connect_side thing came in. This should make both my way and the way using connect_side work. Also removed some debugging cruft
2013-10-15 05:45:07 +02:00
end
function tube_autoroute(pos)
local active = {0, 0, 0, 0, 0, 0}
local nctr = minetest.get_node(pos)
if not is_tube(nctr.name) then return end
local adjustments = {
{ x=-1, y=0, z=0 },
{ x=1, y=0, z=0 },
{ x=0, y=-1, z=0 },
{ x=0, y=1, z=0 },
{ x=0, y=0, z=-1 },
{ x=0, y=0, z=1 }
}
-- xm = 1, xp = 2, ym = 3, yp = 4, zm = 5, zp = 6
local positions = {}
local nodes = {}
for i,adj in ipairs(adjustments) do
positions[i] = {x=pos.x+adj.x, y=pos.y+adj.y, z=pos.z+adj.z}
nodes[i] = minetest.get_node(positions[i])
end
for i,node in ipairs(nodes) do
local idef = minetest.registered_nodes[node.name]
-- handle the tubes themselves
if is_tube(node.name) then
active[i] = 1
-- handle new style connectors
elseif idef.tube and idef.tube.connect_sides then
Multiple updates: 1) Refactor autoplace, There was a lot of redundant code and like a dozen unneccessary string scans for every node next to every tube placed! I put it all into indexed tables and loops instead of bizarre and unexplainable variable names and copy and pasted code. There was also no support for notifying a chest when an item has been taken from it by a filter, so I added something for that. I also thought it prudent to fall back on the allow_metadata_inventory_take function should a special can_remove not exist. In fact if can_insert doesn't exist, it calls allow_metadata_inventory_put instead. I also added a thing for allowing pipes to attach to nodes of other modules, without having to hard code type all those node names into autoplace.lua. Basically node.tube.collects(i,param2) and i is the direction from the pipe and param2 is the param2 of what it's pointing at. I also abstracted the inscrutable correlation between i and param2 by trial and error (and the paramwand mod) into understandable functions. There was no pipeworks namespace so I created it, and put these functions into pipeworks.collects (as distinguished from a node.tube.collects function, which uses those functions) And now it's too late to cart my old clothes to the thrift store, dangit. 2) My "node.tube.collects" idea might be redundant with the node.tube.connect_sides thing, though possibly more versatile so I'll leave it in. 3) I was using node.tube.connects and fancy functions for checking if it's the sides or top or whatnot, and this connect_side thing came in. This should make both my way and the way using connect_side work. Also removed some debugging cruft
2013-10-15 05:45:07 +02:00
local dir = adjustments[i]
if idef.tube.connect_sides[nodeside(node, {x=-dir.x, y=-dir.y, z=-dir.z})] then active[i] = 1 end
Multiple updates: 1) Refactor autoplace, There was a lot of redundant code and like a dozen unneccessary string scans for every node next to every tube placed! I put it all into indexed tables and loops instead of bizarre and unexplainable variable names and copy and pasted code. There was also no support for notifying a chest when an item has been taken from it by a filter, so I added something for that. I also thought it prudent to fall back on the allow_metadata_inventory_take function should a special can_remove not exist. In fact if can_insert doesn't exist, it calls allow_metadata_inventory_put instead. I also added a thing for allowing pipes to attach to nodes of other modules, without having to hard code type all those node names into autoplace.lua. Basically node.tube.collects(i,param2) and i is the direction from the pipe and param2 is the param2 of what it's pointing at. I also abstracted the inscrutable correlation between i and param2 by trial and error (and the paramwand mod) into understandable functions. There was no pipeworks namespace so I created it, and put these functions into pipeworks.collects (as distinguished from a node.tube.collects function, which uses those functions) And now it's too late to cart my old clothes to the thrift store, dangit. 2) My "node.tube.collects" idea might be redundant with the node.tube.connect_sides thing, though possibly more versatile so I'll leave it in. 3) I was using node.tube.connects and fancy functions for checking if it's the sides or top or whatnot, and this connect_side thing came in. This should make both my way and the way using connect_side work. Also removed some debugging cruft
2013-10-15 05:45:07 +02:00
end
end
-- all sides checked, now figure which tube to use.
nsurround = ""
for i,n in ipairs(active) do
nsurround = nsurround .. n
end
local newname=string.sub(nctr.name,1,-7)..nsurround
if newname == nctr.name then return end
local meta=minetest.get_meta(pos)
local meta0=meta:to_table() -- XXX: hacky_swap_node
nctr.name = newname
minetest.add_node(pos, nctr)
local meta=minetest.get_meta(pos)
meta:from_table(meta0)
local nctr = minetest.get_node(pos)
end
minetest.register_on_placenode(function(pos, newnode, placer, oldnode, itemstack)
if minetest.registered_items[newnode.name]
and minetest.registered_items[newnode.name].tube
and minetest.registered_items[newnode.name].tube.connect_sides then
pipeworks.scan_for_tube_objects(pos)
end
end)
minetest.register_on_dignode(function(pos, oldnode, digger)
if minetest.registered_items[oldnode.name]
and minetest.registered_items[oldnode.name].tube
and minetest.registered_items[oldnode.name].tube.connect_sides then
pipeworks.scan_for_tube_objects(pos)
end
end)