----- Original Message -----
From: "Orion Poplawski"
a difference in the sign of 0. The -0 result appears to be more "correct"
in that that is what is present in both the C and perl x86_64 tests. This
is pretty baffling to me. Thoughts?
Which version of perl ?
I know that 5.14.0 (and probably later, too) can be particularly bad at
honouring the sign of zero.
For example:
#########################
use warnings;
use Devel::Peek;
use Inline C => <<'EOC';
SV * foo() {
double d = -0.;
return newSVnv(d);
}
EOC
$x = foo();
print $x, "\n";
Dump($x);
#########################
On my build of 5.14.0, the print() will claim that the value is "0", even
though the Dump() correctly shows that the the NV slot contains "-0".
Earlier builds of perl seem fine, however.
If the particular perl was built using one compiler, and
PDL-Graphics-PLplot (or even plplot) was built using a different compiler,
then there's also the possibility that you're seeing a bug in the way that
one of those compilers handles the signed zero .... but I'd mainly be
suspicious of perl.
I think those failures you see invalidate the tests, rather than demonstrate
a failing of the module itself.
Cheers,
Rob
_______________________________________________
Perldl mailing list
[email protected]
http://mailman.jach.hawaii.edu/mailman/listinfo/perldl
