Closed GoogleCodeExporter closed 9 years ago
Original comment by ed.summers
on 24 May 2010 at 9:23
Original comment by ed.summers
on 24 May 2010 at 9:25
Aside from discussions in the atom-syntax mailing list I'm not aware of
anything we
might be able to reference here.
Original comment by hadrien....@gmail.com
on 24 May 2010 at 10:53
A link to said discussion would be helpful I think.
Original comment by ed.summers
on 24 May 2010 at 11:07
Here are some of the discussions:
* http://markmail.org/thread/sybmn6byftbnfa7f
* http://markmail.org/thread/omds3mvvbjnkdwrb
* http://www.ietf.org/rfc/rfc3236.txt
* http://intertwingly.net/wiki/pie/PaceProfileAttribute
* http://buzzword.org.uk/2009/draft-inkster-profile-parameter-00.html
I believe that the syntax in RFC3236 and James' proposal isn't valid against
§4.3 of RFC4288.
"""
Parameter names have the syntax as media type names and values:
parameter-name = reg-name
Note that this syntax is somewhat more restrictive than what is
allowed by the ABNF in [RFC2045] and amended by [RFC2231].
"""
I don't understand why a link to these discussions in the specification itself
would be useful.
Original comment by abdela...@gmail.com
on 24 May 2010 at 2:30
Original comment by abdela...@gmail.com
on 25 May 2010 at 5:41
Original comment by abdela...@gmail.com
on 25 May 2010 at 5:44
I guess we can't really reference anything authoritative for opds' use of the
profile
parameter with the application/atom+xml media type. It sounds like various
people
have talked about it, but it never ended up in RFC 4287.
Working through this I agree that we are covered by RFC 4288 section 4.3: we
aren't
changing the functionality of the application/atom+xml media type in any way,
we're
just adding additional information. But it is kind of unfortunate that adding a
profile parameter means that apps like Firefox, Google Reader, etc are no
longer able
to autodiscover a feed. I also noticed that feedbooks doesn't appear to be
using the
type and profile parameters in the Content-type response header, was that
intentionally done because doing so breaks some things?
"""
curl -I http://www.feedbooks.com/books/recent.atom
HTTP/1.1 200 OK
Server: nginx/0.7.65
Date: Thu, 27 May 2010 06:53:26 GMT
Content-Type: application/atom+xml; charset=utf-8
Connection: keep-alive
Vary: Accept-Encoding
Status: 200 OK
ETag: "d898628705b5777b310923d6fb171b62"
X-Runtime: 62
Cache-Control: private, max-age=0, must-revalidate
Content-Length: 39740
Set-Cookie:
_sn=BAh7BjoPc2Vzc2lvbl9pZCIlMmRiNDllYTkzYzBkYzE4ZWQzM2I5MjJlMjkxMDZhMjQ%3D--
6e6aacf4fe6faba8649b1fdfc816a384c40595e5; domain=.feedbooks.com; path=/;
HttpOnly
"""
Original comment by ed.summers
on 27 May 2010 at 6:56
Using the profile parameter in our atom:link would break compatibility with
existing
clients for sure. With "Content-Type" I'm not so sure but with dozens of
thousands of
people using our feeds on a daily basis, I need to be careful.
Original comment by hadrien....@gmail.com
on 27 May 2010 at 11:28
I believe we've addressed Ed's concerns. Reopen if that isn't true.
Original comment by abdela...@gmail.com
on 4 Aug 2010 at 3:30
Original issue reported on code.google.com by
ed.summers
on 24 May 2010 at 9:23