Corrects the float_equality_without_abs lint
Fixes an issue in the `float_equality_without_abs` lint. The lint suggestion was configured in a way that it lints the whole error and not just the subtraction part. In the current configuration the lint would suggest to change the expression in a wrong way, e.g.
```rust
let _ = (a - b) < f32::EPSILON; // before
let _ = (a - b).abs(); // after
```
This was dicovered by @flip1995. (See discussion of PR #5952).
Also the suggestion is now formatted via `utils::sugg`.
changelog: none
Fix `let_and_return` bad suggestion
Add a cast to the suggestion when the return expression has adjustments.
These adjustments are lost when the suggestion is applied.
This is similar to the problem in issue #4437.
Closes#5729
changelog: Fix `let_and_return` bad suggestion
Fix fp in `borrow_interior_mutable_const`
fixes#5796
changelog: fix false positive in `borrow_interior_mutable_const` when referencing a field behind a pointer.
Fix incorrect suggestion when `clone_on_ref_ptr` is triggered in macros
In the lint `clone_on_ref_ptr`, if the `span` is in a macro, don't expand it for suggestion.
Fixes: #2076
changelog: none
r? @ebroto
Re-enable len_zero for ranges now that `is_empty` is stable on them
Fixes#5956
Completed stabilization PR: https://github.com/rust-lang/rust/pull/75132
changelog: len_zero: re-enable linting ranges now that range_is_empty is stable
The fmt test will cause clippy dev to be compiled and run. The clippy dev
dependencies are currently stored at `target/debug/deps` when this happens.
This location sometimes causes conflicts with ui tests which result in the
build error "thread 'compile_test' panicked at 'Found multiple rlibs for crate
`regex` ...".
This commit forces the clippy_dev dependencies to be stored under
`clippy_dev/target/` which seems to resolve this problem.
stable_sort_primitive: print the type that is being sorted in the lint message
changelog: stable_sort_primitive: print the type that is being sorted in the lint message
Fix cargo dev new_lint for late lint passes
Since 30c046ede4, `LateLintPass` has only one lifetime parameter.
I'm not sure how to easily test this, I think adding this kind of tests would be probably part of #5394
changelog: none
Add a cast to the suggestion when the return expression has adjustments.
These adjustments are lost when the suggestion is applied.
This is similar to the problem in issue #4437.
Closes#5729
Don't lint if it has always inline attribute
Don't trigger the lint `trivially_copy_pass_by_ref` if it has `#[inline(always)]` attribute.
Note: I am not particularly familiar with `inline` impacts, so I implemented this the way that if only `#[inline]` attribute is here (without `always`), the lint will still trigger. Also, it will still trigger if it has `#[inline(never)]`.
Just tell me if it sounds too much conservative.
Fixes: #5876
changelog: none
Widen understanding of prelude import
Prelude imports are exempt from wildcard import warnings. Until now only
imports of the form
```
use ...::prelude::*;
```
were considered. This change makes it so that the segment `prelude` can
show up anywhere, for instance:
```
use ...::prelude::v1::*;
```
Fixes#5917
changelog: Allow `prelude` to appear in any segment of the import path in [`wildcard_imports`]
Prelude imports are exempt from wildcard import warnings. Until now only
imports of the form
```
use ...::prelude::*;
```
were considered. This change makes it so that the segment `prelude` can
show up anywhere, for instance:
```
use ...::prelude::v1::*;
```
Fixes#5917
