Completely re-write the lightmap generation code:
- Follow the general lightmapper code structure from 4.0.
- Use proper path tracing to compute the global illumination.
- Use atlassing to merge all lightmaps into a single texture (done by @RandomShaper)
- Use OpenImageDenoiser to improve the generated lightmaps.
- Take into account alpha transparency in material textures.
- Allow baking environment lighting.
- Add bicubic lightmap filtering.
There is some minor compatibility breakage in some properties and methods
in BakedLightmap, but lightmaps generated in previous engine versions
should work fine out of the box.
The scene importer has been changed to generate `.unwrap_cache` files
next to the imported scene files. These files *SHOULD* be added to any
version control system as they guarantee there won't be differences when
re-importing the scene from other OSes or engine versions.
This work started as a Google Summer of Code project; Was later funded by IMVU for a good amount of progress;
Was then finished and polished by me on my free time.
Co-authored-by: Pedro J. Estébanez <pedrojrulez@gmail.com>
This change does two things:
1. Properly update the internal shape data using _update_in_shape_owner
when updating a shape (in 2D it was resetting one way collision)
2. Avoid unnecessary updates when calling set_shape with the same shape,
which happens each time a shape property is modified
(e.g shape.extents.x = ...)
Fixes#45090
(cherry picked from commit 4b43cd17c5)
Happy new year to the wonderful Godot community!
2020 has been a tough year for most of us personally, but a good year for
Godot development nonetheless with a huge amount of work done towards Godot
4.0 and great improvements backported to the long-lived 3.2 branch.
We've had close to 400 contributors to engine code this year, authoring near
7,000 commit! (And that's only for the `master` branch and for the engine code,
there's a lot more when counting docs, demos and other first-party repos.)
Here's to a great year 2021 for all Godot users 🎆
(cherry picked from commit b5334d14f7)
Complete rewrite of spatial partitioning using a bounding volume hierarchy rather than octree.
Switchable in project settings between using octree or BVH for rendering and physics.
This allows the user to query the AnimationNodeStateMachinePlayback's current
play position and total length of current animation state. These methods are currently
used in the editor plugin, but can also be useful for querying general playback state
information.
Added documentation for AnimationNodeStateMachinePlayback's `get_current_play_position`
(cherry picked from commit 674fb52f52)
Body transforms from physics are used to setup the joint and they are
only updated before the physics step by default.
Without forcing the transform update, joints could use a previous
position if the body's position was set after it was added to the scene.
3D physics is not affected by this issue.
(cherry picked from commit 11bee25de4)
See #36285 which mistakenly added documentation for the whole C++ API, while
some of it is meant to be and stay private as it's not exposed to scripts.
The access modifiers and method prefix were not used properly.
Cleanup code, and rename wrong `group_name` parameters to `method`, as it's a
method name which is being broadcast.
This is a very old class from pre-open source days, chances are that it was
just forgotten and not meant to be kept as is and undocumented.
(cherry picked from commit 79ba70f7ee)
Partially revert change allowing sprite get_rect snapping to be controlled by `pixel_snap` again rather than `transform_snap` (to prevent breaking compatibility). Adds a final `use_camera_snap` project setting to allow snapping viewports as in reduz original PR.
See #43689.
Also 'fixed' some spelling for behavior in publicly visible strings.
(Sorry en_GB, en_CA, en_AU, and more... Silicon Valley won the tech spelling
war.)
(cherry picked from commit a655de89e3)
Prevents `Timer` to prematurely start and timeout immediately if internal
processing is enabled manually with `Timer.set_process_internal(true)` or
`Timer.set_physics_process_internal(true)`.
Even if the internal processing is enabled manually, the user still has to
actually start the timer with `start()` method explicitly.
(cherry picked from commit afcb6f38db)
The return type for `_make_custom_tooltip` is clarified as Control, and users
should make sure to return a visible node for proper size calculations.
Moreover in the current master branch, a PopupPanel will be added as parent
to the provided tooltip to make it a sub-window.
Clarifies documentation for `Control._make_custom_tooltip`, and shows how to
use the (until now undocumented) "TooltipPanel" and "TooltipLabel" theme types
to style tooltips.
Fixes#39677.
(cherry picked from commit c5d8dafec4)
Previously, when the mouse was released after dragging a scrollbar,
its highlight was not dropped (if the mouse cursor was still inside
the viewport). This seems to be because the currently hovered control
only gets updated when the mouse is moved.
This commit fixes the dropping of the cosmetic highlight by running
the check for whether the currently hovered control has changed on
mouse-clicks, in addition to to the existing mouse-movements.
(cherry picked from commit e8804b9978)
Every NOTIFICATION_PROCESS the spatial_editor_plugin.cpp is calling set_use_fxaa which is causing a redraw_request(). Same with debanding.
