747 Commits

• Turn tests green, for a short while, maybe (#37218)

  ## [Rewrite `BackgroundMusicManager` to not run into framework
  breakage](https://github.com/ppy/osu/commit/622216d8911832c39fa4e126b2810e4e0f46cbf7)
  
  The attempted proper fix to this was
  https://github.com/ppy/osu-framework/pull/6727. Unfortunately when
  presented with [the framework
  bump](https://github.com/ppy/osu/pull/37217) with that change, CI says
  "you're stupid" and fails on some disposal idiocy that of course is
  undebuggable and irreproducible:
  
  The active test run was aborted. Reason: Test host process crashed :
  Unhandled exception. System.AggregateException: One or more errors
  occurred. (Object reference not set to an instance of an object.)
  ---> System.NullReferenceException: Object reference not set to an
  instance of an object.
  at osu.Framework.Audio.Sample.SampleChannelBass.Dispose(Boolean
  disposing)
  at System.Threading.ExecutionContext.RunInternal(ExecutionContext
  executionContext, ContextCallback callback, Object state)
  	--- End of stack trace from previous location ---
  at System.Threading.ExecutionContext.RunInternal(ExecutionContext
  executionContext, ContextCallback callback, Object state)
  at System.Threading.Tasks.Task.ExecuteWithThreadLocal(Task&
  currentTaskSlot, Thread threadPoolThread)
  	   --- End of inner exception stack trace ---
  	   at osu.Framework.Audio.AudioCollectionManager`1.UpdateChildren()
  	   at osu.Framework.Audio.AudioCollectionManager`1.UpdateChildren()
  	   at osu.Framework.Audio.AudioCollectionManager`1.UpdateChildren()
  	   at osu.Framework.Audio.AudioCollectionManager`1.UpdateChildren()
  	   at osu.Framework.Threading.AudioThread.OnExit()
  at osu.Framework.Threading.GameThread.setExitState(GameThreadState
  exitState)
  	   at osu.Framework.Threading.GameThread.RunSingleFrame()
  at osu.Framework.Threading.GameThread.<createThread>g__runWork|70_0()
  at System.Threading.ExecutionContext.RunInternal(ExecutionContext
  executionContext, ContextCallback callback, Object state)
  	--- End of stack trace from previous location ---
  at System.Threading.ExecutionContext.RunInternal(ExecutionContext
  executionContext, ContextCallback callback, Object state)
  
  
  (https://github.com/ppy/osu/actions/runs/24019928154/job/70046733058?pr=37217#step:5:119)
  
  I no longer have the energy for any of this shit.
  
  @nekodex would appreciate if you could check that I actually haven't
  broken anything with the bgm here. Seems okay to me in test scenes at
  least.
  
  ## [Apply lowest-effort maybe-fixing changes to a bunch of flaking
  tests](https://github.com/ppy/osu/commit/7bd3ca4adfcce5b90add11565a13f3fe9177ad5e)
  
  None of the failures are reproducible locally, of course. I'm tired of
  this. If anyone else wants to subject themselves to actually
  investigating any of these, by all means, godspeed and good luck.

  Bartłomiej Dach · 2026-04-06 17:59:39 +09:00
  809298ddeb
• Assorted test deflaking (#37191)

  A few quick ones from https://github.com/ppy/osu/issues/37190.
  
  ## [Rewrite `TestSceneDeleteLocalScore` to have less context menu
  containers (and hopefully no longer
  flake)](https://github.com/ppy/osu/commit/ea0bc5b72374056d1749f8058f4874b7ef66a241)
  
  As seen in
  https://github.com/ppy/osu/actions/runs/23890777748#user-content-r3s3.
  
  This is a speculative fix but I'm feeling somewhat confident about this
  one.
  
  `BeatmapLeaderboardWedge` has TWO separate `ContextMenuContainer`s
  itself, and the test mentioned here was bringing a third. I have a
  feeling that the test flaking may have something to do with the fact
  that the test logic would attempt to click a menu item on specifically
  ONE of the three context menus. My bet is that when it fails, it's
  because it's trying the wrong one, but I don't have reproduction.
  
  ## [Wait for text to appear in flaking
  `TestSceneDrawableRoomPlaylist.TestSelectableMouseHandling`
  test](https://github.com/ppy/osu/commit/9d62eea8b56a0eb013e8116db0d38424b500876f)
  
  As seen in
  https://github.com/ppy/osu/actions/runs/23888446400#user-content-r1s1.
  
