-For inspector refresh, the inspector now detects if a property change by polling a few times per second and then does update the control if so. This process is very cheap.
-For property list refresh, a new signal (property_list_changed) was added to Object. _change_notify() is replaced by notify_property_list_changed()
-Changed all objects using the old method to the signal, or just deleted the calls to _change_notify(<property>) since they are unnecesary now.
-Rendering server now uses a split RID allocate/initialize internally, this allows generating RIDs immediately but initialization to happen later on the proper thread (as rendering APIs generally requiere to call on the right thread).
-RenderingServerWrapMT is no more, multithreading is done in RenderingServerDefault.
-Some functions like texture or mesh creation, when renderer supports it, can register and return immediately (so no waiting for server API to flush, and saving staging and command buffer memory).
-3D physics server changed to be made multithread friendly.
-Added PhysicsServer3DWrapMT to use 3D physics server from multiple threads.
-Disablet Bullet (too much effort to make multithread friendly, this needs to be fixed eventually).
-Always use temporal reproject, it just loos way better than any other filter.
-By always using termporal reproject, the shadowmap reduction can be done away with, massively improving performance.
-Disadvantage of temporal reproject is update latency so..
-Made sure a gaussian filter runs in XY after fog, this allows to keep stability and lower latency.
- Based on C++11's `thread` and `thread_local`
- No more need to allocate-deallocate or check for null
- No pointer anymore, just a member variable
- Platform-specific implementations no longer needed (except for the few cases of non-portable functions)
- Simpler for `NO_THREADS`
- Thread ids are now the same across platforms (main is 1; others follow)
-SDFGI direct light is done over many frames
-SDFGI Changed settings for rays/frame
-SDFGI Misc optimizations
-SDFGI Bug fix on probe scroll
-GIProbe was not working, got it to work again
-GIProbe dynamic objects were not working, fixed
-Added a half size GI option.
Fix Raycast3D node render debug not showing in editor camera preview.
Use dynamic mesh update to change the ray on-the-fly without too much
extra cost when collision debug is enabled.
Fixes#43571
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
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 🎆
This makes it possible to view far away objects without
having to tweak any settings. This results in a more usable
editor when working on large-scale levels.
This change should have no impact on performance, but note that
Z-fighting will be visible at a distance. This can be made less
visible by increasing the Znear value (however, doing so will cause
nearby surfaces to disappear).
This change was also applied to the editor, but it will only
apply to newly created scenes.
This also changes the default camera settings in the glTF importer
to match the Camera node's defaults.
-Happens on import by default for all models
-Just works (tm)
-Biasing can be later adjusted per node or per viewport (as well as globally)
-Disabled AABB.get_support test because its broken
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.
-Changed how mesh data is organized, hoping to make it more efficient on Vulkan and GLES.
-Removed compression, it now always uses the most efficient format.
-Added support for custom arrays (up to 8 custom formats)
-Added support for 8 weights in skeleton data.
-Added a simple optional versioning system for imported assets, to reimport if binary is newer
-Fixes #43979 (I needed to test)
WARNING:
-NOT backwards compatible with previous 4.x-devel, will most likely never be, but it will force reimport scenes due to version change.
-NOT backwards compatible with 3.x scenes, this will be eventually re-added.
-Skeletons not working any longer, will fix in next 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.)
* Removed the pointers to PhysicalBone in the code, as they were unused.
* Forward ported the SkeletonIK bone scaling fix I made from Godot 3.2 to Godot 4.0.
* Fixed issue where the root bone in the IK chain would not rotate correctly.
* The issue turned out to be the update_chain function being called in solve. This would override the root bone transform incorrectly and that would cause it not to rotate after just a single solve. Removing the update_chain function fixes the issue and based on my testing there are no adverse effects.
* While the old fix on this PR (prior to a force push) required a hack fix, this new fix does not!
* Removed the update_chain function. This change doesn't appear to have any adverse effects in any of the projects I tested (including with animations, Skeleton3D or otherwise, from AnimationPlayer nodes!)
* Fixed issue where the scale of the Skeleton node would change the position of the target, causing it not to work with skeletons that have a global scale of anything but 1.
