Re: Patch for 9.1: initdb -C option

On Jul 23, 2010, at 6:36 AM, Robert Haas wrote:

> 2010/7/23 KaiGai Kohei <kaigai(at)ak(dot)jp(dot)nec(dot)com>:
>> Sorry for the confusion.
>>
>> What I wanted to say is the patch itself is fine but we need to make consensus
>> before the detailed code reviewing.
>
> I guess we probably need some more people to express an opinion, then.
> Do you have one?
>
> I'm not sure I do, yet. I'd like to hear the patch author's response
> to Itagaki Takahiro's question upthread: "Why don't you use just "echo
> 'options' >> $PGDATA/postgresql.conf" ? Could you explain where the
> -C options is better than initdb + echo?"

At this point, I have no real preference for this patch; it is just as easy to echo line >> datadir/postgresql.conf, so perhaps that makes this patch somewhat pointless. I suppose there's a shaky argument to be made for Windows compatibility, but I'm sure there's also an equivalent functionality to be found in the windows shell.

Reception to this idea has seemed pretty lukewarm, although I think Peter expressed some interest. Some of the previous linked correspondence in the review referred to some of the proposed split configuration file mechanisms. My particular implementation is fairly limited to the idea of a single configuration file, so compared to some of the other proposed approaches including split .conf files, it may not cover the same ground.

Like I said in the original submission, I found it helpful for the programmatic configuration of a number of simultaneous node, but if it's not generally useful to the community at large, I'll understand if it's punted.

Regards,

David
--
David Christensen
End Point Corporation
david(at)endpoint(dot)com