# Objective
- There are multiple instances of `let Some(x) = ... else { None };`
throughout the project.
- Because `Option<T>` implements
[`Try`](https://doc.rust-lang.org/stable/std/ops/trait.Try.html), it can
use the question mark `?` operator.
## Solution
- Use question mark operator instead of `let Some(x) = ... else { None
}`.
---
There was another PR that did a similar thing a few weeks ago, but I
couldn't find it.
I just implemented this to record a video for the new blog post, but I
figured it would also make a good dedicated example. This also allows us
to remove a lot of code from the 2d/3d gizmo examples since it
supersedes this portion of code.
Depends on: https://github.com/bevyengine/bevy/pull/11699
# Objective
I tried using `insert_gizmo_group` to configure my physics gizmos in a
bevy_xpbd example, but was surprised to see that nothing happened. I
found out that the method does *not* overwrite gizmo groups that have
already been initialized (with `init_gizmo_group`). This is unexpected,
since methods like `insert_resource` *do* overwrite.
## Solution
Insert the configuration even if it has already been initialized.
# Objective
`RenderMeshInstance::material_bind_group_id` is only set from
`queue_material_meshes::<M>`. this field is used (only) for determining
batch groups, so some items may be batched incorrectly if they have
never been in the camera's view or if they don't use the Material
abstraction.
in particular, shadow views render more meshes than the main camera, and
currently batch some meshes where the object has never entered the
camera view together. this is quite hard to trigger, but should occur in
a scene with out-of-view alpha-mask materials (so that the material
instance actually affects the shadow) in the path of a light.
this is also a footgun for custom pipelines: failing to set the
material_bind_group_id will result in all meshes being batched together
and all using the closest/furthest material to the camera (depending on
sort order).
## Solution
- queue_shadows now sets the material_bind_group_id correctly
- `MeshPipeline` doesn't attempt to batch meshes if the
material_bind_group_id has not been set. custom pipelines still need to
set this field to take advantage of batching, but will at least render
correctly if it is not set
# Objective
sysinfo was updated to 0.30 in #11071. Ever since then the `cpu` field
of the `SystemInfo` struct that gets printed every time one starts an
bevy app has been empty. This is because the following part of the
sysinfo migration guide was overlooked:
---
### `Cpu` changes
Information like `Cpu::brand`, `Cpu::vendor_id` or `Cpu::frequency` are
not set on the "global" CPU.
---
## Solution
- Get the CPU brand information from a specific CPU instead. In this
case, just choose the first one. It's theoretically possible for
different CPUs to have different names, but in practice this doesn't
really happen I think. Even Intel's newer hybrid processors use a
uniform name for all CPUs in my experience.
- We can use this opportunity to also update our `sysinfo::System`
initialization here to only fetch the information we're interested in.
Make the renamings/changes regarding texture atlases a bit less
confusing by calling `TextureAtlasLayout` a layout, not a texture atlas.
Co-authored-by: Alice Cecile <alice.i.cecile@gmail.com>
# Objective
It can sometimes be useful to combine several meshes into one. This
allows constructing more complex meshes out of simple primitives without
needing to use a 3D modeling program or entity hierarchies.
This could also be used internally to increase code reuse by using
existing mesh generation logic for e.g. circles and using that in
cylinder mesh generation logic to add the top and bottom of the
cylinder.
**Note**: This is *not* implementing CSGs (Constructive Solid Geometry)
or any boolean operations, as that is much more complex. This is simply
adding the mesh data of another mesh to a mesh.
## Solution
Add a `merge` method to `Mesh`. It appends the vertex attributes and
indices of `other` to `self`, resulting in a `Mesh` that is the
combination of the two.
For example, you could do this:
```rust
let mut cuboid = Mesh::from(shape::Box::default());
let mut cylinder = Mesh::from(shape::Cylinder::default());
let mut torus = Mesh::from(shape::Torus::default());
cuboid.merge(cylinder);
cuboid.merge(torus);
```
This would result in `cuboid` being a `Mesh` that also has the cylinder
mesh and torus mesh. In this case, they would just be placed on top of
each other, but by utilizing #11454 we can transform the cylinder and
torus to get a result like this:
https://github.com/bevyengine/bevy/assets/57632562/557402c6-b896-4aba-bd95-312e7d1b5238
This is just a single entity and a single mesh.
# Objective
- Fixes#11782.
