Only the end bounds of ranges can actually be included or excluded. This
commit changes the SpannedRange type to reflect that. Update `Kind::value`
to and `Kind::cmp` for this change. `Kind::cmp` gets flipped to check value
first and then the bound details and is much shorter.
This unbounded case never actually happens because `all_ranges(..)` uses
the scrutinee type bounds for open ranges. Switch to our own `Bound`
enum so that we don't have this case.
Introduce `expr_visitor` and `expr_visitor_no_bodies`
changelog: none
A couple utils that satisfy a *lot* of visitor use cases. Factoring in every possible usage would be really big so I just focused on cleaning clippy_utils.
TraitKind -> Trait
TyAliasKind -> TyAlias
ImplKind -> Impl
FnKind -> Fn
All `*Kind`s in AST are supposed to be enums.
Tuple structs are converted to braced structs for the types above, and fields are reordered in syntactic order.
Also, mutable AST visitor now correctly visit spans in defaultness, unsafety, impl polarity and constness.
Replace `in_macro` usage with `from_expansion`
changelog: none
Generally replace `in_macro(span)` with `span.from_expansion()`. If we're just trying to avoid expanded code, this seems more appropriate because any kind of expanded code is prone to false positives. One place I did not touch is `macro_use.rs`. I think this lint could use a rewrite so I moved `in_macro` there, the only place it is still used.
Move non_ascii_literal to restriction
It feels like the more apt category, since cases where you'd want it enabled would be pretty specific
changelog: Move [`non_ascii_literal`] to `restriction`
Prevent clippy::needless_lifetimes false positive in async function definition
Scan `OpaqueDef` bounds for lifetimes as well. Those `OpaqueDef` instances are generated while desugaring an `async` function definition.
This fixes#7893
changelog: Prevent [`clippy::needless_lifetimes`] false positive in `async` function definition
Fix manual_assert and match_wild_err_arm for `#![no_std]` and Rust 2021
Rust 2015 `std::panic!` has a wrapping block while `core::panic!` and Rust 2021 `std::panic!` does not. See rust-lang/rust#88919 for details.
Note that the test won't pass until clippy changes in rust-lang/rust#88860 is synced.
---
changelog: Fix [`manual_assert`] and [`match_wild_err_arm`] for `#![no_std]` and Rust 2021.
Fixes#7723
Unseparated literal suffix
Closes#7658
Since `literal_suffix` style is opinionated, we should disable by default and only enforce if it's stated as so.
changelog: [`unseparated_literal_suffix`] is renamed to `literal_suffix`, adds a new configuration `literal-suffix-style` to enforce a certain style writing literal_suffix. Possible values for `literal-suffix-style`: `"separated"`, `"unseparated"`
avoid linting `possible_truncation` on bit-reducing operations
---
*Please write a short comment explaining your change (or "none" for internal only changes)*
changelog: avoid linting `possible_truncation` on bit-reducing operations
`unseparated_literal_suffix`
This commit adds a configuration `literal-suffix-style` to enforce a
specific style for unseparated_literal_suffix. The configuration accepts
two values:
- "separated"
enforce all literals to be written separately (e.g. `123_i32`)
- "unseparated"
enforce all literals to be written as unseparated (e.g. `123i32`)
Not specifying a value means that there is no preference on style and
any style should not be warned.
Disable "if_not_else" lints from firing on else-ifs
Fixes#7892
1. Convert `['if_not_else']` to `LateLintPass` and use `clippy_utils::is_else_clause` for checking.
2. Update tests.
changelog: [`if_not_else`] now ignores else if statements.
1. Convert IfNotElse to LateLintPass and use clippy_utils::is_else_clause for checking.
2. Handle the case where the span comes from desugaring.
3. Update tests.
Ignore references to type aliases in ptr_arg
Works using the fact that the hir path will point to a TyAlias, rather than being resolved to the underlying type
Fixes#7699
changelog: [`ptr_arg`] No longer lints references to type aliases
Add unit-hash lint
changelog: [`unit_hash`] Add lint for hashing unit values
This will lint for situations where the end user is attempting to hash a unit value (`()`), as the implementation in `std` simply [does nothing][impl]. Closes#7159 .
