Commit graph

3652 commits

Author SHA1 Message Date
Zicklag
f9c56b321d Enable Constructing ReflectComponent/Resource (#6257)
# Objective

- Fixes #6206

## Solution

- Create a constructor for creating `ReflectComponent` and `ReflectResource`

---

## Changelog

> This section is optional. If this was a trivial fix, or has no externally-visible impact, you can delete this section.

### Added

- Created constructors for `ReflectComponent` and `ReflectResource`, allowing for advanced scripting use-cases.
2022-10-17 14:01:50 +00:00
JoJoJet
89c4b77bdd Add a method for accessing the width of a Table (#6249)
# Objective

There is currently no good way of getting the width (# of components) of a table outside of `bevy_ecs`.

# Solution

Added the methods `Table::{component_count, component_capacity}`
For consistency and clarity, renamed `Table::{len, capacity}` to `entity_count` and `entity_capacity`.

## Changelog

- Added the methods `Table::component_count` and `Table::component_capacity`
- Renamed `Table::len` and `Table::capacity` to `entity_count` and `entity_capacity`

## Migration Guide

Any use of `Table::len` should now be `Table::entity_count`. Any use of `Table::capacity` should now be `Table::entity_capacity`.
2022-10-17 13:47:02 +00:00
Lena Milizé
73605f43b6 Replace the bool argument of Timer with TimerMode (#6247)
As mentioned in #2926, it's better to have an explicit type that clearly communicates the intent of the timer mode rather than an opaque boolean, which can be only understood when knowing the signature or having to look up the documentation.

This also opens up a way to merge different timers, such as `Stopwatch`, and possibly future ones, such as `DiscreteStopwatch` and `DiscreteTimer` from #2683, into one struct.

Signed-off-by: Lena Milizé <me@lvmn.org>

# Objective

Fixes #2926.

## Solution

Introduce `TimerMode` which replaces the `bool` argument of `Timer` constructors. A `Default` value for `TimerMode` is `Once`.

---

## Changelog

### Added

- `TimerMode` enum, along with variants `TimerMode::Once` and `TimerMode::Repeating`

### Changed

- Replace `bool` argument of `Timer::new` and `Timer::from_seconds` with `TimerMode`
- Change `repeating: bool` field of `Timer` with `mode: TimerMode`

## Migration Guide

- Replace `Timer::new(duration, false)` with `Timer::new(duration, TimerMode::Once)`.
- Replace `Timer::new(duration, true)` with `Timer::new(duration, TimerMode::Repeating)`.
- Replace `Timer::from_seconds(seconds, false)` with `Timer::from_seconds(seconds, TimerMode::Once)`.
- Replace `Timer::from_seconds(seconds, true)` with `Timer::from_seconds(seconds, TimerMode::Repeating)`.
- Change `timer.repeating()` to `timer.mode() == TimerMode::Repeating`.
2022-10-17 13:47:01 +00:00
mike
a0f1468108 Add iter_entities to World #6228 (#6242)
# Objective

- Add a way to iterate over all entities from &World

## Solution

- Added a function `iter_entities` on World which returns an iterator of `Entity` derived from the entities in the `World`'s `archetypes`

---

## Changelog

- Added a function `iter_entities` on World, allowing iterating over all entities in contexts where you only have read-only access to the World.
2022-10-17 13:47:00 +00:00
Sergi-Ferrez
05c7babba2 Clarify bevy::ui::Node field and documentation (#5995)
# Objective
Fixes #5820

## Solution

Change field name and documentation from `bevy::ui::Node` struct

---

## Changelog

`bevy::ui::Node` `size` field has renamed to `calculated_size`

## Migration Guide

All references to the old `size` name has been changed, to access `bevy::ui::Node` `size` field use `calculated_size`
2022-10-17 13:27:24 +00:00
ira
92ba6224b9 Use SpatialBundle/TransformBundle in examples (#6002)
Does what it do

