* Fix VirtualizingStackPanel offset when scrolling to variable-sized items
* Remove the `isScrollIntoView` parameter.
* Tweak now-failing test.
Change the the test to check that the element is outside of the viewport, instead of asserting its exact coordinates - that's the important part.
* Retrigger CI.
Azure Pipelines never queued a build for 3d81ceb9.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0169P8BipcMkn44hvhpQ2UNF
---------
Co-authored-by: grokys <grokys@users.noreply.github.com>
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
* test(VirtualizingStackPanel): cover focused item virtualization
A focused item remains attached so keyboard focus can be preserved after it leaves the realized range. It must nevertheless stop rendering until it is realized again; otherwise an estimated position can surface it in the viewport.
* fix(VirtualizingStackPanel): hide cached focused items
A focused container stays attached after it leaves the realized range so keyboard focus survives. Its estimated position can overlap the viewport when item sizes vary, causing it to render and receive pointer input as a ghost item.\n\nSuppress rendering and hit testing while the container is cached, then restore its original state before re-realization or recycling. This preserves focus and template-provided opacity.
* fix(VirtualizingStackPanel): restore current focused container state
Focused-container suppression previously replayed the opacity and hit-test values observed when virtualization began. That could hide updates from bindings, styles, or application code while the container was retained.\n\nUse disposable animation-priority overrides for the temporary suppression so disposing them reveals the property's current underlying value.
* fix(VirtualizingStackPanel): keep retained focused container out of the viewport
The container of a focused item that leaves the realized range is retained
to preserve keyboard focus, and arranged at a position estimated from the
average realized element size. With differing item sizes that estimate can
place the container so that it overlaps the viewport, where it renders as a
ghost item and intercepts pointer input (#17935).
An unrealized item before the realized range always ends at or before the
range's start, and one after it always starts at or after the range's end.
Clamp the estimated arrange position to that invariant so the retained
container can never overlap the realized elements, replacing the previous
approach of masking the container with animation-priority opacity and
hit-test overrides.
* Add failing test for #20688
Unit test created from issue description
* Add a failing test for wrong CacheLength handling
All items would be realized which is obviously wrong
* fix test setup used StackPanel instead of expcected
VirtualizingStackPanel
* Make sure the test actually fails
* update comment
* Fix for focusedElement and focusedIndex
* add another unit test
* Fixes for new test cases
* Addressing Review
* Update tests to match new behavior
* only recycle focused element if it is not null
* Address review
* Address copilot review
* add failing test
* fix StartU estimation
* remove unused sample file
* Add failing test for #20688
Unit test created from issue description
* Add a failing test for wrong CacheLength handling
All items would be realized which is obviously wrong
* fix test setup used StackPanel instead of expcected
VirtualizingStackPanel
* Make sure the test actually fails
* update comment
* Fix for focusedElement and focusedIndex
* add another unit test
* Fixes for new test cases
* Addressing Review
* Update tests to match new behavior
* only recycle focused element if it is not null
* Address review
* Address copilot review
* Extracted IInputRoot out of TopLevel
* Move some input handling out of TopLevel
* Remove old class, make layout manager private
* Removed IRenderRoot
* Make VisualTreeAttachmentEventArgs a bit more sensible
* Move ILayoutRoot to PresentationSource
# Conflicts:
# tests/Avalonia.Controls.UnitTests/TabControlTests.cs
* Updated some VisualRoot / GetVisualRoot usages
* Updated more XxxRoot usages
* More Root usages
* Addressed review
* Hurr-durr xml
* More fixes
* Maybe fix android compilation
* API diff
* Yet another cast
* I had to use MSIL analysis to detect those casts
* Fixed automation
* Fix PointerOverPreProcessor
* Fix?
* Removed yet another cast to Visual
* The amount of random downcasts is astonishing
* Maybe fix mac
* Addressed review
* remove margin deflation in BringDescendantIntoView and update related test to test cases with margin and no margin
* create transform relative to presenter, instead of child.
* update tests
* fix tests
* Add a failing test for #15712.
* Validate StartU at the start of a measure pass.
If any container U size has changed since the last layout pass then `StartU` must be considered unstable as the average container height will have changed.
* Correctly position focused element.
If the focused element has been moved outside the visible viewport due to a realized container size change, then we need to ensure it's positioned correctly.