  Speculative. Banking on the fact that it takes time to load the sprite
  texts.
  
  ## [Remove all tests from `TestSceneWikiMarkdownContainer` dependent on
  existence of wiki on
  dev](https://github.com/ppy/osu/commit/8f319f91c98510b84e915bb61ec147618a584ce1)
  
  Deletes a flaky as seen in
  https://github.com/ppy/osu/actions/runs/23878899702#user-content-r2s2.
  
  The year is 2026 and LLM scrapers hammer [the
  entire](https://sourcehut.org/blog/2025-04-15-you-cannot-have-our-users-data/)
  [internet](https://blog.metabrainz.org/2025/12/11/we-cant-have-nice-things-because-of-ai-scrapers/)
  all over to scrape whatever ounce of Human Content there is left to feed
  the Moloch so that it can regurgitate it back in the form of The Most
  Average Speech You've Ever Read. We are not immune to this, and as such
  the LLM homunculi have hit the dev.ppy.sh wiki instance enough times for
  it to just completely [be banished to the
  blagole](https://www.youtube.com/watch?v=AfA_2Ku1aJY).
  
  Which means I get to freely delete flaky tests that should never have
  been running as part of CI because they're completely useless now and
  it's not like we're ever turning them back on again.
  
  ## [Use equality check in
  `TestSceneBeatmapCarouselScrolling.TestScrollPositionMaintainedOnRemove_SecondSelected`
  that's less sensitive to floating
  point](https://github.com/ppy/osu/commit/6f2a1de58f7e816830c430e89117ee4c594f529d)
  
  As seen in
  https://github.com/ppy/osu/actions/runs/23826420794#user-content-r0s1.

  Bartłomiej Dach · 2026-04-04 00:17:31 +09:00
  090fe44f86
• Fix song select dropdowns not being navigable with keyboard (#37158)

  - Closes https://github.com/ppy/osu/issues/31684
  
  Uses the global action variant of up and down for dropdown menus. It is
  already used in multiplayer/playlist room/beatmap navigation and
  gameplay menu overlays.
  
  Comment wording is derived from:
  
  https://github.com/ppy/osu/blob/fc817627e56b7db1c46e2f00fc9a3bd14af51211/osu.Game/Graphics/UserInterface/FocusedTextBox.cs#L73-L74

  Joseph Madamba · 2026-04-01 20:59:47 +09:00
  556be552c3
• Remove duplicated CollectionDropdown class (#37020)

  Arguably we should just nuke the now playing overlay playlist window
  completely as the functionality was gimped when search was removed. Now
  it doesn't seek to the currently playing track, nor play sequentially,
  nor do anything useful.
  
  ---------
  
  Co-authored-by: Bartłomiej Dach <dach.bartlomiej@gmail.com>

  Dean Herbert · 2026-03-18 10:42:01 +01:00
  f84caf1182
• Fix a few flaky tests

  See failures on https://github.com/ppy/osu/runs/66073876886#r3s1.

  Dean Herbert · 2026-03-07 23:19:13 +09:00
  14ae2312ff
• Fix crash when clicking panels during recycling (#36457)

  Closes #36246
  
  This PR overrides `HandlePositionalInput` to reject all positional input
  when `Item` is null.
  
  ---------
  
  Co-authored-by: Dean Herbert <pe@ppy.sh>

  Eeli Ogata · 2026-03-06 09:36:39 +01:00
  e911217980
• Move v2 files to final location

  This contains only renames and namespace updates.

  Dean Herbert · 2026-03-03 16:45:15 +09:00
  c28c64940a
• Add special not regarding collection dropdown
  Dean Herbert · 2026-03-03 16:45:15 +09:00
  b4f40639a1
• Remove all song select v1 files

  This contains no code changes (that would need review), only file
  deletion and extraction where required in a few odd cases.

