Safari may enter a microphone permission check loop due to buggy behavior
in the Permissions API. When permission isn't permanently denied, gUM
requests fail with a NotAllowedError for a few seconds. During this time,
the permission state remains 'prompt' instead of transitioning to 'denied'
and back to 'prompt' after the timeout.
This leads to an issue where, on retrying while in 'prompt' + blocked,
the client loops through gUM checks via: 1) checking permission in the API,
2) receiving 'prompt', so trying gUM, 3) gUM fails, 4) returning to the
modal and checking permission again because the API still says 'prompt'.
Additionally, the `isUsingAudio` flag incorrectly counts the local echo
test/audio settings modal as "using audio," which toggles the flag on/off,
triggering the useEffect that causes the loop more frequently.
To fix this, remove the unnecessary AudioModal permission check that
causes the loop. Also, exclude "isEchoTest" from the `isUsingAudio` flag.
When listen only mode is deactivated and an user joins audio, an incorrect
remount of AudioSettings can trigger a spurious mute toggle. This happens
because AudioManager clears the `isConnecting` flag before setting the
`isConnected` flag. This creates a brief period where audio is flagged as
"disconnected," leading to a remount and unmount cycle that causes unwanted
mute/unmute actions due to AudioSettings' logic of muting/unmuting
active devices.
Ensure the `isConnected` flag is set before clearing the
`isConnecting` flag, preventing audio from being incorrectly flagged as
disconnected.
When going from "no mic" -> mic via the unmute action, the client isn't
unmuting itself after confirming the change. This is caused by not
waiting the liveChangeInputDevice method (which is a Promise) to be
fully executed before unmounting the AudioSettings modal -- the one
responsible for triggering the unmute. Since it unmounts before the
device is changed, the unmute action will be ignored because the device
is still "listen-only" (no mic).
Properly unmute audio when transitioning from "no mic" -> "mic" via the
unmute trigger by waiting for liveChangeInputDevice to resolve.
Additionally, some general improvements to UI/UX:
- Display the AudioSettings modal title when gUM is on prompt mode
- Add specific subtitles to the AudioSettings modal to 1) warn that no
mic is selected 2) Give a hint that the user can test their devices
- Always honor settings.yml's "initialHearingState" state (whether
local echo feedback should be played by default in AudioSettings)
This is a rework of the audio join procedure whithout the explict listen
only separation in mind. It's supposed to be used in conjunction with
the transparent listen only feature so that the distinction between
modes is seamless with minimal server-side impact. An abridged list of
changes:
- Let the user pick no input device when joining microphone while
allowing them to set an input device on the fly later on
- Give the user the option to join audio with no input device whenever
we fail to obtain input devices, with the option to try re-enabling
them on the fly later on
- Add the option to open the audio settings modal (echo test et al)
via the in-call device selection chevron
- Rework the SFU audio bridge and its services to support
adding/removing tracks on the fly without renegotiation
- Rework the SFU audio bridge and its services to support a new peer
role called "passive-sendrecv". That role is used by dupled peers
that have no active input source on start, but might have one later
on.
- Remove stale PermissionsOverlay component from the audio modal
- Rework how permission errors are detected using the Permissions API
- Rework the local echo test so that it uses a separate media tag
rather than the remote
- Add new, separate dialplans that mute/hold FreeSWITCH channels on
hold based on UA strings. This is orchestrated server-side via
webrtc-sfu and akka-apps. The basic difference here is that channels
now join in their desired state rather than waiting for client side
observers to sync the state up. It also mitigates transparent listen
only performance edge cases on multiple audio channels joining at
the same time.
The old, decoupled listen only mode is still present in code while we
validate this new approach. To test this, transparentListenOnly
must be enabled and listen only mode must be disable on audio join so
that the user skips straight through microphone join.
* Refactor: Make bundle using webpack
* Fix: restore after install codes and a few settings
* Fix: build script folder permission
* Refactor: Remove support to async import on audio bridges
* Upgrade npm using nvm
* Avoid questions on npm ci execution
* Let npm ci install dev dependencies (as we need the build tools here)
* Fix: enconding
* Fix: old lock files
* Remove: bbb-config dependency to bbb-html5 service, bbb-html5 isn't a service anymore
* Fix: TS errors
* Fix: eslint
* Fix: chat styles
* npm install with "lockfileVersion": 3 (newer npm)
* build: allow nodejs 22
* node 22; drop meteor from CI and bbb-conf
* TEMP: use bbb-install without mongo but with node 22 and newer image
* build: relax nodejs condition to not trip 22.6
* build: ensure dir /usr/share/bigbluebutton/nginx exists
* init sites-available/bbb; drop disable-transparent-
* nginx complaining of missing file and ;
* TMP: print status of services
* WIP: tweak nginx location to debug
* Fix: webcam widgets alignments
* akka-apps -- update location of settings.yml
* build: add locales path for nginx
* docs and config changes for removal of meteor
* Fix: build encoding and locales enpoint folder path
* build: set wss url for media
* Add: Enable minimizer and modify to Terser
* Fix: TS errors
---------
Co-authored-by: Tiago Jacobs <tiago.jacobs@gmail.com>
Co-authored-by: Anton Georgiev <anto.georgiev@gmail.com>
Co-authored-by: Anton Georgiev <antobinary@users.noreply.github.com>
* refactor(storage): replace Tracker.Dependency with observer hook
* fix(storage): set initial value
* refactor(storage): stop using Meteor's Session singleton
The audio troubleshooting modal has very microphone-specific strings,
which might confuse users trying to join listen only.