Example:
```rust
().hash(&mut state);
// Should (probably) be replaced with:
0_u8.hash(&mut state);
```
[impl]: a5f164faad/library/core/src/hash/mod.rs (L656)
Register the generated lints from `cargo dev new_lint`
How to register a lint was something that took me a couple reads to figure out, this will hopefully make that easier. It appends the created lint to the end of the list when running `cargo dev new_lint`
changelog: none
Fix `question_mark` FP on custom error type
Closes#7859#7840 aims to ignore `question_mark` when the return type is custom, which is [covered here](df65291edd/tests/ui/question_mark.rs (L144-L149)). But this fails when there is a call in conditional predicate
changelog: [`question_mark`] Fix false positive when there is call in conditional predicate
Clean up tests/ui/rename.rs
Part one of #7057, cleaning up `tests/ui/rename.rs`. `tests/ui/deprecated.rs` will be updated in a subsequent PR.
changelog: none
new lint: string-slice
This is a restriction lint to highlight code that should have tests containing non-ascii characters. See #6623.
changelog: new lint: [`string-slice`]
Fix `match_str_case_mismatch` on uncased chars
False positives would result because `char::is_lowercase` and friends will return `false` for non-alphabetic chars and alphabetic chars lacking case (such as CJK scripts). Care also has to be taken for handling titlecase characters (`Dz`) and lowercased chars with no uppercase equivalent (`ʁ`).
For example, when verifying lowercase:
* Check `!any(char::is_ascii_uppercase)` instead of `all(char::is_ascii_lowercase)` for ASCII.
* Check that `all(|c| c.to_lowercase() == c)` instead of `all(char::is_lowercase)` for non-ASCII
Fixes#7863.
changelog: Fix false positives in [`match_str_case_mismatch`] on uncased characters
Update `str` utils to prevent ICEs and FNs
This PR reworks some string handling for lints regarding enum naming. I hope the refactoring will prevent future ICEs and help with new bug free implementations.
It might be better to review this PR by going through the commits, as `clippy_utils::camel_case` was renamed to `clippy_utils::str_utils` and then changed further. GH sadly doesn't really make the changes that obvious 🙃
Not too much more to say. Have a nice day 🌞
---
Fixes: rust-lang/rust-clippy#7869
changelog: ICE Fix: [`enum_variant_names`] #7869
Don't mark for loop iter expression as desugared
We typically don't mark spans of lowered things as desugared. This helps Clippy rightly discern when code is (not) from expansion. This was discovered by ``@flip1995`` at https://github.com/rust-lang/rust-clippy/pull/7789#issuecomment-939289501.
missing_safety_doc: Handle 'implementation safety' headers as well
We hit some FPs on this in `yoke`, it's somewhat normal to mark trait impl safety with "implementation safety". We could also broaden the check for headers which contain the word "safety" somehow, or split out impl safety stuff to only apply to traits.
changelog: handle 'implementation safety' headers in `missing_safety_doc`
Fix FP: no lint when cast is coming from `signum` method call for `cast_possible_truncation` lint
Fixes a FP when cast is coming from `signum` method call
fixes: #5395
changelog: [`cast_possible_truncation`] Fix FP when cast is coming from `signum` method call
Warn on structs with a trailing zero-sized array but no `repr` attribute
Closes#2868
changelog: Implement ``[`trailing_empty_array`]``, which warns if a struct is defined where the last field is a zero-sized array but there are no `repr` attributes. Zero-sized arrays aren't very useful in Rust itself, so such a struct is likely being created to pass to C code or in some other situation where control over memory layout matters. Either way, a `repr` attribute is needed.
FIx FP in `missing_safety_doc` lint
Fix FP where lint souldn't fire if any parent has `#[doc(hidden)]` attribute
fixes: #7347
changelog: [`missing_safety_doc`] Fix FP if any parent has `#[doc(hidden)]` attribute
The bug was dues to the constant bytes being compared instead of their
values. This meant that negative values were being treated as larger
than some positive values.
Fixes#7829
avoid `eq_op` in test code
Add a check to `eq_op` that will avoid linting in functions annotated with `#[test]`
---
*Please write a short comment explaining your change (or "none" for internal only changes)*
changelog: avoid `eq_op` in test functions
Some "parenthesis" and "parentheses" fixes
"Parenthesis" is the singular (e.g. one `(` or one `)`) and "parentheses" is the plural (multiple `(` or `)`s) and this is not hard to mix up so here are some fixes for that.
