bevy/crates/bevy_reflect/derive/src/utility.rs
Gino Valente 6183b56b5d
bevy_reflect: Reflect remote types (#6042)
# Objective

The goal with this PR is to allow the use of types that don't implement
`Reflect` within the reflection API.

Rust's [orphan
rule](https://doc.rust-lang.org/book/ch10-02-traits.html#implementing-a-trait-on-a-type)
prevents implementing a trait on an external type when neither type nor
trait are owned by the implementor. This means that if a crate,
`cool_rust_lib`, defines a type, `Foo`, then a user cannot use it with
reflection. What this means is that we have to ignore it most of the
time:

```rust
#[derive(Reflect)]
struct SomeStruct {
  #[reflect(ignore)]
  data: cool_rust_lib::Foo
}
```

Obviously, it's impossible to implement `Reflect` on `Foo`. But does it
*have* to be?

Most of reflection doesn't deal with concrete types— it's almost all
using `dyn Reflect`. And being very metadata-driven, it should
theoretically be possible. I mean,
[`serde`](https://serde.rs/remote-derive.html) does it.

## Solution

> Special thanks to @danielhenrymantilla for their help reviewing this
PR and offering wisdom wrt safety.

Taking a page out of `serde`'s book, this PR adds the ability to easily
use "remote types" with reflection. In this context, a "remote type" is
the external type for which we have no ability to implement `Reflect`.

This adds the `#[reflect_remote(...)]` attribute macro, which is used to
generate "remote type wrappers". All you have to do is define the
wrapper exactly the same as the remote type's definition:

```rust
// Pretend this is our external crate
mod cool_rust_lib {
  #[derive(Default)]
  struct Foo {
    pub value: String
  }
}

#[reflect_remote(cool_rust_lib::Foo)]
struct FooWrapper {
  pub value: String
}
```

> **Note:** All fields in the external type *must* be public. This could
be addressed with a separate getter/setter attribute either in this PR
or in another one.

The macro takes this user-defined item and transforms it into a newtype
wrapper around the external type, marking it as `#[repr(transparent)]`.
The fields/variants defined by the user are simply used to build out the
reflection impls.

Additionally, it generates an implementation of the new trait,
`ReflectRemote`, which helps prevent accidental misuses of this API.

Therefore, the output generated by the macro would look something like:

```rust
#[repr(transparent)]
struct FooWrapper(pub cool_rust_lib::Foo);

impl ReflectRemote for FooWrapper {
  type Remote = cool_rust_lib::Foo;

  // transmutation methods...
}

// reflection impls...
// these will acknowledge and make use of the `value` field
```

Internally, the reflection API will pass around the `FooWrapper` and
[transmute](https://doc.rust-lang.org/std/mem/fn.transmute.html) it
where necessary. All we have to do is then tell `Reflect` to do that. So
rather than ignoring the field, we tell `Reflect` to use our wrapper
using the `#[reflect(remote = ...)]` field attribute:

```rust
#[derive(Reflect)]
struct SomeStruct {
  #[reflect(remote = FooWrapper)]
  data: cool_rust_lib::Foo
}
```

#### Other Macros & Type Data

Because this macro consumes the defined item and generates a new one, we
can't just put our macros anywhere. All macros that should be passed to
the generated struct need to come *below* this macro. For example, to
derive `Default` and register its associated type data:

```rust
//  GOOD
#[reflect_remote(cool_rust_lib::Foo)]
#[derive(Default)]
#[reflect(Default)]
struct FooWrapper {
  pub value: String
}

//  BAD
#[derive(Default)]
#[reflect_remote(cool_rust_lib::Foo)]
#[reflect(Default)]
struct FooWrapper {
  pub value: String
}
```

#### Generics

Generics are forwarded to the generated struct as well. They should also
be defined in the same order:

```rust
#[reflect_remote(RemoteGeneric<'a, T1, T2>)]
struct GenericWrapper<'a, T1, T2> {
  pub foo: &'a T1,
  pub bar: &'a T2,
}
```

> Naming does *not* need to match the original definition's. Only order
matters here.

> Also note that the code above is just a demonstration and doesn't
actually compile since we'd need to enforce certain bounds (e.g. `T1:
Reflect`, `'a: 'static`, etc.)

#### Nesting

And, yes, you can nest remote types:

```rust
#[reflect_remote(RemoteOuter)]
struct OuterWrapper {
  #[reflect(remote = InnerWrapper)]
  pub inner: RemoteInner
}

#[reflect_remote(RemoteInner)]
struct InnerWrapper(usize);
```

#### Assertions

This macro will also generate some compile-time assertions to ensure
that the correct types are used. It's important we catch this early so
users don't have to wait for something to panic. And it also helps keep
our `unsafe` a little safer.

