[`map_entry`]: Check insert expression for map use
The lint makes sure that the map is not used (borrowed) before the call to `insert`. Since the lint creates a mutable borrow on the map with the `Entry`, it wouldn't be possible to replace such code with `Entry`. However, expressions up to the `insert` call are checked, but not expressions for the arguments of the `insert` call itself. This commit fixes that.
Fixes#11935
----
changelog: [`map_entry`]: Fix false positive when borrowing the map in the `insert` call
If the whole cast expression is a unary expression (`(*x as T)`) or an
addressof expression (`(&x as T)`), then not surrounding the suggestion
into a block risks us changing the precedence of operators if the cast
expression is followed by an operation with higher precedence than the
unary operator (`(*x as T).foo()` would become `*x.foo()`, which changes
what the `*` applies on).
The same is true if the expression encompassing the cast expression is a
unary expression or an addressof expression.
The lint supports the latter case, but missed the former one. This PR
fixes that.
Fixes#11968
The lint makes sure that the map is not used (borrowed) before the call
to `insert`. Since the lint creates a mutable borrow on the map with the
`Entry`, it wouldn't be possible to replace such code with `Entry`.
However, expressions up to the `insert` call are checked, but not
expressions for the arguments of the `insert` call itself. This commit
fixes that.
Fixes#11935
Fix sign-handling bugs and false negatives in `cast_sign_loss`
**Note: anyone should feel free to move this PR forward, I might not see notifications from reviewers.**
changelog: [`cast_sign_loss`]: Fix sign-handling bugs and false negatives
This PR fixes some arithmetic bugs and false negatives in PR #11883 (and maybe earlier PRs).
Cc `@J-ZhengLi`
I haven't updated the tests yet. I was hoping for some initial feedback before adding tests to cover the cases listed below.
Here are the issues I've attempted to fix:
#### `abs()` can return a negative value in release builds
Example:
```rust
i32::MIN.abs()
```
https://play.rust-lang.org/?version=stable&mode=release&edition=2021&gist=022d200f9ef6ee72f629c0c9c1af11b8
Docs: https://doc.rust-lang.org/std/primitive.i32.html#method.abs
Other overflows that produce negative values could cause false negatives (and underflows could produce false positives), but they're harder to detect.
#### Values with uncertain signs can be positive or negative
Any number of values with uncertain signs cause the whole expression to have an uncertain sign, because an uncertain sign can be positive or negative.
Example (from UI tests):
```rust
fn main() {
foo(a: i32, b: i32, c: i32) -> u32 {
(a * b * c * c) as u32
//~^ ERROR: casting `i32` to `u32` may lose the sign of the value
}
println!("{}", foo(1, -1, 1));
}
```
https://play.rust-lang.org/?version=nightly&mode=debug&edition=2021&gist=165d2e2676ee8343b1b9fe60db32aadd
#### Handle `expect()` the same way as `unwrap()`
Since we're ignoring `unwrap()` we might as well do the same with `expect()`.
This doesn't seem to have tests but I'm happy to add some like `Some(existing_test).unwrap() as u32`.
#### A negative base to an odd exponent is guaranteed to be negative
An integer `pow()`'s sign is only uncertain when its operants are uncertain. (Ignoring overflow.)
Example:
```rust
((-2_i32).pow(3) * -2) as u32
```
This offsets some of the false positives created by one or more uncertain signs producing an uncertain sign. (Rather than just an odd number of uncertain signs.)
#### Both sides of a multiply or divide should be peeled recursively
I'm not sure why the lhs was peeled recursively, and the rhs was left intact. But the sign of any sequence of multiplies and divides is determined by the signs of its operands. (Ignoring overflow.)
I'm not sure what to use as an example here, because most expressions I want to use are const-evaluable.
