Squamous Eventful Plugin Suggestions
维护者通常 3 天内回复
还没有人认领这个 Issue。
评估
调研方向
The request proposes several new Eventful hooks: projectile item removal or landing, flow creation, and queued ranged weapon actions, plus a question about projectile firer data. Start by reading Eventful's existing event definitions and the DFHack hooks that expose projectile, flow, and unit action changes. Done means agreeing on which hooks and guarantees are needed, implementing the selected events, and adding coverage for them; the issue does not name specific files or tests.
由索引模型根据 Issue 内容生成。
描述
As per the prior discord conversation here is the stuff that would really benefit me currently. This is gonna be a little scatterbrained but here goes:
- Something that fires when a proj_itemst is removed from world.projectiles.all. Right now, my magazine system has to, when it is "fired", set its velocity to 0 and drop it at the shooters feet, tracking it the entire time, until it lands, at which point it is deleted and a copy added to the shooter's inventory (while this is going on, a bullet is spawned from the magazine, and the newly spawned magazine has its bullet count decreased by 1). This is what that snippet of code looks like:
local function try_return(info, attempt)
if not S.enabled then return end
local item = df.item.find(info.item_id)
if not item then return end
local still_proj = false
pcall(function() still_proj = item_is_projectile(info.item_id) end)
-- this scans all of world.projectiles.all each call
if still_proj then
if attempt <= 30 then
dfhack.timeout(1, "ticks", function() try_return(info, attempt + 1) end)
end
return
end
-- ... inventory stuff (return mag to container)
end
-- Called for every single shot:
dfhack.timeout(2, "ticks", function() try_return(ret_info, 1) end)
local function try_remove(item_id, attempt)
if not S.enabled then return end
local item = df.item.find(item_id)
if not item then return end
local still_proj = false
pcall(function() still_proj = item_is_projectile(item_id) end)
if still_proj then
if attempt <= 30 then
dfhack.timeout(1, "ticks", function() try_remove(item_id, attempt + 1) end)
end
return
end
pcall(dfhack.items.remove, item)
end
-- Called for every single shot:
dfhack.timeout(2, "ticks", function() try_remove(item.id, 1) end)
If there was something, like, say, onProjItemLanded(item_id), it could maybe look like this:
-- ammo.lua enable():
eventful.onProjItemLanded["ammo"] = function(item_id)
local info = S.pending_returns[item_id]
if not info then return end
S.pending_returns[item_id] = nil
-- ... do the inventory work directly, no polling
end
-- shotgun.lua enable():
eventful.onProjItemLanded["shotgun"] = function(item_id)
if not S.pending_removals[item_id] then return end
S.pending_removals[item_id] = nil
pcall(dfhack.items.remove, df.item.find(item_id))
end
==================================================================================
Q: is proj.firer (the unit pointer on a proj_itemst) guaranteed to be non-nil when onProjItemCheckMovement fires?
A (that i got in discord): probably not, since projectiles can also be created by things being thrown into the air by various other mechanics
o currently i have to do this to track every shot, and it happens every tick:
local function find_firer_id(bow_id)
if bow_id < 0 then return -1 end
for _, unit in ipairs(df.global.world.units.active) do
if dfhack.units.isActive(unit) and not dfhack.units.isDead(unit) then
for _, inv_item in ipairs(unit.inventory) do
if inv_item.item.id == bow_id then
return unit.id
end
end
end
end
return -1
end
But if proj.firer could be guaranteed to be set at onProjItemCheckMovement time, I can make it look like this (or even something cleaner):
local firer_unit_id = -1
pcall(function()
if projectile.firer then
firer_unit_id = projectile.firer.id
end
end)
==================================================================================
- Something like "onFlowCreated", firing when a flow is created on the map either via dfhack.maps.spawnFlow() or by DF mechanics. Right now I have a script that prunes excess fire so all the flamethrowers and incendiary grenades I have don't completely torch the map each time. The check looks like this:
local CHECK_INTERVAL = 50 -- ticks between scans
local function check_fire()
if not enabled then return end
call_tick = call_tick + 1
local seen = {}
for _, block in ipairs(df.global.world.map.map_blocks) do
for _, flow in ipairs(block.flows) do
if flow.type == df.flow_type.Fire then
local wx = block.map_pos.x + flow.pos.x
local wy = block.map_pos.y + flow.pos.y
local wz = block.map_pos.z
local key = pos_key(wx, wy, wz)
seen[key] = true
if not fire_ages[key] then
fire_ages[key] = call_tick
elseif (call_tick - fire_ages[key]) >= GRACE_CALLS then
flow.expanding = false
end
end
end
end
-- ... prune stale entries
dfhack.timeout(CHECK_INTERVAL, "ticks", check_fire)
end
But if there was something like that, I could make it look like this:
eventful.enableEvent(eventful.eventType.FLOW_CREATED, 1)
eventful.onFlowCreated["firecontrol"] = function(pos, flow_type)
if not enabled then return end
if flow_type ~= df.flow_type.Fire then return end
local key = pos_key(pos.x, pos.y, pos.z)
if not fire_ages[key] then
fire_ages[key] = call_tick
end
end
- Another thing I just ran into, something like onShootRangedWeapon), which would fire when a unit queues a ShootRangedWeapon action, would be huge for me. I could cut down a lot of uses of repeat-util.
- 主要语言
- C++
- 星标
- 2.1k
- 派生
- 509
- 平均合并
- 2 天 19 小时
- 30 天内合并 PR
- 25
环境准备
- 没有 Dockerfile 或 Docker Compose 文件
- 有 Pull Request 模板
- 阅读贡献指南
从这里开始
- 先读完整个 Issue,再读项目的贡献指南。
- 在 Issue 下留言说明你要接手 —— 这能避免两个人做同样的事。
- Fork 仓库,在一个分支上完成修改。
- 提交 Pull Request,并在描述里引用这个 Issue 编号。
DFHack/dfhack 的其他 Issue
-
难度 2/5 1-3 小时 新手友好度 78/100
维护者通常 3 天内回复
-
难度 2/5 1-3 小时 新手友好度 66/100
维护者通常 3 天内回复
-
难度 5/5 一周以上 新手友好度 35/100
维护者通常 3 天内回复
-
RemoteFortressReader: GetCreatureRaws sends CP437 (not UTF-8) strings, which breaks protobuf decoding可能已有人在做 @dwmuller 于 2 天前认领。 未关闭
难度 2/5 1-3 小时 新手友好度 20/100
维护者通常 3 天内回复
-
难度 5/5 一周以上 新手友好度 25/100
维护者通常 3 天内回复