Commit graph

507 commits

Author SHA1 Message Date
Carter Anderson
01aedc8431 Spawn now takes a Bundle (#6054)
# Objective

Now that we can consolidate Bundles and Components under a single insert (thanks to #2975 and #6039), almost 100% of world spawns now look like `world.spawn().insert((Some, Tuple, Here))`. Spawning an entity without any components is an extremely uncommon pattern, so it makes sense to give spawn the "first class" ergonomic api. This consolidated api should be made consistent across all spawn apis (such as World and Commands).

## Solution

All `spawn` apis (`World::spawn`, `Commands:;spawn`, `ChildBuilder::spawn`, and `WorldChildBuilder::spawn`) now accept a bundle as input:

```rust
// before:
commands
  .spawn()
  .insert((A, B, C));
world
  .spawn()
  .insert((A, B, C);

// after
commands.spawn((A, B, C));
world.spawn((A, B, C));
```

All existing instances of `spawn_bundle` have been deprecated in favor of the new `spawn` api. A new `spawn_empty` has been added, replacing the old `spawn` api.  

By allowing `world.spawn(some_bundle)` to replace `world.spawn().insert(some_bundle)`, this opened the door to removing the initial entity allocation in the "empty" archetype / table done in `spawn()` (and subsequent move to the actual archetype in `.insert(some_bundle)`).

This improves spawn performance by over 10%:
![image](https://user-images.githubusercontent.com/2694663/191627587-4ab2f949-4ccd-4231-80eb-80dd4d9ad6b9.png)

To take this measurement, I added a new `world_spawn` benchmark.

Unfortunately, optimizing `Commands::spawn` is slightly less trivial, as Commands expose the Entity id of spawned entities prior to actually spawning. Doing the optimization would (naively) require assurances that the `spawn(some_bundle)` command is applied before all other commands involving the entity (which would not necessarily be true, if memory serves). Optimizing `Commands::spawn` this way does feel possible, but it will require careful thought (and maybe some additional checks), which deserves its own PR. For now, it has the same performance characteristics of the current `Commands::spawn_bundle` on main.

**Note that 99% of this PR is simple renames and refactors. The only code that needs careful scrutiny is the new `World::spawn()` impl, which is relatively straightforward, but it has some new unsafe code (which re-uses battle tested BundlerSpawner code path).** 

---

## Changelog

- All `spawn` apis (`World::spawn`, `Commands:;spawn`, `ChildBuilder::spawn`, and `WorldChildBuilder::spawn`) now accept a bundle as input
- All instances of `spawn_bundle` have been deprecated in favor of the new `spawn` api
- World and Commands now have `spawn_empty()`, which is equivalent to the old `spawn()` behavior.  

## Migration Guide

```rust
// Old (0.8):
commands
  .spawn()
  .insert_bundle((A, B, C));
// New (0.9)
commands.spawn((A, B, C));

// Old (0.8):
commands.spawn_bundle((A, B, C));
// New (0.9)
commands.spawn((A, B, C));

// Old (0.8):
let entity = commands.spawn().id();
// New (0.9)
let entity = commands.spawn_empty().id();

// Old (0.8)
let entity = world.spawn().id();
// New (0.9)
let entity = world.spawn_empty();
```
2022-09-23 19:55:54 +00:00
JoJoJet
fb74ca3d46 Add ambiguity detection tests (#6053)
# Objective

- Add unit tests for ambiguity detection reporting.
- Incremental implementation of #4299.

## Solution

- Refactor ambiguity detection internals to make it testable. As a bonus, this should make it easier to extend in the future.

## Notes

* This code was copy-pasted from #4299 and modified. Credit goes to @alice-i-cecile and @afonsolage, though I'm not sure who wrote what at this point.
2022-09-22 20:01:54 +00:00
targrub
a09dd034a2 Fix CI issues arising from use of Rust 1.64 (#6067)
## Objective

Fixes https://github.com/bevyengine/bevy/issues/6063

## Solution

- Use `then_some(x)` instead of `then( || x)`.
- Updated error logs from `bevy_ecs_compile_fail_tests`.

## Migration Guide

From Rust 1.63 to 1.64, a new Clippy error was added; now one should use `then_some(x)` instead of `then( || x)`.
2022-09-22 16:56:43 +00:00
Carter Anderson
cd15f0f5be Accept Bundles for insert and remove. Deprecate insert/remove_bundle (#6039)
# Objective

Take advantage of the "impl Bundle for Component" changes in #2975 / add the follow up changes discussed there.