* We can skip check if StartU is already unstable.
* Don't invalidate virt. panels more than necessary.
* Add another virt panel test.
And revert the expected results for another test to the way they were at the beginning of this PR.
* Tweak container size estimation.
Use the desired size of _measured_ containers instead of the bounds: a layout pass may not had completed on the containers yet, so the bounds may not be up-to-date. Was easier to move the estimation methods out of `RealizedStackElements` and into `VirtualizingStackPanel` itself in order to do this, and arguably makes more sense.
* Add more tests for ScrollIntoView.
* Improve ScrollIntoView.
Take into account the element we're scrolling to when calculating the anchor element for realization.
This fixes a case when elements have different widths and the BringIntoView wants to scroll horizontally due to custom TargetRect
Co-authored-by: Max Katz <maxkatz6@outlook.com>
* Add failing unit test for scenario 1 in #11878.
* Set TabOnceActiveElement on realized container.
Fixes scenario 1 in #11878.
* Use TabOnceActiveElement to decide focused element.
Fixes scenario #3 in #11878.
* Added failing unit tests for #11666.
* Handle items replaced with its own logic.
Don't rely on a remove/insert - it creates more work than needed.
* Fix some off-by-N errors in RealizedStackElements.
Fixes#11666
* Don't alter state in properties.
The `IValueEntry.HasValue` and `ValueFrame.IsActive` properties could alter state, which meant that when inspecting objects with these properties in a debugger, the state got altered by observing it. Make them methods.
* Deleted unused file.
* Add failing test for #12381.
* Use the index from the event in EvaluateIsActive.
Break the rules in this case.
Fixes#12381
---------
Co-authored-by: Max Katz <maxkatz6@outlook.com>
- animation/layout/render cycle is now managed from a central location
- animations are now throttled if animation/layout/render pass takes longer than a frame which previously caused a soft-freeze with input not being processed
- the public API is trimmed to make sure that we can make other planned changes during the 11.x support cycle
"Changelog":
- IClock is hidden and is planned to be replaced later
- Animator classes are hidden and are planned to be refactored later
- IAnimation members are hidden, it's supposed to be a marker interface for Style.Animations collection now, to start animations manually use Animation.RunAsync
- Sealed several classes in Avalonia.Animation namespace
- Spring class is removed from the public API (it wasn't possible to use it directly in a meaningful way anyway)
- Sealed brushes, transforms, effects and drawings
- Removed separate dispatcher priorities for Layout and Composition, everything now happens from a central place with Render priority (same as WPF)
- - some private "hook" priorities are added for now, those will be removed later
- IRenderLoop is hidden and removed from locator
- IRenderer is hidden (the plan is to remove that concept later)
- - Renderer.Start/Stop exposed as StartRendering/StopRendering on the toplevel (will be on a CompositionTarget/PresentationSource-like type later)
- - Renderer.Diagnistics exposed as RendererDiagnostics (same)
- - Renderer is no longer created by the platform code and is created by TopLevel itself
- - From the user-code hit-testing should be done by VisualExtensions.GetVisual(s)At, which has the same features
- - For unit tests a separate IHitTester interface is added which can be changed for a particular toplevel
- ILayoutManager is hidden
- - LayoutManager.ExecuteLayoutPass() exposed as TopLevel.UpdateLayout()
- Custom animators now have a separate base class that only deals with interpolation
Minor improvements:
- Compositor has a mode that doesn't use DispatcherTimers, useful for unit tests
- Introduced ScopedTestBase that auto-resets the locator when test is finished
- Check `CanFocus` in `FocusManager.SetFocusedElement`
- Check for `IsEffectivelyVisible` in `CanFocus`
- Clear focus when control made invisible
- Update tests that relied on unfocusable controls being focused
`ItemsControl` now works more like WPF, in that there are separate `Items` and `ItemsSource` properties. For backwards compatibility `Items` can still be set, though the setter is deprecated. `Items` needed to be changed from `IEnumerable` to `IList` though.
Julien Lebosquain
Yoyo
Nathan Nguyen
Alexander Marek
Tim
Nikita Tsukanov
Emmanuel Hansen
mpylon
dbriard
Steven Kirk
Bartosz Korczyński
Tom Edwards
Steven Kirk