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.
In general they are more confusing to users because they expect
inheritance to fully override parent methods. This behavior can be
enabled by script writers using a simple super() call.
- Makes all boolean setters/getters consistent.
- Fixes bug where `glow_hdr_bleed_scale` was not used.
- Split CameraEffects to their own source file.
- Reorder all Environment method and properties declarations,
definitions and bindings to be consistent with each other
and with the order of property bindings.
- Bind missing enum values added with SDFGI.
- Remove unused SDFGI enhance_ssr boolean.
- Sync doc changes after SDFGI merge and other misc changes.
* Added helper functions to Skeleton3D for converting transforms from bone space to global space, and vice versa.
* Updated the Skeleton3D class reference.
* Changed the icon used for bones in the Skeleton3D inspector to use BoneAttachement3D's icon.
* Changed the Skeleton3D inspector to use EditorPropertyTransform and EditorPropertyVector3 when possible.
* Placed the Transform/Matrix for each bone in a sub-section, so it is visually similar to the Node3D inspector.
Fixes#36372 as Path2D/Path3D's `curve` property no longer uses a Curve
instance as default value, but instead it gets a (unique) default Curve
instance when created through the editor (CreateDialog).
ClassDB gets a sanity check to ensure that we don't do the same mistake
for other properties in the future, but instead use the dedicated
property usage hint.
Fixes#36372.
Fixes#36650.
Supersedes #36644 and #36656.
Co-authored-by: Thakee Nathees <thakeenathees@gmail.com>
Co-authored-by: simpuid <utkarsh.email@yahoo.com>
I couldn't find a tool that enforces it, so I went the manual route:
```
find -name "thirdparty" -prune \
-o -name "*.cpp" -o -name "*.h" -o -name "*.m" -o -name "*.mm" \
-o -name "*.glsl" > files
perl -0777 -pi -e 's/\n}\n([^#])/\n}\n\n\1/g' $(cat files)
misc/scripts/fix_style.sh -c
```
This adds a newline after all `}` on the first column, unless they
are followed by `#` (typically `#endif`). This leads to having lots
of places with two lines between function/class definitions, but
clang-format then fixes it as we enforce max one line of separation.
This doesn't fix potential occurrences of function definitions which
are indented (e.g. for a helper class defined in a .cpp), but it's
better than nothing. Also can't be made to run easily on CI/hooks so
we'll have to be careful with new code.
Part of #33027.
Which means that reduz' beloved style which we all became used to
will now be changed automatically to remove the first empty line.
This makes us lean closer to 1TBS (the one true brace style) instead
of hybridating it with some Allman-inspired spacing.
There's still the case of braces around single-statement blocks that
needs to be addressed (but clang-format can't help with that, but
clang-tidy may if we agree about it).
Part of #33027.
Using `clang-tidy`'s `modernize-use-default-member-init` check and
manual review of the changes, and some extra manual changes that
`clang-tidy` failed to do.
Also went manually through all of `core` to find occurrences that
`clang-tidy` couldn't handle, especially all initializations done
in a constructor without using initializer lists.
-Added LocalVector (needed it)
-Added stb_rect_pack (It's pretty cool, we could probably use it for other stuff too)
-Fixes and changes all around the place
-Added library for 128 bits fixed point (required for Delaunay3D)
This reverts commit ec7b481170.
This was wrong, `d` is not a distance but the `d` constant in the
parametric equation `ax + by + cz = d` describing the plane.
Part of #33027, also discussed in #29848.
Enforcing the use of brackets even on single line statements would be
preferred, but `clang-format` doesn't have this functionality yet.
A vertical FOV of 75 degrees is roughly equivalent to a 91 degree
horizontal FOV on a 4:3 display (~107.51 degrees on 16:9),
which is close to the typical default FOV used in PC games.
Note that this doesn't apply to the in-editor camera which keeps its
FOV to 70. This is because it doesn't display in fullscreen;
its viewport only displays in the center of the editor (roughly).
This means the viewport won't cover the viewer's eyes as much. Therefore,
the editor camera FOV should be slightly lower to account for this.
