System Advanced Indexing & Data Storage

[W!†A_cRaft]

error message when trying to save the map:

Spoiler

 

[Romek]

Update your Jasshelper.

There's a link in this thread. (Or click here)

 

[W!†A_cRaft]

i downloaded the 0.A.2.7 version, how do I update? simply copy the files into the correct directory?

or I only need to copy jasshelper.exe and forget the rest?

 

[Darthfett]

i downloaded the 0.A.2.7 version, how do I update? simply copy the files into the correct directory?

or I only need to copy jasshelper.exe and forget the rest?

When installing a new jasshelper version in newgen pack, replace jasshelper.exe and clijasshelper.exe in the folder called "bin" inside jass newgen pack's folder.

.

 

[W!†A_cRaft]

I cannot replace somethign that doesnt exist.....

In :
...Previous directoyry\JassNewGen1.5\bin\
there is no jasshelper.exe nor clijasshelper.exe

so i cannot replace them with somethign if they do not exist, i can put these two (jasshelper.exe and clijasshelper.exe from the following directory:
...Previous directory\0.A.2.7.jasshelper\executable\

but than what am i suppose to do with the rest of the stuff that is in
0.A.2.7.JassHelper folder? simply delete them or what?

Picture of the NewGen1.5/bin directory (there is no jasshelper.exe nor clijasshelper.exe)


picture of 0.A.2.7.jasshelper directory (what to do with rest of the stuff?)

 

[Jesus4Lyf]

Should be JassNewgenPack\jasshelper, not JassNewgenPack\bin.

 

[W!†A_cRaft]

okay got it, now should i copy everything or only jasshelper.exe and clijasshelper.exe?

 

[Jesus4Lyf]

>only jasshelper.exe and clijasshelper.exe
That.

 

[Kenny]

I was just reading up on AutoIndex, and found that it hooked RemoveUnit() and all that. Does this accomplish the same thing without the use of hooks? Or did you just not implement that?

Also, I was just wondering.. besides the AIDS structs, what else does AIDS provide that something like AutoIndex doesn't. Like efficiency perks or something.

(Just trying to get an opinion on the two systems, people at wc3c seem to hate anything that isn't in their database, and also things that have unusual and/or funny names)

 

[Jesus4Lyf]

>I was just reading up on AutoIndex, and found that it hooked RemoveUnit() and all that. Does this accomplish the same thing without the use of hooks? Or did you just not implement that?

That. I don't need hooks.

>Also, I was just wondering.. besides the AIDS structs, what else does AIDS provide that something like AutoIndex doesn't. Like efficiency perks or something.

AIDS has lots of efficiency perks that AutoIndex doesn't have. For example, getting an AIDS struct for a unit is a native call, and you can typecast between AIDS structs for a unit. AutoIndex allows access to a unit just before it is removed from the game, AIDS trades that off for struct/index locking and efficiency (I've never needed access on removal personally - you can access all the unit's data in AIDS, but not the unit itself). Truth is I could add it if I felt like it. I don't see the point.

>(Just trying to get an opinion on the two systems, people at wc3c seem to hate anything that isn't in their database, and also things that have unusual and/or funny names)

In reality, the systems are very close. The only difference in practice is efficiency vs accessing the unit before it leaves the game. AIDS structs are never allocated and deallocated. Actually, the methods for onCreate/onDestroy are called for the unit ID. In AutoIndex, to achieve a similar thing, they make a struct which is allocated/deallocated as the unit enters/leaves the map, and this stuff is fired through a trigger evaluate loop. AIDS fires it through a single trigger eval (diff between TT and KT2 stuff, and some).

I do it for the cheap (lightweight) unit data structs. :thup:

 

[Kenny]

Thanks for the detailed response. Grim's rants on some threads kinda got me a bit confused. I think the biggest difference was the periodic removal vs undefend removal, which has now been implemented.

Even with some of the "drawbacks" of AIDS structs (such as no array members and all that crap), I still find AIDS structs a pretty simple way to go about doing many unit based systems, such as spawn systems. And I am aware that I can use array members if I use a work around method.

Anywho, looks like I will be using AIDS. Plus it is always good to support local systems.

Oh and:

>I was just reading up on AutoIndex, and found that it hooked RemoveUnit() and all that. Does this accomplish the same thing without the use of hooks? Or did you just not implement that?

That. I don't need hooks.

That is really interesting. And also a major plus for me. I never thought it necessary to hook [ljass]RemoveUnit()[/ljass] and all that for some unit indexing.

Edit:

Some more questions:

If I was to use the undefend method for unit-leaves events, the AIDS_onDestroy method would fire just before the unit leaves, allowing me to still do stuff with the unit (such as add special effects and stuff) just before it leaves, correct?

