Snippet TrigWrap

[kingkingyyk3]

////////////////////////////////////////
//      ~TrigWrap v1.0.1~
//          by kingking
//  
//  TrigWrap is a trigger wrapper
//  that enables trigger recycling.
//
//  Functions :
//  call Trigger.create() -> Trigger
//  call Trigger.getTriggering() -> Trigger
//  set <vars>.event = whichEvent
//  call <vars>.unregister()
//  call <vars>.destroy()
//  set <vars>.data = integer
//  set <vars>.condition = conditionfunc
//  set <vars>.action = code
//  set <vars>.enable = true/false
//  <vars>.data -> integer
//  <vars>.trig -> trigger
//
//  Precautions :
//  Do not destroy the trigger manually,
//  use .destroy() instead.
//
//  Requires : Jasshelper 0.A.2.9 or newer, Event
/////////////////////////////////////
library TrigWrap requires Event
    
    struct Trigger
        private static hashtable hasht
        trigger trig
        EventReg eventReg
        Event registeredEvent
        
        static method create takes nothing returns thistype
            local thistype this = thistype.allocate()
            if this.trig == null then
                set this.trig = CreateTrigger()
                call SaveInteger(thistype.hasht,GetHandleId(this.trig),0,this)
            else
                call EnableTrigger(this.trig)
            endif
            return this
        endmethod
        
        static method getTriggering takes nothing returns thistype
            return LoadInteger(thistype.hasht,GetHandleId(GetTriggeringTrigger()),0)
        endmethod
        
        method operator condition= takes conditionfunc func returns nothing
            call TriggerAddCondition(this.trig,func)
        endmethod
        
        method operator action= takes code func returns nothing
            call TriggerAddAction(this.trig,func)
        endmethod
        
        method operator data= takes integer data returns nothing
            if this.eventReg != 0 then
                set this.eventReg.data = data
            endif
        endmethod
        
        method operator data takes nothing returns integer
            return this.eventReg.data
        endmethod
        
        method operator enable= takes boolean flag returns nothing
            if flag then
                call EnableTrigger(this.trig)
            else
                call DisableTrigger(this.trig)
            endif
        endmethod
        
        method operator event= takes Event whichEvent returns nothing
            debug if this.registeredEvent > 0 then
            debug   call BJDebugMsg("TrigWrap error : Trying to register trigger to more than 1 events")
            debug endif
            set this.registeredEvent = whichEvent
            set this.eventReg = whichEvent.register(this.trig)
        endmethod
        
        method unregister takes nothing returns nothing
            call this.registeredEvent.unregister(this.trig)
            set this.eventReg = 0
        endmethod
        
        method destroy takes nothing returns nothing
            if this.registeredEvent > 0 then
                call this.registeredEvent.unregister(this.trig)
            endif
            call DisableTrigger(this.trig)
            call TriggerClearConditions(this.trig)
            call TriggerClearActions(this.trig)
            set this.eventReg = 0
            set this.registeredEvent = 0
        endmethod
        
        static method onInit takes nothing returns nothing
            set thistype.hasht = InitHashtable()
        endmethod
    endstruct
    
endlibrary

 

[Lyerae]

[ljass]// Cons : You only can add 1 condition and 1 action for trigger.[/ljass]
Not a fan of that line.

How 'bout a demo script, to show us how it's used, exactly.

 

[Sevion]

@Lyerae

Ditz. I don't really like that you can only add 1 condition and 1 action.

I don't really see how this could be useful considering you can only add one action and condition.

Really, all this is doing is adding the ability to attach data to a trigger and confining you to one action and one condition?

I could easily replicate this with a hashtable and not confine myself to that.

A simple hashtable wrapper is all this is.

Besides, you can easily remove that "con".

Also, [ljass]struct Trigger extends array[/ljass]

 

[kingkingyyk3]

Huh? I think my info is leading people to misunderstand what is it. I will add example.