## Solution

- Change `insert` and `remove` to accept a Bundle instead of a Component (for both Commands and World)
- Deprecate `insert_bundle`, `remove_bundle`, and `remove_bundle_intersection`
- Add `remove_intersection`

---

## Changelog

- Change `insert` and `remove` now accept a Bundle instead of a Component (for both Commands and World)
- `insert_bundle` and `remove_bundle` are deprecated
 

## Migration Guide

Replace `insert_bundle` with `insert`:
```rust
// Old (0.8)
commands.spawn().insert_bundle(SomeBundle::default());
// New (0.9)
commands.spawn().insert(SomeBundle::default());
```

Replace `remove_bundle` with `remove`:
```rust
// Old (0.8)
commands.entity(some_entity).remove_bundle::<SomeBundle>();
// New (0.9)
commands.entity(some_entity).remove::<SomeBundle>();
```

Replace `remove_bundle_intersection` with `remove_intersection`:
```rust
// Old (0.8)
world.entity_mut(some_entity).remove_bundle_intersection::<SomeBundle>();
// New (0.9)
world.entity_mut(some_entity).remove_intersection::<SomeBundle>();
```

Consider consolidating as many operations as possible to improve ergonomics and cut down on archetype moves:
```rust
// Old (0.8)
commands.spawn()
  .insert_bundle(SomeBundle::default())
  .insert(SomeComponent);

// New (0.9) - Option 1
commands.spawn().insert((
  SomeBundle::default(),
  SomeComponent,
))

// New (0.9) - Option 2
commands.spawn_bundle((
  SomeBundle::default(),
  SomeComponent,
))
```

## Next Steps

Consider changing `spawn` to accept a bundle and deprecate `spawn_bundle`.
2022-09-21 21:47:53 +00:00
Federico Rinaldi
d9e99cd80c Fix API docs for Commands methods (#5955)
# Objective

The doc comments for `Command` methods are a bit inconsistent on the format, they sometimes go out of scope, and most importantly they are wrong, in the sense that they claim to perform the action described by the command, while in reality, they just push a command to perform the action.

- Follow-up of #5938.
- Related to #5913.

## Solution

- Where applicable, only stated that a `Command` is pushed.
- Added a “See also” section for similar methods.
- Added a missing “Panics” section for `Commands::entity`.
- Removed a wrong comment about `Commands::get_or_spawn` returning `None` (It does not return an option).
- Removed polluting descriptions of other items.
- Misc formatting changes.

## Future possibilities

Since the `Command` implementors (`Spawn`, `InsertBundle`, `InitResource`, ...) are public, I thought that it might be appropriate to describe the action of the command there instead of the method, and to add a `method → command struct` link to fill the gap.

If that seems too far-fetched, we may opt to make them private, if possible, or `#[doc(hidden)]`.
2022-09-21 17:37:57 +00:00
Daniel McNab
1a2aedd165 Implement Bundle for Component. Use Bundle tuples for insertion (#2975)
@BoxyUwU this is your fault. 

Also cart didn't arrive in time to tell us not to do this.

# Objective

- Fix #2974

## Solution

- The first commit just does the actual change
- Follow up commits do steps to prove that this method works to unify as required, but this does not remove `insert_bundle`.

## Changelog

### Changed
Nested bundles now collapse automatically, and every `Component` now implements `Bundle`.
This means that you can combine bundles and components arbitrarily, for example:
```rust
// before:
.insert(A).insert_bundle(MyBBundle{..})
// after:
.insert_bundle((A, MyBBundle {..}))
```

Note that there will be a follow up PR that removes the current `insert` impl and renames `insert_bundle` to `insert`.

### Removed
The `bundle` attribute in `derive(Bundle)`.

## Migration guide

In `derive(Bundle)`, the `bundle` attribute has been removed. Nested bundles are not collapsed automatically. You should remove `#[bundle]` attributes.

Co-authored-by: Carter Anderson <mcanders1@gmail.com>
2022-09-20 20:17:08 +00:00
targrub
d0e294c86b Query filter types must be ReadOnlyWorldQuery (#6008)
# Objective

Fixes Issue #6005.

## Solution

Replaced WorldQuery with ReadOnlyWorldQuery on F generic in Query filters and QueryState to restrict its trait bound.

## Migration Guide

Query filter (`F`) generics are now bound by `ReadOnlyWorldQuery`, rather than `WorldQuery`. If for some reason you were requesting `Query<&A, &mut B>`, please use `Query<&A, With<B>>` instead.
2022-09-18 23:52:01 +00:00
Federico Rinaldi
deeab3fc90 Optimize use statement (#5992)
Just a very small `use` statement thing. Check the changed file.
2022-09-15 17:05:09 +00:00
James Liu
5d821fe1a7 Start running systems while prepare_systems is running (#4919)
# Objective
While using the ParallelExecutor, systems do not actually start until `prepare_systems` completes. In stages where there are large numbers of "empty" systems with very little work to do, this delay adds significant overhead, which can add up over many stages.