Co-authored-by: devil-ira <justthecooldude@gmail.com>
2022-10-13 12:53:18 +00:00
Martin Svanberg
000e6e2874 Fix documentation for looking_at/look_at (#4696)
Bevy's coordinate system is right-handed Y up, so +Z points towards my nose and I'm looking in the -Z direction. Therefore, `Transform::looking_at/look_at` must be pointing towards -Z. Or am I wrong here?
2022-10-12 20:18:20 +00:00
François
ccf7c65a78 dynamic scene builder (#6227)
# Objective

- make it easier to build dynamic scenes

## Solution

- add a builder to create a dynamic scene from a world. it can extract an entity or an iterator of entities
- alternative to #6013, leaving the "hierarchy iteration" part to #6185 which does it better
- alternative to #6004 
- using a builder makes it easier to chain several extractions
2022-10-12 01:48:16 +00:00
Alice Cecile
c0a93aa7a4 Rename system chaining to system piping (#6230)
# Objective

> System chaining is a confusing name: it implies the ability to construct non-linear graphs, and suggests a sense of system ordering that is only incidentally true. Instead, it actually works by passing data from one system to the next, much like the pipe operator.

> In the accepted [stageless RFC](https://github.com/bevyengine/rfcs/blob/main/rfcs/45-stageless.md), this concept is renamed to piping, and "system chaining" is used to construct groups of systems with ordering dependencies between them.

Fixes #6225.

## Changelog

System chaining has been renamed to system piping to improve clarity (and free up the name for new ordering APIs). 

## Migration Guide

The `.chain(handler_system)` method on systems is now `.pipe(handler_system)`.
The `IntoChainSystem` trait is now `IntoPipeSystem`, and the `ChainSystem` struct is now `PipeSystem`.
2022-10-11 15:21:12 +00:00
Michel van der Hulst
6ce7ce208e Change UI coordinate system to have origin at top left corner (#6000)
# Objective
Fixes #5572

## Solution

Approach is to invert the Y-axis of the UI Camera by changing the UI projection matrix to render the UI upside down.

After that I'm trying to fix all issues, that pop up:
- interaction expected the "old" position
- images and text were displayed upside-down
- baseline of text was based on the top of the glyph instead of bottom

... probably a lot more.

---

Result when running examples:
<details>
    <summary>Button example</summary>

main branch:
![button main](https://user-images.githubusercontent.com/4232644/190856087-61dd1d98-42b5-4238-bd97-149744ddfeba.png)
this pr:
![button pr](https://user-images.githubusercontent.com/4232644/190856097-3f4bc97a-ed15-4e97-b7f1-2b2dd6bb8b14.png)

</details>

<details>
    <summary>Text example</summary>

m
![text main](https://user-images.githubusercontent.com/4232644/192142831-4cf19aa1-f49a-485e-af7b-374d6f5c396c.png)
ain branch: 


this pr:
![text pr fixed](https://user-images.githubusercontent.com/4232644/192142829-c433db3b-32e1-4ee8-b493-0b4a4d9c8e70.png)


</details>

<details>
    <summary>Text debug example</summary>

main branch:
![text_debug main](https://user-images.githubusercontent.com/4232644/192142822-940aefa6-e502-410b-8da4-5570f77b5df2.png)

this pr:
![text_debug pr fixed](https://user-images.githubusercontent.com/4232644/194547010-8c968f5c-5a71-4ffc-871d-790c06d48016.png)

</details>

<details>
    <summary>Transparency UI example</summary>

main branch:
![transparency_ui main](https://user-images.githubusercontent.com/4232644/190856172-328c60fe-3622-4598-97d5-2f1595db13b3.png)


this pr:
![transperency_ui pr](https://user-images.githubusercontent.com/4232644/190856179-a2dafb99-41ea-45a9-9dd6-400fa3ef24b9.png)

</details>

<details>
    <summary>UI example</summary>

**ui example**
main branch:
![ui main](https://user-images.githubusercontent.com/4232644/192142812-e20ba31a-6841-46d9-a785-4198cf22dc99.png)

this pr:
![ui pr fixed](https://user-images.githubusercontent.com/4232644/192142788-cc0b74e0-7710-4faa-b5a2-60270a5da77c.png)

</details>

## Changelog
UI coordinate system and cursor position was changed from bottom left origin, y+ up to top left origin, y+ down.

