1190 Commits

• Eagerly connect to latest server instance for best online experience (#37506)

  Client side requirements for making the client connect as soon as
  possible, based on how the client is being used. This is especially
  important with the introduction of ranked play: previously the worst
  case scenario would be that you couldn't join a multiplayer room (or
  spectate a user) and this was [automatically
  handled](https://github.com/ppy/osu/blob/f66e2c432fdb08db46477c4fa08ca74e551d037f/osu.Game/Online/Multiplayer/OnlineMultiplayerClient.cs#L115-L121)
  mostly*, but now, if you leave the game open for a while, you can
  potentially be stuck queueing in ranked play with no users remaining on
  your server.
  
  Some samples of how this looks follow. Do note that the client is
  showing "Server is shutting down" errors. This is only going to happen
  in local debug environments – In production, when you reconnect to the
  endpoint you will always get a non-shutting-down instance.
  
  Idle scenario:
  
  
  https://github.com/user-attachments/assets/dd47fdf6-8d49-48e3-a19f-b196a581070b
  
  Non-idle scenario:
  
  
  https://github.com/user-attachments/assets/dfc8a41a-83fb-4b08-94b4-9595faf88294
  
  * Spectator isn't handled properly, as one example.

  Dean Herbert · 2026-05-05 16:06:00 +09:00
  51b4e89773
• Implement duels for ranked play (#37556)

  I've seen this suggested quite a bit and is a pretty easy implementation
  all things considered.
  
  For now, while on the queue screen, you can open up the dashboard
  overlay and select another player to duel. This will bring you into an
  unranked lobby.
  
  
  https://github.com/user-attachments/assets/712897a9-9350-4741-899d-59662c722e43

  Dan Balasescu · 2026-04-29 16:12:27 +09:00
  a3b8b9aee9
• Add additional information to the matchmaking queue screen (#37229)

  - [x] Depends on https://github.com/ppy/osu/pull/37226
  - [x] Depends on https://github.com/ppy/osu-server-spectator/pull/464
  
  This adds two new components to the queue screen:
  - A listing of the most recently completed matches (global).
  - A rank distribution graph.
  
  It looks something like this (fake data / test scene):
  
  <img width="1669" height="1005" alt="image"
  src="https://github.com/user-attachments/assets/caa57119-4267-4c6e-9898-2f414de865bf"
  />
  
  It's completely dev-design(TM), but I used Lichess as inspiration for
  the graph, and the original design document as inspiration for the
  panels.
  
  As for the history, because these are _completed_ matches one of the
  player life points will always be 0, but I've designed it so as to
  possibly support showing ongoing matches too in the future. It's only
  supported for ranked play right now, though there is no reason we
  couldn't track quick play rooms too (it's just... I'm not sure how to
  design the panels for quick play).
  
  ---------
  
  Co-authored-by: Dean Herbert <pe@ppy.sh>

  Dan Balasescu · 2026-04-14 10:54:54 +09:00
  ac620ee375
• Fix lack of encode-decode stability when writing out mania beatmaps with some key counts (#37256)

  Closes https://github.com/ppy/osu/issues/37232.
  
  The actual fix is
  https://github.com/ppy/osu/commit/e959b20517497a093d3c00a17457c5d36bf57651;
  everything else is window dressing / test harness to ensure I don't try
  and do a wrong change like https://github.com/ppy/osu/pull/37251 did. I
  recommend reviewing commit-by-commit.
  
  See [this desmos](https://www.desmos.com/calculator/a5yjpacvxa) for
  visual explanation of change, I think it does a better job at explaining
  this than any words I could type here.
  
  Of note:
  
  - In the end this did only affect 14K but that should never be assumed
  when floating point is involved.
  - Test cases generated here were generated in stable manually.
  - Except for 11 / 13 / 15 / 17K which are not officially supported and
  which don't work in lazer due to orthogonal reasons (see comment added
  in this PR in `ManiaBeatmapConverter`), decoding in lazer was always
  fine.
  - My worry was that the old encoding method before this PR could
  potentially cause stable to move a note from one column to another but
  thankfully that is not the case. The old method of encoding columns as X
  positions does not cause issues wherein lazer reads them back
  differently than stable after encode.
  
  I checked this by checking out `master`, re-encoding all of the test
  stair-pattern nK beatmaps added in this PR on `master`, exporting that
  as compatibility, re-importing to stable, and cross-checking that the
  decoded beatmap is visually the same on lazer and on stable.
  
