Stashed errors used to be counted as errors, but could then be
cancelled, leading to `ErrorGuaranteed` soundness holes. #120828 changed
that, closing the soundness hole. But it introduced other difficulties
because you sometimes have to account for pending stashed errors when
making decisions about whether errors have occured/will occur and it's
easy to overlook these.
This commit aims for a middle ground.
- Stashed errors (not warnings) are counted immediately as emitted
errors, avoiding the possibility of forgetting to consider them.
- The ability to cancel (or downgrade) stashed errors is eliminated, by
disallowing the use of `steal_diagnostic` with errors, and introducing
the more restrictive methods `try_steal_{modify,replace}_and_emit_err`
that can be used instead.
Other things:
- `DiagnosticBuilder::stash` and `DiagCtxt::stash_diagnostic` now both
return `Option<ErrorGuaranteed>`, which enables the removal of two
`delayed_bug` calls and one `Ty::new_error_with_message` call. This is
possible because we store error guarantees in
`DiagCtxt::stashed_diagnostics`.
- Storing the guarantees also saves us having to maintain a counter.
- Calls to the `stashed_err_count` method are no longer necessary
alongside calls to `has_errors`, which is a nice simplification, and
eliminates two more `span_delayed_bug` calls and one FIXME comment.
- Tests are added for three of the four fixed PRs mentioned below.
- `issue-121108.rs`'s output improved slightly, omitting a non-useful
error message.
Fixes#121451.
Fixes#121477.
Fixes#121504.
Fixes#121508.
Opportunistically resolve regions when processing region outlives obligations
Due to the matching in `TypeOutlives` being structural, we should attempt to opportunistically resolve regions before processing region obligations. Thanks ``@lcnr`` for finding this.
r? lcnr
Add `StructurallyRelateAliases` to allow instantiating infer vars with rigid aliases.
Change `instantiate_query_response` to be infallible in the new solver. This requires canonicalization to not hide any information used by the query, so weaken
universe compression. It also modifies `term_is_fully_unconstrained` to allow
region inference variables in a higher universe.
Rollup of 7 pull requests
Successful merges:
- #121435 (Account for RPITIT in E0310 explicit lifetime constraint suggestion)
- #121490 (Rustdoc: include crate name in links for local primitives)
- #121520 (delay cloning of iterator items)
- #121522 (check that simd_insert/extract indices are in-bounds)
- #121531 (Ignore less tests in debug builds)
- #121539 (compiler/rustc_target/src/spec/base/apple/tests.rs: Avoid unnecessary large move)
- #121542 (update stdarch)
r? `@ghost`
`@rustbot` modify labels: rollup
Account for RPITIT in E0310 explicit lifetime constraint suggestion
When given
```rust
trait Original {
fn f() -> impl Fn();
}
trait Erased {
fn f(&self) -> Box<dyn Fn()>;
}
impl<T: Original> Erased for T {
fn f(&self) -> Box<dyn Fn()> {
Box::new(<T as Original>::f())
}
}
```
emit do not emit an invalid suggestion restricting the `Trait::{opaque}` type in a `where` clause:
```
error[E0310]: the associated type `<T as Original>::{opaque#0}` may not live long enough
--> $DIR/missing-static-bound-from-impl.rs:11:9
|
LL | Box::new(<T as Original>::f())
| ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
| |
| the associated type `<T as Original>::{opaque#0}` must be valid for the static lifetime...
| ...so that the type `impl Fn()` will meet its required lifetime bounds
```
Partially address #119773. Ideally we'd suggest modifying `Erased::f` instead.
r? `@compiler-errors`
region unification: update universe of region vars
necessary for #119106. see inline comment for why this is necessary
r? `@compiler-errors` `@BoxyUwU`
fix generalizer unsoundness
I ended up getting confused while trying to flip the variances when flipping the order. Should be all right now.
This is only exploitable when generalizing if the `ambient_variance` of the relation is `Contravariant`. This can currently only be the case in the NLL generalizer which only rarely generalizes, causing us to miss this regression. Very much an issue with #121462 however.
When given
```rust
trait Original {
fn f() -> impl Fn();
}
trait Erased {
fn f(&self) -> Box<dyn Fn()>;
}
impl<T: Original> Erased for T {
fn f(&self) -> Box<dyn Fn()> {
Box::new(<T as Original>::f())
}
}
```
avoid suggestion to restrict the `Trait::{opaque}` type in a `where` clause:
```
error[E0310]: the associated type `<T as Original>::{opaque#0}` may not live long enough
--> $DIR/missing-static-bound-from-impl.rs:11:9
|
LL | Box::new(<T as Original>::f())
| ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
| |
| the associated type `<T as Original>::{opaque#0}` must be valid for the static lifetime...
| ...so that the type `impl Fn()` will meet its required lifetime bounds
```
CC #119773.
Without doing so we use the same candidate cache entry
for `?0: Trait<?1>` and `?0: Trait<?0>`. These goals are different
and we must not use the same entry for them.
we don't track them when canonicalizing or when freshening,
resulting in instable caching in the old solver, and issues when
instantiating query responses in the new one.
Top level error handling
The interactions between the following things are surprisingly complicated:
- `emit_stashed_diagnostics`,
- `flush_delayed`,
- normal return vs `abort_if_errors`/`FatalError.raise()` unwinding in the call to the closure in `interface::run_compiler`.
This PR disentangles it all.
r? `@oli-obk`
Currently `has_errors` excludes lint errors. This commit changes it to
include lint errors.
The motivation for this is that for most places it doesn't matter
whether lint errors are included or not. But there are multiple places
where they must be includes, and only one place where they must not be
included. So it makes sense for `has_errors` to do the thing that fits
the most situations, and the new `has_errors_excluding_lint_errors`
method in the one exceptional place.
The same change is made for `err_count`. Annoyingly, this requires the
introduction of `err_count_excluding_lint_errs` for one place, to
preserve existing error printing behaviour. But I still think the change
is worthwhile overall.
Make --verbose imply -Z write-long-types-to-disk=no
When shortening the type it is necessary to take into account the `--verbose` flag, if it is activated, we must always show the entire type and not write it in a file.
Fixes: https://github.com/rust-lang/rust/issues/119130
Convert `delayed_bug`s to `bug`s.
I have a suspicion that quite a few delayed bug paths are impossible to reach, so I did an experiment.
I converted every `delayed_bug` to a `bug`, ran the full test suite, then converted back every `bug` that was hit. A surprising number were never hit.
This is too dangerous to merge. Increased coverage (fuzzing or a crater run) would likely hit more cases. But it might be useful for people to look at and think about which paths are genuinely unreachable.
r? `@ghost`
I have a suspicion that quite a few delayed bug paths are impossible to
reach, so I did an experiment.
I converted every `delayed_bug` to a `bug`, ran the full test suite,
then converted back every `bug` that was hit. A surprising number were
never hit.
The next commit will convert some more back, based on human judgment.
Overhaul `Diagnostic` and `DiagnosticBuilder`
Implements the first part of https://github.com/rust-lang/compiler-team/issues/722, which moves functionality and use away from `Diagnostic`, onto `DiagnosticBuilder`.
Likely follow-ups:
- Move things around, because this PR was written to minimize diff size, so some things end up in sub-optimal places. E.g. `DiagnosticBuilder` has impls in both `diagnostic.rs` and `diagnostic_builder.rs`.
- Rename `Diagnostic` as `DiagInner` and `DiagnosticBuilder` as `Diag`.
r? `@davidtwco`
return `ty::Error` when equating `ty::Error`
This helps iron out a difference in diagnostics between `Sub` and `Equate` relations, which I'm currently trying to unify.
r? oli-obk
Drive-by `DUMMY_SP` -> `Span` and fmt changes
Noticed these while doing something else. There's no practical change, but it's preferable to use `DUMMY_SP` as little as possible, particularly when we have perfectlly useful `Span`s available.
Currently many diagnostic modifier methods are available on both
`Diagnostic` and `DiagnosticBuilder`. This commit removes most of them
from `Diagnostic`. To minimize the diff size, it keeps them within
`diagnostic.rs` but changes the surrounding `impl Diagnostic` block to
`impl DiagnosticBuilder`. (I intend to move things around later, to give
a more sensible code layout.)
`Diagnostic` keeps a few methods that it still needs, like `sub`,
`arg`, and `replace_args`.
The `forward!` macro, which defined two additional methods per call
(e.g. `note` and `with_note`), is replaced by the `with_fn!` macro,
which defines one additional method per call (e.g. `with_note`). It's
now also only used when necessary -- not all modifier methods currently
need a `with_*` form. (New ones can be easily added as necessary.)
All this also requires changing `trait AddToDiagnostic` so its methods
take `DiagnosticBuilder` instead of `Diagnostic`, which leads to many
mechanical changes. `SubdiagnosticMessageOp` gains a type parameter `G`.
There are three subdiagnostics -- `DelayedAtWithoutNewline`,
`DelayedAtWithNewline`, and `InvalidFlushedDelayedDiagnosticLevel` --
that are created within the diagnostics machinery and appended to
external diagnostics. These are handled at the `Diagnostic` level, which
means it's now hard to construct them via `derive(Diagnostic)`, so
instead we construct them by hand. This has no effect on what they look
like when printed.
