# Objective
This PR fixes a crash that happens when an asset failure event is
processed after the asset has already been dropped.
```
2024-07-03T17:12:16.847178Z ERROR bevy_asset::server: Encountered HTTP status 404 when loading asset
thread 'main' panicked at bevy/crates/bevy_asset/src/server/info.rs:593:18:
```
## Solution
- Update `process_asset_fail` to match the graceful behavior in
`process_asset_load` (it does not assume the state still exists).
---
## Changelog
- Fixed a rare crash that happens when an asset failed event is
processed after the asset has been dropped.
# Objective
Allow random sampling from the surfaces of triangle meshes.
## Solution
This has two parts.
Firstly, rendering meshes can now yield their collections of triangles
through a method `Mesh::triangles`. This has signature
```rust
pub fn triangles(&self) -> Result<Vec<Triangle3d>, MeshTrianglesError> { //... }
```
and fails in a variety of cases — the most obvious of these is that the
mesh must have either the `TriangleList` or `TriangleStrip` topology,
and the others correspond to malformed vertex or triangle-index data.
With that in hand, we have the second piece, which is
`UniformMeshSampler`, which is a `Vec3`-valued
[distribution](https://docs.rs/rand/latest/rand/distributions/trait.Distribution.html)
that samples uniformly from collections of triangles. It caches the
triangles' distribution of areas so that after its initial setup,
sampling is allocation-free. It is constructed via
`UniformMeshSampler::try_new`, which looks like this:
```rust
pub fn try_new<T: Into<Vec<Triangle3d>>>(triangles: T) -> Result<Self, ZeroAreaMeshError> { //... }
```
It fails if the collection of triangles has zero area.
The sum of these parts means that you can sample random points from a
mesh as follows:
```rust
let triangles = my_mesh.triangles().unwrap();
let mut rng = StdRng::seed_from_u64(8765309);
let distribution = UniformMeshSampler::try_new(triangles).unwrap();
// 10000 random points from the surface of my_mesh:
let sample_points: Vec<Vec3> = distribution.sample_iter(&mut rng).take(10000).collect();
```
## Testing
Tested by instantiating meshes and sampling as demonstrated above.
---
## Changelog
- Added `Mesh::triangles` method to get a collection of triangles from a
mesh.
- Added `UniformMeshSampler` to `bevy_math::sampling`. This is a
distribution which allows random sampling over collections of triangles
(such as those provided through meshes).
---
## Discussion
### Design decisions
The main thing here was making sure to have a good separation between
the parts of this in `bevy_render` and in `bevy_math`. Getting the
triangles from a mesh seems like a reasonable step after adding
`Triangle3d` to `bevy_math`, so I decided to make all of the random
sampling operate at that level, with the fallible conversion to
triangles doing most of the work.
Notably, the sampler could be called something else that reflects that
its input is a collection of triangles, but if/when we add other kinds
of meshes to `bevy_math` (e.g. half-edge meshes), the fact that
`try_new` takes an `impl Into<Vec<Triangle3d>>` means that those meshes
just need to satisfy that trait bound in order to work immediately with
this sampling functionality. In that case, the result would just be
something like this:
```rust
let dist = UniformMeshSampler::try_new(mesh).unwrap();
```
I think this highlights that most of the friction is really just from
extracting data from `Mesh`.
It's maybe worth mentioning also that "collection of triangles"
(`Vec<Triangle3d>`) sits downstream of any other kind of triangle mesh,
since the topology connecting the triangles has been effectively erased,
which makes an `Into<Vec<Triangle3d>>` trait bound seem all the more
natural to me.
---------
Co-authored-by: Alice Cecile <alice.i.cecile@gmail.com>
# Objective
Fixes https://github.com/bevyengine/bevy/issues/13972
## Solution
Added 3 new attributes to the `Component` macro.
## Testing
Added `component_hook_order_spawn_despawn_with_macro_hooks`, that makes
the same as `component_hook_order_spawn_despawn` but uses a struct, that
defines it's hooks with the `Component` macro.
---
---------
Co-authored-by: Gino Valente <49806985+MrGVSV@users.noreply.github.com>
# Objective
- `EmptyPathStream` is only used in android and wasm32
- This now makes rust nightly warn
## Solution
- flag the struct to only be present when needed
- also change how `MorphTargetNames` is used because that makes rust
happier?
# Objective
- Emit an event regardless of scene type (`Scene` and `DynamicScene`).
- Also send the `InstanceId` along.
Follow-up to #11002.
Fixes#2218.
## Solution
- Send `SceneInstanceReady` regardless of scene type.
- Make `SceneInstanceReady::parent` `Option`al.
- Add `SceneInstanceReady::id`.
---
## Changelog
### Changed
- `SceneInstanceReady` is now sent for `Scene` as well.
`SceneInstanceReady::parent` is an `Option` and
`SceneInstanceReady::id`, an `InstanceId`, is added to identify the
corresponding `Scene`.
## Migration Guide
- `SceneInstanceReady { parent: Entity }` is now `SceneInstanceReady {
id: InstanceId, parent: Option<Entity> }`.
# Objective
- `Parallel::drain()` has an unused type parameter `B` than can be
removed.
- Caught [on
Discord](https://discord.com/channels/691052431525675048/692572690833473578/1259004180560085003)
by Andrew, thanks!
## Solution
- Remove it! :)
## Testing
- `Parallel::drain()` should still function exactly the same.
---
## Changelog
- Removed unused type parameter in `Parallel::drain()`.
## Migration Guide
The type parameter of `Parallel::drain()` was unused, so it is now
removed. If you were manually specifying it, you can remove the bounds.
```rust
// 0.14
// Create a `Parallel` and give it a value.
let mut parallel: Parallel<Vec<u8>> = Parallel::default();
*parallel.borrow_local_mut() = vec![1, 2, 3];
for v in parallel.drain::<u8>() {
// ...
}
// 0.15
let mut parallel: Parallel<Vec<u8>> = Parallel::default();
*parallel.borrow_local_mut() = vec![1, 2, 3];
// Remove the type parameter.
for v in parallel.drain() {
// ...
}
```
# Objective
Looks like I accidentally disabled the reflection compile fail tests in
#13152. These should be re-enabled.
## Solution
Re-enable reflection compile fail tests.
## Testing
CI should pass. You can also test locally by navigating to
`crates/bevy_reflect/compile_fail/` and running:
```
cargo test --target-dir ../../../target
```
# Objective
Both `Material` and `MaterialExtension` (base and extension) can derive
Debug, so there's no reason to not allow `ExtendedMaterial` to derive it
## Solution
- Describe the solution used to achieve the objective above.
Add `Debug` to the list of derived traits
## Testing
- Did you test these changes? If so, how?
I compiled my test project on latest commit, making sure it actually
compiles
- How can other people (reviewers) test your changes? Is there anything
specific they need to know?
Create an ExtendedMaterial instance, try to `println!("{:?}",
material);`
Co-authored-by: NWPlayer123 <NWPlayer123@users.noreply.github.com>
# Replace ab_glyph with the more capable cosmic-text
Fixes#7616.
Cosmic-text is a more mature text-rendering library that handles scripts
and ligatures better than ab_glyph, it can also handle system fonts
which can be implemented in bevy in the future
Rebase of https://github.com/bevyengine/bevy/pull/8808
## Changelog
Replaces text renderer ab_glyph with cosmic-text
The definition of the font size has changed with the migration to cosmic
text. The behavior is now consistent with other platforms (e.g. the
web), where the font size in pixels measures the height of the font (the
distance between the top of the highest ascender and the bottom of the
lowest descender). Font sizes in your app need to be rescaled to
approximately 1.2x smaller; for example, if you were using a font size
of 60.0, you should now use a font size of 50.0.
## Migration guide
- `Text2dBounds` has been replaced with `TextBounds`, and it now accepts
`Option`s to the bounds, instead of using `f32::INFINITY` to inidicate
lack of bounds
- Textsizes should be changed, dividing the current size with 1.2 will
result in the same size as before.