Since this changes the default value, this may break existing projects
slightly.
For the record, this was already done in
https://github.com/godotengine/godot-demo-projects/pull/260
for the official demo projects.
It changed name as part of the DisplayServer and input refactoring
in #37317, with the rationale that input no longer goes through the
main loop, so the previous Input singleton now only does filtering.
But the gains in consistency are quite limited in the renaming, and
it breaks compatibility for all scripts and tutorials that access
the Input singleton via the scripting language. A temporary option
was suggested to keep the scripting singleton named `Input` even if
its type is `InputFilter`, but that adds inconsistency and breaks C#.
Fixesgodotengine/godot-proposals#639.
Fixes#37319.
Fixes#37690.
Also implemented decal atlas, so projectors and other stuff can be added.
Sidenote: Had to make RID hashable, so some unrelated includes changed
in order to include it in hashfuncs.h
- Made shadow bias size independent, so it will remain when changing light or camera size.
- Implemented normal offset bias, which greatly enhances quality.
- Added transmission to subsurface scattering
- Reimplemented shadow filter modes
Closes#17260
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.
For us, it practically only changes the fact that `A<A<int>>` is now
used instead of the C++03 compatible `A<A<int> >`.
Note: clang-format 10+ changed the `Standard` arguments to fully
specified `c++11`, `c++14`, etc. versions, but we can't use `c++17`
now if we want to preserve compatibility with clang-format 8 and 9.
`Cpp11` is still supported as deprecated alias for `Latest`.
In the case where a ConcavePolygonShape is used as a shape for a RigidBody
in another mode than static, a configuration warning will appear in the
editor.
Those were problematic as they call a method of their parent class,
but callable_mp does not allow that unless it's public.
To solve it, we declare a local class that calls the parent class'
method, which now needs to be protected to be accessible in the
derived class.
It's tedious work...
Some can't be ported as they depend on private or protected methods
of different classes, which is not supported by callable_mp (even if
it's a class inherited by the current one).
Remove now unnecessary bindings of signal callbacks in the public API.
There might be some false positives that need rebinding if they were
meant to be public.
No regular expressions were harmed in the making of this commit.
(Nah, just kidding.)
Main:
- It's now implemented thanks to `<mutex>`. No more platform-specific implementations.
- `BinaryMutex` (non-recursive) is added, as an alternative for special cases.
- Doesn't need allocation/deallocation anymore. It can live in the stack and be part of other classes.
- Because of that, it's methods are now `const` and the inner mutex is `mutable` so it can be easily used in `const` contexts.
- A no-op implementation is provided if `NO_THREADS` is defined. No more need to add `#ifdef NO_THREADS` just for this.
- `MutexLock` now takes a reference. At this point the cases of null `Mutex`es are rare. If you ever need that, just don't use `MutexLock`.
- Thread-safe utilities are therefore simpler now.
Misc.:
- `ScopedMutexLock` is dropped and replaced by `MutexLock`, because they were pretty much the same.
- Every case of lock, do-something, unlock is replaced by `MutexLock` (complex cases where it's not straightfoward are kept as as explicit lock and unlock).
- `ShaderRD` contained an `std::mutex`, which has been replaced by `Mutex`.
- Renames PackedIntArray to PackedInt32Array.
- Renames PackedFloatArray to PackedFloat32Array.
- Adds PackedInt64Array and PackedFloat64Array.
- Renames Variant::REAL to Variant::FLOAT for consistency.
Packed arrays are for storing large amount of data and creating stuff like
meshes, buffers. textures, etc. Forcing them to be 64 is a huge waste of
memory. That said, many users requested the ability to have 64 bits packed
arrays for their games, so this is just an optional added type.
For Variant, the float datatype is always 64 bits, and exposed as `float`.
We still have `real_t` which is the datatype that can change from 32 to 64
bits depending on a compile flag (not entirely working right now, but that's
the idea). It affects math related datatypes and code only.
Neither Variant nor PackedArray make use of real_t, which is only intended
for math precision, so the term is removed from there to keep only float.
