Received: from h97.c194.tor.velocet.net (H97.C194.tor.velocet.net [216.138.194.97]) by postgresql.org (Postfix) with ESMTP id 129AC475A9D for ; Mon, 4 Mar 2002 16:43:27 -0500 (EST) Received: (qmail 43077 invoked by uid 0); 4 Mar 2002 21:43:27 -0000 Received: from nat.inquent.com (HELO jester) (216.6.14.45) by 192.168.1.11 with RC4-MD5 encrypted SMTP; 4 Mar 2002 21:43:27 -0000 Message-ID: <00f801c1c3c5$b00d7490$b002000a@jester> From: "Rod Taylor" To: , "Tom Lane" Cc: , References: <200203042004.g24K4lk79699@lists.unixathome.org> <7581.1015274302@sss.pgh.pa.us> Subject: Re: [HACKERS] Uniqueness of rule, constraint, and trigger names Date: Mon, 4 Mar 2002 16:40:16 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2600.0000 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000 X-Archive-Number: 200203/79 X-Sequence-Number: 6960 If pgupgrade can be fixed to work, perhaps it could set off warnings on items that need to be corrected in a 'schema upgradability test' which will ensure that the user can upgrade it properly -- it shouldn't upgrade if it can't guarentee an upgrade will succeed. This should include full schema test (that whole bad schema entry stuff that pg_dump is supposed to work around) too. Something I'm thinking about digging into. -- Rod Taylor This message represents the official view of the voices in my head ----- Original Message ----- From: "Tom Lane" To: Cc: ; Sent: Monday, March 04, 2002 3:38 PM Subject: Re: [HACKERS] [SQL] Uniqueness of rule, constraint, and trigger names > "Dan Langille" writes: > > On 4 Mar 2002 at 14:24, Tom Lane wrote: > >> ... but I'm a little worried about the possibility of errors > >> in loading schemas from existing databases, where there might be > >> non-unique constraint names. > > > Create a tool to generate unique constraint names during a dump. > > And then all we need is a time machine, so we can make existing > instances of pg_dump contain the tool? It's not that easy ... > > I am not sure that there's really a problem here, because I don't > think duplicate constraint names will be generated during plain > CREATE operations. However, an ALTER TABLE might leave you with > a problem. Hard to tell if this is critical enough to worry about. > > regards, tom lane > > ---------------------------(end of broadcast)--------------------------- > TIP 3: if posting/reading through Usenet, please send an appropriate > subscribe-nomail command to majordomo@postgresql.org so that your > message can get through to the mailing list cleanly >