| Age | Commit message (Collapse) | Author | Lines |
|---|
|
Handle another negated literal in `eat_token_lit`.
Extends the change from #139653, which was on expressions, to literals.
Fixes #140098.
r? ``@petrochenkov``
|
|
Shourya742:2025-04-21-auto-diff-fails-on-impl-block, r=ZuseZ4
Fix auto diff failing on inherent impl blocks
closes: #139557
r? ``@ZuseZ4``
|
|
improve diagnostic for raw pointer field access with ->
This PR enhances the error messages emitted by the Rust compiler when users attempt to use the `->` operator for field access on raw pointers or when dereferencing is needed. The changes aim to provide clearer guidance, by suggesting the correct use of the `.` operator and explicit dereferencing.
**Before:**
```
help: `xs` is a raw pointer; try dereferencing it
|
LL | (*xs)->count += 1;
| ++ +
```
**Now:**
```
help: use `.` on a dereferenced raw pointer instead
|
LL - xs->count += 1;
LL + (*xs).count += 1;
|
```
I added extra clarification in the message. Since this error occurs in the parser, we can't be certain that the type is a raw pointer. That's why the message includes only a small note in brackets. (In contrast, the message above is emitted in HIR, where we *can* check whether it's a raw pointer.)
**Before:**
```
--> main.rs:11:11
|
11 | xs->count += 1;
| ^^
|
= help: the . operator will dereference the value if needed
```
**After:**
```
--> main.rs:11:11
|
11 | xs->count += 1;
| ^^
|
= help: the `.` operator will automatically dereference the value, except if the value is a raw pointer
```
|
|
Add tests for two untested cases of placeholder relations
During work on #130227, I discovered several situations not covered by any previously existing UI test. This commit introudces tests to cover that.
r? lcnr
|
|
During work on #130227, I discovered several situations not covered by any
previously existing UI test. This commit introudces tests to cover that.
|
|
Stabilize let chains in the 2024 edition
# Stabilization report
This proposes the stabilization of `let_chains` ([tracking issue], [RFC 2497]) in the [2024 edition] of Rust.
[tracking issue]: https://github.com/rust-lang/rust/issues/53667
[RFC 2497]: https://github.com/rust-lang/rfcs/pull/2497
[2024 edition]: https://doc.rust-lang.org/nightly/edition-guide/rust-2024/index.html
## What is being stabilized
The ability to `&&`-chain `let` statements inside `if` and `while` is being stabilized, allowing intermixture with boolean expressions. The patterns inside the `let` sub-expressions can be irrefutable or refutable.
```Rust
struct FnCall<'a> {
fn_name: &'a str,
args: Vec<i32>,
}
fn is_legal_ident(s: &str) -> bool {
s.chars()
.all(|c| ('a'..='z').contains(&c) || ('A'..='Z').contains(&c))
}
impl<'a> FnCall<'a> {
fn parse(s: &'a str) -> Option<Self> {
if let Some((fn_name, after_name)) = s.split_once("(")
&& !fn_name.is_empty()
&& is_legal_ident(fn_name)
&& let Some((args_str, "")) = after_name.rsplit_once(")")
{
let args = args_str
.split(',')
.map(|arg| arg.parse())
.collect::<Result<Vec<_>, _>>();
args.ok().map(|args| FnCall { fn_name, args })
} else {
None
}
}
fn exec(&self) -> Option<i32> {
let iter = self.args.iter().copied();
match self.fn_name {
"sum" => Some(iter.sum()),
"max" => iter.max(),
"min" => iter.min(),
_ => None,
}
}
}
fn main() {
println!("{:?}", FnCall::parse("sum(1,2,3)").unwrap().exec());
println!("{:?}", FnCall::parse("max(4,5)").unwrap().exec());
}
```
The feature will only be stabilized for the 2024 edition and future editions. Users of past editions will get an error with a hint to update the edition.
closes #53667
## Why 2024 edition?
Rust generally tries to ship new features to all editions. So even the oldest editions receive the newest features. However, sometimes a feature requires a breaking change so much that offering the feature without the breaking change makes no sense. This occurs rarely, but has happened in the 2018 edition already with `async` and `await` syntax. It required an edition boundary in order for `async`/`await` to become keywords, and the entire feature foots on those keywords.
In the instance of let chains, the issue is the drop order of `if let` chains. If we want `if let` chains to be compatible with `if let`, drop order makes it hard for us to [generate correct MIR]. It would be strange to have different behaviour for `if let ... {}` and `if true && let ... {}`. So it's better to [stay consistent with `if let`].
In edition 2024, [drop order changes] have been introduced to make `if let` temporaries be lived more shortly. These changes also affected `if let` chains. These changes make sense even if you don't take the `if let` chains MIR generation problem into account. But if we want to use them as the solution to the MIR generation problem, we need to restrict let chains to edition 2024 and beyond: for let chains, it's not just a change towards more sensible behaviour, but one required for correct function.