## Solution
- Remove the run condition for `apply_global_wireframe_material`, since
it prevent detecting when meshes are added or the `NoWireframe` marker
component is removed from an entity. Alternatively this could be done by
using a run condition like "added `Handle<Mesh>` or removed
`NoWireframe` or `WireframeConfig` changed" but this seems less clear to
me than directly letting the queries on
`apply_global_wireframe_material` do the filtering.
# Objective
-
[`crossbeam::scope`](https://docs.rs/crossbeam/latest/crossbeam/fn.scope.html)
is soft-deprecated in favor of the standard library's implementation.
## Solution
- Replace reference in `TaskPool`'s docs to mention `std:🧵:scope`
instead.
# Objective
- Fixes#11695
## Solution
- Added `delta: Option<Vec2>` to `bevy_window::CursorMoved`. `delta` is
an `Option` because the `CursorMoved` event does get fired when the
cursor was outside the window area in the last frame. In that case there
is no cursor position from the last frame to compare with the current
cursor position.
---
## Changelog
- Added `delta: Option<Vec2>` to `bevy_window::CursorMoved`.
## Migration Guide
- You need to add `delta` to any manually created `CursorMoved` struct.
---------
Co-authored-by: Kanabenki <lucien.menassol@gmail.com>
Co-authored-by: James Liu <contact@jamessliu.com>
# Objective
- Fixes#11638
- See
[here](https://github.com/bevyengine/bevy/issues/11638#issuecomment-1920508465)
for details on the cause of this issue.
## Solution
- Modified `AssetLoaders` to capture possibility of multiple
`AssetLoader` registrations operating on the same `Asset` type, but
different extensions.
- Added an algorithm which will attempt to resolve via `AssetLoader`
name, then `Asset` type, then by extension. If at any point multiple
loaders fit a particular criteria, the next criteria is used as a tie
breaker.
# Objective
At the start of every schedule run, there's currently a guaranteed piece
of overhead as the async executor spawns the MultithreadeExecutor task
onto one of the ComputeTaskPool threads.
## Solution
Poll the executor once to immediately schedule systems without waiting
for the async executor, then spawn the task if and only if the executor
does not immediately terminate.
On a similar note, having the executor task immediately start executing
a system in the same async task might yield similar results over a
broader set of cases. However, this might be more involved, and may need
a solution like #8304.
# Objective
When applying a command, we currently use double indirection for the
world reference `&mut Option<&mut World>`. Since this is used across a
`fn` pointer boundary, this can't get optimized away.
## Solution
Reborrow the world reference and pass `Option<&mut World>` instead.
# Objective
Scheduling low cost systems has significant overhead due to task pool
contention and the extra machinery to schedule and run them. Following
the example of #7728, `asset_events` is good example of this kind of
system, where there is no work to be done when there are no queued asset
events.
## Solution
Put a run condition on it that checks if there are any queued events.
## Performance
Tested against `many_foxes`, we can see a slight improvement in the
total time spent in `UpdateAssets`. Also noted much less volatility due
to not being at the whim of the OS thread scheduler.
---------
Co-authored-by: Alice Cecile <alice.i.cecile@gmail.com>
# Objective
The code in `bevy_reflect_derive` could use some cleanup.
## Solution
Took some of the changes in #11659 to create a dedicated PR for cleaning
up the field and container attribute logic.
#### Updated Naming
I renamed `ReflectTraits` and `ReflectFieldAttr` to
`ContainerAttributes` and `FieldAttributes`, respectively. I think these
are clearer.
#### Updated Parsing
##### Readability
The parsing logic wasn't too bad before, but it was getting difficult to
read. There was some duplicated logic between `Meta::List` and
`Meta::Path` attributes. Additionally, all the logic was kept inside a
large method.
To simply things, I replaced the nested meta parsing with `ParseStream`
parsing. In my opinion, this is easier to follow since it breaks up the
large match statement into a small set of single-line if statements,
where each if-block contains a single call to the appropriate attribute
parsing method.
##### Flexibility
On top of the added simplicity, this also makes our attribute parsing
much more flexible. It allows us to more elegantly handle custom where
clauses (i.e. `#[reflect(where T: Foo)]`) and it opens the door for more
non-standard attribute syntax (e.g. #11659).
##### Errors
This also allows us to automatically provide certain errors when
parsing. For example, since we can use `stream.lookahead1()`, we get
errors like the following for free:
```
error: expected one of: `ignore`, `skip_serializing`, `default`
--> crates/bevy_reflect/src/lib.rs:1988:23
|
1988 | #[reflect(foo)]
| ^^^
```
---
## Changelog
> [!note]
> All changes are internal to `bevy_reflect_derive` and should not
affect the public API[^1].
