Re: IEEE 754 in GPC

lanceboyle@qwest.net wrote:

What is the relationship between GPC and IEEE 754 floating point?

Platform dependent, see http://www.gnu-pascal.de/crystal/gpc/en/thread10316.html.

THINK and CW (both Mac dialects) allowed e.g.

a := Inf;
a := -Inf;
a := NaN;

but these are rejected by GPC as undeclared identifiers.

You can use the nan from fp.pas in the Mac OS X Pascal Interfaces

{ * nan() * * Availability: * Non-Carbon CFM: in MathLib 1.0 and later * CarbonLib: in CarbonLib 1.0 and later * Mac OS X: in version 10.0 and later }

function nan(tagp: ConstCStringPtr): Double; external name 'nan';

or the BSD one (see "man nan" or http://developer.apple.com/documentation/Darwin/Reference/ManPages/man3/nan.3.html which looks to be essentially the same.

NAN(3) BSD Library Functions Manual NAN(3)

NAME nan -- generate a quiet NaN

SYNOPSIS #include <math.h>

double nan(const char *tagp);

long double nanl(const char *tagp);

float nanf(const char *tagp);

The gcc compiler does have Inf and NaN, see e.g. the compiler sources gcc-42/gcc-4.2-20060408/gcc/testsuite/gcc.c-torture/execute/ieee, builtin-nan-1.c, inf-1.c and inf-2.c

double n1 = __builtin_nan("0x1"); double n2 = __builtin_nan("0X1");

float fi = __builtin_inff(); double di = __builtin_inf(); long double li = __builtin_infl();

float fh = __builtin_huge_valf(); double dh = __builtin_huge_val(); long double lh = __builtin_huge_vall();

I believe you can't use gcc builtins directly from gpc, except by linking in a few lines of C.

I can get the same effect by e.g.

a := MaxReal * 2.0; {inf}
a := 0.0 / 0.0;     {nan}

but this seems like a hack.

Also, the old SANE (Apple's most excellent 754 implementation from way back--I think the guru Kaplan was involved, as he was with the also-most-excellent HP-15c pocket calculator) allowed returning specific numeric codes to allow determining what caused a NaN. I remember writing an arctan2 function and in the case of arctan2(0.0, 0.0), I assigned a returm value of something like NaN(85), which was somehow a NaN even though the construct looks like a function. I _think_ I could then query that particular NaN to get a clue about where it arose. I lost this when porting to CW, however.

Yes, he wrote an amusing foreword for the Apple Numerics Manual, Second Edition (Addison Wesley 1988).

Regards,

Adriaan van Os