Re: upgrade from V5R2 to V5R3.

Hi Lim,

Is this the actual code? I.e. did you really have a variable named "num" that was uninitialized and assigned to itself??

Or is this "simplified" code to make it easier for us to understand? I'm a little concerned that you might've changed the logic a little bit when you "simplified" it.

int num;
while (num=num) {
. . . . . . }

This code says the following:

a) Define an integer named "num".

b) Note: "num" has not been set to anything. Unlike RPG, C does not automatically set variables to zero when they're created, so "num" could potentially have ANY value when this program starts. It might be zero, it might not. Are you sure that in the real program, nothing ever set the value of "num"?

c) Assign the value of num to itself. This is where I'm wondering if you changed the logic of the code a bit. It makes no sense to assign a variable to itself.

d) loop while the value that was assigned in step C isn't 0.

Are you sure your code wasn't more like this:

int x,y;
... other code is here. it sets the value of x...
while (y=x) {
. . .
}

That would make far more sense.

At any rate, I can't imagine that this is really a release upgrade thing. The behavior of a while loop with assignment, like the one you posted, has been well-defined since the 1970's. If V5R2 handled it wrong, there'd be a lot of software that didn't work.

Also, I really can't see why you'd want to do this:

int num;
while (num==num) {
}

That would be silly. Why define a variable named "num" and then check it against itself in a loop? It'd always be equal to itself! Seems a little weird to do something like that. If you really want an infinite loop, then code it a little more clearly.

while (1) {
... do indefinitely ...
}

or

for (;;) {
... do indefinitely ...
}