5730 Commits

• Fix subscription leak from composer reloads (#36828)

  The reproduction scenario for the subscription leak is as follows:
  
  1. Switch to a scrolling ruleset (anything but osu! from the standard
  ones).
  2. Select a beatmap to edit.
  3. Load the composer.
  4. Go to timing tab.
  5. Change a timing point.
  6. Go back to the composer.
  
  At this point, `EditorChangeHandler.OnStateChange` will have multiple of
  the same delegate in the invocation list.
  
  <img width="691" height="311" alt="Screenshot 2026-03-05 at 11 15 55"
  src="https://github.com/user-attachments/assets/57788341-9573-48f1-b360-f21036891081"
  />
  
  That in turn is caused by the fact that changing a timing point *does*
  incur a full reload of the composer via the following flow:
  
  
  https://github.com/ppy/osu/blob/15b6e28ebe888b1a87574891be1a0db3b04093b7/osu.Game/Rulesets/Edit/ScrollingHitObjectComposer.cs#L145
  https://github.com/ppy/osu/blob/64a29313a852d50095ae4b7ea8f22fde23aa634f/osu.Game/Screens/Edit/Editor.cs#L1137-L1145
  
  This flow is my "fault"; see https://github.com/ppy/osu/pull/28444. The
  reason why a full composer reload is used is not clear to my
  recollection at this time, but it is likely because it's just the least
  likely to fail. A smarter solution that wouldn't require a full reload
  would also entail checking that there exists a safe insertion point that
  allows replacing timing points in a way that will reflect everywhere it
  must. Including all of the `IScrollingAlgorithm` machinery and such.
  
  In general it is not uncommon in the codebase to not bother to clean up
  some event callbacks if it is implicitly or explicitly guaranteed that
  both objects bound by the callback will always get disposed in tandem at
  the same time. This *was* true with this particular flow to a point,
  which was until that full composer reload was implemented.
  
  <details>
  <summary>To address the elephant in the room</summary>
  
  Someone will inevitably notice https://github.com/ppy/osu/pull/36824
  which was a clanked pull request pointing out this leak. And then
  someone will inevitably call this "AI discrimination"! *Gasp!*
  
  So first of all, let me stop you right there. Yes, as far as I am
  _personally_ concerned, it is "AI discrimination". I invoke the full
  force of the Butlerian Jihad.
  
  The clank army's goal is to eradicate my job and make me work in an
  Amazon warehouse instead. Or, if not that, at least my job is to be rid
  of all remnants of fun I still get from it and for me to be reduced to
  that one guy from the meme "i guess we're doin circles now". You know
  the one.
  
  I resent this. You attack me directly. I do not perceive the need to
  meet you halfway or be civil.
  
  That said, I have too much respect for the users of this software to
  leave reports of potentially real issues unchecked. So I did check, and
  it was real. And you know what? Good job to the clanker. It did what it
  was designed to do: it parsed a code file, recognised a hole in a
  pattern it was designed to recognise, and invoked forms of language
  given to it to communicate this to the meatbag that opened that PR.
  
  And here's the thing: my primary issue is with that meatbag that opened
  that PR. That meatbag served no functional purpose in any of this. The
  meatbag took a hose that spews 90% water and 10% raw sewage at random
  intervals and pointed it at my house directly, claiming that they just
  want to clean it. At no point did the meatbag appear to have the common
  decency to pull out a container, pour some magic liquid out, check if
  there's sewage in it, and filter it out if there is any. But no, that
  would take *effort* and *thought*, would it not? The *effort* and
  *thought* that is required of *me* to *review* the clanker's work?
  
  The PR had no reproduction scenario, and had testing checkboxes that
  were presumably meant for *me* to check off. Why is it *my* job to
  figure all of this out rather than the submitter meatbag's?
  
  I do *not* have obligations towards spew-hose-pointing meatbags. Point
  that hose at your own backyard at your peril.
  
  If you *actually manage* to get the clanker to filter out *all* of the
  spew without fail itself, my only win condition is gone. But it is not
  yet that time. So at least have the decency to check for the spew
  yourself, rather than telling the clanker to put checkboxes in the PR
  descriptions telling *me* to check for it.
  </details>

  Bartłomiej Dach · 2026-03-05 22:37:50 +09:00
  43897457fb
• Improve usability of sample bank toggles (#36753)

  - [x] Depends on https://github.com/ppy/osu/pull/36741 for merge
  conflict avoidance
  
  RFC, cc @OliBomby
  
  ## [Adjust behaviour of automatic bank assignment during
  placement](https://github.com/ppy/osu/commit/547f55e9b3ded668fe6e1c8865a2d625e64a2f45)
  
  Diatribe time!
  