Inspired by #89958
Add `format_in_format_args` and `to_string_in_format_args` lints
Fixes#7667 and #7729
I put these in `perf` since that was one of `@jplatte's` suggestions, and `redundant_clone` (which I consider to be similar) lives there as well.
However, I am open to changing the category or anything else.
r? `@camsteffen`
changelog: Add `format_in_format_args` and `to_string_in_format_args` lints
Do not expand macros in equatable_if_let suggestion
Fixes#7781
Let's use Hacktoberfest as a motivation to start contributing PRs myself again :)
changelog: [`equatable_if_let`]: No longer expands macros in the suggestion
Allow giving reasons for `disallowed_types`
Similar to #7609 but for the `disallowed_type` lint. The permitted form of configuration is the same as for `disallowed_methods`.
changelog: Allow giving reasons for [`disallowed_type`]
Fix FP in external macros for `mut_mut` lint
Fix FP in `mut_mut` lint when type is defined in external macros.
fixes: #6922
changelog: [`mut_mut`] Fix FP when type is defined in external macros
Refactor `clippy::match_ref_pats` to check for multiple reference patterns
fixes#7740
When there is only one pattern, to begin with, i.e. a single deref(`&`), then in such cases we suppress `clippy::match_ref_pats`.
This is done by checking the count of the reference pattern and emitting `clippy::match_ref_pats` when more than one pattern is present.
Removed certain errors in the `stderr` tests as they would not be triggered.
changelog: Refactor `clippy::match_ref_pats` to check for multiple reference patterns
Before this lint didn't trigger on macros. With rust-lang/rust#88175
this isn't enough anymore. In this PR a `WhileLoop` desugaring kind was
introduced. This overrides the span of expanded expressions when
lowering the while loop. So if a while loop is in a macro, the
expressions that it expands to are no longer marked with
`ExpnKind::Macro`, but with `ExpnKind::Desugaring`. In general, this is
the correct behavior and the same that is done for `ForLoop`s. It just
tripped up this lint.
Restriction lint for function pointer casts
The existing lints for function pointer casts cover the cases where a cast is non-portable or would result in truncation, however there's currently no way to forbid numeric function pointer casts entirely.
I've added a new lint `fn_to_numeric_cast_any`, which allows one to ban _all_ numeric function pointer casts, including to `usize`. This is useful if you're writing high-level Rust and want to opt-out of potentially surprising behaviour, avoiding silent bugs from forgotten parentheses, e.g.
```rust
fn foo() -> u32 {
10
}
fn main() {
let _ = foo as usize; // oops, forgot to call foo and got a random address instead!
}
```
~~I'm open to suggestions for the naming of the lint, because `fn_to_numeric_cast_any` is a bit clunky. Ideally I'd call this lint `fn_to_numeric_cast`, but that name is already taken for the more specific lint~~. We've stuck with `fn_to_numeric_cast_any` to avoid renaming the existing lint, or choosing a different name that's too generic (like `fn_cast`).
I'm also open to changing the suggestion behaviour, as adding parentheses is only one of many possible ways to fix the lint.
changelog: add ``[`fn_to_numeric_cast_any`]`` restriction lint
Make `shadow_reuse` suggestion less verbose
Closes#7764
Make `shadow_reuse` suggestion less verbose.
changelog: [`shadow_reuse`] does not warn shadowing statement
fix bug for large_enum_variants
Fix the discussion problem in the issue of https://github.com/rust-lang/rust-clippy/issues/7666#issuecomment-919654291
About the false positive problem of case:
```rust
enum LargeEnum6 {
A,
B([u8;255]),
C([u8;200]),
}
```
changelog: Fix largest_enum_variant wrongly identifying the second largest variant.
Rustup
This needs a review this time. Especially 521bf8f0fa cc `@camsteffen` I think this is necessary now, because `itertools` is no longer a dependency of `clippy_utils` and therefore this path can't be found 🤔
( I forgot about the sync last week. I should get to document this process better, so other people can do it when I'm not around )
changelog: none
bump clippy crates to edition 2021
Also helps with dogfooding edition 2021 a bit. :)
Tests passed locally.