There are lots of new `allow` markers for `untranslatable_diagnostics`
and `diagnostics_outside_of_impl`. This is because
`#[rustc_lint_diagnostics]` annotations were present on the `Diagnostic`
modifier methods, but missing from the `DiagnosticBuilder` modifier
methods. They're now present.
deduplicate infer var instantiation
Having 3 separate implementations of one of the most subtle parts of our type system is not a good strategy if we want to maintain a sound type system ✨ while working on this I already found some subtle bugs in the existing code, so that's awesome 🎉 cc #121159
This was necessary as I am not confident in my nll changes in #119106, so I am first cleaning this up in a separate PR.
r? `@BoxyUwU`
Noticed these while doing something else. There's no practical change, but it's preferable to use `DUMMY_SP` as little as possible, particularly when we have perfectlly useful `Span`s available.
There are lots of functions that modify a diagnostic. This can be via a
`&mut Diagnostic` or a `&mut DiagnosticBuilder`, because the latter type
wraps the former and impls `DerefMut`.
This commit converts all the `&mut Diagnostic` occurrences to `&mut
DiagnosticBuilder`. This is a step towards greatly simplifying
`Diagnostic`. Some of the relevant function are made generic, because
they deal with both errors and warnings. No function bodies are changed,
because all the modifier methods are available on both `Diagnostic` and
`DiagnosticBuilder`.
fixes#117448
For example unnecessary imports in std::prelude that can be eliminated:
```rust
use std::option::Option::Some;//~ WARNING the item `Some` is imported redundantly
use std::option::Option::None; //~ WARNING the item `None` is imported redundantly
```
errors: only eagerly translate subdiagnostics
Subdiagnostics don't need to be lazily translated, they can always be eagerly translated. Eager translation is slightly more complex as we need to have a `DiagCtxt` available to perform the translation, which involves slightly more threading of that context.
This slight increase in complexity should enable later simplifications - like passing `DiagCtxt` into `AddToDiagnostic` and moving Fluent messages into the diagnostic structs rather than having them in separate files (working on that was what led to this change).
r? ```@nnethercote```
we already use `instantiate_const_var`. This does lose some debugging
info for nll because we stop populating the `reg_var_to_origin` table with
`RegionCtxt::Existential(None)`, I don't think that matters however.
Supporting this adds additional complexity to one of the most involved
parts of the type system, so I really don't think it's worth it.
Implement intrinsics with fallback bodies
fixes#93145 (though we can port many more intrinsics)
cc #63585
The way this works is that the backend logic for generating custom code for intrinsics has been made fallible. The only failure path is "this intrinsic is unknown". The `Instance` (that was `InstanceDef::Intrinsic`) then gets converted to `InstanceDef::Item`, which represents the fallback body. A regular function call to that body is then codegenned. This is currently implemented for
* codegen_ssa (so llvm and gcc)
* codegen_cranelift
other backends will need to adjust, but they can just keep doing what they were doing if they prefer (though adding new intrinsics to the compiler will then require them to implement them, instead of getting the fallback body).
cc `@scottmcm` `@WaffleLapkin`
### todo
* [ ] miri support
* [x] default intrinsic name to name of function instead of requiring it to be specified in attribute
* [x] make sure that the bodies are always available (must be collected for metadata)
Subdiagnostics don't need to be lazily translated, they can always be
eagerly translated. Eager translation is slightly more complex as we need
to have a `DiagCtxt` available to perform the translation, which involves
slightly more threading of that context.
This slight increase in complexity should enable later simplifications -
like passing `DiagCtxt` into `AddToDiagnostic` and moving Fluent messages
into the diagnostic structs rather than having them in separate files
(working on that was what led to this change).
Signed-off-by: David Wood <david@davidtw.co>
For some cases where it's clear that an error has already occurred,
e.g.:
- there's a comment stating exactly that, or
- things like HIR lowering, where we are lowering an error kind
The commit also tweaks some comments around delayed bug sites.
`cargo update`
Run `cargo update`, with some pinning and fixes necessitated by that. This *should* unblock #112865
There's a couple of places where I only pinned a dependency in one location - this seems like a bit of a hack, but better than duplicating the FIXME across all `Cargo.toml` where a dependency is introduced.
cc `@Nilstrieb`
Dejargonize `subst`
In favor of #110793, replace almost every occurence of `subst` and `substitution` from rustc codes, but they still remains in subtrees under `src/tools/` like clippy and test codes (I'd like to replace them after this)
The check within changed from `delay_span_bug` to `delay_good_path_bug`
in #110476, and removing the check altogether was considered. It's a
very weak sanity check and gets in the way of removing good path delayed
bugs altogether, so this PR just removes it.
Allow restricted trait impls under `#[allow_internal_unstable(min_specialization)]`
This is a follow-up to #119963 and a companion to #120866, though it can land independently from the latter.
---
We have several compiler crates that only enable `#[feature(min_specialization)]` because it is required by their expansions of `newtype_index!`, in order to implement traits marked with `#[rustc_specialization_trait]`.
This PR allows those traits to be implemented internally by macros with `#[allow_internal_unstable(min_specialization)]`, without needing specialization to be enabled in the enclosing crate.
Harmonize `AsyncFn` implementations, make async closures conditionally impl `Fn*` traits
This PR implements several changes to the built-in and libcore-provided implementations of `Fn*` and `AsyncFn*` to address two problems:
1. async closures do not implement the `Fn*` family traits, leading to breakage: https://crater-reports.s3.amazonaws.com/pr-120361/index.html
2. *references* to async closures do not implement `AsyncFn*`, as a consequence of the existing blanket impls of the shape `AsyncFn for F where F: Fn, F::Output: Future`.
In order to fix (1.), we implement `Fn` traits appropriately for async closures. It turns out that async closures can:
* always implement `FnOnce`, meaning that they're drop-in compatible with `FnOnce`-bound combinators like `Option::map`.
* conditionally implement `Fn`/`FnMut` if they have no captures, which means that existing usages of async closures should *probably* work without breakage (crater checking this: https://github.com/rust-lang/rust/pull/120712#issuecomment-1930587805).
In order to fix (2.), we make all of the built-in callables implement `AsyncFn*` via built-in impls, and instead adjust the blanket impls for `AsyncFn*` provided by libcore to match the blanket impls for `Fn*`.
These crates all needed specialization for `newtype_index!`, which will no
longer be necessary when the current nightly eventually becomes the next
bootstrap compiler.
Fix `ErrorGuaranteed` unsoundness with stash/steal.
When you stash an error, the error count is incremented. You can then use the non-zero error count to get an `ErrorGuaranteed`. You can then steal the error, which decrements the error count. You can then cancel the error.
Example code:
```
fn unsound(dcx: &DiagCtxt) -> ErrorGuaranteed {
let sp = rustc_span::DUMMY_SP;
let k = rustc_errors::StashKey::Cycle;
dcx.struct_err("bogus").stash(sp, k); // increment error count on stash
let guar = dcx.has_errors().unwrap(); // ErrorGuaranteed from error count > 0
let err = dcx.steal_diagnostic(sp, k).unwrap(); // decrement error count on steal
err.cancel(); // cancel error
guar // ErrorGuaranteed with no error emitted!
}
```
This commit fixes the problem in the simplest way: by not counting stashed errors in `DiagCtxt::{err_count,has_errors}`.
However, just doing this without any other changes leads to over 40 ui test failures. Mostly because of uninteresting extra errors (many saying "type annotations needed" when type inference fails), and in a few cases, due to delayed bugs causing ICEs when no normal errors are printed.
To fix these, this commit adds `DiagCtxt::stashed_err_count`, and uses it in three places alongside `DiagCtxt::{has_errors,err_count}`. It's dodgy to rely on it, because unlike `DiagCtxt::err_count` it can go up and down. But it's needed to preserve existing behaviour, and at least the three places that need it are now obvious.
r? oli-obk
Invert diagnostic lints.
That is, change `diagnostic_outside_of_impl` and `untranslatable_diagnostic` from `allow` to `deny`, because more than half of the compiler has been converted to use translated diagnostics.
This commit removes more `deny` attributes than it adds `allow` attributes, which proves that this change is warranted.
r? ````@davidtwco````
When you stash an error, the error count is incremented. You can then
use the non-zero error count to get an `ErrorGuaranteed`. You can then
steal the error, which decrements the error count. You can then cancel
the error.
Example code:
```
fn unsound(dcx: &DiagCtxt) -> ErrorGuaranteed {
let sp = rustc_span::DUMMY_SP;
let k = rustc_errors::StashKey::Cycle;
dcx.struct_err("bogus").stash(sp, k); // increment error count on stash
let guar = dcx.has_errors().unwrap(); // ErrorGuaranteed from error count > 0
let err = dcx.steal_diagnostic(sp, k).unwrap(); // decrement error count on steal
err.cancel(); // cancel error
guar // ErrorGuaranteed with no error emitted!
}
```
This commit fixes the problem in the simplest way: by not counting
stashed errors in `DiagCtxt::{err_count,has_errors}`.
However, just doing this without any other changes leads to over 40 ui
test failures. Mostly because of uninteresting extra errors (many saying
"type annotations needed" when type inference fails), and in a few
cases, due to delayed bugs causing ICEs when no normal errors are
printed.
