Use `approx_ty_size` for `large_enum_variant`
This builds upon #9373 to use the approximate size of each variant for `large_enum_variant`. This allows us to lint in situations where an `enum` contains generics but is still guaranteed to have a large variant on an at-least basis, e.g. with `(T, [u8; 512])`.
* I've changed the wording from "is ... bytes" to "contains at least" because
* the size is now an approximate lower bound (e.g. `512` in the example above). The actual size is larger due to `T`, including due to `T`'s memory layout.
* the discriminant is not taken into account in the message. This comes up with variants like `A(T)`, which are "is at least 0 bytes" otherwise, which may be misleading.
* If the second-largest variant has no fields, there is a special case "carries no data" instead of "is at least 0 bytes".
* A variant like `A(T)` is "at least 0 bytes", which is technically true, yet we don't distinguish between "indeterminate" and truly "ZST".
* The generics-tests that were there before now lint while they didn't lint before. AFAICS this is correct.
I guess the above is correct-ish. However, I use the `SubstsRef` that I got via `cx.tcx.type_of(item.def_id)` to solve for generics in the variants. Is this even applicable, since we start from an - [ ] `ItemKind`?
changelog: none
Fix `mut_mutex_lock` when Mutex is behind immutable deref
I *think* the problem here is the `if let ty::Ref(_, _, Mutability::Mut) = cx.typeck_results().expr_ty(recv).kind()` line tries to check if the `Mutex` can be mutably borrowed (there already is a test for `Arc<Mutex<_>>`), but gets bamboozled by the `&mut Arc` indirection. And I *think* checking the deref-adjustment to filter immutable-adjust (the deref through the `Arc`, starting from `&mut Arc`) is the correct fix.
Fixes#9415
changelog: Fix `mut_mutex_lock` when Mutex is behind immutable deref
Don't use `hir_ty_to_ty` in `result_large_err`
fixes#9414
This occurs starting with 2022-09-01. I checked that this does fix the ICE on rust-lang/rust@9353538. Not sure which pr caused the late-bound region to leak through `hir_ty_to_ty`.
changelog: None
Fix `suboptimal_float` not linting on `{const}.powf({const})`
There used to be an early return if the receiver was an effective const but the method was not linted, not taking into account later cases where the receiver and the arguments are both effective consts for different methods. Removed the early return.
Fixes#9402Fixes#9201
changelog: Fix `suboptimal_flops`, `imprecise_flops` not linting on `{const}.powf({const})` et al
Fix the emission order of `trait_duplication_in_bounds`
Makes the lint emit in source order rather than whatever order the hash map happens to be in. This is currently blocking the sync into rustc.
changelog: None
Initial implementation `result_large_err`
This is a shot at #6560, #4652, and #3884. The lint checks for `Result` being returned from functions/methods where the `Err` variant is larger than a configurable threshold (the default of which is 128 bytes). There has been some discussion around this, which I'll try to quickly summarize:
* A large `Err`-variant may force an equally large `Result` if `Err` is actually bigger than `Ok`.
* There is a cost involved in large `Result`, as LLVM may choose to `memcpy` them around above a certain size.
* We usually expect the `Err` variant to be seldomly used, but pay the cost every time.
* `Result` returned from library code has a high chance of bubbling up the call stack, getting stuffed into `MyLibError { IoError(std::io::Error), ParseError(parselib::Error), ...}`, exacerbating the problem.
This PR deliberately does not take into account comparing the `Ok` to the `Err` variant (e.g. a ratio, or one being larger than the other). Rather we choose an absolute threshold for `Err`'s size, above which we warn. The reason for this is that `Err`s probably get `map_err`'ed further up the call stack, and we can't draw conclusions from the ratio at the point where the `Result` is returned. A relative threshold would also be less predictable, while not accounting for the cost of LLVM being forced to generate less efficient code if the `Err`-variant is _large_ in absolute terms.
We lint private functions as well as public functions, as the perf-cost applies to in-crate code as well.
In order to account for type-parameters, I conjured up `fn approx_ty_size`. The function relies on `LateContext::layout_of` to compute the actual size, and in case of failure (e.g. due to generics) tries to come up with an "at least size". In the latter case, the size of obviously wrong, but the inspected size certainly can't be smaller than that. Please give the approach a heavy dose of review, as I'm not actually familiar with the type-system at all (read: I have no idea what I'm doing).
The approach does, however flimsy it is, allow us to successfully lint situations like
```rust
pub union UnionError<T: Copy> {
_maybe: T,
_or_perhaps_even: (T, [u8; 512]),
}
// We know `UnionError<T>` will be at least 512 bytes, no matter what `T` is
pub fn param_large_union<T: Copy>() -> Result<(), UnionError<T>> {
Ok(())
}
```
I've given some refactoring to `functions/result_unit_err.rs` to re-use some bits. This is also the groundwork for #6409
The default threshold is 128 because of https://github.com/rust-lang/rust-clippy/issues/4652#issue-505670554
`lintcheck` does not trigger this lint for a threshold of 128. It does warn for 64, though.