## Migration Guide
All flex layout should be inverted (ColumnReverse => Column, FlexStart => FlexEnd, WrapReverse => Wrap)
System where dealing with cursor position should be changed to account for cursor position being based on the top left instead of bottom left
2022-10-11 12:51:44 +00:00
François
13dcdba05f use bevy default texture format if the surface is not yet available (#6233)
# Objective

- Fix #6231

## Solution

- In case no supported format is found, try to use Bevy default instead of panicking
2022-10-11 12:32:03 +00:00
Emerson MX
7673db731e Make TouchInput and ForceTouch serializable (#6191)
Closes #6021
2022-10-10 23:59:27 +00:00
ira
1ca1c8c39e Fix RemoveChildren command (#6192)
# Objective

`RemoveChildren` could remove the `Parent` component from children belonging to a different parent, which breaks the hierarchy.

This change looks a little funny because I'm reusing the events to avoid needing to clone the parent's `Children`.

Co-authored-by: devil-ira <justthecooldude@gmail.com>
2022-10-10 23:40:32 +00:00
François
b4accebe10 scenes: simplify return type of iter_instance_entities (#5994)
# Objective

- Taking the API improvement out of #5431 
- `iter_instance_entities` used to return an option of iterator, now it just returns an iterator

---

## Changelog

- If you use `SceneSpawner::iter_instance_entities`, it no longer returns an `Option`. The iterator will be empty if the return value used to be `None`
2022-10-10 23:09:08 +00:00
targrub
9a597b758e Adding Debug implementations for App, Stage, Schedule, Query, QueryState, etc. (#6214)
# Objective

- Adding Debug implementations for App, Stage, Schedule, Query, QueryState.
- Fixes #1130.

## Solution

- Implemented std::fmt::Debug for a number of structures.

---

## Changelog

Also added Debug implementations for ParallelSystemExecutor, SingleThreadedExecutor, various RunCriteria structures, SystemContainer, and SystemDescriptor.

Opinions are sure to differ as to what information to provide in a Debug implementation.  Best guess was taken for this initial version for these structures.


Co-authored-by: targrub <62773321+targrub@users.noreply.github.com>
2022-10-10 20:59:38 +00:00
Torstein Grindvik
55d126cab9 Add globals struct to mesh2d (#6222)
See commit message.
I noticed I couldn't use `globals.time` when using `Material2d`.

I copied the solution from 8073362039 , and now `Material2d` works for me.

Perhaps some of these struct definitions could be shared in the future, but for now I've just copy pasted it (it looked like the `View` struct was done that way).

Ping @IceSentry , I saw a comment on the linked commit that you intended to do this work at some point in the future.
2022-10-10 19:23:43 +00:00
Charles
740ae9a37f remove mandatory mesh attributes (#6127)
# Objective

- It's possible to create a mesh without positions or normals, but currently bevy forces these attributes to be present on any mesh.

## Solution

- Don't assume these attributes are present and add a shader defs for each attributes
- I updated 2d and 3d meshes to use the same logic.

---

## Changelog

- Meshes don't require any attributes

# Notes
I didn't update the pbr.wgsl shader because I'm not sure how to handle it. It doesn't really make sense to use it without positions or normals.
2022-10-10 17:58:15 +00:00
Light Ning
aebd760711 bevy_input: Fix process touch event (#4352)
# Objective

- `process_touch_event` in `bevy_input` don't update position info. `TouchPhase::Ended` and `TouchPhase::Cancelled` should use the position info from `pressed`. Otherwise, it'll not update. The position info is updated from `TouchPhase::Moved`.

## Solution

- Use updated touch info. 

---

## Changelog

> This section is optional. If this was a trivial fix, or has no externally-visible impact, feel free to skip this section.

- Fixed: bevy_input, fix process touch event, update touch info
2022-10-10 17:43:10 +00:00
JoJoJet
2cff2278ca Add a method for mapping Mut<T> -> Mut<U> (#6199)
# Objective

When designing an API, you may wish to provide access only to a specific field of a component or resource. The current options for doing this in safe code are

* `*Mut::into_inner`, which flags a change no matter what.
* `*Mut::bypass_change_detection`, which misses all changes.

