* moving locales folder to /public and applying changes needed to serve locales as static files
* better dev/prod check
* transifex pull script changes to match new locales directory + ignore locales with less than 100 lines
* fix local/prod locales path
* merge fallback messages
* applies new locale changes to legacy client
`bbb-html5.nginx` file content should also be changed to the following:
```
location /html5client/locales {
alias /usr/share/meteor/bundle/programs/web.browser/app/locales;
}
location /html5client/compatibility {
alias /usr/share/meteor/bundle/programs/web.browser/app/compatibility;
}
location /html5client/resources {
alias /usr/share/meteor/bundle/programs/web.browser/app/resources;
}
location /html5client/svgs {
alias /usr/share/meteor/bundle/programs/web.browser/app/svgs;
}
location /html5client/fonts {
alias /usr/share/meteor/bundle/programs/web.browser/app/fonts;
}
location /html5client {
# proxy_pass http://127.0.0.1:4100; # use for development
proxy_pass http://poolhtml5servers; # use for production
proxy_http_version 1.1;
proxy_set_header Upgrade $http_upgrade;
proxy_set_header Connection "Upgrade";
}
```
This is a workaround that shaves off some edge cases around wrongfully done camera reconnections and outbound cameras getting stuck in the UI even though they arent being shared anymore
Problem: setReconnectionTimeout was being called in the first candidate generation to set the negotiation/reconnection timeout up. That caused some browsers or specific scenarios (mainly envs without STUN) to establish the negotiation (playStart) before generating any useful out-of-band candidates (relay). That would cause the timeout to be set AFTER it is supposed to be cleared due to success (playStart), making the webcam drop after a while
So I moved the setReconnectionTimeout call to a safer spot: right after the first negotiation requisition goes out to bbb-webrtc-sfu
There's some scenarios that video errors triggers multiple toast notifications
that don't have any mapped defined message feedback so they all drop to the default
permission error. This leaves the impression that something is broken at the toast
container.
Since those messages don't bring up much information about the problem we can avoid sending
them until we don't have a more informative one to notify the user.
Depending on the sequence of events on the server, user sometimes is
added to the users collection with the name information, and sometimes this
information is updated in a second moment. Once we need the username
to show the push notification, we observe changes on this field for both
added and changed event.
Closes#11307
When validating tokens, the dummyUser created at the beginning is set
with validated=true. This means there won't be the state change that used
to occur from validated:false to validated:true (which detects the moment
joined the meeting) , therefore the alert code that expects for this change
won't run.
To fix this for audio alerts, we now detect when user join by observing
additions in Users's collection. This is actually good because we start
observing this only once (in componentDidMount), differently we used to do,
by calling this every time the tracker was activated.
To distinguish between the user addition that initially populates user's
collection from those that happens after user join (which are the ones we
want), we store the initial state (at componentDidMount) and compare it
with new additions. If the user added is present at the initial state,
then it is an addition to populates the collection, otherwise this is a real
user addition (happened after user joined the meeting)
Partially fixes#11399
Moderators are able to send a message to the meeting's guest lobby. This new
event reaches bbb-web and is sent to the guest user with her/his status response
while polling. All guest users that are waiting for acceptance will be able to
read this message.
enableGuestLobbyMessage is disabled by default.
It is safer to look at subscriptions's state to users instead of depending
on the initial state of the collection at mounting time. Although
i couldn't force any race conditions on this, there's a chance it could
happen. The fix avoid this problem.
This complements #11533
Currently this information is lost everytime breakout-room component is
unmounted, causing the panel to shows wrong information during next renders
Fixes#11333
Replaces the PR #10604, which has been temporarily fixed a problem in 2.3a2, that is, the slide size gets crazy when turning on and off fullscreen mode. Since #10604 was just a workaround, I tried to fix it in a way making more sense.
### Closes Issue(s)
closes#10459 (already closed by #10604) and #10589
### Motivation
#10604 is fixing the problem by calling "slide change" event, although the slide has not been changed actually. I tried to fix the same problem by adding a "full screen change" event listener in the layout manager, which I think is more elegant.
Without 'exact' match, the browser fallbacks to the default inputDeviceId
This prevents the error (input device error) when breakout is ended and we try
to skipCheck the microphone when user returns to main room (assuming the
user had the microphone active before joining breakout room).
Debounce could make the behaviour of "not always ringing when a guest
enters" feel "buggy". Throttle makes it clearer that it happens to
prevent it from ringing too frequently.
Also move Throttle time from magic number to constant for
a meaningful variable name.