But if `p()` is [a non-const function that returns a positive value](https://doc.rust-lang.org/std/primitive.i32.html#method.isqrt), and if the lint handles unary negation, these should all lint:
```rust
fn peel_all(x: i32) {
(-p(x) * -p(x) * -p(x)) as u32;
((-p(x) * -p(x)) * -p(x)) as u32;
(-p(x) * (-p(x) * -p(x))) as u32;
}
```
#### The right hand side of a Rem doesn't change the sign
Unlike Mul and Div,
> Given remainder = dividend % divisor, the remainder will have the same sign as the dividend.
https://doc.rust-lang.org/reference/expressions/operator-expr.html#arithmetic-and-logical-binary-operators
I'm not sure what to use as an example here, because most expressions I want to use are const-evaluable.
But if `p()` is [a non-const function that returns a positive value](https://doc.rust-lang.org/std/primitive.i32.html#method.isqrt), and if the lint handles unary negation, only the first six expressions should lint.
The expressions that start with a constant should lint (or not lint) regardless of whether the lint supports `p()` or unary negation, because only the dividend's sign matters.
Example:
```rust
fn rem_lhs(x: i32) {
(-p(x) % -1) as u32;
(-p(x) % 1) as u32;
(-1 % -p(x)) as u32;
(-1 % p(x)) as u32;
(-1 % -x) as u32;
(-1 % x) as u32;
// These shouldn't lint:
(p(x) % -1) as u32;
(p(x) % 1) as u32;
(1 % -p(x)) as u32;
(1 % p(x)) as u32;
(1 % -x) as u32;
(1 % x) as u32;
}
```
#### There's no need to bail on other expressions
When peeling, any other operators or expressions can be left intact and sent to the constant evaluator.
If these expressions can be evaluated, this offsets some of the false positives created by one or more uncertain signs producing an uncertain sign. If not, they end up marked as having uncertain sign.
[`read_line_without_trim`]: detect string literal comparison and `.ends_with()` calls
This lint now also realizes that a comparison like `s == "foo"` and calls such as `s.ends_with("foo")` will fail if `s` was initialized by a call to `Stdin::read_line` (because of the trailing newline).
changelog: [`read_line_without_trim`]: detect string literal comparison and `.ends_with()` calls
r? `@giraffate` assigning you because you reviewed #10970 that added this lint, so this is kinda a followup PR ^^
fix suggestion error in [`useless_vec`]
fixes: #12101
---
changelog: fix suggestion error in [`useless_vec`]
r+ `@matthiaskrgr` since they opened the issue?
Empty docs
Fixes https://github.com/rust-lang/rust-clippy/issues/9931
changelog: [`empty_doc`]: Detects documentation that is empty.
changelog: Doc comment lints now trigger for struct field and enum variant documentation
When encountering code such as:
```
Box::new(outer::Inner::default())
```
clippy would suggest replacing with `Box::<Inner>::default()`, dropping
the `outer::` segment. This behavior is incorrect and that commit fixes
it.
What it does is it checks the contents of the `Box::new` and, if it is
of the form `A::B::default`, does a text replacement, inserting `A::B`
in the `Box`'s quickfix generic list.
If the source does not match that pattern (including `Vec::from(..)`
or other `T::new()` calls), we then fallback to the original code.
Fixes#11927
Look for `implied_bounds_in_impls` in more positions
With this, we lint `impl Trait` implied bounds in more positions:
- Type alias impl trait
- Associated type position impl trait
- Argument position impl trait
- these are not opaque types, but instead are desugared to `where` clauses, so we need extra logic for finding them (`check_generics`), however the rest of the logic is the same
Before this, we'd only lint RPIT `impl Trait`s.
"Hide whitespaces" and reviewing commits individually might make this easier
changelog: [`implied_bounds_in_impls`]: start linting implied bounds in APIT, ATPIT, TAIT
Merge `single_call_fn` post-crate visitor into lint pass
The `single_call_fn` lint worked by first collecting a list of function definitions in the lint pass, then populating the list of uses for each function in a second visitor after the crate is checked.