[generate correct MIR]: https://github.com/rust-lang/rust/issues/104843
[stay consistent with `if let`]: https://github.com/rust-lang/rust/pull/103293#issuecomment-1293408574
[drop order changes]: https://github.com/rust-lang/rust/issues/124085
## Introduction considerations
As edition 2024 is very new, this stabilization PR only makes it possible to use let chains on 2024 without that feature gate, it doesn't mark that feature gate as stable/removed. I would propose to continue offering the `let_chains` feature (behind a feature gate) for a limited time (maybe 3 months after stabilization?) on older editions to allow nightly users to adopt edition 2024 at their own pace. After that, the feature gate shall be marked as *stabilized*, not removed, and replaced by an error on editions 2021 and below.
## Implementation history
* History from before March 14, 2022 can be found in the [original stabilization PR] that was reverted.
* https://github.com/rust-lang/rust/pull/94927
* https://github.com/rust-lang/rust/pull/94951
* https://github.com/rust-lang/rust/pull/94974
* https://github.com/rust-lang/rust/pull/95008
* https://github.com/rust-lang/rust/pull/97295
* https://github.com/rust-lang/rust/pull/98633
* https://github.com/rust-lang/rust/pull/99731
* https://github.com/rust-lang/rust/pull/102394
* https://github.com/rust-lang/rust/pull/100526
* https://github.com/rust-lang/rust/pull/100538
* https://github.com/rust-lang/rust/pull/102998
* https://github.com/rust-lang/rust/pull/103405
* https://github.com/rust-lang/rust/pull/103293
* https://github.com/rust-lang/rust/pull/107251
* https://github.com/rust-lang/rust/pull/110568
* https://github.com/rust-lang/rust/pull/115677
* https://github.com/rust-lang/rust/pull/117743
* https://github.com/rust-lang/rust/pull/117770
* https://github.com/rust-lang/rust/pull/118191
* https://github.com/rust-lang/rust/pull/119554
* https://github.com/rust-lang/rust/pull/129394
* https://github.com/rust-lang/rust/pull/132828
* https://github.com/rust-lang/reference/pull/1179
* https://github.com/rust-lang/reference/pull/1251
* https://github.com/rust-lang/rustfmt/pull/5910
[original stabilization PR]: https://github.com/rust-lang/rust/pull/94927
## Adoption history
### In the compiler
* History before March 14, 2022 can be found in the [original stabilization PR].
* https://github.com/rust-lang/rust/pull/115983
* https://github.com/rust-lang/rust/pull/116549
* https://github.com/rust-lang/rust/pull/116688
### Outside of the compiler
* https://github.com/rust-lang/rust-clippy/pull/11750
* [rspack](https://github.com/web-infra-dev/rspack)
* [risingwave](https://github.com/risingwavelabs/risingwave)
* [dylint](https://github.com/trailofbits/dylint)
* [convex-backend](https://github.com/get-convex/convex-backend)
* [tikv](https://github.com/tikv/tikv)
* [Daft](https://github.com/Eventual-Inc/Daft)
* [greptimedb](https://github.com/GreptimeTeam/greptimedb)
## Tests
<details>
### Intentional restrictions
[`partially-macro-expanded.rs`](https://github.com/rust-lang/rust/blob/4adafcf40aa6064d2bbcb44bc1a50b3b1e86e5e0/tests/ui/rfcs/rfc-2294-if-let-guard/partially-macro-expanded.rs), [`macro-expanded.rs`](https://github.com/rust-lang/rust/blob/4adafcf40aa6064d2bbcb44bc1a50b3b1e86e5e0/tests/ui/rfcs/rfc-2294-if-let-guard/macro-expanded.rs): it is possible to use macros to expand to both the pattern and the expression inside a let chain, but not to the entire `let pat = expr` operand.
[`parens.rs`](https://github.com/rust-lang/rust/blob/4adafcf40aa6064d2bbcb44bc1a50b3b1e86e5e0/tests/ui/rfcs/rfc-2294-if-let-guard/parens.rs): `if (let pat = expr)` is not allowed in chains
[`ensure-that-let-else-does-not-interact-with-let-chains.rs`](https://github.com/rust-lang/rust/blob/4adafcf40aa6064d2bbcb44bc1a50b3b1e86e5e0/tests/ui/rfcs/rfc-2497-if-let-chains/ensure-that-let-else-does-not-interact-with-let-chains.rs): `let...else` doesn't support chaining.
### Overlap with match guards
[`move-guard-if-let-chain.rs`](https://github.com/rust-lang/rust/blob/4adafcf40aa6064d2bbcb44bc1a50b3b1e86e5e0/tests/ui/rfcs/rfc-2294-if-let-guard/move-guard-if-let-chain.rs): test for the `use moved value` error working well in match guards. could maybe be extended with let chains that have more than one `let`
[`shadowing.rs`](https://github.com/rust-lang/rust/blob/4adafcf40aa6064d2bbcb44bc1a50b3b1e86e5e0/tests/ui/rfcs/rfc-2294-if-let-guard/shadowing.rs): shadowing in if let guards works as expected
[`ast-validate-guards.rs`](https://github.com/rust-lang/rust/blob/4adafcf40aa6064d2bbcb44bc1a50b3b1e86e5e0/tests/ui/rfcs/rfc-2497-if-let-chains/ast-validate-guards.rs): let chains in match guards require the match guards feature gate
### Simple cases from the early days
PR #88642 has added some tests with very simple usages of `let else`, mostly as regression tests to early bugs.