For example, a wrapper definition that does not match its corresponding
remote type will result in an error:

```rust
mod external_crate {
  pub struct TheirStruct(pub u32);
}

#[reflect_remote(external_crate::TheirStruct)]
struct MyStruct(pub String); // ERROR: expected type `u32` but found `String`
```

<details>
<summary>Generated Assertion</summary>

```rust
const _: () = {
  #[allow(non_snake_case)]
  #[allow(unused_variables)]
  #[allow(unused_assignments)]
  #[allow(unreachable_patterns)]
  #[allow(clippy::multiple_bound_locations)]
  fn assert_wrapper_definition_matches_remote_type(
    mut __remote__: external_crate::TheirStruct,
  ) {
    __remote__.0 = (|| -> ::core::option::Option<String> { None })().unwrap();
  }
};
```

</details>

Additionally, using the incorrect type in a `#[reflect(remote = ...)]`
attribute should result in an error:

```rust
mod external_crate {
  pub struct TheirFoo(pub u32);
  pub struct TheirBar(pub i32);
}

#[reflect_remote(external_crate::TheirFoo)]
struct MyFoo(pub u32);

#[reflect_remote(external_crate::TheirBar)]
struct MyBar(pub i32);

#[derive(Reflect)]
struct MyStruct {
  #[reflect(remote = MyBar)] // ERROR: expected type `TheirFoo` but found struct `TheirBar`
  foo: external_crate::TheirFoo
}
```

<details>
<summary>Generated Assertion</summary>

```rust
const _: () = {
    struct RemoteFieldAssertions;
    impl RemoteFieldAssertions {
        #[allow(non_snake_case)]
        #[allow(clippy::multiple_bound_locations)]
        fn assert__foo__is_valid_remote() {
            let _: <MyBar as bevy_reflect::ReflectRemote>::Remote = (|| -> ::core::option::Option<external_crate::TheirFoo> {
              None
            })().unwrap();
        }
    }
};
```

</details>

### Discussion

There are a couple points that I think still need discussion or
validation.

- [x] 1. `Any` shenanigans

~~If we wanted to downcast our remote type from a `dyn Reflect`, we'd
have to first downcast to the wrapper then extract the inner type. This
PR has a [commit](b840db9f74cb6d357f951cb11b150d46bac89ee2) that
addresses this by making all the `Reflect::*any` methods return the
inner type rather than the wrapper type. This allows us to downcast
directly to our remote type.~~

~~However, I'm not sure if this is something we want to do. For
unknowing users, it could be confusing and seemingly inconsistent. Is it
worth keeping? Or should this behavior be removed?~~

I think this should be fine. The remote wrapper is an implementation
detail and users should not need to downcast to the wrapper type. Feel
free to let me know if there are other opinions on this though!

- [x] 2. Implementing `Deref/DerefMut` and `From`

~~We don't currently do this, but should we implement other traits on
the generated transparent struct? We could implement `Deref`/`DerefMut`
to easily access the inner type. And we could implement `From` for
easier conversion between the two types (e.g. `T: Into<Foo>`).~~ As
mentioned in the comments, we probably don't need to do this. Again, the
remote wrapper is an implementation detail, and should generally not be
used directly.
     
- [x] 3. ~~Should we define a getter/setter field attribute in this PR
as well or leave it for a future one?~~ I think this should be saved for
a future PR

- [ ] 4. Any foreseeable issues with this implementation?

#### Alternatives

One alternative to defining our own `ReflectRemote` would be to use
[bytemuck's
`TransparentWrapper`](https://docs.rs/bytemuck/1.13.1/bytemuck/trait.TransparentWrapper.html)
(as suggested by @danielhenrymantilla).

This is definitely a viable option, as `ReflectRemote` is pretty much
the same thing as `TransparentWrapper`. However, the cost would be
bringing in a new crate— though, it is already in use in a few other
sub-crates like bevy_render.

I think we're okay just defining `ReflectRemote` ourselves, but we can
go the bytemuck route if we'd prefer offloading that work to another
crate.

