Re: Regression tests fail with musl libc because libpq.so can't be loaded

Tom Lane:
> Thomas Munro <thomas(dot)munro(at)gmail(dot)com> writes:
>> Just a thought: if we want to go this way, do we need a new exec call?
>> We already control the initial exec in pg_ctl.c.
>
> I'm resistant to assuming the postmaster is launched through pg_ctl.
> systemd, for example, might well prefer not to do that, not to
> mention all the troglodytes still using 1990s launch scripts.

Right, the systemd example in the docs is not using pg_ctl.

But, it should be easily possible to have:
- pg_ctl call postgres with a padded argv0
- postgres call itself with padding, if it wasn't called with that already

This way, there would be no additional exec call when started through
pg_ctl, but one more call when started directly.

Best,

Wolfgang