## Solution

Add the method `map_unchanged`.

### Example

```rust
// When run, zeroes the translation of every entity.
fn reset_all(mut transforms: Query<&mut Transform>) {
    for transform in &mut transforms {
        // We pinky promise not to modify `t` within the closure.
        let translation = transform.map_unchanged(|t| &mut t.translation);
        // Only reset the translation if it isn't already zero.
        translation.set_if_not_equal(Vec2::ZERO);
    }
}
```

---

## Changelog

+ Added the method `map_unchanged` to types `Mut<T>`, `ResMut<T>`, and `NonSendMut<T>`.
2022-10-10 17:06:31 +00:00
ira
eb0a9e1586 Remove Transform::apply_non_uniform_scale (#6133)
This is a holdover from back when `Transform` was backed by a private `Mat4` two years ago.
Not particularly useful anymore :)

## Migration Guide
`Transform::apply_non_uniform_scale` has been removed.
It can be replaced with the following snippet:
```rust
transform.scale *= scale_factor;
```


Co-authored-by: devil-ira <justthecooldude@gmail.com>
2022-10-10 16:50:18 +00:00
ira
9423cb6a8d Rename Transform::mul_vec3 to transform_point and improve docs (#6132)
The docs ended up quite verbose :v

Also added a missing `#[inline]` to `GlobalTransform::mul_transform`.

I'd say this resolves #5500

# Migration Guide
`Transform::mul_vec3` has been renamed to `transform_point`.

Co-authored-by: devil-ira <justthecooldude@gmail.com>
2022-10-10 16:50:17 +00:00
cathalogue
2cde4c73ed Update linux_dependencies.md (#6205)
for nix build, pkgconfig has been renamed to pkg-config. Very small fix :>

# Objective

- Describe the objective or issue this PR addresses.
- If you're fixing a specific issue, say "Fixes #X".

## Solution

- Describe the solution used to achieve the objective above.

---

## Changelog

> This section is optional. If this was a trivial fix, or has no externally-visible impact, you can delete this section.

- What changed as a result of this PR?
- If applicable, organize changes under "Added", "Changed", or "Fixed" sub-headings
- Stick to one or two sentences. If more detail is needed for a particular change, consider adding it to the "Solution" section
  - If you can't summarize the work, your change may be unreasonably large / unrelated. Consider splitting your PR to make it easier to review and merge!

## Migration Guide

> This section is optional. If there are no breaking changes, you can delete this section.

- If this PR is a breaking change (relative to the last release of Bevy), describe how a user might need to migrate their code to support these changes
- Simply adding new functionality is not a breaking change.
- Fixing behavior that was definitely a bug, rather than a questionable design choice is not a breaking change.
2022-10-10 16:34:24 +00:00
Zicklag
a18e2b1a7f Reflect Default for GlobalTransform (#6200)
# Objective

Make `GlobalTransform` constructible from scripts, in the same vein as #6187.

## Solution

- Use the derive macro to reflect default

---

## Changelog

> This section is optional. If this was a trivial fix, or has no externally-visible impact, you can delete this section.

- `GlobalTransform` now reflects the `Default` trait.
2022-10-10 16:34:23 +00:00
Noah
6ae46f6403 Fixes Camera not being serializable due to missing registrations in core functionality. (#6170)
…

# Objective

- Fixes Camera not being serializable due to missing registrations in core functionality. 
- Fixes #6169

## Solution

- Updated Bevy_Render CameraPlugin with registrations for Option<Viewport> and then Bevy_Core CorePlugin with registrations for ReflectSerialize and ReflectDeserialize for type data Range<f32> respectively according to the solution in #6169



Co-authored-by: Noah <noahshomette@gmail.com>
2022-10-10 16:34:22 +00:00
JoJoJet
3321d68a75 Add methods for silencing system-order ambiguity warnings (#6158)
# Background

Incremental implementation of #4299. The code is heavily borrowed from that PR.

# Objective

The execution order ambiguity checker often emits false positives, since bevy is not aware of invariants upheld by the user.

