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

[gekko]

guess I found a bug (or misunderstood something).

code:

JASS:

library libbug

    private function init takes nothing returns nothing
        call init1.execute()
    endfunction
    
    private function init1 takes nothing returns nothing
    endfunction

endlibrary

JassHelper says

"init1 is not a type that allows . syntax"

any ideas?

2. having to manually import the AI for all 4 races in 30+ maps is a very tedious job. it would be nice if when using the 4 parameter command line mode that clijasshelper would import the resulting .ai file into the map (and replace any existing one with the same name). this would work like using the command line version for triggers where the actual map is updated with the compiled triggers, but this would import the AI into the map.

I don't know how the import section is stored in the map, but there would need to be a way to specify the path (i.e. scripts/human.ai etc) for the file being imported.

this way we could just use a batch file to update all the maps with the edited ai (as we do when using the command line version when creating triggers).

if this is not possible or important enough to add to clijasshelper, do you know where the map file format (the import section) is posted so that maybe I can try write a command line importer?

Maybe you wanna try my tool GMSI. You can add imports there with just a function call. And you can insert the script into all your maps with just one script execution. One click, AI in every map you want. :)

[Troll-Brain]

Originally Posted by gekko

guess I found a bug (or misunderstood something).

code:

JASS:

library libbug

    private function init takes nothing returns nothing
        call init1.execute()
    endfunction
    
    private function init1 takes nothing returns nothing
    endfunction

endlibrary

JassHelper says

"init1 is not a type that allows . syntax"

any ideas?

you have to use a private keyword like that :

JASS:

library libbug

    private keyword init1

    private function init takes nothing returns nothing
        call init1.execute()
    endfunction
    
    private function init1 takes nothing returns nothing
    endfunction

endlibrary

Or change the position of the functions like that :

JASS:

library libbug
    
    private function init1 takes nothing returns nothing
    endfunction
    
    private function init takes nothing returns nothing
        call init1.execute()
    endfunction
    
endlibrary

[daPanda]

Originally Posted by gekko

Maybe you wanna try my tool GMSI. You can add imports there with just a function call. And you can insert the script into all your maps with just one script execution. One click, AI in every map you want. :)

wow! very cool tool this will solve a LOT of tedious work in other areas that have been driving me nuts (some of the things you mention on your website).

I'll give it a whirl and see how it works with updating the AI imports.
Thanks!

[Dark_Dragon]

vex since you where asked to do linebreaks with semicolons why dont you insted do this?

JASS:

set i = 0
loopuntil (i >= 5)
    call BJDebugMsg(I2S(i))
    set i = i + 1
endloop

its compiled to:

JASS:

set i = 0
loop
 exitwhen (i >= 5)
    call BJDebugMsg(I2S(i))
    set i = i + 1
endloop

that would be awesome and btw why not add this (i know that you dont like c that much but it would be cool)

JASS:

for (integer i = 0, i < 5, i+1)
   call IMsg(i)
endfor

//compiled to

local integer i

// for detected (integer i not jet loaded so load it now)
set i = 0
loop
exitwhen not (i < 5)
   call IMsg(i)
   set i = i + 1
endloop

each time i load a for and ask for integer you must check is it already declared.

JASS:

local integer i
for (i = 0, i < 5, i+1)
   call IMsg(i)
endfor

this would as well be acceptable.

i think that for loop is really missing the jass, so its a suggestion to add it!
it would be good if you add at leat one of them loopuntil or for... if you add both it would be even better...

Greets!
~Dark Dragon

[Bobo_The_Kodo]

Thats not a struct Litany, >.>

[daPanda]

Originally Posted by Dark_Dragon

i think that for loop is really missing the jass, so its a suggestion to add it!
it would be good if you add at leat one of them loopuntil or for... if you add both it would be even better...

Greets!
~Dark Dragon

I'd vote for the for() loop if it came down to choice. much easier and less ambiguous, IMO. As for implementation of it and checking for existing local variables, you could just substitute the user defined name with name mangling (auto generated integer var name).

[C2H3NaO2]

Originally Posted by daPanda

I'd vote for the for() loop if it came down to choice. much easier and less ambiguous, IMO. As for implementation of it and checking for existing local variables, you could just substitute the user defined name with name mangling (auto generated integer var name).

That is really ugly in jass syntax! I would prefer a pascal like syntax.

JASS:

local integer i
for i = 1 to 1
call BJDebugMsg(I2S(i))
endfor

[Dark_Dragon]

Originally Posted by C2H3NaO2

That is really ugly in jass syntax! I would prefer a pascal like syntax.

JASS:

local integer i
for i = 1 to 1
call BJDebugMsg(I2S(i))
endfor

its not like yours is bad i can live with it as well but where is increment?

EDIT: hmm as i remember syntax would be:

JASS:

local integer i

for i = 1 to 10 step 1
endfor

//or

for i = 10 to 0 step -2
endfor

questions for vex!

1) do you like "for loop"? like you said you dont like i++ stuff
2) is any of this syntaxes good enough?
3) do you think you will ever implement any syntax for "for loop"?

its importand to know the answers coz i dont want to keep suggesting it if it bothers u!

