From: Tom Lane <tgl@sss.pgh.pa.us>
To: Japin Li <japinli@hotmail.com>
cc: Andres Freund <andres@anarazel.de>,
        PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Subject: Re: Cannot find a working 64-bit integer type on Illumos
In-reply-to: <ME3P282MB3166D876E9CB83009727A5ECB6312@ME3P282MB3166.AUSP282.PROD.OUTLOOK.COM>
References: <ME3P282MB3166F9D1F71F787929C0C7E7B6312@ME3P282MB3166.AUSP282.PROD.OUTLOOK.COM> <20240322165305.6zrtxcmzdrywvmsu@awork3.anarazel.de> <ME3P282MB3166D876E9CB83009727A5ECB6312@ME3P282MB3166.AUSP282.PROD.OUTLOOK.COM>
Comments: In-reply-to Japin Li <japinli@hotmail.com>
	message dated "Sat, 23 Mar 2024 01:11:31 +0800"
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <944093.1711128356.1@sss.pgh.pa.us>
Content-Transfer-Encoding: quoted-printable
Date: Fri, 22 Mar 2024 13:25:56 -0400
Message-ID: <944094.1711128356@sss.pgh.pa.us>
Archived-At: <https://www.postgresql.org/message-id/944094.1711128356%40sss.pgh.pa.us>
Precedence: bulk

Japin Li <japinli@hotmail.com> writes:
> On Sat, 23 Mar 2024 at 00:53, Andres Freund <andres@anarazel.de> wrote:
>> Likely there's an unrelated warning triggering the configure test to
>> fail. We'd need to see config.log to see what that is.

> Thanks for your quick reply. Attach the config.log.

Yup:

conftest.c:139:5: error: no previous prototype for 'does_int64_work' [-Wer=
ror=3Dmissing-prototypes]
  139 | int does_int64_work()
      |     ^~~~~~~~~~~~~~~
cc1: all warnings being treated as errors
configure:17003: $? =3D 1
configure: program exited with status 1

This warning is harmless normally, but breaks the configure probe if
you enable -Werror.

No doubt we could improve that test snippet so that it does not
trigger that warning.  But trying to make configure safe for -Werror
seems like a fool's errand, for these reasons:

* Do you really want to try to make all of configure's probes proof
against every compiler warning everywhere?

* Many of the test snippets aren't readily under our control, as they
are supplied by Autoconf.

* In the majority of cases, any such failures would be silent, as
configure would just conclude that the feature it is probing for
isn't there.  So even finding there's a problem would be difficult.

The short answer is that Autoconf is not designed to support -Werror
and it's not worth it to try to make it do so.

			regards, tom lane