Re: Determining the minimum representable increment/decrement possible?

For an Inline::C solution, the following works ok for me with gcc on Windows:

use strict;
use warnings;

use Inline C => Config =>
 BUILD_NOISY => 1,
 ;

use Inline C => <<'EOC';

double nxtafter(double in, double dir) {
  return nextafter(in, dir);
}

EOC

my $pos_inf = (99 ** 99) ** 99;
my $neg_inf = -$pos_inf;

my $next_down = nxtafter(1.9041105342991877e+258, $neg_inf);

printf "%.16e\n", $next_down;

my $next_up = nxtafter(-8.2727285363069939e-293, $pos_inf);

printf "%.16e\n", $next_up;

__END__
Outputs:
1.9041105342991875e+258
-8.2727285363069927e-293
[download]

With the latest MS compiler that I have (14.00.40310.41) I need to replace "nextafter" with "_nextafter" in order to get the script to compile.
But it then produces incorrect results anyway:

1.9041105342991884e+258
-7.9999999999999937e-293
[download]

Perhaps a more recent MS compiler fares better.

Cheers,
Rob

Comment on Re: Determining the minimum representable increment/decrement possible? Select or Download Code

Replies are listed 'Best First'.
Re^2: Determining the minimum representable increment/decrement possible? by BrowserUk (Patriarch) on Jun 16, 2016 at 13:24 UTC
I'm probably being thick here, but where does nextafter() come from? With the rise and rise of 'Social' network sites: 'Computers are making people easier to use everyday' Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error. "Science is about questioning the status quo. Questioning authority". I knew I was on the right track :) In the absence of evidence, opinion is indistinguishable from prejudice. Not understood.	[reply]
Re^3: Determining the minimum representable increment/decrement possible? by syphilis (Archbishop) on Jun 16, 2016 at 14:03 UTC
... where does nextafter() come from? As pryrt noted, it's in math.h - which perl.h already (indirectly) includes. With my MS compiler, I found matches for the string "nextafter" in: `C:\_64\Platform_SDK\Include\crt\float.h C:\_64\Platform_SDK\Include\crt\fpieee.h C:\_64\Platform_SDK\Include\crt\math.h C:\_64\Platform_SDK\src\crt\float.h C:\_64\Platform_SDK\src\crt\fpieee.h` [download] Cheers, Rob	[reply] [d/l]
Re^3: Determining the minimum representable increment/decrement possible? by pryrt (Abbot) on Jun 16, 2016 at 13:43 UTC
it looks like it's in math.h -- I didn't know about it until syphilis's post, either... I love how much I learn here. :-)	[reply]
Re^3: Determining the minimum representable increment/decrement possible? by fullermd (Vicar) on Jul 05, 2016 at 07:29 UTC
For completeness, `nextafter()` is specified as part of `math.h` in C99 (section 7.12.11.3 in the N1256 draft).	[reply] [d/l] [select]