Review the Help screen so that listen only scenarios are more generic.
As a bonus, review the unknownError locale with a more actionable text.
Listen only has a built-in retry routine on join failures that's
convoluted half-broken. It stems from the Kurento era where it could
fail randomly due to a myriad of reasons.
Production logs indicate that the retry is seldom used nowadays in
mediasoup-based environments. The presence of the retry also breaks
the error troubleshooting modal when actual failures happening, leaving
users in the dark about what's happening.
Remove the listen only retry code from AudioManager and bubble up any
join failure to the callers.
In scenarios where the join audio flow skips echo test, NotAllowedError
(and any other errors) are all being mashed together under a generic
MEDIA_ERROR object.
Properly handle specific errors in audio-manager so they're correctly
render in the audio modal help screen.
WebRTC-based stats generation in the connection status modal is broken
on Firefox >= 125. A broken type check coupled with a new partially
implemented RTCIceTransport dictionary causes and undefined function
call when fetching the selected candidate pair. Since that error is
unhandled, collection breaks.
Correctly check for the getSelectedCandidatePair method availability in
RTCIceTransport so that it skips to pair inference from getStats if
necessary.
Audio exit toasts are fired in some redundant situations, e.g.: when the
error help screen is toast.
Change the logic a bit so that it's only fired when the audio help modal
won't be shown, i.e.: when audio had succesfully connected.
There are some situations where previously set deviceIds (
local/session storage) may become stale. This causes an unexpected
behavior where audio is temporarily borked until the user clears their
local storage.
This issue has been seen more recently on Safari endpoints when switching
back-and-forth breakout rooms in environments running under iframes.
Also seen randomly on endpoints with virtual input devices.
This centralizes audio gUM calling into a single method that retries the
gUM procedure without pre-set deviceIds only if the initial call fails
due with an OverconstrainedError - hopefully circumventing the issue.
Extract the deviceId again from the stream to guarantee consistency
between stream DID vs chosen DID. That's necessary in scenarios where,
eg, there's no default/pre-set deviceId ('') and the browser's
default device has been altered by the user (browser default != system's
default).
There's no rollback procedure in case a device switch fails right now,
nor does the code entrypoints that call the switching procedures wait
for resolution or failure before marking the new device as chosen. That
may cause inconsistent states in a couple of ways:
- No rollback: switch fails, audio is still on but no actual
microphone input is being transmitted
- Not waiting for resolutions: inconsistent chosen devices on failures
Device switching errors are also not surfaced to the end user
This commit:
- Adds device rollback and proper resolution/failure response
awaits to try and make the state a bit more consistent.
- Centralizes the input device switching code to be reused between
different bridges
- Centralizes device ID state management in audio-manager to try and
mantain them a bit more consistent across the board
- Surface device switching failures to the end user
- Guarantee device IDs are set to the session storage on all
appropriate scenarios
- Remove the old listen only bridge (kurento.js), superseded by the equivalent
and equally stable (AS FAR AS LISTEN ONLY IS CONCERNED) sfu-audio-bridge
- Rename FullAudioBridge.js -> sfu-audio-bridge.js
* A more generic name that better represents the capabilities and
the nature of the bridge
* The bridge name identifier in configuration is still the same
('fullaudio')
- Remove the FreeSWITCH listen only fallback
- Temporarily disable the "trickle ICE" pair gathering feature used
in SIP.js (which was always experimental, nonstandard and disabled
by default)
- Updates to settings.yml keys in places where relevant
"default" is not an universally valid default value for deviceIds which was causing issues with Firefox and Safari in some specific scenarios where exact deviceId constraints were being used
Seems to have been introduced by a partial merge commit
There were a bunch of style changes introduced by that partial commit as well; I kept those changes to avoid introducing further conflicts between v2.4-2.5...
When joining breakouts, we now wait for the bridge to be loaded before
automatically start user's audio.
This problems happens only on fullaudio bridge