  This is fallout of the discussion about auto bank in
  https://github.com/ppy/osu/issues/36705.
  
  Auto bank in lazer as written before this commit is confused. On stable,
  auto bank is closer to "no bank", as in "go look up the current sample
  timing point, get the bank of that, and use that". lazer has no timing
  points anymore, but people still want auto bank. So what do?
  
  Auto bank for normal samples is somewhat sane still. It only works
  during placement, and will just copy the normal bank of the previous
  object - if one exists. That said, one *might not* exist, but the
  resulting object will still have its normal sample created with
  `editorAutoBank: true`. That is largely cosmetic and without
  consequences, but this commit fixes that.
  
  Auto bank for *addition* samples, however... Hoo boy.
  
  - For placed objects, auto bank means "take the normal sample, read its
  bank, and use that". Simple enough, right?
  - Hoooooowever. During placement, auto bank before this commit used to
  mean "look at the *previous object*, check if it has an addition sound
  and then use its bank, if not use *the previous object's* normal sample
  and then use its bank" which is a completely different thing with its
  own implications. Like, say, what happens if the previous object uses
  the auto addition bank too? What should be copied over? Should it be the
  notion of "auto bank" in that the addition bank should match the normal
  bank, or should it be the literal bank that the previous object is
  using?
  
  This change attempts to define this unambiguously. "Auto additions bank"
  means "the same bank as the normal bank of this object", full stop.
  
  ## [Do not touch sample toggle state if there are no selected
  objects](https://github.com/ppy/osu/commit/052cde5987e48800ec68ab2528c7e0ce3140e6e0)
  
  Fixes issue described in
  https://github.com/ppy/osu/issues/36705#issuecomment-3953917163 wherein
  opening a sample popover will disable addition bank toggles and toggle
  off all addition samples.
  
  ---------
  
  Co-authored-by: Dean Herbert <pe@ppy.sh>

  Bartłomiej Dach · 2026-02-26 15:18:29 +09:00
  7e2771c3f0
• Fix reversing straight perfect curve sliders positioning them weirdly (#36698)
  Arthur Araujo · 2026-02-20 16:57:42 +01:00
  c144cf188a
• Adjust nightcore sample playback to not stop randomly and be closer to stable (#36702)

  - Closes https://github.com/ppy/osu/issues/30293
  - Fixes https://osu.ppy.sh/community/forums/topics/2179339?n=1
  
  Aside from fixing the off-by-one error that I mentioned in
  https://github.com/ppy/osu/issues/30293#issuecomment-2413801663, this
  also:
  
  - Brings back the behaviour wherein if timing points are arranged very
  weird and nightcore would play e.g. two first beats in a timing point
  back-to-back, the second timing point is silent.
  - Brings back the behaviour wherein the finish sample only plays if
  `OmitFirstBarLine` on the timing point is disabled.
  
  However:
  
  - This does not bring back the behaviour wherein hat samples only play
  if the slider tick rate is even because that only kind of makes sense in
  common time, and if common time is mixed with waltz time or other time
  signatures, it just gets weird.
  - Also stable has zero attempt for compensating for waltz time anyway,
  lazer's behaviour is bespoke, so that is not going to match any way you
  cut it.
  
  My testing procedure essentially consisted of getting stable to log when
  it was playing nightcore samples and cross-checking the first 30sec or
  so of https://osu.ppy.sh/beatmapsets/534385#osu/1131956 (check out the
  timing of that beatmap, for something ranked it is DEEPLY messed up).
  
  I guess I can add test cases if deemed required but I already wasted
  much more time than I would have liked here...

  Bartłomiej Dach · 2026-02-20 12:53:12 +09:00
  1e8b9a7639
• Fix extra lives in Easy mod potentially getting reapplied during gameplay (#36678)

  Closes https://github.com/ppy/osu/issues/36676.

