On Tue, Dec 16, 2014 at 3:18 PM, Martin Thomson <martin.thomson@gmail.com> wrote: > > You probably need to get someone with more DOM clue than I to look at > this, but the use of a Promise-bearing setter for sinkId might be > incompatible with established notions of using HTML attributes for > this sort of thing. As currently specified, I assume that you could > not support <video sinkId="foo"/>. > We considered that, but ultimately decided on the Promise approach because it provides a reasonable failure semantic. And, of course, because I love Promises. > This also increases the fingerprinting surface (by a miniscule amount) > if you can tell which of the enumerated audio output devices is the > default just by creating an audio/video element and checking the value > of sinkId. I guess you could address that my making this only read > what was written. > That seems like a good idea. > > I like the "do nothing" option for web audio. > Chris Wilson mentioned he prefers the sink ID to be passed as a param to the AudioContext ctor. I updated the document with this as a new option. > > On 16 December 2014 at 14:49, Justin Uberti <juberti@google.com> wrote: > > View it at: https://w3c.github.io/mediacapture-output/ > > File issues, submit changes at: > https://github.com/w3c/mediacapture-output/ >
Received on Wednesday, 17 December 2014 01:03:03 UTC