  This is important to check because if this wasn't the case, we'd
  potentially have cases of actual online beatmaps (remember that we have
  BSS now) wherein a beatmap plays differently on stable than on lazer due
  to notes moving between columns, and would need to screen for this being
  the case and potentially apply corrective / reconciliatory action.

  Bartłomiej Dach · 2026-04-11 00:38:22 +09:00
  6231e06ebe
• Do not send replay frames to spectator server if initial begin play invocation failed (#37159)

  RFC
  
  Until now, if the initial `BeginPlaySession()` call failed, the client
  would continue operating as if it didn't - it would still continue to
  send frames and call `EndPlaySession()` at the end of a session.
  
  Server-side, two things generally can happen after this:
  
  - The sent frames and the `EndPlaySession()` call are
  [completely](https://github.com/ppy/osu-server-spectator/blob/7bab117e9d161455485368f63a0607a9e53f9f8a/osu.Server.Spectator/Hubs/Spectator/SpectatorHub.cs#L122-L125)
  [ignored](https://github.com/ppy/osu-server-spectator/blob/7bab117e9d161455485368f63a0607a9e53f9f8a/osu.Server.Spectator/Hubs/Spectator/SpectatorHub.cs#L153-L157)
  as no-ops, or
  - A hub filter (like `ClientVersionChecker`) that failed the initial
  `BeginPlaySession()` call continues to fail the calls to
  `SendFrameData()` and `EndPlaySession()`, all the while creating a storm
  in logs, because it needs to throw `HubException`s to communicate to
  users that they need to update their game, and the exceptions can't be
  silenced from logs because they look like every other failure.
  
  To that end, this has two goals: reduce useless network traffic, and
  reduce noise in spectator server logs after the client version checks
  were recently reactivated.
  
  Probably needs tests, but unsure if everyone's going to be on board with
  this to begin with to be quite frank, so I'm leaving tests for when I'm
  told this needs tests.

  Bartłomiej Dach · 2026-04-09 03:40:32 +09:00
  ca44d8233d
• Add models for improvements to matchmaking lobby (#37226)

  1. Gives `MatchmakingJoinLobby` parameters.
  2. Adds additional data to lobby status update models.
  
  A further PR will build upon (2) to add more data to the queue screen.

  Dan Balasescu · 2026-04-08 18:43:51 +09:00
  8c6818e275
• Fix ranked play test failures from adding null playlist items (#37189)

  Example:
  https://github.com/ppy/osu/actions/runs/23900675414/job/69696255970?pr=37178#step:5:38
  
  Regressed in https://github.com/ppy/osu/pull/37172, cc @LiquidPL
  
  Would fail in multiple tests. I'm not going to spend time figuring out
  exactly why, I'm just going to guess that not all tests bother to set up
  the relevant playlist items for the cards or whatever.
  
  Some of the failing tests are flaky but not because the `item` here
  isn't sometimes null in those cases. It's always null, but the callbacks
  are probably scheduled or whatever and therefore have a chance to never
  run. Also some of the failures appear to cascade / spill from other
  tests as well.

  Bartłomiej Dach · 2026-04-03 22:11:50 +09:00
  796877a594
• Update room state on card played in TestMultiplayerClient (#37172)

  Noticed this not updating when working on tests.

  Krzysztof Gutkowski · 2026-04-02 16:19:02 +09:00
  2d85966074
• Rework ranked play chat to reduce overall area (#37097)

  In discussions, we've come to the conclusion to attempt to use a
  chat-bubble system to minimise the effective area of the chat. In
  particular, the results screen doesn't give us enough space to display
  the full chat box without overlapping the main screen content.
  
  This PR both adds the chat to the results screen, and reworks it to use
  such a bubble system (not sure what to call it, IM style?).
  
  
  https://github.com/user-attachments/assets/a8a88c51-8a9d-4a03-92b6-621112a15a41
  
  - New messages are previewed for 3 seconds.
  - When focusing and unfocusing the textbox, the history moves into
  expanded state (show the most recent 10 messages) or collapsed state
  (fade messages out ASAP).
  
  This is a bit of an initial implementation to get a feel of how it
  behaves, and there's more that can be done such as adding colours,
  improving the transforms, perhaps adding it to the intro screen
  (post-animation) but the structure's a bit weird atm.
  