- `TextSettings` struct is removed
- Feature `subpixel_alignment` has been removed since cosmic-text
already does this automatically
- TextBundles and things rendering texts requires the `CosmicBuffer`
Component on them as well
## Suggested followups:
- TextPipeline: reconstruct byte indices for keeping track of eventual
cursors in text input
- TextPipeline: (future work) split text entities into section entities
- TextPipeline: (future work) text editing
- Support line height as an option. Unitless `1.2` is the default used
in browsers (1.2x font size).
- Support System Fonts and font families
- Example showing of animated text styles. Eg. throbbing hyperlinks
---------
Co-authored-by: tigregalis <anak.harimau@gmail.com>
Co-authored-by: Nico Burns <nico@nicoburns.com>
Co-authored-by: sam edelsten <samedelsten1@gmail.com>
Co-authored-by: Dimchikkk <velo.app1@gmail.com>
Co-authored-by: Alice Cecile <alice.i.cecile@gmail.com>
Co-authored-by: Rob Parrett <robparrett@gmail.com>
# Objective
- After #14017 , I noticed that the drawcall increased 10x in the
`many_buttons`, causing the `UIPassNode `to increase from 1.5ms to 6ms.
This is because our UI batching is very fragile.
## Solution
- skip extract UiImage when its texture is default
## Performance
many_buttons UiPassNode
# Objective
- Standard Material is starting to run out of samplers (currently uses
13 with no additional features off, I think in 0.13 it was 12).
- This change adds a new feature switch, modelled on the other ones
which add features to Standard Material, to turn off the new anisotropy
feature by default.
## Solution
- feature + texture define
## Testing
- Anisotropy example still works fine
- Other samples work fine
- Standard Material now takes 12 samplers by default on my Mac instead
of 13
## Migration Guide
- Add feature pbr_anisotropy_texture if you are using that texture in
any standard materials.
---------
Co-authored-by: John Payne <20407779+johngpayne@users.noreply.github.com>
# Objective
The bounds for query iterators are quite intimidating.
## Solution
With Rust 1.79, [associated type
bounds](https://github.com/rust-lang/rust/pull/122055/) stabilized,
which can simplify the bounds slightly.
# Objective
The Bevy API around manipulating hierarchies removes `Children` if the
operation results in an entity having no children. This means that
`Children` is guaranteed to hold actual children. However, the following
code unexpectedly inserts empty `Children`:
```rust
commands.entity(entity).with_children(|_| {});
```
This was discovered by @Jondolf:
https://discord.com/channels/691052431525675048/1124043933886976171/1257660865625325800
## Solution
- `with_children` is now a noop when no children were passed
## Testing
- Added a regression test
# Objective
The `AssetReader` trait allows customizing the behavior of fetching
bytes for an `AssetPath`, and expects implementors to return `dyn
AsyncRead + AsyncSeek`. This gives implementors of `AssetLoader` great
flexibility to tightly integrate their asset loading behavior with the
asynchronous task system.
However, almost all implementors of `AssetLoader` don't use the async
functionality at all, and just call `AsyncReadExt::read_to_end(&mut
Vec<u8>)`. This is incredibly inefficient, as this method repeatedly
calls `poll_read` on the trait object, filling the vector 32 bytes at a
time. At my work we have assets that are hundreds of megabytes which
makes this a meaningful overhead.
## Solution
Turn the `Reader` type alias into an actual trait, with a provided
method `read_to_end`. This provided method should be more efficient than
the existing extension method, as the compiler will know the underlying
type of `Reader` when generating this function, which removes the
repeated dynamic dispatches and allows the compiler to make further
optimizations after inlining. Individual implementors are able to
override the provided implementation -- for simple asset readers that
just copy bytes from one buffer to another, this allows removing a large
amount of overhead from the provided implementation.
Now that `Reader` is an actual trait, I also improved the ergonomics for
implementing `AssetReader`. Currently, implementors are expected to box
their reader and return it as a trait object, which adds unnecessary
boilerplate to implementations. This PR changes that trait method to
return a pseudo trait alias, which allows implementors to return `impl
Reader` instead of `Box<dyn Reader>`. Now, the boilerplate for boxing
occurs in `ErasedAssetReader`.
## Testing
I made identical changes to my company's fork of bevy. Our app, which
makes heavy use of `read_to_end` for asset loading, still worked
properly after this. I am not aware if we have a more systematic way of
testing asset loading for correctness.
---
## Migration Guide
The trait method `bevy_asset::io::AssetReader::read` (and `read_meta`)
now return an opaque type instead of a boxed trait object. Implementors
of these methods should change the type signatures appropriately
```rust
impl AssetReader for MyReader {
// Before
async fn read<'a>(&'a self, path: &'a Path) -> Result<Box<Reader<'a>>, AssetReaderError> {
let reader = // construct a reader
Box::new(reader) as Box<Reader<'a>>
}
// After
async fn read<'a>(&'a self, path: &'a Path) -> Result<impl Reader + 'a, AssetReaderError> {
// create a reader
}
}
```
`bevy::asset::io::Reader` is now a trait, rather than a type alias for a
trait object. Implementors of `AssetLoader::load` will need to adjust
the method signature accordingly
```rust
impl AssetLoader for MyLoader {
async fn load<'a>(
&'a self,
// Before:
reader: &'a mut bevy::asset::io::Reader,
// After:
reader: &'a mut dyn bevy::asset::io::Reader,
_: &'a Self::Settings,
load_context: &'a mut LoadContext<'_>,
) -> Result<Self::Asset, Self::Error> {
}
```
Additionally, implementors of `AssetReader` that return a type
implementing `futures_io::AsyncRead` and `AsyncSeek` might need to
explicitly implement `bevy::asset::io::Reader` for that type.
```rust
impl bevy::asset::io::Reader for MyAsyncReadAndSeek {}
```
# Objective
Add ability to de-register events from the EventRegistry (and the
associated World).
The initial reasoning relates to retaining support for Event hot
reloading in `dexterous_developer`.
## Solution
Add a `deregister_events<T: Event>(&mut world)` method to the
`EventRegistry` struct.
## Testing
Added an automated test that verifies the event registry adds and
removes `Events<T>` from the world.
---------
Co-authored-by: BD103 <59022059+BD103@users.noreply.github.com>
# Objective
It's not always obvious what the default value for `RenderLayers`
represents. It is documented, but since it's an implementation of a
trait method the documentation may or may not be shown depending on the
IDE.
## Solution
Add documentation to the `none` method that explicitly calls out the
difference.
---------
Co-authored-by: Alice Cecile <alice.i.cecile@gmail.com>
Also bumps `accesskit_winit` to 0.22 and fixes one breaking change.
# Objective
- `accesskit` has been updated recently to 0.16!
## Solution
- Update `accesskit`, as well as `accesskit_winit`.
- [`accesskit`
changelog](552032c839/common/CHANGELOG.md (0160-2024-06-29))
- [`accesskit_winit`
changelog](552032c839/platforms/winit/CHANGELOG.md (0220-2024-06-29))
- Fix one breaking change where `Role::StaticText` has been renamed to
`Role::Label`.
## Testing
- The test suite should cover most things.
- It would be good to test this with an example, but I don't know how.
---
## Changelog
- Update `accesskit` to 0.16 and `accesskit_winit` to 0.22.
## Migration Guide
`accesskit`'s `Role::StaticText` variant has been renamed to
`Role::Label`.
# Objective
- Though Rodio will eventually be replaced with Kira for `bevy_audio`,
we should not let it languish.
## Solution
- Bump Rodio to 0.19.
- This is [the
changelog](27f2b42406/CHANGELOG.md (version-0190-2024-06-29)).
No apparent breaking changes, only 1 feature and 1 fix.
## Testing
- Run an example that uses audio, on both native and WASM.
---
## Changelog
- Bumped Rodio to 0.19.
# Objective
The `BuildChildren` and `BuildWorldChildren` traits are mostly
identical, so I decided to try and merge them. I'm not sure of the
history, maybe they were added before GATs existed.
