Initial support for auto traits with default bounds
This PR is part of ["MCP: Low level components for async drop"](https://github.com/rust-lang/compiler-team/issues/727)
Tracking issue: #138781
Summary: https://github.com/rust-lang/rust/pull/120706#issuecomment-1934006762
### Intro
Sometimes we want to use type system to express specific behavior and provide safety guarantees. This behavior can be specified by various "marker" traits. For example, we use `Send` and `Sync` to keep track of which types are thread safe. As the language develops, there are more problems that could be solved by adding new marker traits:
- to forbid types with an async destructor to be dropped in a synchronous context a trait like `SyncDrop` could be used [Async destructors, async genericity and completion futures](https://sabrinajewson.org/blog/async-drop).
- to support [scoped tasks](https://without.boats/blog/the-scoped-task-trilemma/) or in a more general sense to provide a [destruction guarantee](https://zetanumbers.github.io/book/myosotis.html) there is a desire among some users to see a `Leak` (or `Forget`) trait.
- Withoutboats in his [post](https://without.boats/blog/changing-the-rules-of-rust/) reflected on the use of `Move` trait instead of a `Pin`.
All the traits proposed above are supposed to be auto traits implemented for most types, and usually implemented automatically by compiler.
For backward compatibility these traits have to be added implicitly to all bound lists in old code (see below). Adding new default bounds involves many difficulties: many standard library interfaces may need to opt out of those default bounds, and therefore be infected with confusing `?Trait` syntax, migration to a new edition may contain backward compatibility holes, supporting new traits in the compiler can be quite difficult and so forth. Anyway, it's hard to evaluate the complexity until we try the system on a practice.
In this PR we introduce new optional lang items for traits that are added to all bound lists by default, similarly to existing `Sized`. The examples of such traits could be `Leak`, `Move`, `SyncDrop` or something else, it doesn't matter much right now (further I will call them `DefaultAutoTrait`'s). We want to land this change into rustc under an option, so it becomes available in bootstrap compiler. Then we'll be able to do standard library experiments with the aforementioned traits without adding hundreds of `#[cfg(not(bootstrap))]`s. Based on the experiments, we can come up with some scheme for the next edition, in which such bounds are added in a more targeted way, and not just everywhere.
Most of the implementation is basically a refactoring that replaces hardcoded uses of `Sized` with iterating over a list of traits including both `Sized` and the new traits when `-Zexperimental-default-bounds` is enabled (or just `Sized` as before, if the option is not enabled).
### Default bounds for old editions
All existing types, including generic parameters, are considered `Leak`/`Move`/`SyncDrop` and can be forgotten, moved or destroyed in generic contexts without specifying any bounds. New types that cannot be, for example, forgotten and do not implement `Leak` can be added at some point, and they should not be usable in such generic contexts in existing code.
To both maintain this property and keep backward compatibility with existing code, the new traits should be added as default bounds _everywhere_ in previous editions. Besides the implicit `Sized` bound contexts that includes supertrait lists and trait lists in trait objects (`dyn Trait1 + ... + TraitN`). Compiler should also generate implicit `DefaultAutoTrait` implementations for foreign types (`extern { type Foo; }`) because they are also currently usable in generic contexts without any bounds.
#### Supertraits
Adding the new traits as supertraits to all existing traits is potentially necessary, because, for example, using a `Self` param in a trait's associated item may be a breaking change otherwise:
```rust
trait Foo: Sized {
fn new() -> Option<Self>; // ERROR: `Option` requires `DefaultAutoTrait`, but `Self` is not `DefaultAutoTrait`
}
// desugared `Option`
enum Option<T: DefaultAutoTrait + Sized> {
Some(T),
None,
}
```
However, default supertraits can significantly affect compiler performance. For example, if we know that `T: Trait`, the compiler would deduce that `T: DefaultAutoTrait`. It also implies proving `F: DefaultAutoTrait` for each field `F` of type `T` until an explicit impl is be provided.
If the standard library is not modified, then even traits like `Copy` or `Send` would get these supertraits.
In this PR for optimization purposes instead of adding default supertraits, bounds are added to the associated items:
```rust
// Default bounds are generated in the following way:
trait Trait {
fn foo(&self) where Self: DefaultAutoTrait {}
}
// instead of this:
trait Trait: DefaultAutoTrait {
fn foo(&self) {}
}
```
It is not always possible to do this optimization because of backward compatibility:
```rust
pub trait Trait<Rhs = Self> {}
pub trait Trait1 : Trait {} // ERROR: `Rhs` requires `DefaultAutoTrait`, but `Self` is not `DefaultAutoTrait`
```
or
```rust
trait Trait {
type Type where Self: Sized;
}
trait Trait2<T> : Trait<Type = T> {} // ERROR: `???` requires `DefaultAutoTrait`, but `Self` is not `DefaultAutoTrait`
```
Therefore, `DefaultAutoTrait`'s are still being added to supertraits if the `Self` params or type bindings were found in the trait header.
