- From: Cullen Jennings <fluffy@cisco.com>
- Date: Wed, 11 Jan 2006 02:10:23 -0800
- To: Julian Reschke <julian.reschke@gmx.de>
- CC: WebDav <w3c-dist-auth@w3.org>
Of those three options, 2 and 3 give clear advice. You could change 1 to also give clear advice by saying the "server should return a 400 and potentially and HTML response". I was not really saying which one was good other than whatever thing you decide, give clear advice on what a server should do and what a client can rely on. On 1/11/06 1:45 AM, "Julian Reschke" <julian.reschke@gmx.de> wrote: > Cullen Jennings wrote: >> On 1/11/06 1:01 AM, "Julian Reschke" <julian.reschke@gmx.de> wrote: >> >>> Cullen Jennings wrote: >>>> I think you should provide implementers advice on what to do in this case. >>> It's already saying: >>> >>>>> User agents may be able to take advantage of this >>>>> to display a meaningful error message to the user. >>> I'm not sure what else we could say... >>> >>> >>> Best regards, Julian >> >> I was thinking of advice that answered the following ... If I am a server >> implementer, and I get a URI other than the request URI that is too big, >> what do I do? > > The server should do whatever the spec says. If we... > >> 1) Remove that paragraph, thus be silent on what servers return in this >> case, which in general would mean 400 Bad Request. This is consistent >> with RFC2616, but doesn't give clients a mean of detecting this error >> condition. > > ...the server should return a 400, and potentially an HTML response. A > non-browser client would likely not know what to do with it. > > If we... > >> 2) Keep Section 11.2 (being inconsistent with RFC2616) > > the server would reply with 414 (potentially confusing clients because > that's not what 414 is for). > > If we... > >> 3) Define a precondition code that could be used with status 400, and >> which would allow not only to describe the general problem, but also >> could include the URI that was actually causing the problem (keep in >> mind that URIs can appear in several places in the request headers, such >> as "Destination" and Tagged-Lists in "If"). The exact proposal for this >> is below. > > ...the server would return something like > > HTTP/1.1 400 BAD REQUEST > Content-Type: application/xml > > <error xmlns="DAV:"> > <uri-size-not-exceeded> > <href>...the offending URI...</href> > </uri-size-not-exceedeed> > </error> > > and a client could then pull out both the reason (condition code) and > the details (URI) from the body. > > And again...: > >> Feedback appreciated, mine would be: >> >> 1) +.5 >> 2) -1 (because of inconsistency with RFC2616) >> 3) +.5 (it doesn't hurt) > > Best regards, Julian
Received on Wednesday, 11 January 2006 10:10:19 UTC