To fix these, this commit adds `DiagCtxt::stashed_err_count`, and uses
it in three places alongside `DiagCtxt::{has_errors,err_count}`. It's
dodgy to rely on it, because unlike `DiagCtxt::err_count` it can go up
and down. But it's needed to preserve existing behaviour, and at least
the three places that need it are now obvious.
Remove unused args from functions
`#[instrument]` suppresses the unused arguments from a function, *and* suppresses unused methods too! This PR removes things which are only used via `#[instrument]` calls, and fixes some other errors (privacy?) that I will comment inline.
It's possible that some of these arguments were being passed in for the purposes of being instrumented, but I am unconvinced by most of them.
Introduce `enter_forall` to supercede `instantiate_binder_with_placeholders`
r? `@lcnr`
Long term we'd like to experiment with decrementing the universe count after "exiting" binders so that we do not end up creating infer vars in non-root universes even when they logically reside in the root universe. The fact that we dont do this currently results in a number of issues in the new trait solver where we consider goals to be ambiguous because otherwise it would require lowering the universe of an infer var. i.e. the goal `?x.0 eq <T as Trait<?y.1>>::Assoc` where the alias is rigid would not be able to instantiate `?x` with the alias as there would be a universe error.
This PR is the first-ish sort of step towards being able to implement this as eventually we would want to decrement the universe in `enter_forall`. Unfortunately its Difficult to actually implement decrementing universes nicely so this is a separate step which moves us closer to the long term goal ✨
Normalize type outlives obligations in NLL for new solver
Normalize the type outlives assumptions and obligations in MIR borrowck. This should fix any of the lazy-norm-related MIR borrowck problems.
Also some cleanups from last PR:
1. Normalize obligations in a loop in lexical region resolution
2. Use `deeply_normalize_with_skipped_universes` in lexical resolution since we may have, e.g. `for<'a> Alias<'a>: 'b`.
r? lcnr
That is, change `diagnostic_outside_of_impl` and
`untranslatable_diagnostic` from `allow` to `deny`, because more than
half of the compiler has be converted to use translated diagnostics.
This commit removes more `deny` attributes than it adds `allow`
attributes, which proves that this change is warranted.
- `emitted_at` isn't used outside the crate.
- `code` and `messages` are public fields, so there's no point have
trivial getters/setters for them.
- `suggestions` is public, so the comment about "functionality on
`Diagnostic`" isn't needed.
Normalize region obligation in lexical region resolution with next-gen solver
This normalizes region obligations when we `resolve_regions`, since they may be unnormalized with deferred projection equality.
It's pretty hard to add tests that exercise this without also triggering MIR borrowck errors (because we don't normalize there yet). I've added one test with two revisions that should test that we both 1. normalize region obligations in the param env, and 2. normalize registered region obligations during lexical region resolution.
Remove various `has_errors` or `err_count` uses
follow up to https://github.com/rust-lang/rust/pull/119895
r? `@nnethercote` since you recently did something similar.
There are so many more of these, but I wanted to get a PR out instead of growing the commit list indefinitely. The commits all work on their own and can be reviewed commit by commit.
Because it's almost always static.
This makes `impl IntoDiagnosticArg for DiagnosticArgValue` trivial,
which is nice.
There are a few diagnostics constructed in
`compiler/rustc_mir_build/src/check_unsafety.rs` and
`compiler/rustc_mir_transform/src/errors.rs` that now need symbols
converted to `String` with `to_string` instead of `&str` with `as_str`,
but that' no big deal, and worth it for the simplifications elsewhere.
Error codes are integers, but `String` is used everywhere to represent
them. Gross!
This commit introduces `ErrCode`, an integral newtype for error codes,
replacing `String`. It also introduces a constant for every error code,
e.g. `E0123`, and removes the `error_code!` macro. The constants are
imported wherever used with `use rustc_errors::codes::*`.
With the old code, we have three different ways to specify an error code
at a use point:
```
error_code!(E0123) // macro call
struct_span_code_err!(dcx, span, E0123, "msg"); // bare ident arg to macro call
\#[diag(name, code = "E0123")] // string
struct Diag;
```
With the new code, they all use the `E0123` constant.
```
E0123 // constant
struct_span_code_err!(dcx, span, E0123, "msg"); // constant
\#[diag(name, code = E0123)] // constant
struct Diag;
```
The commit also changes the structure of the error code definitions:
- `rustc_error_codes` now just defines a higher-order macro listing the
used error codes and nothing else.
- Because that's now the only thing in the `rustc_error_codes` crate, I
moved it into the `lib.rs` file and removed the `error_codes.rs` file.
- `rustc_errors` uses that macro to define everything, e.g. the error
code constants and the `DIAGNOSTIC_TABLES`. This is in its new
`codes.rs` file.
When encountering a type mismatch error involving `dyn Trait`, mention
the existence of boxed trait objects if the other type involved
implements `Trait`.
Partially addresses #102629.
Provide structured suggestion to use trait objects in some cases of `if` arm type divergence
```
error[E0308]: `if` and `else` have incompatible types
--> $DIR/suggest-box-on-divergent-if-else-arms.rs:15:9
|
LL | let _ = if true {
| _____________-
LL | | Struct
| | ------ expected because of this
LL | | } else {
LL | | foo()
| | ^^^^^ expected `Struct`, found `Box<dyn Trait>`
LL | | };
| |_____- `if` and `else` have incompatible types
|
= note: expected struct `Struct`
found struct `Box<dyn Trait>`
help: `Struct` implements `Trait` so you can box it to coerce to the trait object `Box<dyn Trait>`
|
LL | Box::new(Struct)
| +++++++++ +
error[E0308]: `if` and `else` have incompatible types
--> $DIR/suggest-box-on-divergent-if-else-arms.rs:20:9
|
LL | let _ = if true {
| _____________-
LL | | foo()
| | ----- expected because of this
LL | | } else {
LL | | Struct
| | ^^^^^^ expected `Box<dyn Trait>`, found `Struct`
LL | | };
| |_____- `if` and `else` have incompatible types
|
= note: expected struct `Box<dyn Trait>`
found struct `Struct`
= note: for more on the distinction between the stack and the heap, read https://doc.rust-lang.org/book/ch15-01-box.html, https://doc.rust-lang.org/rust-by-example/std/box.html, and https://doc.rust-lang.org/std/boxed/index.html
help: store this in the heap by calling `Box::new`
|
LL | Box::new(Struct)
| +++++++++ +
error[E0308]: `if` and `else` have incompatible types
--> $DIR/suggest-box-on-divergent-if-else-arms.rs:25:9
|
LL | fn bar() -> impl Trait {
| ---------- the found opaque type
...
LL | let _ = if true {
| _____________-
LL | | Struct
| | ------ expected because of this
LL | | } else {
LL | | bar()
| | ^^^^^ expected `Struct`, found opaque type
LL | | };
| |_____- `if` and `else` have incompatible types
|
= note: expected struct `Struct`
found opaque type `impl Trait`
help: `Struct` implements `Trait` so you can box both arms and coerce to the trait object `Box<dyn Trait>`
|
LL ~ Box::new(Struct) as Box<dyn Trait>
LL | } else {
LL ~ Box::new(bar())
|
error[E0308]: `if` and `else` have incompatible types
--> $DIR/suggest-box-on-divergent-if-else-arms.rs:30:9
|
LL | fn bar() -> impl Trait {
| ---------- the expected opaque type
...
LL | let _ = if true {
| _____________-
LL | | bar()
| | ----- expected because of this
LL | | } else {
LL | | Struct
| | ^^^^^^ expected opaque type, found `Struct`
LL | | };
| |_____- `if` and `else` have incompatible types
|
= note: expected opaque type `impl Trait`
found struct `Struct`
help: `Struct` implements `Trait` so you can box both arms and coerce to the trait object `Box<dyn Trait>`
|
LL ~ Box::new(bar()) as Box<dyn Trait>
LL | } else {
LL ~ Box::new(Struct)
|
```
Partially address #102629.
When encountering
```rust
let _ = if true {
Struct
} else {
foo() // -> Box<dyn Trait>
};
```
if `Struct` implements `Trait`, suggest boxing the then arm tail expression.
Part of #102629.
We have several methods indicating the presence of errors, lint errors,
and delayed bugs. I find it frustrating that it's very unclear which one
you should use in any particular spot. This commit attempts to instill a
basic principle of "use the least general one possible", because that
reflects reality in practice -- `has_errors` is the least general one
and has by far the most uses (esp. via `abort_if_errors`).
Specifics:
- Add some comments giving some usage guidelines.
- Prefer `has_errors` to comparing `err_count` to zero.
- Remove `has_errors_or_span_delayed_bugs` because it's a weird one: in
the cases where we need to count delayed bugs, we should really be
counting lint errors as well.
- Rename `is_compilation_going_to_fail` as
`has_errors_or_lint_errors_or_span_delayed_bugs`, for consistency with
`has_errors` and `has_errors_or_lint_errors`.
- Change a few other `has_errors_or_lint_errors` calls to `has_errors`,
as per the "least general" principle.
This didn't turn out to be as neat as I hoped when I started, but I
think it's still an improvement.
Expose Obligations created during type inference.
This PR is a first pass at exposing the trait obligations generated and solved for during the type-check progress. Exposing these obligations allows for rustc plugins to use the public interface for proof trees (provided by the next gen trait solver).