## Solution
- Add an associated type to `BuildChildren` which reflects the prior
differences between the `BuildChildren` and `BuildWorldChildren` traits.
- Add `ChildBuild` trait that is the bounds for
`BuildChildren::Builder`, with impls for `ChildBuilder` and
`WorldChildBuilder`.
- Remove `BuildWorldChildren` trait and replace it with an impl of
`BuildChildren` for `EntityWorldMut`.
## Testing
I ran several of the examples that use entity hierarchies, mainly UI.
---
## Changelog
n/a
## Migration Guide
n/a
# Objective
- Add the `AccumulatedMouseMotion` and `AccumulatedMouseScroll`
resources to make it simpler to track mouse motion/scroll changes
- Closes#13915
## Solution
- Created two resources, `AccumulatedMouseMotion` and
`AccumulatedMouseScroll`, and a method that tracks the `MouseMotion` and
`MouseWheel` events and accumulates their deltas every frame.
- Also modified the mouse input example to show how to use the
resources.
## Testing
- Tested the changes by modifying an existing example to use the newly
added resources, and moving/scrolling my trackpad around a ton.
---------
Co-authored-by: Alice Cecile <alice.i.cecile@gmail.com>
Co-authored-by: Gino Valente <49806985+MrGVSV@users.noreply.github.com>
# Objective
Fixes#13995.
## Solution
Override the default `Ctrl+C` handler with one that sends `AppExit`
event to every app with `TerminalCtrlCHandlerPlugin`.
## Testing
Tested by running the `3d_scene` example and hitting `Ctrl+C` in the
terminal.
---
## Changelog
Handles `Ctrl+C` in the terminal gracefully.
## Migration Guide
If you are overriding the `Ctrl+C` handler then you should call
`TerminalCtrlCHandlerPlugin::gracefully_exit` from your handler. It will
tell the app to exit.
Part of #13681
# Objective
gLTF Assets shouldn't be duplicated between Assets resource and node
children.
Also changed `asset_label` to be a method as [per previous PR
comment](https://github.com/bevyengine/bevy/pull/13558).
## Solution
- Made GltfNode children be Handles instead of asset copies.
## Testing
- Added tests that actually test loading and hierarchy as previous ones
unit tested only one function and that makes little sense.
- Made circular nodes an actual loading failure instead of a warning
no-op. You [_MUST NOT_ have cycles in
gLTF](https://registry.khronos.org/glTF/specs/2.0/glTF-2.0.html#nodes-and-hierarchy)
according to the spec.
- IMO this is a bugfix, not a breaking change. But in an extremely
unlikely event in which you relied on invalid behavior for loading gLTF
with cyclic children, you will not be able to do that anymore. You
should fix your gLTF file as it's not valid according to gLTF spec. For
it to for work someone, it had to be bevy with bevy_animation flag off.
---
## Changelog
### Changed
- `GltfNode.children` are now `Vec<Handle<GltfNode>>` instead of
`Vec<GltfNode>`
- Having children cycles between gLTF nodes in a gLTF document is now an
explicit asset loading failure.
## Migration Guide
If accessing children, use `Assets<GltfNode>` resource to get the actual
child object.
#### Before
```rs
fn gltf_print_first_node_children_system(gltf_component_query: Query<Handle<Gltf>>, gltf_assets: Res<Assets<Gltf>>, gltf_nodes: Res<Assets<GltfNode>>) {
for gltf_handle in gltf_component_query.iter() {
let gltf_root = gltf_assets.get(gltf_handle).unwrap();
let first_node_handle = gltf_root.nodes.get(0).unwrap();
let first_node = gltf_nodes.get(first_node_handle).unwrap();
let first_child = first_node.children.get(0).unwrap();
println!("First nodes child node name is {:?)", first_child.name);
}
}
```
#### After
```rs
fn gltf_print_first_node_children_system(gltf_component_query: Query<Handle<Gltf>>, gltf_assets: Res<Assets<Gltf>>, gltf_nodes: Res<Assets<GltfNode>>) {
for gltf_handle in gltf_component_query.iter() {
let gltf_root = gltf_assets.get(gltf_handle).unwrap();
let first_node_handle = gltf_root.nodes.get(0).unwrap();
let first_node = gltf_nodes.get(first_node_handle).unwrap();
let first_child_handle = first_node.children.get(0).unwrap();
let first_child = gltf_nodes.get(first_child_handle).unwrap();
println!("First nodes child node name is {:?)", first_child.name);
}
}
```
# Objective
Allow combining render layers with a more-ergonomic syntax than
`RenderLayers::from_iter(a.iter().chain(b.iter()))`.
## Solution
Add the `or` operation (and corresponding `const` method) to allow
computing the union of a set of render layers. While we're here, also
added `and` and `xor` operations. Someone might find them useful
## Testing
Added a simple unit test.
# Objective
We're able to reflect types sooooooo... why not functions?
The goal of this PR is to make functions callable within a dynamic
context, where type information is not readily available at compile
time.
For example, if we have a function:
```rust
fn add(left: i32, right: i32) -> i32 {
left + right
}
```
And two `Reflect` values we've already validated are `i32` types:
```rust
let left: Box<dyn Reflect> = Box::new(2_i32);
let right: Box<dyn Reflect> = Box::new(2_i32);
```
We should be able to call `add` with these values:
```rust
// ?????
let result: Box<dyn Reflect> = add.call_dynamic(left, right);
```
And ideally this wouldn't just work for functions, but methods and
closures too!
Right now, users have two options:
1. Manually parse the reflected data and call the function themselves
2. Rely on registered type data to handle the conversions for them
For a small function like `add`, this isn't too bad. But what about for
more complex functions? What about for many functions?
At worst, this process is error-prone. At best, it's simply tedious.
And this is assuming we know the function at compile time. What if we
want to accept a function dynamically and call it with our own
arguments?
It would be much nicer if `bevy_reflect` could alleviate some of the
problems here.
## Solution
Added function reflection!
This adds a `DynamicFunction` type to wrap a function dynamically. This
can be called with an `ArgList`, which is a dynamic list of
`Reflect`-containing `Arg` arguments. It returns a `FunctionResult`
which indicates whether or not the function call succeeded, returning a
`Reflect`-containing `Return` type if it did succeed.
Many functions can be converted into this `DynamicFunction` type thanks
to the `IntoFunction` trait.
Taking our previous `add` example, this might look something like
(explicit types added for readability):
```rust
fn add(left: i32, right: i32) -> i32 {
left + right
}
let mut function: DynamicFunction = add.into_function();
let args: ArgList = ArgList::new().push_owned(2_i32).push_owned(2_i32);
let result: Return = function.call(args).unwrap();
let value: Box<dyn Reflect> = result.unwrap_owned();
assert_eq!(value.take::<i32>().unwrap(), 4);
```
And it also works on closures:
```rust
let add = |left: i32, right: i32| left + right;
let mut function: DynamicFunction = add.into_function();
let args: ArgList = ArgList::new().push_owned(2_i32).push_owned(2_i32);
let result: Return = function.call(args).unwrap();
let value: Box<dyn Reflect> = result.unwrap_owned();
assert_eq!(value.take::<i32>().unwrap(), 4);
```
As well as methods:
```rust
#[derive(Reflect)]
struct Foo(i32);
impl Foo {
fn add(&mut self, value: i32) {
self.0 += value;
}
}
let mut foo = Foo(2);
let mut function: DynamicFunction = Foo::add.into_function();
let args: ArgList = ArgList::new().push_mut(&mut foo).push_owned(2_i32);
function.call(args).unwrap();
assert_eq!(foo.0, 4);
```
### Limitations
While this does cover many functions, it is far from a perfect system
and has quite a few limitations. Here are a few of the limitations when
using `IntoFunction`:
1. The lifetime of the return value is only tied to the lifetime of the
first argument (useful for methods). This means you can't have a
function like `(a: i32, b: &i32) -> &i32` without creating the
`DynamicFunction` manually.
