bevy/crates/bevy_reflect
Carter Anderson b3a1db60f2 Proper prehashing (#3963)
For some keys, it is too expensive to hash them on every lookup. Historically in Bevy, we have regrettably done the "wrong" thing in these cases (pre-computing hashes, then re-hashing them) because Rust's built in hashed collections don't give us the tools we need to do otherwise. Doing this is "wrong" because two different values can result in the same hash. Hashed collections generally get around this by falling back to equality checks on hash collisions. You can't do that if the key _is_ the hash. Additionally, re-hashing a hash increase the odds of collision!
 
#3959 needs pre-hashing to be viable, so I decided to finally properly solve the problem. The solution involves two different changes:

1. A new generalized "pre-hashing" solution in bevy_utils: `Hashed<T>` types, which store a value alongside a pre-computed hash. And `PreHashMap<K, V>` (which uses `Hashed<T>` internally) . `PreHashMap` is just an alias for a normal HashMap that uses `Hashed<T>` as the key and a new `PassHash` implementation as the Hasher. 
2. Replacing the `std::collections` re-exports in `bevy_utils` with equivalent `hashbrown` impls. Avoiding re-hashes requires the `raw_entry_mut` api, which isn't stabilized yet (and may never be ... `entry_ref` has favor now, but also isn't available yet). If std's HashMap ever provides the tools we need, we can move back to that. The latest version of `hashbrown` adds support for the `entity_ref` api, so we can move to that in preparation for an std migration, if thats the direction they seem to be going in. Note that adding hashbrown doesn't increase our dependency count because it was already in our tree.

In addition to providing these core tools, I also ported the "table identity hashing" in `bevy_ecs` to `raw_entry_mut`, which was a particularly egregious case.

The biggest outstanding case is `AssetPathId`, which stores a pre-hash. We need AssetPathId to be cheaply clone-able (and ideally Copy), but `Hashed<AssetPath>` requires ownership of the AssetPath, which makes cloning ids way more expensive. We could consider doing `Hashed<Arc<AssetPath>>`, but cloning an arc is still a non-trivial expensive that needs to be considered. I would like to handle this in a separate PR. And given that we will be re-evaluating the Bevy Assets implementation in the very near future, I'd prefer to hold off until after that conversation is concluded.
2022-02-18 03:26:01 +00:00
..
bevy_reflect_derive small and mostly pointless refactoring (#2934) 2022-02-13 22:33:55 +00:00
src Proper prehashing (#3963) 2022-02-18 03:26:01 +00:00
Cargo.toml Proper prehashing (#3963) 2022-02-18 03:26:01 +00:00
README.md Add readme as docs to relevant crates. (#2575) 2021-12-18 22:59:55 +00:00

Bevy Reflect

This crate enables you to dynamically interact with Rust types:

  • Derive the Reflect traits
  • Interact with fields using their names (for named structs) or indices (for tuple structs)
  • "Patch" your types with new values
  • Look up nested fields using "path strings"
  • Iterate over struct fields
  • Automatically serialize and deserialize via Serde (without explicit serde impls)
  • Trait "reflection"

Features

Derive the Reflect traits

// this will automatically implement the Reflect trait and the Struct trait (because the type is a struct)
#[derive(Reflect)]
struct Foo {
    a: u32,
    b: Bar,
    c: Vec<i32>,
    d: Vec<Bar>,
}

// this will automatically implement the Reflect trait and the TupleStruct trait (because the type is a tuple struct)
#[derive(Reflect)]
struct Bar(String);

#[derive(Reflect)]
struct Baz {
    value: f32,
}

// We will use this value to illustrate `bevy_reflect` features
let mut foo = Foo {
    a: 1,
    b: Bar("hello".to_string()),
    c: vec![1, 2],
    d: vec![Baz { value: 3.14 }],
};

Interact with fields using their names

assert_eq!(*foo.get_field::<u32>("a").unwrap(), 1);

