In the past we used a PanResponder to detect user gestures in the sheet
to show a reduced version or a full-height version of it, and also to
close it.

There is an obvious conflic between the gestures and scrolling, which
didn't work all that great, but we could live with it.

After reactions were introduced we no longer rendered the 2 different
heights, so that functionaligy stopped being used but the PanResponder
still remained there.

This commit removes it completely and sets a max height of 75% on any
BottomSheet, so any tap outside will close it.

* fix(face-landmarks): filter face detections based on detection score

* fix: add blank line and semi column

* Send speaker stats for brk room

* Fix comments component loaded

https://github.com/jitsi/lib-jitsi-meet/compare/v1448.0.0+3df2abec...v1450.0.0+462996fc

added car mode and polished many words

Fixes an issue - not able to stop the screen sharing in presenter
mode from the screen sharing button.

* fix(face-landmarks): set max detected faces up to 4

* add(face-landmarks): config for max faces detected

* fix(config.js): default value for capture interval face-landmarks

* add missing coma

A minor wording change to prevent confusion: On logout one is asked
whether to stop the conference but the conference is only left by the
participant.

When there is a password and lobby enabled, participants cannot enter password as the display name is stealing the focus.
When there is just password set the same field steals the focus from the password prompt.

https://github.com/jitsi/lib-jitsi-meet/compare/v1439.0.0+42f5b4bf...v1448.0.0+3df2abec

As luajwtjitsi is in jitsi-meet-prosody move its dependencies.

They "force" a dark mode but we already have a dark palette, so avoid
the system overriding the base colors.

Fixes: https://github.com/jitsi/jitsi-meet/issues/9981
Fixes: https://github.com/jitsi/jitsi-meet/issues/8781

Ref: https://github.com/jitsi/jitsi-meet/issues/11611

* fix(multi-stream) RN Add listeners for track streaming status updates on large-video.
Fixes an issue where video on large-video is not being rendered when there is no filmstrip, i.e., there is only 1 remote participant in the call with source-name signaling enabled. Also do not show the screensharing indicator on the camera thumbnail when a virtual SS tile is created for the local screenshare.

* squash: add a comment

Fix auto focus on menu open
Fix unable to use space in input

* fix(face-landmarks): human helper tensor disposal and async functions

* fix(face-landmarks): rename functions in interface

Fixes: https://github.com/jitsi/jitsi-meet/issues/11563

Only dispatch raise hand if hand was not already raised

* fix(filmstrip) Add handlers for track streaming status on RN.
This is needed for switching between video and avatar when the track's streaming status changes in source-name signaling mode.

* squash: Add comment.

https://github.com/jitsi/lib-jitsi-meet/compare/v1438.0.0+b9aa0e59...v1439.0.0+42f5b4bf

The device selection initialization on the prejoin use case was handled
like the welcome page. This was introducing issues with selecting the
stored devices and not the ones used, enabling the device selection when
it will fail and others.

https://github.com/jitsi/lib-jitsi-meet/compare/v1437.0.0+966fcd93...v1438.0.0+b9aa0e59

On iOS Safari when the tracks for the previews are created the old ones
are auto destroyed which leads many issues like stop sending media while
the devie selectioin is displayed, error messages, etc.