On 18/10/2010, at 4:31 PM, Roy T. Fielding wrote: > On Oct 17, 2010, at 9:28 PM, Mark Nottingham wrote: > >> Roy, what's your precise definition of successful? 2xx, or 2xx + 3xx? >> >> E.g., 303 See Other seems like it's squarely within the intent of cache invalidation, since Location is included. > > Note that this is just for invalidation of an existing cached > response to a prior GET upon receiving a state-changing request > for the same URI. For example, in the case > > GET /x --> 200 with form to fill out > POST /x --> 303 to /y > > I would not expect the first GET to be invalidated by the 303 > even if the 303 is marked cacheable. Ah. That's an interesting point. The question here, though, is whether /y should also be invalidated; since 2616 goes to pretty extensive lengths to say that the URL indicated by Location is to be invalidated, I don't see why it shouldn't be... -- Mark Nottingham http://www.mnot.net/Received on Tuesday, 19 October 2010 00:29:13 GMT
This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 27 April 2012 06:51:29 GMT