On Mon, Oct 15, 2012 at 9:18 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> Phil Sorber <phil@omniti.com> writes:
>> I would also like it to have a regression test
>> which none of those seem to have.
>
> [ shrug... ] There is nothing in the current regression infrastructure
> that would work for this, so that desire is pie-in-the-sky regardless of
> where you put it in the source tree. Also, PQping itself is exercised
> in every buildfarm run as part of "pg_ctl start", so I don't feel a real
> strong need to test pg_ping separately.
My plan was to borrow heavily from the pg_upgrade test. I want to
verify the exit status based on known database state as presumably
people would be using this for monitoring/load balancing, etc. Hoping
to prevent silly breakage like the help output from returning an
'Accepting Connections' exit status.