This commit moves the front end of the rendering pipeline to a retained
model when GPU preprocessing is in use (i.e. by default, except in
constrained environments). `RenderMeshInstance` and `MeshUniformData`
are stored from frame to frame and are updated only for the entities
that changed state. This was rather tricky and requires some careful
surgery to keep the data valid in the case of removals.
This patch is built on top of Bevy's change detection. Generally, this
worked, except that `ViewVisibility` isn't currently properly tracked.
Therefore, this commit adds proper change tracking for `ViewVisibility`.
Doing this required adding a new system that runs after all
`check_visibility` invocations, as no single `check_visibility`
invocation has enough global information to detect changes.
On the Bistro exterior scene, with all textures forced to opaque, this
patch improves steady-state `extract_meshes_for_gpu_building` from
93.8us to 34.5us and steady-state `collect_meshes_for_gpu_building` from
195.7us to 4.28us. Altogether this constitutes an improvement from 290us
to 38us, which is a 7.46x speedup.
This patch is only lightly tested and shouldn't land before 0.15 is
released anyway, so I'm releasing it as a draft.
# Objective
- Required by #16622 due to differing implementations of `System` by
`FunctionSystem` and `ExclusiveFunctionSystem`.
- Optimize the memory usage of instances of `apply_deferred` in system
schedules.
## Solution
By changing `apply_deferred` from being an ordinary system that ends up
as an `ExclusiveFunctionSystem`, and instead into a ZST struct that
implements `System` manually, we save ~320 bytes per instance of
`apply_deferred` in any schedule.
## Testing
- All current tests pass.
---
## Migration Guide
- If you were previously calling the special `apply_deferred` system via
`apply_deferred(world)`, don't.
# Objective
- Fixes#16208
## Solution
- Added an associated type to `Component`, `Mutability`, which flags
whether a component is mutable, or immutable. If `Mutability= Mutable`,
the component is mutable. If `Mutability= Immutable`, the component is
immutable.
- Updated `derive_component` to default to mutable unless an
`#[component(immutable)]` attribute is added.
- Updated `ReflectComponent` to check if a component is mutable and, if
not, panic when attempting to mutate.
## Testing
- CI
- `immutable_components` example.
---
## Showcase
Users can now mark a component as `#[component(immutable)]` to prevent
safe mutation of a component while it is attached to an entity:
```rust
#[derive(Component)]
#[component(immutable)]
struct Foo {
// ...
}
```
This prevents creating an exclusive reference to the component while it
is attached to an entity. This is particularly powerful when combined
with component hooks, as you can now fully track a component's value,
ensuring whatever invariants you desire are upheld. Before this would be
done my making a component private, and manually creating a `QueryData`
implementation which only permitted read access.
<details>
<summary>Using immutable components as an index</summary>
```rust
/// This is an example of a component like [`Name`](bevy::prelude::Name), but immutable.
#[derive(Clone, Copy, PartialEq, Eq, PartialOrd, Ord, Hash, Component)]
#[component(
immutable,
on_insert = on_insert_name,
on_replace = on_replace_name,
)]
pub struct Name(pub &'static str);
/// This index allows for O(1) lookups of an [`Entity`] by its [`Name`].
#[derive(Resource, Default)]
struct NameIndex {
name_to_entity: HashMap<Name, Entity>,
}
impl NameIndex {
fn get_entity(&self, name: &'static str) -> Option<Entity> {
self.name_to_entity.get(&Name(name)).copied()
}
}
fn on_insert_name(mut world: DeferredWorld<'_>, entity: Entity, _component: ComponentId) {
let Some(&name) = world.entity(entity).get::<Name>() else {
unreachable!()
};
let Some(mut index) = world.get_resource_mut::<NameIndex>() else {
return;
};
index.name_to_entity.insert(name, entity);
}
fn on_replace_name(mut world: DeferredWorld<'_>, entity: Entity, _component: ComponentId) {
let Some(&name) = world.entity(entity).get::<Name>() else {
unreachable!()
};
let Some(mut index) = world.get_resource_mut::<NameIndex>() else {
return;
};
index.name_to_entity.remove(&name);
}
// Setup our name index
world.init_resource::<NameIndex>();
// Spawn some entities!
let alyssa = world.spawn(Name("Alyssa")).id();
let javier = world.spawn(Name("Javier")).id();
// Check our index
let index = world.resource::<NameIndex>();
assert_eq!(index.get_entity("Alyssa"), Some(alyssa));
assert_eq!(index.get_entity("Javier"), Some(javier));
// Changing the name of an entity is also fully capture by our index
world.entity_mut(javier).insert(Name("Steven"));
// Javier changed their name to Steven
let steven = javier;
// Check our index
let index = world.resource::<NameIndex>();
assert_eq!(index.get_entity("Javier"), None);
assert_eq!(index.get_entity("Steven"), Some(steven));
```
</details>
Additionally, users can use `Component<Mutability = ...>` in trait
bounds to enforce that a component _is_ mutable or _is_ immutable. When
using `Component` as a trait bound without specifying `Mutability`, any
component is applicable. However, methods which only work on mutable or
immutable components are unavailable, since the compiler must be
pessimistic about the type.
## Migration Guide
- When implementing `Component` manually, you must now provide a type
for `Mutability`. The type `Mutable` provides equivalent behaviour to
earlier versions of `Component`:
```rust
impl Component for Foo {
type Mutability = Mutable;
// ...
}
```
- When working with generic components, you may need to specify that
your generic parameter implements `Component<Mutability = Mutable>`
rather than `Component` if you require mutable access to said component.