---
changelog: bump edition from 2018 to 2021
Demote float_cmp to pedantic
See this issue: https://github.com/rust-lang/rust-clippy/issues/7666
This is one of the most frequently suppressed lints. It is deny-by-default. It is not actually clearly wrong, as there are many instances where direct float comparison is actually desirable. It is only after operating on floats that they may lose precision, and that depends greatly on the operation. As most correctness lints have a much higher standard of error, being based on hard and fast binary logic, this should not be amongst them.
A linter is not a substitute for observing the math carefully and running tests, and doing the desirable thing is even more likely to lead one to want exact comparisons.
changelog: Demote [`float_cmp`] from correctness to pedantic lints
Be explicit about using Binder::dummy
This is somewhat of a late followup to the binder refactor PR. It removes `ToPredicate` and `ToPolyTraitImpls` that hide the use of `Binder::dummy`. While this does make code a bit more verbose, it allows us be more careful about where we create binders.
Another alternative here might be to add a new trait `ToBinder` or something with a `dummy()` fn. Which could still allow grepping but allows doing something like `trait_ref.dummy()` (but I also wonder if longer-term, it would be better to be even more explicit with a `bind_with_vars(ty::List::empty())` *but* that's not clear yet.
r? ``@nikomatsakis``
Don't lint `suspicious_else_formatting` inside proc-macros
fixes: #7650
I'll add a test for this one soon.
changelog: Don't lint `suspicious_else_formatting` inside proc-macros
Expand box_vec lint to box_collection
fixed#7451
changelog: Expand `box_vec` into [`box_collection`], and have it error on all sorts of boxed collections
Change `while_let_on_iterator` suggestion to use `by_ref()`
It came up in the discussion #7659 that suggesting `iter.by_ref()` is a clearer suggestion than `&mut iter`. I personally think they're equivalent, but if `by_ref()` is clearer to people then that should be the suggestion.
changelog: Change `while_let_on_iterator` suggestion when using `&mut` to use `by_ref()`
New lint: `same_name_method`
changelog: ``[`same_name_method`]``
fix: https://github.com/rust-lang/rust-clippy/issues/7632
It only compares a method in `impl` with another in `impl trait for`
It doesn't lint two methods in two traits.
I'm not sure my approach is the best way. I meet difficulty in other approaches.
Fix various redundant_closure bugs
changelog: Fix various false negatives and false positives for [`redundant_closure`]
Closes#3071Closes#4002
This lint is full of weird nuances and this is basically a re-write to tighten up the logic.
Encode spans relative to the enclosing item
The aim of this PR is to avoid recomputing queries when code is moved without modification.
MCP at https://github.com/rust-lang/compiler-team/issues/443
This is achieved by :
1. storing the HIR owner LocalDefId information inside the span;
2. encoding and decoding spans relative to the enclosing item in the incremental on-disk cache;
3. marking a dependency to the `source_span(LocalDefId)` query when we translate a span from the short (`Span`) representation to its explicit (`SpanData`) representation.
Since all client code uses `Span`, step 3 ensures that all manipulations
of span byte positions actually create the dependency edge between
the caller and the `source_span(LocalDefId)`.
This query return the actual absolute span of the parent item.
As a consequence, any source code motion that changes the absolute byte position of a node will either:
- modify the distance to the parent's beginning, so change the relative span's hash;
- dirty `source_span`, and trigger the incremental recomputation of all code that
depends on the span's absolute byte position.
With this scheme, I believe the dependency tracking to be accurate.
For the moment, the spans are marked during lowering.
I'd rather do this during def-collection,
but the AST MutVisitor is not practical enough just yet.
The only difference is that we attach macro-expanded spans
to their expansion point instead of the macro itself.
Fix result order for `manual_split_once` when `rsplitn` is used
fixes: #7656
changelog: Fix result order for `manual_split_once` when `rsplitn` is used
rustc: use more correct span data in for loop desugaring
Fixes#82462
Before:
help: consider adding semicolon after the expression so its temporaries are dropped sooner, before the local variables declared by the block are dropped
|
LL | for x in DroppingSlice(&*v).iter(); {
| +
After:
help: consider adding semicolon after the expression so its temporaries are dropped sooner, before the local variables declared by the block are dropped
|
LL | };
| +
This seems like a reasonable fix: since the desugared "expr_drop_temps_mut" contains the entire desugared loop construct, its span should contain the entire loop construct as well.