  Bartłomiej Dach · 2026-02-16 19:02:59 +09:00
  810edebe87
• Merge pull request #36639 from bdach/actually-all-valid-hit-results

  Refactor hit result methods on `Ruleset`

  Dean Herbert · 2026-02-10 17:07:04 +09:00
  c59e5bf335
• Perform extra checks when loading rulesets (#36641)
  Bartłomiej Dach · 2026-02-10 13:34:43 +09:00
  655d725f0c
• Adjust contract of Ruleset.GetValidHitResults() to always enumerate all results used
  Bartłomiej Dach · 2026-02-09 12:48:53 +01:00
  15c49a729e
• Publicise Ruleset.GetValidHitResults()
  Bartłomiej Dach · 2026-02-09 12:00:30 +01:00
  3678ea0d0f
• Rename Ruleset.GetHitResults{ -> ForDisplay}()
  Bartłomiej Dach · 2026-02-09 12:00:15 +01:00
  dbce8886f8
• Improve adjusting mods settings values with keyboard (#36090)

  - closes https://github.com/ppy/osu/issues/36016
  
  Co-authored-by: Dean Herbert <pe@ppy.sh>

  Denis Titovets · 2026-02-09 11:35:01 +09:00
  60d98f0afd
• Standardise ruleset error logging to always include exception in logs

  Supersedes and closes https://github.com/ppy/osu/pull/36420.

  Dean Herbert · 2026-01-22 15:48:51 +09:00
  fdd13cd7e1
• Specialise mod setting hover text in song select scoreboard (#36391)

  - resolves #35992
  
  **Changes**
  
  Introduced GetSettingTooltipText, which returns the value of the
  respective setting. The function can be overwritten, allowing to
  implement specific behavior. This allows you to define that a value of 0
  for the “Muted” mod will instead display “always muted.”
  In the same step, I also adjusted the "NoScope" mod, where we had the
  same problem.
  Wherever we want to implement specialization like this, we can simply
  overwrite GetSettingTooltipText to customize the behavior for the mod
  settings of the mod.
  
  **Result**
  <img width="2560" height="1440" alt="2026-01-19-122307_hyprshot"
  src="https://github.com/user-attachments/assets/20e3aa9a-aa6f-4284-9cf1-3092f52c021d"
  />
  <img width="2560" height="1440" alt="2026-01-19-122324_hyprshot"
  src="https://github.com/user-attachments/assets/6f3fac5b-2a5d-4dd9-a56c-4b09c02cbcca"
  />
  <img width="2560" height="1440" alt="2026-01-19-155307_hyprshot"
  src="https://github.com/user-attachments/assets/f9ee75d3-c200-4536-9ee9-d20ddbd9fa44"
  />
  
  Signed-off-by: Linus Genz <linuslinuxgenz@gmail.com>

  Linus Genz · 2026-01-20 16:51:23 +09:00
  dfa7ac2082
• Fix toasts showing "no key bound" for operations which can't have keys bound

  Supersedes and closes https://github.com/ppy/osu/pull/35781.
  Closes https://github.com/ppy/osu/issues/36294.

  Dean Herbert · 2026-01-15 17:52:30 +09:00
  b4ba327c1c
• Fix editor crashing on undoing after hit object placement

  Happens when initial timing is undone.
  
  Regressed in
  https://github.com/ppy/osu/commit/12170df80add87e8f292545f67566f8d25465efc.
  Closes https://github.com/ppy/osu/issues/36228.

  Dean Herbert · 2026-01-05 15:26:14 +09:00
  9916e5bb1c
• Merge branch 'master' into custom-sample-set-selection
  Dean Herbert · 2025-12-17 19:30:24 +09:00
  61bf68f336
• Add basic colour indication as to when placements are valid

  Unsure about this one, but I find the preceding commit to be very
  lacking in explaining to the user why the editor don't work. Shining
  some things red may help aid understanding.

  Bartłomiej Dach · 2025-12-04 14:37:56 +01:00
  c6cc92315c
• Disallow placing hit objects before first timing point

  Because they can break stable. See
  https://github.com/ppy/osu/issues/31591#issuecomment-3575270120 for
  detailed rationale.

  Bartłomiej Dach · 2025-12-04 14:15:52 +01:00
  12170df80a
• Merge pull request #35757 from peppy/settings-wank

  Adjust settings buttons and general section to feel better

  Bartłomiej Dach · 2025-11-25 08:46:00 +01:00
  9e3486d4e6
• Reduce abstractions of rounded button
  Dean Herbert · 2025-11-25 00:55:42 +09:00
  b0762fc8ec
• Adjust element spacing in editor toolboxes
  Bartłomiej Dach · 2025-11-19 13:38:34 +01:00
  397041099e
• Use new sliders-with-text-input in editor toolboxes

  Addresses https://github.com/ppy/osu/discussions/35732.
  
  And yes, I renamed "perfect curve threshold" to "bias" so that the text
  can fit. Sue me.

