Oh awesome!  I was looking at the wrong branch of our Shindig code.
Our production stuff is 1.0-incubating with some backports and I was
looking at that instead of the actual Shindig trunk.  This is indeed
fixed on trunk and I'll be backporting that change to our version.
Phew, that's a relief.  Thanks.

Rich

On Tue, Aug 18, 2009 at 8:36 PM, Kevin Brown<[email protected]> wrote:
> On Tue, Aug 18, 2009 at 5:35 PM, Richard Wallace <
> [email protected]> wrote:
>
>> On Tue, Aug 18, 2009 at 4:27 PM, Richard
>> Wallace<[email protected]> wrote:
>> > Hello,
>> >
>> > I've got a situation where a request to an OAuth protected resource is
>> > generating a 400 response and populating the body of the response with
>> > additional details.  But, when OAuthRequest sees the 400 response
>> > status code it decides to go ahead and ignore the body of the response
>> > from the service provider and replace it with it's own.  So, I have 2
>> > questsions:
>> >
>> > 1) Why would it ever need to do that?  401 and 403 responses that are
>> > OAuth specific are handled in the fetchData() method with the
>> > checkForProtocolProblem(response) call.  If the response is not an
>> > OAuth error, why would Shindig need to replace the actual response
>> > body with one of it's own creation or do any extra work at all?
>> >
>> > 2) Is there a way to disable this behaviour?
>> >
>> > Thanks,
>> > Rich
>> >
>>
>> As a bit of a follow-up.  I tried eliminating the sending of the
>> request trace when it's not an OAuth error and found that the JSON
>> body value still isn't being populated with the content of the
>> original response.  Looking at MakeRequestHandler, it only sets the
>> body value if the response was a 200!  This is not good because it is
>> totally valid, and expected, that web services will send entities when
>> the response status code is something like "201 Created", "400 Bad
>> Request", and "409 Conflict".
>
>
> Are you using the PHP or java code? This issue was fixed in the java code
> almost 3 months ago, so it should have even been covered in the last
> release.
>
>
>>
>>
>> Rich
>>
>

Reply via email to