[`then-else-blocks.rs`](https://github.com/rust-lang/rust/blob/4adafcf40aa6064d2bbcb44bc1a50b3b1e86e5e0/tests/ui/rfcs/rfc-2497-if-let-chains/then-else-blocks.rs)
[`ast-lowering-does-not-wrap-let-chains.rs`](https://github.com/rust-lang/rust/blob/4adafcf40aa6064d2bbcb44bc1a50b3b1e86e5e0/tests/ui/rfcs/rfc-2497-if-let-chains/ast-lowering-does-not-wrap-let-chains.rs)
[`issue-90722.rs`](https://github.com/rust-lang/rust/blob/4adafcf40aa6064d2bbcb44bc1a50b3b1e86e5e0/tests/ui/rfcs/rfc-2497-if-let-chains/issue-90722.rs)
[`issue-92145.rs`](https://github.com/rust-lang/rust/blob/4adafcf40aa6064d2bbcb44bc1a50b3b1e86e5e0/tests/ui/rfcs/rfc-2497-if-let-chains/issue-92145.rs)
### Drop order/MIR scoping tests
[`issue-100276.rs`](https://github.com/rust-lang/rust/blob/4adafcf40aa6064d2bbcb44bc1a50b3b1e86e5e0/tests/ui/drop/issue-100276.rs): let expressions on RHS aren't terminating scopes
[`drop_order.rs`](https://github.com/rust-lang/rust/blob/4adafcf40aa6064d2bbcb44bc1a50b3b1e86e5e0/tests/ui/drop/drop_order.rs): exhaustive temporary drop order test for various Rust constructs, including let chains
[`scope.rs`](https://github.com/rust-lang/rust/blob/4adafcf40aa6064d2bbcb44bc1a50b3b1e86e5e0/tests/ui/rfcs/rfc-2294-if-let-guard/scope.rs): match guard scoping test
[`drop-scope.rs`](https://github.com/rust-lang/rust/blob/4adafcf40aa6064d2bbcb44bc1a50b3b1e86e5e0/tests/ui/rfcs/rfc-2294-if-let-guard/drop-scope.rs): another match guard scoping test, ensuring that temporaries in if-let guards live for the arm
[`drop_order_if_let_rescope.rs`](https://github.com/rust-lang/rust/blob/4adafcf40aa6064d2bbcb44bc1a50b3b1e86e5e0/tests/ui/drop/drop_order_if_let_rescope.rs): if let rescoping on edition 2024, including chains
[`mir_let_chains_drop_order.rs`](https://github.com/rust-lang/rust/blob/4adafcf40aa6064d2bbcb44bc1a50b3b1e86e5e0/tests/ui/mir/mir_let_chains_drop_order.rs): comprehensive drop order test for let chains, distinguishes editions 2021 and 2024.
[`issue-99938.rs`](https://github.com/rust-lang/rust/blob/4adafcf40aa6064d2bbcb44bc1a50b3b1e86e5e0/tests/ui/rfcs/rfc-2497-if-let-chains/issue-99938.rs), [`issue-99852.rs`](https://github.com/rust-lang/rust/blob/4adafcf40aa6064d2bbcb44bc1a50b3b1e86e5e0/tests/ui/mir/issue-99852.rs) both bad MIR ICEs fixed by #102394
### Linting
[`irrefutable-lets.rs`](https://github.com/rust-lang/rust/blob/4adafcf40aa6064d2bbcb44bc1a50b3b1e86e5e0/tests/ui/rfcs/rfc-2497-if-let-chains/irrefutable-lets.rs): trailing and leading irrefutable let patterns get linted for, others don't. The lint is turned off for `else if`.
[`issue-121070-let-range.rs`](https://github.com/rust-lang/rust/blob/4adafcf40aa6064d2bbcb44bc1a50b3b1e86e5e0/tests/ui/lint/issue-121070-let-range.rs): regression test for false positive of the unused parens lint, precedence requires the `()`s here
### Parser: intentional restrictions
[`disallowed-positions.rs`](https://github.com/rust-lang/rust/blob/2128d8df0e858edcbe6a0861bac948b88b7fabc3/tests/ui/rfcs/rfc-2497-if-let-chains/disallowed-positions.rs): `let` in expression context is rejected everywhere except at the top level
[`invalid-let-in-a-valid-let-context.rs`](https://github.com/rust-lang/rust/blob/4adafcf40aa6064d2bbcb44bc1a50b3b1e86e5e0/tests/ui/rfcs/rfc-2497-if-let-chains/invalid-let-in-a-valid-let-context.rs): nested `let` is not allowed (let's are no legal expressions just because they are allowed in `if` and `while`).
