为何Rust函数可满足Bevy的IntoSystemDescriptor<Params> trait？

问题 #

我有以下几个Rust函数：

fn add_people(mut commands: Commands) {
    commands.spawn((Person, Name("A".to_string())));
    commands.spawn((Person, Name("B".to_string())));
    commands.spawn((Person, Name("C".to_string())));
    commands.spawn((Person, Name("D".to_string())));
}

fn greet_people(query: Query<&Name, With<Person>>) {
    for name in query.iter() {
        println!("hello {}!", name.0);
    }
}
fn hello_world() {
    println!("hello world");
}

这些函数的类型分别为fn(Commands)、fn(Query<&Name, With<Person>>)、fn()，在Bevy框架中可直接传入add_startup_system和add_system方法（方法定义如下）：

pub fn add_startup_system<Params>(&mut self, system: impl IntoSystemDescriptor<Params>)
pub fn add_system<Params>(&mut self, system: impl IntoSystemDescriptor<Params>)

示例调用代码：

App::new()
    .add_plugins(DefaultPlugins)
    .add_startup_system(hello_world)
    .add_startup_system(add_people)
    .add_system(greet_people)
    .run();

IntoSystemDescriptor<Params> trait的定义如下：

pub trait IntoSystemDescriptor<Params> {
    fn into_descriptor(self) -> SystemDescriptor;
    /// Assigns a run criteria to the system. Can be a new descriptor or a label of a
    /// run criteria defined elsewhere.
    fn with_run_criteria<Marker>(
        self,
        run_criteria: impl IntoRunCriteria<Marker>,
    ) -> SystemDescriptor;

    /// Assigns a label to the system; there can be more than one, and it doesn't have to be unique.
    fn label(self, label: impl SystemLabel) -> SystemDescriptor;

    /// Specifies that the system should run before systems with the given label.
    fn before<Marker>(self, label: impl AsSystemLabel<Marker>) -> SystemDescriptor;

    /// Specifies that the system should run after systems with the given label.
    fn after<Marker>(self, label: impl AsSystemLabel<Marker>) -> SystemDescriptor;

    /// Marks this system as ambiguous with any system with the specified label.
    /// This means that execution order between these systems does not matter,
    /// which allows [some warnings](crate::schedule::ReportExecutionOrderAmbiguities) to be silenced.
    fn ambiguous_with<Marker>(self, label: impl AsSystemLabel<Marker>) -> SystemDescriptor;

    /// Specifies that this system should opt out of
    /// [execution order ambiguity detection](crate::schedule::ReportExecutionOrderAmbiguities).
    fn ignore_all_ambiguities(self) -> SystemDescriptor;

    /// Specifies that the system should run with other exclusive systems at the start of stage.
    fn at_start(self) -> SystemDescriptor;

    /// Specifies that the system should run with other exclusive systems after the parallel
    /// systems and before command buffer application.
    fn before_commands(self) -> SystemDescriptor;

    /// Specifies that the system should run with other exclusive systems at the end of stage.
    fn at_end(self) -> SystemDescriptor;
}

我无法理解为何这些函数类型会实现IntoSystemDescriptor<Params> trait，二者看起来并无关联，希望有人解释原因，或提供相关Rust特性的说明。

回答 #

核心原因是Rust的**Blanket Implementations（通用blanket实现）**特性，Bevy框架借助这个特性，为所有符合条件的函数类型自动实现了IntoSystemDescriptor<Params> trait。

具体细节：

• Bevy内部定义了一个泛型trait实现，针对所有满足「函数参数符合系统参数规范」的函数类型，大致结构如下：

  impl<F, Params> IntoSystemDescriptor<Params> for F
  where
      F: IntoSystem<Params, ()>, // IntoSystem是Bevy的核心trait，负责将普通函数转换为可调度的系统
  {
      fn into_descriptor(self) -> SystemDescriptor {
          // 内部逻辑：将函数包装为框架可识别的SystemDescriptor实例
      }
      // 其他trait方法的默认实现...
  }
  

• 你写的三个函数都满足IntoSystem<Params, ()>的约束：Bevy会自动解析函数参数（比如Commands、Query都是框架支持的系统参数类型），把普通函数转换成可被调度的系统对象，进而自动实现IntoSystemDescriptor。

简单来说，Bevy通过Rust的泛型和trait系统，完成了普通函数到框架系统描述符的自动适配，不需要你手动为每个函数编写trait实现，这就是你能直接把函数传给add_startup_system和add_system的原因。

内容的提问来源于stack exchange，提问作者alice pro