RegisterCallback

[zork]

I just wrote myself a function that triggers a callback on any given event. On top of that you have access to two sets of arguments. One which you define on init. One which is passed along on event.

Benefit is that you only need one event frame for any future event callback.

--rLib:RegisterCallback
function rLib:RegisterCallback(event, callback, ...)
  if not self.eventFrame then
    self.eventFrame = CreateFrame("Frame")
    function self.eventFrame:OnEvent(event, ...)
      for callback, args in next, self.callbacks[event] do
        callback(args, ...)
      end
    end
    self.eventFrame:SetScript("OnEvent", self.eventFrame.OnEvent)
  end
  if not self.eventFrame.callbacks then self.eventFrame.callbacks = {} end
  if not self.eventFrame.callbacks[event] then self.eventFrame.callbacks[event] = {} end
  self.eventFrame.callbacks[event][callback] = {...}
  self.eventFrame:RegisterEvent(event)
end
 
--init
local function OnLogin(...)
  print(...)
end
rLib:RegisterCallback("PLAYER_LOGIN",OnLogin,"hello1","hello2")

Arguments you pass along on init are the first argument in the triggered callback packed as a table.

Any thoughts?

[SDPhantom]

It's a bit outdated, but Spydon uses one of my first implementations of an event library. More recently, I've been trying to weigh the impact between the overhead of events firing being redirected to a table of callbacks versus individual frames that are created for their own purposes responding to their own unique list of events. In all, it's one thing if you have code that has no frames responding to events versus code that already has frames tying into existing resources.