Re: I am done - Mailing list pgsql-hackers

From Tom Lane
Subject Re: I am done
Date
Msg-id 28633.1031010948@sss.pgh.pa.us
Whole thread Raw
In response to Re: I am done  (Peter Eisentraut <peter_e@gmx.net>)
List pgsql-hackers
Peter Eisentraut <peter_e@gmx.net> writes:
> Perhaps one should consider removing the autocommit option.  It's no use
> if it's there but everything breaks when you turn it on.

As far as I'm concerned, it's in there for one reason only (as far as
7.3 goes): so that we can run the NIST SQL compliance tests.  Anyone who
wants to use it in production at this point is doing so at their own
risk.

In practice, as long as we fix libpq's startup SET commands, the major
problems will just be with large object support, which is not mainstream
usage either; so I'm prepared to live with it for a release or so.  It's
not like there aren't any other combinations of PG features that don't
play nice together.
        regards, tom lane


pgsql-hackers by date:

Previous
From: Tom Lane
Date:
Subject: Re: reindex of toast tables
Next
From: Tom Lane
Date:
Subject: Re: Future of --enable-recode?