Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Path: utzoo!utgpu!water!watmath!clyde!rutgers!ames!ptsfa!vixie!paul From: paul@vixie.UUCP Newsgroups: comp.mail.uucp Subject: Re: whether to prefix myhost! onto the From: or not.. Message-ID: <603@vixie.UUCP> Date: Thu, 23-Apr-87 05:44:33 EST Article-I.D.: vixie.603 Posted: Thu Apr 23 05:44:33 1987 Date-Received: Sat, 25-Apr-87 04:07:59 EST References: <16238@amdcad.AMD.COM> <600@vixie.UUCP> <870@xanth.UUCP> Reply-To: paul@vixie.UUCP (Paul Vixie Esq) Organization: Vixie Enterprises, San Francisco Lines: 49 In article <870@xanth.UUCP> kyle@xanth.UUCP (kyle jones) writes: >> I am vixie!paul. > >According the RFC822 this is would not be from user "paul" on host >"vixie", but rather from user "vixie!paul" on the local host! Ooops. Silly of me to list myself that way in an article calling for greater adherence to published standards! However, I seem to recall the context being a lengthy explaination of what could happen to a route if various intermediate hosts did or didn't prepend their host name -- so listing it in host!user instead of user@host.domain seemed appropriate. >paul@vixie.uucp seems the best alternative. As others will no doubt point out, this is bad and evil. paul@vixie is okay -- barely, but a trivial case to rewrite. Many older sendmail.cf's use .UUCP as their mock domain for UUCP traffic -- so an address input as vixie.UUCP can cause confusion -- it becomes vixie.UUCP.UUCP internally, and after that, chaos. Newer template .cf's use .UUX to address this, but the older ones remain. (SMAIL fixes it!) > Internet only sites won't grok >"vixie!paul" or "paul@vixie.uucp", but RFC822 compliant UUCP sites [...] >are much happier with the user@host.uucp form. Why? Because mailers >at such sites tend to do one of two things when given a From: line containing >just host!user > > :-) 1) they rewrite it to user@host.uucp >or :-( 2) they tack on their own host/domain name (e.g. host!user@mtv.com) Or, :-) 3) they rewrite to a user-host-domain triple regardless of input format and then try to do something intelligent with it (SMAIL wins again) Thanks for pointing out #2, though: '!' is not an officially recognized character! As far as the RFC's are concerned, '!' is part of whichever side of the '@' is appears on, and is not guaranteed to be parsed at all. Since the DDN is more or less RFC-compliant, this makes the case for treating a!b@c as a!c!b pretty tenuous (try nonexistent) -- it's c!a!b. >kyle jones, old dominion university computer science >ARPA: kyle@xanth.cs.odu.edu CSNET: kyle@odu.csnet >UUCP: kyle@xanth.uucp -or- {sun,harvard,mit-eddie}!xanth!kyle -- Paul A. Vixie {ptsfa, crash, winfree}!vixie!paul 329 Noe Street dual!ptsfa!vixie!paul@ucbvax.Berkeley.EDU San Francisco CA 94116 paul@vixie.UUCP (415) 864-7013