🚀 go-pugleaf

In article <1992Dec09.011729.20954@Veritas.COM>, joshua@Veritas.COM (Joshua Levy) writes:
> Consider the following code:
>    int same(int b) { return b ; }
>    if (X)
>    if (same(!!(X)))
>    if (same((int)(X))))
>    X can be any legal C code including something like "f == 4.0"
>    where f is a float or a double.
> 
> If I have an ANSI C compiler, are either of the second two if statements
> identical in behavior to the first if statement?

Yes, the second one is.

First, note that your same() function is, at best, a long-winded way to
write an (int) cast and, at worst, a noop.  Thus, the second and third
if statements are equivalent to:
	if ((int)!!(X))
	if ((int)(int)(X))
Since the result of the ! operator is already an int and you already had
an explicit (int) cast in the third if statement, same() is a complete
noop in both of the cases which then degerate to:
	if (!!(X))
	if ((int)(X))

Now, let's write out the implicit comparisons with 0 explicitly:
	if ((X) != 0)
	if (!((X) == 0))  => if ((X) != 0)
	if ((int)(X) != 0)

It's now obvious that the first and second statements are identical and
the third satement is also identical *if X has an integral value*.  If X
has a non-integral value like 0.1, the first two statements will be true
but the third will be false.

> I'm also interested in knowing if either of the second two if statements
> will have different behavior in any common K&R C compiler.

No, but a K&R compiler will not automatically convert the argument to
same() to int the way an ANSI compiler will (given a prototype).
----
Larry Jones, SDRC, 2000 Eastman Dr., Milford, OH  45150-2789  513-576-2070
larry.jones@sdrc.com  or  ...uunet!sdrc!larry.jones
I sure wish I could get my hands on some REAL dynamite. -- Calvin