2. Only 15 arguments are currently supported. If the first argument is a
(mutable) reference, this number increases to 16.
3. Manual implementations of `Reflect` will need to implement the new
`FromArg`, `GetOwnership`, and `IntoReturn` traits in order to be used
as arguments/return types.
And some limitations of `DynamicFunction` itself:
1. All arguments share the same lifetime, or rather, they will shrink to
the shortest lifetime.
2. Closures that capture their environment may need to have their
`DynamicFunction` dropped before accessing those variables again (there
is a `DynamicFunction::call_once` to make this a bit easier)
3. All arguments and return types must implement `Reflect`. While not a
big surprise coming from `bevy_reflect`, this implementation could
actually still work by swapping `Reflect` out with `Any`. Of course,
that makes working with the arguments and return values a bit harder.
4. Generic functions are not supported (unless they have been manually
monomorphized)
And general, reflection gotchas:
1. `&str` does not implement `Reflect`. Rather, `&'static str`
implements `Reflect` (the same is true for `&Path` and similar types).
This means that `&'static str` is considered an "owned" value for the
sake of generating arguments. Additionally, arguments and return types
containing `&str` will assume it's `&'static str`, which is almost never
the desired behavior. In these cases, the only solution (I believe) is
to use `&String` instead.
### Followup Work
This PR is the first of two PRs I intend to work on. The second PR will
aim to integrate this new function reflection system into the existing
reflection traits and `TypeInfo`. The goal would be to register and call
a reflected type's methods dynamically.
I chose not to do that in this PR since the diff is already quite large.
I also want the discussion for both PRs to be focused on their own
implementation.
Another followup I'd like to do is investigate allowing common container
types as a return type, such as `Option<&[mut] T>` and `Result<&[mut] T,
E>`. This would allow even more functions to opt into this system. I
chose to not include it in this one, though, for the same reasoning as
previously mentioned.
### Alternatives
One alternative I had considered was adding a macro to convert any
function into a reflection-based counterpart. The idea would be that a
struct that wraps the function would be created and users could specify
which arguments and return values should be `Reflect`. It could then be
called via a new `Function` trait.
I think that could still work, but it will be a fair bit more involved,
requiring some slightly more complex parsing. And it of course is a bit
more work for the user, since they need to create the type via macro
invocation.
It also makes registering these functions onto a type a bit more
complicated (depending on how it's implemented).
For now, I think this is a fairly simple, yet powerful solution that
provides the least amount of friction for users.
---
## Showcase
Bevy now adds support for storing and calling functions dynamically
using reflection!
```rust
// 1. Take a standard Rust function
fn add(left: i32, right: i32) -> i32 {
left + right
}
// 2. Convert it into a type-erased `DynamicFunction` using the `IntoFunction` trait
let mut function: DynamicFunction = add.into_function();
// 3. Define your arguments from reflected values
let args: ArgList = ArgList::new().push_owned(2_i32).push_owned(2_i32);
// 4. Call the function with your arguments
let result: Return = function.call(args).unwrap();
// 5. Extract the return value
let value: Box<dyn Reflect> = result.unwrap_owned();
assert_eq!(value.take::<i32>().unwrap(), 4);
```
## Changelog
#### TL;DR
- Added support for function reflection
- Added a new `Function Reflection` example:
ba727898f2/examples/reflection/function_reflection.rs (L1-L157)
#### Details
Added the following items:
- `ArgError` enum
- `ArgId` enum
- `ArgInfo` struct
- `ArgList` struct
- `Arg` enum
- `DynamicFunction` struct
- `FromArg` trait (derived with `derive(Reflect)`)
- `FunctionError` enum
- `FunctionInfo` struct
- `FunctionResult` alias
- `GetOwnership` trait (derived with `derive(Reflect)`)
- `IntoFunction` trait (with blanket implementation)
- `IntoReturn` trait (derived with `derive(Reflect)`)
- `Ownership` enum
- `ReturnInfo` struct
- `Return` enum
---------
Co-authored-by: Periwink <charlesbour@gmail.com>
# Objective
- When no wgpu backend is selected, there should be a clear explanation.
- Fix a regression in 0.14 when not using default features. I hit this
compile failure when trying to build bevy_framepace for 0.14.0-rc.4
```
error[E0432]: unresolved import `crate::core_3d::DEPTH_TEXTURE_SAMPLING_SUPPORTED`
--> /Users/aevyrie/.cargo/registry/src/index.crates.io-6f17d22bba15001f/bevy_core_pipeline-0.14.0-rc.4/src/dof/mod.rs:59:19
|
59 | Camera3d, DEPTH_TEXTURE_SAMPLING_SUPPORTED,
| ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ no `DEPTH_TEXTURE_SAMPLING_SUPPORTED` in `core_3d`
|
note: found an item that was configured out
--> /Users/aevyrie/.cargo/registry/src/index.crates.io-6f17d22bba15001f/bevy_core_pipeline-0.14.0-rc.4/src/core_3d/mod.rs:53:11
|
53 | pub const DEPTH_TEXTURE_SAMPLING_SUPPORTED: bool = false;
| ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
note: found an item that was configured out
--> /Users/aevyrie/.cargo/registry/src/index.crates.io-6f17d22bba15001f/bevy_core_pipeline-0.14.0-rc.4/src/core_3d/mod.rs:63:11
|
63 | pub const DEPTH_TEXTURE_SAMPLING_SUPPORTED: bool = true;
| ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
```
## Solution
- Ensure that `DEPTH_TEXTURE_SAMPLING_SUPPORTED` is either `true` or
`false`, it shouldn't be completely missing.
## Testing
- Building on WASM without default features, which now seemingly no
longer includes webgl, will panic on startup with a message saying that
no wgpu backend was selected. This is much more helpful than the compile
time failure:
```
No wgpu backend feature that is implemented for the target platform was enabled
```
- I can see an argument for making this a compile time failure, however
the current failure mode is very confusing for novice users, and
provides no clues for how to fix it. If we want this to fail at compile
time, we should do it in a way that fails with a helpful message,
similar to what this PR acheives.
# Objective
- Fixes#14059
- `morphed_skinned_mesh_layout` is the same as
`morphed_skinned_motion_mesh_layout` but shouldn't have the skin / morph
from previous frame, as they're used for motion
## Solution
- Remove the extra entries
## Testing
- Run with the glTF file reproducing #14059, it works
# Objective / Solution
Make it possible to construct `Material2dBindGroupId` for custom 2D
material pipelines by making the inner field public.
---
The 3D variant (`MaterialBindGroupId`) had this done in
e79b9b62ce
# Objective
- Fixes a correctness error introduced in
https://github.com/bevyengine/bevy/pull/14013 ...
## Solution
I've been playing around a lot of with the access code and I realized
that I introduced a soundness error when trying to simplify the code.
When we have a `Or<(With<A>, With<B>)>` filter, we cannot call
```
let mut intermediate = FilteredAccess::default();
$name::update_component_access($name, &mut intermediate);
_new_access.append_or(&intermediate);
```
because that's just equivalent to adding the new components as `Or`
clauses.
For example if the existing `filter_sets` was `vec![With<C>]`, we would
then get `vec![With<C>, With<A>, With<B>]` which translates to `A or B
or C`.
Instead what we want is `(A and B) or (A and C)`, so we need to have
each new OR clause compose with the existing access like so:
```
let mut intermediate = _access.clone();
// if we previously had a With<C> in the filter_set, this will become `With<C> AND With<A>`
$name::update_component_access($name, &mut intermediate);
_new_access.append_or(&intermediate);
```
## Testing
- Added a unit test that is broken in main, but passes in this PR
# Objective
`sickle_ui` needs `PartialEq` on components to turn them into animatable
style attributes.
## Solution
All properties of Outline is already `PartialEq`, add derive on
`Outline` as well.