- The entity entry API has had to have some changes made to minimise
friction when working with immutable components. Methods which
previously returned a `Mut<T>` will now typically return an
`OccupiedEntry<T>` instead, requiring you to add an `into_mut()` to get
the `Mut<T>` item again.
## Draft Release Notes
Components can now be made immutable while stored within the ECS.
Components are the fundamental unit of data within an ECS, and Bevy
provides a number of ways to work with them that align with Rust's rules
around ownership and borrowing. One part of this is hooks, which allow
for defining custom behavior at key points in a component's lifecycle,
such as addition and removal. However, there is currently no way to
respond to _mutation_ of a component using hooks. The reasons for this
are quite technical, but to summarize, their addition poses a
significant challenge to Bevy's core promises around performance.
Without mutation hooks, it's relatively trivial to modify a component in
such a way that breaks invariants it intends to uphold. For example, you
can use `core::mem::swap` to swap the components of two entities,
bypassing the insertion and removal hooks.
This means the only way to react to this modification is via change
detection in a system, which then begs the question of what happens
_between_ that alteration and the next run of that system?
Alternatively, you could make your component private to prevent
mutation, but now you need to provide commands and a custom `QueryData`
implementation to allow users to interact with your component at all.
Immutable components solve this problem by preventing the creation of an
exclusive reference to the component entirely. Without an exclusive
reference, the only way to modify an immutable component is via removal
or replacement, which is fully captured by component hooks. To make a
component immutable, simply add `#[component(immutable)]`:
```rust
#[derive(Component)]
#[component(immutable)]
struct Foo {
// ...
}
```
When implementing `Component` manually, there is an associated type
`Mutability` which controls this behavior:
```rust
impl Component for Foo {
type Mutability = Mutable;
// ...
}
```
Note that this means when working with generic components, you may need
to specify that a component is mutable to gain access to certain
methods:
```rust
// Before
fn bar<C: Component>() {
// ...
}
// After
fn bar<C: Component<Mutability = Mutable>>() {
// ...
}
```
With this new tool, creating index components, or caching data on an
entity should be more user friendly, allowing libraries to provide APIs
relying on components and hooks to uphold their invariants.
## Notes
- ~~I've done my best to implement this feature, but I'm not happy with
how reflection has turned out. If any reflection SMEs know a way to
improve this situation I'd greatly appreciate it.~~ There is an
outstanding issue around the fallibility of mutable methods on
`ReflectComponent`, but the DX is largely unchanged from `main` now.
- I've attempted to prevent all safe mutable access to a component that
does not implement `Component<Mutability = Mutable>`, but there may
still be some methods I have missed. Please indicate so and I will
address them, as they are bugs.
- Unsafe is an escape hatch I am _not_ attempting to prevent. Whatever
you do with unsafe is between you and your compiler.
- I am marking this PR as ready, but I suspect it will undergo fairly
major revisions based on SME feedback.
- I've marked this PR as _Uncontroversial_ based on the feature, not the
implementation.
---------
Co-authored-by: Alice Cecile <alice.i.cecile@gmail.com>
Co-authored-by: Benjamin Brienen <benjamin.brienen@outlook.com>
Co-authored-by: Gino Valente <49806985+MrGVSV@users.noreply.github.com>
Co-authored-by: Nuutti Kotivuori <naked@iki.fi>
# Objective
Make documentation of a component's required components more visible by
moving it to the type's docs
## Solution
Change `#[require]` from a derive macro helper to an attribute macro.
Disadvantages:
- this silences any unused code warnings on the component, as it is used
by the macro!
- need to import `require` if not using the ecs prelude (I have not
included this in the migration guilde as Rust tooling already suggests
the fix)
---
## Showcase
---------
Co-authored-by: Alice Cecile <alice.i.cecile@gmail.com>
Co-authored-by: JMS55 <47158642+JMS55@users.noreply.github.com>
This patch adds the infrastructure necessary for Bevy to support
*bindless resources*, by adding a new `#[bindless]` attribute to
`AsBindGroup`.
Classically, only a single texture (or sampler, or buffer) can be
attached to each shader binding. This means that switching materials
requires breaking a batch and issuing a new drawcall, even if the mesh
is otherwise identical. This adds significant overhead not only in the
driver but also in `wgpu`, as switching bind groups increases the amount
of validation work that `wgpu` must do.
*Bindless resources* are the typical solution to this problem. Instead
of switching bindings between each texture, the renderer instead
supplies a large *array* of all textures in the scene up front, and the
material contains an index into that array. This pattern is repeated for
buffers and samplers as well. The renderer now no longer needs to switch
binding descriptor sets while drawing the scene.
Unfortunately, as things currently stand, this approach won't quite work
for Bevy. Two aspects of `wgpu` conspire to make this ideal approach
unacceptably slow:
1. In the DX12 backend, all binding arrays (bindless resources) must
have a constant size declared in the shader, and all textures in an
array must be bound to actual textures. Changing the size requires a
recompile.
2. Changing even one texture incurs revalidation of all textures, a
process that takes time that's linear in the total size of the binding
array.
This means that declaring a large array of textures big enough to
encompass the entire scene is presently unacceptably slow. For example,
if you declare 4096 textures, then `wgpu` will have to revalidate all
4096 textures if even a single one changes. This process can take
multiple frames.