The changes proposed track *all* obligations during the type-check process, this is desirable to not only look at the trees of failed obligations, but also those of successfully proved obligations. This feature is placed behind an unstable compiler option `track-trait-obligations` which should be used together with the `next-solver` option. I should note that the main interface is the function `inspect_typeck` made public in `rustc_hir_typeck/src/lib.rs` which allows the caller to provide a callback granting access to the `FnCtxt`.
r? `@lcnr`
Pass each obligation to an fn callback with its respective inference context. This avoids needing to keep around copies of obligations or inference contexts.
Specify usability of inspect_typeck in comment.
Rework how diagnostic lints are stored.
`Diagnostic::code` has the type `DiagnosticId`, which has `Error` and
`Lint` variants. Plus `Diagnostic::is_lint` is a bool, which should be
redundant w.r.t. `Diagnostic::code`.
Seems simple. Except it's possible for a lint to have an error code, in
which case its `code` field is recorded as `Error`, and `is_lint` is
required to indicate that it's a lint. This is what happens with
`derive(LintDiagnostic)` lints. Which means those lints don't have a
lint name or a `has_future_breakage` field because those are stored in
the `DiagnosticId::Lint`.
It's all a bit messy and confused and seems unintentional.
This commit:
- removes `DiagnosticId`;
- changes `Diagnostic::code` to `Option<String>`, which means both
errors and lints can straightforwardly have an error code;
- changes `Diagnostic::is_lint` to `Option<IsLint>`, where `IsLint` is a
new type containing a lint name and a `has_future_breakage` bool, so
all lints can have those, error code or not.
r? `@oli-obk`
Make sure to instantiate placeholders correctly in old solver
When creating the query substitution guess for an input placeholder type like `!1_T` (in universe 1), we were guessing the response substitution with something like `!0_T`. This failed to unify with `!1_T`, causing an ICE.
This PR reworks the query substitution guess code to work a bit more like the new solver. I'm *pretty* sure this is correct, though I'd really appreciate some scrutiny from someone (*cough* lcnr) who knows a bit more about query instantiation :)
Fixes#119941
r? lcnr
`OutputTypeParameterMismatch` -> `SignatureMismatch`
I'm probably missing something that made this rename more complicated. What did you end up getting stuck on when renaming this selection error, `@lcnr?`
**also** I renamed the `FulfillmentErrorCode` variants. This is just churn but I wanted to do it forever. I can move it out of this PR if desired.
r? lcnr
Silence some follow-up errors [3/x]
this is one piece of the requested cleanups from https://github.com/rust-lang/rust/pull/117449
Keep error types around, even in obligations.
These help silence follow-up errors, as we now figure out that some types (most notably inference variables) are equal to an error type.
But it also allows figuring out more types in the presence of errors, possibly causing more errors.
`Diagnostic::code` has the type `DiagnosticId`, which has `Error` and
`Lint` variants. Plus `Diagnostic::is_lint` is a bool, which should be
redundant w.r.t. `Diagnostic::code`.
Seems simple. Except it's possible for a lint to have an error code, in
which case its `code` field is recorded as `Error`, and `is_lint` is
required to indicate that it's a lint. This is what happens with
`derive(LintDiagnostic)` lints. Which means those lints don't have a
lint name or a `has_future_breakage` field because those are stored in
the `DiagnosticId::Lint`.
It's all a bit messy and confused and seems unintentional.
This commit:
- removes `DiagnosticId`;
- changes `Diagnostic::code` to `Option<String>`, which means both
errors and lints can straightforwardly have an error code;
- changes `Diagnostic::is_lint` to `Option<IsLint>`, where `IsLint` is a
new type containing a lint name and a `has_future_breakage` bool, so
all lints can have those, error code or not.
In #119606 I added them and used a `_mv` suffix, but that wasn't great.
A `with_` prefix has three different existing uses.
- Constructors, e.g. `Vec::with_capacity`.
- Wrappers that provide an environment to execute some code, e.g.
`with_session_globals`.
- Consuming chaining methods, e.g. `Span::with_{lo,hi,ctxt}`.
The third case is exactly what we want, so this commit changes
`DiagnosticBuilder::foo_mv` to `DiagnosticBuilder::with_foo`.
Thanks to @compiler-errors for the suggestion.
We have `span_delayed_bug` and often pass it a `DUMMY_SP`. This commit
adds `delayed_bug`, which matches pairs like `err`/`span_err` and
`warn`/`span_warn`.
Because it takes an error code after the span. This avoids the confusing
overlap with the `DiagCtxt::struct_span_err` method, which doesn't take
an error code.
unify query canonicalization mode
Exclude from canonicalization only the static lifetimes that appear in the param env because of #118965 . Any other occurrence can be canonicalized safely AFAICT.
r? `@lcnr`
The existing uses are replaced in one of three ways.
- In a function that also has calls to `emit`, just rearrange the code
so that exactly one of `delay_as_bug` or `emit` is called on every
path.
- In a function returning a `DiagnosticBuilder`, use
`downgrade_to_delayed_bug`. That's good enough because it will get
emitted later anyway.
- In `unclosed_delim_err`, one set of errors is being replaced with
another set, so just cancel the original errors.
This works for most of its call sites. This is nice, because `emit` very
much makes sense as a consuming operation -- indeed,
`DiagnosticBuilderState` exists to ensure no diagnostic is emitted
twice, but it uses runtime checks.
For the small number of call sites where a consuming emit doesn't work,
the commit adds `DiagnosticBuilder::emit_without_consuming`. (This will
be removed in subsequent commits.)
Likewise, `emit_unless` becomes consuming. And `delay_as_bug` becomes
consuming, while `delay_as_bug_without_consuming` is added (which will
also be removed in subsequent commits.)
All this requires significant changes to `DiagnosticBuilder`'s chaining
methods. Currently `DiagnosticBuilder` method chaining uses a
non-consuming `&mut self -> &mut Self` style, which allows chaining to
be used when the chain ends in `emit()`, like so:
```
struct_err(msg).span(span).emit();
```
But it doesn't work when producing a `DiagnosticBuilder` value,
requiring this:
```
let mut err = self.struct_err(msg);
err.span(span);
err
```
This style of chaining won't work with consuming `emit` though. For
that, we need to use to a `self -> Self` style. That also would allow
`DiagnosticBuilder` production to be chained, e.g.:
```
self.struct_err(msg).span(span)
```
However, removing the `&mut self -> &mut Self` style would require that
individual modifications of a `DiagnosticBuilder` go from this:
```
err.span(span);
```
to this:
```
err = err.span(span);
```
There are *many* such places. I have a high tolerance for tedious
refactorings, but even I gave up after a long time trying to convert
them all.
Instead, this commit has it both ways: the existing `&mut self -> Self`
chaining methods are kept, and new `self -> Self` chaining methods are
added, all of which have a `_mv` suffix (short for "move"). Changes to
the existing `forward!` macro lets this happen with very little
additional boilerplate code. I chose to add the suffix to the new
chaining methods rather than the existing ones, because the number of
changes required is much smaller that way.
This doubled chainging is a bit clumsy, but I think it is worthwhile
because it allows a *lot* of good things to subsequently happen. In this
commit, there are many `mut` qualifiers removed in places where
diagnostics are emitted without being modified. In subsequent commits:
- chaining can be used more, making the code more concise;
- more use of chaining also permits the removal of redundant diagnostic
APIs like `struct_err_with_code`, which can be replaced easily with
`struct_err` + `code_mv`;
- `emit_without_diagnostic` can be removed, which simplifies a lot of
machinery, removing the need for `DiagnosticBuilderState`.
Use `resolutions(()).effective_visiblities` to avoid cycle errors in `report_object_error`
Inside of `report_object_error`, using the `effective_visibilities` query causes cycles since it calls `type_of`, which itself may call `typeck`, which may end up reporting its own object-safety errors.
Fixes#119346Fixes#119502
Tweak suggestions for bare trait used as a type
```
error[E0782]: trait objects must include the `dyn` keyword
--> $DIR/not-on-bare-trait-2021.rs:11:11
|
LL | fn bar(x: Foo) -> Foo {
| ^^^
|
help: use a generic type parameter, constrained by the trait `Foo`
|
LL | fn bar<T: Foo>(x: T) -> Foo {
| ++++++++ ~
help: you can also use `impl Foo`, but users won't be able to specify the type paramer when calling the `fn`, having to rely exclusively on type inference
|
LL | fn bar(x: impl Foo) -> Foo {
| ++++
help: alternatively, use a trait object to accept any type that implements `Foo`, accessing its methods at runtime using dynamic dispatch
|
LL | fn bar(x: &dyn Foo) -> Foo {
| ++++
error[E0782]: trait objects must include the `dyn` keyword
--> $DIR/not-on-bare-trait-2021.rs:11:19
|
LL | fn bar(x: Foo) -> Foo {
| ^^^
|
help: use `impl Foo` to return an opaque type, as long as you return a single underlying type
|
LL | fn bar(x: Foo) -> impl Foo {
| ++++
help: alternatively, you can return an owned trait object
|
LL | fn bar(x: Foo) -> Box<dyn Foo> {
| +++++++ +
```
Fix#119525:
```
error[E0038]: the trait `Ord` cannot be made into an object
--> $DIR/bare-trait-dont-suggest-dyn.rs:3:33
|
LL | fn ord_prefer_dot(s: String) -> Ord {
| ^^^ `Ord` cannot be made into an object
|
note: for a trait to be "object safe" it needs to allow building a vtable to allow the call to be resolvable dynamically; for more information visit <https://doc.rust-lang.org/reference/items/traits.html#object-safety>
--> $SRC_DIR/core/src/cmp.rs:LL:COL
|
= note: the trait cannot be made into an object because it uses `Self` as a type parameter
::: $SRC_DIR/core/src/cmp.rs:LL:COL
|
= note: the trait cannot be made into an object because it uses `Self` as a type parameter
help: consider using an opaque type instead
|
LL | fn ord_prefer_dot(s: String) -> impl Ord {
| ++++
```
`Diagnostic` has 40 methods that return `&mut Self` and could be
considered setters. Four of them have a `set_` prefix. This doesn't seem
necessary for a type that implements the builder pattern. This commit
removes the `set_` prefixes on those four methods.