## Solution
Immediately and synchronously signal the start of systems that can run without dependencies inside `prepare_systems` instead of waiting for the first executor iteration after `prepare_systems` completes. Any system that is dependent on them still cannot run until after `prepare_systems` completes, but there are a large number of unconstrained systems in the base engine where this is a general benefit in almost every case.

## Performance

This change was tested against `many_foxes` in the default configuration. As this change is sensitive to the overhead around scheduling systems, the spans for measuring system timing, system overhead, and system commands were all commented out for these measurements.

The median stage timings between `main` and this PR are as follows:

|stage|main|this PR|
|:--|:--|:--|
|First|75.54 us|61.61 us|
|LoadAssets|51.05 us|42.32 us|
|PreUpdate|54.6 us|55.56 us|
|Update|61.89 us|51.5 us|
|PostUpdate|7.27 ms|6.71 ms|
|AssetEvents|47.82 us|35.95 us|
|Last|39.19 us|37.71 us|
|reserve_and_flush|57.83 us|48.2 us|
|Extract|1.41 ms|1.28 ms|
|Prepare|554.49 us|502.53 us|
|Queue|216.29 us|207.51 us|
|Sort|67.03 us|60.99 us|
|Render|1.73 ms|1.58 ms|
|Cleanup|33.55 us|30.76 us|
|Clear Entities|18.56 us|17.05 us|
|**full frame**|**11.9 ms**|**10.91 ms**|

For the first few stages, the benefit is small but cumulative over each. For PostUpdate in particular, this allows `parent_update` to run while prepare_systems is running, which is required for the animation and transform propagation systems, which dominate the time spent in the stage, but also frontloads the contention as the other "empty" systems are also running while `parent_update` is running. For Render, where there is just a single large exclusive system, the benefit comes from not waiting on a spuriously scheduled task on the task pool to kick off the system: it's immediately scheduled to run.
2022-09-13 19:28:13 +00:00
Boxy
404b4fc0eb lifetime related cleanup in entity_ref.rs (#5611)
# Objective

EntityMut::world takes &mut self instead of &self I don't see any reason for this.
EntityRef is overly restrictive with fn world and could return &'w World

---

## Changelog

- EntityRef now implements Copy and Clone
- EntityRef::world is now fn(&self) -> &'w World instead of fn(&mut self) -> &World
- EntityMut::world is now fn(&self) -> &World instead of fn(&mut self) -> &World
2022-09-12 04:34:52 +00:00
PROMETHIA-27
05afbc6815 Remove Sync bound from Local (#5483)
# Objective

Currently, `Local` has a `Sync` bound. Theoretically this is unnecessary as a local can only ever be accessed from its own system, ensuring exclusive access on one thread. This PR removes this restriction.

## Solution

- By removing the `Resource` bound from `Local` and adding the new `SyncCell` threading primative, `Local` can have the `Sync` bound removed.

## Changelog

### Added

- Added `SyncCell` to `bevy_utils`

### Changed

- Removed `Resource` bound from `Local`
- `Local` is now wrapped in a `SyncCell`

## Migration Guide

- Any code relying on `Local<T>` having `T: Resource` may have to be changed, but this is unlikely.

Co-authored-by: PROMETHIA-27 <42193387+PROMETHIA-27@users.noreply.github.com>
2022-09-12 04:15:55 +00:00
Federico Rinaldi
fc07557913 Clarify Commands API docs (#5938)
# Objective

- Make people stop believing that commands are applied immediately (hopefully).
- Close #5913.
- Alternative to #5930.

## Solution

I added the clause “to perform impactful changes to the `World`” to the first line to subliminally help the reader accept the fact that some operations cannot be performed immediately without messing up everything.

Then I explicitely said that applying a command requires exclusive `World` access, and finally I proceeded to show when these commands are automatically applied.

I also added a brief paragraph about how commands can be applied manually, if they want.

