Re: [ietf-smtp] Is this a valid Message-ID header ?

[Keith Moore <[e-mail]>]

On 3/4/20 11:34 PM, John Levine wrote:

> I read IETF mailing lists on the IETF's IMAP server* and of late I've been
> seeing some awfully funky stuff.
>
> This is an actual Message-ID header in one of the messages in dnsop.
> Is that valid?  Even though the MIME decodes to a an ASCII message ID
> in the <string@domain> format, I think the answer is no.  That's not
> what RFC 5322 sec 3.6 allows.
>
> Message-ID: =?utf-8?q?=3CBN7PR11MB25474DF04998FF3AA9E0B80BC9E40=40BN7PR11MB2?= =?utf-8?q?547=2Enamprd11=2Eprod=2Eoutlook=2Ecom=3E?=

No, it's not valid.   The above field body scans as "atom atom", and a 
msg-id cannot begin with two atoms without some other symbol (e.g. ".") 
in between.

RFC 2047 encoded-words were designed to be compatible with the 
(then-current) RFC 822 grammar, so that they wouldn't break existing 
parsers.   That means that any decoding of encoded-words for 
presentation occurs after (header field-specific) parsing. Also, section 
5 of RFC 2047 limits use of encoded-words to very specific places in the 
message header.   Any text that parses as atom that appears elsewhere in 
a message header, should be treated as an ordinary atom, even if it 
looks like an encoded-word.

Keith