#### Trait objects
Trait objects requires explicit `+ Trait` bound to implement corresponding trait which is not backward compatible:
```rust
fn use_trait_object(x: Box<dyn Trait>) {
foo(x) // ERROR: `foo` requires `DefaultAutoTrait`, but `dyn Trait` is not `DefaultAutoTrait`
}
// implicit T: DefaultAutoTrait here
fn foo<T>(_: T) {}
```
So, for a trait object `dyn Trait` we should add an implicit bound `dyn Trait + DefaultAutoTrait` to make it usable, and allow relaxing it with a question mark syntax `dyn Trait + ?DefaultAutoTrait` when it's not necessary.
#### Foreign types
If compiler doesn't generate auto trait implementations for a foreign type, then it's a breaking change if the default bounds are added everywhere else:
```rust
// implicit T: DefaultAutoTrait here
fn foo<T: ?Sized>(_: &T) {}
extern "C" {
type ExternTy;
}
fn forward_extern_ty(x: &ExternTy) {
foo(x); // ERROR: `foo` requires `DefaultAutoTrait`, but `ExternTy` is not `DefaultAutoTrait`
}
```
We'll have to enable implicit `DefaultAutoTrait` implementations for foreign types at least for previous editions:
```rust
// implicit T: DefaultAutoTrait here
fn foo<T: ?Sized>(_: &T) {}
extern "C" {
type ExternTy;
}
impl DefaultAutoTrait for ExternTy {} // implicit impl
fn forward_extern_ty(x: &ExternTy) {
foo(x); // OK
}
```
### Unresolved questions
New default bounds affect all existing Rust code complicating an already complex type system.
- Proving an auto trait predicate requires recursively traversing the type and proving the predicate for it's fields. This leads to a significant performance regression. Measurements for the stage 2 compiler build show up to 3x regression.
- We hope that fast path optimizations for well known traits could mitigate such regressions at least partially.
- New default bounds trigger some compiler bugs in both old and new trait solver.
- With new default bounds we encounter some trait solver cycle errors that break existing code.
- We hope that these cases are bugs that can be addressed in the new trait solver.
Also migration to a new edition could be quite ugly and enormous, but that's actually what we want to solve. For other issues there's a chance that they could be solved by a new solver.
Move methods from `Map` to `TyCtxt`, part 5.
This eliminates all methods on `Map`. Actually removing `Map` will occur in a follow-up PR.
A follow-up to #137504.
r? `@Zalathar`
Various local trait item iteration cleanups
Adding a trait impl for `Foo` unconditionally affected all queries that are interested in a completely independent trait `Bar`. Perf has no effect on this. We probably don't have a good perf test for this tho.
r? `@compiler-errors`
I am unsure about 9d05efb66f as it doesn't improve anything wrt incremental, because we still do all the checks for valid `Drop` impls, which subsequently will still invoke many queries and basically keep the depgraph the same.
I want to do
9549077a47/compiler/rustc_middle/src/ty/trait_def.rs (L141)
but would leave that to a follow-up PR, this one changes enough things as it is
Skip suggest impl or dyn when poly trait is not a real trait
Fixes#139174
When `poly_trait_ref` is not a real trait, we should stop suggesting `impl` and `dyn` to avoid false positives. 3 cases were added to the ui test.
0b45675cfc/compiler/rustc_hir_analysis/src/hir_ty_lowering/lint.rs (L88-L93)
In the first commit, I submitted the test and passed it. In the second commit, I modified the code and we can see the changes in the test.
r? compiler
Remove ScopeDepth
The scope depth was tracked, but never seemed to be used for anything.
Every single place that used `(Scope, ScopeDepth)`, matched it on `(p, _)`.
Do not rely on `type_var_origin` in `OrphanCheckErr::NonLocalInputType`
The ordering of ty var unification means that we may end up with a root variable whose ty var origin is from another item's params.
Let's not rely on this by just unifying the infer vars with the params of the impl + resolving. It's kinda goofy but it's clearer IMO.
Fixes#132826.
r? `@fmease` or `@lcnr`
Use `Option<Ident>` for lowered param names.
Parameter patterns are lowered to an `Ident` by `lower_fn_params_to_names`, which is used when lowering bare function types, trait methods, and foreign functions. Currently, there are two exceptional cases where the lowered param can become an empty `Ident`.
- If the incoming pattern is an empty `Ident`. This occurs if the parameter is anonymous, e.g. in a bare function type.
- If the incoming pattern is neither an ident nor an underscore. Any such parameter will have triggered a compile error (hence the `span_delayed_bug`), but lowering still occurs.
This commit replaces these empty `Ident` results with `None`, which eliminates a number of `kw::Empty` uses, and makes it impossible to fail to check for these exceptional cases.
Note: the `FIXME` comment in `is_unwrap_or_empty_symbol` is removed. It actually should have been removed in #138482, the precursor to this PR. That PR changed the lowering of wild patterns to `_` symbols instead of empty symbols, which made the mentioned underscore check load-bearing.
r? ``@compiler-errors``
Parameter patterns are lowered to an `Ident` by
`lower_fn_params_to_names`, which is used when lowering bare function
types, trait methods, and foreign functions. Currently, there are two
exceptional cases where the lowered param can become an empty `Ident`.
