Fix issue #2907.
Update the "borrow box" lint to avoid recommending the following
conversion:
```
// Old
pub fn f(&mut Box<T>) {...}
// New
pub fn f(&mut T) {...}
```
Given a mutable reference to a box, functions may want to change
"which" object the Box is pointing at.
This change avoids recommending removing the "Box" parameter
for mutable references.
changelog: Don't trigger [`borrow_box`] lint on `&mut Box` references
Cleanup: `node_id` -> `hir_id`
This removes some more `node_id` terminology from Clippy and replaces one occurrence of `as_local_node_id` with `as_local_hir_id`, which should be doing the same for that particular case.
changelog: none
Don't trigger toplevel_ref_arg for `for` loops
The lint suggests turning `for ref x in 0..10 {` into `for ref x in let x = &0..10; {`.
---
changelog: none
Update the "borrow box" lint to avoid recommending the following
conversion:
```
// Old
pub fn f(&mut Box<T>) {...}
// New
pub fn f(&mut T) {...}
```
Given a mutable reference to a box, functions may want to change
"which" object the Box is pointing at.
This change avoids recommending removing the "Box" parameter
for mutable references.
Add lint named implicit_saturating_sub
Fixes: #5399
I've made a basic skeleton of the lint, would love more feedback on how to make it better.
changelog: Add lint [`implicit_saturating_sub`]
fix redundant_pattern_matching lint
- now it handles `while let` case (related to #5462)
- better suggestions in `if let` case
changelog: Fix suggestion in `redundant_pattern_matching` and also apply this lint to the `while let` case
add lint futures_not_send
changelog: add lint futures_not_send
fixes#5379
~Remark: one thing that can (should?) still be improved is to directly include the error message from the `Send` check so that the programmer stays in the flow. Currently, getting the actual error message requires a restructuring of the code to make the `Send` constraint explicit.~
It now shows all unmet constraints for allowing the Future to be Send.
Fixes issue #4892.
First contribution here 😊 ! Do not hesitate to correct me.
This PR is related to issue #4892 .
# Summary
```rust
-literal.method_call(args)
```
The main idea is to not trigger `clippy::precedence` when the method call is an odd function.
# Example
```rust
// should trigger lint
let _ = -1.0_f64.abs() //precedence of method call abs() and neg ('-') is ambiguous
// should not trigger lint
let _ = -1.0_f64.sin() // sin is an odd function => -sin(x) = sin(-x)
```
# Theory
Rust allows following literals:
- char
- string
- integers
- floats
- byte
- bool
Only integers/floats implements the relevant `std::ops::Neg`.
Following odd functions are implemented on i[8-128] and/or f[32-64]:
- `asin`
- `asinh`
- `atan`
- `atanh`
- `cbrt`
- `fract`
- `round`
- `signum`
- `sin`
- `sinh`
- `tan`
- `tanh `
- `to_degrees`
- `to_radians`
# Implementation
As suggested by `flip1995` in [comment](https://github.com/rust-lang/rust-clippy/issues/4892#issuecomment-568249683), this PR add a whitelist of odd functions and compare method call to the the whitelist before triggering lint.
changelog: Don't trigger [`clippy::precedence`] on odd functions.
question_mark: don't add `as_ref()` for a call expression
If a call returns a `!Copy` value, it does so regardless of whether `as_ref()` is added. For example, `foo.into_option().as_ref()?` can be simplified to `foo.into_option()?`.
---
changelog: Improved `question_mark` lint suggestion so that it doesn't add redundant `as_ref()`
unit_arg suggestion is not really machine-applicable
This lint suggests replacing any code returining `()`-type with `()` literal, which can change the behavior. e.g. `Ok(do_something())` -> `OK(())`.
---
changelog: none
Do not lint in macros for match lints
Don't lint in macros for match lints, more precisely in `check_pat` and `check_local` where it was not the case.
changelog: none
fixes: #5362
[fix] Minor typo in GH Actions 'clippy_dev'
changelog: minor: fixes file name in Github Actions workflow `clippy_dev.yml`
I was looking for something else and ended up on that typo, just thought I would fix it.
add --fix support to `cargo-clippy`
Prior to this we had started work on integrating clippy as a subcommand directly into cargo in the form of `cargo clippy-preview` and `cargo fix --clippy`. In the course of that work it was decided that the best approach would be to strictly add the features clippy needed to cargo in order to insert `clippy-driver` only for workspace crates. This was accomplished by adding a `RUSTC_WORKSPACE_WRAPPER` env variable to cargo that will override the normal `RUSTC_WRAPPER` when both are present and the current crate is a workspace crate.
This change adds support to clippy to use this by setting the `RUSTC_WORKSPACE_WRAPPER` env variable instead `RUSTC_WRAPPER` and by detecting `--fix` as an arg and swapping out the `check` cargo command for `fix` when it is present.
WIP, here are the current issues that I still need to resolve
- [x] Detect if we're running on nightly rust
- [x] Set `RUSTC_WORKSPACE_WRAPPER` on nightly, and `RUSTC_WRAPPER` on stable
- [x] Error out on stable when `--fix` is specified, because stable currently hasn't landed the PR for `RUSTC_WORKSPACE_WRAPPER` so if we set this it just runs check and silently fails
- [ ] Update the help text
- [ ] The current plan is to shell out to `cargo check --help` and then postprocess the output to mention clippy instead of check where appropriate and to add the extra info about `--fix` and the `-- -A lint` options.
- [x] tests?
changelog: add `--fix` arg to `cargo-clippy`
