Hello, how would you model an image archive where the client could retrieve: - images in their original format - thumbnails in different size (small, medium,...
I model /image/2 to represent the image resource, and each resolution that exists as an individual document gets its own url too at /image/2/{format} By...
Conneg could be used for this use case for non-HTML clients. However, conneg is based on each of these representations having a different media type, and you...
... Arguably, a 201 response with a Location header and an empty entity body should be treated like a 303 status. Regards, -- Aristotle Pagaltzis //...
I wonder if extensions to a mime type in this instance would be allowed? Provided the server is implemented correctly, requesting image/jpeg;size=medium should...
It should be up to the user to decide if they want to see what they just created. Nothing prevents you from returning the view of what you just created in the...
... I think it would be better to extend COPY accordingly. Oh wait, Microsoft already did (I think for Exchange)... if you're interested, I can try to locate...
... Cool. There's still something I'm stuck on, which is, aside from my AtomPub bias showing, imo a copy operation is to the target server not the thing being...
... How about: GETPUT /collection/slug Host: www.example.org Allow-Rename: t Content-Type: message/http Content-Length <...> Authorization: <...> GET...
Brian Smith
brian@...
Jul 4, 2008 5:04 pm
10978
... Note that I'm not saying this is the best way to do it. But, it's possible, and if there's enough interest we (the AtomPub/HTTP/WebDAV community) should be...
... That’s not a solution, it’s a bad workaround. If a user fills in a New Thread page on a web forum and then sends me a POST, and I create a new thread,...
... The registration of each MIME type specifies the parameters you may attach to it. You can’t make up new ones willy nilly. Regards, -- Aristotle Pagaltzis...
... The semantics of WebDAV are backways imo - I don't say "copy yourself to there", I say "copy this to that". Aside from that it seems the latter fits closer...
... The reason to have PATCH is that if the only way to patch a resource is to PUT the patch, then clients have no way to *store* patch documents (ie. use PUT...
... Indeed. It should have been defined as COPY from, with a Source: instead of a Destination: server. We still could to that. ... Agreed. I think we only...
... Aren't WebDAV COPY requests are sent to the source so that content negotiation works? Using WebDAV COPY you can hint as to which variant you want copied....
Brian Smith
brian@...
Jul 5, 2008 6:19 pm
10987
... That's an interesting question. I have been involved in WebDAV since 2001, and the WG never has explored content negotiation / variant handling. I think...
... The one thing that makes me vaguely uncomfortable is that the request in the body MUST be a GET request. I think I would rather end up pulling the source...
HTTP injection, cool! Who would have thought that we could invent a new attack vector so easily! Security issues aside, there are quite some obstacles when...
Stefan Eissing
stefan.eissing@...
Jul 7, 2008 6:51 am
10990
... There is no reason any server has to process GETPUTs that would involve an actual over-the-network HTTP GET. If I wrote server code to handle GETPUT I...
... Following redirects? Having the cookies for that namespace on your server? Etc. I understand that you want to resolve the URI only in your very own...
Stefan Eissing
stefan.eissing@...
Jul 7, 2008 8:16 am
10992
My issue would be with special casing any status code, it obfuscate them quite a bit. Just like we have a different code for a success no content and a success...
... That was just a case of word choice on my part. What about “201 with Location and an empty entity” seems ambiguous to you? What other way could this...