  ---------
  
  Co-authored-by: Dean Herbert <pe@ppy.sh>

  Dan Balasescu · 2026-03-31 03:45:47 +09:00
  fc817627e5
• Various tournament-minded improvements for multiplayer (#37000)

  https://github.com/user-attachments/assets/042311f1-b8c3-479b-a173-13b93fe2c5cc
  
  - `/roll` command is now supported in multiplayer chat for all players
  (don't need to be a referee).
  - Referees are shown in the room with a special status.
  - Tournament mode rooms can be locked, which prevents users from
  changing team (and slot, whenever those get brought back). When the room
  is locked, the user team indicator shows a padlock icon on top to
  indicate the lock state.
  
  ---
  
  - Related: https://github.com/ppy/osu-server-spectator/issues/406
  
  Grab-bag because I really don't think splitting this into 3 PRs is very
  helpful.
  
  I was going to add an animation for rolling but I had a go on Friday and
  my attempt got more or less the same reception as a wet fart so I'm not
  trying again. Someone else can if so inclined. I have completely lost
  trust in my design senses.
  
  Contrary to stable the roll results are completely ephemeral and go away
  when the room is re-entered. This could be both considered good and bad.
  Not for me to say.

  Bartłomiej Dach · 2026-03-20 18:56:45 +09:00
  43c53b5a8d
• Fix cancelling a restart for update still quitting the game after one minutes (#36950)

  Closes #36816.
  
  - [x] Self review / testing.

  Dean Herbert · 2026-03-18 16:48:07 +09:00
  e0ccdfae94
• Implement ranked play (#36819)

  I don't really have much to say here. Instead, I'll give a brief history
  rundown that lists many pages of documentation you can read, if
  interested.
  
  - Started off as BTMC + Happy24 (Vivi)'s ["The
  Vision"](https://docs.google.com/document/d/1p1IpPmd2RICp8G4OqkCSs7u8Ug8FbFv8qqP0mfSrHf0/edit?tab=t.0#heading=h.fol093d9f9xi)
  - Initial
  [designs](https://www.figma.com/design/f5qqC57t9jFlgpzhRqUNVX/The-Vision?node-id=0-1&p=f)
  were led by Vivi.
  - Designs
  [morphed](https://www.figma.com/design/vtFmLrXKvWNyYiRjTezFTM/Untitled--Copy-?node-id=0-1&p=f)
  during development into what's currently present, led by @minetoblend.
  - There is some more ongoing work creating a [game design
  document](https://docs.google.com/document/d/1iffJFCsIBfYF0D4ogItSBEj6YBmbp-rdCpItAeaJiTA/edit?tab=t.0).
  
  **tl;dr:** Create something with the mechanics of a trading card game
  within osu!. The name of this is "ranked play".
  
  ---
  
  To be frank, a lot of stuff is missing here. Some of it I don't want to
  mention, because the point of this exercise is to get the system into
  the hands of players, gather feedback especially around mechanics, and
  discuss any further direction with the team.
  
  I am expecting a blanket approval on all of the new code, with
  particular attention to changes in existing components that I'll point
  out in a self review.
  
  There is also some [ongoing
  work](https://github.com/smoogipoo/osu/pulls) that may arrive in this
  branch prior to being merged.
  
  ---------
  
  Co-authored-by: maarvin <minetoblend@gmail.com>
  Co-authored-by: Marvin <m.schuerz@hautzy.com>
  Co-authored-by: Jamie Taylor <me@nekodex.net>
  Co-authored-by: ArijanJ <arijanj@proton.me>
  Co-authored-by: Dean Herbert <pe@ppy.sh>
  Co-authored-by: Tim Oliver <git@tim.dev>
  Co-authored-by: Joseph Madamba <madamba.joehu@outlook.com>
  Co-authored-by: Bartłomiej Dach <dach.bartlomiej@gmail.com>
  Co-authored-by: nil <25884226+voidstar0@users.noreply.github.com>
  Co-authored-by: Ботников Максим <mr.botnikoff@ya.ru>
  Co-authored-by: Denis Titovets <den232titovets@yandex.ru>
  Co-authored-by: Michael Middlezong <119022671+mmiddlezong@users.noreply.github.com>
  Co-authored-by: SupDos <6813986+SupDos@users.noreply.github.com>
  Co-authored-by: failaip12 <86018517+failaip12@users.noreply.github.com>

  Dan Balasescu · 2026-03-07 02:30:50 +09:00
  0988552567
• Move v2 files to final location

  This contains only renames and namespace updates.

