Re: consistent naming of components

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: "Robert Dyas" <rdyas(at)adelphia(dot)net>
Cc: pgsql-hackers(at)postgresql(dot)org
Subject: Re: consistent naming of components
Date: 2001-10-27 17:11:32
Message-ID: 2693.1004202692@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

"Robert Dyas" <rdyas(at)adelphia(dot)net> writes:
> [ rename and move just about everything in sight ]

Sorry, but I don't think this is going to happen. We'd be breaking
a heck of a lot of user applications, startup scripts, etc to achieve
(IMHO) very little of value. Renaming psql->pgsql would alone break
more user scripts than I care to think about.

> change data location /var/lib/pgsql/data to /var/pgsql
> move .conf files from /var/lib/pgsql/data to /etc/pgsql

The present sources do not have any hardwired notion of where things
should go. If you care to install things in those directories, you
can --- but you won't get far insisting that everyone else should do
likewise. Preferred filesystem organization varies across platforms.
Even if it didn't, there are situations such as running multiple
postmasters (eg, setting up a test version) in which some instances
*must* have a nonstandard location.

You might possibly be able to talk the RPM maintainer into changing
his ideas of where the RPMs should install stuff --- but I believe
he thinks he's following the Linux filesystem layout standard
(FHS? forget what it's called exactly). In any case, breaking
backwards compatibility won't be an easy sell.

> Going a bit further in reorganization, if the config files always lived in
> an /etc/pgsql directory, then pgsqld (aka postmaster) could start with zero
> parameters and zero environment variables (true?),

Again, see multiple-postmaster issue. AFAICT you are proposing to
remove flexibility that is *necessary* for some people. (Like me
... I currently have three postmasters of different vintages running
on this machine ...)

regards, tom lane

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Peter Eisentraut 2001-10-27 17:44:05 HISTORY file
Previous Message Tom Lane 2001-10-27 16:46:11 Re: [patch] helps fe-connect.c handle -EINTR more gracefully