Add test for issue #101532: dead code warnings in const _
Closesrust-lang/rust#101532
Adds a test for dead code warnings in `const _`.
This was already fixed, just adding coverage to avoid regressions.
Avoid ICE in explicit reference cast suggestion for unrelated leaf pr…
Fixesrust-lang/rust#154403
The explicit reference cast suggestion in was enabled based on `main_trait_predicate` being `From`/`TryFrom`, but it then extracted the source type from `leaf_trait_predicate`.
That is only valid when the leaf obligation is also part of the `From`/`TryFrom` conversion family.
Improve doc comment unicode guidance
Fixesrust-lang/rust#153096
This PR does not suggest HTML bidi markup, because doc comments are still Rust source first, not only rendered rustdoc output. HTML only helps in the rendered documentation rather than making them explicit in source.
move many tests out of `ui/unsafe`
`ui/unsafe` is a pretty big and generic directory. This PR moves some tests from it to `ui/union` and some others to a new `rustc_layout_scalar_valid_range` directory.
r? @Kivooeo
Remove divergence check from check_expr_array
Fixesrust-lang/rust#153695.
`check_expr_array` currently assumes it should only be entered with` self.diverges == Diverges::Maybe`, but that assumption does not appear to hold in all valid cases. A never-pattern parameter can seed a function or closure body with inherited `Diverges::Always`, and exprs in that body are still typecked.
Make `rustc_hir_analysis` not depend on `rustc_lint`.
`rustc_hir_analysis` depends on `rustc_lint` in just a single function: `emit_delayed_lint`, which is used by the
"emit_ast_lowering_delayed_lints" checking section within `rustc_hir_analysis::check_crate`.
This commit moves that function and section to out of `rustc_hir_analysis::check_crate`, into `rustc_interface`, eliminating the dependency. This seems reasonable because the delayed lint errors aren't really related to HIR analysis, they were in there just because HIR analysis is what follows AST lowering.
This means `rustc_hir_analysis` and `rustc_lint` can both start compiling as soon as `rustc_trait_selection` finishes. This also changes the error order in one test, which doesn't matter.
The commit also changes `emit_delayed_lint` to `emit_delayed_lints`, factoring out some code duplicated in rustdoc.
r? @davidtwco
Rollup of 2 pull requests
Successful merges:
- rust-lang/rust#154229 (Ensure `ErasedData` only implements appropriate auto traits)
- rust-lang/rust#154409 (Update `try_blocks` to a new tracking issue number)
Suggest using equality comparison instead of pattern matching on non-structural constant in pattern
When encountering a pattern containing a non-structural constant (not marked as `#[derive(PartialEq)]` to make it suitable for pattern matching, `C` in the examples below), we would previously not provide additional guidance. With this PR, the `help` in the following examples are added:
```
error: constant of non-structural type `partial_eq::S` in a pattern
--> $DIR/suggest_equality_comparison_instead_of_pattern_matching.rs:16:18
|
LL | struct S;
| -------- `partial_eq::S` must be annotated with `#[derive(PartialEq)]` to be usable in patterns
...
LL | const C: S = S;
| ---------- constant defined here
...
LL | Some(C) => {}
| ^ constant of non-structural type
|
note: the `PartialEq` trait must be derived, manual `impl`s are not sufficient; see https://doc.rust-lang.org/stable/std/marker/trait.StructuralPartialEq.html for details
--> $DIR/suggest_equality_comparison_instead_of_pattern_matching.rs:5:5
|
LL | impl PartialEq<S> for S {
| ^^^^^^^^^^^^^^^^^^^^^^^
help: add a condition to the match arm checking for equality
|
LL - Some(C) => {}
LL + Some(binding) if binding == C => {}
|
```
```
error: constant of non-structural type `partial_eq::S` in a pattern
--> $DIR/suggest_equality_comparison_instead_of_pattern_matching.rs:22:18
|
LL | struct S;
| -------- `partial_eq::S` must be annotated with `#[derive(PartialEq)]` to be usable in patterns
...
LL | const C: S = S;
| ---------- constant defined here
...
LL | let Some(C) = Some(S) else { return; };
| ^ constant of non-structural type
|
note: the `PartialEq` trait must be derived, manual `impl`s are not sufficient; see https://doc.rust-lang.org/stable/std/marker/trait.StructuralPartialEq.html for details
--> $DIR/suggest_equality_comparison_instead_of_pattern_matching.rs:5:5
|
LL | impl PartialEq<S> for S {
| ^^^^^^^^^^^^^^^^^^^^^^^
help: check for equality instead of pattern matching
|
LL - let Some(C) = Some(S) else { return; };
LL + if Some(C) == Some(S) { return; };
|
```
The suggestion accounts for a few conditions:
- if the type is not from the local crate and has no `PartialEq` impl, the user can't make it structural, so we don't provide the suggestion
- regardless of whether the type is local or remote, if it has a manual `PartialEq`, explain that with a derived `PartialEq` you could use equality
- if the type is local and has no impl, suggest adding a derived `PartialEq` and use equality check instead of pattern matching
- when suggesting equality, account for `if-let` to suggest chaining (edition dependent), `match` arm with a present `if` check, `match` arm without an existing `if` check
- when encountering `let-else`, we suggest turning it into an `if` expression instead (this doesn't check for additional bindings beyond the constant, which would suggest incorrect code in some more complex cases).
Fixrust-lang/rust#42753.