  Bartłomiej Dach · 2025-11-19 13:22:20 +01:00
  4b59a4657f
• Fix several issues with incorrect sample playback (#35685)

  * Add failing test coverage for layered hit samples not playing in mania when beatmap is converted
  
  Adding the `osu.Game.Rulesets.Osu` reference to the mania test project
  is required so that `HitObjectSampleTest` base logic doesn't die on
  
  https://github.com/ppy/osu/blob/f0aeeeea966f06add12cf2bca3dd48dac8573e82/osu.Game/Tests/Beatmaps/HitObjectSampleTest.cs#L88-L91
  
  * Fix layered hit sounds not playing on converted beatmaps in mania
  
  Compare
  https://github.com/peppy/osu-stable-reference/blob/f9e58b4864a10f801393199e7652b2192c7342c3/osu!/GameplayElements/HitObjects/HitObject.cs#L476-L477.
  
  In case of converted beatmaps, the last condition there
  (`BeatmapManager.Current.PlayMode != PlayModes.OsuMania`) fails,
  and thus layered hitsounds are allowed to play.
  
  * Add failing test coverage for mania beatmap conversion assigning wrong samples to spinners
  
  * Fix mania beatmap conversion assigning wrong samples to spinners
  
  A spinner is never `IHasRepeats`. It was a dead condition, leading to
  the hitobject generating fallback `NodeSamples`, which in particular
  feature a silent tail which stable doesn't do.
  
  Noticeably, stable also appears to force the head of the generated hold
  note to have no addition sounds:
  
  https://github.com/peppy/osu-stable-reference/blob/f9e58b4864a10f801393199e7652b2192c7342c3/osu!/GameplayElements/HitObjects/Mania/SpinnerMania.cs#L86-L89
  
  * Add failing test coverage for file hit sample not falling back to plain samples if file missing
  
  * Allow `FileHitSampleInfo` to fall back to standard samples if the file is not found (or not allowed to be looked up)
  
  I'm honestly not 100% as to how closely this matches stable because I
  reached the point wherein I'd rather not look at stable code anymore, so
  as long as this passes tests I'm fine to wait for someone else to report
  new breakage.
  
  * Use alternative workaround for lack of osu! ruleset assembly in mania test project
  
  * Fix encode stability test failures

  Bartłomiej Dach · 2025-11-15 16:19:08 +09:00
  bd4ed49c06
• Merge pull request #35249 from dnfd1/mania-difficultychange-limits

  Adjust extended OD limits for mania difficulty change mod to reflect HR and EZ values

  Dean Herbert · 2025-10-28 15:05:26 +09:00
  98829bf857
• Inherit sample set info from previous object on placement
  Bartłomiej Dach · 2025-10-24 12:09:30 +02:00
  511921c1f7
• Adjust gameplay sample models to support custom sample sets

  This is a set of model changes which is supposed to facilitate support
  for custom sample sets to the beatmap editor that is on par with stable.
  
  It is the minimal set of changes. Because of this, it can probably be
  considered "ugly" or however else you want to put it - but before you
  say that, I want to try and pre-empt that criticism by explaining where
  the problems lie.
  
  Problem #1: duality in sample models
  ---
  
  There is currently a weird duality of what a `HitObject`'s samples will
  be.
  
  - If an object has just been placed in the editor, and not saved /
    decoded yet, it will use `HitSampleInfo`.
  
  - If an object has already been encoded to the beatmap at least once, it
    will use `ConvertHitObjectParser.LegacyHitSampleInfo`.
  
  As long as that state of affairs remains, `HitSampleInfo` must be able
  to represent anything that `LegacyHitSampleInfo` can, if feature parity
  is to be achieved.
  
  Problem 2: The 0 & 1 sample banks
  ---
  
  Custom sample banks of 2 and above are a pretty clean affair. They map to
  a suffix on the sample filename, and said samples are allowed to be
  looked up from the beatmap skin. `Suffix` already exists in
  `HitSampleInfo`.
  
  However, the 1 custom sample bank is evil. It uses *non-suffixed*
  samples, *allows lookups from the beatmap skins*, contrary to no bank /
  bank 0, which *also* uses non-suffixed samples, but *doesn't* allow them
  to be looked up from the beatmap skin.
  
  This is why `HitSampleInfo.UseBeatmapSamples` has been called to
  existence - without it there is no way to represent the ability of using
  or not using the beatmap skin assets.
  
  As has been stated previously in discussions about this feature, it's
  both a *mapping* and a *skinning* concern.
  
  There are many things you could do about either of these problems, but I
  am pretty sure tackling either one is going to take *many* more lines of
  code than this commit does. Which is why this is the starting point of
  negotiation.

