Pete Resnick <presnick(_at_)qualcomm(_dot_)com> writes:
Also, Charles alone does not rough consensus make. If there are others
who strongly feel that these changes should be made (and strongly would
be good; we're trying to push this along through the standards track,
and changes at this point would be less than desirable), please speak up
and indicate why the current specification does "the wrong thing".
I strongly agree with the message ID changes, and agree in general with
the rest (although not as strongly). The space after colon may warrant a
SHOULD rather than a MUST, though, following your point about being very
serious about the RFC 2119 definitions. (It's truly a MUST in netnews for
interoperability reasons -- headers without a space after a colon won't
work at all with most common implementations since RFC 1036 has always
required the space -- but that doesn't necessarily mean it is in e-mail as
well.)
Again, potentially breaking some netnews implementations is not by
itself a reason to remove this IMO. Are there known non-netnews
implementations that break because of this? (And why are netnews
implementations so non-robust in the first place, unable to parse
comments and quoted-pairs?)
RFC 1036 doesn't mention comments in the References header, and while RFC
1036 references RFC 822 as well, when there's no real use of comments in
References and no netnews implementation has ever generated them, it's not
particularly suprising to me that few netnews implementations have ever
bothered to expect them and strip them. Netnews has been using References
in a structured and protocol-significant fashion for a lot longer than
e-mail.
As for quoted-pairs in message IDs, message ID parsing is at the heart of
the critical path of a netnews server and a typical high-performance
server matches message IDs hundreds or thousands of times a second.
Simplification of this code path both has a leveraged effect on
performance since it's in the heart of the processing loop (not that this
sort of parsing is the performance-critical operation, but everything adds
up) and makes the core of a server more robust. Properly parsing all the
bizarre bits allowed in historic mail message IDs is kind of nuts, and
since RFC 1036, netnews has always required a more restrictive syntax for
the field. Message ID is far, far more important in netnews than it has
been historically in e-mail.
--
Russ Allbery (rra(_at_)stanford(_dot_)edu)
<http://www.eyrie.org/~eagle/>