---

### Further possibilities

If you agree, we can also change the text of the method documentation (in a separate PR) to stress about enqueueing an action instead of just performing it. For example, in `Commands::spawn`:

> Creates a new `Entity`

would be changed to something like:

> Issues a `Command` to spawn a new `Entity`

This may even have a greater effect, since when typing in an IDE, the docs of the method pop up and the programmer can read them on the fly.
2022-09-12 01:06:09 +00:00
Jakob Hellermann
6f2cc0b30e relax Sized bounds around change detection types (#5917)
# Objective

I wanted to run the code
```rust
let reflect_resource: ReflectResource = ...;
let value: Mut<dyn Reflect> = reflect_resource.reflect(world);
value.deref();
// ^ ERROR: deref method doesn't exist because `dyn Reflect` doesnt satisfy `: Sized`.
```

## Solution

Relax `Sized` bounds in all the methods and trait implementations for `Mut` and friends.
2022-09-09 21:26:36 +00:00
Alice Cecile
ca3fa9dd6f Move ambiguity detection into its own file (#5918)
# Objective

This code is very disjoint, and the `stage.rs` file that it's in is already very long.

All I've done is move the code and clean up the compiler errors that result.

Followup to #5916, split out from #4299.
2022-09-09 18:44:47 +00:00
Alice Cecile
c96b7ffb50 Remove ambiguity sets (#5916)
# Objective

Ambiguity sets are used to ignore system order ambiguities between groups of systems. However, they are not very useful: they are clunky, poorly integrated, and generally hampered by the difficulty using (or discovering) the ambiguity detector.

As a first step to the work in #4299, we're removing them.

## Migration Guide

Ambiguity sets have been removed.
2022-09-09 17:21:50 +00:00
Alice Cecile
54e32ee681 Add a change detection bypass and manual control over change ticks (#5635)
# Objective

- Our existing change detection API is not flexible enough for advanced users: particularly those attempting to do rollback networking.
- This is an important use case, and with adequate warnings we can make mucking about with change ticks scary enough that users generally won't do it.
- Fixes #5633.
- Closes #2363.

## Changelog

- added `ChangeDetection::set_last_changed` to manually mutate the `last_change_ticks` field"
- the `ChangeDetection` trait now requires an `Inner` associated type, which contains the value being wrapped.
- added `ChangeDetection::bypass_change_detection`, which hands out a raw `&mut Inner`

## Migration Guide

Add the `Inner` associated type and new methods to any type that you've implemented `DetectChanges` for.
2022-09-09 16:26:52 +00:00
robem
7a92555233 Update WorldQueryGats doc with type aliases (#5898)
Make API users aware that the type aliases `QueryItem` and `QueryFetch` can be used instead of the more bloated alternative with `WorldQueryGats`.

Fixes #5842
2022-09-06 21:24:40 +00:00
Gabriel Bourgeois
092bb71bcf Clean up taffy nodes when UI node entities are removed (#5886)
# Objective

Clean up taffy nodes when the associated UI node gets removed. The current UI code will keep the taffy nodes around forever.

## Solution

Use `RemovedComponents<Node>` to iterate over nodes that are no longer valid UI nodes or that have been despawned, and remove them from taffy and the internal hash map.

## Implementation Notes

Do note that using `despawn()` instead of `despawn_recursive()` on a UI node that has children will result in a [warnings spam](https://github.com/bevyengine/bevy/blob/main/crates/bevy_ui/src/flex/mod.rs#L120) since the children will not be part of a proper UI hierarchy anymore.

---

## Changelog

- Fixed memory leak when nodes are removed in bevy_ui
2022-09-05 21:50:31 +00:00
ira
76ae6f4c6e Miscellaneous code-quality improvements. (#5860)
Does what it do.

Co-authored-by: devil-ira <justthecooldude@gmail.com>
2022-09-05 00:30:21 +00:00
Moulberry
927441d048 Add From<EntityMut> for EntityRef (fixes #5459) (#5461)
Add From<EntityMut> for EntityRef (fixes #5459)
2022-09-03 18:06:42 +00:00
JoJoJet
697d297b55 Remove last uses of string-labels (#5420)
# Objective

* Related: #4341
* Remove all remaining uses of stringly-typed labels in the repo. Right now, it's just a bunch of tests and examples.
2022-09-03 18:06:41 +00:00
Federico Rinaldi
dfeb63e7b8 Update Query methods documentation (#5742)
# Objective

- Increase consistency across documentation of `Query` methods.
- Fixes #5506

## Solution

- See #4989. This PR is derived from it. It just includes changes to the `Query` methods' docs.
2022-09-02 16:33:19 +00:00
Federico Rinaldi
7511c9bfaa Update Query struct docs (#5741)
# Objective

- Update `Query` docs with better terminology
- add some performance remarks (Fixes #4742)

## Solution

- See #4989. This PR is derived from it. It just includes changes to the `Query` struct docs.
2022-09-02 12:57:39 +00:00
Federico Rinaldi
59bf3c4cc9 Improve WorldQuery docs (#5740)
# Objective

- Update docs to `WorldQuery`

## Solution

- See #4989. This PR is derived from it, and limited to the `WorldQuery` item docs.
2022-09-02 12:35:24 +00:00
James O'Brien
f9853cbbc2 Add get_entity to Commands (#5854)
# Objective

- Fixes #5850 

## Solution

- As described in the issue, added a `get_entity` method on `Commands` that returns an `Option<EntityCommands>`

## Changelog
- Added the new method with a simple doc test
- I have re-used `get_entity` in `entity`, similarly to how `get_single` is used in `single` while additionally preserving the error message
- Add `#[inline]` to both functions

Entities that have commands queued to despawn system will still return commands when `get_entity` is called but that is representative of the fact that the entity is still around until those commands are flushed.

