🚀 go-pugleaf

pjh@cs.tut.fi (Haavisto Petri) writes:
>I just spent several hours debugging a simple program and before
>crying Compiler Bug I would like to make sure that there is one.
>I don't have the standard, but it seems clear to me that given an
>expression of type double with a positive value, the following
>should always be true:
>		(long) (expression) == (long) floor(expression)
>Does the standard guarantee this or is there a loop hole that has to
>do with finite precision arithmetic and which allows the compiler to
>do what it chooses?

The most obvious thought is:  do you have a proper prototype for floor()
in scope at that point?  If not, you can expect random results.

The second most obvious thought:
"If the integral part of the double cannot be represented by the integral
type, the behavior is undefined."  So, if the expression's value is 5.0E+9
on a system with 32-bit longs, then acceptable runtime behavior would include
growing long red tentacles and slapping you silly, or returning the current
time, or whatever (especially including calculating two different numbers,
and there are plausible reasons why it could).

If floor() is in scope and properly declared, and the expression is in the
range of the representation of long, then you have a compiler bug (which 
will surprise no one :-).