To work around this problem, this PR groups bindless resources into
small *slabs* and maintains a free list for each. The size of each slab
for the bindless arrays associated with a material is specified via the
`#[bindless(N)]` attribute. For instance, consider the following
declaration:
```rust
#[derive(AsBindGroup)]
#[bindless(16)]
struct MyMaterial {
#[buffer(0)]
color: Vec4,
#[texture(1)]
#[sampler(2)]
diffuse: Handle<Image>,
}
```
The `#[bindless(N)]` attribute specifies that, if bindless arrays are
supported on the current platform, each resource becomes a binding array
of N instances of that resource. So, for `MyMaterial` above, the `color`
attribute is exposed to the shader as `binding_array<vec4<f32>, 16>`,
the `diffuse` texture is exposed to the shader as
`binding_array<texture_2d<f32>, 16>`, and the `diffuse` sampler is
exposed to the shader as `binding_array<sampler, 16>`. Inside the
material's vertex and fragment shaders, the applicable index is
available via the `material_bind_group_slot` field of the `Mesh`
structure. So, for instance, you can access the current color like so:
```wgsl
// `uniform` binding arrays are a non-sequitur, so `uniform` is automatically promoted
// to `storage` in bindless mode.
@group(2) @binding(0) var<storage> material_color: binding_array<Color, 4>;
...
@fragment
fn fragment(in: VertexOutput) -> @location(0) vec4<f32> {
let color = material_color[mesh[in.instance_index].material_bind_group_slot];
...
}
```
Note that portable shader code can't guarantee that the current platform
supports bindless textures. Indeed, bindless mode is only available in
Vulkan and DX12. The `BINDLESS` shader definition is available for your
use to determine whether you're on a bindless platform or not. Thus a
portable version of the shader above would look like:
```wgsl
#ifdef BINDLESS
@group(2) @binding(0) var<storage> material_color: binding_array<Color, 4>;
#else // BINDLESS
@group(2) @binding(0) var<uniform> material_color: Color;
#endif // BINDLESS
...
@fragment
fn fragment(in: VertexOutput) -> @location(0) vec4<f32> {
#ifdef BINDLESS
let color = material_color[mesh[in.instance_index].material_bind_group_slot];
#else // BINDLESS
let color = material_color;
#endif // BINDLESS
...
}
```
Importantly, this PR *doesn't* update `StandardMaterial` to be bindless.
So, for example, `scene_viewer` will currently not run any faster. I
intend to update `StandardMaterial` to use bindless mode in a follow-up
patch.
A new example, `shaders/shader_material_bindless`, has been added to
demonstrate how to use this new feature.
Here's a Tracy profile of `submit_graph_commands` of this patch and an
additional patch (not submitted yet) that makes `StandardMaterial` use
bindless. Red is those patches; yellow is `main`. The scene was Bistro
Exterior with a hack that forces all textures to opaque. You can see a
1.47x mean speedup.
## Migration Guide
* `RenderAssets::prepare_asset` now takes an `AssetId` parameter.
* Bin keys now have Bevy-specific material bind group indices instead of
`wgpu` material bind group IDs, as part of the bindless change. Use the
new `MaterialBindGroupAllocator` to map from bind group index to bind
group ID.
# Objective
Fixes#15941
## Solution
Created https://crates.io/crates/variadics_please and moved the code
there; updating references
`bevy_utils/macros` is deleted.
## Testing
cargo check
## Migration Guide
Use `variadics_please::{all_tuples, all_tuples_with_size}` instead of
`bevy::utils::{all_tuples, all_tuples_with_size}`.
# Objective
- Fixes: #15663
## Solution
- Add an `is_forward_plane` method to `Aabb`, and a `contains_aabb`
method to `Frustum`.
## Test
- I have created a frustum with an offset along with three unit tests to
evaluate the `contains_aabb` algorithm.
## Explanation for the Test Cases
- To facilitate the code review, I will explain how the frustum is
created. Initially, we create a frustum without any offset and then
create a cuboid that is just contained within it.
<img width="714" alt="image"
src="https://github.com/user-attachments/assets/a9ac53a2-f8a3-4e09-b20b-4ee71b27a099">
- Secondly, we move the cuboid by 2 units along both the x-axis and the
y-axis to make it more general.
## Reference
- [Frustum
Culling](https://learnopengl.com/Guest-Articles/2021/Scene/Frustum-Culling#)
- [AABB Plane
intersection](https://gdbooks.gitbooks.io/3dcollisions/content/Chapter2/static_aabb_plane.html)
---------
Co-authored-by: IQuick 143 <IQuick143cz@gmail.com>
# Objective
- Refactor
## Solution
- Refactor
## Testing
- Ran 3d_scene
---
## Migration Guide
`RenderCreation::Manual` variant fields are now wrapped in a struct
called `RenderResources`
# Objective
`TemporaryRenderEntity` currently uses `SparseSet` storage, but doesn't
seem to fit the criteria for a component that would benefit from this.
Typical usage of `TemporaryRenderEntity` (and all current usages of it
in engine as far as I can tell) would be to spawn an entity with it once
and then iterate over it once to despawn that entity.
`SparseSet` is said to be useful for insert/removal perf at the cost of
iteration perf.
## Solution
Use the default table storage
## Testing
Possibly this could show up in stress tests like `many_buttons`. I
didn't do any benchmarking.
# Objective
Original motivation was a bundle I am migrating that is `Copy` which
needs to be synced to the render world. It probably doesn't actually
*need* to be `Copy`, so this isn't critical or anything.
I am continuing to use this bundle while bundles still exist to give
users an easier migration path.
## Solution
These ZSTs might as well be `Copy`. Add `Copy` derives.
PR #15164 made Bevy consider the center of the mesh to be the center of
the axis-aligned bounding box (AABB). Unfortunately, this breaks
crossfading in many cases. LODs may have different AABBs and so the
center of the AABB may differ for different LODs of the same mesh. The
crossfading, however, relies on all LODs having *precisely* the same
position.