- Renamed `ReflectTraits` to `ContainerAttributes`
- Renamed `ReflectMeta::traits` to `ReflectMeta::attrs`
- Renamed `ReflectFieldAttr` to `FieldAttributes`
- Updated parsing logic for field/container attribute parsing
- Now uses a `ParseStream` directly instead of nested meta parsing
- General code cleanup of the field/container attribute modules for
`bevy_reflect_derive`
[^1]: Does not include errors, which may look slightly different.
---------
Co-authored-by: Alice Cecile <alice.i.cecile@gmail.com>
# Objective
Bevy's change detection functionality is invaluable for writing robust
apps, but it only works in the context of systems and exclusive systems.
Oftentimes it is necessary to detect changes made in earlier code
without having to place the code in separate systems, but it is not
currently possible to do so since there is no way to set the value of
`World::last_change_tick`.
`World::clear_trackers` allows you to update the change tick, but this
has unintended side effects, since it irreversibly affects the behavior
of change and removal detection for the entire app.
## Solution
Add a method `World::last_change_tick_scope`. This allows you to set
`last_change_tick` to a specific value for a region of code. To ensure
that misuse doesn't break unrelated functions, we restore the world's
original change tick at the end of the provided scope.
### Example
A function that uses this to run an update loop repeatedly, allowing
each iteration of the loop to react to changes made in the previous loop
iteration.
```rust
fn update_loop(
world: &mut World,
mut update_fn: impl FnMut(&mut World) -> std::ops::ControlFlow<()>,
) {
let mut last_change_tick = world.last_change_tick();
// Repeatedly run the update function until it requests a break.
loop {
// Update once.
let control_flow = world.last_change_tick_scope(last_change_tick, |world| {
update_fn(world)
});
// End the loop when the closure returns `ControlFlow::Break`.
if control_flow.is_break() {
break;
}
// Increment the change tick so the next update can detect changes from this update.
last_change_tick = world.change_tick();
world.increment_change_tick();
}
}
```
---
## Changelog
+ Added `World::last_change_tick_scope`, which allows you to specify the
reference for change detection within a certain scope.
# Objective
- Part of #11590.
## Solution
- Fix `unsafe_op_in_unsafe_fn` for `bevy_dynamic_plugin`.
---
## Changelog
- Added further restrictions to the safety requirements of
`bevy_dynamic_plugin::dynamically_load_plugin`.
---
I had a few issues, specifically with the safety comment on
`dynamically_load_plugin`. There are three different unsafe functions
called within the function body, and they all need their own
justification / message.
Also, would it be unsound to call `dynamically_load_plugin` multiple
times on the same file? I feel the documentation needs to be more clear.
# Objective
Reduce the size of `bevy_utils`
(https://github.com/bevyengine/bevy/issues/11478)
## Solution
Move `EntityHash` related types into `bevy_ecs`. This also allows us
access to `Entity`, which means we no longer need `EntityHashMap`'s
first generic argument.
---
## Changelog
- Moved `bevy::utils::{EntityHash, EntityHasher, EntityHashMap,
EntityHashSet}` into `bevy::ecs::entity::hash` .
- Removed `EntityHashMap`'s first generic argument. It is now hardcoded
to always be `Entity`.
## Migration Guide
- Uses of `bevy::utils::{EntityHash, EntityHasher, EntityHashMap,
EntityHashSet}` now have to be imported from `bevy::ecs::entity::hash`.
- Uses of `EntityHashMap` no longer have to specify the first generic
parameter. It is now hardcoded to always be `Entity`.
# Objective
Loading some textures from the days of yonder give me errors cause the
mipmap level is 0
## Solution
Set a minimum of 1
## Changelog
Make mipmap level at least 1
# Objective
Use `GamepadButtonType` with library that requires `Ord`.
## Motivation
`KeyCode` derives `Ord` that I'm using with a trie for recognizing
[input
sequences](https://github.com/shanecelis/bevy-input-sequence/tree/trie).
I would like to do the same for `GamepadButtonType` but am stymied by it
not deriving `Ord`.
## Solution
This PR add derivations PartialOrd and Ord for `GamepadButtonType`.
## Workaround
If deriving `Ord` is not possible, I'd be happy to know how I might
coerce `GamepadButtonType` into a `u32` or something else that is `Ord`,
so I can wrap `GamepadButtonType` in a newtype. I suppose serializing
with serde may work or reflect?
