[Webkit-unassigned] [Bug 49736] New: For speech input events, send an object containing all the recognition results and metadata.

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=49736

           Summary: For speech input events, send an object containing all
                    the recognition results and metadata.
           Product: WebKit
           Version: 528+ (Nightly build)
          Platform: All
        OS/Version: All
            Status: UNCONFIRMED
          Severity: Normal
          Priority: P2
         Component: Forms
        AssignedTo: webkit-unassigned at lists.webkit.org
        ReportedBy: satish at chromium.org


Currently the speech input event 'onwebkitspeechchange' receives no parameters. WebCore already has the full list of recognition results and populates only the top result into the input field automatically. In this patch we pass a SpeechInputEvent object to the event handler, and the JS code can access the full list of results via "event.results.length", "results[i].utterance" and "results[i].confidence".

To test multiple recognition results, I have replaced the 'setMockSpeechInputResult' methods with 'addMockSpeechInputResult' which can be called multiple times to add multiple results. This method also takes in the result's confidence value. The mock object in webCore has been updated, along with existing layout tests and a new test added.

I have placed some newly added speech input IDLs and headers under WebCore/page since that seemed most appropriate and previously added files are located there. I'm happy to move them if there is a better suggestion.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.