  Dean Herbert · 2026-03-03 16:45:14 +09:00
  1a8d2855d4
• Move components only used by beatmap set overlay local to namespace
  Dean Herbert · 2026-03-03 16:45:13 +09:00
  ce542b03db
• Fix test scene leaks through RealmRulesetStore/RealmAccess
  Dan Balasescu · 2025-10-09 15:28:15 +09:00
  79c367d208
• Fix intermittent collection dropdown tests
  Dan Balasescu · 2025-09-23 18:23:41 +09:00
  37e661b27e
• Add test coverage
  Bartłomiej Dach · 2025-09-03 12:05:03 +02:00
  15d73ce07e
• Fix test
  Bartłomiej Dach · 2025-07-30 14:06:38 +02:00
  1319a5499b
• Fix tests (they're using LocalisableStrings now)
  Denis Titovets · 2025-07-10 21:58:05 +03:00
  d5fe7e20be
• Fix tests #2
  Denis Titovets · 2025-07-09 18:14:34 +03:00
  7c6f0d1d11
• Update existing tests to work with Song Select v2
  Dean Herbert · 2025-07-01 17:25:38 +09:00
  397b9b3994
• Remove no longer used method
  Dan Balasescu · 2025-06-25 18:31:35 +09:00
  d42fcabcf4
• Fix SHOCKING test

  What the actual heck is that magic number stupidity.
  
  I'm not looking into why the test falls over so badly that it apparently
  dies on some main menu logic just because assertions fail, because the
  main menu logic is for hold-to-song-select-v2 and is thus temporary.

  Bartłomiej Dach · 2025-06-13 10:34:14 +02:00
  f4bf2ae7a5
• Fix flaky collections test
  Dan Balasescu · 2025-06-09 21:53:19 +09:00
  83ce468c57
• Add debounce to difficulty calculation operations in churn scenarios
  Dean Herbert · 2025-05-19 16:16:33 +09:00
  192a1642d4
• Fix test in old song select not working on Apple platforms
  Salman Alshamrani · 2025-05-03 05:38:00 +03:00
  109b29c1da
• Update existing test cases
  Salman Alshamrani · 2025-04-30 02:24:37 +03:00
  dcbb7209df
• Adjust namespaces and naming
  Dean Herbert · 2025-04-22 17:03:28 +09:00
  ec854f7b7f
• Move collection dropdown test coverage to isolated test scene
  Salman Alshamrani · 2025-04-17 09:09:46 -04:00
  a6a8e2a44f
• Fix garbage data in test case
  Bartłomiej Dach · 2025-04-17 11:37:28 +02:00
  5f4afe156f
• Make some test methods static for future reuse
  Dean Herbert · 2025-04-16 19:41:49 +09:00
  2ac1b89037
• Rename song select v2 classes and namespaces

  This aims to bring some conformity to naming to make it easier to
  understand component structure for new components.
  
  Renames are pulled out of the song select v2 changes and are more
  relevant there due to many new classes being added.
  
  - `V2` suffix is dropped, with v2 components being moved to a relevant V2 namespace.
  - Related classes have a prefix of the area they are used.
  - Experimenting with using partial/nested classes in the song select v2 implementation.
    Not committing to this yet but want to see how it plays out.
  - Moved base carousel components to a generic namespace to avoid confusion with actual beatmap carousel implementation.

  Dean Herbert · 2025-04-16 18:34:53 +09:00
  0aff50fbf5
• Add failing test
  Bartłomiej Dach · 2025-04-16 10:27:05 +02:00
  64b9d4642a
• Merge pull request #32494 from bdach/yet-another-leaderboard-rewrite-attempt

  Extract leaderboard fetch logic from song select beatmap leaderboard drawable

  Dean Herbert · 2025-04-16 16:06:07 +09:00
  57a81658f8
• Reduce bleed in tests to allow them to pass
  Bartłomiej Dach · 2025-04-11 10:13:49 +02:00
  9911e0819e
• Fix carousel handling of bleed areas

  The idea of specifying "bleed" is to make the carousel aware of its
  vertical display area. The top bleed is under the filter control; bottom
  beneath the toolbar. At the end of the day, the point of panel X offset
  incursion, and the scroll target for current selection, should be at the
  centre of the screen.
  
  The fixes match code which already exists in the previous
  implementation. Basically, without incorporating `BleedTop` into
  calculations a second time, the centre position would not match
  expectations (of being the centre including bleed).

  Dean Herbert · 2025-04-11 00:05:51 +09:00
  f9112066d3
• Merge branch 'master' into yet-another-leaderboard-rewrite-attempt
  Bartłomiej Dach · 2025-04-08 12:54:27 +02:00
  567c7cb798
• Split tracker into per-system implementations
  Dan Balasescu · 2025-03-25 14:56:12 +09:00
  53ece9396a
• Extract leaderboard fetch logic from song select beatmap leaderboard drawable

  RFC. Another attempt at this.
  
