Trigger Event - Leak or not

[--Thanatos--]

^topic.

 

[DBZ-123]

Trigger Events dont leak ( well thats what i think )

 

[--Thanatos--]

Yes... thats also what I thought.

But my spell map's handle id keeps increasing lol I don't understand. I don't think I made any other leak.

Any other answers?

 

[TriggerHappy]

Each handle != a leak.

Do you events have any null boolexpr's?

I heard they leak, but some believe they don't.

 

[--Thanatos--]

used Condition(function GetTrue) for the boolexpr.

EDIT: Event Leaks.
Proven by the map I uploaded.

Now what must I do for my map...
How many handles can we have in a map ?

EDIT again: eh, not. They are destroyed when the trigger is destroyed -_-

 

[saw792]

Events don't leak because without the event the trigger won't fire. It's like saying that the hero unit you control on a map is a leak. Sure, it's a handle, but it isn't a leak because you need it to be there. Events will only leak if they become redundant.

 

[Troll-Brain]

They leak if you try to one more than an event to one trigger, even if you destroy the trigger it will leak, but it will be fine if you don't need to destroy the trigger, or if you destroy the trigger and there were only one event (you need to remove the actions before destroy the trigger)

For the null boolxecpr plz stop speaking about that if you're too lazy or not enough competent to test it yourself.
It leaks only with GroupEnum ...

 

[--Thanatos--]

It somewhat doesn't leak (reset handleId) after I destroyed the trigger?

Maybe it's destroyed that way? (whatever, what on Earth?)

 

[Sooda]

Where came multiple events on trigger leak info?

 

[Troll-Brain]

Where came multiple events on trigger leak info?

Just tested myself, you can also try.

 

[--Thanatos--]

... You sure only one event on a trigger that doesn't leak ? (I mean, 30 event on a trigger leaks even when the trigger got destroyed?)

 

[Troll-Brain]

... You sure only one event on a trigger that doesn't leak ? (I mean, 30 event on a trigger leaks even when the trigger got destroyed?)

it's easy to test, with a short timer callback, like 0.01 s, create a trigger add it X events, destroy the trigger and don't forget to null the variable if it is a local.
Then check the memory used by the war3 process (war3 in a window).

If the memory doesn't grow up, then it doesn't leak (war3 must not be paused in any way)

 

[Tom Jones]

I find it hard to believe that trigger events leaks. I think you've been fooled by the engine's recycling of handle ids. Here's how I would test it, plus three other tests concerning triggers (NewGen required):

*Edit*
Yes, I am aware that the handle count jumps sometimes, but note it is only when the first occurence of a handle type are created thus increasing the handle id index.

 

[Sooda]

> Then check the memory used by the war3 process (war3 in a window)

That isn't valid method to judge memory usage. It' s as good as betoken answer by observing clouds in sky.

 

[Troll-Brain]

> Then check the memory used by the war3 process (war3 in a window)

That isn't valid method to judge memory usage. It' s as good as betoken answer by observing clouds in sky.

In a real map yes, but it is valid in a solo test map with this only test, you just need to care about war3 never be paused and you don't use a timer with a too much short periodic time.

Anyway i don't care that you don't believe me, i will live with that.

 

[SFilip]

> it is valid in a solo test map
Only if you know how WC3's memory allocation works. Which you don't.

> It leaks only with GroupEnum
It leaks only if it's null in some enum function.

 

[Troll-Brain]

> it is valid in a solo test map
Only if you know how WC3's memory allocation works. Which you don't.

Well, this is my last sentence about it, i can perfectly know if something leak or not in a TEST map with ONLY ONE test WITHOUT ANY other thing.
You can test the leak by yourself with an other more accurate method.

> It leaks only with GroupEnum
It leaks only if it's null in some enum function.

That is what i wanted to say, sorry if it wasn't clear.