Discussion New Language - Discussion

[Lyerae]

In another thread, someone commented about having someone else work on vJASS (at least that's what I think he implied), so I replied that I had been planning to make a language built specifically as a "holy crap, I'm bored so I'm making this" kind of project.

But for an hour or so, I started thinking.
We could really, REALLY use a new language. cJASS is broken last I heard, and it seems vJASS has a lot of bugs (I don't pay attention to this stuff unless it directly affects me, so meh. :/ ), and each are developed by a single person.

I personally believe this is very inefficent, and by making a new language, with a team of people working on it, we can create a brand-new language, which has minimal bugs and/or issues, has frequent updates, cross-platform, and has the features people want.

So I mainly want to ask this: Would you want a brand new language, which has all the features that vJASS currently has (and cJASS, possibly.), built into a fast compiler, which has as few as possible bugs (hopefully, at least.).

(Now, despite what I had said about having a team work on this, I'd be more than happy develop the language and build the parser myself. I just need input on features.)

PS: This will be written in C, and will be as cross-platform as possible.

 

[Laiev]

acceptable but what exactly you want to make? i mean, new methods to do things and those things already destroy/remove/null things to prevent bugs? or i don't get what you want :banghead:

because some extra language = more things to people learn

 

[Dinowc]

I'm fine with vJass right now

not really in the mood to learn everything all over again

 

[Jesus4Lyf]

I'm fine with vJass right now

not really in the mood to learn everything all over again

That, and if Vex can't do it, what makes you think anyone can... (not even because he's good, just because he's someone.)

 

[Lyerae]

Because Vexorian is lazy, maybe? Or is unwilling to add things.

I'd make this as close as possible to vJASS, in terms of features, so it wouldn't be hard to learn (though I'll likely have a more features).

 

[Laiev]

like <example> everything you just start any variable with location/group, then start some script which will check when the endfunction come and already remove/destroy that location/group and null then without any other command?

 

[Lyerae]

Erm, sorry. Not sure I understand what you mean.
Your saying something about automatic leak-checking?
I could possibly do that, if I can find an efficent way to do so.

 

[Laiev]

yes, is a leak-checker-remove


something like this:

function Test takes nothing returns nothing
    local location point

this will run other script in the 'new language'

'save' location point in P

then continue the trigger

...
    call ...
    call ...
endfunction

when the save map read the 'endfunction' the script will search for EVERY thing saved in the script and will clear and remove <just like newgen, when you save the map, it convert every vjass to jass>

'load' location point from P
call RemoveLocation (point)
set point = null

 

[Lyerae]

> 'new language'

Just call it "X Code" or something like that for now. Easier than "new language".

And yes, that should be perfectly possible, and it's very likely I'll add it.

 

[Nestharus]

Personally, I'd be happy if I could code jass using an interpreted language and have access to a lexical object that contained the environment : |.

Also, one time extraction of map archive depending on what I'm using (war3map.j, etc) and one import.

If the core is around this idea, that'd be awesome. The interpreted language could be written using #s to signify preprocessing. Also the scripts should know where they are in the lexical environment (static text). Interscript communication (all one file), and etc.

The language you'd write would work in this core and be like a standard distribution for extending JASS.

Make it like this and that will be golden : ). I suggest Lua as Grimoire already uses it... ><

And use this distribution please: http://luaforwindows.luaforge.net/

And be sure to rewrite stuff like ObjectMerger to work with 1 import concept (should be inside of the core).

I'm fine with vJass right now

not really in the mood to learn everything all over again

lol... vjass is so screwy. cJASS and Zinc are more screwy though ^^.

 

[Lyerae]

Well, I was personally thinking of creating a whole new language, with vJASS like features (of course, I'd add structs, a library-like implementation, storage enhancers, etc), and then adding some way to add new scripting languages, such as Lua, or Euphoria (the one you've been talking about), and others.

I'm not 100% sure on how to do this yet. I've got a lot of research to do.

 

[Romek]

Don't forget that the only language Wc3 can read is vanilla JASS. Whatever you do will need to be able to compile to that; nothing more, nothing less.

 

[Lyerae]

Yes, thank you. I'll be sure it makes valid code. Even if it takes months of working on it.

 

[Nestharus]

Yes, thank you. I'll be sure it makes valid code. Even if it takes months of working on it.

if it takes months, sc2 will be out within 2-3 months of you doing your first release, hence the reason I mentioned doing a core that could be applied to anything and working off of that.. ><

Make your stuff reusable or you'll regret it later : D.

 

[Lyerae]

Well, I've been thinking of making the parser read a simple XML file (or something like it), which actually defines the language, and have the parser output code based on the XML files definitions.

How's that sound?

 

[Nestharus]

uh huh, but you should replace preprocessing directives in all languages (#) with an interpreted language that's part of the core.. : |, as well as lexical environment.

 

[Lyerae]

Explain?

 

[tooltiperror]

Make C++ loops.

     for(int x = 0; X<10; X++)
          {
               call BJDebugMsg(I2S(x))
          }

Also, maybe a global variable keyword?

 function global takes nothing returns nothing
     global integer x = 0
 endfunction

Or, maybe, a global function call.

 function global takes variable variable, string name, value value, returns nothing
     globals
        variable name = value
     endglobals
 endfunction

Something to that effect, I think it would be possible, you just have to add some fancy stuff. Still an idea.

Edit: Me, and maybe Weep, could beta test for Macs, if you go through with it.

EditEdit: And prettier text macros, please? Maybe, a keyword, or a better syntax? Maybe with asteriks?

 *textmacro example *unit
     call KillUnit(*unit*)
 *endtextmacro

 

[Jesus4Lyf]

Or, maybe, a global function call.

 function global takes variable variable, string name, value value, returns nothing
     globals
        variable name = value
     endglobals
 endfunction

I like how that doesn't make sense.

>yes, is a leak-checker-remove
I like how that changes what the map does. Nothing in WC3 actually leaks, thanks to the FogState hashtable --> agent thing. What if you purposely "leak" a sequence of locations, so that you can iterate through handle ids later and retrieve them for use?

 

[tooltiperror]

Of course it doesn`t.

It was almost thinking without logic, if you will.

The idea was thinking of C in JASS with the JASS and C misfigured.