The suggestion currently is the following, which is just a placeholder for discussion to be had. I did have the computed size in a `span_label`. However, that might cause both ui-tests here and lints elsewhere to become flaky wrt to their output (as the size is platform dependent).
```
error: the `Err`-variant returned via this `Result` is very large
--> $DIR/result_large_err.rs:36:34
|
LL | pub fn param_large_error<R>() -> Result<(), (u128, R, FullyDefinedLargeError)> {
| ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ The `Err` variant is unusually large, at least 128 bytes
```
changelog: Add [`result_large_err`] lint
Fix missing parens in `suboptimal_flops` suggestion
Fixes#9391. The problem is simple enough, I didn't check if the same problem occurs elsewhere, though.
changelog: fix missing parenthesis in `suboptimal_flops` suggestion
New lint: Raw slice pointer cast
Adds a lint to check for a raw slice being created and cast back to a pointer, suggesting `ptr::slice_from_raw_parts`, to identify UB such as https://github.com/SimonSapin/rust-typed-arena/pull/54.
```
changelog: [`cast_slice_from_raw_parts`]: Add lint to check for `slice::from_raw_parts(.., ..) as *const _`
```
Don't lint literal `None` from expansion
This addresses https://github.com/rust-lang/rust-clippy/pull/9288#issuecomment-1229398524: If the literal `None` is from expansion, we never lint. This is correct because e.g. replacing the call to `option_env!` with whatever that macro expanded to at the time of linting is certainly wrong.
changelog: Don't lint [`partialeq_to_none`] for macro-expansions
Ignore `match_like_matches_macro` when there is comment
Closes#9164
changelog: [`match_like_matches_macro`] is ignored when there is some comment inside the match block.
Also add `span_contains_comment` util to check if given span contains comments.
Implemented `suspicious_to_owned` lint to check if `to_owned` is called on a `Cow`
changelog: Add lint ``[`suspicious_to_owned`]``
-----------------
Hi,
posting this unsolicited PR as I've been burned by this issue :)
Being unsolicited, feel free to reject it or reassign a different lint level etc.
This lint checks whether `to_owned` is called on `Cow<'_, _>`. This is done because `to_owned` is very similarly named to `into_owned`, but the effect of calling those two methods is completely different (one makes the `Cow::Borrowed` into a `Cow::Owned`, the other just clones the `Cow`). If the cow is then passed to code for which the type is not checked (e.g. generics, closures, etc.) it might slip through and if the cow data is coming from an unsafe context there is the potential for accidentally cause undefined behavior.
Even if not falling into this painful case, there's really no reason to call `to_owned` on a `Cow` other than confusing people reading the code: either `into_owned` or `clone` should be called.
Note that this overlaps perfectly with `implicit_clone` as a warning, but `implicit_clone` is classified pedantic (while the consequences for `Cow` might be of a wider blast radius than just pedantry); given the overlap, I set-up the lint so that if `suspicious_to_owned` triggers `implicit_clone` will not trigger. I'm not 100% sure this is done in the correct way (I tried to copy what other lints were doing) so please provide feedback on it if it isn't.
### Checklist
- \[x] Followed [lint naming conventions][lint_naming]
- \[x] Added passing UI tests (including committed `.stderr` file)
- \[x] `cargo test` passes locally
- \[x] Executed `cargo dev update_lints`
- \[x] Added lint documentation
- \[x] Run `cargo dev fmt`
Don't lint `needless_return` if `return` has attrs
Fixes#9361
The lint used to have a mechanic to allow `cfg`-attrs on naked `return`-statements. This was well-intentioned, yet we can have any kind of attribute, e.g. `allow`, `expect` or even custom `derive`. So the mechanic was simply removed. We now never lint on a naked `return`-statement that has attributes on it.
Turns out that the ui-test had a Catch22 in it: In `check_expect()` the `#[expect(clippy::needless_return)]` is an attribute on the `return` statement that can and will be rustfixed away without side effects. But any other attribute would also have been removed, which is what #9361 is about. The test proved the wrong thing. Removed the test, the body is tested elsewhere as well.
changelog: Ignore [`needless_return`] on `return`s with attrs
This is done because `to_owned` is very similarly named to `into_owned`, but the
effect of calling those two methods is completely different. This creates
confusion (stemming from the ambiguity of the 'owned' term in the context of
`Cow`s) and might not be what the writer intended.
new lint
This fixes#6576
If you added a new lint, here's a checklist for things that will be
checked during review or continuous integration.
- \[x] Followed [lint naming conventions][lint_naming]
- \[x] Added passing UI tests (including committed `.stderr` file)
- \[x] `cargo test` passes locally
- \[x] Executed `cargo dev update_lints`
- \[x] Added lint documentation
- \[x] Run `cargo dev fmt`
---
changelog: add [`multi_assignments`] lint
