ietf-smtp
[Top] [All Lists]

Re: TINW was RFC2821b is-01 Issue 1: trailing dot in Domain

2007-03-28 02:09:46

Tony Finch writes:
On Tue, 27 Mar 2007, Arnt Gulbrandsen wrote:
2. I see quite a few addresses of the form 'localpart(_at_)hostname'. Say between 1% and 1ppm of messages.

Yes, I see about 1 per 1000 on my MXs today. About 85% of them are @exe or
@dll so probably an artefact of one particular broken bit of ratware.

Well, really, .com, .exe, .dll, what's the difference? ;)

If 'localpart(_at_)tld' is guaranteed not to be valid, a smarthost's life is simpler: 'localpart(_at_)something' is either a local address (not subject to IETF rules) or invalid.

Private domains are not necessarily single-label, so your suggestion isn't a general-purpose simplification and in fact probably *adds* complexity: If your MTA can handle multi-label private domains correctly then it should be able to handle single-label private domains without any special cases.

Private arrangements _can_ be anything. In actual practice some are much more common than others, and senders generating 'localpart(_at_)hostname' is the one I would prefer not to fight.

Disclaimer: I'm not saying that 'localpart(_at_)hostname' is a good idea or anything like that. I think it's a terrible idea and I wish people would stop doing it. I'm only saying that if 2821bis keeps 'localpart(_at_)tld' invalid, 'localpart(_at_)hostname' will be less bothersome.

Arnt