[No] suport for returns on functions not only in the end

[ibm5155]

Well i was wandering if you´d mind inplementing a return that is not required a return only on the end of a function, example

this is a method that i don´t like to use

Code: Select all • Expand view

function int ibm_negativo(int a)
{
  int x;
   if (a<0)
      x=1;
   else x=0;
    return x;
}

here is how it should be (like in c)

Code: Select all • Expand view

function int ibm_negativo(int a)
{
  int x;
   if (a<0)
      return 1;
   else return 0;
}

but in doom builder i can´t compile without a return on the end , so, i think this could help others with functions too, because it doen´t require a little part of memory for cearing x for acumulate just 1 or 0, and even the performance could be helped (if enter on the first if, then he return 1 and stop the function, not testing the rest...)

I´m not going to talk it´s a easy thing to do, but if acs similar to c, i really don´t thing it´s a difficult thing to do, just a little modify on the acs compiler and voala \o/

[Isle]

you could just use the second example and get rid of the else. when it hits a return it skips the rest of the function.

[Graf Zahl]

i really don´t thing it´s a difficult thing to do, just a little modify on the acs compiler and voala \o/

The problem is, that the ACC compiler is not sanely programmed. One can only hack it this far and no further.
I'll leave the decision to Randy though.

[Apothem]

but the else statement in that code is basically what would happen even if you left the else out. When you use return it terminates the function anyway. I fail to see the point of this

[randi]

ACC doesn't maintain enough state to know that both paths through the if return values, so yes it is a difficult thing to do (relatively speaking, anyway). As Isle said, in this particular case, you can just get rid of the else, and the result will be a working function that does the same thing.

[ibm5155]

i didn´t try that method, thanks randi ^^