From: | Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us> |
---|---|
To: | PostgreSQL-development <pgsql-hackers(at)postgreSQL(dot)org> |
Cc: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
Subject: | Re: pg_dump and backslash escapes |
Date: | 2006-05-17 13:35:31 |
Message-ID: | 200605171335.k4HDZVT13531@candle.pha.pa.us |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
The basic issue is that we need to set standard_conforming_strings to
'off' for loading into newer releases, >= 8.2, but that SET command is
going to generate an error even dumping/loading into the same version of
PostgreSQL, like 7.3 to 7.3. I don't think we want that, do we? I agree
we can have errors when doing cross-version dumps, but have we ever
generated errors when dumping/reloading into the same version?
We don't need to set escape_string_warning because it is just a warning
message and the warning will only happen when loading into 8.2 or later,
so we could skip that part of the patch. But, I figured as long as we
are suppressing warnings at that point, might as well add that too.
---------------------------------------------------------------------------
Bruce Momjian wrote:
> Tom Lane wrote:
> > Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us> writes:
> > > I have seen no reply to my suggestion below, so I assume it is the way
> > > people want to go for 7.3, 7.4, and 8.0.
> >
> > I'm not particularly for it, if that's what you meant, and certainly not
> > for hacking up old branches that way. For one thing, you can't
> > retroactively cause servers that are already out there to not spit
> > errors for GUC variables they've not heard of; and even if you had such
> > a time-travel machine at hand, it's far from clear that it'd be a good
> > idea.
> >
> > The pg_dump philosophy for cross-version updates is generally that the
> > dump should load if you are willing to ignore errors and press on. Not
> > that there will never be errors. See for example our previous handling
> > of the without_oids business, or search_path, or tablespaces.
>
> So, we should SET the variables and allow people to get the errors on
> load? And not supress them from the client and server logs? Is that
> better than suppressing them?
>
> --
> Bruce Momjian http://candle.pha.pa.us
> EnterpriseDB http://www.enterprisedb.com
>
> + If your life is a hard drive, Christ can be your backup. +
>
> ---------------------------(end of broadcast)---------------------------
> TIP 6: explain analyze is your friend
>
--
Bruce Momjian http://candle.pha.pa.us
EnterpriseDB http://www.enterprisedb.com
+ If your life is a hard drive, Christ can be your backup. +
From | Date | Subject | |
---|---|---|---|
Next Message | Zdenek Kotala | 2006-05-17 13:44:22 | What default is - SET behavior |
Previous Message | Martijn van Oosterhout | 2006-05-17 12:53:05 | Re: Return results for PQexec vs PQexecP* |