And if I was to use the periodic method for the above situation, I wouldn't be able to do so, right?

Also, what is your reasoning behind not using an AIDS module? I guess then users could have their onInit method back (although it would feel awkward having AIDS_onCreate, AIDS_onDestroy, AIDS_onFilter but no AIDS_onInit ).

 

[Jesus4Lyf]

>If I was to use the undefend method for unit-leaves events, the AIDS_onDestroy method would fire just before the unit leaves, allowing me to still do stuff with the unit (such as add special effects and stuff) just before it leaves, correct?

Nope. Regardless which method you use, you can't access a unit just before it leaves using AIDS. Only the unit's data.

I can change this but I haven't seen any need yet.

>Also, what is your reasoning behind not using an AIDS module?
At the time of release, there were still issues with modules, and they took up your onInit method regardless. The textmacro would compile on the JassHelper version that came with NewGen, and most people didn't update their JassHelper beyond that. I intend to release an AIDS module in the future, and have the textmacro simply implement the module. Even with the AIDS module, AIDS_onInit will work for backwards compatability. I might even release a second module that allows non [LJASS]extends array[/LJASS] structs. (Actually, I intend to.)

 

[Kenny]

>Nope. Regardless which method you use, you can't access a unit just before it leaves using AIDS. Only the unit's data.

Oh.. Fair enough. What I meant was something like:

struct Example extends array
     //! runtextmacro AIDS()

    real x
    real y

    static method AIDS_filter takes unit u returns boolean
        return GetUnitTypeId(u) == 'hfoo'
    endmethod

    method AIDS_onDestroy takes nothing returns nothing
        call DestroyEffect(AddSpecialEffect("SomeBigEffect.mdl",this.x,this.y)) // So can I do this?
        call DestroyEffect(AddSpecialEffectTarget("SomeBigEffect.mdl",this.unit,"chest")) // And I can't do this?
    endmethod

    method AIDS_onCreate takes nothing returns nothing
        set this.x = GetUnitX(this.unit)
        set this.y = GetUnitY(this.unit)
        // Lets just pretend that x and y get updated every second or so because I am lazy.
    endmethod

endstruct

Also, is it necessary to have some of the functions in AIDS (such as RegisterOnEnter() etc) available to users (via the public keyword)? I just found it odd.

Oh and I am looking forward to the AIDS module.

 

[Jesus4Lyf]

>[LJASS]// So can I do this?[/LJASS]
Yep.
>[LJASS]// And I can't do this?[/LJASS]
Yep.

>is it necessary to have some of the functions in AIDS (such as RegisterOnEnter() etc) available to users (via the public keyword)?

They'll be made private finally when I release the module. All part of what I mentioned before as to why I used a textmacro.

 

[Kenny]

Just wanted to give a heads up about static ifs outside of functions being removed in the future:

Vexorian's post.

More specifically:

Static ifs outside functions will get removed. They DO NOT work fine right now the library phase will get confused by them.

Seems like that could screw with your "choose one or the other" deal you have going right now.

 

[Jesus4Lyf]

Nah, well, I saw that, but it shouldn't matter. It just means extra functions will exist, but nothing is named the same anyway.

If it throws syntax errors or something I'll fix and release, but else I'll just fix it with the next release (whenever that is). But I did note that, thanks.

 

[Azlier]

There is one way in which this system can fail.

Units with a default autocast ability (ex: Priests with heal) receive the healon order when created and fire the associated triggers before the enter region event even fires. Yes, even when you abuse the boolexpr in the event.

Which means that triggers that are registered for orders with no targets have the possibility of getting an unindexed unit .

Insane, no?

 

[Jesus4Lyf]

At least AIDS provides a GetUnitIndex function which will work for that situation (but would break all AIDS structs for that unit ). Hm.

Perhaps there is a way for me to capture that. I'll have to take a look some time. :thup:

 

[Azlier]

There is a way to capture that, with an event that detects autocast orders. Where are we gonna get one of those? Hmm.

 

[Jesus4Lyf]

>Where are we gonna get one of those? Hmm.
*Laughs* :banghead:
I'll need to investigate. Makes sense to use yours if that is indeed what I need, however. I really should reverse the order Event fires its triggers...
I think its back to front at the moment, off memory (meaning using your sys would solve nothing since AIDS would fire after anything using it, but at least I wrote what your sys runs on (Event)... it's all RIGGED).

Investigating AIDS with autocast orders, exposing T32 members, reversing Event firing, adding .unregister(trig) to Event, I'll get around to it!