I wrote:
> I googled a bit and found these recommendations on stackoverflow:
> my $inf = 9**9**9;
> my $neginf = -9**9**9;
> my $nan = -sin(9**9**9);
> These do seem to produce the desired results, at least on the
> couple of Perl versions I checked, including 5.8.3.
Nope, buildfarm shows that still doesn't work everywhere.
If we had a transform function that converted SQL float8 directly
to Perl NV, we could use that to produce NVs containing inf/nan.
But that would be a pretty ridiculous amount of test scaffolding
to create to test what, in the end, is two lines of very
straightforward and unlikely-to-break code.
My recommendation is just to drop these test cases. Getting them
to work on all old Perl versions is more trouble than they're worth.
regards, tom lane