This attribute is now part of the standard we target so we no longer
need compiler-specific hacks.
Also enables -Wimplicit-fallthrough for Clang now that we can properly
support it. It's already on by default for GCC's -Wextra.
Fixes new warnings raised by Clang's -Wimplicit-fallthrough.
Fix -Wunused-variable, -Wunused-but-set-variable and -Wswitch warnings
raised by GCC 8 and 9.
Fix -Wunused-function, -Wunused-private-field and
-Wtautological-constant-out-of-range-compare raised by Clang.
Fix MSVC 2019 warning C4804 (unsafe use of type 'bool' in comparison
operation).
GCC -Wcpp warnings/Clang -W#warnings (`#warning`) are no longer raising
errors and will thus not abort compilation with `werror=yes`.
Treat glslang headers are system headers to avoid raising warnings.
Re-enables us to build with `werror=yes` on Linux and macOS, thus
catching warnings that would be introduced by new code.
Fixes#36132.
Lots of internal API changes and some docstrings were lost in the conversion.
I manually salvaged many of them but for all the rendering-related ones, an
additional pass is needed.
Added missing enum bindings in BaseMaterial3D and VisualServer.
-Texture renamed to Texture2D
-TextureLayered as base now inherits 2Darray, cubemap and cubemap array
-Removed all references to flags in textures (they will go in the shader)
-Texture3D gone for now (will come back later done properly)
-Create base rasterizer for RenderDevice, RasterizerRD
Fixes#26637.
Fixes#19900.
The viewport_size returned by get_viewport_size was previously incorrect, being half the correct value. The function is renamed to get_viewport_half_extents, and now returns a Vector2.
Code which called this function has also been modified accordingly.
This PR also fixes shadow culling when using ortho cameras, because the correct input for CameraMatrix::set_orthogonal should be the full HEIGHT from get_viewport_half_extents, and not half the width.
It also fixes state.ubo_data.viewport_size in rasterizer_scene_gles3.cpp to be the width and the height of the viewport in pixels as stated in the documentation, rather than the current value which is half the viewport extents in worldspace, presumed to be a bug.
When there is no collision with a floor the get_floor_normal() function
should return the zero vector to be consistent with get_floor_velocity().
Renames floor_normal to up_direction in all bindings.
Updates the documentation of get_floor_normal() and get_floor_velocity()
to make it clear when the values are valid. Updates the documentation for
move_and_slide() and move_and_slide_with_snap() to use the new up_direction
parameter name.
Updating the floor velocity with the body's current linear velocity
discards the velocity component provided by the body's angular
rotation. Without the current contact point there is no way to calculate
the current velocity component provided by the body's angular rotation
therefore we need to use the velocity calculated at the time of the
collision.
Fixes#34807.
Happy new year to the wonderful Godot community!
We're starting a new decade with a well-established, non-profit, free
and open source game engine, and tons of further improvements in the
pipeline from hundreds of contributors.
Godot will keep getting better, and we're looking forward to all the
games that the community will keep developing and releasing with it.
In the 3D version:
- Partially revert #20908 that was reverted in the 2D version as part
of #21653. This ensures that the Vector returned is always perpendicular
to the surface collided with; and not the floor_normal Vector passed to
the function when on a floor.
- Include an update of the floor velocity before multiplying by the time
delta, which was added to the 2D version as part of commit 13a8014.
In the 2D version:
- Use the Vector2.slide() function instead of Vector2.tangent() to adjust
the amount of motion the stop_on_slope undoes to ensure that it is in the
right direction. This is a implementation of the 3D approach from #30588.
- Combine the !found_collision and motion == Vector2() checks for break.
- Other minor formating changes to make the functions look identical.
Also renamed some variables to align with their use.
It was triggering a warning in bullet followed with a crash in some cases.
WARNING: assert_no_constraints: A body with a joints is destroyed. Please check the implementation in order to destroy the joint before the body.
At: modules/bullet/rigid_body_bullet.cpp:465
A z_depth of 0 returns the camera position, which is not really useful.
This also makes the API breakage from 3.1 clearer as 3.1 code will now
fail to compile, so users will have to adapt and use the new parameter.
