Move more query system code
Towards the goal of eliminating `rustc_query_system`, this commit moves some code from `rustc_query_system` to `rustc_middle` and `rustc_query_impl`, and from `rustc_middle` to `rustc_query_impl`.
r? @Zalathar
Move the needs-drop check for `arena_cache` queries out of macro code
This is slightly simpler than before, because now the macro only needs to call a single function, and can just unconditionally supply `tcx` and a typed arena.
There should be no actual change to compiler behaviour.
Move `impl Interner for TyCtxt` to its own submodule
This impl is several hundred lines of mostly self-contained, mostly boilerplate code that can be extracted out of the dauntingly large `rustc_middle::ty::context` module.
- The trait and its impl were introduced by https://github.com/rust-lang/rust/pull/97287.
---
There should be no change to compiler behaviour.
Requiring `fn` in the macro syntax makes it a little more obvious that the
macro declares functions with those names, and makes them easier to grep for.
Update mgca to use `type const` syntax instead of the `#[type_const]` attribute.
This PR changes the `#[type_const]` attribute to the `type const` syntax for https://github.com/rust-lang/rust/issues/132980.
This will fixes https://github.com/rust-lang/rust/issues/151273 and similar issues, since we need to check `type const` of items before expansion. The move to add a syntax was mentioned here: https://github.com/rust-lang/rust/pull/151289#issuecomment-3765241397
The first part of this PR adds support by allowing `type const <IDENT>: <TYPE> { = <EXPR> };` syntax in `rustc_parse/src/parser/item.rs`.
The next part since the AST item does not contain enough information to determine if we have a `type const` was rework `ConstItemRhs` into `ConstItemRhsKind` to store the information since we no longer have the attribute acting as a source of extra data/metadata.
The hir node `ConstItemRhsKind` current shape mostly works, except in the case of `TraitItemKind` where it is an option. I initially went about giving `hir::ConstItemRhsKind` a similar form the AST, but it touches a lot more lines of code and files so because of that, the less invasive option was to add a simple boolean flag to `TraitItemKind`.
The forth part of this PR includes adding a query I called `is_rhs_type_const` so that we can handle both local and foreign def_ids.
The fifth aspect of the PR is adding a `mgca_type_const_syntax` feature gate that is checked before expansion. The standard mgca feature gate is ran after expansion. This feature gate allows for conditional compilation (e.g #[cfg(..)]) of the `type const` syntax in nightly without `min_generic_const_args` being enabled.
The last bit is updating all the the tests that used the `#[type_const]` attribute to use the new syntax that failed because of the changes. This is the bulk of touched/edited files in the PR.
r? @BoxyUwU
@rustbot label +F-associated_const_equality +F-min_generic_const_args
Revert, but without type const.
Update symbol for feature err, then update suggestion output, and lastly update tests that change because of those.
Update these new tests with the correct syntax, and few existing tests with the new outputs the merge with main added.
Fix for tidyfmt and some errors when manually resolving a merge conflicts.
Update these tests to use update error messages and type const syntax.
Update comments and error message to use new syntax instead of old type_const attribute.
Remove the type_const attribute
update some more tests to use the new syntax.
Update these test cases.
update feature gate test
Change gate logic for `mgca_type_const_syntax` to work also if `min_generic_const_args` is enabled.
Create a new feature gate that checks for the feature before expansion.
Make rustfmt handle the `type const` syntax correctly.
Add a convience method to check if a RhsKind is type const.
Rename `Const` discriminant to `Body` for `ConstItemRhsKind`
Give the `TraitItemKind` flag an enum instead of a simple bool to better describe what the flag is for.
Update formatting for these match statements.
Update clippy test to use type const syntax.
Update test to use type const syntax.
update rustfmt to match ast items.
Update clippy to match ast and hir items.
Few more test cases that used old attribute, instead of 'type const'
Update to match the output from the feature gate checks.
tidyfmt adjustments.
Update the is_type_const, so I can constrain record!(..) in encoder.rs
Update conditional compilation test.
Move the feature gate to after expansion to allow for cfg(...) to work.
Update some more tests to use the new syntax.
Update type const tests in associated-const-bindings to use new syntax.
Don't check based off the attribute, but the item here.
Update some tests outside of the const_generics folder that were using #[type_const]
update the tests in associated consts that use #[type_const] to use type const
Update these mgca tests with the type const syntax.
Add a flag to TraitItemKind for detecting type const for now. Maybe later change ItemConstRhs to have optional consts but that touches a lot more lines of code.
Don't need into for these now that it's a query.
Add is_type_const query to handle foreign def ids.
update this test to use type const syntax.
Fix logic here, we only want to lower if there is expression in this case.