  Bartłomiej Dach · 2025-10-22 14:58:08 +02:00
  c361e6d3c2
• Update framework
  Dean Herbert · 2025-10-21 17:24:33 +09:00
  0b84916c3e
• Allow OD to be overrided
  dnfd1 · 2025-10-08 21:27:49 -07:00
  348713d83d
• Disallow zero-length sliders from specifying a non-zero number of repeats (#35220)
  Bartłomiej Dach · 2025-10-08 16:01:04 +09:00
  1a522a19f1
• Re-remove duplicate error functions
  Dean Herbert · 2025-10-07 16:30:40 +09:00
  743a94bd22
• Merge branch 'master' into pp-dev
  Dean Herbert · 2025-10-07 16:11:54 +09:00
  bda5737135
• Merge pull request #34934 from Valerus9/fixtestdouble

  Fix TestDouble failing on systems where the decimal separator isn't a dot

  Bartłomiej Dach · 2025-09-29 16:46:32 +09:00
  5336bcd71d
• Merge pull request #35120 from LumpBloom7/fix-toolbox-tooltip

  Fix composition tool tooltip not changing text when enabled

  Bartłomiej Dach · 2025-09-29 16:04:15 +09:00
  ba3d7392bf
• Merge branch 'master' into fixtestdouble
  Bartłomiej Dach · 2025-09-29 08:43:49 +02:00
  9257e62bac
• Use slightly nicer way of invarianting rate adjust setting value
  Bartłomiej Dach · 2025-09-29 08:39:01 +02:00
  70f683e7fa
• Handle disabled text in composition tool button
  Derrick Timmermans · 2025-09-26 09:28:04 +02:00
  c66cc5ebb1
• Adjust matchmaking naming, namespaces, xmldoc (#35123)

  * Adjust matchmaking naming, namespaces, xmldoc
  
  * Change partial filenames to use `.` instead of `_` separator

  Dean Herbert · 2025-09-26 12:17:28 +09:00
  e1ba1b45b0
• Fix composition tool tooltip not changing text when enabled
  Derrick Timmermans · 2025-09-25 10:44:05 +02:00
  df795da070
• Remove databasing of MechanicalDifficulty and ReadingDifficulty attributes (#35028)

  * Remove databasing of `MechanicalDifficulty` and `ReadingDifficulty` attributes
  
  * Update attribute IDs

  James Wilson · 2025-09-15 13:46:00 +03:00
  2749184c38
• Fix testtooltip failure
  Valerus9 · 2025-09-13 06:44:29 +02:00
  e34b0659da
• Merge branch 'master' into pp-dev
  Dan Balasescu · 2025-09-07 19:22:08 +09:00
  ffab8342f2
• Make OsuGame dependency optional for sanity
  Dean Herbert · 2025-09-02 21:40:52 +09:00
  a1105ba16f
• Change bass workaround fix to use game clock intead of another-audio-clock

  paper trail:
  https://github.com/ppy/osu/pull/34890#issuecomment-3244549790

  Dean Herbert · 2025-09-02 21:39:12 +09:00
  bee6c32b83
• Fix gameplay freezing on stutter frames / long load times

  Closes https://github.com/ppy/osu/issues/34732.
  
  May hotfix for this one.

  Dean Herbert · 2025-09-02 17:37:33 +09:00
  677beb4251
• Revert "Adjust bass invalid data threshold"

  This reverts commit ddce11fbc8.

  Dean Herbert · 2025-09-02 17:36:08 +09:00
  a8ef57ad0a
• Merge pull request #34830 from minetoblend/fix/hitobject-lifetime-delay

  Fix hitobject drawables becoming visible 1 frame too late

  Dan Balasescu · 2025-08-29 18:38:24 +09:00
  d5575b4037
• Fix obviously incorrect conditional

  Co-authored-by: Bartłomiej Dach <dach.bartlomiej@gmail.com>

  Dean Herbert · 2025-08-29 16:01:49 +09:00
  9e77a5b050
• Fix some mods showing tooltips when settings are default
  Dean Herbert · 2025-08-29 14:08:18 +09:00
  e83f3d5e77
• Return early to avoid creating mod description strings unnecessarily
  Dean Herbert · 2025-08-29 14:08:06 +09:00
  f2f5cf19a2
• Fix operations in PooledDrawableWithLifetimeContainer.CheckChildrenLife being in wrong order

  Previously CompositeDrawable.CheckChildrenLife() would be run before lifetimeManager.Update() which lead to the new drawables being inserted into the container but not being made alive immediately, leading to the drawable not becoming visibile until the next update loop.

  marvin · 2025-08-28 23:36:29 +02:00
  bb9f9e4d35