Up until now, mobile was oblivious to participants' presence state. Presence
state handling is required (probably, amongst other things) for "call flows".
So, let's add it! This is done by gathering the presence state when a
participant first joins, and handling subsequent changes.
Fix the "mute ping pong" for once and for all. This patch takes a new approach
to the problem: it keeps track of the user generated CallKit transaction ations
and avoids calling the delegate method in those cases.
This results in a much cleaner and easier to understand handling of the flow: if
the delegate method is called it means the user tapped on the mute button. When
we sync the muted state in JS with CallKit the delegate method won't be called
at all, thus avoiding the ping-pong altogether.
In addition, make sure all CallKit methods run in the UI thread. CallKit will
call our delegate methods in the UI thread too, thsu there is no need to
synchronize access to the listener / pending action sets.
* feat(settings): setting dialog
- Move device selection, profile edit, language select, moderator
options, and server auth into one modal with tabs.
- Remove side panel profile and settings and logic used to update
them.
- Pipe server auth status into redux to display in the settings
dialog.
- Change filmstrip only device selection popup to use the new
stateless settings dialog component.
* squash: do not show profile tab if not guest
* squash: profile button not clickable if no profile to show
* squash: nits
* ref: Settings dialog.
Emits CONFERENCE_FAILED in response to CONNECTION_FAILED event
which then triggers JitsiConference.leave() through the middleware
processing. Also base/conference state will be adjusted. It is to have
a consistent redux state in which both connection and conference are
failed. It could happen that in a buggy environment the XMPP connection
is dropped, but the media is still flowing which would result in weird
user experience.
The change to mobile/external-api is required to not emit
CONFERENCE_FAILED for CONNECTION_FAILED if the conference has been
started, because base/conference state will still hold conference
instances which are to be ended by other means and result in the
appropriate event (which will adjust the base/conference state).
Currently the listeners for disconnected and failed connection events
are unsubscribed as soon as the connection is established, so
the CONNECTION_DISCONNECTED is never triggered which would clear the
'connection' field. This commit will clear the 'connection' state on
CONNECTION_WILL_CONNECT. It's needed anyway given that there's no
guarantee on when and if the async disconnect operation will finish.
One issue caused by the 'connection' not cleared was that
CONNECTION_FAILED was not reduced correctly and the reload screen was
not displayed for the following scenario:
1. Join and leave any working conference.
2. Turn off network connectivity on the device.
3. Wait for CONNECTION_FAILED. The reload screen will not be displayed,
because CONNECTION_FAILED is not reduced correctly, because the old
'connection' value is still there.
Your truly introduced this regression in
8c7a3f16b1, alas.
The audio only mode is used to set the CallKit call type. This affects the
behavior on the recent calls entries (calls are marked either as audio or video
calls).
Sync both at the start and for transitions. The previous code was working by
chance (in a way): when the CallKit UI is presented the local video is muted,
which triggers a SET_VIDEO_MUTED action, at which point the audio-only mode was
checked for. Now we are more explicit and act on SET_AUDIO_MUTED.
Read the muted state from the track itself instead of from base/media. This
avoid expressing the incorrect desire when the call starts muted because
permission was never granted.
In the current middleware logic, when the local participant becomes
dominant speaker, a new participant can be selected to receive
high quality video from. This means large-video could potentially
do a switch to another participant when the local participant
becomes dominant speaker. Prevent such behavior.
* Adds an option to disable features based on token data.
Reverts changes from b84e910086, removes disableDesktopSharing option and an interface_config option.
* Disable recording button based on token features data.
Hide recording if local participant isGuest and roles based on token.
When enableUserRolesBasedOnToken is enabled we were not hiding the record button for guests.
* Adds filtering of jibri iqs and rayo based on features.
Moves feature checking in separate utility function.
Renames utility method.
* Adds a footer text when outbound-call is not feature enabled.
* Fixes comments.
Use it unless the connection is not ACTIVE. We don't really care if it's
recovering or whatever, if it's not active it has problems, so that's that.
This fixes a potential edge case in which the connection remains in RESTORING
state for some time.
* ref(toolbar): show recording features based on explicit configs
* squash: bring back button configs, use final config names
* squash: update interfaceConfig comment, remove unused config whitelist
* squash: change order of button enabled checks to reduce diff
* squash: fileRecording -> fileRecordings
* feat(recording): add sounds for when recording starts and stops
* squash: use constants, play sounds for file only
* squash: rename recordingStopped.mp3 -> recordingOff.mp3
* squash: flip var declaration for alpha order
Layout:
Use an absolute-fill view as the background with the sidebar on top of. This
greatly simplifies styling, as there is no need to calculate how large the
backdrop needs to be.
