The recent "groupers" which extracted out timeline grouping logic forgot to
pass through the last event state for read marker computation. This causes the
read marker to become visible when e.g. returning to room if it was last placed
inside a grouped set of events (currently room creation and membership events).
Regressed by https://github.com/matrix-org/matrix-react-sdk/pull/4059
Related to https://github.com/vector-im/riot-web/issues/12338
We send a parameter to prevent being immediately logged in after
registration, but we regressed it from snake case to camelcase
during the course of cross-signing dev.
Fixes https://github.com/vector-im/riot-web/issues/12865
For https://github.com/vector-im/riot-web/issues/12877
Original error:
```
Warning: Render methods should be a pure function of props and state; triggering nested component updates from render is not allowed. If necessary, trigger nested updates in componentDidUpdate.
Check the render method of PersistedElement.
in PersistedElement (created by AppTile)
in div (created by AppTile)
in div (created by AppTile)
in AppTile (created by AppsDrawer)
in div (created by AppsDrawer)
in div (created by AppsDrawer)
in AppsDrawer (created by AuxPanel)
in div (created by AutoHideScrollbar)
in AutoHideScrollbar (created by AuxPanel)
in AuxPanel (created by RoomView)
in div (created by RoomView)
in div (created by MainSplit)
in MainSplit (created by RoomView)
in ErrorBoundary (created by RoomView)
in main (created by RoomView)
in RoomView (created by LoggedInView)
in div (created by LoggedInView)
in DragDropContext (created by LoggedInView)
in div (created by LoggedInView)
in LoggedInView (created by MatrixChat)
in ErrorBoundary (created by MatrixChat)
in MatrixChat
```
Fixes https://github.com/vector-im/riot-web/issues/12929
Note: we don't need this to fix conferences in our hosted instance (riot.im or modular.im), but it is a common thing for self-hosters, including sometimes ourselves, to accidentally make mistakes on.
The logic determining if a batch of events is already in the database
has been reported to be faulty. Since we know that when we add initial
checkpoints they need to go to the start of the timeline we can mark
them to do a full crawl which will skip the test if the events are
already in the database.
This is a quick fix for this and checkpoints that are added after a gap
in the timeline might incorrectly conclude the same resulting in missing
events if there was a gap and the issue is triggered.
J. Ryan Stinnett