  Dean Herbert · 2026-03-03 16:45:15 +09:00
  c28c64940a
• Refactor TestSceneScreenFooter to test entire OsuScreens (#36718)

  Part of the `ScreenFooter` refactor, which intends to move the footer
  content handling to `OsuScreen`. This commit updates the `ScreenFooter`
  test to operate on entire `OsuScreen`s, in order to better test the
  entire flow of pushing a screen, and having it create and add its own
  content to the footer.
  
  This should be 80-90% identical to the original test case structure
  wise, just that instead of manually manipulating the footer with
  `SetButtons()`, various screens with the appropriate buttons are being
  moved around the screen stack.
  
  Additionally this adds some more tests handling common use cases, as
  well as removes `TestLoadOverlayAfterFooterIsDisplayed()`, since as far
  as I understand the behaviour described in it doesn't actually happen in
  production code. From what I can see, Screens instantiate their overlays
  in `load()`, and then register them in `LoadComplete()`. There seems to
  be no case where a `ShearedOverlayContainer` is created in the middle of
  a screen's lifecycle.

  Krzysztof Gutkowski · 2026-03-01 22:36:07 +09:00
  b88cba0829
• Migrate sheared overlay tests to ScreenTestScene (#36736)

  Part of the screen footer refactor.
  
  Once footer content is being managed by `OsuScreen`, the current tests
  which simply create the tested overlay and `ScreenFooter` in a container
  will no longer work.
  
  This PR refactors them to use `ScreenTestScene` with the setup being
  creating a dedicated testing `OsuScreen` which does the bare minimum to
  create the tested overlay and necessary components (eg.
  `FooterButtonFreeModsV2` for `TestSceneFreeModsOverlay`).
  
  Most of the changes here can be described as
  `%s/<...>Overlay/screen.Overlay/g`, with some minor touchups as
  necessary, given that we're now testing a more complete flow which
  checks more things that were previously not handled by the tests.
  
  ## [Move footer to front in
  ScreenTestScene](https://github.com/ppy/osu/commit/f8740e0403b3c0badd60d394c737f2aa912a9ed6)
  
  Self-explanatory. Without it the footer would show below the actual
  overlay, breaking tests depending on manual input. For the sake of tests
  not breaking in CI, both #36718 and this have this included - would
  prefer the former to be merged first since it was already reviewed
  there.
  
  ## `TestSceneModSelectOverlay`
  
  There were a few tests (`TestColumnHidingOnIsValidChange`,
  `TestColumnHidingOnTextFilterChange`, and
  `TestHidingOverlayClearsTextSearch`) that would create a custom overlay
  instance instead of the globally provided one. I've tested both and the
  tests run fine with the default overlay, so they're now using that
  instead.
  
  ## `TestSceneFreeModSelectOverlay`
  
  Updated to use footer v2.
  
  ---------
  
  Co-authored-by: Dean Herbert <pe@ppy.sh>

  Krzysztof Gutkowski · 2026-02-28 04:22:58 +09:00
  105342e5bf
• Add type parameter to pool retrieval method
  Dan Balasescu · 2026-01-28 13:42:58 +09:00
  c1fc1edd9c
• Add metadata endpoint to refresh friend listing (#36386)

  Alternative to / closes https://github.com/ppy/osu/pull/36341
  See also: https://github.com/ppy/osu-server-spectator/pull/415
  
  This is a simple solution by adding a spectator endpoint to refresh the
  friend listing. A more complicated form of this is to make
  adding/removing friends only via spectator, but that would require
  osu-web changes too.