# Objective
- Get rid of unwraps in winit fullscreen handling code, which are the
source of some crashes.
- Fix#11275
## Solution
- Replace the unwraps with warnings. Ignore the fullscreen request, do
nothing instead.
# Objective
It would be useful to be able to inspect a `QueryState`'s accesses so we
can detect when the data it accesses changes without having to iterate
it. However there are two things preventing this:
* These accesses are unnecessarily encapsulated.
* `Has<T>` indirectly accesses `T`, but does not register it.
## Solution
* Expose accesses and matches used by `QueryState`.
* Add the notion of "archetypal" accesses, which are not accessed
directly, but whose presence in an archetype affects a query result.
---------
Co-authored-by: Alice Cecile <alice.i.cecile@gmail.com>
# Objective
I want to keep track of despawned entities.
I am aware of
[`RemovedComponents`](https://docs.rs/bevy/0.12.1/bevy/ecs/prelude/struct.RemovedComponents.html).
However, the docs don't explicitly mention that despawned entities are
also included in this event iterator.
I searched through the bevy tests to find `removal_tracking` in
`crates/bevy_ecs/src/system/mod.rs` that confirmed the behavior:
```rust
...
assert_eq!(
removed_i32.read().collect::<Vec<_>>(),
&[despawned.0],
"despawning causes the correct entity to show up in the 'RemovedComponent' system parameter."
);
...
```
## Solution
- Explicitly mention this behavior in docs.
This fixes a `FIXME` in `extract_meshes` and results in a performance
improvement.
As a result of this change, meshes in the render world might not be
attached to entities anymore. Therefore, the `entity` parameter to
`RenderCommand::render()` is now wrapped in an `Option`. Most
applications that use the render app's ECS can simply unwrap the
`Option`.
Note that for now sprites, gizmos, and UI elements still use the render
world as usual.
## Migration guide
* For efficiency reasons, some meshes in the render world may not have
corresponding `Entity` IDs anymore. As a result, the `entity` parameter
to `RenderCommand::render()` is now wrapped in an `Option`. Custom
rendering code may need to be updated to handle the case in which no
`Entity` exists for an object that is to be rendered.
# Objective
`bevy_utils` only requires aHash 0.8.3, which is broken on Rust 1.7.6:
```
error: could not compile `ahash` (lib) due to 1 previous error
error[E0635]: unknown feature `stdsimd`
```
See https://github.com/tkaitchuck/aHash/issues/200
This is fixed in aHash 0.8.7, so require at least that version
(Cargo.lock is already up to date).
# Objective
The deprecation message of `bevy::render::mesh::shape::Quad` says that
you should use `bevy_math`'s `Quad` instead. But it doesn't exist.
## Solution
Mention the correct primitive: `Rectangle`
> Follow up to #11600 and #10588
@mockersf expressed some [valid
concerns](https://github.com/bevyengine/bevy/pull/11600#issuecomment-1932796498)
about the current system this PR attempts to fix:
The `ComputedTextureSlices` reacts to asset change in both `bevy_sprite`
and `bevy_ui`, meaning that if the `ImageScaleMode` is inserted by
default in the bundles, we will iterate through most 2d items every time
an asset is updated.
# Solution
- `ImageScaleMode` only has two variants: `Sliced` and `Tiled`. I
removed the `Stretched` default
- `ImageScaleMode` is no longer part of any bundle, but the relevant
bundles explain that this additional component can be inserted
This way, the *absence* of `ImageScaleMode` means the image will be
stretched, and its *presence* will include the entity to the various
slicing systems
Optional components in bundles would make this more straigthfoward
# Additional work
Should I add new bundles with the `ImageScaleMode` component ?
# Objective
The screenshots generated by the `screenshot` example need to be removed
manually before commiting changes.
## Solution
`.gitignore` the screenshots generated by the `screenshot` example. I
also added the `**/` prefix so the screenshots get ignored regardless of
whether of where the example is run from.
# Objective
- Encoding many GPU commands (such as in a renderpass with many draws,
such as the main opaque pass) onto a `wgpu::CommandEncoder` is very
expensive, and takes a long time.
- To improve performance, we want to perform the command encoding for
these heavy passes in parallel.
## Solution
- `RenderContext` can now queue up "command buffer generation tasks"
which are closures that will generate a command buffer when called.
- When finalizing the render context to produce the final list of
command buffers, these tasks are run in parallel on the
`ComputeTaskPool` to produce their corresponding command buffers.
