Received: from localhost (daemon@localhost) by CS.UTK.EDU with SMTP (cf v2.9s-UTK) id NAA06463; Fri, 9 May 1997 13:31:54 -0400 Received: by CS.UTK.EDU (bulk_mailer v1.7); Fri, 9 May 1997 13:30:17 -0400 Received: by CS.UTK.EDU (cf v2.9s-UTK) id NAA06181; Fri, 9 May 1997 13:30:14 -0400 Received: from moon.nbn.com (moon.nbn.com [199.4.65.1]) by CS.UTK.EDU with ESMTP (cf v2.9s-UTK) id NAA06137; Fri, 9 May 1997 13:30:05 -0400 Received: from candle.brasslantern.com (schaefer@zagzig.zanshin.com [206.155.48.241]) by moon.nbn.com (8.8.2/8.8.2/MOON) with ESMTP id JAA03264; Fri, 9 May 1997 09:14:26 -0700 (PDT) Received: (from schaefer@localhost) by candle.brasslantern.com (8.7.6/8.7.3) id JAA03121; Fri, 9 May 1997 09:13:51 -0700 From: "Bart Schaefer" Message-Id: <970509091350.ZM3120@candle.brasslantern.com> Date: Fri, 9 May 1997 09:13:50 -0700 In-Reply-To: Comments: In reply to Jacob Palme "MBOX format or Berkeley Mailbox Format - a standard?" (May 8, 11:44am) References: X-Mailer: Z-Mail (4.0b.820 20aug96) To: Jacob Palme , drums@cs.utk.edu Subject: Re: MBOX format or Berkeley Mailbox Format - a standard? MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii On May 8, 11:44am, Jacob Palme wrote: } Subject: MBOX format or Berkeley Mailbox Format - a standard? } } Many e-mail systems use a format for mailboxes called the MBOX format } or Berkeley Mailbox Format. This is the format with } } >From e-mail-address } } as a separator line between messages. } } Does anyone know where an authorative definition of this format } can be found? The closest definition that I know of for the base format is in RFC976, the UUCP Mail Interchange Format Standard. MBOX/Berkeley/Bezerk format is essentially just a series of UUCP-formatted messages concatenated one after the other, separated by a single newline. See 976 section 2.4 for the definition of a "From " line. In a mailbox, the "remote from system" part is usually (but not always) dropped. The problem is that there are at least two widely-used variants of this format; the second most common is the SunOS/SysV addition of a header called Content-Length that gives the octet count from the end of the message header to the blank line before the next message. In this format, nothing within that octet range is a separator line, no matter what it looks like. This is intended to permit forwarded mailboxes and binaries to be stored concatenated in the same file with ordinary text messages. Content-Length is supposed to be added (and recomputed) at delivery time, but in practice it leaks out into other mail systems and causes no end of grief, precisely because people mail and/or copy their mailboxes or NFS mount them from a Sun server. -- Bart Schaefer Brass Lantern Enterprises http://www.well.com/user/barts http://www.nbn.com/people/lantern