On Thursday, January 7, 2016 at 9:22:15 AM UTC-7, Jan-Ivar Bruaroey wrote: > Look closely at the Firefox permission dialog here, and you'll see it > doesn't actually let users change the camera or microphone (because the > site uses `exact` constraints).
Interesting--is that "exact" constraint documented somewhere? Does Chrome support it as well? This would roughly solve my problem, because it'd be possible to know as a developer if they opted not to share something because the resulting MediaStreamTrack would contain no corresponding track. > > Even if they didn't select "always allow," does it really need to > > prompt with every change? > > The primary purpose of the prompt is to request user permission. Firefox > is being protective of users here by not forcing "always allow" on > users. The selector is a secondary feature. I am not advocating "always allow". My point: is it really necessary for someone to "allow" repeatedly on the same page load? Example: I browse to www.myVideoConferencingSite.com, which requests access to mic and camera. The user grants it. Later, the user *changes* mic and camera. This causes another request to access mic and camera, which is tedious since the user already granted permission to mic/camera. I see the exact subject being discussed in https://bugzilla.mozilla.org/show_bug.cgi?id=1212996 , but the reality is Chrome already functions the way people are arguing amounts to "collateral grant." I disagree with this because it's a questionable assertion that users see a "privacy" difference from one mic/camera to another. Also, there are very real implications for sites using WebRTC, because users effectively aren't streaming media while the change is occurring. On Chrome it's a quick blip, but on Firefox it can be seconds or more. Hopefully this situation will be improved with the ORTC spec and the ability to change camera/mic without affecting streaming, but for Firefox, the user is literally gone from a live stream for as long as it takes them to navigate that extra UI. I'll post in that defect. Thanks for the response. _______________________________________________ dev-media mailing list [email protected] https://lists.mozilla.org/listinfo/dev-media
