Time  Nick         Message
18:19 saml         http://stackoverflow.com/questions/3241326/set-more-than-one-http-header-with-the-same-name   looks like normal for Set-Cookie
18:18 saml         how come there are two Set-Cookie: header in the response?
18:18 saml         curl -vso /dev/null -H 'Prefer: girl=pretty' www.google.com
18:13 saml         Prefer:
18:13 saml         nice
18:12 Jarda        I only use 204 for DELETE requests
18:11 trygvis      perhaps this is what you should use: http://tools.ietf.org/html/rfc7240#section-4.2
18:07 saml         trying to find a good reason to reject
18:07 saml         and made pull request
18:07 saml         this guy changed to 204
18:07 saml         i did too
18:05 Jarda        I always include the updated response
18:02 saml         or just 204 ?
18:02 saml         do I include updated representation in the response of PUT?
12:35 pdurbin      I wonder if there's a concept of "empty but could be available" in http://transit-format.org
11:58 trygvis      sometimes stuff trickle down from the hypermedia into http, like the Link header
11:58 trygvis      which I mean should be put into the payload instead
11:58 trygvis      sort of, yes. but if you're inventing your own semantics and stuff that into http you more or less end up with your own protocol
11:52 quimrstorres i think the purpose of using http as an applicational protocol is to use as much as its semantics as possible
11:51 quimrstorres i think i don't totally agree with that last statement trygvis
11:34 trygvis      the application itself shouldn't care about the headers, it's not supposed to know its using http or pidgins
11:34 trygvis      probably never
11:33 quimrstorres so, when should i use mechanisms like the warning header to indicate something to the client?
11:28 trygvis      yeah, that's definitely what you should do
11:24 tmoore       yes
11:21 quimrstorres and by that you mean you'd send that info in the payload, right?
11:20 tmoore       I'd just make that part of the response format
11:19 quimrstorres with a warning header or something?
11:18 quimrstorres yep, thanks. would it make sense if the response indicated, in some way, that some of the metrics are not supported?
11:17 trygvis      yep. so always 200 OK
11:17 quimrstorres my question was how should i respond to the request. but as you said, from the resource point of view it's full response
11:17 trygvis      did you want a different response code if it didn't include all of the metrics?
11:16 trygvis      what was your question again?
11:15 quimrstorres yeah, you're right
11:14 trygvis      there is a difference between a partial response and a full response saying it doesn't support every part of the request
11:14 trygvis      ok, but from a resource point of view it includes everything
11:13 quimrstorres only if the server supports all of the metrics requested
11:13 trygvis      or not?
11:13 trygvis      so it does return all of the content
11:12 quimrstorres rted metrics and some indication that some of them were not returned.
11:12 quimrstorres hey there! i am struggling to find the right way to respond to an http request on my web api. i have a resource that, in some cases, it only returns part of the content. it is not a collection resource or anything. basically it is a resource that returns some metrics information, but it is possible that some of the metrics requested don't exist (the server does not support them). i'd like to return the value of the suppo