Integer/Real glitch #2?

[bowser499]

Hello again. Gonna write a post about JASS bug again.
When I've analyzed SirNikolas's version of FICS, I've found a lulz. Here is the code:

    function Trig_FICS_BinSearch takes integer id returns integer
        local integer first = 0
        local integer last = udg_FICS_Length
        local integer mid
        if id >= udg_FICS_Ids[0] and id <= udg_FICS_Ids[last - 1] then
            loop
                exitwhen first >= last
                set mid = (first + last) / 2
                if id <= udg_FICS_Ids[mid] then
                    set last = mid
                else
                    set first = mid + 1
                endif
            endloop
            if id == udg_FICS_Ids[last] then
                return udg_FICS_Num[last]
            endif
        endif
        return 0
    endfunction

I understand that this is a binary searching.
But can someone explain which sorcery made this line: set mid = (first + last) / 2 compile?
It must give an error since the result of the division may be real and I see: local integer mid in the beginning! integer is not real. But / 2 is compiling!
I wonder, how... Bug or a feature?

 

[TriggerHappy]

It just rounds the number up.

Notice you if add a decimal to the end of the 2(2.), you will get a compile error.

This is no bug/glitch

 

[chobibo]

Feature of type integer.

 

[Malhorne]

It rounds to the number below if I remember well...

By the way it is the same in C....

 

[chobibo]

I don't think it rounds values, I think it truncates the fractional part.

script

library IntegerTest initializer init
    
    function FUCKMEPLZ takes nothing returns nothing
        local integer a=R2I(1.2)
        local integer b=R2I(1.8)
        call DisplayTimedTextToPlayer(GetLocalPlayer(),0,0,60.00," a holds value "+I2S(a))
        call DisplayTimedTextToPlayer(GetLocalPlayer(),0,0,60.00," b holds value "+I2S(b))
    endfunction
    
    function init takes nothing returns nothing
        call TimerStart(CreateTimer(),1.00,false,function FUCKMEPLZ)
    endfunction
endlibrary

Screenshot!

 

[WaterKnight]

It just rounds the number up.

3 / 2 -> 1
3 / 4 -> 0
-3 / 4 -> 0
-3 / 2 -> -1

3 / -2 -> -1
3 / -4 -> 0
-3 / -4 -> 0
-3 / -2 -> 1

It takes the next integer whose abs is smaller/how many times the divisor fits wholly in the dividend. Or as chobibo described.

 

[Maker]

Yes, it rounds down.

It would be kind of silly if you could only perform division with even numbers.

 

[muzzel]

Dont you know the expression "Its not a bug, its a feature!".

Well, this is literally the case here...

And about rounding, its simply a "DIV" operation (some languages support that, including WurstScript). Its equal to performing a real division and cutting off the digits behind the decimal point.

 

[chobibo]

I thought it truncates it, darn it. so it's like floor()?

 

[Maker]

I'd say truncating is more accurate that to say rounds down.

 

[TriggerHappy]

You guys are right, haha.

I tend to get things mixed up a lot.

 

[SA Dashie]

I actually like using this "feature" alot.

The same would obviously apply to.

local real r = R2I(2.6)
//r will now equal 2