### Parser: recovery
[`issue-103381.rs`](https://github.com/rust-lang/rust/blob/4adafcf40aa6064d2bbcb44bc1a50b3b1e86e5e0/tests/ui/parser/issues/issue-103381.rs): Graceful recovery of incorrect chaining of `if` and `if let`
[`semi-in-let-chain.rs`](https://github.com/rust-lang/rust/blob/4adafcf40aa6064d2bbcb44bc1a50b3b1e86e5e0/tests/ui/parser/semi-in-let-chain.rs): Ensure that stray `;`s in let chains give nice errors (`if_chain!` users might be accustomed to `;`s)
[`deli-ident-issue-1.rs`](https://github.com/rust-lang/rust/blob/4adafcf40aa6064d2bbcb44bc1a50b3b1e86e5e0/tests/ui/parser/deli-ident-issue-1.rs), [`brace-in-let-chain.rs`](https://github.com/rust-lang/rust/blob/4adafcf40aa6064d2bbcb44bc1a50b3b1e86e5e0/tests/ui/parser/brace-in-let-chain.rs): Ensure that stray unclosed `{`s in let chains give nice errors and hints
### Misc
[`conflicting_bindings.rs`](https://github.com/rust-lang/rust/blob/4adafcf40aa6064d2bbcb44bc1a50b3b1e86e5e0/tests/ui/pattern/usefulness/conflicting_bindings.rs): the conflicting bindings check also works in let chains. Personally, I'd extend it to chains with multiple let's as well.
[`let-chains-attr.rs`](https://github.com/rust-lang/rust/blob/4adafcf40aa6064d2bbcb44bc1a50b3b1e86e5e0/tests/ui/expr/if/attrs/let-chains-attr.rs): attributes work on let chains
### Tangential tests with `#![feature(let_chains)]`
[`if-let.rs`](https://github.com/rust-lang/rust/blob/4adafcf40aa6064d2bbcb44bc1a50b3b1e86e5e0/tests/coverage/branch/if-let.rs): MC/DC coverage tests for let chains
[`logical_or_in_conditional.rs`](https://github.com/rust-lang/rust/blob/4adafcf40aa6064d2bbcb44bc1a50b3b1e86e5e0/tests/mir-opt/building/logical_or_in_conditional.rs): not really about let chains, more about dropping/scoping behaviour of `||`
[`stringify.rs`](https://github.com/rust-lang/rust/blob/4adafcf40aa6064d2bbcb44bc1a50b3b1e86e5e0/tests/ui/macros/stringify.rs): exhaustive test of the `stringify` macro
[`expanded-interpolation.rs`](https://github.com/rust-lang/rust/blob/4adafcf40aa6064d2bbcb44bc1a50b3b1e86e5e0/tests/ui/unpretty/expanded-interpolation.rs), [`expanded-exhaustive.rs`](https://github.com/rust-lang/rust/blob/4adafcf40aa6064d2bbcb44bc1a50b3b1e86e5e0/tests/ui/unpretty/expanded-exhaustive.rs): Exhaustive test of `-Zunpretty`
[`diverges-not.rs`](https://github.com/rust-lang/rust/blob/4adafcf40aa6064d2bbcb44bc1a50b3b1e86e5e0/tests/ui/rfcs/rfc-0000-never_patterns/diverges-not.rs): Never type, mostly tangential to let chains
</details>
## Possible future work
* There is proposals to allow `if let Pat(bindings) = expr {}` to be written as `if expr is Pat(bindings) {}` ([RFC 3573]). `if let` chains are a natural extension of the already existing `if let` syntax, and I'd argue orthogonal towards `is` syntax.
* https://github.com/rust-lang/lang-team/issues/297
* One could have similar chaining inside `let ... else` statements. There is no proposed RFC for this however, nor is it implemented on nightly.
* Match guards have the `if` keyword as well, but on stable Rust, they don't support `let`. The functionality is available via an unstable feature ([`if_let_guard` tracking issue]). Stabilization of let chains affects this feature in so far as match guards containing let chains now only need the `if_let_guard` feature gate be present instead of also the `let_chains` feature (NOTE: this PR doesn't implement this simplification, it's left for future work).
[RFC 3573]: https://github.com/rust-lang/rfcs/pull/3573
[`if_let_guard` tracking issue]: https://github.com/rust-lang/rust/issues/51114
## Open questions / blockers
- [ ] bad recovery if you don't put a `let` (I don't think this is a blocker): [#117977](https://github.com/rust-lang/rust/issues/117977)
- [x] An instance where a temporary lives shorter than with nested ifs, breaking compilation: [#103476](https://github.com/rust-lang/rust/issues/103476). Personally I don't think this is a blocker either, as it's an edge case. Edit: turns out to not reproduce in edition 2025 any more, due to let rescoping. regression test added in #133093
- [x] One should probably extend the tests for `move-guard-if-let-chain.rs` and `conflicting_bindings.rs` to have chains with multiple let's: done in 133093
- [x] Parsing rejection tests: addressed by https://github.com/rust-lang/rust/pull/132828
- [x] [Style](https://rust-lang.zulipchat.com/#narrow/channel/346005-t-style/topic/let.20chains.20stabilization.20and.20formatting): https://github.com/rust-lang/rust/pull/139456
- [x] https://github.com/rust-lang/rust/issues/86730 explicitly mentions `let_else`. I think we can live with `let pat = expr` not evaluating as `expr` for macro_rules macros, especially given that `let pat = expr` is not a legal expression anywhere except inside `if` and `while`.
- [x] Documentation in the reference: https://github.com/rust-lang/reference/pull/1740
- [x] Add chapter to the Rust 2024 [edition guide]: https://github.com/rust-lang/edition-guide/pull/337
- [x] Resolve open questions on desired drop order.
