On May 22, 2013, at 12:03 PM, Sergey Beryozkin <sberyozkin_at_talend.com> wrote:
> Hi Marek
>
> I'm seeing few tests like this one:
>
> We have 2 providers, lets say two MBRs, one with "*/*", another one with "text/plain". The filter sets Content-Type to */*. MBR with "*/*" is expected to be selected.
>
> I've got temporarily confused re why MBR with text/plain has to be selected, it definitely loses to MBR with */* on the distance factor.
>
> This is why I asked. But I think I've sorted it out now.
>
> If we have Content-Type: */* and one MBR with */* then if we depend on a distance factor then it has to win over MBRs with more specific media types because the distance here is 0.
Before I read any further, HOW can one have a content type set to */* ???
Marek
>
> However, here is a text from 4.2.3:
>
> "When choosing an entity provider an implementation sorts the available providers according to the media
> types they declare support for. Sorting of media types follows the general rule: x/y < x/* < */*, i.e. a
> provider that explicitly lists a media types is sorted before a provider that lists */*."
>
> Which is what CXF does. So in the above example,
>
> MBR (text/plain) is on top of the list, MBR (text/*) comes next (we also assume both MBR have the same Java type supported).
>
> And we have Content-Type: */*. We start from the top of the list.
>
> And appears to me MBR (text/plain) is to be selected, simply because it is the first one in the list, and text/plain intersects with */*.
>
> Furthermore, if we apply those S functions which are used in the selection of methods and response media types
> S(*/*, */*) -> */*;d=0
> S(*/*, text/plain) -> text/plain;d=2
>
> text/plain;d=2 wins as text/plain is more specific
>
> Do you agree ?
>
> Thanks, Sergey
>
>
>
>
>
> On 21/05/13 14:51, Marek Potociar wrote:
>> I do not think that more generic providers would win over more specific ones. What use case do you have in mind? Let's try to step through the algorithm.
>>
>> Marek
>>
>> On May 20, 2013, at 11:30 PM, Sergey Beryozkin <sberyozkin_at_talend.com> wrote:
>>
>>> Hi
>>>
>>> what is a distance factor between
>>>
>>> */* and */*.
>>>
>>> Is it 0 ? This means it will let methods or providers declaring no Consumes or Produces win over methods/providers declaring more specific
>>> media types values.
>>>
>>> Sergey
>>
>
>