## Solution

Title

---

## Changelog

+ Added methods `SystemDescriptor::ignore_all_ambiguities` and `::ambiguous_with`. These allow you to silence warnings for specific system-order ambiguities.

## Migration Guide

***Note for maintainers**: This should replace the migration guide for #5916*

Ambiguity sets have been replaced with a simpler API.

```rust
// These systems technically conflict, but we don't care which order they run in.
fn jump_on_click(mouse: Res<Input<MouseButton>>, mut transforms: Query<&mut Transform>) { ... }
fn jump_on_spacebar(keys: Res<Input<KeyCode>>, mut transforms: Query<&mut Transform>) { ... }

//
// Before

#[derive(AmbiguitySetLabel)]
struct JumpSystems;

app
  .add_system(jump_on_click.in_ambiguity_set(JumpSystems))
  .add_system(jump_on_spacebar.in_ambiguity_set(JumpSystems));

//
// After

app
  .add_system(jump_on_click.ambiguous_with(jump_on_spacebar))
  .add_system(jump_on_spacebar);

```
2022-10-10 16:34:21 +00:00
VitalyR
f5322cd757 get proper texture format after the renderer is initialized, fix #3897 (#5413)
# Objective
There is no Srgb support on some GPU and display protocols with `winit` (for example, Nvidia's GPUs with Wayland). Thus `TextureFormat::bevy_default()` which returns `Rgba8UnormSrgb` or `Bgra8UnormSrgb` will cause panics on such platforms. This patch will resolve this problem. Fix https://github.com/bevyengine/bevy/issues/3897.

## Solution

Make `initialize_renderer` expose `wgpu::Adapter` and `first_available_texture_format`, use the `first_available_texture_format` by default.

## Changelog

* Fixed https://github.com/bevyengine/bevy/issues/3897.
2022-10-10 16:10:05 +00:00
TimJentzsch
1738527902 Make the default background color of NodeBundle transparent (#6211)
# Objective

Closes #6202.

The default background color for `NodeBundle` is currently white.
However, it's very rare that you actually want a white background color.
Instead, you often want a background color specific to the style of your game or a transparent background (e.g. for UI layout nodes).

## Solution

`Default` is not derived for `NodeBundle` anymore, but explicitly specified.
The default background color is now transparent (`Color::NONE.into()`) as this is the most common use-case, is familiar from the web and makes specifying a layout for your UI less tedious.

---

## Changelog

- Changed the default `NodeBundle.background_color` to be transparent (`Color::NONE.into()`).

## Migration Guide

If you want a `NodeBundle` with a white background color, you must explicitly specify it:

Before:

```rust
let node = NodeBundle {
    ..default()
}
```

After:

```rust
let node = NodeBundle {
    background_color: Color::WHITE.into(),
    ..default()
}
```
2022-10-09 21:03:05 +00:00
CatThingy
5e71d7f833 Call mesh2d_tangent_local_to_world with the right arguments (#6209)
# Objective

Allow `Mesh2d` shaders to work with meshes that have vertex tangents
## Solution

Correctly pass `mesh.model` into `mesh2d_tangent_local_to_world`
2022-10-09 16:21:42 +00:00
Hennadii Chernyshchyk
ca3e6e6797 Impl Reflect for PathBuf and OsString (#6193)
# Objective

`Reflect` impl is missing for `PathBuf` and `OsString`. Closes #6166.

## Solution

Add implementations.

---

## Changelog

### Added

`Reflect` impls for `PathBuf` and `OsString`.
2022-10-08 17:02:21 +00:00
Dan Kov
cf86f275a9 Fix doc for Timer::percent_left (#6198)
# Objective

- Fix a mistake in documentation.
2022-10-08 14:51:21 +00:00
Gabriel Bourgeois
6b75589e2c Fix inconsistent children removal behavior (#6017)
# Objective

Fixes #6010

## Solution

As discussed in #6010, this makes it so the `Children` component is removed from the entity whenever all of its children are removed. The behavior is now consistent between all of the commands that may remove children from a parent, and this is tested via two new test functions (one for world functions and one for commands).