These can be fixed be checking for noop state changes.
It can be enabled in the Project Settings
(`rendering/quality/filters/use_debanding`). It's disabled
by default as it has a small performance impact and can make
PNG screenshots much larger (due to how dithering works).
As a result, it should be enabled only when banding is noticeable enough.
Since debanding requires a HDR viewport to work, it's only supported
in the GLES3 backend.
This fixes a regression from PR #40313 (support for software skinning in MeshInstance).
Before, the base mesh was always updated on load even if not skinning
was used, which caused mesh instance materials to be reset on the
rendering side.
Now the base mesh is set only when it has been modified, or when
switching software skinning on or off. In this case the mesh instance
materials are always updated properly afterwards.
My mistake I missed this before the PR got merged:
is_visible_in_tree() should be avoided being called per frame unless absolutely necessary, because it is a recursive function that traverses the scene tree. It should be used when deciding on rare occasions whether to switch on or off skeleton processing, but it is better to use the cheaper is_visible() check on the per frame update.
Don't apply lighting to objects when they have a lightmap texture and
the light is set to BAKE_ALL. This prevents applying the same direct
light twice on the same object and makes setting up scenes with mixed
lighting much easier.
Option in MeshInstance to enable software skinning, in order to test
against the current USE_SKELETON_SOFTWARE path which causes problems
with bad performance.
Co-authored-by: lawnjelly <lawnjelly@gmail.com>
These values are only meaningful in the range 0 to 1.
Make sure the editor enforces reasonable values.
Fixes#42202.
(cherry picked from commit 6c18baee9c)
When changing the texture region for a StyleBox, the regions was not updating automatically in the Texture editor.
(cherry picked from commit 7d7727bade)
InterpolatedCamera has already been removed from the `master` branch.
This adds a deprecation notice to inform people about the upcoming removal
in Godot 4.0.
Its functionality could be replicated in a GDScript add-on with relative
ease.
Fixes#41006 (regression from #39164).
The original alignment fix was limited to PROCESS_DRAW mode, which
caused some discrepancies with PROCESS_POINTER mode.
Now only PROCESS_CACHE is excluded with a condition a few lines above.
(cherry picked from commit b783fa1416)
LineEdit should not return early when processing KEY_ENTER, so it can
consume the event properly.
Regression introduced by mistake while fixing enter events for Android
(PR #40487 - c0b394572f)
(cherry picked from commit 5c63dec36e)
Updated set_max_length() function to actually pull a substring of the current text so it's not all thrown away when the new max length is shorter than the current length.
(cherry picked from commit 71febfd6e2)
The constant is already exposed in GDScript, but not in C++.
This information is useful for implementing animated texture
resource importers via modules.
(cherry picked from commit 528056a3c5)
Changed CPU velocity calculation for EMISSION_SHAPE_DIRECTED_POINTS
to follow the same logic as in the GPU version:
mat2 rotm;
rotm[0] = texelFetch(emission_texture_normal, emission_tex_ofs, 0).xy;
rotm[1] = rotm[0].yx * vec2(1.0, -1.0);
VELOCITY.xy = rotm * VELOCITY.xy;
Now both CPUParticles2D & CPUParticles3D (z disabled) show the same results
as their GPU counterparts and take the initial velocity settings into account.
(cherry picked from commit 1c231cacb3)
Adding support for disabling virtual keyboard on mobile platforms, in
order to make it consistent with LineEdit.
It allows implementing a custom virtual keyboard.
See: https://github.com/godotengine/godot/pull/41139
Add any leftover fractional pixels to an error accumulator. When the
accumulator is greater or equal to one, add one pixel to the current
Node's size and subtract one from the accumulator.
Closes#36522
(cherry picked from commit 04ea6ec88d)
This is a revert of 9d78274e06, which was an attempt to fix#21431, but in the end it seems a different problem was the root of the issue.
Renewing focus steal allowance every time is needed on Windows.
Depending on the device implementation, editor actions could be
received with different action ids or not at all for multi-line.
Added a parameter to virtual keyboards to properly handle single-line
and multi-line cases in all situations.
Single-line:
Input type set to text without multiline to make sure actions are sent.
IME options are set to DONE action to force action id consistency.
Multi-line:
Input type set to text and multiline to make sure enter triggers new lines.
Actions are disabled by the multiline flag, so '\n' characters are
handled in text changed callbacks.
Emit `node_selected` and `node_deselected` signals where appropriate
when selecting or unselecting multiple GraphEdit nodes at once.
(cherry picked from commit c1d5c7727c)
Add optional parameter to specify whether applying rotation to the
PathFollow3D's Transform is necessary, preventing erroneous updates.