Doing another pass through the crate shouldn't be needed, and we should be able to do it in the same lint pass, by looking for path references to functions only and then processing them post-crate.
Other changes:
- `FxHashMap` -> `FxIndexMap` so that we emit warnings in a consistent order, as we see them (making the diff a bit confusing to look at, because warnings were moved around)
- no longer storing a `Vec<Span>` per function: an enum representing "seen once" or "seen more than once" should be enough (only the first element is used later)
- "used here" help is now a note
I also noticed that it lints on trait methods with a default implementation, but not on regular trait methods without a body (because that's what `check_fn` does). I'm not sure if that's useful though, maybe we shouldn't lint trait methods at all? It's not like you can avoid it sometimes (but then again it's a restriction lint). Either way, I left the behavior where it was before so that there are no functional changes made in this PR and it's purely a refactor. I can change it though
changelog: none
FIX(12243): redundant_guards
Fixed#12243
changelog: Fix[`redundant_guards`]
I have made a correction so that no warning does appear when y.is_empty() is used within a constant function as follows.
```rust
pub const fn const_fn(x: &str) {
match x {
// Shouldn't lint.
y if y.is_empty() => {},
_ => {},
}
}
```
Add new `unnecessary_get_then_check` lint
No issue linked to this as far as I can see. It's a lint I discovered that could be added when I worked on another lint.
r? `@llogiq`
changelog: Add new `unnecessary_get_then_check` lint
A warning is now suppressed when "<str_va> if <str_var>.is_empty" is used in a constant function.
FIX: instead of clippy_util::in_const
FIX: Merged `redundant_guards_const_fn.rs` into `redundant_guards.rs`.
Extend `unnecessary_to_owned` to handle `Borrow` trait in map types
Fixes https://github.com/rust-lang/rust-clippy/issues/8088.
Alternative to #12315.
r? `@y21`
changelog: Extend `unnecessary_to_owned` to handle `Borrow` trait in map types
----
UPDATE: add async block into test.
FIX: no_effect
Fixed asynchronous function parameter names with underscores so that warnings are not displayed when underscores are added to parameter names
ADD: test case
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`.
This mostly works well, and eliminates a couple of delayed bugs.
One annoying thing is that we should really also add an
`ErrorGuaranteed` to `proc_macro::bridge::LitKind::Err`. But that's
difficult because `proc_macro` doesn't have access to `ErrorGuaranteed`,
so we have to fake it.
Ignore imported items in `min_ident_chars`
Suppress the `min_ident_chars` warning for items whose name we cannot control. Do not warn for `use a::b`, but warn for `use a::b as c`, since `c` is a local identifier.
Fixes#12232
---
*Please write a short comment explaining your change (or "none" for internal only changes)*
changelog: [`min_ident_chars`]: Do not warn on non-local identifiers
[`incompatible_msrv`]: allow expressions that come from desugaring
Fixes#12273
changelog: [`incompatible_msrv`]: don't lint on the `IntoFuture::into_future` call desugared by `.await`
Suppress the `min_ident_chars` warning for items whose name we cannot
control. Do not warn for `use a::b`, but warn for `use a::b as c`, since
`c` is a local identifier.
Fixes#12232
[`implied_bounds_in_impls`]: avoid linting on overlapping associated tys
Fixes#11880
Before this change, we were simply ignoring associated types (except for suggestion purposes), because of an incorrect assumption (see the comment that I also removed).
For something like
```rs
trait X { type T; }
trait Y: X { type T; }
// Can't constrain `X::T` through `Y`
fn f() -> impl X<T = i32> + Y<T = u32> { ... }
```
We now avoid linting if the implied bound (`X<T = i32>`) "names" associated types that also exists in the implying trait (`trait Y`). Here that would be the case.
But if we only wrote `impl X + Y<T = u32>` then that's ok because `X::T` was never constrained in the first place.
