Vitaly Burovoy <vitaly.burovoy@gmail.com> writes:
> On 1/31/16, Tom Lane <tgl@sss.pgh.pa.us> wrote:
>> 2. POSIX:2008 only requires that "The isinf() macro shall return a
>> non-zero value if and only if its argument has an infinite value."
> Ok, then I'll use "is_infinite" from "float.c".
Yeah, that's good.
> But why functions' (in "src/port/isinf.c") behavior are different? It
> is a bit confusing…
Probably the authors of those different implementations were making it
match the behavior of whatever isinf() they had locally. I don't feel
a need to change isinf.c --- it should be zero-maintenance at this
point, especially if we suspect it is no longer used anywhere.
regards, tom lane