Add new lint `iter_not_returning_iterator`
Add new lint [`iter_not_returning_iterator`] to detect method `iter()` or `iter_mut()` returning a type not implementing `Iterator`
changelog: Add new lint [`iter_not_returning_iterator`]
Avoid slice indexing in Clippy (down with the ICEs)
While working on #7569 I got about 23 lint reports where we can avoid slice indexing by destructing the slice early. This is a preparation PR to avoid fixing them in the lint PR. (The implementation already takes about 300 lines without tests 😅). Either way, this should hopefully be easy to review 🙃
---
changelog: none
Provide `layout_of` automatically (given tcx + param_env + error handling).
After #88337, there's no longer any uses of `LayoutOf` within `rustc_target` itself, so I realized I could move the trait to `rustc_middle::ty::layout` and redesign it a bit.
This is similar to #88338 (and supersedes it), but at no ergonomic loss, since there's no funky `C: LayoutOf<Ty = Ty>` -> `Ty: TyAbiInterface<C>` generic `impl` chain, and each `LayoutOf` still corresponds to one `impl` (of `LayoutOfHelpers`) for the specific context.
After this PR, this is what's needed to get `trait LayoutOf` (with the `layout_of` method) implemented on some context type:
* `TyCtxt`, via `HasTyCtxt`
* `ParamEnv`, via `HasParamEnv`
* a way to transform `LayoutError`s into the desired error type
* an error type of `!` can be paired with having `cx.layout_of(...)` return `TyAndLayout` *without* `Result<...>` around it, such as used by codegen
* this is done through a new `LayoutOfHelpers` trait (and so is specifying the type of `cx.layout_of(...)`)
When going through this path (and not bypassing it with a manual `impl` of `LayoutOf`), the end result is that only the error case can be customized, the query itself and the success paths are guaranteed to be uniform.
(**EDIT**: just noticed that because of the supertrait relationship, you cannot actually implement `LayoutOf` yourself, the blanket `impl` fully covers all possible context types that could ever implement it)
Part of the motivation for this shape of API is that I've been working on querifying `FnAbi::of_*`, and what I want/need to introduce for that looks a lot like the setup in this PR - in particular, it's harder to express the `FnAbi` methods in `rustc_target`, since they're much more tied to `rustc` concepts.
r? `@nagisa` cc `@oli-obk` `@bjorn3`
Path remapping: Make behavior of diagnostics output dependent on presence of --remap-path-prefix.
This PR fixes a regression (#87745) with `--remap-path-prefix` where the flag stopped causing diagnostic messages to be remapped as well. The regression was introduced in https://github.com/rust-lang/rust/pull/83813 where we erroneously assumed that remapping of diagnostic messages was not desired anymore (because #70642 partially undid that functionality with nobody objecting).
The issue is fixed by making `--remap-path-prefix` remap diagnostic messages again, including for paths that have been remapped in upstream crates (e.g. the standard library). This means that "sysroot-localization" (implemented in #70642) is also disabled if `rustc` is invoked with `--remap-path-prefix`. The assumption is that once someone starts explicitly remapping paths they also don't want paths to their local Rust installation in their build output.
In the future we might want to give more fine-grained control over this behavior via compiler flags (see https://github.com/rust-lang/rfcs/pull/3127 for a related RFC). For now this PR is intended as a regression fix.
This PR is an alternative to https://github.com/rust-lang/rust/pull/88191, which makes diagnostic messages be remapped unconditionally. That approach, however, would effectively revert #70642.
Fixes https://github.com/rust-lang/rust/issues/87745.
cc `@cbeuw`
r? `@ghost`
Fix `option_if_let_else`
fixes: #5822fixes: #6737fixes: #7567
The inference from #6137 still exists so I'm not sure if this should be moved from the nursery. Before doing that though I'd almost want to see this split into two lints. One suggesting `map_or` and the other suggesting `map_or_else`.