For the reference, in 3.1, the z_depth was hardcoded to the near plane.
Closes#33493.
Make sure particles are processed during the same frame when visibility is set to on, in case they are still active from before and need to be restarted.
Fixed#33476
Particles were processed only on the next frame after the emission started, causing a one frame delay in rendering. Now the first process cycle is started during the same frame, which makes them consistent with Particles & Particles2D.
Fixes#32890
Adds a new NavigationMesh property to select which objects will be taken
into account for the generation.
By default it will use all the NavigationMeshInstance children to keep
compatibility. The new modes allow to build the NavigationMesh from
all the nodes belonging to a specific group, and optionally include
their children too.
When moving KinematicBody2D from one scene to another and not freeing
the old scene, the first call to move_and_slide() in the new scene will
generate an error because KinematicBody2D keeps internaly a
RID on_floor_body of a body resource in the old scene which no more has
a space assigned.
To fix this, on_floor_body is set to empty RID in response to
NOTIFICATION_ENTER_TREE notification of KinematicBody2D and
KinematicBody. Also all other data related to move_and_slide() is reset:
floor, ceiling, wall flags, colliders vector, floor_velocity.
This fixes#31416.
This was a regression in 3.1 and later from the new inspector, where
PROPERTY_HINT_SPRITE_FRAME was not fully re-implemented. It's meant to
be a normal PROPERTY_HINT_RANGE which also automatically increments its
value when keyed in the animation player.
To avoid code duplication, I made the frames properties use the actual
PROPERTY_HINT_RANGE and introduced a PROPERTY_USAGE_KEYING_INCREMENTS
usage flag instead.
Also added support for SCons project-absolute paths (starting with #) and
warning about duplicates in add_source_files(), and fixed
default_controller_mappings.gen.cpp being included twice after first build
due to *.cpp globbing.
Part of #30270.
With the default camera node settings, this makes directional shadows
look consistent between the editor and the running project.
The original issue occurs because the editor camera defaults to a
Z-far value of 500, whereas the Camera node defaults to a Z-far
value of 100. Since the directional shadow maximum distance is clamped
to the Z-far value, it caused the running project's effective shadow
distance to be lower compared to the editor (100 instead of 200).
This partially addresses #13575.
- Refer to properties explicitly when possible
- When multiple warnings are returned, always separate them by one
blank line to make them easier to distinguish
- Improve grammar and formatting
For clarity, assign-to-release idiom for PoolVector::Read/Write
replaced with a function call.
Existing uses replaced (or removed if already handled by scope)
The original shader code uses a phase (ratio from 0 to 1 for the particle
lifetime) for the randomness ratio computations, and this code was ported
over but converted to time computations.
The seeding/cycle logic was thus invalid, so we're going back to phase
for these computations, thus fixing the previous non-working time/emission
randomness property.
Part of #29692. Follow-up to #26859.
The tangential acceleration for both CPUParticles2D and CPUParticles had been
badly converted from their GPU counterpart (ParticlesMaterial).
This fixes it and ensures that both GPU and CPU particles behave the same with
regard to tangential acceleration.
It's not necessary, but the vast majority of calls of error macros
do have an ending semicolon, so it's best to be consistent.
Most WARN_DEPRECATED calls did *not* have a semicolon, but there's
no reason for them to be treated differently.
As mentioned in
https://github.com/godotengine/godot/pull/26897#issuecomment-491178089
the look-at scaling issue solved by PR #26897 happens also in another
look-at method.
Spatial::look_at_from_position() also does not have same input checking
Spatial::look_at() has. Therefore, I fixed it too at same time.
* Expose EditorNavigationMeshGenerator as an engine singleton so users
can generate navmesehes from `tool` scripts.
* Add support for generating navmeshes from static colliders. All
collision shapes are supported except for Plane (since Plane is an
infinite collider and navmeshes need to have finite geometry).
* When using static colliders as a geometry source, a layer mask can be
specified to ignore certain colliders.
* Don't rely on global transform. It still should give the exact same
results but allows for building navmeshes on nodes that are not in the
tree (useful in `tool` scripts).
