Re: pg_ctl options checking

On Mon, 2006-04-17 at 15:12 -0400, Tom Lane wrote:
> Peter Eisentraut <peter_e(at)gmx(dot)net> writes:
> > Bruce Momjian wrote:
> >> For a command as significant as pg_ctl, I can't see how making it
> >> _convenient_ is a good argument.
>
> > Well, loss of convenience is one argument in opposition to this change
> > but I don't see any argument in _favor_ of this change other than
> > "let's reject these option combinations", some of which seem perfectly
> > valid.
>
> Ignoring irrelevant arguments is a time-honored Unix tradition that
> contributes significantly to the usefulness of cc, for example.
> Would you be happy if cc rejected -D when being used only to link, say?
>
> I hadn't thought about this when Simon submitted the patch, but I'm
> with Peter: we should not reject arguments just because they're not
> relevant. If you can make a case that particular combinations strongly
> suggest user error, then let's reject those cases ... but not a blanket
> prohibition.

AFAICS -l -o on stop and -m on start could be ignored

Mixing options between register and non-registration commands definitely
indicates user error. So does mixing up -w and -W

--
Simon Riggs
EnterpriseDB http://www.enterprisedb.com/