## Testing
- used `sickle_ui` to test if it can be made animatable
As reported in #14004, many third-party plugins, such as Hanabi, enqueue
entities that don't have meshes into render phases. However, the
introduction of indirect mode added a dependency on mesh-specific data,
breaking this workflow. This is because GPU preprocessing requires that
the render phases manage indirect draw parameters, which don't apply to
objects that aren't meshes. The existing code skips over binned entities
that don't have indirect draw parameters, which causes the rendering to
be skipped for such objects.
To support this workflow, this commit adds a new field,
`non_mesh_items`, to `BinnedRenderPhase`. This field contains a simple
list of (bin key, entity) pairs. After drawing batchable and unbatchable
objects, the non-mesh items are drawn one after another. Bevy itself
doesn't enqueue any items into this list; it exists solely for the
application and/or plugins to use.
Additionally, this commit switches the asset ID in the standard bin keys
to be an untyped asset ID rather than that of a mesh. This allows more
flexibility, allowing bins to be keyed off any type of asset.
This patch adds a new example, `custom_phase_item`, which simultaneously
serves to demonstrate how to use this new feature and to act as a
regression test so this doesn't break again.
Fixes#14004.
## Changelog
### Added
* `BinnedRenderPhase` now contains a `non_mesh_items` field for plugins
to add custom items to.
# Objective
`StaticSystemParam` should delegate all `SystemParam` methods to the
inner param, but it looks like it was missed when the new `queue()`
method was added in #10839.
## Solution
Implement `StaticSystemParam::queue()` to delegate to the inner param.
The comment was incorrect - we are already looking at the pyramid
texture so we do not need to transform the size in any way. Doing that
resulted in a mip that was too fine to be selected in certain cases,
which resulted in a 2x2 pixel footprint not actually fully covering the
cluster sphere - sometimes this could lead to a non-conservative depth
value being computed which resulted in the cluster being marked as
invisible incorrectly.
This change updates meshopt-rs to 0.3 to take advantage of the newly
added sparse simplification mode: by default, simplifier assumes that
the entire mesh is simplified and runs a set of calculations that are
O(vertex count), but in our case we simplify many small mesh subsets
which is inefficient.
Sparse mode instead assumes that the simplified subset is only using a
portion of the vertex buffer, and optimizes accordingly. This changes
the meaning of the error (as it becomes relative to the subset, in our
case a meshlet group); to ensure consistent error selection, we also use
the ErrorAbsolute mode which allows us to operate in mesh coordinate
space.
Additionally, meshopt 0.3 runs optimizeMeshlet automatically as part of
`build_meshlets` so we no longer need to call it ourselves.
This reduces the time to build meshlet representation for Stanford Bunny
mesh from ~1.65s to ~0.45s (3.7x) in optimized builds.
# Objective
Implement `BuildChildrenTransformExt` for `EntityWorldMut`, which is
useful when working directly with a mutable `World` ref.
## Solution
I realize this isn't the most optimal implementation in that it doesn't
reuse the existing entity location for the child, but it is terse and
reuses the existing code. I can address that if needed.
## Testing
I only tested locally. There are no tests for `set_parent_in_place` and
`remove_parent_in_place` currently, but I can add some.
---
## Changelog
`BuildChildrenTransformExt` implemented for `EntityWorldMut`.
# Objective
When using combinators such as `EntityCommand::with_entity` to build
commands, it can be easy to forget to apply that command, leading to
dead code. In many cases this doesn't even lead to an unused variable
warning, which can make these mistakes difficult to track down
## Solution
Annotate the method with `#[must_use]`
Co-authored-by: Alice Cecile <alice.i.cecile@gmail.com>
# Objective
- Second part of #13900
- based on #13905
## Solution
- check_dir_light_mesh_visibility defers setting the entity's
`ViewVisibility `so that Bevy can schedule it to run in parallel with
`check_point_light_mesh_visibility`.
- Reduce HashMap lookups for directional light checking as much as
possible
- Use `par_iter `to parallelize the checking process within each system.
---------
Co-authored-by: Kristoffer Søholm <k.soeholm@gmail.com>
# Objective
The trait method `with_entity` is used to add an `EntityCommand` to the
command queue. Currently this method returns `WithEntity<C>` which pairs
a command with an `Entity`. By replacing this explicit type with an
opaque type, implementors can override this default implementation by
returning a custom command or closure that does the same thing with a
lower memory footprint.
# Solution
Return an opaque type from the method. As a bonus this file is now
cleaner without the `WithEntity` boilerplate
# Objective
Numerous people have been confused that Bevy runs slowly, when the
reason is that the `llvmpipe` software rendered is being used.
## Solution
Printing a warning could reduce the confusion.
# Objective
- Fixes#13728
## Solution
- add a new feature `smaa_luts`. if enables, it also enables `ktx2` and
`zstd`. if not, it doesn't load the files but use placeholders instead
- adds all the resources needed in the same places that system that uses
them are added.
# Objective
Fixes https://github.com/bevyengine/bevy/issues/13993
PR inspired by https://github.com/bevyengine/bevy/pull/14007 to
accomplish the same thing, but maybe in a clearer fashion.
@Gingeh feel free to take my changes and add them to your PR, I don't
want to steal any credit
---------
Co-authored-by: Gingeh <39150378+Gingeh@users.noreply.github.com>
Co-authored-by: Bob Gardner <rgardner@inworld.ai>
Co-authored-by: Martín Maita <47983254+mnmaita@users.noreply.github.com>
# Objective
In Bevy 0.13, `BackgroundColor` simply tinted the image of any
`UiImage`. This was confusing: in every other case (e.g. Text), this
added a solid square behind the element. #11165 changed this, but
removed `BackgroundColor` from `ImageBundle` to avoid confusion, since
the semantic meaning had changed.
However, this resulted in a serious UX downgrade / inconsistency, as
this behavior was no longer part of the bundle (unlike for `TextBundle`
or `NodeBundle`), leaving users with a relatively frustrating upgrade
path.
Additionally, adding both `BackgroundColor` and `UiImage` resulted in a
bizarre effect, where the background color was seemingly ignored as it
was covered by a solid white placeholder image.
Fixes#13969.
## Solution
Per @viridia's design:
> - if you don't specify a background color, it's transparent.
> - if you don't specify an image color, it's white (because it's a
multiplier).
> - if you don't specify an image, no image is drawn.
> - if you specify both a background color and an image color, they are
independent.
> - the background color is drawn behind the image (in whatever pixels
are transparent)
As laid out by @benfrankel, this involves:
1. Changing the default `UiImage` to use a transparent texture but a
pure white tint.
2. Adding `UiImage::solid_color` to quickly set placeholder images.
3. Changing the default `BorderColor` and `BackgroundColor` to
transparent.
4. Removing the default overrides for these values in the other assorted
UI bundles.
5. Adding `BackgroundColor` back to `ImageBundle` and `ButtonBundle`.
6. Adding a 1x1 `Image::transparent`, which can be accessed from
`Assets<Image>` via the `TRANSPARENT_IMAGE_HANDLE` constant.
Huge thanks to everyone who helped out with the design in the linked
issue and [the Discord
thread](https://discord.com/channels/691052431525675048/1255209923890118697/1255209999278280844):
this was very much a joint design.
@cart helped me figure out how to set the UiImage's default texture to a
transparent 1x1 image, which is a much nicer fix.
## Testing
I've checked the examples modified by this PR, and the `ui` example as
well just to be sure.
## Migration Guide
- `BackgroundColor` no longer tints the color of images in `ImageBundle`
or `ButtonBundle`. Set `UiImage::color` to tint images instead.
- The default texture for `UiImage` is now a transparent white square.
Use `UiImage::solid_color` to quickly draw debug images.
- The default value for `BackgroundColor` and `BorderColor` is now
transparent. Set the color to white manually to return to previous
behavior.
# Objective
- Fixes#13811 (probably, I lost my test code...)
## Solution
- Turns out that Queue and PrepareAssets are _not_ ordered. We should
probably either rethink our system sets (again), or improve the
documentation here. For reference, I've included the current ordering
below.