Implement constant propagation on top of MIR SSA analysis
This implements the idea I proposed in https://github.com/rust-lang/rust/pull/110719#issuecomment-1718324700
Based on https://github.com/rust-lang/rust/pull/109597
The value numbering "GVN" pass formulates each rvalue that appears in MIR with an abstract form (the `Value` enum), and assigns an integer `VnIndex` to each. This abstract form can be used to deduplicate values, reusing an earlier local that holds the same value instead of recomputing. This part is proposed in #109597.
From this abstract representation, we can perform more involved simplifications, for example in https://github.com/rust-lang/rust/pull/111344.
With the abstract representation `Value`, we can also attempt to evaluate each to a constant using the interpreter. This builds a `VnIndex -> OpTy` map. From this map, we can opportunistically replace an operand or a rvalue with a constant if their value has an associated `OpTy`.
The most relevant commit is [Evaluated computed values to constants.](2767c4912e)"
r? `@oli-obk`
rework `-Zverbose`
implements the changes described in https://github.com/rust-lang/compiler-team/issues/706
the first commit is only a name change from `-Zverbose` to `-Zverbose-internals` and does not change behavior. the second commit changes diagnostics.
possible follow up work:
- `ty::pretty` could print more info with `--verbose` than it does currently. `-Z verbose-internals` shows too much info in a way that's not helpful to users. michael had ideas about this i didn't fully understand: https://rust-lang.zulipchat.com/#narrow/stream/233931-t-compiler.2Fmajor-changes/topic/uplift.20some.20-Zverbose.20calls.20and.20rename.20to.E2.80.A6.20compiler-team.23706/near/408984200
- `--verbose` should imply `-Z write-long-types-to-disk=no`. the code in `ty_string_with_limit` should take `--verbose` into account (apparently this affects `Ty::sort_string`, i'm not familiar with this code). writing a file to disk should suggest passing `--verbose`.
r? `@compiler-errors` cc `@estebank`
`IntoDiagnostic` defaults to `ErrorGuaranteed`, because errors are the
most common diagnostic level. It makes sense to do likewise for the
closely-related (and much more widely used) `DiagnosticBuilder` type,
letting us write `DiagnosticBuilder<'a, ErrorGuaranteed>` as just
`DiagnosticBuilder<'a>`. This cuts over 200 lines of code due to many
multi-line things becoming single line things.
subtype_predicate: remove unnecessary probe
There is no reason to probe here. The failure either results in an actual type error, in which cases the probe is useless, or it is used inside of evaluate, in which case we're already inside of the `fn evaluation_probe`, so it is also not necessary.
Yeet unnecessary param envs
We don't need to pass in param-envs around in the lexical region resolution code (or in `MatchAgainstFreshVars` in the solver), since it is only used to eval some consts in `structurally_relate_tys` which I removed.
This is in preparation for normalizing the outlives clauses in `ParamEnv` for the new trait solver.
r? lcnr
This commit replaces this pattern:
```
err.into_diagnostic(dcx)
```
with this pattern:
```
dcx.create_err(err)
```
in a lot of places.
It's a little shorter, makes the error level explicit, avoids some
`IntoDiagnostic` imports, and is a necessary prerequisite for the next
commit which will add a `level` arg to `into_diagnostic`.
This requires adding `track_caller` on `create_err` to avoid mucking up
the output of `tests/ui/track-diagnostics/track4.rs`. It probably should
have been there already.
Collect lang items from AST, get rid of `GenericBound::LangItemTrait`
r? `@cjgillot`
cc #115178
Looking forward, the work to remove `QPath::LangItem` will also be significantly more difficult, but I plan on doing it as well. Specifically, we have to change:
1. A lot of `rustc_ast_lowering` for things like expr `..`
2. A lot of astconv, since we actually instantiate lang and non-lang paths quite differently.
3. A ton of diagnostics and clippy lints that are special-cased via `QPath::LangItem`
Meanwhile, it was pretty easy to remove `GenericBound::LangItemTrait`, so I just did that here.
cache param env canonicalization
Canonicalize ParamEnv only once and store it. Then whenever we try to canonicalize `ParamEnvAnd<'tcx, T>` we only have to canonicalize `T` and then merge the results.
Prelimiary results show ~3-4% savings in diesel and serde benchmarks.
Best to review commits individually. Some commits have a short description.
Initial implementation had a soundness bug (https://github.com/rust-lang/rust/pull/117749#issuecomment-1840453387) due to cache invalidation:
- When canonicalizing `Ty<'?0>` we first try to resolve region variables in the current InferCtxt which may have a constraint `?0 == 'static`. This means that we register `Ty<'?0> => Canonical<Ty<'static>>` in the cache, which is obviously incorrect in another inference context.
- This is fixed by not doing region resolution when canonicalizing the query *input* (vs. response), which is the only place where ParamEnv is used, and then in a later commit we *statically* guard against any form of inference variable resolution of the cached canonical ParamEnv's.
r? `@ghost`
This doesn't change behavior.
It should prevent unintentional resolution of inference variables
during canonicalization, which previously caused a soundness bug.
See PR description for more.
ParamEnv is canonicalized in *queries input* rather than query response.
In such case we don't "preserve universes" of canonical variable.
This means that `universe_map` always has the default value, which is
wasteful to store in the cache.
Renamings:
- find -> opt_hir_node
- get -> hir_node
- find_by_def_id -> opt_hir_node_by_def_id
- get_by_def_id -> hir_node_by_def_id
Fix rebase changes using removed methods
Use `tcx.hir_node_by_def_id()` whenever possible in compiler
Fix clippy errors
Fix compiler
Apply suggestions from code review
Co-authored-by: Vadim Petrochenkov <vadim.petrochenkov@gmail.com>
Add FIXME for `tcx.hir()` returned type about its removal
Simplify with with `tcx.hir_node_by_def_id`
remove redundant imports
detects redundant imports that can be eliminated.
for #117772 :
In order to facilitate review and modification, split the checking code and removing redundant imports code into two PR.
r? `@petrochenkov`
detects redundant imports that can be eliminated.
for #117772 :
In order to facilitate review and modification, split the checking code and
removing redundant imports code into two PR.
recurse into refs when comparing tys for diagnostics
before:
after:
this diff from the test suite is also quite nice imo:
```diff
`@@` -4,8 +4,8 `@@` error[E0308]: mismatched types
LL | debug_assert_eq!(iter.next(), Some(value));
| ^^^^^^^^^^^ expected `Option<<I as Iterator>::Item>`, found `Option<&<I as Iterator>::Item>`
|
- = note: expected enum `Option<<I as Iterator>::Item>`
- found enum `Option<&<I as Iterator>::Item>`
+ = note: expected enum `Option<_>`
+ found enum `Option<&_>`
```
Unify `TraitRefs` and `PolyTraitRefs` in `ValuePairs`
I did this recently with `FnSigs` and `PolyFnSigs` but didn't think to do it with `TraitRefs` and `PolyTraitRefs`.
Currently we always do this:
```
use rustc_fluent_macro::fluent_messages;
...
fluent_messages! { "./example.ftl" }
```
But there is no need, we can just do this everywhere:
```
rustc_fluent_macro::fluent_messages! { "./example.ftl" }
```
which is shorter.
The `fluent_messages!` macro produces uses of
`crate::{D,Subd}iagnosticMessage`, which means that every crate using
the macro must have this import:
```
use rustc_errors::{DiagnosticMessage, SubdiagnosticMessage};
```
This commit changes the macro to instead use
`rustc_errors::{D,Subd}iagnosticMessage`, which avoids the need for the
imports.
Remove `HirId` from `QPath::LangItem`
Remove `HirId` from `QPath::LangItem`, since there was only *one* use-case (`ObligationCauseCode::AwaitableExpr`), which we can instead recover by walking the HIR tree.
Move EagerResolution to rustc_infer::infer::resolve
`EagerResolver` fits better in `rustc_infer::infer::resolver`.
Started to disentagle #118118 that has a lot of unrelated things.
r? `@compiler-errors` `@lcnr`
Cache flags for `ty::Const`
Not sure if this has been attempted yet, but worth a shot. It does make the code simpler in `rustc_type_ir`, since we can assume that consts have a `flags` method that is no-cost.
r? `@ghost`
Remove `PredicateKind::ClosureKind`
We don't need the `ClosureKind` predicate kind -- instead, `Fn`-family trait goals are left as ambiguous, and we only need to make progress on `FnOnce` projection goals for inference purposes.