I haven't really thought about how this interacts with GATs, but I think it's fine. Fine as in, it might create false negatives, but hopefully no false positives.
(The diff is slightly annoying because of formatting things. Really the only thing that changed in the if chain is extracting the `implied_by_def_id` which is needed for getting associated types from the trait, and of course actually checking for overlap)
cc `@Jarcho` ? idk if you want to review this or not. I assume you looked into this code a bit to find this bug.
changelog: [`implied_bounds_in_impls`]: avoid linting when associated type from supertrait can't be constrained through the implying trait bound
[`mem_replace_with_default`] No longer triggers on unused expression
changelog:[`mem_replace_with_default`]: No longer triggers on unused expression
Change [`mem_replace_with_default`] to not trigger on unused expression because the lint from `#[must_use]` handle this case better.
fixes: #5586
Minor refactor format-args
* Move all linting logic into a single format implementations struct
This should help with the future format-args improvements.
**NOTE TO REVIEWERS**: use "hide whitespace" in the github diff -- most of the code has shifted, but relatively low number of lines actually modified.
Followig up from #12274
r? `@xFrednet`
---
changelog: none
Fix async closures in CTFE
First commit renames `is_coroutine_or_closure` into `is_closure_like`, because `is_coroutine_or_closure_or_coroutine_closure` seems confusing and long.
Second commit fixes some forgotten cases where we want to handle `TyKind::CoroutineClosure` the same as closures and coroutines.
The test exercises the change to `ValidityVisitor::aggregate_field_path_elem` which is the source of #120946, but not the change to `UsedParamsNeedSubstVisitor`, though I feel like it's not that big of a deal. Let me know if you'd like for me to look into constructing a test for the latter, though I have no idea what it'd look like (we can't assert against `TooGeneric` anywhere?).
Fixes#120946
r? oli-obk cc ``@RalfJung``
fix: ICE when array index exceeds usize
fixes#12253
This PR fixes ICE in `indexing_slicing` as it panics when the index of the array exceeds `usize`.
changelog: none
Don't allow derive macros to silence `disallowed_macros`
fixes#12254
The implementation is a bit of a hack, but "works". A derive expanding to another derive won't work properly, but we shouldn't be linting those anyways.
changelog: `disallowed_macros`: Don't allow derive macros to silence their own expansion
Assert that params with the same *index* have the same *name*
Found this bug when trying to build libcore with the new solver, since it will canonicalize two params with the same index into *different* placeholders if those params differ by name.
Minor refactor format-impls
Move all linting logic into a single format implementations struct
This should help with the future format-args improvements.
TODO: do the same with format_args.rs, perhaps in the same PR
**NOTE TO REVIEWERS**: use "hide whitespace" in the github diff -- most of the code has shifted, but relatively low number of lines actually modified.
changelog: none
Refactor `implied_bounds_in_impls` lint
Some refactors in `implied_bounds_in_impls` that I wanted to make while working on something else in that file, but I found them "large" enough that I didn't want them in the same PR and instead wanted them reviewed separately (since itd just be distracting).
This just splits up the two phases of "collect all the supertraits from each of the `impl Trait` bounds" and "find those `impl Trait` bounds that are mentioned in one of the previously-collected supertraits" into separate functions. Before, this was all in a single function.
Reviewing it commit by commit might make it easier. I can squash it down later.
changelog: none
stop linting [`blocks_in_conditions`] on `match` with weird attr macro case
should fixes: #12016
---
changelog: [`blocks_in_conditions`] - fix FP on `match` with weird attr macro
This might not be the best solution, as the root cause (i think?) is the `span` of block was incorrectly given by the compiler?
I'm open to better solutions
* Move all linting logic into a single format implementations struct
This should help with the future format-args improvements.
TODO: do the same with format_args.rs, perhaps in the same PR
A lot of cases of the "noise" cases of `similar_names` come from two
idents with a different first letter, which is easy enough to
differentiate visually but causes this lint to be raised.
