Received: from sss2.sss.pgh.pa.us (sss.pgh.pa.us [209.114.166.2]) by hub.org (8.9.3/8.9.3) with ESMTP id SAA15964 for ; Fri, 21 Apr 2000 18:54:47 -0400 (EDT) (envelope-from tgl@sss.pgh.pa.us) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss2.sss.pgh.pa.us (8.9.3/8.9.3) with ESMTP id SAA23688; Fri, 21 Apr 2000 18:54:23 -0400 (EDT) To: Peter Eisentraut cc: Jan Wieck , pgsql-sql@postgresql.org Subject: Re: TOAST (was: BLOB) In-reply-to: References: Comments: In-reply-to Peter Eisentraut message dated "Sat, 22 Apr 2000 00:04:01 +0200" Date: Fri, 21 Apr 2000 18:54:22 -0400 Message-ID: <23685.956357662@sss.pgh.pa.us> From: Tom Lane X-Archive-Number: 200004/247 Peter Eisentraut writes: >> I'll need alot of help to make all our existing types >> toastable, > I'm wondering how transparent all of this will be. What is involved in > making existing types toastable? How does that affect user defined > datatypes now and in the future? I'd like to think that we can fold the TOAST support into the argument-access macros that I plan to be defining for the fmgr rewrite. See previous rants on this subject, eg 14-Jun-99, 23-Oct-99 (in pgsql-hackers archives). Net result for user-defined-datatype authors will be "if you revise your routines, they will be easier to read, more portable, and will support TOASTed values. If you don't, they'll still work about as well (or poorly) as they did before." regards, tom lane