W3C home > Mailing lists > Public > public-webrtc@w3.org > September 2015

Re: RTCSessionDescription and RTCIceCandidate

From: Harald Alvestrand <harald@alvestrand.no>
Date: Mon, 14 Sep 2015 14:03:59 -0700
To: public-webrtc@w3.org
Message-ID: <55F7363F.7090701@alvestrand.no>
On 09/14/2015 01:53 PM, Martin Thomson wrote:
> On 14 September 2015 at 13:47, Harald Alvestrand <harald@alvestrand.no> wrote:
>> For RTCIceCandidate, I think one of Peter's PRs suggestd a bunch of new
>> fields for it (and we agreed, I think, that if we have both the string
>> form and the attribute form, all of the fields would have to be readonly).
> Yes, but that is orthogonal to this.  If we accept either .candidate
> or the { set of things Peter described } this still works.

The only reason it matters is that there's no such thing as a readonly
dictionary (I think - dictionaries are supposed to have copy semantics
rather than object semantics, and when this matters always makes my head
hurts.)
>
>> Note (in case we adopt this): I think we can make the argument to
>> pc.addCandidate be "RTCIcecandidate or DOMString" for as long as we want
>> to preserve compatibility for. Doesn't have to be in the spec either.
> We can't do that because each candidate needs to be matched to a media
> section with mLineIndex or mid.
>
Argh, yes.

I don't see any huge advantage of using "dictionary" over "interface",
but that may be just me.

-- 
Surveillance is pervasive. Go Dark.
Received on Monday, 14 September 2015 21:04:30 UTC

This archive was generated by hypermail 2.3.1 : Monday, 23 October 2017 15:19:46 UTC