[Snippet] TimedHandles

[PurgeandFire]

Hmm right, so you could do the same stuff with other agent types ?
http://www.hiveworkshop.com/forums/triggers-scripts-269/fun-stuff-images-210359/
Don't take me wrong, i'm not saying that would be useful, just trying to know.

At least it makes sense for me now why people said you don't need to null agents, even if i still highly doubt you don't need it, for let's say timer, just as an agent example.
I have a doubt for item/unit/pool as an handle example, but for the others it makes sense.

I'm not sure what you are asking. You don't need to null pointers to agents that are never destroyed. You should null pointers to agents that may be destroyed.

You don't need to null pointers to regular handles though, because they are not reference counted (e.g. texttags), even if you destroy them. They don't suffer the handle ID bug.

 

[Troll-Brain]

Nevermind, basically i asked if every agent type had no reference counter, but i suppose the answer is obvious, according to the comment on common.j (i was not aware of it before edo494 told about it)

 

[TriggerHappy]

Updated, code is a bit more convuluted now but I think the new feature may be useful.

-1.0.3
Added option for single timer
Macros moved to the top

It may be better in some cases to use a single timer, like when you plan on having many instances.

 

[TriggerHappy]

Updated.

-1.0.4
NULL_SAFETY option added
Single timer option set to true by default
function wrapper now returns struct instance

I wasn't sure if handle destructors checked for null internally, so I added the option.

 

[Luorax]

// Check if the handle is null before destroying it
private constant boolean NULL_SAFETY = true

If I took the trouble to implement null safety, I'd rather do it in the constructor, cuz I mean what's the point in keeping a null value in the list?

 

[TriggerHappy]

If I took the trouble to implement null safety, I'd rather do it in the constructor, cuz I mean what's the point in keeping a null value in the list?

What?

 

[Luorax]

call DestroyEffectTimed(null, 5)

This is going to create a new instance, a one-shot timer with a timeout of 5 seconds, and once it expires, it's going to recycle the timer and the instance. And all of this for what? A null value? It's not going to anything useful, just waste your CPU time.

It gets even worse with single timer enabled, it's going to add quite a lot to the loop's execution time for utterly no reason.

All I was trying to say is that if I wanted to handle null values properly ("I took the trouble to implement null safety"), I would sort them out before instanciation ("I'd do it in the constructor").

(Was it really that hard to interpret?)

 

[TriggerHappy]

You realize handles can destroyed outside of this script right? lol.

call DestroyEffectTimed(e, 10)
call DestroyEffect(e)

Also, you can't honestly be complaining that the null check is going to negatively affect performance.

 

[Luorax]

//! zinc

library NullTest
{
    struct Foo
    {
        effect e;
        
        private static method onInit()
        {
            thistype this=thistype.allocate();
            
            this.e=AddSpecialEffect("",0.,0.);
            
            if (this.e==null)
            {
                BJDebugMsg("e is null");
            }
            else
            {
                BJDebugMsg("e is not null");
            }
            
            DestroyEffect(this.e);
            
            if (this.e==null)
            {
                BJDebugMsg("e is null");
            }
            else
            {
                BJDebugMsg("e is not null");
            }
            
        }
    }
}

//! endzinc

Output:
e is not null
e is not null

But W/E. If you want null to be a valid argument, then so shall it be.

P.s: and I was not complaining about performance, I was complaining about a lot of BS being done for no reason (when a null is passed, that is)

 

[TriggerHappy]

Lol i dont see why youre getting so frustrated

Anyway im on my phone right now, but did u try adding a TriggerSleepAction? Variables may not be dereferenced instantly. (obviously they aren't)

 

[Luorax]

Even with a 5 second TSA it outputs "e is not null" for the third time too.

 

[TriggerHappy]

Ya my dumb mistake ill fix it

 

[TriggerHappy]

Updated.

-1.0.5
Removed needless null checking

 

[Kazeon]

OMG, so useful