Update built-in macros to use ConstItemRhsKind
Update more instance of the old ConstItemRhs.
Rename ConstItemKind to ConstItemRhsKind, I noticed there is a typed called ConstantItemKind, so add the Rhs to the name to avoid confusion.
Update lower to use ConstItemKind
Add an other helper method to check if the rhs kinda has an expr.
Update item parse to use ConstItemKind enum.
Felt the field name could a be little clear when editing a few other things.
Change the ConstItem struct see know if we have a type const or regular const.
Make sure this syntax is properly feature gated.
Remove `QueryDispatcher`
`QueryDispatcher` is a trait that existed purely because `rustc_query_system` had code that didn't have access to `TyCtxt`. That is no longer the case, so the trait can be removed.
r? @Zalathar
Implement MVP for opaque generic const arguments
This is meant to be the interim successor to generic const expressions.
Essentially, const item RHS's will be allowed to do arbitrary const
operations using generics. The limitation is that these const items will
be treated opaquely, like ADTs in nominal typing, such that uses of them
will only be equal if the same const item is referenced. In other words,
two const items with the exact same RHS will not be considered equal.
I also added some logic to check feature gates that depend on others
being enabled (like oGCA depending on mGCA).
### Coherence
During coherence, OGCA consts should be normalized ambiguously because
they are opaque but eventually resolved to a real value. We don't want
two OGCAs that have the same value to be treated as distinct for
coherence purposes. (Just like opaque types.)
This actually doesn't work yet because there are pre-existing
fundamental issues with equate relations involving consts that need to
be normalized. The problem is that we normalize only one layer of the
const item and don't actually process the resulting anon const. Normally
the created inference variable should be handled, which in this case
would cause us to hit the anon const, but that's not happening.
Specifically, `visit_const` on `Generalizer` should be updated to be
similar to `visit_ty`.
r? @BoxyUwU
This is meant to be the interim successor to generic const expressions.
Essentially, const item RHS's will be allowed to do arbitrary const
operations using generics. The limitation is that these const items will
be treated opaquely, like ADTs in nominal typing, such that uses of them
will only be equal if the same const item is referenced. In other words,
two const items with the exact same RHS will not be considered equal.
I also added some logic to check feature gates that depend on others
being enabled (like oGCA depending on mGCA).
= Coherence =
During coherence, OGCA consts should be normalized ambiguously because
they are opaque but eventually resolved to a real value. We don't want
two OGCAs that have the same value to be treated as distinct for
coherence purposes. (Just like opaque types.)
This actually doesn't work yet because there are pre-existing
fundamental issues with equate relations involving consts that need to
be normalized. The problem is that we normalize only one layer of the
const item and don't actually process the resulting anon const. Normally
the created inference variable should be handled, which in this case
would cause us to hit the anon const, but that's not happening.
Specifically, `visit_const` on `Generalizer` should be updated to be
similar to `visit_ty`.
Start cutting down `rustc_query_system`
The query system is implemented in `rustc_query_system`, `rustc_middle`, and `rustc_query_impl`. `rustc_query_system` is hamstrung by not having access to `TyCtxt`, and there seems to be consensus to eliminate it. It's contents can be moved into the other two crates. Moving as much stuff as possible to `rustc_query_impl` is preferred, because `rustc_middle` is already so big.
This PR starts this process. It moves one small function to `rustc_middle` and a good chunk of code to `rustc_query_impl`.
Once `rustc_query_system` is gone (or at least shrunk down a lot more) some of the traits like `DepContext`, `QueryContext`, and `QueryDispatcher` will be removable.
r? @Zalathar
Remove some unnecessary `try`-related type annotations
I left a few, like
```rust
let result: Result<_, ModError<'_>> = try {
```
where it felt like seeing it might still be useful for the reader.
Feel free to push back on any of these changes if you think they should be left alone.
GVN: Only propagate borrows from SSA locals
Fixes https://github.com/rust-lang/rust/issues/141313. This is a more principled fix than https://github.com/rust-lang/rust/pull/147886.
Using a reference that is not a borrowing of an SSA local at a new location may be UB.
The PR has two major changes.
The first one, when introducing a new dereference at a new location, is that the reference must point to an SSA local or be an immutable argument. `dereference_address` has handled SSA locals.
The second one, if we cannot guard to the reference point to an SSA local in `visit_assign`, we have to rewrite the value to opaque. This avoids unifying the following dereferences that also are references:
```rust
let b: &T = *a;
// ... `a` is allowed to be modified. `c` and `b` have different borrowing lifetime.
// Unifying them will extend the lifetime of `b`.
let c: &T = *a;
```
See also https://github.com/rust-lang/rust/issues/130853.