`map_or_else` tends to have longer expressions for both branches so it doesn't end up much shorter than a match expression in practice. It also seems most people find it harder to read. `map_or` at least has the terseness benefit of being on one line most of the time, especially when the `None` branch is just a literal or path expression.
changelog: `break` and `continue` statments local to the would-be closure are allowed in `option_if_let_else`
changelog: don't lint in const contexts in `option_if_let_else`
changelog: don't lint when yield expressions are used in `option_if_let_else`
changelog: don't lint when the captures made by the would-be closure conflict with the other branch in `option_if_let_else`
changelog: don't lint when a field of a local is used when the type could be pontentially moved from in `option_if_let_else`
changelog: in some cases, don't lint when scrutinee expression conflicts with the captures of the would-be closure in `option_if_let_else`
Add module_style lint to style
changelog: Add new [`module_style`] style lint
This is a configurable (no mod file/mod file) lint that determines if `mod.rs` is used consistently or if `mod.rs` is never used (using the new mod layout).
Don't report function calls as unnecessary operation if used in array index
Attempts to fix: #7412
changelog: Don't report function calls used in indexing as unnecessary operation. [`unnecessary_operation`]
Add tests for disallowed_mod in ui-cargo test section
Use correct algorithm to determine if mod.rs is missing
Move to two lints and remove config option
Switch lint names so they read "warn on ..."
Emit the same help info for self_named_mod_file warnings
Bail when both lints are Allow
Reword help message for both module_style lints
Add new lint `negative_feature_names` and `redundant_feature_names`
Add new lint [`negative_feature_names`] to detect feature names with prefixes `no-` or `not-` and new lint [`redundant_feature_names`] to detect feature names with prefixes `use-`, `with-` or suffix `-support`
changelog: Add new lint [`negative_feature_names`] and [`redundant_feature_names`]
New lint `manual_split_once`
This is a WIP because it still needs to recognize more patterns. Currently handles:
```rust
s.splitn(2, ' ').next();
s.splitn(2, ' ').nth(0)
s.splitn(2, ' ').nth(1);
s.splitn(2, ' ').skip(0).next();
s.splitn(2, ' ').skip(1).next();
s.splitn(2, ' ').next_tuple(); // from itertools
// as well as `unwrap()` and `?` forms
```
Still to do:
```rust
let mut iter = s.splitn(2, ' ');
(iter.next().unwrap(), iter.next()?)
let mut iter = s.splitn(2, ' ');
let key = iter.next().unwrap();
let value = iter.next()?;
```
Suggestions on other patterns to check for would be useful. I've done a search on github for uses of `splitn`. Still have yet to actually look through the results.
There's also the question of whether the lint shouold trigger on all uses of `splitn` with two values, or only on recognized usages. The former could have false positives where it couldn't be replaced, but I'm not sure how common that would be.
changelog: Add lint `manual_split_once`
* `break` and `continue` statments local to the would-be closure are allowed
* don't lint in const contexts
* don't lint when yield expressions are used
* don't lint when the captures made by the would-be closure conflict with the other branch
* don't lint when a field of a local is used when the type could be pontentially moved from
* in some cases, don't lint when scrutinee expression conflicts with the captures of the would-be closure
Uplift the invalid_atomic_ordering lint from clippy to rustc
This is mostly just a rebase of https://github.com/rust-lang/rust/pull/79654; I've copy/pasted the text from that PR below.
r? `@lcnr` since you reviewed the last one, but feel free to reassign.
---
This is an implementation of https://github.com/rust-lang/compiler-team/issues/390.
As mentioned, in general this turns an unconditional runtime panic into a (compile time) lint failure. It has no false positives, and the only false negatives I'm aware of are if `Ordering` isn't specified directly and is comes from an argument/constant/whatever.
As a result of it having no false positives, and the alternative always being strictly wrong, it's on as deny by default. This seems right.
In the [zulip stream](https://rust-lang.zulipchat.com/#narrow/stream/233931-t-compiler.2Fmajor-changes/topic/Uplift.20the.20.60invalid_atomic_ordering.60.20lint.20from.20clippy/near/218483957) `@joshtriplett` suggested that lang team should FCP this before landing it. Perhaps libs team cares too?