* Update navigation gizmos after every new bake.
This work has been kindly sponsored by IMVU.
It always normalized basis after look_at() computation.
Now it applies previous scale back, in order to avoid
distortions when global scale was different of (1,1,1).
fix#10003 and #19000
Related to #17924
Adds the ability to directly add disabled shapes to a collision object. Before this commit a shape has always been assumed to be enabled and had to be disabled in an extra step.
The bake mode property of lights previously didn't affect GI probes.
This change makes the GI probe ignore lights that have their bake mode
set to disabled.
Made AudioFrame and Vector2 equivalent for casting.
Added ability to obtain the playback object from stream players.
Added ability to obtain effect instance from audio server.
It seems to stay compatible with formatting done by clang-format 6.0 and 7.0,
so contributors can keep using those versions for now (they will not undo those
changes).
FBX support and MMD (pmx) support.
Normals, Albedo, Metallic, and Roughness through Arnold 5 Materials for Maya FBX.
Maya FBX Stingray PBS support.
Importing FBX static meshes work.
Importing FBX animations is a work in progress.
Supports FBX 4 bone influence animations.
Supports FBX blend shapes.
MMDs do not have an associated animation import yet.
Sponsored by IMVU Inc.
Adds `FALLTHROUGH` macro to specify when a fallthrough is intentional.
Can be replaced by `[[fallthrough]]` if/when we switch to C++17.
The warning is now enabled by default for GCC on `extra` warnings level
(part of GCC's `-Wextra`). It's not enabled in Clang's `-Wextra` yet,
but we could enable it manually once we switch to C++11. There's no
equivalent feature in MSVC for now.
Fixes#26135.
1. Consider 'own_world' as well as 'world' to stop propagating enter/exit world notifications.
2. Clean & fix handling of camera currency.
This fixes some random crashes and error logs in the editor; namely
- when enabling/disabling own world in a Viewport;
- when switching back from a subscene displayed into a main scene's Viewport;
- when exiting the editor after any of them;
- memory corruption (can that explain certain other seemingly unrelated crash reports?).
This also fixes situations where a Viewport and its main Camera get out of sync about which World is relevant to them.
With this change finally one can use compound collisions (like those created
by Gridmaps) without serious performance issues. The previous KinematicBody
code for Bullet was practically doing a whole bunch of unnecessary
calculations. Gridmaps with fairly large octant sizes (in my case 32) can get
up to 10000x speedup with this change (literally!). I expect the FPS demo to
get a fair speedup as well.
List of fixes and improvements:
- Fixed a general bug in move_and_slide that affects both GodotPhysics and
Bullet, where ray shapes would be ignored unless the stop_on_slope parameter
is disabled. Not sure where that came from, but looking at the 2D physics
code it was obvious there's a difference.
- Enabled the dynamic AABB tree that Bullet uses to allow broadphase collision
tests against individual shapes of compound shapes. This is crucial to get
good performance with Gridmaps and in general improves the performance
whenever a KinematicBody collides with compound collision shapes.
- Added code to the broadphase collision detection code used by the Bullet
module for KinematicBodies to also do broadphase on the sub-shapes of
compound collision shapes. This is possible thanks to the dynamic AABB
tree that was previously disabled and it's the change that provides the
biggest performance boost.
- Now broadphase test is only done once per KinematicBody in Bullet instead of
once per each of its shapes which was completely unnecessary.
- Fixed the way how the ray separation results are populated in Bullet which
was completely broken previously, overwriting previous results and similar
non-sense.
- Fixed ray shapes for good now. Previously the margin set in the editor was
not respected at all, and the KinematicBody code for ray separation was
complete bogus, thus all previous attempts to fix it were mislead.
- Fixed an obvious bug also in GodotPhysics where an out-of-bounds index was
used in the ray result array.
There are a whole set of other problems with the KinematicBody code of Bullet
which cost performance and may cause unexpected behavior, but those are not
addressed in this change (need to keep it "simple").
Not sure whether this fixes any outstanding Github issues but I wouldn't be
surprised.