To address this problem, this PR adds a new field, `use_aabb`, to
`VisibilityRange`, which makes the AABB center point behavior opt-in.
@BenjaminBrienen first noticed this issue when reviewing PR #16286. That
PR contains a video showing the effects of this regression on the
`visibility_range` example. This commit fixes that example.
## Migration Guide
* The `VisibilityRange` component now has an extra field, `use_aabb`.
Generally, you can safely set it to false.
We have an early-out to avoid updating `RenderVisibilityRanges` when a
`VisibilityRange` component is *modified*, but not when one is
*removed*. This means that removing `VisibilityRange` from an entity
might not update the rendering.
This PR fixes the issue by adding a check for removed
`VisibilityRange`s.
# Objective
- bevy_render (poorly) implements gcd (which should be in bevy_math but
theres not enough justification to have it there either anyways cus its
just one usage)
## Solution
- hardcoded LUT replacement for the one usage
## Testing
- verified the alternative implementation of 4/gcd(4,x) agreed with
original for 0..200
# Objective
- Fix part of #15920
## Solution
- Keep track of the last written amount of bytes, and bind only that
much of the buffer.
## Testing
- Did you test these changes? If so, how? No
- Are there any parts that need more testing?
- How can other people (reviewers) test your changes? Is there anything
specific they need to know?
- If relevant, what platforms did you test these changes on, and are
there any important ones you can't test?
---
## Migration Guide
- Fixed a bug with StorageBuffer and DynamicStorageBuffer binding data
from the previous frame(s) due to caching GPU buffers between frames.
# Objective
- Fixes#16285
- Inactive camera are keeping the component `ViewUniformOffset` from
when they were active, still matching some queries trying to render to
them
## Solution
- Remove component `ViewUniformOffset` from cameras that are inactive
## Testing
- Ran example `render_primitives` and switched camera
# Objective
UI Anti-aliasing is incorrectly implemented. It always uses an edge
radius of 0.25 logical pixels, and ignores the physical resolution. For
low dpi screens 0.25 is is too low and on higher dpi screens the
physical edge radius is much too large, resulting in visual artifacts.
## Solution
Multiply the distance by the scale factor in the `antialias` function so
that the edge radius stays constant in physical pixels.
## Testing
To see the problem really clearly run the button example with `UiScale`
set really high. With `UiScale(25.)` on main if you examine the button's
border you can see a thick gradient fading away from the edges:
<img width="127" alt="edgg"
src="https://github.com/user-attachments/assets/7c852030-c0e8-4aef-8d3e-768cb2464cab">
With this PR the edges are sharp and smooth at all scale factors:
<img width="127" alt="edge"
src="https://github.com/user-attachments/assets/b3231140-1bbc-4a4f-a1d3-dde21f287988">
# Objective
_If I understand it correctly_, we were checking mesh visibility, as
well as re-rendering point and spot light shadow maps for each view.
This makes it so that M views and N lights produce M x N complexity.
This PR aims to fix that, as well as introduce a stress test for this
specific scenario.
## Solution
- Keep track of what lights have already had mesh visibility calculated
and do not calculate it again;
- Reuse shadow depth textures and attachments across all views, and only
render shadow maps for the _first_ time a light is encountered on a
view;
- Directional lights remain unaltered, since their shadow map cascades
are view-dependent;
- Add a new `many_cameras_lights` stress test example to verify the
solution
## Showcase
110% speed up on the stress test
83% reduction of memory usage in stress test
### Before (5.35 FPS on stress test)
<img width="1392" alt="Screenshot 2024-09-11 at 12 25 57"
src="https://github.com/user-attachments/assets/136b0785-e9a4-44df-9a22-f99cc465e126">
### After (11.34 FPS on stress test)
<img width="1392" alt="Screenshot 2024-09-11 at 12 24 35"
src="https://github.com/user-attachments/assets/b8dd858f-5e19-467f-8344-2b46ca039630">
## Testing
- Did you test these changes? If so, how?
- On my game project where I have two cameras, and many shadow casting
lights I managed to get pretty much double the FPS.
- Also included a stress test, see the comparison above
- Are there any parts that need more testing?
- Yes, I would like help verifying that this fix is indeed correct, and
that we were really re-rendering the shadow maps by mistake and it's
indeed okay to not do that
- How can other people (reviewers) test your changes? Is there anything
specific they need to know?
- Run the `many_cameras_lights` example
- On the `main` branch, cherry pick the commit with the example (`git
cherry-pick --no-commit 1ed4ace01`) and run it
- If relevant, what platforms did you test these changes on, and are
there any important ones you can't test?
- macOS
---------
Co-authored-by: François Mockers <francois.mockers@vleue.com>
# Objective
Fixes#15940
## Solution
Remove the `pub use` and fix the compile errors.
Make `bevy_image` available as `bevy::image`.
## Testing
Feature Frenzy would be good here! Maybe I'll learn how to use it if I
have some time this weekend, or maybe a reviewer can use it.
## Migration Guide
Use `bevy_image` instead of `bevy_render::texture` items.
---------
Co-authored-by: chompaa <antony.m.3012@gmail.com>
Co-authored-by: Carter Anderson <mcanders1@gmail.com>
# Objective
- wgpu 0.20 made workgroup vars stop being zero-init by default. this
broke some applications (cough foresight cough) and now we workaround
it. wgpu exposes a compilation option that zero initializes workgroup
memory by default, but bevy does not expose it.