- The general idea is that the node graph will run in serial, but in a
node, instead of doing rendering work, you can add tasks to do render
work in parallel with other node's tasks that get ran at the end of the
graph execution.
## Nodes Parallelized
- `MainOpaquePass3dNode`
- `PrepassNode`
- `DeferredGBufferPrepassNode`
- `ShadowPassNode` (One task per view)
## Future Work
- For large number of draws calls, might be worth further subdividing
passes into 2+ tasks.
- Extend this to UI, 2d, transparent, and transmissive nodes?
- Needs testing - small command buffers are inefficient - it may be
worth reverting to the serial command encoder usage for render phases
with few items.
- All "serial" (traditional) rendering work must finish before parallel
rendering tasks (the new stuff) can start to run.
- There is still only one submission to the graphics queue at the end of
the graph execution. There is still no ability to submit work earlier.
## Performance Improvement
Thanks to @Elabajaba for testing on Bistro.
TLDR: Without shadow mapping, this PR has no impact. _With_ shadow
mapping, this PR gives **~40 more fps** than main.
---
## Changelog
- `MainOpaquePass3dNode`, `PrepassNode`, `DeferredGBufferPrepassNode`,
and each shadow map within `ShadowPassNode` are now encoded in parallel,
giving _greatly_ increased CPU performance, mainly when shadow mapping
is enabled.
- Does not work on WASM or AMD+Windows+Vulkan.
- Added `RenderContext::add_command_buffer_generation_task()`.
- `RenderContext::new()` now takes adapter info
- Some render graph and Node related types and methods now have
additional lifetime constraints.
## Migration Guide
`RenderContext::new()` now takes adapter info
- Some render graph and Node related types and methods now have
additional lifetime constraints.
---------
Co-authored-by: Elabajaba <Elabajaba@users.noreply.github.com>
Co-authored-by: François <mockersf@gmail.com>
# Objective
While profiling around to validate the results of #9172, I noticed that
`present_frames` can take a significant amount of time. Digging into the
cause, it seems like we're creating a new `QueryState` from scratch
every frame. This involves scanning the entire World's metadata instead
of just updating its view of the world.
## Solution
Use a `SystemState` argument to cache the `QueryState` to avoid this
construction cost.
## Performance
Against `many_foxes`, this seems to cut the time spent in
`present_frames` by nearly almost 2x. Yellow is this PR, red is main.
# Objective
#11431 and #11688 implemented meshing support for Bevy's new geometric
primitives. The next step is to deprecate the shapes in
`bevy_render::mesh::shape` and to later remove them completely for 0.14.
## Solution
Deprecate the shapes and reduce code duplication by utilizing the
primitive meshing API for the old shapes where possible.
Note that some shapes have behavior that can't be exactly reproduced
with the new primitives yet:
- `Box` is more of an AABB with min/max extents
- `Plane` supports a subdivision count
- `Quad` has a `flipped` property
These types have not been changed to utilize the new primitives yet.
---
## Changelog
- Deprecated all shapes in `bevy_render::mesh::shape`
- Changed all examples to use new primitives for meshing
## Migration Guide
Bevy has previously used rendering-specific types like `UVSphere` and
`Quad` for primitive mesh shapes. These have now been deprecated to use
the geometric primitives newly introduced in version 0.13.
Some examples:
```rust
let before = meshes.add(shape::Box::new(5.0, 0.15, 5.0));
let after = meshes.add(Cuboid::new(5.0, 0.15, 5.0));
let before = meshes.add(shape::Quad::default());
let after = meshes.add(Rectangle::default());
let before = meshes.add(shape::Plane::from_size(5.0));
// The surface normal can now also be specified when using `new`
let after = meshes.add(Plane3d::default().mesh().size(5.0, 5.0));
let before = meshes.add(
Mesh::try_from(shape::Icosphere {
radius: 0.5,
subdivisions: 5,
})
.unwrap(),
);
let after = meshes.add(Sphere::new(0.5).mesh().ico(5).unwrap());
```
# Objective
- Try not to drop the render world on the render thread, and drop the
main world after the render world.
- The render world has a drop check that will panic if it is dropped off
the main thread.
## Solution
- Keep track of where the render world is and wait for it to come back
when the channel resource is dropped.
---
## Changelog
- Wait for the render world when the main world is dropped.
## Migration Guide
- If you were using the pipelined rendering channels,
`MainToRenderAppSender` and `RenderToMainAppReceiver`, they have been
combined into the single resource `RenderAppChannels`.
