If user activates an OpenID authentication method by clicking on it, leaves the URL field empty and clicks the Login button, warning message will appear.
In other case, temporary progress frame will be activated and discovery process initiated. If no usable OpenID services will be found with the URL provided a message will be displayed and authentication frame will be activated again.
In case of successful discovery process, an OpenID frame will be activated asking a user to click Next in order to authenticate via OpenID. If user clicks on that button, an authentication page on his server will open in a new popup window and another OpenID frame will be activated. In a popup window user will have to authenticate himself and will be able to either grand or deny identity validation. In a frame, user can do the following:
- click on the Done button to complete his authentication process;
- click on the Cancel button to close WCB and return to the initial state;
- click on the Try another URL link to cancel current authentication process and start another one.
If user clicks on the Done button the system will try to finish the authentication process and, in case of success user will be logged in and current page will be updated.
In other case, a message will be displayed and authentication frame will be activated again.
Anton: OpenID authentication (both discovery and completion parts) works as described. Both parts are unit testable: to achieve this we will have to emulate OpenID responses (#540).
