JassHelper 0.A.2.B - A vJass and Zinc 2 Jass compiler

[Barade]

Sorry, but in my opinion the following features are useless:
- Added delegate.
- Added extends array.
- Added the colon operator.

The only thing they do is making the code more unreadable.

offtopic: I could not see the website navigation anymore, using Firefox 3.0.1.

[PurplePoot]

Originally Posted by Barade

Sorry, but in my opinion the following features are useless:
- Added delegate.
- Added extends array.
- Added the colon operator.

The only thing they do is making the code more unreadable.

offtopic: I could not see the website navigation anymore, using Firefox 3.0.1.

I'd agree with colon and somewhat on delegate, but extends array I can see a point to (Heck, take Vex's example, it makes sense well enough).

You're basically just packaging your global arrays into struct names.

EDIT: Would it be possible to overload : separately from []? If so, it would be more useful.

[Vexorian]

The only thing they do is making the code more unreadable.

I disagree, as a matter of fact, : and delegate have been made for the sole purpose of making code more readable. I am already using : to make code a lot more readable because I am not typing stuff backwards anymore, and delegate will really simplify code in something I will release later.

Oh, and don't mess with extends array, it definitely it is baseless to call it useless since it is one of the most useful things to be added to vJass in a while.

Either way, like other vJass features, nobody is forcing anybody to use them.

EDIT: Would it be possible to overload : separately from []? If so, it would be more useful.

No, I really doubt it would be more useful.

[TheDamien]

The below does not appear to error:

JASS:

scope Test initializer Init
    function interface what takes real r returns nothing

    function waffle takes what f returns nothing
        call f.evaluate(5.6)
    endfunction

    function foo1 takes real v returns nothing
        call BJDebugMsg(R2S(v))
    endfunction

    function foo2 takes real v returns real
        call BJDebugMsg(R2S(v))
        return 0.
    endfunction

    function foo3 takes string s returns nothing
        call BJDebugMsg("WUT")
    endfunction
    
    private function Init takes nothing returns nothing
        call waffle(foo1)
        call waffle(foo2)
        call waffle(foo3)
    endfunction
endscope

[Vexorian]

And it shouldn't.

If you wanted validation use what.foo3

[Sooda]

I got vJASS which crashes World Editor on save, in the very end of save process:

JASS:

library CreateBug initializer Init
    globals
        private constant integer number = '0000'
        private constant integer number = 0
    endglobals

    private function Init takes nothing return nothing
    endfunction
endlibrary

Try to save map and World Editor will crash in the very end of save process. If 'return' is replaced with 'returns' JASSHelper starts to give error messages again.

PM-d Vexorian about this issue and he told it should not be caused by JASSHelper.

I'm using Jass New Generation Pack 5a.

[Vexorian]

PM-d Vexorian about this issue and he told it should not be caused by JASSHelper.

It isn't, I can compile that just fine with jasshelper, try newgen pack's thread or something.

[Sooda]

You are right, I found out that when I save my map, JassHelper (which should convert vJass to Jass?) puts map's vJass directly into war3map.j.

I compared two maps and found it out, really strange. I have JassHelper enabled and from Grimoire disabled WE Syntax Check.

I will attach map and extracted war3map.j for prove. I think when you open map and save it will save fine. Something wrong with my WE or JNGP.

All worked fine before, I got Win XP Prof with SP3, JNGP 5a and double installation of both RoC and TFT (In Program Files under Warcraft III and Warcraft III 2). One copy is for playing in battle.net and another is for editing maps.

war3map download link from Rapidshare: http://rapidshare.com/files/142568295/war3map.j.html

EDIT:
WE (version 1.21b, 6052) crashes when reaches to 'Generating Map Archive', I unticked all options from both Grimoire and JassHelper, still happens. When I first save Grimoire gives error 'can't find map path' and on second save crashes at 'Generating Map Archive' start.

EDIT: Reinstalling games didn't solve problem. I have RoC downloaded from Blizzard Account and TFT is on original CD, maybe that causes problems, but it worked perfectly fine with such install combo before.

EDIT: Running jasshelper.exe gave error about missing bin\SFmpq.dll, sfmpq.dll was in jasshelper folder, I created bin folder and moved sfmpq.dll into it, changed name to SFmpq.dll and ran jasshelper.exe again, now it gives error "Missing arguments: <commonj> <blizzardj> <mappath>, I can only click OK. Do I have to run jasshelper.exe from cmd line? If yes how?

EDIT: Ah cmd line, I googled how to use it.

EDIT: Good news, I just compiled my map with cmd line!
Used these parameters:

Code:

jasshelper.exe "C:\common.j" "C:\Blizzard.j" "C:\war3map.j" "C:\Archi.w3x"

EDIT: Anyway to fix WE Jasshelper? Or I have to use cmd line from now on?

[cohadar]

Jasshelper command line not working on windows

not even --about works.

[Vexorian]

It would be nonsense, if grimoire can call jasshelper, then jasshelper's command line works...

[PurplePoot]

In the movableEffect example struct in Operator Overloading, you use effectpath twice, where one should be effectpath=.

[cohadar]

I found out what was the problem.
grimoire expects sfmpq.dll to be in same directory as jasshelper
while jasshelper when used alone expects bin/sfmpq.dll

EDIT:
can you please add command line option for macro preprocessor only.

[Vexorian]

Yes, should add an explanation about sfmpq.dll for command line people.

What kind macro exactly?

Sure you don't want --warcity?

[cohadar]

--warcity does not evaluate //! runtextmacro

I need to call macro preprocessor only.

[Vexorian]

I could make one that does novjass, import and textmacro. Only textmacro would be harder.