## Solution
- expose the compilation option wgpu gives us
## Testing
- ran examples: 3d_scene, compute_shader_game_of_life, gpu_readback,
lines, specialized_mesh_pipeline. they all work
- confirmed fix for our own problems
---
</details>
## Migration Guide
- add `zero_initialize_workgroup_memory: false,` to
`ComputePipelineDescriptor` or `RenderPipelineDescriptor` structs to
preserve 0.14 functionality, add `zero_initialize_workgroup_memory:
true,` to restore bevy 0.13 functionality.
# Objective
#14273 changed `Msaa` to be a component rather than a resource. However,
the documentation still says that it is a resource. This tripped me up
during migration to 0.15 until I looked at the type definition.
Additionally, the docs have some unnecessary repetition and some grammar
mistakes, and they don't link to camera documentation.
## Solution
Fix up the docs!
# Objective
- Fixes#16122
When the wayland feature is not enabled, xwayland is used on wayland.
Nvidia drivers are somewhat bugged on linux and return outdated surfaces
on xwayland for seemingly no reason. Oftentimes at startup we get into
an infine loop where the surface is permanently outdated and nothing (or
sometimes only the first frame) is drawn on the screen.
## Solution
After experimenting I found that we can safely call configure again and
the issue seems to resolve itsef. After this change I couldn't reproduce
the original issue after many tries. More testing is probably needed
though.
The main issue is that `get_current_texture` fails sometimes because the
surface remains outdated even after configuring. It would be better to
just properly handle and never panic when `get_current_texture` fails.
This way we always call configure when outdated and bail when getting
the swapchain fails instead of crashing. The number of special cases is
also reduced.
## Testing
I tested the example "rotation" manually by trying to move around.
It works with X11 and Xwayland and the non panicing code paths didn't
change so other platforms aren't affected.
# Objective
- Display message for `AsBindGroupError::InvalidSamplerType` was not
correctly displaying the binding index
## Solution
- Simple typo fix
## Testing
- Tested locally
# Objective
Bevy seems to want to standardize on "American English" spellings. Not
sure if this is laid out anywhere in writing, but see also #15947.
While perusing the docs for `typos`, I noticed that it has a `locale`
config option and tried it out.
## Solution
Switch to `en-us` locale in the `typos` config and run `typos -w`
## Migration Guide
The following methods or fields have been renamed from `*dependants*` to
`*dependents*`.
- `ProcessorAssetInfo::dependants`
- `ProcessorAssetInfos::add_dependant`
- `ProcessorAssetInfos::non_existent_dependants`
- `AssetInfo::dependants_waiting_on_load`
- `AssetInfo::dependants_waiting_on_recursive_dep_load`
- `AssetInfos::loader_dependants`
- `AssetInfos::remove_dependants_and_labels`
# Objective
Switch to retained render world causes the extracted cameras in render
world to not be removed until camera in main world is despawned. When
extracting data from main world inactive cameras are skipped. Therefore
camera that was active and became inactive has a retained
`ExtractedCamera` component from previous frames (when it was active)
and is processed the same way as if it were active (there is no `active`
field on `ExtractedCamera`). This breakes switching between cameras in
`render_primitives` example.
Fixes#15822
## Solution
Fix it by removing `ExtractedCamera` and related components from
inactive cameras.
Note that despawning inactive camera seems to be bad option because they
are spawned using `SyncToRenderWorld` component.
## Testing
Switching camera in `render_primitives` example now works correctly.
---------
Co-authored-by: akimakinai <105044389+akimakinai@users.noreply.github.com>
# Objective
Fixes#15791.
As raised in #11022, scaling orthographic cameras is confusing! In Bevy
0.14, there were multiple completely redundant ways to do this, and no
clear guidance on which to use.
As a result, #15075 removed the `scale` field from
`OrthographicProjection` completely, solving the redundancy issue.
However, this resulted in an unintuitive API and a painful migration, as
discussed in #15791. Users simply want to change a single parameter to
zoom, rather than deal with the irrelevant details of how the camera is
being scaled.
## Solution
This PR reverts #15075, and takes an alternate, more nuanced approach to
the redundancy problem. `ScalingMode::WindowSize` was by far the biggest
offender. This was the default variant, and stored a float that was
*fully* redundant to setting `scale`.
All of the other variants contained meaningful semantic information and
had an intuitive scale. I could have made these unitless, storing an
aspect ratio, but this would have been a worse API and resulted in a
pointlessly painful migration.
In the course of this work I've also:
- improved the documentation to explain that you should just set `scale`
to zoom cameras
- swapped to named fields for all of the variants in `ScalingMode` for
more clarity about the parameter meanings
- substantially improved the `projection_zoom` example
- removed the footgunny `Mul` and `Div` impls for `ScalingMode`,
especially since these no longer have the intended effect on
`ScalingMode::WindowSize`.
- removed a rounding step because this is now redundant 🎉
## Testing
I've tested these changes as part of my work in the `projection_zoom`
example, and things seem to work fine.
## Migration Guide
`ScalingMode` has been refactored for clarity, especially on how to zoom
orthographic cameras and their projections:
- `ScalingMode::WindowSize` no longer stores a float, and acts as if its
value was 1. Divide your camera's scale by any previous value to achieve
identical results.
- `ScalingMode::FixedVertical` and `FixedHorizontal` now use named
fields.
---------
Co-authored-by: MiniaczQ <xnetroidpl@gmail.com>
Fixes#15834
## Migration Guide
The APIs of `Time`, `Timer` and `Stopwatch` have been cleaned up for
consistency with each other and the standard library's `Duration` type.
The following methods have been renamed:
- `Stowatch::paused` -> `Stopwatch::is_paused`
- `Time::elapsed_seconds` -> `Time::elasped_secs` (including `_f64` and
`_wrapped` variants)
# Objective
- `C: ExtractComponent` inserts `C::Out` instead of `C`, so we need to
remove `C::Out`. cc #15904.