- The `prepare_meshlet_meshes_X` systems need to run after
`prepare_assets::<PreparedMaterial<M>>`, and have also been moved to
QueueMeshes.
```rust
schedule.configure_sets(
(
ExtractCommands,
ManageViews,
Queue,
PhaseSort,
Prepare,
Render,
Cleanup,
)
.chain(),
);
schedule.configure_sets((ExtractCommands, PrepareAssets, Prepare).chain());
schedule.configure_sets(QueueMeshes.in_set(Queue).after(prepare_assets::<GpuMesh>));
schedule.configure_sets(
(PrepareResources, PrepareResourcesFlush, PrepareBindGroups)
.chain()
.in_set(Prepare),
);
```
## Testing
- Ambiguity checker to make sure I don't have ambiguous system ordering
# Objective
#12469 changed the `Debug` impl for `Entity`, making sure it's actually
accurate for debugging. To ensure that its can still be readily logged
in error messages and inspectors, this PR added a more concise and
human-friendly `Display` impl.
However, users found this form too verbose: the `to_bits` information
was unhelpful and too long. Fixes#13980.
## Solution
- Don't include `Entity::to_bits` in the `Display` implementation for
`Entity`. This information can readily be accessed and logged for users
who need it.
- Also clean up the implementation of `Display` for `DebugName`,
introduced in https://github.com/bevyengine/bevy/pull/13760, to simply
use the new `Display` impl (since this was the desired format there).
## Testing
I've updated an existing test to verify the output of `Entity::display`.
---------
Co-authored-by: Kristoffer Søholm <k.soeholm@gmail.com>
# Objective
- Fix issue #13821
## Solution
- Rewrote the test to ensure that it actually tests the functionality
correctly. Then made the par_read function correctly change the values
of self.reader.last_event_count.
## Testing
- Rewrote the test for par_read to run the system schedule twice,
checking the output each time
---------
Co-authored-by: Martín Maita <47983254+mnmaita@users.noreply.github.com>
# Objective
- Fixes https://github.com/bevyengine/bevy/issues/14003
## Solution
- Expose an API to perform updates for a specific sub-app, so we can
avoid mutable borrow the app twice.
## Testing
- I have tested the API by modifying the code in the `many_lights`
example with the following changes:
```rust
impl Plugin for LogVisibleLights {
fn build(&self, app: &mut App) {
let Some(render_app) = app.get_sub_app_mut(RenderApp) else {
return;
};
render_app.add_systems(Render, print_visible_light_count.in_set(RenderSet::Prepare));
}
fn finish(&self, app: &mut App) {
app.update_sub_app_by_label(RenderApp);
}
}
```
---
## Changelog
- add the `update_sub_app_by_label` API to `App` and `SubApps`.
---------
Co-authored-by: Jan Hohenheim <jan@hohenheim.ch>
# Objective
- When writing "in game" debugging tools, quite often you need the name
of an entity (for example an entity tree). DebugName is the usual way of
doing that.
- A recent change to Entity's Debug implementation meant it was no
longer a minimal {index}v{generation} but instead a more verbose auto
generated Debug.
- This made DebugName's Debug implementation also verbose
## Solution
- I changed DebugName to derive Debug automatically and added a new
(preferred) Display implementation for it which is the preferred name
for an entity. If the entity has a Name component its the contents of
that, otherwise it is {index}v{generation} (though this does not use
Display of the Entity as that is more verbose than this).
## Testing
- I've added a new test in name.rs which tests the Display
implementation for DebugName by using to_string.
---
## Migration Guide
- In code which uses DebugName you should now use the Display
implementation rather than the Debug implementation (ie {} instead of
{:?} if you were printing it out).
---------
Co-authored-by: John Payne <20407779+johngpayne@users.noreply.github.com>
Co-authored-by: Jan Hohenheim <jan@hohenheim.ch>
Co-authored-by: Andres O. Vela <andresovela@users.noreply.github.com>
fixes#13944
I literally just added `Did you forget to call SystemState::apply?` to
the error message. I tested it with the code snipped from the Issue and
yeah it works
# Objective
Fixes#13982
## Solution
~~Adds a new field to `bevy_log::LogPlugin`: `ansi: bool`~~
Documents the use of `std::env::set_var("NO_COLOR", "1");` to disable
colour output in terminals.
## Testing
Yes, all tests passed when I ran `cargo run -p ci -- test` and `cargo
run -p ci -- lints`
I have only tested the code on my Mac, though I doubt this change would
have any affect on other platforms.
---
# Objective
Closes#5943. Seems like Assets v2 solved this one.
## Solution
Added a test to confirm that using `Reflect::clone_value` and then
`FromReflect::from_reflect` on a `Handle<T>` both increment the strong
count.
## Testing
A new test was added to confirm behavior.
---------
Co-authored-by: Gino Valente <49806985+MrGVSV@users.noreply.github.com>
# Objective
Tight, in-frame generation, re-parenting, despawning, etc., UI
operations could sometime lead taffy to panic (invalid SlotMap key used)
when an entity with an invalid state later despawned.
Fixes#12403
## Solution
Move the `remove_entities` call after children updates.
## Testing
`sickle_ui` had a case that always caused the panic. Tested before this
change, after this change, and before the change again to make sure the
error is there without the fix. The fix worked. Test steps and used
commit described in issue #12403.
I have also ran every bevy UI example, though none of them deal with
entity re-parenting or removal. No regression detected on them.
Tested on Windows only.
* Fixes https://github.com/bevyengine/bevy/issues/13813
* Fixes https://github.com/bevyengine/bevy/issues/13810
Tested a combined scene with both regular meshes and meshlet meshes
with:
* Regular forward setup
* Forward + normal/motion vector prepasses
* Deferred (with depth prepass since that's required)
* Deferred + depth/normal/motion vector prepasses
Still broken:
* Using meshlet meshes rendering in deferred and regular meshes
rendering in forward + depth/normal prepass. I don't know how to fix
this at the moment, so for now I've just add instructions to not mix
them.
# Objective
`with_event` will result in unsafe casting of event data of the given
type to the type expected by the Observer system. This is inherently
unsafe.
## Solution
Flag `Observer::with_event` and `ObserverDescriptor::with_events` as
unsafe. This will not affect normal workflows as `with_event` is
intended for very specific (largely internal) use cases.
This _should_ be backported to 0.14 before release.
---
## Changelog
- `Observer::with_event` is now unsafe.
- Rename `ObserverDescriptor::with_triggers` to
`ObserverDescriptor::with_events` and make it unsafe.
# Objective
- Fixes#13702
- When creating a new window, its scale was changed to match the one
returned by winit, but its size was not which resulted in an incorrect
size until the event with the correct size was received, at least 1
frame later
## Solution
- Apply the window scale to its size when creating it
I updated my 'main' branch, which accidentally closed the original PR
#13747. I'm reopening the this from an actual branch on my repo like I
should have done in the first place. Here's the original info from the
first PR:
* * *
# Problem
The `file_watcher` feature panics if the file_watcher's path "assets" is
not present. I stumbled upon this behavior when I was actually testing
against `embedded_watcher`. I had no "assets" directory and didn't need
one for [my project](https://github.com/shanecelis/bevy_plane_cut).
```text
$ cargo run --example simple; # Runs fine.
$ cargo run --example simple --feature embedded_watcher; # Panics
thread 'main' panicked at /Users/shane/.cargo/registry/src/index.crates.io-6f17d22bba15001f/bevy_asset-0.14.0-rc.2/src/io/source.rs:503:21:
Failed to create file watcher from path "assets", Error { kind: PathNotFound, paths: ["/Users/shane/Projects/bevy_plane_cut/assets"] }
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace
```
# Opinion
If a project runs without panicing, then adding the `file_watcher`
feature shouldn't cause it to panic.
# Suggested Solution
This PR suggests if the "assets" path does not exist, emit a warning
stating that the file watcher could not be created and why. All other
errors will be treated as before with a panic and a message.
