Re: RTCStatsType value for datachannel stata

On 15 January 2015 at 15:29, Cullen Jennings (fluffy) <> wrote:
>> On Jan 15, 2015, at 4:15 PM, Martin Thomson <> wrote:
>> On 15 January 2015 at 13:06, Jan-Ivar Bruaroey <> wrote:
>>> On 1/15/15 1:54 PM, Cullen Jennings (fluffy) wrote:
>>>> On a related topic to this, I think we should change this from an enum to
>>>> a string because of the issues where we can't extent an enum in webIDL
>>> I believe we're only in trouble if we take RTCStatsType as input somewhere,
>>> I don't think we do.
>> Yes.  The hack for constraints was necessary to allow applications to
>> provide input values that the browser wasn't ready for yet in specific
>> contexts.  There, we didn't want the browser to fail immediately
>> because the values there are sort-of-guidelines.  RTCStatsType is pure
>> output, so no issue there.
> Agree with that but I'm unaware of the WebIDL to extend an existing enum when someone defines a new output type in a new spec a few years from now.  Particularly when multiple specs extend the same enum.

I think that you can just say "these strings are added to enum X".
And not worry about getting fancy.  Do that a few times and Cameron
will eventually fix WebIDL for you.

Received on Thursday, 15 January 2015 23:35:23 UTC