…_settings

Cleanup Android input on render thread settings

…ened-bottom-panel-after-restoring-to-side

Fix Toggle Last Opened Bottom Panel not working after restoring FileSystem Dock to the side

…importation-settings

Fix missing options in Project Import Defaults

Fix LCD font AA modulation in RendererRD.

…n-parent

Fix container minimum size with hidden parent

…-crash

Fix editor inspector crashing when the old object is no longer valid

…-chain-call-setter

GDScript: Call setter on simple setter chain without getter

…plicate-doc-update

Clarify that `Resource.duplicate(true)` doesn't duplicate subresources inside `Array` or `Dictionary` properties

Fix unclear `PROPERTY_USAGE_STORAGE`/`EDITOR` description

Fix SSR orientation issues when using orthogonal camera

…oop-fix

Lightmapper: Ensure the atlas is big enough to fit padded UV maps

Make errors on `RefCounted.free()` more accurate

ThorVG: Update to 0.14.2

D3D12: Avoid enabling depth bounds test if unsupported

Add `SwitchVector4D` shortcut to visual shader members

Only print physics sync warning on dev builds to avoid warning spam from using `move_and_slide`

Re-add `Skeleton3D::animate_physical_bones` property

…signed-docs

GDExtension docs: Advise bindings to use `uint64_t` for bitfields

[CI] Upload build cache before running tests

…identifier

GDScript: Restore support for `Token::UNDERSCORE` in identifiers

…t_what_even_is_a_parent

Clarify `is_visible_in_tree()` in CanvasItem

Fix NavigationPolygon not updating consistently on rebakes

Fix typo in AABB docs

…-fix

Fix texture compression descriptions in RenderingDevice docs

Long press is used to simulate right-click events for finger touch and stylus. The previous logic also caused it to trigger for mouse input, which is not needed since the user can instead use the mouse right click button.

This update disables long press as right click events for mouse input.

…ng-driver-initial-size-fix

Windows: Fix wrong initial size passed to rendering driver

Restore default clear color after displaying boot splash

…rams

Fix error when exporting files without import parameters

Fix physics tick count in `Input.action_press` and `Input.action_release`

…utton_not_enabling

Fix issue preventing enabling the remote button for Android/iOS

…al-link

Add tutorial link to Parallax2D doc

…r-with-files

Fix removing a folder that contains a file is not removed from the FileSystem Dock

…pec-constant

Use a spec constant to control whether the MultiMesh branch is used in the vertex shader.

…-abs

Fix to restore `library_path` as absolute path

Make ShaderLanguage's parser recognize sampler passed in from array

…updating_touchscreen_settings

[Android Editor] Resolve issues with the editor touchscreen settings

Wayland: Commit surface on window creation

Wayland: Check for suspended flag when unsuspending

…n_map

Fix creating default action map

Fix potential null access in `TextEdit`

…el-round-halfway-to-positive

Change GUI controls pixel snap to round halfway towards positive infinity (`floor(x + 0.5)`)

[Winink] Check pointer button states.

[Windows] Fix reading keyboard layout names.

…f_them_pixels

Clarify `clipboard_get_image()`

Some improvements to TileSetScenesCollectionSource docs

…ceholder-with-resource

Fix resources being skipped in InstancePlaceholder

Make sure that selected tile source is remembered

…_just_saying

Don't attempt to re-import broken files if they didn't change

…scroll

[macOS] Fix code completion list scrolling with trackpad.

Fix game window stops responding when debugger pauses

…ayback-finished-signal-redux

Fix audio samples not being able to be "finished"

…-error-column

Editor: Consider tabs when calculating column for jump to error

Allocated XImages are improperly free'd with XFree.
The X11 documentation says that XImage should use
XDestroyImage to free both the image structure and
the data pointed to by the image structure.

Also fix a potential use-after-free bug.

…ction

Fix incorrect `CAMERA_DIRECTION_WORLD` calculation

…playback

[Audio] Prevent crash in interactive playback

…ocky

RenderingDevice: Leave handling of compressed block granularity to the driver

…ot_compat_broken_with_hdr

Fix black `get_texture()` on viewport in compatibility mode with HDR enabled

…-loop-fix-always

Lightmapper: Prevent infinite loop while blitting lightmaps into an atlas

…rray

[Core] Fix TypedArray encoding when full objects is disabled

[Windows] Add `libSynchronization` to link flags.

…ory-leak

Linux/X11: Fix memory leak from created screen images

…me-on-scene-root

Disable unique name in scene when making a node the root node

Add warning when use FSR1 on renderer other than forward plus

Improve Curve preview colors for consistency

…touchscreen devices

Prior to this fix, scrolling via mouse drag on touchscreen devices, and
drag&drop operation on a `TreeItem` element would conflict with each other
preventing the drag scroll from being released when the mouse button is
released.

The issue is addressed by disabling drag&drop when drag scrolling is ongoing.

…lease

[CI] Run unit tests on desktop release templates

Fix empty region in AtlasTexture

Always render when XR is enabled, even if no OS windows can draw

Fix input lock issue when drag scrolling on a `Tree` element on touchscreen devices

…fix-depended-parsers

Autocompletion: Register depended parsers with the main parser

Few fixes for `VisualShaderNodeRotationByAxis`

Fix internal connection count decreased if visual shader node removed

…down

Speed up `GDScriptLanguage::finish`

…ssing

Fix handling of missing bin file in gltf separate

… disabled.

Fixes godotengine#94274

…ns on random physics objects

Fixes godotengine#92333

This looks correct, and fixes the ASAN assertion I'm currently getting
in my program.

This is about not letting the resource format loader set the error code directly on the task anymore. Instead, it's stored locally and assigned only when it is right to do so.

Otherwise, other tasks may see an error code in the current one before it's state having transitioned to errored. While this, besides the technically true data race, may not be a problem in practice, it causes surprising situations during debugging as it breaks assumptions.