Re: Question: UseSTD3ASCIIRules on RHS of IMA
"Adam M. Costello" <[email protected]>
| Newsgroups | gmane.ietf.imaa |
|---|---|
| Message-ID | <[email protected]> |
Roy Badami <[email protected]> wrote: > Should IMAA make any statement as to whether UseSTD3ASCIIRules MUST > or SHOULD be set when processessing any IDN on the RHS of an IMA, or > should this decision be left up to the implementor? Simon Josefsson <[email protected]> wrote: > All application must already know the answer to this problem, as the > same problem existed before IDNA. If the RFCs are unclear, that > should be fixed independently of IMAA. > > I'm not sure I see any need to enforce anything with regard to this > more than already is, and in particular not how IMAA would be the > right place for it. I agree with Simon. IMAA explicitly points to IDNA for handling the domain part of the mail address, and IDNA already says as much as we dare on the subject. Here's a taste of why we don't want stick our necks out: RFC-821 forbids 3foo.org. RFC-2821 allows 3foo.org, but forbids foo+.org. RFC-822 and RFC-2822 allow foo+.org. It's a mess, and we don't want to make it our job to try to make sense of it. We know from experience with IDNA that any attempt at clarification will be too controversial to reach consensus. Roy Badami <[email protected]> wrote: > Further minor wrinkle: if the application chooses not to enforce > hostname rules (ie not to set UseSTD3ASCIIRules), it still needs to > enforce those retrictions on domain names that occur as a result of > the 822/2822 syntax (ie a label can't contain specials). If the mail address is being put into a message header, then it ought to enforce those restrictions, yes. If the mail address is being put into an SMTP command, then it's the 821/2821 syntax that ought to be enforced. The destination might be neither of those contexts, but something else with its own formal syntax rules. > Does IMAA need to say anything about the application of the 822/2822 > rules (in the same way that IDNA talks about the application of the > STD3 rules)? It might be good to generalize "quote the local part if necessary" to something like "handle any context-dependent syntax rules", and give the same example about quotation sometimes being needed in message headers and SMTP commands, and maybe give another example about character restrictions in those contexts. AMC