Re: Database restore speed

And how do we compose the binary data on the client? Do we trust that the client encoding conversion logic is identical to the backend's? If there is a difference, what happens if the same file loaded from different client machines has different results? Key conflicts when loading a restore from one machine and not from another?
- Luke
--------------------------
Sent from my BlackBerry Wireless Device

-----Original Message-----
From: Mitch Skinner <mitch(at)egcrc(dot)net>
To: Luke Lonergan <LLonergan(at)greenplum(dot)com>
CC: Stephen Frost <sfrost(at)snowman(dot)net>; David Lang <dlang(at)invendra(dot)net>; Steve Oualline <soualline(at)stbernard(dot)com>; pgsql-performance(at)postgresql(dot)org <pgsql-performance(at)postgresql(dot)org>
Sent: Fri Dec 02 22:26:06 2005
Subject: Re: [PERFORM] Database restore speed

On Fri, 2005-12-02 at 13:24 -0800, Luke Lonergan wrote:
> It's a matter of safety and generality - in general you
> can't be sure that client machines / OS'es will render the same conversions
> that the backend does in all cases IMO.

Can't binary values can safely be sent cross-platform in DataRow
messages? At least from my ignorant, cursory look at printtup.c,
there's a binary format code path. float4send in utils/adt/float.c uses
pq_sendfloat4. I obviously haven't followed the entire rabbit trail,
but it seems like it happens.

IOW, why isn't there a cross-platform issue when sending binary data
from the backend to the client in query results? And if there isn't a
problem there, why can't binary data be sent from the client to the
backend?

Mitch