Re: Event handlers in RTCPeerConnection

Section 14 does include some events, but the only linking between
those and the classes that generate then is in prose.

On 26 September 2012 13:41, Randell Jesup <randell-ietf@jesup.org> wrote:
> On 9/26/2012 3:56 PM, Martin Thomson wrote:
>>
>> None of the event handlers for RTCPeerConnection have specific types
>> associated with the event.
>>
>> This might be OK for the "stuff just happened events":
>>               attribute EventHandler          onnegotationneeded;
>>               attribute EventHandler          onopen;
>>
>> Each of the following produces an event that includes additional
>> attributes:
>>               attribute EventHandler          onicecandidate;
>>   - candidate : RTCIceCandidate
>>               attribute EventHandler          onstatechange;
>>   - state : RTCPeerState
>>               attribute EventHandler          onaddstream;
>>   - stream : MediaStream
>>               attribute EventHandler          onremovestream;
>>   - stream : MediaStream ?
>>               attribute EventHandler          ongatheringchange;
>>   - state : RTCGatheringState
>>               attribute EventHandler          onicechange;
>>   - state : RTCIceState
>>               attribute EventHandler          onidentityresult;
>>   - assertion : RTCIdentityAssertion
>>   - verified : boolean
>
>
>
> attribute EventHandler ondatachannel;
>   - channel: DataChannel
>
> onconnection, onclosedconnection: no params (and poorly named)
>
> In DataChannel objects:
> onopen, onclose: no params
> onerror: not defined yet - should try to mirror WebSockets
> onmessage: Object (DOMString, Blob or ArrayBuffer) - same as WebSockets
>
>
>
>>
>> Rather than using the generic EventHandler for each of these events, a
>> specific, typed callback function for each these would make
>> documentation easier.  That callback would take a subclass of Event
>> that included the extra attributes.
>>
>> Obviously, for some of these it could be just a matter of following
>> through e.target.iceState or something like that, but others contain
>> vital information.  It would be nice if this could be more clearly
>> documented.
>>
>> --Martin
>>
>
>
> --
> Randell Jesup
> randell-ietf@jesup.org
>
>

Received on Wednesday, 26 September 2012 21:18:51 UTC