Show the device that's actually in use

[dbkr]

Feeds the current track (if any), and therefore device it's using into the device selection dropdown and implement a some somewhat complex logic to observe when the 'default' device has changed to a different actual devbice between when we opened it and now, and display the device we're actually using as selected instead.

The comments should explain more, including the caveat that we can't do the same detective work if the device is one of several devices on the same bit of hardware. This is because is uses the same groupId comparison technique used by Livekit in normalizeDeviceId to do a smiliar thing.

Fixes #1646

[codecov-commenter]

Codecov Report

All modified lines are covered by tests ✅

Comparison is base (3e0bc8a) 29.94% compared to head (aff01d0) 29.94%.
Report is 10 commits behind head on livekit.

Additional details and impacted files

@@           Coverage Diff            @@
##           livekit    #1670   +/-   ##
========================================
  Coverage    29.94%   29.94%           
========================================
  Files           47       47           
  Lines         1877     1877           
  Branches       328      328           
========================================
  Hits           562      562           
  Misses        1274     1274           
  Partials        41       41           

Flag	Coverage Δ
unittests	29.94% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[toger5]

I am wondering if the other PR does obsolete this?
In the other PR the track is always updated in case the default device changes. So the track in use should always be matching the device and hence have the same/correct groupId.

I am wondering if this PR in combination with the other one would lead to the following behaviour:

• default OS device is changed (actual-device-id:a -> b)
• device id in the selected variable is still default
• the group id of chromes default device changed to the groupId of b (also the label got updated to conatining the description of b)
• track is restarted (becasue we do the group id check an see that default is a new device) (other PR) in this case its b now.
• We reach the logic introduced here,
• we pass the conditions (device.selectedId = "default")
• we find (d) => d.groupId === usedGroupId && d.deviceId !== "default" (but there will be two devices with the correct groupId because the default device and the same device with deviceId = "someActualUUIDasTheID1234")
• the found device will be the one where groupId == usedGroupId and deviceId will be the NOT "default" so the selector does not select the default device anymore
• chainging the device in the os again (a -> b) works as normal but I am not sure EC still updates the device since now it does not check anymore if the group Id of the defualt device has changed.

So I am wondering if this would break EC following the system setting on chrome after the first system setting change.

[toger5]

Similar to the last pr. Could the label also work?

[dbkr]

This might involve comparing here which is tricky because chrome prepends 'Default: ' so we'd have to work around that and however it i18ns on the user's system.

[toger5]

Thinking about this topic does livekit already check if the order of the devices changes on firefox?

A Promise that is fulfilled with an array of MediaDeviceInfo objects. Each object in the array describes one of the available media input and output devices. The order is significant — the default capture devices will be listed first.

I am wondering if the ux would be much less broken with the following approach:

• we have a checkbox: os default settings
• if it is checked we only show what device we get as the first device of enumerateDevices and we react to order changes and update the media device accordnigly. (on chrome we always use id="default" and update using the groupId or label trick)
• if it is unchecked a dropdown selector will appear to choose the device. We then NEVER use "default" as the device id and never react on os cahnges and instead only react to changes in the dropdown.

[dbkr]

I am wondering if the other PR does obsolete this?
Yes - hopefully :) Although the plan basically was to make sure the dropdown always shows the truth first, then implement the device switching, so even if the device switching doesn't work for whatever reason, it still shows the truth.

I am wondering if this PR in combination with the other one would lead to the following behaviour [...]

It will switch back to device a when you change the default device in the OS a -> b -> a (I tested). The group ID of the default device should change just the same in either case.

I am wondering if the ux would be much less broken with the following approach: [...]

Possibly! Do you think we should pause this and wait for product / design input?

[toger5]

It will switch back to device a when you change the default device in the OS a -> b -> a (I tested). The group ID of the default device should change just the same in either case.

Then I have a thinko, just to confirm, you tested with both PR's being merged at the same.

[toger5]

Yes - hopefully :) Although the plan basically was to make sure the dropdown always shows the truth first, then implement the device switching, so even if the device switching doesn't work for whatever reason, it still shows the truth.

Considering this,

Possibly! Do you think we should pause this and wait for product / design input?

I think pushing this to the next release and asking design is a good idea.

[CLAassistant]

All committers have signed the CLA.

[hughns]

Making draft as this is stale.