---

## Changelog

* Added the `#[reflect_remote(...)]` attribute macro to allow `Reflect`
to be used on remote types
* Added `ReflectRemote` trait for ensuring proper remote wrapper usage
2024-08-12 19:12:53 +00:00

455 lines
15 KiB
Rust

//! General-purpose utility functions for internal usage within this crate.
use crate::derive_data::ReflectMeta;
use bevy_macro_utils::{
fq_std::{FQAny, FQOption, FQSend, FQSync},
BevyManifest,
};
use proc_macro2::{Ident, Span, TokenStream};
use quote::{quote, ToTokens};
use syn::parse::{Parse, ParseStream, Peek};
use syn::punctuated::Punctuated;
use syn::{spanned::Spanned, LitStr, Member, Path, Token, Type, WhereClause};
/// Returns the correct path for `bevy_reflect`.
pub(crate) fn get_bevy_reflect_path() -> Path {
BevyManifest::get_path_direct("bevy_reflect")
}
/// Returns the "reflected" ident for a given string.
///
/// # Example
///
/// ```
/// # use proc_macro2::Ident;
/// # // We can't import this method because of its visibility.
/// # fn get_reflect_ident(name: &str) -> Ident {
/// # let reflected = format!("Reflect{name}");
/// # Ident::new(&reflected, proc_macro2::Span::call_site())
/// # }
/// let reflected: Ident = get_reflect_ident("Hash");
/// assert_eq!("ReflectHash", reflected.to_string());
/// ```
pub(crate) fn get_reflect_ident(name: &str) -> Ident {
let reflected = format!("Reflect{name}");
Ident::new(&reflected, Span::call_site())
}
/// Helper struct used to process an iterator of `Result<Vec<T>, syn::Error>`,
/// combining errors into one along the way.
pub(crate) struct ResultSifter<T> {
items: Vec<T>,
errors: Option<syn::Error>,
}
/// Returns a [`Member`] made of `ident` or `index` if `ident` is None.
///
/// Rust struct syntax allows for `Struct { foo: "string" }` with explicitly
/// named fields. It allows the `Struct { 0: "string" }` syntax when the struct
/// is declared as a tuple struct.
///
/// ```
/// # fn main() {
/// struct Foo { field: &'static str }
/// struct Bar(&'static str);
/// let Foo { field } = Foo { field: "hi" };
/// let Bar { 0: field } = Bar { 0: "hello" };
/// let Bar(field) = Bar("hello"); // more common syntax
/// # }
/// ```
///
/// This function helps field access in context where you are declaring either
/// a tuple struct or a struct with named fields. If you don't have a field name,
/// it means you need to access the struct through an index.
pub(crate) fn ident_or_index(ident: Option<&Ident>, index: usize) -> Member {
ident.map_or_else(
|| Member::Unnamed(index.into()),
|ident| Member::Named(ident.clone()),
)
}
/// Options defining how to extend the `where` clause for reflection.
pub(crate) struct WhereClauseOptions<'a, 'b> {
meta: &'a ReflectMeta<'b>,
active_fields: Box<[Type]>,
}
impl<'a, 'b> WhereClauseOptions<'a, 'b> {
pub fn new(meta: &'a ReflectMeta<'b>) -> Self {
Self {
meta,
active_fields: Box::new([]),
}
}
pub fn new_with_fields(meta: &'a ReflectMeta<'b>, active_fields: Box<[Type]>) -> Self {
Self {
meta,
active_fields,
}
}
/// Extends the `where` clause for a type with additional bounds needed for the reflection impls.
///
/// The default bounds added are as follows:
/// - `Self` has the bounds `Any + Send + Sync`
/// - Type parameters have the bound `TypePath` unless `#[reflect(type_path = false)]` is present
/// - Active fields have the bounds `TypePath` and either `PartialReflect` if `#[reflect(from_reflect = false)]` is present
/// or `FromReflect` otherwise (or no bounds at all if `#[reflect(no_field_bounds)]` is present)
///
/// When the derive is used with `#[reflect(where)]`, the bounds specified in the attribute are added as well.
///
/// # Example
///
/// ```ignore (bevy_reflect is not accessible from this crate)
/// #[derive(Reflect)]
/// struct Foo<T, U> {
/// a: T,
/// #[reflect(ignore)]
/// b: U
/// }
/// ```
///
/// Generates the following where clause:
///
/// ```ignore (bevy_reflect is not accessible from this crate)
/// where
/// // `Self` bounds:
/// Self: Any + Send + Sync,
/// // Type parameter bounds:
/// T: TypePath,
/// U: TypePath,
/// // Field bounds
/// T: FromReflect + TypePath,
/// ```
///
/// If we had added `#[reflect(where T: MyTrait)]` to the type, it would instead generate:
///