This is similar to how we do confirmation of `Fn`-family trait and projection goals in the new trait solver, which also doesn't use the `ClosureKind` predicate.
Some hacky logic is added in the second commit so that we can keep the error messages the same.
Remove `-Zperf-stats`.
The included measurements have varied over the years. At one point there were quite a few more, but #49558 deleted a lot that were no longer used. Today there's just four, and it's a motley collection that doesn't seem particularly valuable.
I think it has been well and truly subsumed by self-profiling, which collects way more data.
r? `@wesleywiser`
The included measurements have varied over the years. At one point there
were quite a few more, but #49558 deleted a lot that were no longer
used. Today there's just four, and it's a motley collection that doesn't
seem particularly valuable.
I think it has been well and truly subsumed by self-profiling, which
collects way more data.
Most notably, this commit changes the `pub use crate::*;` in that file
to `use crate::*;`. This requires a lot of `use` items in other crates
to be adjusted, because everything defined within `rustc_span::*` was
also available via `rustc_span::source_map::*`, which is bizarre.
The commit also removes `SourceMap::span_to_relative_line_string`, which
is unused.
When we encounter a `dyn Trait` that isn't object safe, look for its
implementors. If there's one, mention using it directly If there are
less than 9, mention the possibility of creating a new enum and using
that instead.
Account for object unsafe `impl Trait on dyn Trait {}`. Make a
distinction between public and sealed traits.
Fix#80194.
Consider alias bounds when computing liveness in NLL (but this time sound hopefully)
This is a revival of #116040, except removing the changes to opaque lifetime captures check to make sure that we're not triggering any unsoundness due to the lack of general existential regions and the currently-existing `ReErased` hack we use instead.
r? `@aliemjay` -- I appreciate you pointing out the unsoundenss in the previous iteration of this PR, and I'd like to hear that you're happy with this iteration of this PR before this goes back into FCP :>
Fixes#116794 as well
---
(mostly copied from #116040 and reworked slightly)
# Background
Right now, liveness analysis in NLL is a bit simplistic. It simply walks through all of the regions of a type and marks them as being live at points. This is problematic in the case of aliases, since it requires that we mark **all** of the regions in their args[^1] as live, leading to bugs like #42940.
In reality, we may be able to deduce that fewer regions are allowed to be present in the projected type (or "hidden type" for opaques) via item bounds or where clauses, and therefore ideally, we should be able to soundly require fewer regions to be live in the alias.
For example:
```rust
trait Captures<'a> {}
impl<T> Captures<'_> for T {}
fn capture<'o>(_: &'o mut ()) -> impl Sized + Captures<'o> + 'static {}
fn test_two_mut(mut x: ()) {
let _f1 = capture(&mut x);
let _f2 = capture(&mut x);
//~^ ERROR cannot borrow `x` as mutable more than once at a time
}
```
In the example above, we should be able to deduce from the `'static` bound on `capture`'s opaque that even though `'o` is a captured region, it *can never* show up in the opaque's hidden type, and can soundly be ignored for liveness purposes.
# The Fix
We apply a simple version of RFC 1214's `OutlivesProjectionEnv` and `OutlivesProjectionTraitDef` rules to NLL's `make_all_regions_live` computation.
Specifically, when we encounter an alias type, we:
1. Look for a unique outlives bound in the param-env or item bounds for that alias. If there is more than one unique region, bail, unless any of the outlives bound's regions is `'static`, and in that case, prefer `'static`. If we find such a unique region, we can mark that outlives region as live and skip walking through the args of the opaque.
2. Otherwise, walk through the alias's args recursively, as we do today.
## Limitation: Multiple choices
This approach has some limitations. Firstly, since liveness doesn't use the same type-test logic as outlives bounds do, we can't really try several options when we're faced with a choice.
If we encounter two unique outlives regions in the param-env or bounds, we simply fall back to walking the opaque via its args. I expect this to be mostly mitigated by the special treatment of `'static`, and can be fixed in a forwards-compatible by a more sophisticated analysis in the future.
## Limitation: Opaque hidden types
Secondly, we do not employ any of these rules when considering whether the regions captured by a hidden type are valid. That causes this code (cc #42940) to fail:
```rust
trait Captures<'a> {}
impl<T> Captures<'_> for T {}
fn a() -> impl Sized + 'static {
b(&vec![])
}
fn b<'o>(_: &'o Vec<i32>) -> impl Sized + Captures<'o> + 'static {}
```
We need to have existential regions to avoid [unsoundness](https://github.com/rust-lang/rust/pull/116040#issuecomment-1751628189) when an opaque captures a region which is not represented in its own substs but which outlives a region that does.
## Read more
Context: https://github.com/rust-lang/rust/pull/115822#issuecomment-1731153952 (for the liveness case)
More context: https://github.com/rust-lang/rust/issues/42940#issuecomment-455198309 (for the opaque capture case, which this does not fix)
[^1]: except for bivariant region args in opaques, which will become less relevant when we move onto edition 2024 capture semantics for opaques.
When encountering an associated item with a type param that could be
constrained, do not look at the parent item if the type param comes from
the associated item.
Fix#117209.
`OptWithInfcx` naming nits, trait bound simplifications
* Use an associated type `Interner` on `InferCtxtLike` to remove a redundant interner parameter (`I: Interner, Infcx: InferCtxtLike<I>` -> `Infcx: InferCtxtLike`).
* Remove double-`Option` between `infcx: Option<Infcx>` and `fn universe_of_ty(&self, ty: ty::InferTy) -> Option<ty::UniverseIndex>`. We don't need the infcx to be optional if we can provide a "noop" (`NoInfcx`) implementation that just always returns `None` for universe index.
* Also removes the `core::convert::Infallible` implementation which I found a bit weird...
* Some naming nits with params.
* I found `InferCtxt` + `InfCtx` and `Infcx` to be a lot of different ways to spell "inference context", so I got rid of the `InfCtx` type parameter name in favor of `Infcx` which is a more standard name.
* I found `OptWithInfcx` to be a bit redundant -> `WithInfcx`.
I'm making these changes because I intend to reuse the `InferCtxtLike` trait for uplifting the canonicalizer into a new trait -- conveniently, the information I need for uplifting the canonicalizer also is just the universe information of a type var, so it's super convenient 😸
r? `@BoxyUwU` or `@lcnr`
Rollup of 6 pull requests
Successful merges:
- #107159 (rand use getrandom for freebsd (available since 12.x))
- #116859 (Make `ty::print::Printer` take `&mut self` instead of `self`)
- #117046 (return unfixed len if pat has reported error)
- #117070 (rustdoc: wrap Type with Box instead of Generics)
- #117074 (Remove smir from triage and add me to stablemir)
- #117086 (Update .mailmap to promote my livename)
r? `@ghost`
`@rustbot` modify labels: rollup
Make `ty::print::Printer` take `&mut self` instead of `self`
based on #116815
This simplifies the code by removing all the `self` assignments and
makes the flow of data clearer - always into the printer.
Especially in v0 mangling, which already used `&mut self` in some
places, it gets a lot more uniform.
This simplifies the code by removing all the `self` assignments and
makes the flow of data clearer - always into the printer.
Especially in v0 mangling, which already used `&mut self` in some
places, it gets a lot more uniform.
These are `Self` in almost all printers except one, which can just store
the state as a field instead. This simplifies the printer and allows for
further simplifications, for example using `&mut self` instead of
passing around the printer.
Relate alias ty with variance
In the new solver, turns out that the subst-relate branch of the alias-relate predicate was relating args invariantly even for opaques, which have variance 💀.
This change is a bit more invasive, but I'd rather not special-case it [here](aeaa5c30e5/compiler/rustc_trait_selection/src/solve/alias_relate.rs (L171-L190)) and then have it break elsewhere. I'm doing a perf run to see if the extra call to `def_kind` is that expensive, if it is, I'll reconsider.
r? ``@lcnr``
Show more information when multiple `impl`s apply
- When there are `impl`s without type params, show only those (to avoid showing overly generic `impl`s).
```
error[E0283]: type annotations needed
--> $DIR/multiple-impl-apply.rs:34:9
|
LL | let y = x.into();
| ^ ---- type must be known at this point
|
note: multiple `impl`s satisfying `_: From<Baz>` found
--> $DIR/multiple-impl-apply.rs:14:1
|
LL | impl From<Baz> for Bar {
| ^^^^^^^^^^^^^^^^^^^^^^
...
LL | impl From<Baz> for Foo {
| ^^^^^^^^^^^^^^^^^^^^^^
= note: required for `Baz` to implement `Into<_>`
help: consider giving `y` an explicit type
|
LL | let y: /* Type */ = x.into();
| ++++++++++++
```
- Lower the importance of `T: Sized`, `T: WellFormed` and coercion errors, to prioritize more relevant errors. The pre-existing deduplication logic deals with hiding redundant errors better that way, and we show errors with more metadata that is useful to the user.
- Show `<SelfTy as Trait>::assoc_fn` suggestion in more cases.
