Fix: ListView scroll when using SwipeControl

[ajpinedam]

GitHub Issue: closes https://github.com/unoplatform/kahua-private/issues/477

PR Type:

🐞 Bugfix

What changed? 🚀

CancelDirectManipulations cancelled every ancestor direct manipulation regardless of axis, so a SwipeControl (TranslateX) inside a ListView killed the parent ScrollContentPresenter's (TranslateY) DM and vertical scrolling died

Fix: Made ConflictsWithRequester axis-aware via a new IDirectManipulationHandler.GetCurrentlyAcceptedModes() peek — a pure-translation requester now only cancels ancestors that share an axis, while multi-pointer (Scale/Rotate) requesters and System/None-mode requesters keep the legacy "cancel everything" behavior so pinch-in-ScrollViewer fix is preserved

related to this change: #23135

PR Checklist ✅

• 🧪 Added Runtime tests, UI tests, or a manual test sample (for bug fixes / features, if applicable)
• 📚 Docs have been added/updated following the documentation template (for bug fixes / features)
• 🖼️ Validated PR Screenshots Compare Test Run results.
• ❗ Contains NO breaking changes
• 👀 Reviewed 2 other open pull requests (optional but appreciated!)

[Copilot]

Pull request overview

This PR fixes a managed-pointer direct manipulation cancellation behavior where a descendant’s CancelDirectManipulations request could incorrectly cancel orthogonal-axis ancestor manipulations (e.g., SwipeControl horizontal manipulation inside a vertically scrolling ListView), breaking vertical scrolling. It introduces an axis-aware conflict check by peeking the ancestor handlers’ currently accepted manipulation modes and adds a regression runtime test.

Changes:

• Make ancestor direct-manipulation cancellation axis-aware via a new handler “accepted modes” preview.
• Extend IDirectManipulationHandler with GetCurrentlyAcceptedModes() (defaulting to All) and implement it for ScrollContentPresenter.
• Add a runtime regression test for vertical scrolling with SwipeControl items inside a ListView.

Reviewed changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 2 comments.

File	Description
src/Uno.UI/UI/Xaml/Internal/InputManager.Pointers.ManagedDirectManip.cs	Adds axis-aware ancestor DM cancellation via ConflictsWithRequester and handler mode peeking.
src/Uno.UI/UI/Xaml/Internal/IDirectManipulationHandler.cs	Adds GetCurrentlyAcceptedModes() with a default implementation to preserve legacy behavior.
src/Uno.UI/UI/Xaml/Controls/ScrollContentPresenter/ScrollContentPresenter.Managed.cs	Implements GetCurrentlyAcceptedModes() by reusing the accepted-mode computation logic.
src/Uno.UI.RuntimeTests/Tests/Uno_UI_Xaml_Core/Given_InputManager.cs	Adds regression test ensuring vertical scroll still works with SwipeControl items.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[unodevops]

🤖 Your WebAssembly Skia Sample App stage site is ready! Visit it here: https://unowasmprstaging.z20.web.core.windows.net/pr-23249/wasm-skia-net9/index.html

[unodevops]

⚠️⚠️ The build 212018 has failed on Uno.UI - CI.

[ajpinedam]

I see test When_DirectManipulationDisabled(TranslateX) is failing, when investigating I got this:

The test is encoding a contract that directly contradicts the premise of the axis‑aware fix on this branch. Here's the mechanics step by step:

What the test asserts (Given_InputManager.cs:1402-1445, introduced Aug 2025 by commit d72aee1 "feat: Add support for ManipulationMode=None"):

• Outer ScrollViewer (200×200, vertical scroll enabled, content 800×800).
• Inner Border with ManipulationMode = TranslateX and a -8000px vertical drag.
• Expectation: sv.VerticalOffset.Should().Be(0) — i.e. setting any non-System ManipulationMode on the descendant must disable every ancestor direct manipulation, regardless of axis. The test name itself — When_DirectManipulationDisabled — codifies that "non-System mode on descendant ⇒ ancestor DMs are off" is a global invariant.

Why the new axis logic produces 620px instead of 0:

1. During press bubble, Border.OnPointerDown (UIElement.Pointers.cs:1199) calls CancelDirectManipulations() because ManipulationMode != System.
2. Walks _directManipulations, finds the SV's SCP DM (ancestor; predicate matches).
3. requesterModes = TranslateX, so requesterAxes = TranslateX, requesterIsMultiPointer = false. The "cancel everything" early-returns don't fire.
4. SCP's ComputeAcceptedManipulationModes() (just confirmed at ScrollContentPresenter.Managed.cs:374-387) returns TranslateY only (200‑wide viewport, no horizontal scroll).
5. (TranslateX) & (TranslateY) == 0 → ConflictsWithRequester returns false → SCP DM survives.
6. Vertical drag flows to SCP, SV scrolls 620px → Be(0) fails with found 620.0.

cc: @dr1rrb @agneszitte

[dr1rrb]

Not sure about the exact interaction result we want (i.e. we need to validate on WinUI), but looks good to me

[dr1rrb]

I see test When_DirectManipulationDisabled(TranslateX) is failing, when investigating I got this:

The test is encoding a contract that directly contradicts the premise of the axis‑aware fix on this branch. Here's the mechanics step by step:

What the test asserts (Given_InputManager.cs:1402-1445, introduced Aug 2025 by commit d72aee1 "feat: Add support for ManipulationMode=None"):

* Outer ScrollViewer (200×200, vertical scroll enabled, content 800×800).

