CALIFFM@BAYLOR.BITNET (Michael Califf) (07/23/90)
Sorry about posting here, feel free to aim me at a better list... At least one of the INTERBIT (Internet-to-BITNET) gateways is rejecting addresses which contain an underscore. The postmaster at the gateway says that this is because his software correctly implements RFC821 which restricts addresses to letters, numbers, and hyphens. I have checked the RFC and it seems to be a little less restrictive than that. Have I missed something? Thanks for any help/pointers, Mike Califf (POSTMAST[ER]) Communications Software Coord Internet: califfm@baylor.edu Baylor University C.C.I.S. Bitnet: CALIFFM@BAYLOR
CALIFFM@BAYLOR.BITNET (Michael Califf) (07/23/90)
Don't Panic - It turns out that the INTERBIT site which was rejecting the messages is rejecting them based on the (illegal) underscore in the domain part of the address, NOT the (legal) underscore in the local-part. I am taking measures locally to remove/replace the underscores from the offending node names. That'll teach me to O.K. node names before fully parsing the RFC's :-) Sorry about that, Mike Califf (POSTMAST[ER]) Communications Software Coord Internet: califfm@baylor.edu Baylor University C.C.I.S. Bitnet: CALIFFM@BAYLOR
postel@VENERA.ISI.EDU (07/24/90)
Hi. You are right, he is wrong. --jon. ----- Begin Included Message ----- Date: Mon, 23 Jul 90 10:38 CDT From: Michael Califf <CALIFFM%BAYLOR.BITNET@ricevm1.rice.edu> Subject: Is underscore legal in the local-part of an address? To: TCP-IP@NIC.DDN.MIL Sorry about posting here, feel free to aim me at a better list... At least one of the INTERBIT (Internet-to-BITNET) gateways is rejecting addresses which contain an underscore. The postmaster at the gateway says that this is because his software correctly implements RFC821 which restricts addresses to letters, numbers, and hyphens. I have checked the RFC and it seems to be a little less restrictive than that. Have I missed something? Thanks for any help/pointers, Mike Califf (POSTMAST[ER]) Communications Software Coord Internet: califfm@baylor.edu Baylor University C.C.I.S. Bitnet: CALIFFM@BAYLOR ----- End Included Message -----
postel@VENERA.ISI.EDU (07/24/90)
Oh. In that case, you are wrong, and he is right. --jon. ----- Begin Included Message ----- Date: Mon, 23 Jul 90 11:19 CDT From: Michael Califf <CALIFFM%BAYLOR.BITNET@ricevm1.rice.edu> Subject: Re: Is underscore legal in the local-part of an address? To: TCP-IP@NIC.DDN.MIL Don't Panic - It turns out that the INTERBIT site which was rejecting the messages is rejecting them based on the (illegal) underscore in the domain part of the address, NOT the (legal) underscore in the local-part. I am taking measures locally to remove/replace the underscores from the offending node names. That'll teach me to O.K. node names before fully parsing the RFC's :-) Sorry about that, Mike Califf (POSTMAST[ER]) Communications Software Coord Internet: califfm@baylor.edu Baylor University C.C.I.S. Bitnet: CALIFFM@BAYLOR ----- End Included Message -----
Guy Middleton <gamiddleton@watmath.waterloo.edu> (07/24/90)
Mike Califf writes: > At least one of the INTERBIT (Internet-to-BITNET) gateways is > rejecting addresses which contain an underscore. The postmaster > at the gateway says that this is because his software correctly > implements RFC821 which restricts addresses to letters, numbers, > and hyphens. I think he may be a bit confused. The elements of the hostname should be letter-digit-hyphen, but the local-part (anything left of the @) may be anything at all, but, if it contains any special characters from the set <>()[]\.,;:@" or control characters, they must be escaped by a \, or else the whole local-part must be enclosed in quotes.
fitz@wang.com (Tom Fitzgerald) (07/25/90)
[I know the original discussion is moot by this time, but anyway...] gamiddleton@watmath.waterloo.edu (Guy Middleton) writes: > The elements of the hostname should be > letter-digit-hyphen, but the local-part (anything left of the @) may be > anything at all, but, if it contains any special characters from the set > <>()[]\.,;:@" or control characters, they must be escaped by a \, or else the > whole local-part must be enclosed in quotes. Periods are legal according to RFC 822. Is this generally true in implementations, or is this considered one of the glitches of RFC 822 that nobody really pays attention to any more? This actually makes a difference to us. A programmer here wrote an e-mail gateway strictly following RFC 822, in the belief that all other mailer implementations in the world followed it strictly as well. This has caused a few problems already... --- Tom Fitzgerald Wang Labs fitz@wang.com 1-508-967-5278 Lowell MA, USA ...!uunet!wang!fitz
Craig_Everhart@TRANSARC.COM (07/25/90)
RFC822 says that periods are legal but somewhat restricted in local-parts. You can't legally have two periods following each other, and you can't have the local-part begin or end with a period. Thus, joe.blow@toaster.com is legal, while joe..blow@toaster.com jane.smith.jr.@toaster.com are not. (The local-part is made up of a sequence of period-separated ``word''s, and an unquoted ``word'' may not contain a period.) As to the ``phrase'' that precedes a route-addr, the syntax is slightly different (a space-separated sequence of ``word''s), and a period anywhere in it needs to be quoted. This is traditionally done by quoting the entire phrase text. Thus, Joe Blow <joe@toaster.com> "Jane Smith Jr." <jane@toaster.com> are legal, but Jane X. Smith <jane@toaster.com> is not. Then again, your programmer who believed that all other mailer implementations strictly follow RFC 822 was a bit naive. ``Liberal in what you accept, conservative in what you generate.'' Craig