From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Thu, 3 May 2001 06:24:11 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Thu, 3 May 2001 06:24:00 -0400 Received: from f00f.stub.clear.net.nz ([203.167.224.51]:29968 "HELO metastasis.f00f.org") by vger.kernel.org with SMTP id ; Thu, 3 May 2001 06:23:50 -0400 Date: Thu, 3 May 2001 22:23:47 +1200 From: Chris Wedgwood To: Alan Cox Cc: Richard Gooch , Ingo Oeser , Jonathan Lundell , linux-kernel@vger.kernel.org Subject: Re: [PATCH] adding PCI bus information to SCSI layer Message-ID: <20010503222347.A26495@metastasis.f00f.org> In-Reply-To: <20010503205754.A26313@metastasis.f00f.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: ; from alan@lxorguk.ukuu.org.uk on Thu, May 03, 2001 at 10:49:02AM +0100 X-No-Archive: Yes Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Thu, May 03, 2001 at 10:49:02AM +0100, Alan Cox wrote: No collisions. There are other good reasons you dont want to do forks on files that way (you need to enumerate them but if you make them directories all hell breaks loose). Hmm... so how else do you support n-forks cleanly preserving existing semantics? I'm not sure I like Linus' suggestion, but I can't think of anything better. --cw (On the whole, I think forks are horrible enough that we shouldn't encourage them further than we do already with .AppleDouble, and similar solutions... but, opinions vary greatly here)