```
error[E0790]: cannot call associated function on trait without specifying the corresponding `impl` type
--> $DIR/cross-return-site-inference.rs:38:16
|
LL | return Err(From::from("foo"));
| ^^^^^^^^^^ cannot call associated function of trait
|
help: use a fully-qualified path to a specific available implementation
|
LL | return Err(</* self type */ as From>::from("foo"));
| +++++++++++++++++++ +
```
Fix#88284.
adjust how closure/generator types are printed
I saw `&[closure@$DIR/issue-20862.rs:2:5]` and I thought it is a slice type, because that's usually what `&[_]` is... it took me a while to realize that this is just a confusing printer and actually there's no slice. Let's use something that cannot be mistaken for a regular type.
This function is now used to check `#[panic_handler]`, `start` lang item, `main`, `#[start]` and intrinsic functions.
The diagnosis produced are now closer to the ones produced by trait/impl method signature mismatch.
move required_consts check to general post-mono-check function
This factors some code that is common between the interpreter and the codegen backends into shared helper functions. Also as a side-effect the interpreter now uses the same `eval` functions as everyone else to get the evaluated MIR constants.
Also this is in preparation for another post-mono check that will be needed for (the current hackfix for) https://github.com/rust-lang/rust/issues/115709: ensuring that all locals are dynamically sized.
I didn't expect this to change diagnostics, but it's just cycle errors that change.
r? `@oli-obk`
It's easier to pass it in to the one method that needs it
(`highlighting_region_vid`) than to store it in the type. This means
`RegionHighlightMode` can impl `Default`.
This fixes the changes brought to codegen tests when effect params are
added to libcore, by not attempting to monomorphize functions that get
the host param by being `const fn`.
Bubble up opaque <eq> opaque operations instead of picking an order
In case we are in `Bubble` mode (meaning every opaque type that is defined in the current crate is treated as if it were in its defining scope), we don't try to register an opaque type as the hidden type of another opaque type, but instead bubble up an obligation to equate them at the query caller site. Usually that means we have a `DefiningAnchor::Bind` and thus can reliably figure out whether an opaque type is in its defining scope. Where we can't, we'll error out, so the default is sound.
With this change we start using `AliasTyEq` predicates in the old solver, too.
fixes https://github.com/rust-lang/rust/issues/108498
But also regresses `tests/ui/impl-trait/anon_scope_creep.rs`. Our use of `Bubble` for `check_opaque_type_well_formed` is going to keep biting us.
r? `@lcnr` `@compiler-errors`
Speed up compilation of `type-system-chess`
[`type-system-chess`](https://github.com/rust-lang/rustc-perf/pull/1680) is an unusual program that implements a compile-time chess position solver in the trait system(!) This PR is about making it compile faster.
r? `@ghost`
Point at return type when it influences non-first `match` arm
When encountering code like
```rust
fn foo() -> i32 {
match 0 {
1 => return 0,
2 => "",
_ => 1,
}
}
```
Point at the return type and not at the prior arm, as that arm has type `!` which isn't influencing the arm corresponding to arm `2`.
Fix#78124.
Point out expectation even if we have `TypeError::RegionsInsufficientlyPolymorphic`
just a minor tweak, since saying "one type is more general than the other" kinda sucks if we don't actually point out two types.
When encountering code like
```rust
fn foo() -> i32 {
match 0 {
1 => return 0,
2 => "",
_ => 1,
}
}
```
Point at the return type and not at the prior arm, as that arm has type
`!` which isn't influencing the arm corresponding to arm `2`.
Fix#78124.
Store the laziness of type aliases in their `DefKind`
Previously, we would treat paths referring to type aliases as *lazy* type aliases if the current crate had lazy type aliases enabled independently of whether the crate which the alias was defined in had the feature enabled or not.
With this PR, the laziness of a type alias depends on the crate it is defined in. This generally makes more sense to me especially if / once lazy type aliases become the default in a new edition and we need to think about *edition interoperability*:
Consider the hypothetical case where the dependency crate has an older edition (and thus eager type aliases), it exports a type alias with bounds & a where-clause (which are void but technically valid), the dependent crate has the latest edition (and thus lazy type aliases) and it uses that type alias. Arguably, the bounds should *not* be checked since at any time, the dependency crate should be allowed to change the bounds at will with a *non*-major version bump & without negatively affecting downstream crates.
As for the reverse case (dependency: lazy type aliases, dependent: eager type aliases), I guess it rules out anything from slight confusion to mild annoyance from upstream crate authors that would be caused by the compiler ignoring the bounds of their type aliases in downstream crates with older editions.
---
This fixes#114468 since before, my assumption that the type alias associated with a given weak projection was lazy (and therefore had its variances computed) did not necessarily hold in cross-crate scenarios (which [I kinda had a hunch about](https://github.com/rust-lang/rust/pull/114253#discussion_r1278608099)) as outlined above. Now it does hold.
`@rustbot` label F-lazy_type_alias
r? `@oli-obk`
Rework upcasting confirmation to support upcasting to fewer projections in target bounds
This PR implements a modified trait upcasting algorithm that is resilient to changes in the number of associated types in the bounds of the source and target trait objects.
It does this by equating each bound of the target trait ref individually against the bounds of the source trait ref, rather than doing them all together by constructing a new trait object.
#### The new way we do trait upcasting confirmation
1. Equate the target trait object's principal trait ref with one of the supertraits of the source trait object's principal.
fdcab310b2/compiler/rustc_trait_selection/src/traits/select/mod.rs (L2509-L2525)
2. Make sure that every auto trait in the *target* trait object is present in the source trait ref's bounds.
fdcab310b2/compiler/rustc_trait_selection/src/traits/select/mod.rs (L2559-L2562)
3. For each projection in the *target* trait object, make sure there is exactly one projection that equates with it in the source trait ref's bound. If there is more than one, bail with ambiguity.
fdcab310b2/compiler/rustc_trait_selection/src/traits/select/mod.rs (L2526-L2557)
* Since there may be more than one that applies, we probe first to check that there is exactly one, then we equate it outside of a probe once we know that it's unique.
4. Make sure the lifetime of the source trait object outlives the lifetime of the target.
<details>
<summary>Meanwhile, this is how we used to do upcasting:</summary>
1. For each supertrait of the source trait object, take that supertrait, append the source object's projection bounds, and the *target* trait object's auto trait bounds, and make this into a new object type:
d12c6e947c/compiler/rustc_trait_selection/src/traits/select/confirmation.rs (L915-L929)
2. Then equate it with the target trait object:
d12c6e947c/compiler/rustc_trait_selection/src/traits/select/confirmation.rs (L936)
This will be a type mismatch if the target trait object has fewer projection bounds, since we compare the bounds structurally in relate:
d12c6e947c/compiler/rustc_middle/src/ty/relate.rs (L696-L698)
</details>
Fixes#114035
Also fixes#114113, because I added a normalize call in the old solver.
r? types
resolve before canonicalization in new solver, ICE if unresolved
Fold the values with a resolver before canonicalization instead of making it happen within canonicalization.
This allows us to filter trivial region constraints from the external constraints.
r? ``@lcnr``
Don't check unnecessarily that impl trait is RPIT
We have this random `return_type_impl_trait` function to detect if a function returns an RPIT which is used in outlives suggestions, but removing it doesn't actually change any diagnostics. Let's just remove it.
Also, suppress a spurious outlives error from a ReError.
Fixes#114274
Map RPITIT's opaque type bounds back from projections to opaques
An RPITIT in a program's AST is eventually translated into both a projection GAT and an opaque. The opaque is used for default trait methods, like:
```
trait Foo {
fn bar() -> impl Sized { 0i32 }
}
```
The item bounds for both the projection and opaque are identical, and both have a *projection* self ty. This is mostly okay, since we can normalize this projection within the default trait method body to the opaque, but it does two things:
1. it leads to bugs in places where we don't normalize item bounds, like `deduce_future_output_from_obligations`
2. it leads to extra match arms that are both suspicious looking and also easy to miss
This PR maps the opaque type bounds of the RPITIT's *opaque* back to the opaque's self type to avoid this quirk. Then we can fix the UI test for #108304 (1.) and also remove a bunch of match arms (2.).
Fixes#108304
r? `@spastorino`
Rollup of 7 pull requests
Successful merges:
- #114099 (privacy: no nominal visibility for assoc fns )
- #114128 (When flushing delayed span bugs, write to the ICE dump file even if it doesn't exist)
- #114138 (Adjust spans correctly for fn -> method suggestion)
- #114146 (Skip reporting item name when checking RPITIT GAT's associated type bounds hold)
- #114147 (Insert RPITITs that were shadowed by missing ADTs that resolve to [type error])
- #114155 (Replace a lazy `RefCell<Option<T>>` with `OnceCell<T>`)
- #114164 (Add regression test for `--cap-lints allow` and trait bounds warning)
r? `@ghost`
`@rustbot` modify labels: rollup
Skip reporting item name when checking RPITIT GAT's associated type bounds hold
Doesn't really make sense to label an item that has a name that users can't really mention. Fixes#114145. Also fixes#113794.
r? `@spastorino`
Introduce `trait DebugWithInfcx` to debug format types with universe info
Seeing universes of infer vars is valuable for debugging but currently we have no way of easily debug formatting a type with the universes of all the infer vars shown. In the future I hope to augment the new solver's proof tree output with a `DebugWithInfcx` impl so that it can show universes but I left that out of this PR as it would be non trivial and this is already large and complex enough.