```text
$ cargo run --example simple --feature embedded_watcher; # Panics
2024-06-08T08:55:11.385249Z WARN bevy_asset::io::source: Skip creating file watcher because path "assets" does not exist.
2024-06-08T08:55:11.385291Z WARN bevy_asset::io::source: AssetSourceId::Default does not have an AssetWatcher configured. Consider enabling the `file_watcher` feature.
```
The second warning is new and I'd prefer it didn't emit under this
condition, but I'll wait to see whether this is actually regarded as a
bug.
# Testing
No tests added. Compiled against my project and it demonstrated the
suggested behavior.
* * *
I changed the second warning to the following when the `file_watcher`
feature is present. When it's not present, it uses the same warning as
before.
```
024-06-09T01:22:16.880619Z WARN bevy_asset::io::source: Skip creating file watcher because path "assets" does not exist.
2024-06-09T01:22:16.880660Z WARN bevy_asset::io::source: AssetSourceId::Default does not have an AssetWatcher configured. Consider adding an "assets" directory.
```
This is an attempt to address issue #13725, which was about the
geometric primitives in the bevy_math crate lacking some detail in the
docs.
# Objective
Fixes#13725
## Solution
Added details to the docstrings. Mostly this consisted of specifying
that the primitives are centered on the origin, or describing how
they're defined (e.g., a circle is the set of all points some distance
from the origin).
## Testing
No testing, since the only changes were to docs.
Fixes#13701
After `winit` upgrade to `0.31`, windows were no longer correctly
resizing. This appears to just have been a simple mistake, where the new
physical size was being sourced from the `winit` window rather than on
the incoming `Window` component.
## Testing
Tested on macOS, but I'm curious whether this was also broken on other
platforms.
# Objective
Fixes#13866
## Solution
Add `insert_before` in **FixedMainScheduleOrder** and
**MainScheduleOrder**, add `insert_startup_before` in
**MainScheduleOrder**, applying the same logic as `insert_after`, except
for parameters naming and insertion index.
This is a followup to https://github.com/bevyengine/bevy/pull/13904
based on the discussion there, and switches two HashMaps that used
meshlet ids as keys to Vec.
In addition to a small further performance boost for `from_mesh` (1.66s
=> 1.60s), this makes processing deterministic modulo threading issues
wrt CRT rand described in the linked PR. This is valuable for debugging,
as you can visually or programmatically inspect the meshlet distribution
before/after making changes that should not change the output, whereas
previously every asset rebuild would change the meshlet structure.
Tested with https://github.com/bevyengine/bevy/pull/13431; after this
change, the visual output of meshlets is consistent between asset
rebuilds, and the MD5 of the output GLB file does not change either,
which was not the case before.
# Objective
- Make primitive meshing behavior consisten across platforms
- Avoid using sizes bigger than `u32` since these aren't even supported
for meshes
## Solution
- Use `u32` instead of `usize` for resolution/subdivisions/segments/etc
fields
---
## Changelog
- Change resolutions in primitive mesh builders from `usize` to `u32`
## Migration Guide
- All primitive mesh builders now take `u32` instead of `usize` for
their resolution/subdivision/segment counts
# Objective
Fixes#13920
## Solution
As described in the issue.
## Testing
Moved a custom transition plugin in example before any of the app-state
methods.
# Objective
Fixes#13933.
## Solution
Changed the return type.
## Testing
Fixed and reused the pre-existing tests for `inspect_entity`.
---
## Migration Guide
- `World::inspect_entity` now returns an `Iterator` instead of a `Vec`.
If you need a `Vec`, immediately collect the iterator:
`world.inspect_entity(entity).collect<Vec<_>>()`
# Objective
Fixes#13845
## Solution
Fix inline docs links inside `init_state` and `insert_state`.
## Testing
- Did you test these changes? If so, how?
Manually checked on `cargo doc` and `rust-analyzer lsp`.
# Objective
Fix a 9-slice asymmetric border issue that
[QueenOfSquiggles](https://blobfox.coffee/@queenofsquiggles/112639035165575222)
found. Here's the behavior before:
<img width="340" alt="the-bug"
src="https://github.com/bevyengine/bevy/assets/54390/81ff1847-b2ea-4578-9fd0-af6ee96c5438">
## Solution
Here's the behavior with the fix.
<img width="327" alt="the-fix"
src="https://github.com/bevyengine/bevy/assets/54390/33a4e3f0-b6a8-448e-9654-1197218ea11d">
## Testing
I used QueenOfSquiggles
[repo](https://github.com/QueenOfSquiggles/my-bevy-learning-project) to
exercise the code. I manually went through a number of variations of the
border and caught a few other issues after the first pass. I added some
code to create random borders and though they often looked funny there
weren't any gaps like before.
### Unit Tests
I did add some tests to `slicer.rs` mostly as an exploratory programming
exercise. So they currently act as a limited, incomplete,
"golden-file"-ish approach. Perhaps they're not worth keeping.
In order to write the tests, I did add a `PartialEq` derive for
`TextureSlice`.
I only tested these changes on macOS.
---
## Changelog
Make 9-slice textures work with asymmetric borders.
# Objective
- Make gizmos behavior consistent across platforms
## Solution
- Use `u32` instead of `usize` for resolution/subdivisions/segments/etc
fields
---
## Changelog
- Change resolutions in gizmos from `usize` to `u32`
## Migration Guide
- All gizmos now take `u32` instead of `usize` for their
resolution/subdivision/segment counts
# Objective
- Primitives should not use poorly defined types like `usize`,
especially since they are serializable
## Solution
- Use `u32` instead of `usize`
- The generic array types do not need to be changed because this size is
not actually stored or serialized anywhere
---
## Migration Guide
- `RegularPolygon` now uses `u32` instead of `usize` for the number of
sides
# Objective
This PR aims to improve error handling for log filters.
Closes#13850
## Solution
I changed the parsing of LogPlugin its filter to lossy, so that it
prints the directives with an error but does not skip them. I decided on
letting it gracefully handle the error instead of panicking to be
consistent with the parsing from an environment variable that it tries
to do before parsing it from the LogPlugin filter.
If the user decides to specify the filter by an environment variable, it
would silently fail and default to the LogPlugin filter value. It now
prints an error before defaulting to the LogPlugin filter value.
Unfortunately, I could not try and loosely set the filter from the
environment variable because the `tracing-subscriber` module does not
expose the function it uses to get the environment variable, and I would
rather not copy its code. We may want to check if the maintainers are
open to exposing the method.
## Testing
Consider the following bevy app, where the second of the 3 filters is
invalid:
```
use bevy::{log::LogPlugin, prelude::*};
fn main() {
App::new().add_plugins(DefaultPlugins
.set(LogPlugin {
filter: "wgpu=error,my_package=invalid_log_level,naga=warn".into(),
..default()
})
).run();
}
```
In the previous situation, it would panic with a non-descriptive error:
"called `Result::unwrap()` on an `Err` value: ParseError { kind:
Other(None) }", while only 1 of the 3 filters is invalid. When running
`cargo run`, it will now use the two valid filters and print an error on
the invalid filter.
> ignoring `my_package=invalid_log_level`: invalid filter directive
This error comes from `tracing-subscriber` and cannot be altered as far
as I can see.
To test setting the log filter through an environment variable, you can
use `RUST_LOG="wgpu=error,my_package=invalid_log_level,naga=warn" cargo
run` to run your app. In the previous situation it would silently fail
and use the LogPlugin filter. It will now print an error before using
the LogPlugin filter.
> LogPlugin failed to parse filter from env: invalid filter directive
## Changelog
- Added warning when using invalid filter in the RUST_LOG environment
variable
- Prevent the app from panicking when setting an invalid LogPlugin
filter
---------
Co-authored-by: Luc Drenth <luc.drenth@ing.com>
# Objective
Fixes#13917
## Solution
Changed `IntoSystemSetConfigs::chain_ignore_deferred`'s return type from
`SystemConfigs` to `SystemSetConfigs`
## Testing
Tried to run the `ecs_guide` example, where `chain` method is replaced
by `chain_ignore_deferred` method
---
# Objective
- Fix a typo in documentation for `Query::single_mut`
## Solution
- Change `item` to `items`
## Testing
- I built the documentation and it looked fine.