Do not raise the lint in these cases, as long as the first character
does not have a lookalike.
Link: https://github.com/rust-lang/rust-clippy/issues/10926
Fix issue #12034: add autofixes for unnecessary_fallible_conversions
fixes#12034
Currently, the `unnecessary_fallible_conversions` lint was capable of autofixing expressions like `0i32.try_into().unwrap()`. However, it couldn't autofix expressions in the form of `i64::try_from(0i32).unwrap()` or `<i64 as TryFrom<i32>>::try_from(0).unwrap()`.
This pull request extends the functionality to correctly autofix these latter forms as well.
changelog: [`unnecessary_fallible_conversions`]: Add autofixes for more forms
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````
Rollup of 9 pull requests
Successful merges:
- #119592 (resolve: Unload speculatively resolved crates before freezing cstore)
- #120103 (Make it so that async-fn-in-trait is compatible with a concrete future in implementation)
- #120206 (hir: Make sure all `HirId`s have corresponding HIR `Node`s)
- #120214 (match lowering: consistently lower bindings deepest-first)
- #120688 (GVN: also turn moves into copies with projections)
- #120702 (docs: also check the inline stmt during redundant link check)
- #120727 (exhaustiveness: Prefer "`0..MAX` not covered" to "`_` not covered")
- #120734 (Add `SubdiagnosticMessageOp` as a trait alias.)
- #120739 (improve pretty printing for associated items in trait objects)
r? `@ghost`
`@rustbot` modify labels: rollup
[`unconditional_recursion`]: compare by `Ty`s instead of `DefId`s
Fixes#12154Fixes#12181 (this was later edited in, so the rest of the description refers to the first linked issue)
Before this change, the lint would work with `DefId`s and use those to compare types. This PR changes it to compare types directly. It fixes the linked issue, but also other false positives I found in a lintcheck run. For example, one of the issues is that some types don't have `DefId`s (primitives, references, etc., leading to possible FNs), and the helper function used to extract a `DefId` didn't handle type parameters.
Another issue was that the lint would use `.peel_refs()` in a few places where that could lead to false positives (one such FP was in the `http` crate). See the doc comment on one of the added functions and also the test case for what I mean.
The code in the linked issue was linted because the receiver type is `T` (a `ty::Param`), which was not handled in `get_ty_def_id` and returned `None`, so this wouldn't actually *get* to comparing `self_arg != ty_id` here, and skip the early-return:
70573af31e/clippy_lints/src/unconditional_recursion.rs (L171-L178)
This alone could be fixed by doing something like `&& get_ty_def_id(ty).map_or(true, |ty_id)| self_arg != ty_id)`, but we don't really need to work with `DefId`s in the first place, I don't think.
changelog: [`unconditional_recursion`]: avoid linting when the other comparison type is a type parameter
Fix false positive in `redundant_type_annotations` lint
This PR changes the `redundant_type_annotations` lint to allow slice type annotations (i.e., `&[u8]`) for byte string literals. It will still consider _array_ type annotations (i.e., `&[u8; 4]`) as redundant. The reasoning behind this is that the type of byte string literals is by default a reference to an array, but, by using a type annotation, you can force it to be a slice. For example:
```rust
let a: &[u8; 4] = b"test";
let b: &[u8] = b"test";
```
Now, the type annotation for `a` will still be linted (as it is still redundant), but the type annotation for `b` will not.
Fixes#12212.
changelog: [`redundant_type_annotations`]: Fix false positive with byte string literals
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.
[`redundant_locals`]: take by-value closure captures into account
Fixes#12225
The same problem in the linked issue can happen to regular closures too, and conveniently async blocks are closures in the HIR so fixing closures will fix async blocks as well.
changelog: [`redundant_locals`]: avoid linting when redefined variable is captured by-value