/// ```ignore (bevy_reflect is not accessible from this crate)
/// where
/// // `Self` bounds:
/// Self: Any + Send + Sync,
/// // Type parameter bounds:
/// T: TypePath,
/// U: TypePath,
/// // Field bounds
/// T: FromReflect + TypePath,
/// // Custom bounds
/// T: MyTrait,
/// ```
///
/// And if we also added `#[reflect(no_field_bounds)]` to the type, it would instead generate:
///
/// ```ignore (bevy_reflect is not accessible from this crate)
/// where
/// // `Self` bounds:
/// Self: Any + Send + Sync,
/// // Type parameter bounds:
/// T: TypePath,
/// U: TypePath,
/// // Custom bounds
/// T: MyTrait,
/// ```
pub fn extend_where_clause(
&self,
where_clause: Option<&WhereClause>,
) -> proc_macro2::TokenStream {
// We would normally just use `Self`, but that won't work for generating things like assertion functions
// and trait impls for a type's reference (e.g. `impl FromArg for &MyType`)
let this = self.meta.type_path().true_type();
let required_bounds = self.required_bounds();
// Maintain existing where clause, if any.
let mut generic_where_clause = if let Some(where_clause) = where_clause {
let predicates = where_clause.predicates.iter();
quote! {where #this: #required_bounds, #(#predicates,)*}
} else {
quote!(where #this: #required_bounds,)
};
// Add additional reflection trait bounds
let predicates = self.predicates();
generic_where_clause.extend(quote! {
#predicates
});
generic_where_clause
}
/// Returns an iterator the where clause predicates to extended the where clause with.
fn predicates(&self) -> Punctuated<TokenStream, Token![,]> {
let mut predicates = Punctuated::new();
if let Some(type_param_predicates) = self.type_param_predicates() {
predicates.extend(type_param_predicates);
}
if let Some(field_predicates) = self.active_field_predicates() {
predicates.extend(field_predicates);
}
if let Some(custom_where) = self.meta.attrs().custom_where() {
predicates.push(custom_where.predicates.to_token_stream());
}
predicates
}
/// Returns an iterator over the where clause predicates for the type parameters
/// if they require one.
fn type_param_predicates(&self) -> Option<impl Iterator<Item = TokenStream> + '_> {
self.type_path_bound().map(|type_path_bound| {
self.meta
.type_path()
.generics()
.type_params()
.map(move |param| {
let ident = &param.ident;
quote!(#ident : #type_path_bound)
})
})
}
/// Returns an iterator over the where clause predicates for the active fields.
fn active_field_predicates(&self) -> Option<impl Iterator<Item = TokenStream> + '_> {
if self.meta.attrs().no_field_bounds() {
None
} else {
let bevy_reflect_path = self.meta.bevy_reflect_path();
let reflect_bound = self.reflect_bound();
// `TypePath` is always required for active fields since they are used to
// construct `NamedField` and `UnnamedField` instances for the `Typed` impl.
// Likewise, `GetTypeRegistration` is always required for active fields since
// they are used to register the type's dependencies.
Some(self.active_fields.iter().map(move |ty| {
quote!(
#ty : #reflect_bound
+ #bevy_reflect_path::TypePath
// Needed for `Typed` impls
+ #bevy_reflect_path::MaybeTyped
// Needed for `GetTypeRegistration` impls
+ #bevy_reflect_path::__macro_exports::RegisterForReflection
)
}))
}
}
/// The `PartialReflect` or `FromReflect` bound to use based on `#[reflect(from_reflect = false)]`.
fn reflect_bound(&self) -> TokenStream {
let bevy_reflect_path = self.meta.bevy_reflect_path();
if self.meta.from_reflect().should_auto_derive() {
quote!(#bevy_reflect_path::FromReflect)
} else {
quote!(#bevy_reflect_path::PartialReflect)
}
}
/// The `TypePath` bounds to use based on `#[reflect(type_path = false)]`.
fn type_path_bound(&self) -> Option<TokenStream> {
if self.meta.type_path_attrs().should_auto_derive() {
let bevy_reflect_path = self.meta.bevy_reflect_path();
Some(quote!(#bevy_reflect_path::TypePath))
} else {
None
}
}
/// The minimum required bounds for a type to be reflected.
fn required_bounds(&self) -> proc_macro2::TokenStream {
quote!(#FQAny + #FQSend + #FQSync)
}
}
impl<T> Default for ResultSifter<T> {
fn default() -> Self {
Self {
items: Vec::new(),
errors: None,
}
}
}
impl<T> ResultSifter<T> {
/// Sift the given result, combining errors if necessary.
