[vJASS] using boolexpr like triggers

[kStiyl]

Is there any downside to using boolexpr like a trigger, instead of using actual trigger? boolexpr seems to use much less memory than creating a trigger.

The code below works just like Nestharus' Event library: pretty much same fps drop and op limit

library SmallEvent

    private module M
        private static method onInit takes nothing returns nothing
            set runner = CreateForce()
            
            call ForceAddPlayer(runner, Player(0))
        endmethod
    endmodule
    struct SmallEvent extends array
        implement M
        
        boolexpr func
        
        private static force runner = null
        private static integer index = 0
        
        method fire takes nothing returns nothing
            call ForceEnumPlayers(runner, .func)
        endmethod
        
        method register takes code c returns nothing
            if .func == null then
                set .func = Filter(c)
            else
                set .func = Or(.func, Filter(c))
            endif
        endmethod
        static method create takes nothing returns thistype
            local thistype this = index + 1
            
            set index = this
            return this
        endmethod
    endstruct
    
endlibrary

 

[Bribe]

The event library should let you unregister the event, otherwise it's just the same as adding some conditions to a trigger and evaluating it. This would be a revolutionary idea if it worked! Unfortunately, it bugs with multiple players and runs the event more than once. Enumerating depletes the force of any players, so when you think you're only enumerating that one player it enumerates all of them.

I made the RealEvent library which is the best of both worlds and is a super short script:

http://www.hiveworkshop.com/forums/submissions-414/snippet-realevent-270158/

 

[kStiyl]

I'm not sure what you mean by 'Enumerating depletes the force of any players, so when you think you're only enumerating that one player it enumerates all of them.' so after enum, the force is depleted = empty?

 

[Bribe]

Enumerating players is like GroupEnums. They add players to the force based on the results of that filter which will run once per player. There exist Enum...Counted natives which, regrettably, do nothing differently due to the engineers forgetting to finish that part of the code.

 

[kStiyl]

Ah understood. Would any other functions using boolexpr stuff be possible candidates? like GroupEnum functions with an environment of only 1 unit

 

[PurplePoot]

To answer the initial question, yes, there are two downsides of boolexprs:

The first, which is sort of trivial, is that it's annoying to return a meaningless boolean at the end of your function.

The second, which is more important, is that boolexprs aren't threaded properly (since they behave more like a function call) and some things (such as TriggerSleepAction) may not work the way you want them to.

 

[Wietlol]

Do they reset operation count to prevent OP limit?

Also... a bit of retoric question... Can you place them in an array? (Normal variables in structs are also arrays.)

 

[Bribe]

Yes, they reset the OP count but do not allow waits.