Received: from localhost by CS.UTK.EDU with SMTP (cf v2.9s-UTK) id TAA04050; Sat, 3 Jun 1995 19:35:56 -0400 X-Resent-To: drums@CS.UTK.EDU ; Sat, 3 Jun 1995 19:35:55 EDT Errors-to: owner-drums@CS.UTK.EDU Received: from munnari.oz.au by CS.UTK.EDU with SMTP (cf v2.9s-UTK) id TAA04043; Sat, 3 Jun 1995 19:35:53 -0400 Received: from mundamutti.cs.mu.OZ.AU by munnari.oz.au with SMTP (5.83--+1.3.1+0.50) id AA07126; Sun, 4 Jun 1995 09:35:20 +1000 (from kre@munnari.OZ.AU) To: John Gardiner Myers Cc: drums@CS.UTK.EDU Subject: Re: address syntax In-Reply-To: Your message of "Fri, 02 Jun 1995 10:02:19 -0400." Date: Sun, 04 Jun 1995 09:35:17 +1000 Message-Id: <1786.802222517@munnari.OZ.AU> From: Robert Elz Date: Fri, 2 Jun 1995 10:02:19 -0400 (EDT) From: John Gardiner Myers Message-ID: > @[ip:v:6]:user@[ip:v:6] It breaks existing correctly written deterministic parsers. The example you gave is what was intended by "heuristic" I suspect, that is, shortcut parsing, not looking at the syntax in detail - if you were, then in stripping the leading parts of the route addr you'd come across the '[' first, and know that a matching ']' was required, and that stripping between those, wherever found, would always be incorrect. In any case, any stripping code that doesn't ensure that the final '@' remains in the address, which seems to be what might go wrong in the example you gave seems terribly fragile to me, similarly any stripping without determining that the address is syntactically correct first. If you're doing the latter now, then the stray ":" chars that might appear would cause your parser to say "fail", and nothing would be stripped. I see three solutions to this problem 1) convince IPng to use some character other than ':'. That has been tried before, however if you wish to try again, the address is ipng@sunroof.eng.sun.com (subscribe by sending "subscribe ipng" in the body of a message to majordomo@sunroof.eng.csun.com) Its not easy as there aren't a lot of available characters, '.' is not available (it is already used in IPv6 addresses), about the only plausible alternative seems to be '-' which turns out to look really ugly when written. 2) Note that this change is likely to happen, and re-write your parsers to accept basically anything (but ']') inside '[' ']' pairs. You have years to get this deployed, you're not seriously likely to see any IPv6 literal addresses inside any joe average mailer anytime in the near future in normal business (someone will stick one in a message to this list just to prove me wrong, but that doesn't count as normal...) 3) Abolish literal addresses altogether. This doesn't seem practical to me, they're too useful when the DNS is broken, or in the case I see a lot, hasn't been set up yet (people send me mail asking for DNS delegations, if their setup is broken, I can't reply without the [] address syntax to get to their mailer). kre