*foo.get_field_mut::<u32>("a").unwrap() = 2;

assert_eq!(foo.a, 2);

"Patch" your types with new values

let mut dynamic_struct = DynamicStruct::default();
dynamic_struct.insert("a", 42u32);
dynamic_struct.insert("c", vec![3, 4, 5]);

foo.apply(&dynamic_struct);

assert_eq!(foo.a, 42);
assert_eq!(foo.c, vec![3, 4, 5]);

Look up nested fields using "path strings"

let value = *foo.get_path::<f32>("d[0].value").unwrap();
assert_eq!(value, 3.14);

Iterate over struct fields

for (i, value: &Reflect) in foo.iter_fields().enumerate() {
    let field_name = foo.name_at(i).unwrap();
    if let Ok(value) = value.downcast_ref::<u32>() {
        println!("{} is a u32 with the value: {}", field_name, *value);
    }
}

Automatically serialize and deserialize via Serde (without explicit serde impls)

let mut registry = TypeRegistry::default();
registry.register::<u32>();
registry.register::<i32>();
registry.register::<f32>();
registry.register::<String>();
registry.register::<Bar>();
registry.register::<Baz>();

let serializer = ReflectSerializer::new(&foo, &registry);
let serialized = ron::ser::to_string_pretty(&serializer, ron::ser::PrettyConfig::default()).unwrap();

let mut deserializer = ron::de::Deserializer::from_str(&serialized).unwrap();
let reflect_deserializer = ReflectDeserializer::new(&registry);
let value = reflect_deserializer.deserialize(&mut deserializer).unwrap();
let dynamic_struct = value.take::<DynamicStruct>().unwrap();

assert!(foo.reflect_partial_eq(&dynamic_struct).unwrap());

Trait "reflection"

Call a trait on a given &dyn Reflect reference without knowing the underlying type!

#[derive(Reflect)]
#[reflect(DoThing)]
struct MyType {
    value: String,
}

impl DoThing for MyType {
    fn do_thing(&self) -> String {
        format!("{} World!", self.value)
    }
}

#[reflect_trait]
pub trait DoThing {
    fn do_thing(&self) -> String;
}

// First, lets box our type as a Box<dyn Reflect>
let reflect_value: Box<dyn Reflect> = Box::new(MyType {
    value: "Hello".to_string(),
});

// This means we no longer have direct access to MyType or its methods. We can only call Reflect methods on reflect_value.
// What if we want to call `do_thing` on our type? We could downcast using reflect_value.downcast_ref::<MyType>(), but what if we
// don't know the type at compile time?

// Normally in rust we would be out of luck at this point. Lets use our new reflection powers to do something cool!
let mut type_registry = TypeRegistry::default()
type_registry.register::<MyType>();

// The #[reflect] attribute we put on our DoThing trait generated a new `ReflectDoThing` struct, which implements TypeData.
// This was added to MyType's TypeRegistration.
let reflect_do_thing = type_registry
    .get_type_data::<ReflectDoThing>(reflect_value.type_id())
    .unwrap();

// We can use this generated type to convert our `&dyn Reflect` reference to a `&dyn DoThing` reference
let my_trait: &dyn DoThing = reflect_do_thing.get(&*reflect_value).unwrap();

// Which means we can now call do_thing(). Magic!
println!("{}", my_trait.do_thing());

// This works because the #[reflect(MyTrait)] we put on MyType informed the Reflect derive to insert a new instance
// of ReflectDoThing into MyType's registration. The instance knows how to cast &dyn Reflect to &dyn MyType, because it
// knows that &dyn Reflect should first be downcasted to &MyType, which can then be safely casted to &dyn MyType

Why make this?

The whole point of Rust is static safety! Why build something that makes it easy to throw it all away?

  • Some problems are inherently dynamic (scripting, some types of serialization / deserialization)
  • Sometimes the dynamic way is easier
  • Sometimes the dynamic way puts less burden on your users to derive a bunch of traits (this was a big motivator for the Bevy project)