pub fn sift(&mut self, result: Result<T, syn::Error>) {
match result {
Ok(data) => self.items.push(data),
Err(err) => {
if let Some(ref mut errors) = self.errors {
errors.combine(err);
} else {
self.errors = Some(err);
}
}
}
}
/// Associated method that provides a convenient implementation for [`Iterator::fold`].
pub fn fold(mut sifter: Self, result: Result<T, syn::Error>) -> Self {
sifter.sift(result);
sifter
}
/// Complete the sifting process and return the final result.
pub fn finish(self) -> Result<Vec<T>, syn::Error> {
if let Some(errors) = self.errors {
Err(errors)
} else {
Ok(self.items)
}
}
}
/// Turns an `Option<TokenStream>` into a `TokenStream` for an `Option`.
pub(crate) fn wrap_in_option(tokens: Option<proc_macro2::TokenStream>) -> proc_macro2::TokenStream {
match tokens {
Some(tokens) => quote! {
#FQOption::Some(#tokens)
},
None => quote! {
#FQOption::None
},
}
}
/// Contains tokens representing different kinds of string.
#[derive(Clone)]
pub(crate) enum StringExpr {
/// A string that is valid at compile time.
///
/// This is either a string literal like `"mystring"`,
/// or a string created by a macro like [`module_path`]
/// or [`concat`].
Const(proc_macro2::TokenStream),
/// A [string slice](str) that is borrowed for a `'static` lifetime.
Borrowed(proc_macro2::TokenStream),
/// An [owned string](String).
Owned(proc_macro2::TokenStream),
}
impl<T: ToString + Spanned> From<T> for StringExpr {
fn from(value: T) -> Self {
Self::from_lit(&LitStr::new(&value.to_string(), value.span()))
}
}
impl StringExpr {
/// Creates a [constant] [`StringExpr`] from a [`struct@LitStr`].
///
/// [constant]: StringExpr::Const
pub fn from_lit(lit: &LitStr) -> Self {
Self::Const(lit.to_token_stream())
}
/// Creates a [constant] [`StringExpr`] by interpreting a [string slice][str] as a [`struct@LitStr`].
///
/// [constant]: StringExpr::Const
pub fn from_str(string: &str) -> Self {
Self::Const(string.into_token_stream())
}
/// Returns tokens for an [owned string](String).
///
/// The returned expression will allocate unless the [`StringExpr`] is [already owned].
///
/// [already owned]: StringExpr::Owned
pub fn into_owned(self) -> proc_macro2::TokenStream {
match self {
Self::Const(tokens) | Self::Borrowed(tokens) => quote! {
::std::string::ToString::to_string(#tokens)
},
Self::Owned(owned) => owned,
}
}
/// Returns tokens for a statically borrowed [string slice](str).
pub fn into_borrowed(self) -> proc_macro2::TokenStream {
match self {
Self::Const(tokens) | Self::Borrowed(tokens) => tokens,
Self::Owned(owned) => quote! {
&#owned
},
}
}
/// Appends a [`StringExpr`] to another.
///
/// If both expressions are [`StringExpr::Const`] this will use [`concat`] to merge them.
pub fn appended_by(mut self, other: StringExpr) -> Self {
if let Self::Const(tokens) = self {
if let Self::Const(more) = other {
return Self::Const(quote! {
::core::concat!(#tokens, #more)
});
}
self = Self::Const(tokens);
}
let owned = self.into_owned();
let borrowed = other.into_borrowed();
Self::Owned(quote! {
#owned + #borrowed
})
}
}
impl Default for StringExpr {
fn default() -> Self {
StringExpr::from_str("")
}
}
impl FromIterator<StringExpr> for StringExpr {
fn from_iter<T: IntoIterator<Item = StringExpr>>(iter: T) -> Self {
let mut iter = iter.into_iter();
match iter.next() {
Some(mut expr) => {
for next in iter {
expr = expr.appended_by(next);
}
expr
}
None => Default::default(),
}
}
}
/// Returns a [`syn::parse::Parser`] which parses a stream of zero or more occurrences of `T`
/// separated by punctuation of type `P`, with optional trailing punctuation.
///
/// This is functionally the same as [`Punctuated::parse_terminated`],
/// but accepts a closure rather than a function pointer.
pub(crate) fn terminated_parser<T, P, F: FnMut(ParseStream) -> syn::Result<T>>(
terminator: P,
mut parser: F,
) -> impl FnOnce(ParseStream) -> syn::Result<Punctuated<T, P::Token>>
where
P: Peek,
P::Token: Parse,
{
let _ = terminator;
move |stream: ParseStream| {
let mut punctuated = Punctuated::new();
loop {
if stream.is_empty() {
break;
}
let value = parser(stream)?;
punctuated.push_value(value);
if stream.is_empty() {
break;
}
let punct = stream.parse()?;
punctuated.push_punct(punct);
}
Ok(punctuated)
}
}