[original reference PR]: https://github.com/rust-lang/reference/pull/1179
[edition guide]: https://github.com/rust-lang/edition-guide
|
|
Extends the change from #139653, which was on expressions, to literals.
Fixes #140098.
|
|
|
|
|
|
r=compiler-errors
Improve diagnostics for pointer arithmetic += and -= (fixes #137391)
**Description**:
This PR improves the diagnostic message for cases where a binary assignment operation like `ptr += offset` or `ptr -= offset` is attempted on `*mut T`. These operations are not allowed, and the compiler previously suggested calling `.add()` or `.wrapping_add()`, which is misleading if not assigned.
This PR updates the diagnostics to suggest assigning the result of `.wrapping_add()` or `.wrapping_sub()` back to the pointer, e.g.:
**Examples**
For this code
```rust
let mut arr = [0u8; 10];
let mut ptr = arr.as_mut_ptr();
ptr += 2;
```
it will say:
```rust
10 | ptr += 2;
| ---^^^^^
| |
| cannot use `+=` on type `*mut u8`
|
help: consider replacing `ptr += offset` with `ptr = ptr.wrapping_add(offset)` or `ptr.add(offset)`
|
10 - ptr += 2;
10 + ptr = ptr.wrapping_add(2);
```
**Related issue**: #137391
cc `@nabijaczleweli` for context (issue author)
|
|
Construct OutputType using macro and print [=FILENAME] help info
Closes #139805
Use define_output_types to define variants of OutputType, as well as refactor all of its methods for clarity. This way no variant is missed when pattern matching or output help messages.
On top of that, I optimized for `emit` help messages.
r? ```@jieyouxu```
|
|
Don't compute name of associated item if it's an RPITIT
Another simple fix for an RPITIT name ICE.
Fixes https://github.com/rust-lang/rust/issues/139941
Fixes #140084
r? nnethercote
|
|
|
|
Rollup of 11 pull requests
Successful merges:
- #134213 (Stabilize `naked_functions`)
- #139711 (Hermit: Unify `std::env::args` with Unix)
- #139795 (Clarify why SGX code specifies linkage/symbol names for certain statics)
- #140036 (Advent of `tests/ui` (misc cleanups and improvements) [4/N])
- #140047 (remove a couple clones)
- #140052 (Fix error when an intra doc link is trying to resolve an empty associated item)
- #140074 (rustdoc-json: Improve test for auto-trait impls)
- #140076 (jsondocck: Require command is at start of line)
- #140107 (rustc-dev-guide subtree update)
- #140111 (cleanup redundant pattern instances)
- #140118 ({B,C}Str: minor cleanup)
r? `@ghost`
`@rustbot` modify labels: rollup
|
|
jsondocck: Require command is at start of line
In one place we use `///``@``` instead of `//``@`.`` The test-runner allowed it, but it probably shouldn't. Ran into by ``@lolbinarycat`` in https://github.com/rust-lang/rust/pull/132748#issuecomment-2816469322:
```
error: unknown disambiguator `?(`
##[error] --> /checkout/tests/rustdoc-json/fns/return_type_alias.rs:3:25
|
3 | ///@ set foo = "$.index[?(``@.name=='Foo')].id"``
| ^^
|
```
Maybe it's also worth erroring on this like we added in #137103
r? ``@GuillaumeGomez``
|
|
rustdoc-json: Improve test for auto-trait impls
The TODO is fixable now due-to #138763. While I was here I realized there's probably a a few more things we should also test.
r? ```@GuillaumeGomez```
|
|
Fix error when an intra doc link is trying to resolve an empty associated item
Fixes https://github.com/rust-lang/rust/issues/140026.
Assigning ```@nnethercote``` since they're the one who wrote the initial change.
I updated rustdoc code instead of compiler's because I think it makes more sense that the caller ensures on their side that the name they're looking for isn't empty.
r? ```@nnethercote```
|
|
Advent of `tests/ui` (misc cleanups and improvements) [4/N]
Some `tests/ui/` housekeeping, to trim down number of tests directly under `tests/ui/`. Part of #133895.
### Review advice
- Best reviewed commit-by-commit.
- I can squash commits before merge, commits are separate to make it easier to review.
|
|
r=tgross35,Amanieu,traviscross
Stabilize `naked_functions`
tracking issue: https://github.com/rust-lang/rust/issues/90957
request for stabilization on tracking issue: https://github.com/rust-lang/rust/issues/90957#issuecomment-2539270352
reference PR: https://github.com/rust-lang/reference/pull/1689
# Request for Stabilization
Two years later, we're ready to try this again. Even though this issue is already marked as having passed FCP, given the amount of time that has passed and the changes in implementation strategy, we should follow the process again.
## Summary
The `naked_functions` feature has two main parts: the `#[naked]` function attribute, and the `naked_asm!` macro.