- If the incoming pattern is an empty `Ident`. This occurs if the
parameter is anonymous, e.g. in a bare function type.
- If the incoming pattern is neither an ident nor an underscore. Any
such parameter will have triggered a compile error (hence the
`span_delayed_bug`), but lowering still occurs.
This commit replaces these empty `Ident` results with `None`, which
eliminates a number of `kw::Empty` uses, and makes it impossible to fail
to check for these exceptional cases.
Note: the `FIXME` comment in `is_unwrap_or_empty_symbol` is removed. It
actually should have been removed in #138482, the precursor to this PR.
That PR changed the lowering of wild patterns to `_` symbols instead of
empty symbols, which made the mentioned underscore check load-bearing.
Move `hir::Item::ident` into `hir::ItemKind`.
`hir::Item` has an `ident` field.
- It's always non-empty for these item kinds: `ExternCrate`, `Static`, `Const`, `Fn`, `Macro`, `Mod`, `TyAlias`, `Enum`, `Struct`, `Union`, Trait`, TraitAalis`.
- It's always empty for these item kinds: `ForeignMod`, `GlobalAsm`, `Impl`.
- For `Use`, it is non-empty for `UseKind::Single` and empty for `UseKind::{Glob,ListStem}`.
All of this is quite non-obvious; the only documentation is a single comment saying "The name might be a dummy name in case of anonymous items". Some sites that handle items check for an empty ident, some don't. This is a very C-like way of doing things, but this is Rust, we have sum types, we can do this properly and never forget to check for the exceptional case and never YOLO possibly empty identifiers (or possibly dummy spans) around and hope that things will work out.
This is step towards `kw::Empty` elimination (#137978).
r? `@fmease`
`hir::Item` has an `ident` field.
- It's always non-empty for these item kinds: `ExternCrate`, `Static`,
`Const`, `Fn`, `Macro`, `Mod`, `TyAlias`, `Enum`, `Struct`, `Union`,
Trait`, TraitAalis`.
- It's always empty for these item kinds: `ForeignMod`, `GlobalAsm`,
`Impl`.
- For `Use`, it is non-empty for `UseKind::Single` and empty for
`UseKind::{Glob,ListStem}`.
All of this is quite non-obvious; the only documentation is a single
comment saying "The name might be a dummy name in case of anonymous
items". Some sites that handle items check for an empty ident, some
don't. This is a very C-like way of doing things, but this is Rust, we
have sum types, we can do this properly and never forget to check for
the exceptional case and never YOLO possibly empty identifiers (or
possibly dummy spans) around and hope that things will work out.
The commit is large but it's mostly obvious plumbing work. Some notable
things.
- A similar transformation makes sense for `ast::Item`, but this is
already a big change. That can be done later.
- Lots of assertions are added to item lowering to ensure that
identifiers are empty/non-empty as expected. These will be removable
when `ast::Item` is done later.
- `ItemKind::Use` doesn't get an `Ident`, but `UseKind::Single` does.
- `lower_use_tree` is significantly simpler. No more confusing `&mut
Ident` to deal with.
- `ItemKind::ident` is a new method, it returns an `Option<Ident>`. It's
used with `unwrap` in a few places; sometimes it's hard to tell
exactly which item kinds might occur. None of these unwraps fail on
the test suite. It's conceivable that some might fail on alternative
input. We can deal with those if/when they happen.
- In `trait_path` the `find_map`/`if let` is replaced with a loop, and
things end up much clearer that way.
- `named_span` no longer checks for an empty name; instead the call site
now checks for a missing identifier if necessary.
- `maybe_inline_local` doesn't need the `glob` argument, it can be
computed in-function from the `renamed` argument.
- `arbitrary_source_item_ordering::check_mod` had a big `if` statement
that was just getting the ident from the item kinds that had one. It
could be mostly replaced by a single call to the new `ItemKind::ident`
method.
- `ItemKind` grows from 56 to 64 bytes, but `Item` stays the same size,
and that's what matters, because `ItemKind` only occurs within `Item`.
Add an opt-out in pretty printing for RTN rendering
Today, we render RPITIT types like `impl Sized { T::method(..) }` when RTN is enabled. This is very useful for diagnostics, since it's often not clear what the `impl Sized` type means by itself, and it makes it clear that that's an RPITIT that can be bounded using RTN syntax. See #115624.
However, since we don't distinguish types that are rendered for the purposes of printing messages vs suggestions, this representation leaks into suggestions and turns into code that can't be parsed. This PR adds a new `with_types_for_suggestion! {}` and `with_types_for_signature! {}` options to the pretty printing architecture to make it clear that we're rendering a type for code suggestions.
This can be applied later as we find that we need it.
make precise capturing args in rustdoc Json typed
close#137616
This PR includes below changes.
- Add `rustc_hir::PreciseCapturingArgKind` which allows the query system to return a arg's data.
- Add `rustdoc::clean::types::PreciseCapturingArg` and change to use it.
- Add `rustdoc-json-types::PreciseCapturingArg` and change to use it.
- Update `tests/rustdoc-json/impl-trait-precise-capturing.rs`.
- Bump `rustdoc_json_types::FORMAT_VERSION`.