---------
Co-authored-by: Alice Cecile <alice.i.cecile@gmail.com>
Co-authored-by: Friz64 <friz64@protonmail.com>
> Follow up to #10588
> Closes#11749 (Supersedes #11756)
Enable Texture slicing for the following UI nodes:
- `ImageBundle`
- `ButtonBundle`
<img width="739" alt="Screenshot 2024-01-29 at 13 57 43"
src="https://github.com/bevyengine/bevy/assets/26703856/37675681-74eb-4689-ab42-024310cf3134">
I also added a collection of `fantazy-ui-borders` from
[Kenney's](www.kenney.nl) assets, with the appropriate license (CC).
If it's a problem I can use the same textures as the `sprite_slice`
example
# Work done
Added the `ImageScaleMode` component to the targetted bundles, most of
the logic is directly reused from `bevy_sprite`.
The only additional internal component is the UI specific
`ComputedSlices`, which does the same thing as its spritee equivalent
but adapted to UI code.
Again the slicing is not compatible with `TextureAtlas`, it's something
I need to tackle more deeply in the future
# Fixes
* [x] I noticed that `TextureSlicer::compute_slices` could infinitely
loop if the border was larger that the image half extents, now an error
is triggered and the texture will fallback to being stretched
* [x] I noticed that when using small textures with very small *tiling*
options we could generate hundred of thousands of slices. Now I set a
minimum size of 1 pixel per slice, which is already ridiculously small,
and a warning will be sent at runtime when slice count goes above 1000
* [x] Sprite slicing with `flip_x` or `flip_y` would give incorrect
results, correct flipping is now supported to both sprites and ui image
nodes thanks to @odecay observation
# GPU Alternative
I create a separate branch attempting to implementing 9 slicing and
tiling directly through the `ui.wgsl` fragment shader. It works but
requires sending more data to the GPU:
- slice border
- tiling factors
And more importantly, the actual quad *scale* which is hard to put in
the shader with the current code, so that would be for a later iteration
[`ScheduleLabel`] derive macro uses "ScheduleName" as the trait name by
mistake. This only affects the error message when a user tries to use
the derive macro on a union type. No other code is affected.
# Objective
- This aims to fix#11755
- After #10812 some pipeline compilation can take more time than before
and all call to `get_render_pipeline` should check the result.
## Solution
- Check `get_render_pipeline` call result for msaa_writeback
- I checked that no other call to `get_render_pipeline` in bevy code
base is missng the checking on the result.
Don't try to create a uniform buffer for light probes if there are no
views.
Fixes the panic on examples that have no views, such as
`touch_input_events`.
# Objective
Fix https://github.com/bevyengine/bevy/issues/11657
## Solution
Add a `ReflectKind` enum, add `Reflect::reflect_kind` which returns a
`ReflectKind`, and add `kind` method implementions to `ReflectRef`,
`ReflectMut`, and `ReflectOwned`, which returns a `ReflectKind`.
I also changed `AccessError` to use this new struct instead of it's own
`TypeKind` struct.
---
## Changelog
- Added `ReflectKind`, an enumeration over the kinds of a reflected type
without its data.
- Added `Reflect::reflect_kind` (with default implementation)
- Added implementation for the `kind` method on `ReflectRef`,
`ReflectMut`, and `ReflectOwned` which gives their kind without any
information, as a `ReflectKind`
# Objective
- Fixes#11740
## Solution
- Turned `Mesh::set_indices` into `Mesh::insert_indices` and added
related methods for completeness.
---
## Changelog
- Replaced `Mesh::set_indices(indices: Option<Indices>)` with
`Mesh::insert_indices(indices: Indices)`
- Replaced `Mesh::with_indices(indices: Option<Indices>)` with
`Mesh::with_inserted_indices(indices: Indices)` and
`Mesh::with_removed_indices()` mirroring the API for inserting /
removing attributes.
- Updated the examples and internal uses of the APIs described above.
## Migration Guide
- Use `Mesh::insert_indices` or `Mesh::with_inserted_indices` instead of
`Mesh::set_indices` / `Mesh::with_indices`.
- If you have passed `None` to `Mesh::set_indices` or
`Mesh::with_indices` you should use `Mesh::remove_indices` or
`Mesh::with_removed_indices` instead.
---------
Co-authored-by: François <mockersf@gmail.com>
# Objective
- System `create_surfaces` needs to happen before `prepare_windows` or
we lose one frame at startup
## Solution
- Specify the ordering, remove the set as it doesn't mean anything there
# Objective
Bevy could benefit from *irradiance volumes*, also known as *voxel
global illumination* or simply as light probes (though this term is not
preferred, as multiple techniques can be called light probes).