A potential `contains_entity` could also be added in this PR if desired, that would effectively be replacing Entities.contains but may be more discoverable if this is a common use case.


Co-authored-by: Carter Anderson <mcanders1@gmail.com>
2022-09-01 22:06:46 +00:00
John
9e34c748c6 Added the ability to get or set the last change tick of a system. (#5838)
# Objective
I'm build a UI system for bevy. In this UI system there is a concept of a system per UI entity. I had an issue where change detection wasn't working how I would expect and it's because when a function system is ran the `last_change_tick` is updated with the latest tick(from world). In my particular case I want to "wait" to update the `last_change_tick` until after my system runs for each entity.

## Solution
Initially I thought bypassing the change detection all together would be a good fix, but on talking to some users in discord a simpler fix is to just expose `last_change_tick` to the end users. This is achieved by adding the following to the `System` trait:
```rust
    /// Allows users to get the system's last change tick.
    fn get_last_change_tick(&self) -> u32;
    /// Allows users to set the system's last change tick.
    fn set_last_change_tick(&mut self, last_change_tick: u32);
```

This causes a bit of weirdness with two implementors of `System`. `FixedTimestep` and `ChainSystem` both implement system and thus it's required that some sort of implementation be given for the new functions. I solved this by outputting a warning and not doing anything for these systems. 

I think it's important to understand why I can't add the new functions only to the function system and not to the `System` trait. In my code I store the systems generically as `Box<dyn System<...>>`. I do this because I have differing parameters that are being passed in depending on the UI widget's system.  As far as I can tell there isn't a way to take a system trait and cast it into a specific type without knowing what those parameters are.

In my own code this ends up looking something like:
```rust
// Runs per entity.
let old_tick = widget_system.get_last_change_tick();
should_update_children = widget_system.run((widget_tree.clone(), entity.0), world);
widget_system.set_last_change_tick(old_tick);


// later on after all the entities have been processed:
for system in context.systems.values_mut() {
    system.set_last_change_tick(world.read_change_tick());
}
```

## Changelog

- Added `get_last_change_tick` and `set_last_change_tick` to `System`'s.
2022-08-31 01:53:15 +00:00
Boxy
ed773dbe30 Misc query.rs cleanup (#5591)
# Objective
- `for_each` methods inconsistently used an actual generic param or `impl Trait` change it to use `impl Trait` always, change them to be consistent
- some methods returned `'w 's` or `'_ '_`, change them to return `'_ 's`

## Solution

- Do what i just said

---

## Changelog

- `iter_unsafe` and `get_unchecked` no longer return borrows tied to `'w`

## Migration Guide

transmute the returned borrow from `iter_unsafe` and `get_unchecked` if this broke you (although preferably find a way to write your code that doesnt need to do this...)
2022-08-30 21:56:00 +00:00
Boxy
df31b7d762 Remove insert_resource_with_id (#5608)
# Objective

remove `insert_resource_with_id` because `insert_resource_by_id` exists and does almost exactly the same thing

blocked on #5587 because otherwise we will leak a resource when it's inserted 

## Solution

remove the function and also add a safety invariant of to `insert_resource_by_id` that the id be valid for the world.

I didn't see any discussion in #4447 about this safety invariant being left off in favor of a panic so I'm curious if there was one or if it just seemed nicer to have less safety invariants for callers to uphold 😅 

---

## Changelog

- safety invariant added to `insert_resource_by_id` requiring the id to be valid for world

## Migration Guide

- audit any calls to `insert_resource_by_id` making sure that the id is valid for the world

Co-authored-by: Carter Anderson <mcanders1@gmail.com>
2022-08-30 20:32:15 +00:00
Jerome Humbert
fe6246dac6 Make most Entity methods const (#5688)
# Objective

Fixes #5687

## Solution

Update the methods on the `Entity` struct to be `const`, so we can
define compile-time constants and more generally use them in a const
context.

