In <3E089CCB(_dot_)7040407(_at_)alex(_dot_)blilly(_dot_)com> Bruce Lilly
<blilly(_at_)erols(_dot_)com> writes:
Pete Resnick wrote:
No, sorry, it doesn't work that way. Just because 2368 (via 1738)
specifies that some structured part of an encoded URL follows 822/2822
syntax for "mailbox" does not mean that a mail header field that
contains such a URL therefore contains a mailbox. The parser for
822/2822/2047 does not know about URL parsing and cannot get past the
"%" encoded bits. There is no intervening 1738 parser.
Exactly so. If you have an encoding inside an encoding, then the main
news/mail protocols need only concern themselves with the outer one. Only
the parts of the implementation that specifically understand that header
(we are talking of headers that contain URIs here) need to know what to do
with the inner one.
That depends on the context. In the case of the List-Owner example and
a number of fairly widely-used MUAs, the URL is properly decoded when
using the List-Owner field content to generate a message to the list
owner, and some even decode the 2047 encoding for display. It is
true that for transport, there is no decoding, but that's axiomatic as
2047 is clearly intended for display purposes and not for processing
during transport.
Surely you're not suggesting that correct use of the List-Owner etc. header
fields to generate messages to the list owner etc. would be to use the raw
URL %-encoded text in the headers?
That is exactly what I would suggest for general use. URIs are already
7-bit clean, so there is no need for gateways to touch them. If some
user-agent that understands the List-Owner-header cares to look inside and
do clever things, then it is welcome, of course.
Now all this could change if IRIs (draft-duerst-iri-02.txt) should get to
the standard stage. That draft contains provisions for encoding them back
into URIs, which is probably what should be done by a gateway that spots
one (and maybe I have to include that case in my text at some stage).
It seems that you're implying that a
To header generated from a mailto URL doesn't contain a mailbox even if it
matches the mailbox syntax and although a To header generated from any other
mechanism may contain a mailbox -- that doesn't sound right either.
Yes, that would be my understanding too, at least so far as what RFC 2047
is supposed to do about deeply nested comments/phrases (unless Keith can
convince us otherwise).
URLs can contain parentheses (which is why there's a problem with the
RFC 2557 ABNF and CFWS).
Which is a good reason for putting URLs inside <...>, and ensuring that
nobody goes looking for comments inside there.
--
Charles H. Lindsey ---------At Home, doing my own thing------------------------
Tel: +44 161 436 6131 Fax: +44 161 436 6133 Web: http://www.cs.man.ac.uk/~chl
Email: chl(_at_)clw(_dot_)cs(_dot_)man(_dot_)ac(_dot_)uk Snail: 5
Clerewood Ave, CHEADLE, SK8 3JU, U.K.
PGP: 2C15F1A9 Fingerprint: 73 6D C2 51 93 A0 01 E7 65 E8 64 7E 14 A4 AB A5