The goal here is to make the various abstractions taking `T: Debug` able to use the codepath for printing out universes, that way we can do `debug!("{:?}", my_x)` and have `my_x` have universes shown, same for the `write!` macro. It's not possible to put the `Infcx: InferCtxtLike<I>` into the formatter argument to `Debug::fmt` so it has to go into the self ty. For this we introduce the type `OptWithInfcx<I: Interner, Infcx: InferCtxtLike<I>, T>` which has the data `T` optionally coupled with the infcx (more on why it's optional later).
Because of coherence/orphan rules it's not possible to write the impl `Debug for OptWithInfcx<..., MyType>` when `OptWithInfcx` is in a upstream crate. This necessitates a blanket impl in the crate defining `OptWithInfcx` like so: `impl<T: DebugWithInfcx> Debug for OptWithInfcx<..., T>`. It is not intended for people to manually call `DebugWithInfcx::fmt`, the `Debug` impl for `OptWithInfcx` should be preferred.
The infcx has to be optional in `OptWithInfcx` as otherwise we would end up with a large amount of code duplication. Almost all types that want to be used with `OptWithInfcx` do not themselves need access to the infcx so if we were to not optional we would end up with large `Debug` and `DebugWithInfcx` impls that were practically identical other than that when formatting their fields we wrap the field in `OptWithInfcx` instead of formatting it alone.
The only types that need access to the infcx themselves are ty/const/region infer vars, everything else is implemented by having the `Debug` impl defer to `OptWithInfcx` with no infcx available. The `DebugWithInfcx` impl is pretty much just the standard `Debug` impl except that instead of recursively formatting fields with `write!(f, "{x:?}")` we must do `write!(f, "{:?}", opt_infcx.wrap(x))`. This is some pretty rough boilerplate but I could not think of an alternative unfortunately.
`OptWithInfcx::wrap` is an eager `Option::map` because 99% of callsites were discarding the existing data in `OptWithInfcx` and did not need lazy evaluation.
A trait `InferCtxtLike` was added instead of using `InferCtxt<'tcx>` as we need to implement `DebugWithInfcx` for types living in `rustc_type_ir` which are generic over an interner and do not have access to `InferCtxt` since it lives in `rustc_infer`. Additionally I suspect that adding universe info to new solver proof tree output will require an implementation of `InferCtxtLike` for something that is not an `InferCtxt` although this is not the primary motivaton.
---
To summarize:
- There is a type `OptWithInfcx` which bundles some data optionally with an infcx with allows us to pass an infcx into a `Debug` impl. It's optional instead of being there unconditionally so that we can share code for `Debug` and `DebugWithInfcx` impls that don't care about whether there is an infcx available but have fields that might care.
- There is a trait `DebugWithInfcx` which allows downstream crates to add impls of the form `Debug for OptWithInfcx<...>` which would normally be forbidden by orphan rules/coherence.
- There is a trait `InferCtxtLike` to allow us to implement `DebugWithInfcx` for types that live in `rustc_type_ir`
This allows debug formatting various `ty::*` structures with universes shown by using the `Debug` impl for `OptWithInfcx::new(ty, infcx)`
---
This PR does not add `DebugWithInfcx` impls to absolutely _everything_ that should realistically have them, for example you cannot use `OptWithInfcx<Obligation<Predicate>>`. I am leaving this to a future PR to do so as it would likely be a lot more work to do.
Rollup of 8 pull requests
Successful merges:
- #113413 (Add needs-triage to all new issues)
- #113426 (Don't ICE in `resolve_bound_vars` when associated return-type bounds are in bad positions)
- #113427 (Remove `variances_of` on RPITIT GATs, remove its one use-case)
- #113441 (miri: check that assignments do not self-overlap)
- #113453 (Remove unused from_method from rustc_on_unimplemented)
- #113456 (Avoid calling report_forbidden_specialization for RPITITs)
- #113466 (Update cargo)
- #113467 (Fix comment of `fn_can_unwind`)
r? `@ghost`
`@rustbot` modify labels: rollup
Remove `variances_of` on RPITIT GATs, remove its one use-case
It doesn't make sense to implement variances on a GAT anyways, since we don't relate GATs with variance:
85bf07972a/compiler/rustc_middle/src/ty/relate.rs (L569-L579)
r? ``@spastorino``
Split `SelectionContext::select` into fns that take a binder and don't
*most* usages of `SelectionContext::select` don't need to use a binder, but wrap them in a dummy because of the signature. Let's split this out into `SelectionContext::{select,poly_select}` and limit the usages of the latter.
Right now, we only have 3 places where we're calling `poly_select` -- fulfillment, internally within the old solver, and the auto-trait finder.
r? `@lcnr`
`TypeParameterDefinition` always require a `DefId`
the `None` case never actually reaches diagnostics so it feels better for diagnostics to be able to rely on the `DefId` being there, cc #113310
Remove chalk support from the compiler
Removes chalk (`-Ztrait-solver=chalk`) from the compiler and prunes any dead code resulting from this, mainly:
* Remove the chalk compatibility layer in `compiler/rustc_traits/src/chalk`
* Remove the chalk flag `-Ztrait-solver=chalk` and its `TraitEngine` implementation
* Remove `TypeWellFormedFromEnv` (and its many `bug!()` match arms)
* Remove the chalk migration mode from compiletest
* Remove the `chalkify` UI tests (do we want to keep any of these, but migrate them to `-Ztrait-solver=next`??)
Fulfills rust-lang/types-team#93.
r? `@jackh726`
Make RPITITs assume/require their parent method's predicates
Removes a FIXME from the `param_env` query where we were manually adding the parent function's predicates to the RPITIT's assumptions.
r? `@spastorino`
[-Ztrait-solver=next, mir-typeck] instantiate hidden types in the root universe
Fixes an ICE in the test `member-constraints-in-root-universe`.
Main motivation is to make #112691 pass under the new solver.
r? ``@compiler-errors``
Various impl trait in assoc tys cleanups
r? `@compiler-errors`
All commits except for the last are pure refactorings. 274dab5bd658c97886a8987340bf50ae57900c39 allows struct fields to participate in deciding whether a function has an opaque in its signature.
best reviewed commit by commit
Add a fully fledged `Clause` type, rename old `Clause` to `ClauseKind`
Does two basic things before I put up a more delicate set of PRs (along the lines of #112714, but hopefully much cleaner) that migrate existing usages of `ty::Predicate` to `ty::Clause` (`predicates_of`/`item_bounds`/`ParamEnv::caller_bounds`).
1. Rename `Clause` to `ClauseKind`, so it's parallel with `PredicateKind`.
2. Add a new `Clause` type which is parallel to `Predicate`.
* This type exposes `Clause::kind(self) -> Binder<'tcx, ClauseKind<'tcx>>` which is parallel to `Predicate::kind` 😸
The new `Clause` type essentially acts as a newtype wrapper around `Predicate` that asserts that it is specifically a `PredicateKind::Clause`. Turns out from experimentation[^1] that this is not negative performance-wise, which is wonderful, since this a much simpler design than something that requires encoding the discriminant into the alignment bits of a predicate kind, or something else like that...
r? ``@lcnr`` or ``@oli-obk``
[^1]: https://github.com/rust-lang/rust/pull/112714#issuecomment-1595653910
Don't consider TAIT normalizable to hidden ty if it would result in impossible item bounds
See test for example where we shouldn't consider it possible to alias-relate a TAIT and hidden type.
r? `@lcnr`
Dont compute `opt_suggest_box_span` span for TAIT
Fixes#112434
Also a couple more commits on top, pruning some dead code and fixing another weird suggestion encountered in the above issue.
Add `-Ztrait-solver=next-coherence`
Flag that conditionally uses the trait solver *only* during coherence, for more testing and/or eventual partial-migration onto the trait solver (in the medium- to long-term).
* This still uses the selection context in some of the coherence methods I think, so it's not "complete". Putting this up for review and/or for further work in-tree.
* I probably need to spend a bit more time making sure that we don't sneakily create any other infcx's during coherence that also need the new solver enabled.
r? `@lcnr`
Emit an error when return-type-notation is used with type/const params
These are not intended to be supported initially, even though the compiler supports them internally...
we can otherwise assign a hidden type to the opaque which
causes ICE if we don't use `take_opaque_types` during
coherence. This is annoying so I didn't bother. Added a test
showing the behavior this prevents.
Each of `{D,Subd}iagnosticMessage::{Str,Eager}` has a comment:
```
// FIXME(davidtwco): can a `Cow<'static, str>` be used here?
```
This commit answers that question in the affirmative. It's not the most
compelling change ever, but it might be worth merging.
This requires changing the `impl<'a> From<&'a str>` impls to `impl
From<&'static str>`, which involves a bunch of knock-on changes that
require/result in call sites being a little more precise about exactly
what kind of string they use to create errors, and not just `&str`. This
will result in fewer unnecessary allocations, though this will not have
any notable perf effects given that these are error paths.
Note that I was lazy within Clippy, using `to_string` in a few places to
preserve the existing string imprecision. I could have used `impl
Into<{D,Subd}iagnosticMessage>` in various places as is done in the
compiler, but that would have required changes to *many* call sites
(mostly changing `&format("...")` to `format!("...")`) which didn't seem
worthwhile.
