wc3edit.net

Alright JASS coders, I have an interesting little puzzle here.

This is a function that exists in DotA Allstars v6.83d-fixed. You can extract the .j and then find this function

Now, this function has a return at the top and then more code below it (which is obviously unreachable) and another return statement at the bottom. It should be functionally equivalent to "return true". Any ideas on why code like this would ever be written? Maybe just to troll people like me?

This sort of JASS compilation trick was supposed to be patched year ago when the return bug was fixed.

Do you guys have any theories on why this compiles?

Mine:


Bonus

DotA Allstars also defines the following global variable:

Later on in function VJ1, it defines the following local:


Why do these variables not collide?
Does JASS actually allow this if they are different datatypes? I feel like it doesn't, and it shouldn't.
This function does get called. So my theory above doesn't apply.
Maybe it allows it if one is an array and one isn't, since compiler confusing is impossible there.

It's not related to not being called. If there's a function wrote there, it'll be compiled. Seems like you can do multiple returns since you return the correct datatype. About the collide thing, it's not related to datatypes. You can do a local integer A and a global integer A that it wont interfer anything.

Interesting - I tested and you're correct. It does allow multiple return values all the time

None of the other languages I code would allow you to do this. I guess it's just another way that JASS is poorly made.

And I guess if you declare a local with the same name as a global, then you just can't modify the global in that method. You can't exactly do "this.variable" in JASS.
I guess that makes sense.