Animation:
Switch to a translateX transform animation. This serves 2 purposes: first,
there seems to be a bug somewhere in React Native 0.51-0.55 where the content
that is being animated starts to be clipped. Very weird! But more importantly,
translateX transmorm animations are supported by the native animation driver!
https://facebook.github.io/react-native/blog/2017/02/14/using-native-driver-for-animated.html8f5ebe5952/Libraries/Animated/src/NativeAnimatedHelper.js (L138-L176)
This makes the animation more performant and buttery smooth.
Some small cleanups are also included here.
Instead of keeping dominant speaker locally, get it from redux and be
updated when the dominant speaker changes. This is in an attempt to mimic
the video layout being reactified and connected to redux.
The visibility of the toolbar can be toggled by interacting with the main screen.
This change allows the toolbar to be configured to be 'always visible'. This voids
the 'toggle' functionality.
Simplify parts of the logic introduced in
11b7144ad0.
Specificaly, using all the state change avoiding functions doesn't give us much
since we need to copy the state for sure.
Makes sure that whenever a conference is left or switched, the local
participant's id will be equal to the default value.
The problem fixed by this commit is a situation where the local
participant may end up sharing the same ID with it's "ghost" when
rejoining a disconnected conference. The most important and easiest to
hit case is when the conference is left after the CONFERENCE_FAILED
event.
Another rare and harder to encounter in the real world issue is
where CONFERENCE_LEFT may come with the delay due to it's asynchronous
nature. The step by step scenario is as follows: trying to leave a
conference, but the network is not doing well, so it takes time,
requests are timing out. After getting back to the welcome page the
the CONFERENCE_LEFT has not arrived yet. The same conference is joined
again and the load config may timeout, but it will be read from the
cache. Now the network gets better and conference is joining which
results in our ghost participant added to the redux state. At this point
there's the root issue: two participants with the same id, because the
local one was neither cleared nor set to the new one yet
(PARTICIPANT_JOINED come, before CONFERENCE_JOINED where we adjust the
id). Then comes CONFERENCE_JOINED and we try to update our local id.
We're updating the ID of both ghost and local participant. It could be
also that the delayed CONFERENCE_LEFT comes for the old conference, but
it's too late and it would update the id for both participants.
The approach here reasons that the ID of the local participant
may be reset as soon as the local participant and, respectively, her ID
is no longer involved in a recoverable JitsiConference of interest to
the user and, consequently, the app.
Co-authored-by: Pawel Domas <pawel.domas@jitsi.org>
Co-authored-by: Lyubo Marinov <lmarinov@atlassian.com>
If the ID of a remote participant was the same as the ID of the local
participant (across multiple conferences), removing the remote
participant on PARTICIPANT_LEFT would remove the local participant.
Like the preceding commit "ref(base/conference): clear the 'conference'
field on WILL_LEAVE", this commit is part of the story how we are to
deal with conferences which take noticeable time to leave.
Like the preceding commit "ref(base/conference): clear the 'conference'
field on WILL_LEAVE", this commit is part of the story how we are to
deal with conferences which take noticeable time to leave.
If a leave is delayed and the leaving JitsiConference manages to sneak a
PARTICIPANT_JOINED in, it may create a remote participant who even
collides with the local participant.
With so many abstractions called conference, I'm not surprised I made a
mistake and my reviewer didn't catch it.
As we are transitioning from remote participants identified by ID alone
to an ID-conference pair, the subsequent commits "Protect against late
PARTICIPANT_JOINED" and "Refine PARTICIPANT_LEFT for ID collisions"
caught the error.
Pawel Domas said:
The conference state field is referring to the current conference in
progress, so it feels like this field should be cleared as soon as we
declare that the conference is being left and the asynchronous process
of leaving the conference starts (which happens on
CONFERENCE_WILL_LEAVE).
Lyubo Marinov said:
Merely setting conference to undefined is incomplete because there are
other redux state properties associated with conference.
What we're doing here is redefining the redux action
CONFERENCE_WILL_LEAVE:
The redux action CONFERENCE_WILL_LEAVE represents the order of the user
to leave a JitsiConference instance. From the user's perspective, there's
no going back (with respect to the instance itself). The app will perform
due clean-up like leaving the associated room, but the instance is no
longer the focus of the attention of the user and, consequently, the app.
Now CONFERENCE_LEFT and CONFERENCE_WILL_LEAVE have more in common than
before and we can do a much better job by expressing the common parts in
source code. Additionally, our source code can be rewritten to better
express the meaning of any of these two actions.
Co-authored-by: Pawel Domas <pawel.domas@jitsi.org>
Co-authored-by: Lyubo Marinov <lmarinov@atlassian.com>
Add ability to provide a display name in the configOverwrite object that
when available it will be used to customize the name of the meeting in
callkit screen and recent call list.
