IsNan and IsInf Functions

[SMcNeill]

If one ever needs to determine if a calculation generates a Nan result or an INF result, they can always use the following little snippet of code:

Save as isnan.h:

Code: (Select All)

#include <stdio.h>
#include <math.h>
#include <float.h>
int IsNan (long double n);
int IsInf (long double n);
int IsNan (long double n) { return -isnan(n); }
int IsInf (long double n) { return -isinf(n); }

And the test code to check for NaN and Inf:

Code: (Select All)

DECLARE LIBRARY "isnan"

    FUNCTION IsNan% (BYVAL n AS _FLOAT)

    FUNCTION IsInf% (BYVAL n AS _FLOAT)

END DECLARE

PRINT IsNan(0 / 0), IsInf(0 / 0)

PRINT IsInf(1 / 0), IsInf(1 / 0)

PRINT IsInf(0 / 1), IsInf(0 / 1)

Notice the output is:
0 / 0 is NaN but is *not* INF.
1 / 0 is NaN and is also INF.
0 / 1 is neither NaN, nor is it INF.

[Dimster]

Thanks for this Steve. Don't mean to highjack this thread but it also demonstrates the use of different kinds of brackets and I was wondering the correct use of them. They are (), [], and {}. The last one in particular I haven't used at all.

[mnrvovrfc]

Only the rounded ones, that you get pressing `SHIFT+9` and `SHIFT+0` have anything to do with this BASIC dialect.

In a few BASIC dialects and other languages, the square brackets is sometimes used like `MID$()` for one character. But not in QB64 at the moment.

Don't let yourself confounded by B.A.M., it was Charlie's choice to use other brackets there and only there.

[SMcNeill]

Updated original post so it works in QB64PE's newest versions.  When we swapped the compiler out a while back, the newer version is a little pickier on syntax than the old one was.