## Solution
- `C` -> `C::Out`
## Testing
- CAS has `<ContrastAdaptiveSharpening as ExtractComponent>::Out =
(DenoiseCas, CasUniform)`. Setting its strength to zero correctly
removes the effect after this change.
# Objective
Closes#15799.
Many rendering people and maintainers are in favor of reverting default
mesh materials added in #15524, especially as the migration to required
component is already large and heavily breaking.
## Solution
Revert default mesh materials, and adjust docs accordingly.
- Remove `extract_default_materials`
- Remove `clear_material_instances`, and move the logic back into
`extract_mesh_materials`
- Remove `HasMaterial2d` and `HasMaterial3d`
- Change default material handles back to pink instead of white
- 2D uses `Color::srgb(1.0, 0.0, 1.0)`, while 3D uses `Color::srgb(1.0,
0.0, 0.5)`. Not sure if this is intended.
There is now no indication at all about missing materials for `Mesh2d`
and `Mesh3d`. Having a mesh without a material renders nothing.
## Testing
I ran `2d_shapes`, `mesh2d_manual`, and `3d_shapes`, with and without
mesh material components.
# Objective
Fixes#15730.
## Solution
As part of #15586, we made a constant to store all the supported image
formats. However since the `ImageFormat` does actually include Hdr and
OpenExr, it also included the `"hdr"` and `"exr"` file extensions. These
are supported by separate loaders though: `HdrTextureLoader` and
`ExrTextureLoader`. This led to a warning about duplicate asset loaders.
Therefore, instead of having the constant for `ImageFormat`, I made the
constant just for `ImageLoader`. This lets us correctly remove `"hdr"`
and `"exr"` from the image formats supported by `ImageLoader`, returning
us to having a single asset loader for every image format.
Note: we could have just removed `hdr` and `exr` from
`ImageFormat::SUPPORTED_FILE_EXTENSIONS`, but this would be very
confusing. Then the list of `ImageFormat`s would not match the list of
supported formats!
## Testing
- I ran the `sprite` example and got no warning! I also replaced the
sprite in that example with an HDR file and everything worked as
expected.
# Objective
Ensure that components that are conditionally extracted do not linger in
the render world when not extracted from the main world.
## Solution
If the `ExtractComponent` returns `None`, we'll remove the render world
component. I think this is the most sensible behavior here. In the
future if there really is a use case for keeping the previous render
component around, we could add a `Option<Self::Out>` parameter for the
previous render component to the method, or something similar. I think
that this follows the principle of least surprise here relative to what
`None` would suggest and the way that render nodes are typically
written. The alternative would be to add an `enabled` field to pretty
much every camera settings component, or duplicate the extraction
condition as #15856 does.
## Testing
`transmission` no longer crashes.
## Migration Guide
Components that implement `ExtractComponent` and return `None` will
cause the extracted component to be removed from the render world.
# Objective
This is a follow-up to #15650. While the core `Image` stuff moved from
`bevy_render` to `bevy_image`, the `ImageLoader` and the
`CompressedImageSaver` remained in `bevy_render`.
## Solution
I moved `ImageLoader` and `CompressedImageSaver` to `bevy_image` and
re-exported everything out from `bevy_render`. The second step isn't
strictly necessary, but `bevy_render` is already doing this for all the
other `bevy_image` types, so I kept it the same for consistency.
Unfortunately I had to give `ImageLoader` a constructor so I can keep
the `RenderDevice` stuff in `bevy_render`.
## Testing
It compiles!
## Migration Guide
- `ImageLoader` can no longer be initialized directly through
`init_asset_loader`. Now you must use
`app.register_asset_loader(ImageLoader::new(supported_compressed_formats))`
(check out the implementation of `bevy_render::ImagePlugin`). This only
affects you if you are initializing the loader manually and does not
affect users of `bevy_render::ImagePlugin`.
## Followup work
- We should be able to move most of the `ImagePlugin` to `bevy_image`.
This would likely require an `ImagePlugin` and a `RenderImagePlugin` or
something though.
# Objective
Another clippy-lint fix: the goal is so that `ci lints` actually
displays the problems that a contributor caused, and not a bunch of
existing stuff in the repo. (when run on nightly)
## Solution
This fixes all but the `clippy::needless_lifetimes` lint, which will
result in substantially more fixes and be in other PR(s). I also
explicitly allow `non_local_definitions` since it is [not working
correctly, but will be
fixed](https://github.com/rust-lang/rust/issues/131643).
A few things were manually fixed: for example, some places had an
explicitly defined `div_ceil` function that was used, which is no longer
needed since this function is stable on unsigned integers. Also, empty
lines in doc comments were handled individually.
## Testing
I ran `cargo clippy --workspace --all-targets --all-features --fix
--allow-staged` with the `clippy::needless_lifetimes` lint marked as
`allow` in `Cargo.toml` to avoid fixing that too. It now passes with all
but the listed lint.
# Objective
#15320 is a particularly painful breaking change, and the new
`RenderEntity` in particular is very noisy, with a lot of `let entity =
entity.id()` spam.
## Solution
Implement `WorldQuery`, `QueryData` and `ReadOnlyQueryData` for
`RenderEntity` and `WorldEntity`.
These work the same as the `Entity` impls from a user-facing
perspective: they simply return an owned (copied) `Entity` identifier.
This dramatically reduces noise and eases migration.
