Re: SQL command to edit postgresql.conf, with comments - Mailing list pgsql-hackers

From Tom Lane
Subject Re: SQL command to edit postgresql.conf, with comments
Date
Msg-id 26639.1286983487@sss.pgh.pa.us
Whole thread Raw
In response to Re: SQL command to edit postgresql.conf, with comments  (Andrew Dunstan <andrew@dunslane.net>)
Responses Re: SQL command to edit postgresql.conf, with comments
Re: SQL command to edit postgresql.conf, with comments
List pgsql-hackers
Andrew Dunstan <andrew@dunslane.net> writes:
> +1. Preserving the comments when you change the value could make the 
> comments totally bogus. Separating machine-generated values into a 
> separate file makes plenty of sense to me.

> Which one wins, though? I can see cases being made for both.

IIRC the proposal was that postgresql.conf (the people-editable file)
would have "include postgresql.auto" in it.  You could put that at
the top, bottom, or even middle to control how the priority works.
So it's user-configurable.  I think the factory default ought to
be to have it at the top, which would result in manually edited
settings (if any) overriding SET PERMANENT.

Basically the way I'd like to see this go is that SET PERMANENT
gets attached on the side of what's there now, and people who are
used to the old way don't have to change their habits at all.
If the new way is as much better as its advocates claim, use of
manual editing of postgresql.conf will soon die off; then at some
future date we could consider whether to remove that file or at
least delete all the comments it contains out-of-the-box.  About
the only change I want to make immediately is that initdb ought
to shove its settings into postgresql.auto instead of mucking with
postgresql.conf.
        regards, tom lane


pgsql-hackers by date:

Previous
From: "Kevin Grittner"
Date:
Subject: Re: leaky views, yet again
Next
From: Tom Lane
Date:
Subject: Re: levenshtein_less_equal (was: multibyte charater set in levenshtein function)