It was handling just conference_failed with password required, but not connection failed with password required.

Since all the auth logic is not ported to React on the web,
`_isWaitingForOwner` will always return `false` because the
`waitForOwner()` action is not (yet) used there.

THis fix always tries to hide the dialog no matter what, which is not a
Bad Thing to do anyway.

There is a related bug remaining, however: if one pressed "I am the
host" and then cancel, it doesn't goo back to the previous dialog, but
it completely kils the meeting. This is a compromise we'll have to live
with for a bit longer.

Fixes: https://github.com/jitsi/jitsi-meet/issues/11413
Fixes: https://github.com/jitsi/jitsi-meet/issues/11432

* fix: Drop duplicate call of wait for owner.

* fix: Fixes leaking listeners while waiting for host to join.

While waiting for the host to join on the dialog we attempt to join over and over again till we are admitted to enter the meeting. While doing that authRequired flag is on, and we were adding listeners on and on.

* feat: Introduces conference join in progress action.

This event is coming from lib-jitsi-meet and is fired when we receive the first presence of series when joining. It is always fired before joined event.

* fix: Moves testing middleware to use CONFERENCE_JOIN_IN_PROGRESS.

* fix: Moves follow-me middleware to use CONFERENCE_JOIN_IN_PROGRESS.

* fix: Moves some polls logic to middleware and use CONFERENCE_JOIN_IN_PROGRESS.

* fix: Moves reactions middleware to use CONFERENCE_JOIN_IN_PROGRESS.

* fix: Moves recordings middleware to use CONFERENCE_JOIN_IN_PROGRESS.

* fix: Moves shared-video middleware to use CONFERENCE_JOIN_IN_PROGRESS.

* fix: Moves videosipgw middleware to use CONFERENCE_JOIN_IN_PROGRESS.

* squash: Fix comments.

* fix: Fixes join in progress on web.

* fix: Moves variable extraction inside handlers.

* fix: Moves variable extraction inside handlers again.

* fix: Moves etherpad middleware to use CONFERENCE_JOIN_IN_PROGRESS.

* fix: Fixes #10796 authentication in conference.

* fixup!

* fixup2!

Co-authored-by: Saúl Ibarra Corretgé <saghul@jitsi.org>

* fix(authentication) login dialog now closes when connection is established

* fix(authentication) fixed shibboleth auth

* fix(authentication) renamed authenticateExternal func to authenticate and updated its logic

* fix(authentication)removed logindialog.js and created actions.any

* fix(authentication) removed focus from externalauthwindow

* fix(authentication) removed private sign from some actions and added openLoginDialog to actions.any

* fix(authentication) exported all from actions.any

* fix(authentication) reverted change regarding externalAuth

* fix(authentication) fixed indentation

Up until now we relied on implicit loading of middlewares and reducers, through
having imports in each feature's index.js.

This leads to many complex import cycles which result in (sometimes) hard to fix
bugs in addition to (often) breaking mobile because a web-only feature gets
imported on mobile too, thanks to the implicit loading.

This PR changes that to make the process explicit. Both middlewares and reducers
are imported in a single place, the app entrypoint. They have been divided into
3 categories: any, web and native, which represent each of the platforms
  respectively.

Ideally no feature should have an index.js exporting actions, action types and
components, but that's a larger ordeal, so this is just the first step in
getting there. In order to both set example and avoid large cycles the app
feature has been refactored to not have an idex.js itself.

🔥🔥🔥

Delay the CONFERENCE_FAILED event until the user cancels the login
dialog using the 'recoverable' event flag.

The feature authentication affords recovery from CONFERENCE_FAILED
caused by AUTHENTICATION_REQUIRED. (Like the feature room-lock afforded
recovery from CONFERENCE_FAILED caused by PASSWORD_REQUIRED.)

Gradually, we exploded the error of CONNECTION_FAILED in multiple
redux state properties. The explosion makes maintenance harder because
the properties have to be updated in sync. Collect them in an object
resembling an Error instance.