* Inner Border with ManipulationMode = TranslateX and a -8000px vertical drag.

* Expectation: sv.VerticalOffset.Should().Be(0) — i.e. setting any non-System ManipulationMode on the descendant must disable every ancestor direct manipulation, regardless of axis. The test name itself — When_DirectManipulationDisabled — codifies that "non-System mode on descendant ⇒ ancestor DMs are off" is a global invariant.

Why the new axis logic produces 620px instead of 0:

1. During press bubble, Border.OnPointerDown (UIElement.Pointers.cs:1199) calls CancelDirectManipulations() because ManipulationMode != System.

2. Walks _directManipulations, finds the SV's SCP DM (ancestor; predicate matches).

3. requesterModes = TranslateX, so requesterAxes = TranslateX, requesterIsMultiPointer = false. The "cancel everything" early-returns don't fire.

4. SCP's ComputeAcceptedManipulationModes() (just confirmed at ScrollContentPresenter.Managed.cs:374-387) returns TranslateY only (200‑wide viewport, no horizontal scroll).

5. (TranslateX) & (TranslateY) == 0 → ConflictsWithRequester returns false → SCP DM survives.

6. Vertical drag flows to SCP, SV scrolls 620px → Be(0) fails with found 620.0.

cc: @dr1rrb @agneszitte

A None should prevent all gestures kinds from any parent, but indeed the tests might be off, a TranslateX shoudl not prevent a parent to handle TranslateY

[ajpinedam]

I see test When_DirectManipulationDisabled(TranslateX) is failing, when investigating I got this:
The test is encoding a contract that directly contradicts the premise of the axis‑aware fix on this branch. Here's the mechanics step by step:
What the test asserts (Given_InputManager.cs:1402-1445, introduced Aug 2025 by commit d72aee1 "feat: Add support for ManipulationMode=None"):

* Outer ScrollViewer (200×200, vertical scroll enabled, content 800×800).

* Inner Border with ManipulationMode = TranslateX and a -8000px vertical drag.

* Expectation: sv.VerticalOffset.Should().Be(0) — i.e. setting any non-System ManipulationMode on the descendant must disable every ancestor direct manipulation, regardless of axis. The test name itself — When_DirectManipulationDisabled — codifies that "non-System mode on descendant ⇒ ancestor DMs are off" is a global invariant.

Why the new axis logic produces 620px instead of 0:

1. During press bubble, Border.OnPointerDown (UIElement.Pointers.cs:1199) calls CancelDirectManipulations() because ManipulationMode != System.

2. Walks _directManipulations, finds the SV's SCP DM (ancestor; predicate matches).

3. requesterModes = TranslateX, so requesterAxes = TranslateX, requesterIsMultiPointer = false. The "cancel everything" early-returns don't fire.

4. SCP's ComputeAcceptedManipulationModes() (just confirmed at ScrollContentPresenter.Managed.cs:374-387) returns TranslateY only (200‑wide viewport, no horizontal scroll).

5. (TranslateX) & (TranslateY) == 0 → ConflictsWithRequester returns false → SCP DM survives.

6. Vertical drag flows to SCP, SV scrolls 620px → Be(0) fails with found 620.0.

cc: @dr1rrb @agneszitte

A None should prevent all gestures kinds from any parent, but indeed the tests might be off, a TranslateX shoudl not prevent a parent to handle TranslateY

Update the TranslateX test so that it allows TranslateY .

[Copilot]

Pull request overview

Copilot reviewed 4 out of 4 changed files in this pull request and generated 3 comments.

Comments suppressed due to low confidence (1)

src/Uno.UI.RuntimeTests/Tests/Uno_UI_Xaml_Core/Given_InputManager.cs:1407

• This test currently expects ManipulationModes.TranslateRailsX to cancel the ancestor ScrollViewer’s vertical scroll (expects VerticalOffset == 0). With the new axis-aware cancellation intent, TranslateRailsX is still an X-only translation request, so a vertical drag should likely continue to scroll the ancestor (similar to the TranslateX row). If ConflictsWithRequester is updated to treat rails as axis-specific, update this expected outcome accordingly.

	[DataRow(ManipulationModes.TranslateX, true)] // Descendant claims X only -> parent's Y-axis scroll must still work.
	[DataRow(ManipulationModes.TranslateY, false)]
	[DataRow(ManipulationModes.TranslateRailsX, false)]
	[DataRow(ManipulationModes.TranslateRailsY, false)]
	[DataRow(ManipulationModes.TranslateInertia, false)]

[unodevops]

🤖 Your WebAssembly Skia Sample App stage site is ready! Visit it here: https://unowasmprstaging.z20.web.core.windows.net/pr-23249/wasm-skia-net9/index.html

[unodevops]

🤖 Your WebAssembly Skia Sample App stage site is ready! Visit it here: https://unowasmprstaging.z20.web.core.windows.net/pr-23249/wasm-skia-net9/index.html