From: Tom Lane <tgl@sss.pgh.pa.us>
To: Alvaro Herrera <alvherre@alvh.no-ip.org>
cc: Stephen Frost <sfrost@snowman.net>, pgsql-hackers@lists.postgresql.org
Subject: Re: pgsql: Fix headerscheck failure in replication/worker_internal.h
In-reply-to: <202111161816.bncq4ckt6ahl@alvherre.pgsql>
References: <202111161816.bncq4ckt6ahl@alvherre.pgsql>
Comments: In-reply-to Alvaro Herrera <alvherre@alvh.no-ip.org>
	message dated "Tue, 16 Nov 2021 15:16:45 -0300"
MIME-Version: 1.0
Content-Type: text/plain; charset="UTF-8"
Content-ID: <2256006.1637093527.1@sss.pgh.pa.us>
Content-Transfer-Encoding: quoted-printable
Date: Tue, 16 Nov 2021 15:12:07 -0500
Message-ID: <2256007.1637093527@sss.pgh.pa.us>
Archived-At: 
 <https://www.postgresql.org/message-id/2256007.1637093527%40sss.pgh.pa.us>
Precedence: bulk

Alvaro Herrera <alvherre@alvh.no-ip.org> writes:
> On 2021-Nov-16, Stephen Frost wrote:
>> Not against possibly changing that but I don=E2=80=99t get the point of=
 including
>> be-gssapi-common.h if it=E2=80=99s not enabled in the build and typical=
ly if GSSAPI
>> is possible and the reason for including be-gssapi-common.h then there=E2=
=80=99s
>> other things that need to be under a ifdef, again, as in auth.c

> BTW, this is exactly why my first suggestion was to add an exclusion
> rule to headerscheck so that be-gssapi-common.h is not verified by that
> script.  After re-reading your response, that looks like a reasonable
> answer too.

I think adding #ifdef ENABLE_GSS as per your prior message is better.
Headers have little business making assumptions about the context in
which they're included --- which is exactly why headerscheck exists ---
so I disagree with Stephen's argument.  In any case I am not in favor of
making random exclusions from that script's testing.

			regards, tom lane