Re: Notes on implementing URI syntax for libpq

Excerpts from Peter Eisentraut's message of Thu Nov 24 22:05:09 +0200 2011:
>
> On tor, 2011-11-24 at 15:43 +0200, Alexander Shulgin wrote:
> > Huh? The service definitions are read from a local pg_service.conf,
> > and are specified by setting PGSERVICE (and PGSERVICEFILE) environment
> > variables, no?
> >
> > What would you do with such URI if you need to other people to connect
> > to the same service? Send them URI along with the pg_service.conf?
>
> A full URI would also rely on host names or IP addresses being the same
> everywhere. It's all a matter of degree ...

True, but it is much more reasonable to expect that hostnames will resolve to the same addresses most of the time (save for zone changes propagation time.)

Still I can imagine where this may be useful, like local networks with shared pg_service.conf files. And since we don't need to do anything special to support the behavior (i.e. postgresql:///?service=foo is going to work out of the box,) this seems to be a non-problem.

--
Alex