  - Supersedes https://github.com/ppy/osu/pull/31881
  - Supersedes / closes https://github.com/ppy/osu/pull/31355
  - Closes https://github.com/ppy/osu/issues/29861
  
  This is a weird diff because I am feeling rather boxed in by all the
  constraints, namely that:
  
  - Leaderboard state should be global state
  - But the global state is essentially managed by song select and namely
    `BeatmapLeaderboard` itself. That's because trying to e.g. not have
    `BeatmapLeaderboard` pass the beatmap and the ruleset to the global
    leaderboard manager is worse, as it essentially introduces two
    parallel paths of execution that need to be somehow merged into one
    (as in I'd have to somehow sync `LeaderboardManager` responding to
    beatmap/ruleset changes with `BeatmapLeaderboard` which is inheritance
    hell)
  - Also local leaderboard fetching is data-push (as in the scores can
    change under the leaderboard manager), and online leaderboard fetching
    is data-pull (as in the scores do not change unless the leaderboard
    manager does something). Also online leaderboard fetching can fail.
    Which is why I need to still have the weird setup wherein there's a
    `FetchWithCriteriaAsync()` (because I need to be able to respond to
    online requests taking time, or failing), but also the
    `BeatmapLeaderboard` only uses the public `Scores` bindable to
    actually read the scores (because it needs to respond to new local
    scores arriving).
  - Another thing to think about here is what happens when a retrieval
    fails because e.g. the user requested friend leaderboards without
    having supporter. With how this diff is written, that special
    condition is handled to `BeatmapLeaderboard`, and
    `LeaderboardManager`'s state will remain as whatever it was before
    that scope change was requested, which may be considered good or it
    may not (I imagine it's better to show scores in gameplay than not in
    this case, but maybe I'm wrong?)

  Bartłomiej Dach · 2025-03-21 12:06:16 +01:00
  0209129618
• Add support for team beatmap leaderboards
  Dean Herbert · 2025-03-13 16:28:20 +09:00
  b0cf5e8bff
• Add visual test coverage
  Bartłomiej Dach · 2025-03-07 08:51:55 +01:00
  33dccfcec8
• Try to fix test
  Bartłomiej Dach · 2025-03-05 12:32:58 +01:00
  7975c301a8
• Add failing test coverage
  Bartłomiej Dach · 2025-03-05 09:21:47 +01:00
  abc4955e81
• User inheritance to avoid Piece structural nightmare
  Dean Herbert · 2025-02-18 16:52:18 +09:00
  88ec204d26
• Merge branch 'master' into carousel-v2-implement-designs
  Dean Herbert · 2025-02-18 15:49:34 +09:00
  40cae4f0a5
• Rename classes for better discoverability / grouping
  Dean Herbert · 2025-02-12 19:34:42 +09:00
  554884710c
• Fix failing test
  Dean Herbert · 2025-02-07 21:46:00 +09:00
  5ace8e911b
• Merge branch 'master' into carousel-v2-implement-designs
  Dean Herbert · 2025-02-07 19:37:06 +09:00
  d8f3dbf988
• Merge pull request #31823 from peppy/carousel-v2-optimisation-pass

  Refactor grouping to be much more efficient

  Bartłomiej Dach · 2025-02-07 10:44:53 +01:00
  a068d88b84
• Merge pull request #31824 from peppy/carousel-v2-group-click-transfer-keyboard

  Always transfer keyboard selection on activation

  Bartłomiej Dach · 2025-02-07 10:12:10 +01:00
  ae179c8feb
• Reduce number of beatmaps added in selection test

  This is because with the new keyboard selection logic, adding too many
  can cause the re-added selection to be off-screen in the headless test
  setup.

  Dean Herbert · 2025-02-07 17:02:53 +09:00
  cef9d2eac5
• Always transfer keyboard selection on activation
  Dean Herbert · 2025-02-07 17:02:53 +09:00
  c935c3154b
• Use random generation in carousel stress test
  Dean Herbert · 2025-02-07 16:09:35 +09:00
  75ef6f6a0e
• Remove stray [Solo]s
  Dean Herbert · 2025-02-07 15:30:31 +09:00
  aa329f397e