Andres Freund <andres@2ndquadrant.com> writes:
> On 2015-01-14 09:34:23 -0500, Tom Lane wrote:
>> Well, that would only fix my problem if we added a configure-time test
>> for whether gcc recognizes "z", which frankly seems like a waste of
>> cycles. I've probably got the last one left in captivity that doesn't.
> Hm. I had kinda assumed that %z support for sprintf and gcc's
> recognition of the format string would coincide and we could just use
> the %z result. But gull's output doesn't actually that way.
It's only reasonable to assume that gcc matches sprintf if gcc is the
native (vendor-supplied) compiler for the platform. That's not the
case on gaur. It used to be very very commonly not the case, though
I think a lot of vendors have now abandoned their proprietary compilers.
If we were to test for this at all, I think we'd need to make the test
separate from the one for sprintf's behavior.
regards, tom lane