Why is the native application launching a browser with a protected resource request? That seems odd.
Note that we currently have no plans of supporting signatures in any of the flows. We are discussing signatures only for using tokens with secrets when accessing protected resources. EHL On 4/8/10 7:08 AM, "George Fletcher" <gffle...@aol.com> wrote: Another use case is where a rich client wants to bootstrap a web session with the same identity (rich client to web SSO). Assuming that the web session will be established via OAuth with signatures, there is no way to fire up the browser with a "signed URL" if the OAuth parameters and signature need to be in a header. As Jon mentions, the concept of allowing a service to create a signed URL and then pass it back to JS running in the browser, or invoking a browser directly is something that we leverage a lot across our rich clients and web applications. I realize that these sorts of use cases are trivial if establishment of the SSO session switches from a signed mechanism to the OAuth WRAP bearer token model. The one nice feature of the signed URL is that it is one time use where the bearer token can be replayed multiple times. Thanks, George Real world use case. Login into the latest AIM client. Click the mail icon/link. On 3/31/10 7:25 AM, Moore, Jonathan wrote: What about a use case where the signature will be generated by one component but the request will be redeemed by another? For example, suppose there is a cross-domain JSONP call that requires authorization; in this case, I might have my client side code hit *my* origin server, obtain a signed URL, and then redeem it by hitting the JSONP resource. This has scaling advantages over having my origin proxy an OAuth request, and doesn't require me to have keys located on the client; I can keep them safely in my data centers. In this case, sending a "ready to redeem" signed request using the query parameter mechanism simplifies the client-side code. Furthermore, in this use case (cross-domain script inclusion), the client doesn't have the means to set the Authorization header (it can only include a <script> element with a URL). A similar use case would be if you wanted to provide signed redirects (similarly useful for cross-domain functionality); browsers aren't going to modify the redirect URL they get back, or add an Authorization header to it. Jon ........ Jon Moore Comcast Interactive Media -----Original Message----- From: oauth-boun...@ietf.org on behalf of Eran Hammer-Lahav Sent: Wed 3/31/2010 12:20 AM To: OAuth WG Subject: [OAUTH-WG] Limiting signed requests to use the Authorizationrequest header Since we have consensus that using signed requests is a more advance use case and will be used by more experienced developer, I would like to suggest we limit sending signed request parameters to the Authorization header (no URI query parameters or form-encoded body). This will not change the ability to send the oauth_token parameter in the query or body when using bearer tokens (as well as in the header). It will only apply to sending signed requests. The makes client request parameter much simpler as the only parameter "invading" the URI or body space of the request is oauth_token. Anything else is limited to the header. Thoughts? If you are not a fan, please reply with a use case. EHL _______________________________________________ OAuth mailing list OAuth@ietf.org https://www.ietf.org/mailman/listinfo/oauth _______________________________________________ OAuth mailing list OAuth@ietf.org https://www.ietf.org/mailman/listinfo/oauth
_______________________________________________ OAuth mailing list OAuth@ietf.org https://www.ietf.org/mailman/listinfo/oauth
