Horton

Re: Adding a pg_servername() function

Tom Lane <tgl@sss.pgh.pa.us>

View thread View on postgr.es
From: Tom Lane <tgl@sss.pgh.pa.us>
To: GF <phabriz@gmail.com>
Cc: Peter Eisentraut <peter@eisentraut.org>, Laetitia Avrot <laetitia.avrot@gmail.com>, Christoph Moench-Tegeder <cmt@burggraben.net>, Matthias van de Meent <boekewurm+postgres@gmail.com>, Pg Hackers <pgsql-hackers@postgresql.org>
Date: 2023-08-09T16:31:35Z
Lists: pgsql-hackers 
GF <phabriz@gmail.com> writes:
> On Wed, 9 Aug 2023 at 16:05, Tom Lane <tgl@sss.pgh.pa.us> wrote:
>> I actually do object to this, because I think the concept of "server
>> name" is extremely ill-defined

> But the gethostname() function is well defined, both in Linux and in
> Windows.

Sure, its call convention is standardized.  But I see nothing in POSIX
saying whether it returns a FQDN or just some random name.  In any
case, the bigger issue is that I don't really want us to expose a
function defined as "whatever gethostname() says".  I think there will
be portability issues on some platforms, and I am dubious that that
definition is what people would want.

One concrete reason why I am doubtful about this is the case of
multiple PG servers running on the same machine.  gethostname()
will be unable to distinguish them.

			regards, tom lane