Re: quiet inline configure check misses a step for clang

On 2014-04-03 10:15:33 -0400, Tom Lane wrote:
> Andres Freund <andres(at)2ndquadrant(dot)com> writes:
> > On 2014-04-03 09:43:20 -0400, Tom Lane wrote:
> >> I object to the latter; you're proposing to greatly increase the warning
> >> noise seen with any compiler that issues a warning for this without caring
> >> about .h vs .c. For somebody who finds gcc -pedantic unusable, I would
> >> think you'd have a bit more sympathy for people using other compilers.
>
> > Yea, but which compilers are that? The only one in the buildfarm I could
> > find a couple weeks back was acc, and there's a flag we could add to the
> > relevant template that silences it. I also don't think that very old
> > platforms won't usually be used for active development, so a louder
> > build there doesn't really have the same impact as noisy builds for
> > actively developed on platforms.
>
> Didn't we already have this discussion last year? The main points
> are all mentioned in
>
> http://www.postgresql.org/message-id/CA+TgmoYJnc4B+8y01gRnAL52GTPbzc3ZsC4SdNW4LGxHqT3Bgg@mail.gmail.com

To which I replied:
http://www.postgresql.org/message-id/20131224141631.GF26564@alap2.anarazel.de

It really seems like an odd behaviour if a compiler behaved that
way. But even if some decade+ old compiler gets this wrong: I am not
going to shed many tears. We're talking about HP-UX's ac++ here. If
binaries get a bit more bloated there...

I really am not trying to win the inline fight here through the
backdoor, I only want to make clang use inlines again. As just written
in the other message, I just don't see any easy and nice way to write
the autoconf test more robustly.

Greetings,

Andres Freund

--
Andres Freund http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services