Documentation was also added to `insert_children`, `push_children`, `add_child` and `remove_children` commands to make this behavior clearer for users.

## Changelog

- Fixed `Children` component not getting removed from entity when all its children are moved to a new parent.

## Migration Guide

- Queries with `Changed<Children>` will no longer match entities that had all of their children removed using `remove_children`.
- `RemovedComponents<Children>` will now contain entities that had all of their children remove using `remove_children`.
2022-10-06 21:39:34 +00:00
Zicklag
cfba7312ef Reflect Default for ComputedVisibility and Handle<T> (#6187)
# Objective

- Reflecting `Default` is required for scripts to create `Reflect` types at runtime with no static type information.
- Reflecting `Default` on `Handle<T>` and `ComputedVisibility` should allow scripts from `bevy_mod_js_scripting` to actually spawn sprites from scratch, without needing any hand-holding from the host-game.

## Solution

- Derive `ReflectDefault` for `Handle<T>` and `ComputedVisiblity`.

---

## Changelog

> This section is optional. If this was a trivial fix, or has no externally-visible impact, you can delete this section.

- The `Default` trait is now reflected for `Handle<T>` and `ComputedVisibility`
2022-10-06 19:31:47 +00:00
François
f00212fd48 make Handle::<T> field id private, and replace with a getter (#6176)
# Objective

- Field `id` of `Handle<T>` is public: https://docs.rs/bevy/latest/bevy/asset/struct.Handle.html#structfield.id
- Changing the value of this field doesn't make sense as it could mean changing the previous handle without dropping it, breaking asset cleanup detection for the old handle and the new one

## Solution

- Make the field private, and add a public getter


Opened after discussion in #6171. Pinging @zicklag 

---

## Migration Guide

- If you were accessing the value `handle.id`, you can now do so with `handle.id()`
2022-10-06 13:33:30 +00:00
Alvin Philips
f2106bb3ce Reduced code duplication in gamepad_viewer example (#6175)
# Objective

- Reduce code duplication in the `gamepad_viewer` example.
- Fixes #6164 

## Solution

- Added a custom Bundle called `GamepadButtonBundle` to avoid repeating similar code throughout the example.
- Created a `new()` method on `GamepadButtonBundle`.



Co-authored-by: Alvin Philips <alvinphilips257@gmail.com>
2022-10-06 13:33:29 +00:00
Emerson MX
087f1c66aa Make bevy_window and bevy_input events serializable (#6180)
Closes #6021
2022-10-06 13:14:23 +00:00
ira
37860a09de Add Camera::viewport_to_world (#6126)
# Objective

Add a method for getting a world space ray from a viewport position.

Opted to add a `Ray` type to `bevy_math` instead of returning a tuple of `Vec3`'s as this is clearer and easier to document
The docs on `viewport_to_world` are okay, but I'm not super happy with them.

## Changelog
* Add `Camera::viewport_to_world`
* Add `Camera::ndc_to_world`
* Add `Ray` to `bevy_math`
* Some doc tweaks

Co-authored-by: devil-ira <justthecooldude@gmail.com>
2022-10-05 22:16:26 +00:00
Fanda Vacek
2362cebcae More explicit help how to cycle the cameras (#6162)
# Objective

Scene viewer example has switch camera keys defined, but only one camera was instantiated on the scene.

## Solution

More explicit help how to cycle the cameras, explaining that more cameras must be present in loaded scene.





Co-authored-by: Fanda Vacek <fvacek@elektroline.cz>
2022-10-05 21:22:11 +00:00
François
720b67396f flaky test: put panicking system in a single threaded stage (#6172)
# Objective

- Fix #5285 

## Solution

- Put the panicking system in a single threaded stage during the test
- This way only the main thread will panic, which is handled by `cargo test`
2022-10-05 16:34:55 +00:00
Noah
26c299bd2a Update window.rs PresentMode docs to clarify which PresentMode will panic and which will fallback (#6160)
# Objective

- Fixes contradictory docs in Window::PresentMode partaining to PresentMode fallback behavior. Fix based on commit history showing the most recent update didn't remove old references to the gracefal fallback for Immediate and Mailbox.
- Fixes #5831 

## Solution

- Updated the docs for Window::PresentMode itself and for each individual enum variant to clarify which will fallback and which will panic.