[ToukoAozaki]

It would be great if default values on interface methods work with null structs... Currently using them with invalid instances likely to return unexpected values or get the thread killed silently

Also, what about making "local" keyword optional? I'd also like to see something like "static auto/local variables" in C...

[akolyt0r]

Originally Posted by ToukoAozaki

It would be great if default values on interface methods work with null structs... Currently using them with invalid instances likely to return unexpected values or get the thread killed silently

have you tried: defaults 0 ?

If this doesnt work aswell, this should really get fixed..

[ToukoAozaki]

Originally Posted by akolyt0r

have you tried: defaults 0 ?

If this doesnt work aswell, this should really get fixed..

JASS:

library A initializer Init
    interface B
        method operator intValue takes nothing returns integer defaults 0
        method operator boolValue takes nothing returns boolean defaults false
    endinterface
    
    private function OnStart takes nothing returns nothing
        local B b = B(0)
        
        call BJDebugMsg("OnStart")
        
        call BJDebugMsg(I2S(b.intValue))
        if b.boolValue then
            call BJDebugMsg("true")
        else
            call BJDebugMsg("false")
        endif
    endfunction
    
    private function Init takes nothing returns nothing
        call TimerStart(CreateTimer(), 0.0, false, function OnStart)
    endfunction
endlibrary

This yields:

JASS:

//Generated method caller for B._get_intValue
function sc__B__get_intValue takes integer this returns integer
    set f__arg_this=this
    //no struct used this interface. We should probably avoid adding it...
 return f__result_integer
endfunction

//Generated method executor for B._get_intValue
function sx__B__get_intValue takes integer this returns nothing
    set f__arg_this=this
endfunction

//Generated method caller for B._get_boolValue
function sc__B__get_boolValue takes integer this returns boolean
    set f__arg_this=this
    //no struct used this interface. We should probably avoid adding it...
 return f__result_boolean
endfunction

//Generated method executor for B._get_boolValue
function sx__B__get_boolValue takes integer this returns nothing
    set f__arg_this=this
endfunction
//Generated destructor of B
function sc__B_destroy takes integer this returns nothing
    if this==null then
            call DisplayTimedTextToPlayer(GetLocalPlayer(),0,0,1000.,"Attempt to destroy a null struct of type: B")
        return
    elseif (si__B_V[this]!=-1) then
            call DisplayTimedTextToPlayer(GetLocalPlayer(),0,0,1000.,"Double free of type: B")
        return
    endif
    //no struct used this interface. We should probably avoid adding it...
    set si__B_V[this]=si__B_F
    set si__B_F=this
endfunction

//library A:
    
    function A___OnStart takes nothing returns nothing
        local integer b= (0)
        
        call BJDebugMsg("OnStart")
        
        call BJDebugMsg(I2S(sc__B__get_intValue(b)))
        if sc__B__get_boolValue(b) then
            call BJDebugMsg("true")
        else
            call BJDebugMsg("false")
        endif
    endfunction
    
    function A___Init takes nothing returns nothing
        call TimerStart(CreateTimer() , 0.0 , false , function A___OnStart)
    endfunction

//library A ends

[akolyt0r]

ah i missunderstood you, i thought that method interfaces which returned structs didnt work somehow .. NVM

[gekko]

Originally Posted by ToukoAozaki

I'd also like to see something like "static auto/local variables" in C...

I also beg for function static variables! How often do I need one or more globals for just one function (to increase performance over local variables, because these are not needed in every branch of the function and it would be insuficient to alloc memory for them each time the function is called, even if they are only used in one of 100 cases when the function is called).
It is so damn ugly to make a globals block above that function and specify all values there. Instead a line inside the function with something like:
"local static integer i"
would be superb, would make the code look much straighter and really emphasize that this global is only for this function/method.

The compiler just had to remove this line from the function and just make a global declaration out of it. (Of course, most function calls would be impossible in static init since its a global declaration)

[Dark_Dragon]

Originally Posted by gekko

I also beg for function static variables! How often do I need one or more globals for just one function (to increase performance over local variables, because these are not needed in every branch of the function and it would be insuficient to alloc memory for them each time the function is called, even if they are only used in one of 100 cases when the function is called).
It is so damn ugly to make a globals block above that function and specify all values there. Instead a line inside the function with something like:
"local static integer i"
would be superb, would make the code look much straighter and really emphasize that this global is only for this function/method.

The compiler just had to remove this line from the function and just make a global declaration out of it. (Of course, most function calls would be impossible in static init since its a global declaration)

JASS:

local static integer i //? that makes no sense
static integer i // thats fine

second thing if this is not a bug then omg!

JASS:

struct xxx
   method onDestroy takes nothing returns nothing
   endmethod
endstruct

local xxx x = xxx.create()
call x.onDestroy()

one guy who i was teaching vjass wrote that and thats how i found about this bug...

[gekko]

Originally Posted by Dark_Dragon

JASS:

local static integer i //? that makes no sense
static integer i // thats fine

I rather ment this:

JASS:

method bla takes nothing returns nothing   
  local static integer i
 //or maybe just
  static integer i
endmethod