The initial screensharing UI landed without any test. We are now polishing this UI with bug 1284877 and bug 1284878, and more importantly we are planning to remove the screensharing whitelist to make the feature available to all websites (bug 1127522). As part of this project I think we need to ensure the UI is well covered by tests.

I think I'm done, but I'll wait to see the color of try results before requesting review.

The colors I see on the try push from comment 19 are promising at this point, so I think it's time to request review. Panos, I picked you but feel free to redirect to whoever has cycles. Before landing I intend to file bugs for the AppConstants.platform checks, to mention the bug numbers in the todo() comments.

Comment on attachment 8816305 [details] [diff] [review] Patch v4 Review of attachment 8816305 [details] [diff] [review]: ----------------------------------------------------------------- ::: browser/base/content/test/webrtc/browser_devices_get_user_media_screen.js @@ +387,5 @@ > + function* check(expected = {}) { > + let shared = Object.keys(expected).join(" and "); > + if (shared) { > + Assert.deepEqual((yield getMediaCaptureState()), expected, > + "expected nothing to be shared"); "nothing" wouldn't be accurate when |expected| has a non-default value, right? ::: browser/base/content/test/webrtc/head.js @@ +247,5 @@ > + return new Promise((resolve, reject) => { > + let mm = _mm(); > + mm.addMessageListener("Test:MessageReceived", function listener({data}) { > + resolve(data); > + mm.removeMessageListener("Test:MessageReceived", listener); Nit: removing the listener before further action is the more common pattern to avoid inadvertent reentrancy. Is there a particular reason you want it like this here?

https://hg.mozilla.org/integration/mozilla-inbound/rev/6e2e9bdc497579d2145440575a9954f3244feccf Bug 1313324 - Cover the screensharing UI with browser chrome test, r=past.