An example of a naked function:
```rust
const THREE: usize = 3;
#[naked]
pub extern "sysv64" fn add_n(number: usize) -> usize {
// SAFETY: the validity of the used registers
// is guaranteed according to the "sysv64" ABI
unsafe {
core::arch::naked_asm!(
"add rdi, {}",
"mov rax, rdi",
"ret",
const THREE,
)
}
}
```
When the `#[naked]` attribute is applied to a function, the compiler won't emit a [function prologue](https://en.wikipedia.org/wiki/Function_prologue_and_epilogue) or epilogue when generating code for this function. This attribute is analogous to [`__attribute__((naked))`](https://developer.arm.com/documentation/100067/0608/Compiler-specific-Function--Variable--and-Type-Attributes/--attribute----naked---function-attribute) in C. The use of this feature allows the programmer to have precise control over the assembly that is generated for a given function.
The body of a naked function must consist of a single `naked_asm!` invocation, a heavily restricted variant of the `asm!` macro: the only legal operands are `const` and `sym`, and the only legal options are `raw` and `att_syntax`. In lieu of specifying operands, the `naked_asm!` within a naked function relies on the function's calling convention to determine the validity of registers.
## Documentation
The Rust Reference: https://github.com/rust-lang/reference/pull/1689
(Previous PR: https://github.com/rust-lang/reference/pull/1153)
## Tests
* [tests/run-make/naked-symbol-visiblity](https://github.com/rust-lang/rust/tree/master/tests/codegen/naked-fn) verifies that `pub`, `#[no_mangle]` and `#[linkage = "..."]` work correctly for naked functions
* [tests/codegen/naked-fn](https://github.com/rust-lang/rust/tree/master/tests/codegen/naked-fn) has tests for function alignment, use of generics, and validates the exact assembly output on linux, macos, windows and thumb
* [tests/ui/asm/naked-*](https://github.com/rust-lang/rust/tree/master/tests/ui/asm) tests for incompatible attributes, generating errors around incorrect use of `naked_asm!`, etc
## Interaction with other (unstable) features
### [fn_align](https://github.com/rust-lang/rust/issues/82232)
Combining `#[naked]` with `#[repr(align(N))]` works well, and is tested e.g. here
- https://github.com/rust-lang/rust/blob/master/tests/codegen/naked-fn/aligned.rs
- https://github.com/rust-lang/rust/blob/master/tests/codegen/naked-fn/min-function-alignment.rs
It's tested extensively because we do need to explicitly support the `repr(align)` attribute (and make sure we e.g. don't mistake powers of two for number of bytes).
## History
This feature was originally proposed in [RFC 1201](https://github.com/rust-lang/rfcs/pull/1201), filed on 2015-07-10 and accepted on 2016-03-21. Support for this feature was added in [#32410](https://github.com/rust-lang/rust/pull/32410), landing on 2016-03-23. Development languished for several years as it was realized that the semantics given in RFC 1201 were insufficiently specific. To address this, a minimal subset of naked functions was specified by [RFC 2972](https://github.com/rust-lang/rfcs/pull/2972), filed on 2020-08-07 and accepted on 2021-11-16. Prior to the acceptance of RFC 2972, all of the stricter behavior specified by RFC 2972 was implemented as a series of warn-by-default lints that would trigger on existing uses of the `naked` attribute; these lints became hard errors in [#93153](https://github.com/rust-lang/rust/pull/93153) on 2022-01-22. As a result, today RFC 2972 has completely superseded RFC 1201 in describing the semantics of the `naked` attribute.
More recently, the `naked_asm!` macro was added to replace the earlier use of a heavily restricted `asm!` invocation. The `naked_asm!` name is clearer in error messages, and provides a place for documenting the specific requirements of inline assembly in naked functions.
The implementation strategy was changed to emitting a global assembly block. In effect, an extern function
```rust
extern "C" fn foo() {
core::arch::naked_asm!("ret")
}
```
is emitted as something similar to
```rust
core::arch::global_asm!(
"foo:",
"ret"
);
extern "C" {
fn foo();
}
```
The codegen approach was chosen over the llvm naked function attribute because:
- the rust compiler can guarantee the behavior (no sneaky additional instructions, no inlining, etc.)
- behavior is the same on all backends (llvm, cranelift, gcc, etc)
Finally, there is now an allow list of compatible attributes on naked functions, so that e.g. `#[inline]` is rejected with an error. The `#[target_feature]` attribute on naked functions was later made separately unstable, because implementing it is complex and we did not want to block naked functions themselves on how target features work on them. See also https://github.com/rust-lang/rust/issues/138568.
relevant PRs for these recent changes
- https://github.com/rust-lang/rust/pull/127853
- https://github.com/rust-lang/rust/pull/128651
- https://github.com/rust-lang/rust/pull/128004
- https://github.com/rust-lang/rust/pull/138570
-
### Various historical notes
#### `noreturn`
[RFC 2972](https://github.com/rust-lang/rfcs/blob/master/text/2972-constrained-naked.md) mentions that naked functions
> must have a body which contains only a single asm!() statement which:
> iii. must contain the noreturn option.
Instead of `asm!`, the current implementation mandates that the body contain a single `naked_asm!` statement. The `naked_asm!` macro is a heavily restricted version of the `asm!` macro, making it easier to talk about and document the rules of assembly in naked functions and give dedicated error messages.