Example added.

 

[Narks]

I like how you use a "z" instead of a "s".

 

[kingkingyyk3]

Hmm, "s" means it have more than 1 trigger. lol.

 

[Sevion]

The example only furthers my reasonings.

The only purpse of this is to allow data attachment to triggers and limit maximum actions and conditions.

You can easily replicate this with a hashtable.

 

[Executor]

The example only furthers my reasonings.

The only purpse of this is to allow data attachment to triggers and limit maximum actions and conditions.

You can easily replicate this with a hashtable.

Agreed. I really don't see the sense of using this lib. Furthermore sometimes attachment by executing could be faster (depending on the usage).

 

[kingkingyyk3]

Furthermore sometimes attachment by executing could be faster (depending on the usage).
Then I need to sacrify the trigger's action. It is not worth.

 

[Executor]

Yes. It wasn't a suggestion to include in your sys, I just pointed out that in some cases triggers with only a condition and a exec attachment make more sense.

BTW: As far as I know, using the condition to do everything related to event reaction makes more sense, because it is faster (?).

 

[Romek]

> Hmm, "s" means it have more than 1 trigger. lol.
Adding 'z' is just a corny alternative to 's'.
They're pronounced almost the same too.

It seems that you're using method operators for the lulz:

.
        method operator data= takes integer whichData returns nothing
            set this.dataz = whichData
        endmethod
        
        method operator trigger takes nothing returns trigger
            return this.triggerz//Please use method in this library to add condition and action, thanks.
        endmethod

        method getData takes nothing returns integer
            return this.dataz
        endmethod

Could all be removed.

.
        static method create takes nothing returns thistype
            set thistype.d = thistype.allocate()
            set thistype.d.triggerz = CreateTrigger()
            set thistype.d.triggerzId = GetHandleId(thistype.d.triggerz)
            call SaveInteger(thistype.ht,thistype.d.triggerzId,0,thistype.d)
            return thistype.d
        endmethod

Do locals offend you?

 

[weaaddar]

On an aside, you must also destroy the boolexpr sent to a condition or else it will leak.

So really you should take the trigger, a boolexpr, and the action. The event will leak ~400bytes last I heard.

 

[kingkingyyk3]

destroy the boolexpr
Hmm, i heard it will leaks if destroy it.

 

[weaaddar]

Destroy both. Destroying the condition func does nothing on its own last I remember :/

 

[Jesus4Lyf]

Lol, destroying it is not stable. If I understand correctly (and I'm fairly sure I do) if you use the same function as a condition for two triggers, then destroying the boolexpr will cause [LJASS]Condition[/LJASS] or [LJASS]Filter[/LJASS] generated boolexprs which have been used in multiple triggers to all fail in those other triggers.

If the boolexprs were created with [LJASS]And[/LJASS], [LJASS]Or[/LJASS] or [LJASS]Not[/LJASS] then they need to be destroyed, are unique, and will leak, but I should think that is the user's job.

I don't see why we need a wrapper for this - it just kind of confuses and restricts and adds a hashtable wrap...

PS. the coolest type of trigger attachment still has to be Azlier's - typecast int to bool and store it inside [LJASS]IsTriggerWaitOnSleeps[/LJASS]. No cleanup! :thup:

So to be blunt... what does this do that ABC can't?

 

[Deaod]

I would never approve of anything that calls DestroyTrigger. Also...this looks a little bit pointless to me.

 

[kingkingyyk3]

Updated. Hmm.. Anyone here does register a trigger more than 1 event? If so, I will add support to it. In this stage, I feel it is very limited because Event in systems are privatized, the authors made some TriggerRegister--Event functions for registration, so I can't access the event and store it.

 

[Jesus4Lyf]

This wasn't graveyarded?
Graveyarded!

In it's current state, this serves as a difficult-to-use interface for dealing with triggers, without solving anything wrong with WC3's.