---

## Changelog

### Added

- Most `Entity` methods are now `const fn`.
2022-08-30 03:16:22 +00:00
Aceeri
d346274e32 Warning message for missing events (#5730)
# Objective
- Reduce debugging burden when using events by telling user when they missed an event.

## Solution

Co-authored-by: Carter Anderson <mcanders1@gmail.com>
2022-08-30 02:23:04 +00:00
Federico Rinaldi
5597fc54d2 Add documentation to QueryCombinationIter (#5739)
# Objective

- Document `QueryCombinationIter`

## Solution

- Describe the item, add usage and examples
- Copy notes about the number of query items generated from the corresponding query methods (they will be removed in #5742 ([motivation]))

## Additional notes

- Derived from #4989 

[motivation]: https://github.com/bevyengine/bevy/pull/4989#issuecomment-1208421496
2022-08-30 00:39:17 +00:00
JoJoJet
584d855fd1 Add a module for common system chain/pipe adapters (#5776)
# Objective

Right now, users have to implement basic system adapters such as `Option` <-> `Result` conversions by themselves. This is slightly annoying and discourages the use of system chaining.

## Solution

Add the module `system_adapter` to the prelude, which contains a collection of common adapters. This is very ergonomic in practice.

## Examples

Convenient early returning.

```rust
use bevy::prelude::*;

App::new()
    // If the system fails, just try again next frame.
    .add_system(pet_dog.chain(system_adapter::ignore))
    .run();

#[derive(Component)]
struct Dog;

fn pet_dog(dogs: Query<(&Name, Option<&Parent>), With<Dog>>) -> Option<()> {
    let (dog, dad) = dogs.iter().next()?;
    println!("You pet {dog}. He/she/they are a good boy/girl/pupper.");
    let (dad, _) = dogs.get(dad?.get()).ok()?;
    println!("Their dad's name is {dad}");
    Some(())
}
```

Converting the output of a system

```rust
use bevy::prelude::*;

App::new()
    .add_system(
        find_name
            .chain(system_adapter::new(String::from))
            .chain(spawn_with_name),
    )
    .run();

fn find_name() -> &'static str { /* ... */ }
fn spawn_with_name(In(name): In<String>, mut commands: Commands) {
    commands.spawn().insert(Name::new(name));
}
```
---

## Changelog

* Added the module `bevy_ecs::prelude::system_adapter`, which contains a collection of common system chaining adapters.
  * `new` - Converts a regular fn to a system adapter.
  * `unwrap` - Similar to `Result::unwrap`
  * `ignore` - Discards the output of the previous system.
2022-08-30 00:17:20 +00:00
Ben Frankel
70106773f2 Fix example in AnyOf docs (#5798) 2022-08-25 20:31:51 +00:00
Ian Chamberlain
cde5ae8104 bevy_ecs: Use 32-bit entity ID cursor on platforms without AtomicI64 (#4452)
# Objective
- Fixes #4451

## Solution
- Conditionally compile entity ID cursor as `AtomicI32` when compiling on a platform that does not support 64-bit atomics.

- This effectively raises the MSRV to 1.60 as it uses a `#[cfg]` that was only just stabilized there. (should this be noted in changelog?)

---

## Changelog
- Added `bevy_ecs` support for platforms without 64-bit atomic ints


## Migration Guide
N/A
2022-08-21 00:45:49 +00:00
Nathan Ward
00323b3048 Better error message for World::resource_scope (#5727)
# Objective

- Fixes #5365 
- The `assert!()` when the resource from `World::resource_scope` is inserted into the world is not descriptive.

## Solution

- Add more context to the assert inside of `World::resource_scope` when the `FnOnce` param inserts the resource.
2022-08-18 18:53:08 +00:00
Aceeri
f0c512731b SystemParam for the name of the system you are currently in (#5731)
# Objective
- Similar to `SystemChangeTick`, probably somewhat useful for debugging messages.

---

## Changelog

- Added `SystemName` which copies the `SystemMeta::name` field so it can be accessed within a system.
2022-08-18 18:31:12 +00:00
JoJoJet
3221e569e0 Remove an outdated workaround for impl Trait (#5659)
# Objective

Rust 1.63 resolved [an issue](https://github.com/rust-lang/rust/issues/83701) that prevents you from combining explicit generic arguments with `impl Trait` arguments.

Now, we no longer need to use dynamic dispatch to work around this.

## Migration Guide

The methods `Schedule::get_stage` and `get_stage_mut` now accept `impl StageLabel` instead of `&dyn StageLabel`.

