To: Grant Finnemore <gaf@ucs.co.za>
cc: Peter Mount <petermount@maidstone.gov.uk>,
	Interfaces <pgsql-interfaces@postgresql.org>
Subject: Re: JDBC 'Unterminated quoted string' 
In-reply-to: <3A0BCC1D.2A2F9F87@ucs.co.za> 
References: 
 <1B3D5E532D18D311861A00600865478CF1B50E@exchange1.nt.maidstone.gov.uk>
	<3A0BCC1D.2A2F9F87@ucs.co.za>
Comments: In-reply-to Grant Finnemore <gaf@ucs.co.za>
	message dated "Fri, 10 Nov 2000 12:21:17 +0200"
Date: Fri, 10 Nov 2000 09:54:43 -0500
Message-ID: <11184.973868083@sss.pgh.pa.us>
From: Tom Lane <tgl@sss.pgh.pa.us>

Grant Finnemore <gaf@ucs.co.za> writes:
> In brief, there seems to be some problem at the backend inserting a
> \0.

char(n) does not support embedded nulls.  Offhand I believe that bytea
is the only datatype that does, and even then you have to write 'em
with a backslash-escape --- ie, send "\000" not an actual null.

Fixing this is not likely to happen soon, if ever, because it would
involve a massive overhaul of the datatype I/O system, with such
side-effects as breaking every user-defined datatype in existence.
Textual representations of datatypes are C strings --- ie,
null-terminated --- and all datatype I/O routines depend on that.

			regards, tom lane