* Correct test for docker container being healthy
* Correct docker healthcheck for busybox wget
* Repeatedly check the health state of the docker container
* Use until loop & rely on timeout-minutes
* Fix check to look at healthy state
---------
Co-authored-by: Michael Telatynski <7t3chguy@gmail.com>
* Track available filters and expose this info from the vm
- Adds a separate hook that tracks the filtered rooms and the available
filters.
- When secondary filters are added, some of the primary filters will be
selectively hidden. So track this info in the vm.
* Write tests
* Fix typescript error
* Fix translation
* Explain what a primary filter is
* Attach the new store to window
* Fetch rooms after space store is ready
If we fetch a list of rooms and then wait for the space store to be
ready, we will need some way of handling the onAction calls we get
while we wait. These calls are dropped now.
* Update quick settings menu to use a11y roles.
* Add e2e test to test quick menu rendering
* Use a testid for now.
* lint lint
* Revert aria changes
* revert managed
* write screenshot
* Implement the favourite filter
* Make the room node capable of dealing with filters
- Holds data to indicate which filters apply
- Provides method to check if a given set of filters apply to this node
- Provides a method to recalculate which filters apply
* Wire up the filtering mechanism in skip list
* Use filters in the store
* Remove else
* Use a set instead of map
* Add space to the bottom of the room summary actions below leave room
* 8x not 6x
* Spacing needs to be within the scoll content, add it to the bottom of the leave action
* Update RoomSummaryCard-test.tsx.snap
* Fix snapshot and add screenshot test
* Refactor submit rageshake so that it uses the new error codes.
* Improve error information given in Bug Report Dialog
* use type
* Refactor with generic error & policy link.
* lint
* lint
* Add BugReportDialog test
* fix time travel
* use waitFor while waiting for fetch to finish
* lint
* Drop error prefix as per https://github.com/matrix-org/rageshake/pull/90/commits/3973bb38ef5253ebc284f6a6a8975dc65dc77b75
* small fixes
* Don't change string here.
* Fixup i18n strings.
* Avoid destroying calls until they are hidden from the UI
We often want calls to exist even when no more participants are left in the MatrixRTC session. So, we should avoid destroying calls as long as they're being presented in the UI; this means that the user has an intent to either join the call or continue looking at an error screen, and we shouldn't interrupt that interaction.
The RoomViewStore is now what takes care of creating and destroying calls, rather than the CallView. In general it seems kinda impossible to safely create and destroy model objects from React lifecycle hooks, so moving this responsibility to a store seemed appropriate and resolves existing issues with calls in React strict mode.
* Wait for a close action before closing a call
This creates a distinction between the user hanging up and the widget being ready to close, which is useful for allowing Element Call to show error screens when disconnected from the call, for example.
* Don't expect a 'close' action in video rooms
These use the returnToLobby option and are expected to remain visible when the user leaves the call.
* prevent user from accidentally triggering multiple identity resets
* apply changes from review and update to latest design
* Use a CSS class and compound variable
* update snapshot
* Update test/unit-tests/components/views/settings/encryption/ResetIdentityPanel-test.tsx
---------
Co-authored-by: Richard van der Hoff <richard@matrix.org>
Co-authored-by: Richard van der Hoff <1389908+richvdh@users.noreply.github.com>
renovate[bot]