Local variables non-local?

06-13-2006, 08:09 PM

My trigger works, looping until the a stag comes into range (64 range). However, whenever one is saved, all currently dead stags are saved, despite all variables used being local variables.

JASS:

function IsStagBool takes nothing returns boolean
    if GetUnitTypeId(GetFilterUnit()) == 'u000' then
        if GetWidgetLife(GetFilterUnit()) > 0.405 then
           return true
        endif
    endif
    return false
endfunction

function Trig_Stag_Dies_Conditions takes nothing returns boolean
    return GetUnitTypeId(GetDyingUnit()) == 'u000'
endfunction

function Trig_Stag_Dies_Actions takes nothing returns nothing
    local player p = GetOwningPlayer(GetDyingUnit())
    local real x = GetUnitX(GetDyingUnit())
    local real y = GetUnitY(GetDyingUnit())
    local unit u = CreateUnit(p, 'u001', x, y, GetUnitFacing(GetDyingUnit()))
    local group g = CreateGroup()
    local boolexpr b = Condition(function IsStagBool)
    call RemoveUnit(GetDyingUnit())
    call GroupEnumUnitsInRect(g, GetPlayableMapRect(), b)
    if IsUnitGroupEmptyBJ(g) then
        call TriggerExecute( gg_trg_Victory )
    endif
    call GroupClear(g)
    loop
        call GroupEnumUnitsInRange(g, x, y, 64, b)
        exitwhen (FirstOfGroup(g) == null) == false
        call GroupClear(g)
        call PolledWait(0.1)
    endloop
    if GetPlayerSlotState(p) == PLAYER_SLOT_STATE_PLAYING then
        call DisplayTextToForce( GetPlayersAll(), ( GetPlayerName(GetOwningPlayer(u)) + " has been rescued." ) )
        call CreateUnit(p, 'u000', x, y, GetUnitFacing(u))
        call PingMinimapForForce( GetPlayersAll(), x, y, 3 )
    endif
    call RemoveUnit(u)
    call DestroyBoolExpr(b)
    call DestroyGroup(g)
    set p = null
    set u = null
    set g = null
endfunction

//===========================================================================
function InitTrig_Stag_Dies takes nothing returns nothing
    set gg_trg_Stag_Dies = CreateTrigger(  )
    call TriggerRegisterAnyUnitEventBJ( gg_trg_Stag_Dies, EVENT_PLAYER_UNIT_DEATH )
    call TriggerAddCondition( gg_trg_Stag_Dies, Condition( function Trig_Stag_Dies_Conditions ) )
    call TriggerAddAction( gg_trg_Stag_Dies, function Trig_Stag_Dies_Actions )
endfunction

Any ideas as to what is causing it?

06-13-2006, 08:21 PM	#2
blu_da_noob	As a quick note: JASS: exitwhen (FirstOfGroup(g) == null) == false Could be: JASS: exitwhen (FirstOfGroup(g) != null) Are all these dead stags in the same place?

06-13-2006, 08:46 PM

Captain Griffen

Quote:

Are all these dead stags in the same place?

No, they're not. It will still happen to all of them even if they are on the other side of the map from each other.

And yes, there are a few things I could optimise a bit (one of my first pure JASS codes), but thanks for pointing it out.

06-13-2006, 09:45 PM

Vexorian

That trigger runs for every stag that dies. So it has plenty of sense all of them are saved when a unit comes in range. You might want to do something so the detected unit is ignored later.

Also there are some issues with the code, replace IsUnitGroupEmptyBJ(g) with a FirstOfGroup check , IsUnitGroupEmpty calls the count function (which is a long process) then compares the result with 0 , it is kind of lame.

Quote:

No, they're not. It will still happen to all of them even if they are on the other side of the map from each other.

Didn't note that, well this is weird, let me check again,

--
Tried with a 10 for the range enum instead of 64?

..
Also if you still think that the locals became non-local try to show a message "(x="+R2S(x)+","y="+R2S(y) whenever a stag is rescued

06-14-2006, 11:58 AM

Captain Griffen

Quote:

Tried with a 10 for the range enum instead of 64?

No, I'll give it a go, but I highly doubt it will make any difference.

Quote:

Also if you still think that the locals became non-local try to show a message "(x="+R2S(x)+","y="+R2S(y) whenever a stag is rescued

Good idea, I'll put in some debug messages to show the state of the variables.

EDIT: Solved the problem. Destroying the boolexpr destroyed it for all currently running triggers. Any ideas why?

06-14-2006, 01:16 PM	#6
Vexorian	hmnn, could you show I2S(H2I(b) ) after the declaration of the boolexpr and see if they aren't equal?

06-16-2006, 03:43 PM	#7
blu_da_noob	Perhaps all boolexpr's of a given function all point to the same thing? :S

06-16-2006, 03:44 PM	#8
Vexorian	blizz wouldn't have made DestroyBoolExpr if that was true