Both debounce or throttle could be used to prevent a spammer from
annoying the moderators.
With debounce, a spammer won't even be noticed by the mods, but
there won't be new notifications for legit guests during the spam.
With throttle, a spammer could still annoy the mods, we would only make
the interval between notifications bigger.
It is a tradeoff. An ideal solution would be preventing spamming from
the same user, but probably unnecessarily complex.
When validating tokens, the dummyUser created at the beginning is set
with validated=true. This means there won't be the state change that used
to occur from validated:false to validated:true (which detects the moment
joined the meeting) , therefore the alert code that expects for this change
won't run.
To fix this for audio alerts, we now detect when user join by observing
additions in Users's collection. This is actually good because we start
observing this only once (in componentDidMount), differently we used to do,
by calling this every time the tracker was activated.
To distinguish between the user addition that initially populates user's
collection from those that happens after user join (which are the ones we
want), we store the initial state (at componentDidMount) and compare it
with new additions. If the user added is present at the initial state,
then it is an addition to populates the collection, otherwise this is a real
user addition (happened after user joined the meeting)
Partially fixes#11399
Without 'exact' match, the browser fallbacks to the default inputDeviceId
This prevents the error (input device error) when breakout is ended and we try
to skipCheck the microphone when user returns to main room (assuming the
user had the microphone active before joining breakout room). Related
to the feature c451666d52
Instead, return empty array when both users and groupChats arrays are empty.
The public chat and user chats may be active when groupChats array is empty.
Fixes#11476
After ending the notification playback, we set the ".src" property to null, which immediately stop the internal player of mobile browser (tested on Chrome for Android - device list is on #11458).
For the specific list of devices, this prevents the internal buffer error "-61" described in #11458.
Fixes#11458.
Locales would be fetched even if you simply open and save settings window (without changing any setting)
We now test it similarly we do on 2.2
Closes#11318
* updating old tests + collecting more snapshots [WIP]
* updates old test suites and collects more visual regressions screenshots
* remove snapshots and their collection temporary
* run tests from packages.json
* update test execution command/export constants from .env to core/constants.js
* update tests/puppeteer/README.md file
* update LOOP_INTERVAL variable call and assign timeouts to the webcam share spec
* redefine waitForSelector func in page.js, update chat test suite specs and add poll chat message test spec
* Merge remote-tracking branch 'upstream/develop' into updating-old-tests-visual-with-visual-regressions
* update webcam test specs collecting videoPreviewTimeout and use it to wait for videoPreview selector
* update custom parameters test suite
* update breakout test suite
* update webcam layout test suite
* update multiusers test suite
* update notifications test suite
* update presentation test suite
* whiteboard test suite
* screenshare test suite
* update sharednotes test suite
* user ELEMENT_WAIT_TIME variable from timeouts constants.js
* list TEST CONSTANTS by category
* add poll test suite and assigns the right unassigned timeouts
* set test pages to headless
After audio reconnection, a muted user would have it's microphone unmuted by default, unless muteOnStart is set to true. This fix this problem.
Fixes#9016
This bug was introduced in #11305, as the system-generated message notifying
the user that the chat history has been cleared does not have an id.
fixes#11349
Associate pads with meetings so session validation is restricted to the
meeting's valid session tokens.
Meteor will dispatch new redis events on shared notes and closed captions
pads creation. This event will go through apps and reach web to populate
a new meeting's pad collection that contains all valid pad id's for that
session. Nginx will use this collection to check if the user's session token
belongs to the pad's authorized users.
Besides these modifications, an extra change will be needed at notes.nginx.
Location /pad/p/ needs to change it's auth_request:
from /bigbluebutton/connection/checkAuthorization;
to /bigbluebutton/connection/validatePad;
When managing Etherpad's pads, Meteor makes API calls to initiate the closed captions
and shared notes modules. The pad id was being mapped to a shorter id than the meeting
id because of a Etherpad lenght limitation.
Changed to something less guessable.
Mainly: track unhealthy fatal states instead of supposedly healthy states
Avoids unwarranted UI reconnecting states or flickering on low end devices or struggling networks
Also added a version of isVideoBroadcasting for presenters which only accounts for the server state, so that we can deal with the presnter`s sharing state with more fidelity in the UI
Done to avoid false positives where the stream would transition to the unhealthy UI state (or flicker between states) due to disconnected also triggering it, which is not fatal. Disconnected has to be used alongside getStats heuristics, which I removed due to issues with it, so I´m hardening the transition to fatal states only