  Dan Balasescu · 2026-01-19 18:34:00 +09:00
  e62a01cf77
• Merge branch 'master' into new-playlists-song-select
  Dan Balasescu · 2025-12-17 13:33:34 +09:00
  67e0348fa8
• Adjust vote-to-skip to be explicit about states
  Dan Balasescu · 2025-12-06 13:23:16 +09:00
  2be50d917a
• Fix failing tests after disallowing object placements before first timing point
  Bartłomiej Dach · 2025-12-05 09:52:32 +01:00
  66ebce8c12
• Merge branch 'master' into new-playlists-song-select
  Dan Balasescu · 2025-11-18 14:08:12 +09:00
  a30f08be80
• Suffix introp methods with "Intro"
  Dan Balasescu · 2025-11-04 11:05:41 +09:00
  f4049c7ec1
• Resolve CI inspections
  Dan Balasescu · 2025-10-31 21:48:24 +09:00
  a9ca4634fc
• Add simple multiplayer skip overlay
  Dan Balasescu · 2025-10-31 21:39:41 +09:00
  b20a41c1e8
• Add support for vote-to-skip in multiplayer
  Dan Balasescu · 2025-10-31 21:39:41 +09:00
  373162df02
• Add interop models
  Dan Balasescu · 2025-10-30 19:04:49 +09:00
  a435dfe93e
• Add very simple countdown timer for quick play stages (#35433)
  Dean Herbert · 2025-10-24 23:07:04 +09:00
  1af462b692
• Use new implementation in ScreenTestScene
  Dan Balasescu · 2025-10-17 19:01:28 +09:00
  4be29425a0
• Add initial screen + test structure
  Dan Balasescu · 2025-10-17 09:16:46 +09:00
  bfa45a23ce
• Use new implementation in ScreenTestScene
  Dan Balasescu · 2025-10-16 20:37:24 +09:00
  c80fe7ab82
• Merge pull request #35251 from smoogipoo/fix-broken-tests

  Fix test failures during individual runs

  Dean Herbert · 2025-10-09 17:09:01 +09:00
  2136d95087
• Fix test failures during individual runs
  Dan Balasescu · 2025-10-09 16:34:55 +09:00
  6eaf91d31a
• Fix test scene leaks through RealmRulesetStore/RealmAccess
  Dan Balasescu · 2025-10-09 15:28:15 +09:00
  79c367d208
• Add ability to jump in quick play
  Dan Balasescu · 2025-09-30 18:38:30 +09:00
  e636a09e0f
• Fix incorrect default room types
  Dan Balasescu · 2025-09-24 14:28:04 +09:00
  83dafb4ef9
• Add stage-changing helper, use in test scenes
  Dan Balasescu · 2025-09-24 14:28:01 +09:00
  9df3bd9a98
• Set initial matchmaking room state
  Dan Balasescu · 2025-09-24 13:42:55 +09:00
  597a06ac38
• Attempt to fix intermittent test
  Dan Balasescu · 2025-09-23 15:09:45 +09:00
  3789010dd5
• Merge branch 'master' into matchmaking
  Dean Herbert · 2025-09-11 15:11:01 +09:00
  619a6e7321
• Query for available pools for selection (#131)
  Dan Balasescu · 2025-09-07 16:17:27 +09:00
  e0c11504a2
• Separate IMatchmakingClient and IMultiplayerClient

  Co-authored-by: Dean Herbert <pe@ppy.sh>

  Dan Balasescu · 2025-09-07 16:17:27 +09:00
  3786efaa5e
• Add matchmaking
  Dan Balasescu · 2025-09-07 16:17:27 +09:00
  111b98ef8e
• Replace old-skin with retro skin in tests
  clayton · 2025-09-05 21:21:52 +09:00
  3c1c537b1b
• Add matchmaking model types required for server-side deploy
  Dan Balasescu · 2025-09-04 18:26:56 +09:00
  815bf9c37b
• Merge branch 'master' into pinned-rooms
  Dean Herbert · 2025-09-01 16:01:50 +09:00
  4e976cff6a
• Move footer to ScreenTestScene
  Dan Balasescu · 2025-08-29 17:49:48 +09:00
  04ba5aa575
• Update tests
  Bartłomiej Dach · 2025-08-22 12:44:08 +02:00
  03e7e2b0d8
• Ignore more potentially incorrect data from BASS

  This is still a workaround but arguably it's something we could leave in
  place without much loss. I think this at least feels better than the
  previous code.
  
  Notably, you could argue the test coverage of the fail case is lower
  since made it implicit that all tests will avoid the "backwards seek"
  detections.  But we never really had tests correctly- fail on the original
  so I don't see any loss of value.

  Dean Herbert · 2025-08-14 19:20:30 +09:00
  82906300b4
• Fix leak due to missing Game disposal
  Dan Balasescu · 2025-07-09 03:37:02 +09:00
  74516a1eaa
• Update existing tests to work with Song Select v2
  Dean Herbert · 2025-07-01 17:25:38 +09:00
  397b9b3994