From owner-freebsd-current Tue Oct 29 16:59:37 1996 Return-Path: owner-current Received: (from root@localhost) by freefall.freebsd.org (8.7.5/8.7.3) id QAA15458 for current-outgoing; Tue, 29 Oct 1996 16:59:37 -0800 (PST) Received: from phaeton.artisoft.com (phaeton.Artisoft.COM [198.17.250.211]) by freefall.freebsd.org (8.7.5/8.7.3) with SMTP id QAA15447 for ; Tue, 29 Oct 1996 16:59:33 -0800 (PST) Received: (from terry@localhost) by phaeton.artisoft.com (8.6.11/8.6.9) id RAA22416; Tue, 29 Oct 1996 17:53:49 -0700 From: Terry Lambert Message-Id: <199610300053.RAA22416@phaeton.artisoft.com> Subject: Re: /var/mail (was: re: Help, permission problems...) To: scrappy@ki.net (Marc G. Fournier) Date: Tue, 29 Oct 1996 17:53:49 -0700 (MST) Cc: terry@lambert.org, MRC@CAC.Washington.EDU, current@FreeBSD.org In-Reply-To: from "Marc G. Fournier" at Oct 29, 96 06:37:04 pm X-Mailer: ELM [version 2.4 PL24] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-current@FreeBSD.org X-Loop: FreeBSD.org Precedence: bulk > The question that *really* needs to be addressed is whether > what Mark is stating is correct (right now, I assume he is) in that > other Uinx variants' mail.local programs use .lock files to deal with > locking. If the general consensus is that this is the case, then IMAP4's > requirement for /var/mail to be 1777 is justified. Actually, this is not a requirement. It's just the default for the code (read the "docs/locking" document for the IMAP4 distribution. Mark wrote the document: it's actually a compilation default that shouldn't be default for FreeBSD. The problem seems to be that the "driver" support is for .lock and flock() (not fcntl()) based locking only. Seems someone in the FreeBSD camp should write an fcntl() aware "driver" for IMAP4. 8-(. > If it is justified, then we should take whatever measures are > required to get rid of the risk associated with having /var/mail 1777, > which, as of yet, I haven't heard exactly what the risk involved is. The risk is that incoming data will get stomped by the reader. Mark discusses this at length in his "Jeremiad" against System V. 8-). Contrary to my previous posting, it looks like he is, indeed, implementing a user agent. It's unclear whether it's also a delivery agent (it seems to be, for a client-initiated modification of a mailbox item!), and if so, why it doesn't go through mail.local -- like it's "supposed to". Ideally, he'd cut down the interface to let sendmail reasonably consume it to stor RFC 822 messages which might have message bodies which are MIME data (sendmail doesn't care, and the interfaces are too complex to presume encapsulation, it seems). > Now, Terry mentions mail readers such as ELM, which, from the > last time I configured/installed it a couple of years ago, actually > gives you a choice of 3 locking methods, none of which are forced > upon you...'.lock' being one of the three, and I think fcntl/flock were > the other two... They are. I'm biased, since I use "elm" with "MetaMail" to make it a good MIME citizen, despite elm being on S. Kramer's "hated client list" (ftp://ftp.cac.washington.edu/mail/imap.software). Regards, Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers.