Re: Testsuite failures of gawk 5.2.0

[arnold]

Hi.

The RiscV issue is really weird. (NaNs are weird.) On x86_64:

$ cat t.c
#include <stdio.h>
#include <math.h>

void
p(double x)
{
        printf("got %g\n", x);
}

int
main(int argc, char **argv)
{
        double d = sqrt(-1.0);

        printf("main: %g\n", d);
        p(d + 0);

        return 0;
}
$ gcc t.c -lm -o t
$ ./t
main: -nan
got -nan

On RiscV, I have to initialize `d' like so:

        double d = -sqrt(-1.0);

$ gcc t.c -lm -o t
$ ./t
main: -nan
got nan

Adding zero changes the sign of the value passed.

Sheesh, what a headache.

I'm not sure yet what to do about this, but at least now I
can point a finger at the hardware.

Arnold