### Before
```rust
let stage = schedule.get_stage_mut::<SystemStage>(&MyLabel)?;
```

### After
```rust
let stage = schedule.get_stage_mut::<SystemStage>(MyLabel)?;
```
2022-08-16 23:40:24 +00:00
Boutillier
de6bef72a1 Fix for bevy CI on main - clippy safety comments on trait. (#5665)
# Objective

Make CI pass on bevy main.

Update to rust-1.63, updated clippy to 1.63 which introduced the following enhancements:
- [undocumented_unsafe_blocks](https://rust-lang.github.io/rust-clippy/master/index.html#undocumented_unsafe_blocks): Now also lints on unsafe trait implementations

This caught two incorrectly written ( but existing) safety comments for unsafe  traits.

## Solution

Fix the comment to use `SAFETY:`
2022-08-13 10:51:19 +00:00
Jakob Hellermann
5595733035 drop old value in insert_resource_by_id if exists (#5587)
# Objective

While trying out the lint `unsafe_op_in_unsafe_fn` I noticed that `insert_resource_by_id` didn't drop the old value if it already existed, and reimplemented `Column::replace` manually for no apparent reason.

## Solution

- use `Column::replace` and add a test expecting the correct drop count

---

## Changelog

- `World::insert_resource_by_id` will now correctly drop the old resource value, if one already existed
2022-08-09 18:05:43 +00:00
Alex
fe97b384a5 fix: typo in system params docs (#5624)
# Objective

- Fix a typo on `SystemParam` docs

## Solution
- added 'be'.
- Hurray my first OSS PR! 

---
2022-08-09 16:53:27 +00:00
Jakob Hellermann
fcb77d6988 remove ReflectMut in favor of Mut<dyn Reflect> (#5630)
# Objective

- `ReflectMut` served no purpose that wasn't met by `Mut<dyn Reflect>` which is easier to understand since you have to deal with fewer types
- there is another `ReflectMut` type that could be confused with this one

## Solution/Changelog

- relax `T: ?Sized` bound in `Mut<T>`
- replace all instances of `ReflectMut` with `Mut<dyn Reflect>`
2022-08-09 16:19:34 +00:00
Nicola Papale
397b6df023 Add into_world_mut to EntityMut (#5586)
# Objective

Provide a safe API to access an `EntityMut`'s `World`.

## Solution

* Add `EntityMut::into_world_mut` for safe access to the entity's world.

---

## Changelog

* Add `EntityMut::into_world_mut` for safe access to the entity's world.
2022-08-08 22:59:18 +00:00
ira
992681b59b Make Resource trait opt-in, requiring #[derive(Resource)] V2 (#5577)
*This PR description is an edited copy of #5007, written by @alice-i-cecile.*
# Objective
Follow-up to https://github.com/bevyengine/bevy/pull/2254. The `Resource` trait currently has a blanket implementation for all types that meet its bounds.

While ergonomic, this results in several drawbacks:

* it is possible to make confusing, silent mistakes such as inserting a function pointer (Foo) rather than a value (Foo::Bar) as a resource
* it is challenging to discover if a type is intended to be used as a resource
* we cannot later add customization options (see the [RFC](https://github.com/bevyengine/rfcs/blob/main/rfcs/27-derive-component.md) for the equivalent choice for Component).
* dependencies can use the same Rust type as a resource in invisibly conflicting ways
* raw Rust types used as resources cannot preserve privacy appropriately, as anyone able to access that type can read and write to internal values
* we cannot capture a definitive list of possible resources to display to users in an editor
## Notes to reviewers
 * Review this commit-by-commit; there's effectively no back-tracking and there's a lot of churn in some of these commits.
   *ira: My commits are not as well organized :')*
 * I've relaxed the bound on Local to Send + Sync + 'static: I don't think these concerns apply there, so this can keep things simple. Storing e.g. a u32 in a Local is fine, because there's a variable name attached explaining what it does.
 * I think this is a bad place for the Resource trait to live, but I've left it in place to make reviewing easier. IMO that's best tackled with https://github.com/bevyengine/bevy/issues/4981.

## Changelog
`Resource` is no longer automatically implemented for all matching types. Instead, use the new `#[derive(Resource)]` macro.

## Migration Guide
Add `#[derive(Resource)]` to all types you are using as a resource.

If you are using a third party type as a resource, wrap it in a tuple struct to bypass orphan rules. Consider deriving `Deref` and `DerefMut` to improve ergonomics.

`ClearColor` no longer implements `Component`. Using `ClearColor` as a component in 0.8 did nothing.
Use the `ClearColorConfig` in the `Camera3d` and `Camera2d` components instead.


