* Prints errors in case of wrong initialization.

Not printing can masks some errors in the code.

* Allow only one Follow Me moderator in a meeting.

* Sends Follow Me state with all presences of the moderator.

This fixes an issue where the moderator sends the Follow Me state and then for example mute or unmute video (this will produce a presence without Follow Me state) and the new comers will not reflect current Follow Me state till a change of it comes.

* Changes fixing comments.

* Changes fixing comments.

Dear reader, I'm not proud at all of what you are about to read, but sometimes
life just gives you lemons, so enjoy some lemonade!

Joining a conference implies first creating the XMPP connection and then joining
the MUC. It's very possible the XMPP connection was made but there was no chance
for the conference to be created.

This patch fixes this case by artificially genrating a conference terminated
event in such case. In order to have all the necessary knowledge for this event
to be sent the connection now keeps track of the conference that runs it.

In addition, there is an even more obscure corner case: it's not impossible to
try to disconnect when there is not even a connection. This was fixed by
creating a fake disconnect event. Alas the location URL is lost at this point,
but it's better than nothing I guess.

This legacy code was added about a year ago to ease the migration between
releases:
631f51d627

I consider this not to be needed anymore.

Use subscribers to detect state change and emit those
out to other participants. Use middleware to register
the command listener.

This reverts commit 9625be1db3.

The Audio.js setRef callback does not behave like react ref callback
in that the former will not have fired before componentDidMount
but the later will have. So for audio output preview, trying to set
sink id on mount will no-op because it does not have a ref yet to
Audio.js, possibly leading to audio output previews playing on
the default speaker device. This generally has not been a user
visible problem due to coincidence; other re-renders necessary
by the parent of audio output preview will have triggered
componentDidUpdates on the audio out preview, which would then
set the sink id on the Audio.js ref it should have received
by then.

Searching for a device (id) by label alone can result in
false results when devices share labels, such as a mic
and speaker having the same label. To prevent such,
specify the device kind to be found instead of iterating
over all device kinds.

When the iFrame api is used to set a preferred audio output using
options passed into the JitsiMeetExternalAPI constructor, no logic
fires to actually change the audio output destination.

Devices of different kinds can have the same id, such as speaker
and mic both being default. Using id only can then lead to
incorrectly setting device descriptions in the current devices
object.