This patch creates an `ObservableMap` for `wasm32-unknown-unknown` which
simply wraps a `BTreeMap`, and without the `stream` method. Indeed, the
first implementation uses `eyeball_im::ObservableVector` which requires
`Send` and `Sync` on its values, which cannot compile to Wasm.
This patch implements `Client::rooms_stream` by forwarding the
result of `matrix_sdk_base::Client::rooms_stream`, and mapping the
`matrix_sdk_base::Room` to `matrix_sdk::Room`.
Running `cargo test -p matrix-sdk-base --features e2e-encryption`
makes the code to fail to compile because `crate::latest_event` isn't
defined. And indeed, it must be defined if `experimental-sliding-sync`
is enabled, but also if `e2e-encryption` is enabled.
This patch updates `Store::rooms` from a
`Arc<StdRwLock<BTreeMap<OwnedRoomId, Room>>>` to a
`Arc<StdRwLock<Rooms>>` where `Rooms` is a new type that mimics a
`BTreeMap` but that is observable. It uses an `ObservableVector`
for saving us the costs of writing a new `ObservableMap` type in
`eyeball-im`. It would have too much implications that are clearly
not necessary. The major one being that an `ObservableMap` must emit
`MapDiff`, but we expect `VectorDiff` everywhere in the SDK where rooms
are observable. It would break too many API and too many projects.
The benchmark is coming, but here are the results (for 10'000 rooms,
extreme case):
* `get_rooms` and `get_rooms_filtered` are twice faster (in practise, it
means 650µs is saved per call),
* `get_room` and `get_or_create_room` are 10% slower (in practise, it
means 8-10ns is lost per call).
Overall, I believe these results are acceptable, and even an improvement
for the first one.
This patch removes the useless `SlidingSyncList::get_room_id` method as
I don't see how it can be useful for anybody. It's mostly dead public
code, let's clean that up.
This might be controversial, but I do strongly prefer explicit over
implicit, in this case: it helps looking at the use cases, when using
the LSP's goto_references().
The HttpError variant was misplaced, as it was always used when the
`user_id()` is missing, which is causing `Error::AuthenticationRequired`
errors everywhere else in the code base. This patch streamlines usage of
`Error::AuthenticationRequired`, and gets rid of the `HttpError`
variant.
This patch renames `Client::get_rooms`, `::get_rooms_filtered` and
`::get_room` to respectively `::rooms`, `::rooms_filtered` and `::room`.
This `get_` prefix isn't really Rust idiomatic.
This patch renames `Store::get_rooms`, `::get_rooms_filtered` and
`::get_room` to respectively `::rooms`, `::rooms_filtered` and `::room`.
This `get_` prefix isn't really Rust idiomatic.
`Store::get_rooms` worked as follows: For each room ID, call
* Take the read lock for `rooms`,
* For each entry in `rooms`, take the room ID (the key), then call `Store::get_room`, which…
* Take the read lock for `rooms`,
* Based on the room ID (the key), read the room (the value) from `rooms`.
So calling `get_rooms` calls `get_room` for each room, which takes the lock every time.
This patch modifies that by fetching the values (the rooms) directly
from `rooms`, without calling `get_room`.
In my benchmark, it took 1.2ms to read 10'000 rooms. Now it takes 542µs.
Performance time has improved by -55.8%.
This patch replaces the `RwLock` by a `Mutex` in
`RoomListService::rooms` because:
1. it removes a race condition,
2. `RwLock` was used because the lock could have been taken for a long
time due to the previous `.await` point. It's no longer the case, so we
can blindly use a `Mutex` here.
This patch makes `RoomListService::room` synchronous. It no longer reads
a `SlidingSyncRoom` from `SlidingSync`, then it not needs to be async
anymore. This patch replaces the `RwLock` of `RoomListService::rooms`
from `tokio::sync` to `std::sync`.
The patch updates all calls to `RoomListService::room` to remove the
`.await` point.
This patch updates `room_list_service::Room::new` to take a `&Client`
and a `&RoomId` instead of a `SlidingSyncRoom`. The `SlidingSyncRoom` is
only used in `Room::default_room_timeline_builder` and is fetched there
from the `SlidingSync` instance lazily. It confines the
`SlidingSyncRoom` to one single method for `Room` now.
Using the resolved homeserver URL causes problems if we need to inspect
the well-known configuration of the homeserver, for example, if the
server name is matrix.org, but the homeserver URL is server.matrix.org,
the well-known might be only available for the former.
This is why we also need to receive the former, i.e. the server name in
the QR code data.
This patch removes `RoomInfo::latest_event`. To get the latest event,
one has to use `RoomListItem::latest_event` because it supports
local events and remote events. It was supposed to be the case of
`Room::room_info` too, but only when the method was called: Once the
`RoomInfo` was created, the latest event inside `RoomInfo`
was static, not dynamic. Also, this code wasn't tested
contrary to `RoomListItem::latest_event` which uses
`matrix_sdk_ui::room_list_service::Room::latest_event` which is itself
tested.
This patch changes the behaviour of the `Timeline` when a duplicated
event is received. Prior to this patch, the old event was removed, and
the one was added. With this patch, the old event is kept, and the new
one is skipped.
This is important for https://github.com/matrix-org/matrix-rust-sdk/pull/3512
where events are mapped to the timeline item indices. When an event is
removed, it becomes difficult to keep the mapping correct.
This patch also adds duplication detection for pagination (with
`TimelineItemPosition::Start`).
This patch removes the `matrix_sdk_ui::timeline::SlidingSyncRoomExt`
trait as it's no longer necessary since `room_list::Room::latest_event`
no longer uses `SlidingSyncRoom` to fetch the latest event.
This patch updates `matrix_sdk_ui::room_list::Room::latest_event`
to fetch the latest event from `matrix_sdk::Room` instead of
`matrix_sdk::sliding_sync::SlidingSyncRoom`. It removes one dependency
to `SlidingSyncRoom` and it simplifies the code.
This avoids a race condition where the caller hasn't set up the initial
items or the listener, and the listener is called *before* the initial
items have been used.
second part of the #3439 breakdown
The context for this API, is for the composer preview an reply without the need of it being actively in the timeline.