For `naked_asm!`, the behavior of the `asm!`'s `noreturn` option is implicit. The `noreturn` option means that it is UB for control flow to fall through the end of the assembly block. With `asm!`, this option is usually used for blocks that diverge (and thus have no return and can be typed as `!`). With `naked_asm!`, the intent is different: usually naked funtions do return, but they must do so from within the assembly block. The `noreturn` option was used so that the compiler would not itself also insert a `ret` instruction at the very end.
#### padding / `ud2`
A `naked_asm!` block that violates the safety assumption that control flow must not fall through the end of the assembly block is UB. Because no return instruction is emitted, whatever bytes follow the naked function will be executed, resulting in truly undefined behavior. There has been discussion whether rustc should emit an invalid instruction (e.g. `ud2` on x86) after the `naked_asm!` block to at least fail early in the case of an invalid `naked_asm!`. It was however decided that it is more useful to guarantee that `#[naked]` functions NEVER contain any instructions besides those in the `naked_asm!` block.
# unresolved questions
None
r? ``@Amanieu``
I've validated the tests on x86_64 and aarch64
|
|
|
|
Relocate tests in `tests/ui`
Part of #133895
Moved tests from a top-level directory into a more appropriate subdirectory.
If there is anything else that could be improved, please let me know!
r? jieyouxu
|
|
Don't ICE on pending obligations from deep normalization in a loop
See the comment I left inline in `compiler/rustc_trait_selection/src/traits/normalize.rs`.
Fixes https://github.com/rust-lang/rust/issues/133868
r? lcnr
|
|
docs(LocalKey<T>): clarify that T's Drop shouldn't panic
Clarify that should a TLS destructor panics, the process will abort.
Also, an abort may be obfuscated as the process can be terminated with `SIGSEGV` or [`STATUS_STACK_BUFFER_OVERRUN`](https://devblogs.microsoft.com/oldnewthing/20190108-00/?p=100655) (i.e., `SIGABRT` is not guaranteed), so explicitly prints that the process was aborted.
Context:
https://users.rust-lang.org/t/status-stack-buffer-overrun-on-windows-without-any-usage-of-unsafe/128417
``@rustbot`` label -T-compiler
|
|
Signed-off-by: xizheyin <xizheyin@smail.nju.edu.cn>
|
|
|
|
|
|
simd intrinsics with mask: accept unsigned integer masks, and fix some of the errors
It's not clear at all why the mask would have to be signed, it is anyway interpreted bitwise. The backend should just make sure that works no matter the surface-level type; our LLVM backend already does this correctly. The note of "the mask may be widened, which only has the correct behavior for signed integers" explains... nothing? Why can't the code do the widening correctly? If necessary, just cast to the signed type first...
Also while we are at it, fix the errors. For simd_masked_load/store, the errors talked about the "third argument" but they meant the first argument (the mask is the first argument there). They also used the wrong type for `expected_element`.
I have extremely low confidence in the GCC part of this PR.
See [discussion on Zulip](https://rust-lang.zulipchat.com/#narrow/channel/257879-project-portable-simd/topic/On.20the.20sign.20of.20masks)
|
|
Signed-off-by: xizheyin <xizheyin@smail.nju.edu.cn>
|
|
|
|
|
|
|
|
|
|
Rollup of 8 pull requests
Successful merges:
- #138934 (support config extensions)
- #139091 (Rewrite on_unimplemented format string parser.)
- #139753 (Make `#[naked]` an unsafe attribute)
- #139762 (Don't assemble non-env/bound candidates if projection is rigid)
- #139834 (Don't canonicalize crate paths)
- #139868 (Move `pal::env` to `std::sys::env_consts`)
- #139978 (Add citool command for generating a test dashboard)
- #139995 (Clean UI tests 4 of n)
r? `@ghost`
`@rustbot` modify labels: rollup
|
|
Do not remove trivial `SwitchInt` in analysis MIR
This PR ensures that we don't prematurely remove trivial `SwitchInt` terminators which affects both the borrow-checking and runtime semantics (i.e. UB) of the code. Previously the `SimplifyCfg` optimization was removing `SwitchInt` terminators when they was "trivial", i.e. when all arms branched to the same basic block, even if that `SwitchInt` terminator had the side-effect of reading an operand which (for example) may not be initialized or may point to an invalid place in memory.
This behavior is unlike all other optimizations, which are only applied after "analysis" (i.e. borrow-checking) is finished, and which Miri disables to make sure the compiler doesn't silently remove UB.
Fixing this code "breaks" (i.e. unmasks) code that used to borrow-check but no longer does, like:
```rust
fn foo() {
let x;
let (0 | _) = x;
}
```
This match expression should perform a read because `_` does not shadow the `0` literal pattern, and the compiler should have to read the match scrutinee to compare it to 0. I've checked that this behavior does not actually manifest in practice via a crater run which came back clean: https://github.com/rust-lang/rust/pull/139042#issuecomment-2767436367
As a side-note, it may be tempting to suggest that this is actually a good thing or that we should preserve this behavior. If we wanted to make this work (i.e. trivially optimize out reads from matches that are redundant like `0 | _`), then we should be enabling this behavior *after* fixing this. However, I think it's kinda unprincipled, and for example other variations of the code don't even work today, e.g.:
```rust
fn foo() {
let x;
let (0.. | _) = x;
}
```
|
|
|
|
|
|
Clean UI tests 4 of n
Cleaned up some tests that have `issue` in the title. I kept the commits to be one per "`issue`" cleanup/rename to make it easier to check. I can rebase to one commit once the changes are approved.