- Since this only affects a doc comment, no further testing should be
necessary.
---
## Changelog
> This section is optional. If this was a trivial fix, or has no
externally-visible impact, you can delete this section.
- Fixed a typo in the documentation for Query.
# Objective
- Fixes#11933.
- Related: #12280.
## Solution
- Specify that, after applying `AmbientLight`, the resulting units are
in cd/m^2.
- This is based on [@fintelia's
comment](https://github.com/bevyengine/bevy/issues/11933#issuecomment-1995427587),
and will need to be verified.
---
## Changelog
- Specified units for `AmbientLight`'s `brightness` field.
This change reworks `find_connected_meshlets` to scale more linearly
with the mesh size, which significantly reduces the cost of building
meshlet representations. As a small extra complexity reduction, it moves
`simplify_scale` call out of the loop so that it's called once (it only
depends on the vertex data => is safe to cache).
The new implementation of connectivity analysis builds edge=>meshlet
list data structure, which allows us to only iterate through
`tuple_combinations` of a (usually) small list. There is still some
redundancy as if two meshlets share two edges, they will be represented
in the meshlet lists twice, but it's overall much faster.
Since the hash traversal is non-deterministic, to keep this part of the
algorithm deterministic for reproducible results we sort the output
adjacency lists.
Overall this reduces the time to process bunny mesh from ~4.2s to ~1.7s
when using release; in unoptimized builds the delta is even more
significant.
This was tested by using https://github.com/bevyengine/bevy/pull/13431
and:
a) comparing the result of `find_connected_meshlets` using old and new
code; they are equal in all steps of the clustering process
b) comparing the rendered result of the old code vs new code *after*
making the rest of the algorithm deterministic: right now the loop that
iterates through the result of `group_meshlets()` call executes in
different order between program runs. This is orthogonal to this change
and can be fixed separately.
Note: a future change can shrink the processing time further from ~1.7s
to ~0.4s with a small diff but that requires an update to meshopt crate
which is pending in https://github.com/gwihlidal/meshopt-rs/pull/42.
This change is independent.
# Objective
- first part of #13900
## Solution
- split `check_light_mesh_visibility `into
`check_dir_light_mesh_visibility `and
`check_point_light_mesh_visibility` for better review
# Objective
- After #12582 , Bevy split visibleEntities into a TypeIdMap for
different types of entities, but the behavior in
`check_light_mesh_visibility `simply calls HashMap::clear(), which will
reallocate memory every frame.
## Testing
cargo run --release --example many_cubes --features bevy/trace_tracy --
--shadows
~10% win in `check_light_mesh_visibilty`
# Objective
The error messages that appear when a value cannot be serialized or
deserialized via reflection could be slightly improved.
When one of these operations fails, some users are confused about how to
resolve the issue. I've spoken with a few who didn't know they could
register `ReflectSerialize` themselves. We should try to clarify this to
some degree in the error messages.
## Solution
Add some more detail to the error messages.
For example, replacing this:
```
Type 'core::ops::RangeInclusive<f32>' did not register ReflectSerialize
```
with this:
```
Type `core::ops::RangeInclusive<f32>` did not register the `ReflectSerialize` type data. For certain types, this may need to be registered manually using `register_type_data`
```
I also added a separate error message if the type was not registered in
the type registry at all:
```
Type `core::ops::RangeInclusive<f32>` is not registered in the type registry
```
## Testing
You can test locally by running:
```
cargo test --package bevy_reflect
```
---
## Changelog
- Added error message for missing type registration when serializing
reflect data
- Changed error message for missing `ReflectSerialize` registration when
serializing reflect data
- Changed error message for missing `ReflectDeserialize` registration
when deserializing reflect data
# Objective
I got little confused by the document of `all_tuples!` because type
names of the parameter `T` and extracted names `Pn` are difference.
## Solution
I fixed type names of the document.
# Objective
While writing code for the `bevy_ecs` I noticed we were using a
unnecessarily stable sort to sort component ids
## Solution
- Sort component ids with a unstable sort
- Comb the bevy_ecs crate for any other obvious inefficiencies.
- Don't clone component vectors when inserting an archetype.
## Testing
I ran `cargo test -p bevy_ecs`. Everything else I leave to CI.
## Profiling
I measured about a 1% speed increase when spawning entities directly
into a world. Since the difference is so small (and might just be noise)
I didn't bother to figure out which of change if any made the biggest
difference.
<details>
<summary> Tracy data </summary>
Yellow is this PR. Red is the commit I branched from.
</details>
<details>
<summary>Methodology</summary>
I created a system that spawn a 1000 entities each with the same 30
components each frame, and then I measured it's run time. The unusually
high number of components was chosen because the standard library [will
use a insertion sort for slices under 20
elements](0de24a5177/library/core/src/slice/sort.rs (L1048-L1049)).
This holds for both stable and unstable sorts.
</details>
# Objective
Fixes#13299
On Linux/X11, changing focus into a winit window will produce winit
KeyboardInput events with a "is_synthetic=true" flag that are not
intended to be used. Bevy erroneously passes them on to the user,
resulting in phantom key presses.
## Solution
This patch properly filters out winit KeyboardInput events with
"is_synthetic=true".
For example, pressing Alt+Tab to focus a bevy winit window results in a
permanently stuck Tab key until the user presses Tab once again to
produce a winit KeyboardInput release event. The Tab key press event
that causes this problem is "synthetic", should not be used according to
the winit devs, and simply ignoring it fixes this problem.
Synthetic key **releases** are still evaluated though, as they are
essential for correct release key handling. For example, if the user
binds the key combination Alt+1 to the action "move the window to
workspace 1", places the bevy game in workspace 2, focuses the game and
presses Alt+1, then the key release event for the "1" key will be
synthetic. If we would filter out all synthetic keys, the bevy game
would think that the 1 key remains pressed forever, until the user
manually presses+releases the key again inside bevy.
Reference:
https://docs.rs/winit/0.30.0/winit/event/enum.WindowEvent.html#variant.KeyboardInput.field.is_synthetic
Relevant discussion: https://github.com/rust-windowing/winit/issues/3543
## Testing
Tested with the "keyboard_input_events" example. Entering/exiting the
window with various keys, as well as changing its workspace, produces
the correct press/release events.
# Objective
- Fixes#13874
## Solution
- Confirm that the `StatesPlugin` is installed when trying to add
states.
- Skipped for state scoped entities, since those will warn about missing
states.
# Objective
Fixes#13854
## Solution
Removed the inaccurate warning. This was done for a few reasons:
- States not existing is now a valid "state" (for lack of a better term)
- Other run conditions don't provide an equivalent warning
# Objective
Allow the use of color definitions from Bevy in other contexts than pure
Bevy apps, e.g. outside ECS use.
In those cases it's nice to not have more dependencies than you need.
## Solution
Hide use of reflection behind a feature flag.
Defaults to on.
## Points to consider
1. This was straightforward _except_ for the
`crates/bevy_color/src/lib.rs` change where I removed `Reflect` as a
bound. That is awkward to have feature gated since features should be
additive. If the bound was added as part of the feature flag, the result
would be _more_ restrictive, and _disable_ impls which did not have the
impl. On the other hand having the reflect bound there unconditionally
would defeat the purpose of the PR. I opted to remove the bound since it
seems overly restrictive anyway.
2. It's possible to hide `encase` and `bytemuck` behind the new feature
flag too (or a separate one). I'm thinking if `bevy-support` is not
desired then it's unlikely that the user has need of those.
---------
Signed-off-by: Torstein Grindvik <torstein.grindvik@muybridge.com>
Co-authored-by: Torstein Grindvik <torstein.grindvik@muybridge.com>
# Objective
- #13846 introduced a bug where text not bound was not displayed
## Solution
- bounds are infinite
- use computed size instead, that already should be using the available
bounds