This still allows unifying non-reference dereferences:
```rust
let a: &T = ...;
let b: T = *a;
// ... a is NOT allowed to be modified.
let c: T = *a;
```
r? @cjgillot
Introduce helper `ty::Generics::has_own_self`
The pattern `generics.has_self && generics.parent.is_none()` only occurs 5 times in rustc+rustdoc at the time of writing but I keep getting reminded/annoyed that there doesn't exist a nice wrapper fn that abstracts it & immediately clarifies the intent. Most recently that happened when working on my open PR RUST-129543 in which I add yet another occurrence of it ([via](ae8c0a5a46/compiler/rustc_hir_analysis/src/collect/resolve_bound_vars.rs (L1771-L1773))).
For context, `generics.has_self` indicates that there is a `Self` type parameter (at position 0) in the "oldest" / "root" generic parameter list in the chain of generic parameter lists (rephrased: it's true if the chain *as a whole* has a `Self` type parameter). `has_own_self` on the other hand indicates that the `own_params` contain a `Self` type parameter which is sometimes needed for offsetting the own parameters or arguments.
Move `rustc_query_system::cache`.
It only defines two types, `Cache` and `WithDepNode`. Neither has anything much to do with queries -- they use `DepNodeIndex`, that's all.
This commit moves the module into `rustc_middle`, to where they are used. It also renames the extremely non-descriptive `Cache` as `WithDepNodeCache`.
r? @Zalathar
Fix ICE in normalizing inherent associated consts with `#[type_const]`
Fixesrust-lang/rust#151027Fixesrust-lang/rust#138089Fixesrust-lang/rust#138226Fixesrust-lang/rust#150960
When an inherent associated const is marked with `#[type_const]`, its generics expect args in the format `[Self, own_params...]`, similar to inherent associated types. However, HIR typeck's `instantiate_value_path` was constructing args in the regular associated const format `[impl_params..., own_params...]`. This mismatch caused ICEs when the `args` were later used in contexts expecting the IAC format, such as user type annotations and `borrowck`'s type ascription.
Consolidate type const checks on `tcx.is_type_const`
A little bit of cleanup; explanation can be found in the reporting issue.
Fixesrust-lang/rust#152124
r? BoxyUwU
Provide more context on trait bounds being unmet due to imperfect derive
When encountering a value that has a borrow checker error where the type was previously moved, when suggesting cloning verify that it is not already being derived. If it is, explain why the `derive(Clone)` doesn't apply:
```
note: if `TypedAddress<T>` implemented `Clone`, you could clone the value
--> $DIR/derive-clone-implicit-bound.rs:6:1
|
LL | #[derive(Clone, Copy)]
| ----- derived `Clone` adds implicit bounds on type parameters
LL | pub struct TypedAddress<T>{
| ^^^^^^^^^^^^^^^^^^^^^^^^-^
| | |
| | introduces an implicit `T: Clone` bound
| consider manually implementing `Clone` for this type
...
LL | let old = self.return_value(offset);
| ------ you could clone this value
```
When encountering a bound coming from a derive macro, suggest manual impl of the trait.
Use the span for the specific param when adding bounds in builtin derive macros, so the diagnostic will point at them as well as the derive macro itself.
```
note: required for `Id<SomeNode>` to implement `PartialEq`
--> $DIR/derive-implicit-bound.rs:5:10
|
LL | #[derive(PartialEq, Eq)]
| ^^^^^^^^^
LL | pub struct Id<T>(PhantomData<T>);
| - unsatisfied trait bound introduced in this `derive` macro
= help: consider manually implementing `PartialEq` to avoid undesired bounds
```
Mention that the trait could be manually implemented in E0599.
Fixrust-lang/rust#108894. Address rust-lang/rust#143714. Address #rust-lang/rust#146515 (but ideally would also suggest constraining the fn bound correctly as well).
Incorporate query description functions into `QueryVTable`
Putting a `desc` function in each query vtable reduces the amount of parameter juggling required when creating query stack frames, because almost all of the necessary information can be found in the vtable.
There should be no change to compiler output.
Some `rustc_query_system` cleanups
Small improvements I found while looking closely at `rustc_query_system`. Best reviewed one commit at a time.
r? @cjgillot
Rename trait `DepNodeParams` to `DepNodeKey`
In query system plumbing, we usually refer to a query's explicit argument value as a “key”.
The first few commits do some preliminary cleanup that would conflict with the rename; the rename itself is in the final commit.
r? nnethercote (or compiler)
It's a tiny module with one trait and a default impl. It's not used in
`rustc_query_system`; all uses and non-default impls are in
`rustc_middle` and `rustc_query_impl`.
This commit moves it into `rustc_middle`, which makes things simpler
overall.