Under the hood, these impls defer to the implementations for `&T` for
everything other than the "call .id() for the user" bit, as they involve
read-only access to component data. Doing it this way (as opposed to
implementing a custom fetch, as tried in the first commit) dramatically
reduces the maintenance risk of complex unsafe code outside of
`bevy_ecs`.
To make this easier (and encourage users to do this themselves!), I've
made `ReadFetch` and `WriteFetch` slightly more public: they're no
longer `doc(hidden)`. This is a good change, since trying to vendor the
logic is much worse than just deferring to the existing tested impls.
## Testing
I've run a handful of rendering examples (breakout, alien_cake_addict,
auto_exposure, fog_volumes, box_shadow) and nothing broke.
## Follow-up
We should lint for the uses of `&RenderEntity` and `&MainEntity` in
queries: this is just less nice for no reason.
---------
Co-authored-by: Trashtalk217 <trashtalk217@gmail.com>
# Objective
- Required components replace bundles, but `SpatialBundle` is yet to be
deprecated
## Solution
- Deprecate `SpatialBundle`
- Insert `Transform` and `Visibility` instead in examples using it
- In `spawn` or `insert` inserting a default `Transform` or `Visibility`
with component already requiring either, remove those components from
the tuple
## Testing
- Did you test these changes? If so, how?
Yes, I ran the examples I changed and tests
- Are there any parts that need more testing?
The `gamepad_viewer` and and `custom_shader_instancing` examples don't
work as intended due to entirely unrelated code, didn't check main.
- How can other people (reviewers) test your changes? Is there anything
specific they need to know?
Run examples, or just check that all spawned values are identical
- If relevant, what platforms did you test these changes on, and are
there any important ones you can't test?
Linux, wayland trough x11 (cause that's the default feature)
---
## Migration Guide
`SpatialBundle` is now deprecated, insert `Transform` and `Visibility`
instead which will automatically insert all other components that were
in the bundle. If you do not specify these values and any other
components in your `spawn`/`insert` call already requires either of
these components you can leave that one out.
before:
```rust
commands.spawn(SpatialBundle::default());
```
after:
```rust
commands.spawn((Transform::default(), Visibility::default());
```
# Objective
In the Render World, there are a number of collections that are derived
from Main World entities and are used to drive rendering. The most
notable are:
- `VisibleEntities`, which is generated in the `check_visibility` system
and contains visible entities for a view.
- `ExtractedInstances`, which maps entity ids to asset ids.
In the old model, these collections were trivially kept in sync -- any
extracted phase item could look itself up because the render entity id
was guaranteed to always match the corresponding main world id.
After #15320, this became much more complicated, and was leading to a
number of subtle bugs in the Render World. The main rendering systems,
i.e. `queue_material_meshes` and `queue_material2d_meshes`, follow a
similar pattern:
```rust
for visible_entity in visible_entities.iter::<With<Mesh2d>>() {
let Some(mesh_instance) = render_mesh_instances.get_mut(visible_entity) else {
continue;
};
// Look some more stuff up and specialize the pipeline...
let bin_key = Opaque2dBinKey {
pipeline: pipeline_id,
draw_function: draw_opaque_2d,
asset_id: mesh_instance.mesh_asset_id.into(),
material_bind_group_id: material_2d.get_bind_group_id().0,
};
opaque_phase.add(
bin_key,
*visible_entity,
BinnedRenderPhaseType::mesh(mesh_instance.automatic_batching),
);
}
```
In this case, `visible_entities` and `render_mesh_instances` are both
collections that are created and keyed by Main World entity ids, and so
this lookup happens to work by coincidence. However, there is a major
unintentional bug here: namely, because `visible_entities` is a
collection of Main World ids, the phase item being queued is created
with a Main World id rather than its correct Render World id.
This happens to not break mesh rendering because the render commands
used for drawing meshes do not access the `ItemQuery` parameter, but
demonstrates the confusion that is now possible: our UI phase items are
correctly being queued with Render World ids while our meshes aren't.
Additionally, this makes it very easy and error prone to use the wrong
entity id to look up things like assets. For example, if instead we
ignored visibility checks and queued our meshes via a query, we'd have
to be extra careful to use `&MainEntity` instead of the natural
`Entity`.
## Solution
Make all collections that are derived from Main World data use
`MainEntity` as their key, to ensure type safety and avoid accidentally
looking up data with the wrong entity id:
```rust
pub type MainEntityHashMap<V> = hashbrown::HashMap<MainEntity, V, EntityHash>;
```
Additionally, we make all `PhaseItem` be able to provide both their Main
and Render World ids, to allow render phase implementors maximum
flexibility as to what id should be used to look up data.
You can think of this like tracking at the type level whether something
in the Render World should use it's "primary key", i.e. entity id, or
needs to use a foreign key, i.e. `MainEntity`.
## Testing
##### TODO:
This will require extensive testing to make sure things didn't break!
Additionally, some extraction logic has become more complicated and
needs to be checked for regressions.
## Migration Guide
With the advent of the retained render world, collections that contain
references to `Entity` that are extracted into the render world have
been changed to contain `MainEntity` in order to prevent errors where a
render world entity id is used to look up an item by accident. Custom
rendering code may need to be changed to query for `&MainEntity` in
order to look up the correct item from such a collection. Additionally,
users who implement their own extraction logic for collections of main
world entity should strongly consider extracting into a different
collection that uses `MainEntity` as a key.
Additionally, render phases now require specifying both the `Entity` and
`MainEntity` for a given `PhaseItem`. Custom render phases should ensure
`MainEntity` is available when queuing a phase item.