Related Issues:
#73494
#133895
r? jieyouxu
|
|
Don't canonicalize crate paths
When printing paths in diagnostic we should favour printing the paths that were passed in rather than resolving all symlinks.
This PR changes the form of the crate path but it should only really affect diagnostics as filesystem functions won't care which path is used. The uncanonicalized path was already used as a fallback for when canonicalization failed.
This is a partial alternative to #139823.
|
|
Don't assemble non-env/bound candidates if projection is rigid
Putting this up for an initial review, it's still missing comments, clean-up, and possibly a tweak to deal with ambiguities in the `BestObligation` folder.
This PR fixes https://github.com/rust-lang/trait-system-refactor-initiative/issues/173. Specifically, we're creating an unnecessary query cycle in normalization by assembling an *impl candidate* even if we know later on during `merge_candidates` that we'll be filtering out that impl candidate.
This PR adjusts the `merge_candidates` to assemble *only* env/bound candidates if we have `TraitGoalProvenVia::ParamEnv | TraitGoalProvenVia::AliasBound`.
I'll leave some thoughts/comments in the code.
r? lcnr
|
|
r=tgross35,traviscross
Make `#[naked]` an unsafe attribute
tracking issue: https://github.com/rust-lang/rust/issues/138997
Per https://github.com/rust-lang/rust/pull/134213#issuecomment-2755984503, the `#[naked]` attribute is now an unsafe attribute (in any edition).
This can only be merged when the above PRs are merged, I'd just like to see if there are any CI surprises here, and maybe there is early review feedback too.
r? ``@traviscross``
|
|
Rewrite on_unimplemented format string parser.
This PR rewrites the format string parser for `rustc_on_unimplemented` and `diagnostic::on_unimplemented`. I plan on moving this code (and more) into the new attribute parsing system soon and wanted to PR it separately.
This PR introduces some minor differences though:
- `rustc_on_unimplemented` on trait *implementations* is no longer checked/used - this is actually never used (outside of some tests) so I plan on removing it in the future.
- for `rustc_on_unimplemented`, it introduces the `{This}` argument in favor of `{ThisTraitname}` (to be removed later). It'll be easier to parse.
- for `rustc_on_unimplemented`, `Self` can now consistently be used as a filter, rather than just `_Self`. It used to not match correctly on for example `Self = "[{integer}]"`
- Some error messages now have better spans.
Fixes https://github.com/rust-lang/rust/issues/130627
|
|
Make rustdoc JSON Span column 1-based, just like line numbers
Fixes https://github.com/rust-lang/rust/issues/139906.
This PR does two things:
1. It makes column 1-indexed as well, just like lines.
2. It updates documentation about them to mention that they are 1-indexed.
I think it's better for coherency to have them both 1-indexed instead of the weird mix we used to have. Docs for `line` and `col` fields can be found [here](https://doc.rust-lang.org/nightly/nightly-rustc/rustc_span/struct.Loc.html#structfield.line).
And finally: it adds a regression test to ensure they are indeed 1-indexed.
r? `@aDotInTheVoid`
|
|
not lint break with label and unsafe block
fixes #137414
we can't label unsafe blocks, so that we can do not lint them
|
|
- Reformat the test.
- Document test intention.
- Move test under `tests/ui/inference/`.
|
|
- `tests/ui/augmented-assignment-feature-gate-cross.rs`:
- This was *originally* to feature-gate overloaded OpAssign
cross-crate, but now let's keep it as a smoke test.
- Renamed as `augmented-assignment-cross-crate.rs`.
- Relocated under `tests/ui/binop/`.
- `tests/ui/augmented-assignments.rs`:
- Documented test intent.
- Moved under `tests/ui/borrowck/`.
- `tests/ui/augmented-assignment-rpass.rs`:
- Renamed to drop the `-rpass` suffix, since this was leftover from
when `run-pass` test suite was a thing.
- Moved under `tests/ui/binop/`.
|
|
- Reworked the test as a *centralized* version of checking that certain
targets correctly require `-C target-cpu` being specified.
- Document test intention.
- Move `amdgpu-require-explicit-cpu.rs` under new dir
`tests/ui/target-cpu/`
- No other ui subdir really fits this "requires `-Ctarget-cpu`" check.
|
|
Implement `pin!()` using `super let`
Tracking issue for super let: https://github.com/rust-lang/rust/issues/139076
This uses `super let` to implement `pin!()`.
This means we can remove [the hack](https://github.com/rust-lang/rust/pull/138717) we had to put in to fix https://github.com/rust-lang/rust/issues/138596.
It also means we can remove the original hack to make `pin!()` work, which used a questionable public-but-unstable field rather than a proper private field.
While `super let` is still unstable and subject to change, it seems safe to assume that future Rust will always have a way to express `pin!()` in a compatible way, considering `pin!()` is already stable.
It'd help [the experiment](https://github.com/rust-lang/rust/issues/139076) to have `pin!()` use `super let`, so we can get some more experience with it.
|
|
|
|
|
|
|
