Date: Mon, 11 Oct 1999 12:05:42 -0600 (MDT)
Message-Id: <199910111805.MAA18307@biology.nmsu.edu>
From: Brook Milligan <brook@biology.nmsu.edu>
To: maillist@candle.pha.pa.us
CC: vev@michvhf.com, pgsql-hackers@postgreSQL.org
In-reply-to: <199910111705.NAA05903@candle.pha.pa.us> (message from Bruce
	Momjian on Mon, 11 Oct 1999 13:05:24 -0400 (EDT))
Subject: Re: [HACKERS] Next release is 7.0(?)
References: <199910111705.NAA05903@candle.pha.pa.us>

   > Would this also be a relevant time to get a regression test to run on
   > a non-installed system?  Couldn't this be done by starting up the
   > database on a different port from within the newly compiled source
   > tree for the purpose of testing?

   Gee, I never even though of that.  What advantage would there be for
   that?

Wouldn't that be useful for developing/debugging/testing a new version
on a machine that runs some other version in production mode?

One example would be someone who wants to update, but wishes to verify
the functioning of the new version prior to blowing away the old
one.  So, one could build the new version, run regression tests,
resolve any issues, THEN deinstall the old version and install the new
already verified one.

Another example is that developers could use a production machine to
tweak new code without having to actually install the test versions.

I'm sure other applications of that flexibility are apparent also.

Cheers,
Brook