vJASS & Zinc Documentation
For the latest documentation about how it works vJASS and Zinc language layers for Warcraft III, please follow these links:
Jasshelper documentation - Zinc documentation - WC3 Optimizer documentation

[Knowledge Construction] JASS mythbusters.

moyack · 12517

0 Members and 1 Guest are viewing this topic.

[Knowledge Construction] JASS mythbusters.
on: August 03, 2012, 05:08:02 PM

Hello Jassers:

I want to start a thread where we can put all the stuff that makes you doubt about jass. Specifically questions related to efficiency and code development.

The idea is that this thread will show in the first post a table with all the information we can gather, and it will serve as a point of reference for support in other questions dine in this site.

Question
Answer
Question 1Answer 1

Let's start with the first question: At the end, what produces leaks and what not?

Dont' forget to provide a test map with the sample code, etc :)


Re: [Knowledge Construction] JASS mythbusters.
Reply #1 on: August 15, 2012, 12:58:23 PM

Sorry but i have no more motivation for such things.
But i suppose this thread could be more or less useful :
http://www.hiveworkshop.com/forums/triggers-scripts-269/benchmarks-truth-out-there-201810/

vJass is already a jass preprocessor, no need to (ab)use vJass features in order to make some inferior vJass preprocessor coded "by hand".


Re: [Knowledge Construction] JASS mythbusters.
Reply #2 on: August 16, 2012, 12:03:57 AM

I really don't have the will to make a test-map for this, but these are all the very leaky things in Warcraft III:

- Locations
- Groups
- Forces
- Special Effects
 
Things that cannot be destroyed:
- Terrain Deformations
- Trackables
- Events

Things that should be nulled when declared locally:
Anything that extends the 'agent' type in common.j.



Re: [Knowledge Construction] JASS mythbusters.
Reply #3 on: August 16, 2012, 05:33:41 AM

And what about players?? they extend agent but in theory they don't leak because they're constants. Is that totally true??


Re: [Knowledge Construction] JASS mythbusters.
Reply #4 on: August 16, 2012, 11:03:13 AM

Note that what Mag said is only a theory, it is hardly a fact.

So we shouldn't null unitpool and itempool  ?! (just as an example)

My opinion about it : just null everything which is not code, string, boolean, integer.
Regardless how the handle is recycled (with a handle reference counter like an unit, or without like an image)

Also about players i've heard that some people null them to fix a leak problem with their maps, however frankly i have never met an enough reliable people to believe this, but that doesn't mean it's not true.

About leak event i'm not sure, maybe it's cleaned when the trigger is destroyed, and well maybe it depends if there is one or more events registred to the same trigger.

PS :

But this is 100 % true : GUI eats childrens.

vJass is already a jass preprocessor, no need to (ab)use vJass features in order to make some inferior vJass preprocessor coded "by hand".


Re: [Knowledge Construction] JASS mythbusters.
Reply #5 on: August 18, 2012, 11:59:57 AM


vJass is already a jass preprocessor, no need to (ab)use vJass features in order to make some inferior vJass preprocessor coded "by hand".


 

Chaos Realm - The world of Game modders and wc3 addicts     WC3JASS.com - The JASS Vault   Jetcraft - A Starcraft II mod