Irradiance volumes are a form of baked global illumination; they work by
sampling the light at the centers of each voxel within a cuboid. At
runtime, the voxels surrounding the fragment center are sampled and
interpolated to produce indirect diffuse illumination.
## Solution
This is divided into two sections. The first is copied and pasted from
the irradiance volume module documentation and describes the technique.
The second part consists of notes on the implementation.
### Overview
An *irradiance volume* is a cuboid voxel region consisting of
regularly-spaced precomputed samples of diffuse indirect light. They're
ideal if you have a dynamic object such as a character that can move
about
static non-moving geometry such as a level in a game, and you want that
dynamic object to be affected by the light bouncing off that static
geometry.
To use irradiance volumes, you need to precompute, or *bake*, the
indirect
light in your scene. Bevy doesn't currently come with a way to do this.
Fortunately, [Blender] provides a [baking tool] as part of the Eevee
renderer, and its irradiance volumes are compatible with those used by
Bevy.
The [`bevy-baked-gi`] project provides a tool, `export-blender-gi`, that
can
extract the baked irradiance volumes from the Blender `.blend` file and
package them up into a `.ktx2` texture for use by the engine. See the
documentation in the `bevy-baked-gi` project for more details as to this
workflow.
Like all light probes in Bevy, irradiance volumes are 1×1×1 cubes that
can
be arbitrarily scaled, rotated, and positioned in a scene with the
[`bevy_transform::components::Transform`] component. The 3D voxel grid
will
be stretched to fill the interior of the cube, and the illumination from
the
irradiance volume will apply to all fragments within that bounding
region.
Bevy's irradiance volumes are based on Valve's [*ambient cubes*] as used
in
*Half-Life 2* ([Mitchell 2006], slide 27). These encode a single color
of
light from the six 3D cardinal directions and blend the sides together
according to the surface normal.
The primary reason for choosing ambient cubes is to match Blender, so
that
its Eevee renderer can be used for baking. However, they also have some
advantages over the common second-order spherical harmonics approach:
ambient cubes don't suffer from ringing artifacts, they are smaller (6
colors for ambient cubes as opposed to 9 for spherical harmonics), and
evaluation is faster. A smaller basis allows for a denser grid of voxels
with the same storage requirements.
If you wish to use a tool other than `export-blender-gi` to produce the
irradiance volumes, you'll need to pack the irradiance volumes in the
following format. The irradiance volume of resolution *(Rx, Ry, Rz)* is
expected to be a 3D texture of dimensions *(Rx, 2Ry, 3Rz)*. The
unnormalized
texture coordinate *(s, t, p)* of the voxel at coordinate *(x, y, z)*
with
side *S* ∈ *{-X, +X, -Y, +Y, -Z, +Z}* is as follows:
```text
s = x
t = y + ⎰ 0 if S ∈ {-X, -Y, -Z}
⎱ Ry if S ∈ {+X, +Y, +Z}
⎧ 0 if S ∈ {-X, +X}
p = z + ⎨ Rz if S ∈ {-Y, +Y}
⎩ 2Rz if S ∈ {-Z, +Z}
```
Visually, in a left-handed coordinate system with Y up, viewed from the
right, the 3D texture looks like a stacked series of voxel grids, one
for
each cube side, in this order:
| **+X** | **+Y** | **+Z** |
| ------ | ------ | ------ |
| **-X** | **-Y** | **-Z** |
A terminology note: Other engines may refer to irradiance volumes as
*voxel
global illumination*, *VXGI*, or simply as *light probes*. Sometimes
*light
probe* refers to what Bevy calls a reflection probe. In Bevy, *light
probe*
is a generic term that encompasses all cuboid bounding regions that
capture
indirect illumination, whether based on voxels or not.
Note that, if binding arrays aren't supported (e.g. on WebGPU or WebGL
2),
then only the closest irradiance volume to the view will be taken into
account during rendering.
[*ambient cubes*]:
https://advances.realtimerendering.com/s2006/Mitchell-ShadingInValvesSourceEngine.pdf
[Mitchell 2006]:
https://advances.realtimerendering.com/s2006/Mitchell-ShadingInValvesSourceEngine.pdf
[Blender]: http://blender.org/
[baking tool]:
https://docs.blender.org/manual/en/latest/render/eevee/render_settings/indirect_lighting.html
[`bevy-baked-gi`]: https://github.com/pcwalton/bevy-baked-gi
### Implementation notes
This patch generalizes light probes so as to reuse as much code as
possible between irradiance volumes and the existing reflection probes.
