* Improve accessibility and testability of Tooltip
Adding a role to Tooltip was motivated by React Testing Library's
reliance on accessibility-related attributes to locate elements.
* Make the ReadyWatchingStore constructor safer
The ReadyWatchingStore constructor previously had a chance to
immediately call onReady, which was dangerous because it was potentially
calling the derived class's onReady at a point when the derived class
hadn't even finished construction yet. In normal usage, I guess this
never was a problem, but it was causing some of the tests I was writing
to crash. This is solved by separating out the onReady call into a start
method.
* Rename 1:1 call components to 'LegacyCall'
to reflect the fact that they're slated for removal, and to not clash
with the new Call code.
* Refactor VideoChannelStore into Call and CallStore
Call is an abstract class that currently only has a Jitsi
implementation, but this will make it easy to later add an Element Call
implementation.
* Remove WidgetReady, ClientReady, and ForceHangupCall hacks
These are no longer used by the new Jitsi call implementation, and can
be removed.
* yarn i18n
* Delete call map entries instead of inserting nulls
* Allow multiple active calls and consolidate call listeners
* Fix a race condition when creating a video room
* Un-hardcode the media device fallback labels
* Apply misc code review fixes
* yarn i18n
* Disconnect from calls more politely on logout
* Fix some strict mode errors
* Fix another updateRoom race condition
* Move all polls processing to events-sdk
This makes polls support the full range of extensible events (both parsing and generation).
* Appease the linter
* Fix & update tests
* Update events-sdk for polls bugfix
* Update events-sdk for typechecking
* Add missing type cast
* Update per review
In 9969b01c5f we stopped updating the sublist whenever we felt like it, which indirectly froze message previews for room tiles (badges, unread state, etc were unaffected because that is managed by a different store). To fix this, we simply have to listen for changes and perform an update.
This means we're abusing the AsyncStoreWithClient to get access to a lifecycle, but overall that seems like a minor crime compared to the time spend abusing the store's state as a map.
With thousands of rooms shown, we can save on average 743ms per preview. The new preview time is 0.12ms on average.
This reverts earlier changes made to textForEvent as they are no longer needed.
This also implements an entire tree of textForEvent-like behaviour as the previews need to be different, which is easiest done with its own stack.