`rustc_hir_analysis` depends on `rustc_lint` in just a single function:
`emit_delayed_lint`, which is used by the
"emit_ast_lowering_delayed_lints" checking section within
`rustc_hir_analysis::check_crate`.
This commit moves that function and section to out of
`rustc_hir_analysis::check_crate`, into `rustc_interface`, eliminating
the dependency. This seems reasonable because the delayed lint errors
aren't really related to HIR analysis. They were in there just because
HIR analysis follows AST lowering.
This means `rustc_hir_analysis` and `rustc_lint` can both start
compiling as soon as `rustc_trait_selection` finishes. This also changes
the error order in one test, which doesn't matter.
The commit also changes `emit_delayed_lint` to `emit_delayed_lints`,
factoring out some code duplicated in rustdoc.
Migrate UI tests
In this pull request, I am migrating the following files and adding a comment at the top, including a link to the issue they were regression tests for:
- `tests/ui/issues/issue-4735.rs` ➝ `tests/ui/drop/drop-noncopyable-raw-pointer.rs`
- `tests/ui/issues/issue-17734.rs` ➝ `tests/ui/codegen/box-str-drop-glue.rs`
r? Kivooeo
rustc_expand: improve diagnostics for non-repeatable metavars
There was an initally opened pr which solve this issue here https://github.com/rust-lang/rust/pull/152679. It got merged but, there was a perf regression. And this new pr is opened to address the problem. The first did the computation of binding and matched_rule and then passed them as owned value down to `diagnostics::emit_frag_parse_err(` but, now this pr address the issue by passing `lhs` and `rules` as borrowed value to from_tts and the move the logic to `diagnostics::emit_frag_parse_err(`.
Fix https://github.com/rust-lang/rust/issues/47452.
Move ui/issues tests to relevant subdirectories
Related to https://github.com/rust-lang/rust/issues/133895 and [Reorganisation of tests/ui/issues for GSOC](https://github.com/rust-lang/google-summer-of-code?tab=readme-ov-file#reorganisation-of-testsuiissues)
This is the first PR in a batch of PRs.
Moves `ui/issues/issue-17546.rs` -> `ui/variants/variant-result-noresult-used-as-type.rs`
Approach:
1. Check linked issue and test contents, determine new target directory
2. Move / rename tests to reflect purpose.
3. Add issue link / comments in separate commit.
4. Rebless if necessary and remove obsolete stderr.
r? @Kivooeo
begin `tests/ui/structs-enums` cleanup
Nearly all tests in this directory are heavily outdated, poorly formatted and have a lot of duplication. This PR is the first of a planned series of PRs to combinine this, `ui/structs` and `ui/enum` into a better structure (`ui/adt` maybe?).
some `tests/ui/macros` cleanup
Move most tests that do not run any code from `//@ run-pass` to `//@ check-pass` and merge the (outdated) `die-macro-*` tests into one file.
Move tests in the statics category
I have moved some tests I feel belong in the statics directory. Please review and let me know if this is the correct way. I think on the first two files I moved, i forgot to turn off rust analyzer and it probably formatted the files, will this be an issue?
delegation: don't propagate synthetic params, remove lifetime hacks
Some small fixes after new delegation lowering was merged: remove lifetime hacks as now we get only early-bound lifetimes from generics, don't propagate synthetic generic params as now we know that they are synthetic. Fixesrust-lang/rust#143498. Part of rust-lang/rust#118212.
r? @petrochenkov
delegation: fix zero-args nested delegation ICE
This PR fixes an ICE when during lowering of nested delegation we need to access information about its parent, who is also inside body of another delegation. As a fix we lower delegation body even if there are no arguments in signature function, in this case we will see an error `this function takes 0 arguments but 1 argument was supplied`. Fixesrust-lang/rust#154332. Part of rust-lang/rust#118212.
r? @petrochenkov
Add macro matcher for `guard` fragment specifier
Tracking issue #153104
This PR implements a new `guard` macro matcher to match `if-let` guards (specifically [`MatchArmGuard`](https://github.com/rust-lang/reference/blob/50a1075e879be75aeec436252c84eef0fad489f4/src/expressions/match-expr.md#match-guards)). In the upcoming PR, we can use this new matcher in the `matches!` and `assert_matches!` macros to support their use with `if-let` guards. (see #152313)
The original `Expr` used to represent a guard has been wrapped in a new `Guard` type, allowing us to carry the span information of the leading `if` keyword. However, it might be even better to include the `if` keyword in the `Guard` type as well? I've left a FIXME comment in the code.
Add `-Zsanitize=kernel-hwaddress`
The Linux kernel has a config option called `CONFIG_KASAN_SW_TAGS` that enables `-fsanitize=kernel-hwaddress`. This is not supported by Rust.
One slightly awkward detail is that `#[sanitize(address = "off")]` applies to both `-Zsanitize=address` and `-Zsanitize=kernel-address`. Probably it was done this way because both are the same LLVM pass. I replicated this logic here for hwaddress, but it might be undesirable.
Note that `#[sanitize(kernel_hwaddress = "off")]` could be supported as an annotation on statics, but since it's also missing for `#[sanitize(hwaddress = "off")]`, I did not add it.
MCP: https://github.com/rust-lang/compiler-team/issues/975
Tracking issue: https://github.com/rust-lang/rust/issues/154171
cc @rcvalle @maurer @ojeda
* Move issue-50411 to tests/ui/mir/inliner-double-elaborate
* Addded the link for the issue to inliner-double-elaborate.rs
* Fix tidy issues
* Renamed to inliner-double-elaborate.rs
Ralf Jung