Co-authored-by: Alice <alice.i.cecile@gmail.com>
Co-authored-by: Alice Cecile <alice.i.cecile@gmail.com>
Co-authored-by: devil-ira <justthecooldude@gmail.com>
Co-authored-by: Carter Anderson <mcanders1@gmail.com>
2022-08-08 21:36:35 +00:00
Marc-Stefan Cassola
765e8d7dca Fix lifetime bound on From impl for NonSendMut -> Mut (#5560)
# Objective

Fixes #5456
2022-08-04 22:09:51 +00:00
Boxy
eabcd27d93 make WorldQuery very flat (#5205)
# Objective

Simplify the worldquery trait hierarchy as much as possible by putting it all in one trait. If/when gats are stabilised this can be trivially migrated over to use them, although that's not why I made this PR, those reasons are:
- Moves all of the conceptually related unsafe code for a worldquery next to eachother
- Removes now unnecessary traits simplifying the "type system magic" in bevy_ecs

---

## Changelog

All methods/functions/types/consts on `FetchState` and `Fetch` traits have been moved to the `WorldQuery` trait and the other traits removed. `WorldQueryGats` now only contains an `Item` and `Fetch` assoc type.

## Migration Guide
Implementors should move items in impls to the `WorldQuery/Gats` traits and remove any `Fetch`/`FetchState` impls
Any use sites of items in the `Fetch`/`FetchState` traits should be updated to use the `WorldQuery` trait items instead


Co-authored-by: Carter Anderson <mcanders1@gmail.com>
2022-08-04 21:51:02 +00:00
PROMETHIA-27
f3b5bf029c Add FromWorld bound to T in Local<T> (#5481)
# Objective

Currently, actually using a `Local` on a system requires that it be `T: FromWorld`, but that requirement is only expressed on the `SystemParam` machinery, which leads to the confusing error message for when the user attempts to add an invalid system. By adding these bounds to `Local` directly, it improves clarity on usage and semantics. 

## Solution

- Add `T: FromWorld` bound to `Local`'s definition

## Migration Guide

- It might be possible for references to `Local`s without `T: FromWorld` to exist, but these should be exceedingly rare and probably dead code. In the event that one of these is encountered, the easiest solutions are to delete the code or wrap the inner `T` in an `Option` to allow it to be default constructed to `None`.
2022-08-01 16:50:11 +00:00
ira
83a9e16158 Replace many_for_each_mut with iter_many_mut. (#5402)
# Objective
Replace `many_for_each_mut` with `iter_many_mut` using the same tricks to avoid aliased mutability that `iter_combinations_mut` uses.

<sub>I tried rebasing the draft PR I made for this before and it died. F</sub>
## Why
`many_for_each_mut` is worse for a few reasons:
1. The closure prevents the use of `continue`, `break`, and `return` behaves like a limited `continue`.
2. rustfmt will crumple it and double the indentation when the line gets too long.
    ```rust
    query.many_for_each_mut(
        &entity_list,
        |(mut transform, velocity, mut component_c)| {
            // Double trouble.
        },
    );
    ```
3. It is more surprising to have `many_for_each_mut` as a mutable counterpart to `iter_many` than `iter_many_mut`.
4. It required a separate unsafe fn; more unsafe code to maintain.
5. The `iter_many_mut` API matches the existing `iter_combinations_mut` API.

Co-authored-by: devil-ira <justthecooldude@gmail.com>
2022-07-30 01:38:13 +00:00
Boxy
be19c696bd Add missing ReadOnly = Self bound (#5462)
# Objective
`ReadOnlyWorldQuery` should have required `Self::ReadOnly = Self` so that calling `.iter()` on a readonly query is equivelent to calling `iter_mut()`.

## Solution

add `ReadOnly = Self` to the definition of `ReadOnlyWorldQuery`

---

## Changelog

ReadOnlyWorldQuery's `ReadOnly` assoc type is now always equal to `Self`

## Migration Guide

Make `Self::ReadOnly = Self` hold
2022-07-27 06:49:36 +00:00
Jerome Humbert
d4787111a3 Conversion of ResMut and NonSendMut to Mut (#5438)
# Objective

Enable treating components and resources equally, which can
simplify the implementation of some systems where only the change
detection feature is relevant and not the kind of object (resource or
component).

## Solution

Implement `From<ResMut<T>>` and `From<NonSendMut<T>>` for
`Mut`. Since the 3 structs are similar, and only differ by their system
param role, the conversion is trivial.

---

## Changelog

Added - `From<ResMut>` and `From<NonSendMut>` for `Mut<T>`.
2022-07-25 16:11:29 +00:00
Rob Parrett
cfee0e882e Fix various typos (#5417)
## Objective

- Fix some typos

## Solution

- Fix em. 
- My favorite was `maxizimed`
2022-07-21 20:46:54 +00:00