Co-authored-by: Noah <noahshomette@gmail.com>
2022-10-05 13:51:32 +00:00
Kewei Huang
2a6b544a0a Document EntityMut::remove() (#6168)
# Objective

- Fixes #5990

## Solution

- Add docs for `EntityMut::remove()` explaining its return value.
2022-10-05 12:21:09 +00:00
robtfm
29098b7a11 fix spot dir nan bug (#6167)
# Objective

fix error with pbr shader's spotlight direction calculation when direction.y ~= 0

## Solution

in pbr_lighting.wgsl, clamp `1-x^2-z^2` to `>= 0` so that we can safely `sqrt` it
2022-10-05 12:00:07 +00:00
JoJoJet
8a268129f9 Deduplicate ambiguity reporting code (#6149)
# Objective

Now that #6083 has been merged, we can clean up some ugly ambiguity detection code.

# Solution

Deduplicate code.
2022-10-03 16:57:31 +00:00
Sludge
ac364e9e28 Register Wireframe type (#6152)
# Objective

The `Wireframe` type implements `Reflect`, but is never registered, making its reflection inaccessible.

## Solution

Call `App::register_type::<Wireframe>()` in the `Plugin::build` implementation of `WireframePlugin`.

---

## Changelog

Fixed `Wireframe` type reflection not getting registered.
2022-10-03 16:37:03 +00:00
ira
3aaf746675 Example cleanup (#6131)
Co-authored-by: devil-ira <justthecooldude@gmail.com>
2022-09-30 13:25:27 +00:00
Cameron
6929d95f7f Replace fixed timestep in alien_cake_addict example with timer (#5760)
# Objective

Examples should use the correct tools for the job.

## Solution

A fixed timestep, by design, can step multiple times consecutively in a single update.

That property used to crash the `alien_cake_addict` example (#2525), which was "fixed" in #3411 (by just not panicking). The proper fix is to use a timer instead, since the system is supposed to spawn a cake every 5 seconds. 

---

A timer guarantees a minimum duration. A fixed timestep guarantees a fixed number of steps per second.
Each one works by essentially sacrificing the other's guarantee.

You can use them together, but no other systems are timestep-based in this example, so the timer is enough.
2022-09-30 11:37:33 +00:00
Nicola Papale
6b8cc2652a Document all StandardMaterial fields (#5921)
# Objective

Add more documentation on `StandardMaterial` and improve
consistency on existing doc.

Co-authored-by: Nicola Papale <nicopap@users.noreply.github.com>
2022-09-28 21:20:29 +00:00
Lucidus115
e8e541e4b7 fix #6062 incorrect links for render module docs (#6099)
# Objective

- Fixes #6062

## Solution

- Change path to `(crate::render::renderer)` from `(bevy_render::renderer)` in `crates/bevy_internal/src/lib.rs`

---
2022-09-28 21:02:26 +00:00
Nicola Papale
4e5b165fa0 Add details about intel linux vulkan driver (#6103)
# Objective

Fixes #6073

# Solution

Add a paragraph about `vulkan-intel` to the archlinux section.
2022-09-28 20:38:43 +00:00
Dawid Piotrowski
0bf7f3153d Allow access to non-send resource through World::resource_scope (#6113)
# Objective

Relaxes the trait bound for `World::resource_scope` to allow non-send resources. Fixes #6037.

## Solution

No big changes in code had to be made. Added a check so that the non-send resources won't be accessed from a different thread.

---

## Changelog
 - `World::resource_scope` accepts non-send resources now
 - `World::resource_scope` verifies non-send access if the resource is non-send
 - Two new tests are added, one for valid use of `World::resource_scope` with a non-send resource, and one for invalid use (calling it from a different thread, resulting in panic)

Co-authored-by: Dawid Piotrowski <41804418+Pietrek14@users.noreply.github.com>
2022-09-28 17:53:58 +00:00
rustui
59f872647d StreamReceiver does not need to be mutable (#6119)
ResMut -> Res
2022-09-28 17:39:36 +00:00