# Objective
- Another step towards #15716
- Remove trait implementations that are dependent on `Handle<T>` being a
`Component`
## Solution
- Remove unused `ExtractComponent` trait implementation for `Handle<T>`
- Remove unused `ExtractInstance` trait implementation for `AssetId`
- Although the `ExtractInstance` trait wasn't used, the `AssetId`s were
being stored inside of `ExtractedInstances` which has an
`ExtractInstance` trait bound on its contents.
I've upgraded the `RenderMaterialInstances` type alias to be its own
resource, identical to `ExtractedInstances<AssetId<M>>` to get around
that with minimal breakage.
## Testing
Tested `many_cubes`, rendering did not explode
# Objective
- Closes#15752
Calling the functions `App::observe` and `World::observe` doesn't make
sense because you're not "observing" the `App` or `World`, you're adding
an observer that listens for an event that occurs *within* the `World`.
We should rename them to better fit this.
## Solution
Renames:
- `App::observe` -> `App::add_observer`
- `World::observe` -> `World::add_observer`
- `Commands::observe` -> `Commands::add_observer`
- `EntityWorldMut::observe_entity` -> `EntityWorldMut::observe`
(Note this isn't a breaking change as the original rename was introduced
earlier this cycle.)
## Testing
Reusing current tests.
# Objective
- Fixes#15285
## Solution
`winit` sends resized to zero events when the window is minimized only
on Windows OS(rust-windowing/winit#2015).
This makes updating window viewport size to `(0, 0)` and panicking when
calculating aspect ratio.
~~So, just skip these kinds of events - resizing to (0, 0) when the
window is minimized - on Windows OS~~
Idially, the camera extraction excludes the cameras whose target size
width or height is zero here;
25bfa80e60/crates/bevy_render/src/camera/camera.rs (L1060-L1074)
but it seems that winit event loop sends resize events after extraction
and before post update schedule, so they might panics before the
extraction filters them out.
Alternatively, it might be possible to change event loop evaluating
order or defer them to the right schedule but I'm afraid that it might
cause some breaking changes, so just skip rendering logics for such
windows and they will be all filtered out by the extractions on the next
frame and thereafter.
## Testing
Running the example in the original issue and minimizing causes panic,
or just running `tests/window/minimising.rs` with `cargo run --example
minimising` panics without this PR and doesn't panics with this PR.
I think that we should run it in CI on Windows OS btw
# Objective
- Fixes (partially) #15701.
## Solution
- Use little-endian bytes over native-endian bytes where applicable.
## Testing
- Ran CI.
## Open Questions
- Should we config-gate these for big-endian targets? It looks like
there are [very few
targets](https://doc.rust-lang.org/nightly/rustc/platform-support.html)
that use big-endian.
# Objective
Fixes#15560
Fixes (most of) #15570
Currently a lot of examples (and presumably some user code) depend on
toggling certain render features by adding/removing a single component
to an entity, e.g. `SpotLight` to toggle a light. Because of the
retained render world this no longer works: Extract will add any new
components, but when it is removed the entity persists unchanged in the
render world.
## Solution
Add `SyncComponentPlugin<C: Component>` that registers
`SyncToRenderWorld` as a required component for `C`, and adds a
component hook that will clear all components from the render world
entity when `C` is removed. We add this plugin to
`ExtractComponentPlugin` which fixes most instances of the problem. For
custom extraction logic we can manually add `SyncComponentPlugin` for
that component.
We also rename `WorldSyncPlugin` to `SyncWorldPlugin` so we start a
naming convention like all the `Extract` plugins.
In this PR I also fixed a bunch of breakage related to the retained
render world, stemming from old code that assumed that `Entity` would be
the same in both worlds.
I found that using the `RenderEntity` wrapper instead of `Entity` in
data structures when referring to render world entities makes intent
much clearer, so I propose we make this an official pattern.
## Testing
Run examples like
```
cargo run --features pbr_multi_layer_material_textures --example clearcoat
cargo run --example volumetric_fog
```
and see that they work, and that toggles work correctly. But really we
should test every single example, as we might not even have caught all
the breakage yet.
---
## Migration Guide
The retained render world notes should be updated to explain this edge
case and `SyncComponentPlugin`
---------
Co-authored-by: Alice Cecile <alice.i.cecile@gmail.com>
Co-authored-by: Trashtalk217 <trashtalk217@gmail.com>
The previous fixes were breaking pretty much everything on main due to
naga-oil complaining about the OIT shader not being loaded, since
apparently webgl is a default feature. This fix is a bit messier, but
properly warns the user and is probably what we should have gone for in
the first place.
# Objective
Bevy supports feature gates for each format it supports, but several
formats that it loads via the `image` crate do not have feature gates.
Additionally, the QOI format is supported by the `image` crate and
wasn't available at all. This fixes that.
## Solution
The following feature gates are added:
* `avif`
* `ff` (Farbfeld)
* `gif`
* `ico`
* `qoi`
* `tiff`
None of these formats are enabled by default, despite the fact that all
these formats appeared to be enabled by default before. Since
`default-features` was disabled for the `image` crate, it's likely that
using any of these formats would have errored by default before this
change, although this probably needs additional testing.
## Testing
The changes seemed minimal enough that a compile test would be
sufficient.
## Migration guide
Image formats that previously weren't feature-gated are now
feature-gated, meaning they will have to be enabled if you use them:
* `avif`
* `ff` (Farbfeld)
* `gif`
* `ico`
* `tiff`
Additionally, the `qoi` feature has been added to support loading QOI
format images.
Previously, these formats appeared in the enum by default, but weren't
actually enabled via the `image` crate, potentially resulting in weird
bugs. Now, you should be able to add these features to your projects to
support them properly.