(cherry picked from commit be3a1769fe)
`ConvexPolygonShape2D` and `ConcavePolygonShape2D` are only meant to be
used directly in code and not in the editor for physics-based use cases
specifically.
Developers are advised to use `CollisionPolygon2D` instead, which does
generate those shapes under the hood, handling polygon convexivity,
proper orientation etc.
(cherry picked from commit dc446203be)
In set_stream() we write to 'playback' while accessing the same object
in _mix_audio() in audio thread. Protect the 'write' part in
set_stream() to avoid possible crash in _mix_audio() function.
(cherry picked from commit e435d57758)
This reverts commit 7f61710183.
See #38868, in its current implementation a small skew value might end up
serialized to scene files due to floating point precision errors, which is
detrimental to VCS.
This can be cherry-picked anew once a fix for #38868 has been found.
Make command-backspace in line edit work like other macOS applications.
If there is a selection, command-backspace deletes the selection.
If there isn't a selection, command-backspace deletes from the cursor to the beginning of the line edit.
This addresses part of godotengine/godot#23548
- The repaint code was moved outside of set_current_tab() and to a "_repaint()" private function
- _on_theme_changed() will now only call _repaint() and update()
This means _on_theme_changed() will do only what it needs (repainting the TabContainer to account for the new theme)
fixes#39498
(cherry picked from commit 5123006f06)
Change error checking in `duplicate_signals()` to check for path to
`p_original`, thus adhering to the method used in `duplicate`, instead
of checking for ownership.
Configured for a max line length of 120 characters.
psf/black is very opinionated and purposely doesn't leave much room for
configuration. The output is mostly OK so that should be fine for us,
but some things worth noting:
- Manually wrapped strings will be reflowed, so by using a line length
of 120 for the sake of preserving readability for our long command
calls, it also means that some manually wrapped strings are back on
the same line and should be manually merged again.
- Code generators using string concatenation extensively look awful,
since black puts each operand on a single line. We need to refactor
these generators to use more pythonic string formatting, for which
many options are available (`%`, `format` or f-strings).
- CI checks and a pre-commit hook will be added to ensure that future
buildsystem changes are well-formatted.
(cherry picked from commit cd4e46ee65)
The crash happens further down when setting an invalid owner in
`Node::_set_owner_nocheck` but I couldn't figure out how to fix it.
But here the proper fix is to catch the invalid scene file early on
and fail loading it.
Part of #17372.
(cherry picked from commit c080ec5da2)
`get_char_size()` is a public virtual function defined in the `Font`
class. Implementations exist for both `BitmapFont` and `Dynamic Font`.
However, it was only exposed to the GDScript API through the Bitmap
Font, and not for Dynamic Font.
This commit exposes the function through `Font` instead.
Fixes#23967
(cherry picked from commit a4413710f9)
The specific case for object reference seems unnecessary, as `RES res = var`
already does the work. The case where REF is invalid is never hit in the case
of already freed objects.
The assignment `res = *r` was causing the resource to be always invalidated
on the 3.2 branch.
(cherry picked from commit 12685df423)
* On press left+command+shift or right+command+shift it should behave like shift+home or shift+end and select the text
* Using home and end events as reference
(cherry picked from commit d8f8a3c606)
This reverts commit bf1cc116e1.
This needs more discussion to avoid breaking some users' expectations.
See #38086 for arguments.
(cherry picked from commit 69f36cea8c)
Skew is x-axis only, because it must be bidirectionally convertible to a 2x3 matrix, but you can subtract it to the rotation to get the effect on y-axis
(cherry picked from commit efb1f7d76b)
(cherry picked from commit 38085f2f6982c491935a434bb45e358dbebe1714)
(cherry picked from commit b9c280b73ff6a13ea490d2da0f2728bcef3038dc)
(cherry picked from commit 895ed2aed7)
Calls to set_amount can increase the size of the particle array, but do not zero the memory, they only set the active flag to false. This uninitialized memory can be sent to the GPU, possibly as NaNs.
RichTextEffect can now have a bbcode string starting like one of the built-in.
It was impossible before as the built-in would take precedence over the custom effect that has the same bbcode start.
Example : [fade] would take precedence over [fade_in]
(cherry picked from commit b4e3042cba)
The stack size of the undo history of a TextEdit was not limited leading
to potential memory leaks when doing lots of operations on a TextEdit.
This commit adds the option gui/common/text_edit_undo_stack_max_size
to the project settings. The first element of the undo stack is popped
if the stack's size exceeds this value ensuring limited memory usage.
The default stack size setting is 1024.
Fixes#37838.
(cherry picked from commit 4a82390aaf)
