Adjust all selected hold notes if they have the same StartTime and Duration

[64ArthurAraujo]

Addresses #36267

The drag now checks all the selected objects in the blueprint container to see if they have the same StartTime and Duration as the dragged note, and if so, adjust them accordingly.

[bdach]

I dunno what to think of any of this really.

Even aside from the implementation wonkage (of which there is some, because of float rounding errors as usual):

• Why does this only apply in the timeline and not in the composer?
• Aside to the above: I've heard it railed over and over by several mania mappers that the top timeline is "useless" for mania and it should be replaced by literally anything else (density graph, or even nothing). So I don't know how productive it even is to be adding behaviours specific to it like this.
• Another thing is: why do the objects need to have the same start time and duration? I could see potentially it being useful - not even in mania! - to perform the same "relative" operation on multiple notes at once, e.g. shorten or lengthen an object's duration by N beats.

And stable isn't any help here because in stable (a) the top timeline shows nothing other than bar lines and (b) you can't even attempt to drag a hold note end when multiple objects are selected because it will deselect everything other than the object you're trying to drag.

Like I dunno. The issue to me is half-baked in spec to begin with. I would have never acted on it without extensive clarification myself. So I can't find myself realistically ever approving this unless the requirements are actually hashed out properly.

[peppy]

I think the behaviour makes sense, but should maybe also apply on the playfield. I'd also be okay with it only working on the timeline because it somewhat feels like it should work there, at very least.

[bdach]

What's the status for re-review on this? There are new changes but apparently none of them relevant to the preceding discussion? There's also fix commits in there like 49b2ae2 apparently but without test coverage?

Very confused.

[64ArthurAraujo]

I moved the loop that checks for objects in the blueprint to OnDragStart, there i save all the objects that match the condition to objsToAdjust (now as a field of DragArea) and clear when the drag stops. Since the loop was running every single moment the drag was ocurring when you selected 2 hold notes one longer, and then dragged the longer the moment they matched the same duration they would start moving together which was weird to me because the user firstly intended to drag a single object, that is commit 49b2ae2.

I modified a existing test to stop the moment they matched and continue decreasing the note's duration until the smallest size (like it was done previously). There is no new assertions because InputManager.MoveMouseTo moved the cursor instantly causing the OnDrag to be triggered only one time, by stopping the drag at the moment both notes have the same duration i can ensure that the OnDrag event is run 2 times and the other note that was not supposed to be modified is not bound to the one being dragged, the expected result is the same "One is smaller, the other is unchanged", that is the next commit ceabbdd. To be more clear do you think this needs to be a separated test case?

One thing that could be pointed out is that objsToAdjust as a field is not used by anyone else other than mania hold notes, so when adjusting a standard slider the OnDragStart will run a useless foreach and maybe save some objects to the field and do nothing with them. That was the best solution i could find to the "notes matching mid-drag issue"
so i ask for some toughts on that approach.

Also i noticed a forgot to write a test to 8c158b6, my bad ill write as soon as possible

[bdach]

The changes are fine I just don't know whether you're done making changes and I can approach re-review.

Also 8ccbe5b is a dubious commit I would just revert. I don't see any reason to apply DRY in tests.

[64ArthurAraujo]

Yep its done, it is working as described without any bugs, please re-review it.
About 8ccbe5b i personally was not liking the previous version especially doing the cast from HitObject to HoldNote on every assert. But i guess we don't need to think that much about this in tests i ll revert it then.

[bdach]

I guess aside from the fixes I just pushed (f107ca9 695cd19) I don't have any other code concerns.

Notably this still doesn't work on the playfield. Will leave it to @peppy to assess whether that's acceptable from the ux angle.

[peppy]

Another thing is: why do the objects need to have the same start time and duration

I also don't get this requirement. I'll do a UX pass on this and see how things feel.

[peppy]

Is this supposed to work? Or am I missing something?

osu.2026-02-25.at.05.45.00.mp4

[bdach]

It worked when I tested it.

[peppy]

Don't sink more time into this for now, I'll investigate as a lower priority. Still not 100% on the behaviour so.