Co-authored-by: Daniel Ornelas <daniel.ob64@gmail.com>
Co-authored-by: Lyubo Marinov <lmarinov@atlassian.com>
This zOrder change fixes the issue that the scrollable filmstrip videos
rendered in front of the local participant, however there is still an
issue that needs to be fixed later: The rendered videos should have
overflow: 'hidden' property applied so then they don't get rendered when
scrolled out of the ScrollView, that property however doesn't seem to
work for Video components.
This fix is based on storing the location URL object we are loading the
configuration for in the redux store. Once the config has been loaded (or it has
failed, for that matter!) we'll check if the current "config URL" is the same we
set, and discard the old one if they don't match.
With some of the preceding commits in the "multiplying remote
thumbnails" story line, I started hitting this error with 100%
reproducibility:
1. Have a remote participant prepared in conferenceA. Web will do as
well.
2. On iOS prepare to join conferenceB in Safari and use the same device
for step 3.
3. Join conferenceA on the iOS device from step 2 with audio-only. The
audio-only is so that avatars are always visible. Wait for the remote
participant prepared in step 1 to appear.
4. Switch to Safari and hit "Continue in the app" to have the app leave
conferenceA and join conferenceB.
What happens:
After the iOS device joins conferenceB in the Jitsi Meet app, the local
participant is on the large video (as expected) but the avatar of the
local participant is the default audo-generated auto-colored
placeholder. That's because this error was hit and the avatar couldn't
be "fetched".
The Jitsi Meet app always has at most 1 conference of primary interest.
It may have to juggle with 2 JitsiConference instances at the same time
if 1 is in the process of being left and one is joining/joined. But the
one which is joining or joined (which we call conference in the
features/base/conference redux state) is the one "of interest", the
other one is "clean up". Consequently, the remote participants of the
conference "of interest" are the remote participants "of interest" and
the others are "clean up". In order to reduce the time during which
there may be multiplying remote thumbnails, clean the remote
participants who are no longer "of interest" up.
As part of the work on fixing the problem with the multiplying
thumbnails, we've associated remote participant w/ JitsiConference.
However, there are periods of time when multiple JitsiConferences are in
the redux state (and that period is going to be shorted by
StateListenerRegistry). In order to give more control to the feature
base/participants, reduce the occurrences of direct access to the
features/base/participants redux state and utilize the feature's
existing read access functions. Which will allow us in the future to
enhance these functions to access participants which are relevant to the
current conference of interest to the user only.
The commit message of "Associate remote participant w/ JitsiConference
(_JOINED)" explains the motivation for this commit.
Practically, _JOINED and _LEFT combined with "Remove remote participants
who are no longer of interest" should alleviate the problem with
multiplying remote participants to an acceptable level of annoyance.
Technically though, a remote participant cannot be identified by an ID
only. The ID is (somewhat) "unique" in the context of a single
JitsiConference instance. So in order to not have to scratch our heads
over an obscure corner, racing case, it's better to always identify
remote participants by the pair id-conference. Unfortunately, that's a
bit of a high order given the existing source code. So I've implemented
the cases which are the easiest so that new source code written with
participantUpdated is more likely to identify a remote participant with
the pair id-conference.
Additionally, the commit "Reduce direct read access to the
features/base/participants redux state" brings more control back to the
functions of the feature base/participants so that one day we can (if we
choose to) do something like, for example:
If getParticipants is called with a conference, it returns the
participants from features/base/participants who are associated with the
specified conference. If no conference is specified in the function
call, then default to the conference which is the primary focus of the
app at the time of the function call. Added to the above, this should
allow us to further reduce the cases in which we're identifying remote
participants by id only and get us even closer to a more "predictable"
behavior in corner, racing cases.
The plan set in motion here is to associate remote participants with the
JitsiConference instances that created them in order to be able to
remove remote participants when a JitsiConference is no longer the
primary focus of the jitsi-meet app. And that's supposed to alleviate a
problem with multiplying remote thumbnails.
Doing all of the above in a single commit is a bit of a high order. So
I'm splitting the whole into multiple successive commits for the
purposes of observability, comprehension. Each commit is supposed to be
safe even if subsequent commits are not accepted, are reverted,
whatever. Obviously, without the successive commits, a commit may be
"unused".
One of the important pieces of the multiplying remote thumbnails "fix"
offered is removing remote participants who are no longer of interest
i.e. PARTICIPANT_LEFT. But in order for _LEFT to be implemented, _JOINED
must be implemented first.
Saúl Ibarra Corretgé
virtuacoplenny
bbaldino
Hristo Terezov
paweldomas
Saúl Ibarra Corretgé
Daniel Ornelas
Lyubo Marinov
Bettenbuk Zoltan
Дамян Минков
Guus der Kinderen