---
Some notes on the code for reviewers / others below
## Changes from clippy
The code is changed from [the implementation in clippy](68cf94f6a6/clippy_lints/src/atomic_ordering.rs) in the following ways:
1. Uses `Symbols` and `rustc_diagnostic_item`s instead of string literals.
- It's possible I should have just invoked Symbol::intern for some of these instead? Seems better to use symbol, but it did require adding several.
2. The functions are moved to static methods inside the lint struct, as a way to namespace them.
- There's a lot of other code in that file — which I picked as the location for this lint because `@jyn514` told me that seemed reasonable.
3. Supports unstable AtomicU128/AtomicI128.
- I did this because it was almost easier to support them than not — not supporting them would have (ideally) required finding a way not to give them a `rustc_diagnostic_item`, which would have complicated an already big macro.
- These don't have tests since I wasn't sure if/how I should make tests conditional on whether or not the target has the atomic... This is to a certain extent an issue of 64bit atomics too, but 128-bit atomics are much less common. Regardless, the existing tests should be *more* than thorough enough here.
4. Minor changes like:
- grammar tweaks ("loads cannot have `Release` **and** `AcqRel` ordering" => "loads cannot have `Release` **or** `AcqRel` ordering")
- function renames (`match_ordering_def_path` => `matches_ordering_def_path`),
- avoiding clippy-specific helper methods that don't exist in rustc_lint and didn't seem worth adding for this case (for example `cx.struct_span_lint` vs clippy's `span_lint_and_help` helper).
## Potential issues
(This is just about the code in this PR, not conceptual issues with the lint or anything)
1. I'm not sure if I should have used a diagnostic item for `Ordering` and its variants (I couldn't figure out how really, so if I should do this some pointers would be appreciated).
- It seems possible that failing to do this might possibly mean there are more cases this lint would miss, but I don't really know how `match_def_path` works and if it has any pitfalls like that, so maybe not.
2. I *think* I deprecated the lint in clippy (CC `@flip1995` who asked to be notified about clippy changes in the future in [this comment](https://github.com/rust-lang/rust/pull/75671#issuecomment-718731659)) but I'm not sure if I need to do anything else there.
- I'm kind of hoping CI will catch if I missed anything, since `x.py test src/tools/clippy` fails with a lot of errors with and without my changes (and is probably a nonsense command regardless). Running `cargo test` from src/tools/clippy also fails with unrelated errors that seem like refactorings that didnt update clippy? So, honestly no clue.
3. I wasn't sure if the description/example I gave good. Hopefully it is. The example is less thorough than the one from clippy here: https://rust-lang.github.io/rust-clippy/master/index.html#invalid_atomic_ordering. Let me know if/how I should change it if it needs changing.
4. It pulls in the `if_chain` crate. This crate was already used in clippy, and seems like it's used elsewhere in rustc, but I'm willing to rewrite it to not use this if needed (I'd prefer not to, all things being equal).
- Deprecate clippy::invalid_atomic_ordering
- Use rustc_diagnostic_item for the orderings in the invalid_atomic_ordering lint
- Reduce code duplication
- Give up on making enum variants diagnostic items and just look for
`Ordering` instead
I ran into tons of trouble with this because apparently the change to
store HIR attrs in a side table also gave the DefIds of the
constructor instead of the variant itself. So I had to change
`matches_ordering` to also check the grandparent of the defid as well.
- Rename `atomic_ordering_x` symbols to just the name of the variant
- Fix typos in checks - there were a few places that said "may not be
Release" in the diagnostic but actually checked for SeqCst in the lint.
- Make constant items const
- Use fewer diagnostic items
- Only look at arguments after making sure the method matches
This prevents an ICE when there aren't enough arguments.
- Ignore trait methods
- Only check Ctors instead of going through `qpath_res`
The functions take values, so this couldn't ever be anything else.
- Add if_chain to allowed dependencies
- Fix grammar
- Remove unnecessary allow
Manual map 7413
fixes: #7413
This only fixes the specific problem from #7413, not the general case. The full fix requires interacting with the borrow checker to determine the lifetime of all the borrows made in the function. I'll open an issue about it later.
changelog: Don't suggest using `map` when the option is borrowed in the match, and also consumed in the arm.
changelog: Locals declared within the would-be closure will not prevent the closure from being suggested in `manual_map` and `map_entry`