This approach was chosen because both techniques share numerous
similarities:
1. Both irradiance volumes and reflection probes are cuboid bounding
regions.
2. Both are responsible for providing baked indirect light.
3. Both techniques involve presenting a variable number of textures to
the shader from which indirect light is sampled. (In the current
implementation, this uses binding arrays.)
4. Both irradiance volumes and reflection probes require gathering and
sorting probes by distance on CPU.
5. Both techniques require the GPU to search through a list of bounding
regions.
6. Both will eventually want to have falloff so that we can smoothly
blend as objects enter and exit the probes' influence ranges. (This is
not implemented yet to keep this patch relatively small and reviewable.)
To do this, we generalize most of the methods in the reflection probes
patch #11366 to be generic over a trait, `LightProbeComponent`. This
trait is implemented by both `EnvironmentMapLight` (for reflection
probes) and `IrradianceVolume` (for irradiance volumes). Using a trait
will allow us to add more types of light probes in the future. In
particular, I highly suspect we will want real-time reflection planes
for mirrors in the future, which can be easily slotted into this
framework.
## Changelog
> This section is optional. If this was a trivial fix, or has no
externally-visible impact, you can delete this section.
### Added
* A new `IrradianceVolume` asset type is available for baked voxelized
light probes. You can bake the global illumination using Blender or
another tool of your choice and use it in Bevy to apply indirect
illumination to dynamic objects.
# Objective
Split up from #11007, fixing most of the remaining work for #10569.
Implement `Meshable` for `Cuboid`, `Sphere`, `Cylinder`, `Capsule`,
`Torus`, and `Plane3d`. This covers all shapes that Bevy has mesh
structs for in `bevy_render::mesh::shapes`.
`Cone` and `ConicalFrustum` are new shapes, so I can add them in a
follow-up, or I could just add them here directly if that's preferrable.
## Solution
Implement `Meshable` for `Cuboid`, `Sphere`, `Cylinder`, `Capsule`,
`Torus`, and `Plane3d`.
The logic is mostly just a copy of the the existing `bevy_render`
shapes, but `Plane3d` has a configurable surface normal that affects the
orientation. Some property names have also been changed to be more
consistent.
The default values differ from the old shapes to make them a bit more
logical:
- Spheres now have a radius of 0.5 instead of 1.0. The default capsule
is equivalent to the default cylinder with the sphere's halves glued on.
- The inner and outer radius of the torus are now 0.5 and 1.0 instead of
0.5 and 1.5 (i.e. the new minor and major radii are 0.25 and 0.75). It's
double the width of the default cuboid, half of its height, and the
default sphere matches the size of the hole.
- `Cuboid` is 1x1x1 by default unlike the dreaded `Box` which is 2x1x1.
Before, with "old" shapes:
Now, with primitive meshing:
I only changed the `3d_shapes` example to use primitives for now. I can
change them all in this PR or a follow-up though, whichever way is
preferrable.
### Sphere API
Spheres have had separate `Icosphere` and `UVSphere` structs, but with
primitives we only have one `Sphere`.
We need to handle this with builders:
```rust
// Existing structs
let ico = Mesh::try_from(Icophere::default()).unwrap();
let uv = Mesh::from(UVSphere::default());
// Primitives
let ico = Sphere::default().mesh().ico(5).unwrap();
let uv = Sphere::default().mesh().uv(32, 18);
```
We could add methods on `Sphere` directly to skip calling `.mesh()`.
I also added a `SphereKind` enum that can be used with the `kind`
method:
```rust
let ico = Sphere::default()
.mesh()
.kind(SphereKind::Ico { subdivisions: 8 })
.build();
```
The default mesh for a `Sphere` is an icosphere with 5 subdivisions
(like the default `Icosphere`).
---
## Changelog
- Implement `Meshable` and `Default` for `Cuboid`, `Sphere`, `Cylinder`,
`Capsule`, `Torus`, and `Plane3d`
- Use primitives in `3d_shapes` example
---------
Co-authored-by: Alice Cecile <alice.i.cecile@gmail.com>
# Objective
- The exported hashtypes are just re-exports from hashbrown, we want to
drop that dependency and (in the future) let the user import their own
choice.
- Fixes#11717
## Solution
- Adding a deprecated tag on the re-exports, so in future releases these
can be safely removed.
