Don't add empty target features for target-cpu=native on macOS
LLVM does not support host feature detection (only host cpu detection) on apple platforms. As such, the returned feature string will be empty. Adding this empty string to the target-features attribute results in a verifier error on LLVM 22.
Fix this by not adding the empty string to the target features. The reason why this was not caught by the target-cpu-native test is that it requires a function that adds *some* target features, otherwise the attribute is omitted entirely. We achieve this with a somewhat peculiar construction that enables `neon` if it's already enabled. (This is to avoid enabling it on softfloat targets.)
Fixes https://github.com/rust-lang/rust/issues/153397.
explicit tail calls: support indirect arguments
tracking issue: https://github.com/rust-lang/rust/issues/112788
After discussion in https://github.com/rust-lang/rust/issues/144855, I was wrong and it is actually possible to support tail calls with `PassMode::Indirect { on_stack: false, .. }` arguments.
Normally an indirect argument with `on_stack: false` would be passed as a pointer into the caller's stack frame. For tail calls, that would be unsound, because the caller's stack frame is overwritten by the callee's stack frame.
Therefore we store the argument for the callee in the corresponding caller's slot. Because guaranteed tail calls demand that the caller's signature matches the callee's, the corresponding slot has the correct type.
To handle cases like the one below, the tail call arguments must first be copied to a temporary, and can only then be copied to the caller's argument slots.
```rust
// A struct big enough that it is not passed via registers.
pub struct Big([u64; 4]);
fn swapper(a: Big, b: Big) -> (Big, Big) {
become swapper_helper(b, a);
}
```
---
I'm not really familiar with MIR and what tricks/helpers we have, so I'm just cobbling this together. Hopefully we can arrive at something more elegant.
Avoid duplicate `requirement` diag args in `RegionOriginNote`
Fixesrust-lang/rust#143872
`RegionOriginNote::WithRequirement` can be emitted multiple times for one
diagnostic. I fixed the ICE by scoping per note fluent args in `RegionOriginNote::WithRequirement` with `store_args` and `restore_args`.
Rename `rustc::pass_by_value` lint as `rustc::disallowed_pass_by_ref`.
The name `pass_by_value` is completely wrong. The lint actually checks for the use of pass by reference for types marked with `rustc_pass_by_value`.
The hardest part of this was choosing the new name. The `disallowed_` part of the name closely matches the following clippy lints:
- `disallowed_macros`
- `disallowed_methods`
- `disallowed_names`
- `disallowed_script_idents`
- `disallowed_types`
The `pass_by_value` part of the name aligns with the following clippy lints:
- `needless_pass_by_value`
- `needless_pass_by_ref_mut`
- `trivially_copy_pass_by_ref`
- `large_types_passed_by_value` (less so)
r? @Urgau
deprecate `Eq::assert_receiver_is_total_eq` and emit FCW on manual impls
The `Eq::assert_receiver_is_total_eq` method is purely meant as an implementation detail by `#[derive(Eq)]` to add checks that all fields of the type the derive is applied to also implement `Eq`.
The method is already `#[doc(hidden)]` and has a comment saying `// This should never be implemented by hand.`.
Unfortunately, it has been stable since 1.0 and there are some cases on GitHub (https://github.com/search?q=assert_receiver_is_total_eq&type=code) where people have implemented this method manually, sometimes even with actual code in the method body (example: https://github.com/Shresht7/codecrafters-redis-rust/blob/31f0ec453c504b4ab053a7b1c3ff548ff36a9db5/src/parser/resp/types.rs#L255).
To prevent further confusion from this, this PR is deprecating the method and adds a FCW when it is manually implemented (this is necessary as the deprecation warning is not emitted when the method is implemented, only when it is called).
This is similar to what was previously done with the `soft_unstable` lint (https://github.com/rust-lang/rust/issues/64266).
See also https://github.com/rust-lang/libs-team/issues/704.
Fix: On wasm targets, call `panic_in_cleanup` if panic occurs in cleanup
Previously this was not correctly implemented. Each funclet may need its own terminate block, so this changes the `terminate_block` into a `terminate_blocks` `IndexVec` which can have a terminate_block for each funclet. We key on the first basic block of the funclet -- in particular, this is the start block for the old case of the top level terminate function.
I also fixed the `terminate` handler to not be invoked when a foreign exception is raised, mimicking the behavior from msvc. On wasm, in order to avoid generating a `catch_all` we need to call `llvm.wasm.get.exception` and `llvm.wasm.get.ehselector`.
The name `pass_by_value` is completely wrong. The lint actually checks
for the use of pass by reference for types marked with
`rustc_pass_by_value`.
The hardest part of this was choosing the new name. The `disallowed_`
part of the name closely matches the following clippy lints:
- `disallowed_macros`
- `disallowed_methods`
- `disallowed_names`
- `disallowed_script_idents`
- `disallowed_types`
The `pass_by_value` part of the name aligns with the following clippy
lints:
- `needless_pass_by_value`
- `needless_pass_by_ref_mut`
- `trivially_copy_pass_by_ref`
- `large_types_passed_by_value` (less so)
Fix attribute parser and kind names.
For the attribute `FooBar` the parser is generally called `FooBarParser` and the kind is called `AttributeKind::FooBar`. This commit renames some cases that don't match that pattern. The most common cases:
- Adding `Rustc` to the front of the parser name for a `rustc_*` attribute.
- Adding `Parser` to the end of a parser name.
- Slight word variations, e.g. `Deprecation` instead of `Deprecated`, `Pointer` instead of `Ptr`, `Stability` instead of `Stable`.
r? @JonathanBrouwer
delete some very old trivial `Box` tests
There's still lots more of these tests; these are just a few duplicated ones I found. None of the tests really test for something interesting except that `Deref{Mut}` trivially works with boxes, but there are still loads of more tests that do the same.
In the past, these tests presumably looked more different to each other than today, with `Box`-like pointers being primitives etc.
This PR is also to check whether such changes (deleting of useless tests) are wanted in general.
Part of rust-lang/rust#133895.
Implement debuginfo for unsafe binder types
Fixes: rust-lang/rust#139462
This treats an unsafe binder like a struct with a single field. This way we'd have the binder's distinct type name while keeping the wrapped value accessible.
Tracking:
- https://github.com/rust-lang/rust/issues/130516
Fix relative path handling for --extern-html-root-url
When `--extern-html-root-url` receives a relative path (like `../`), rustdoc uses it as a literal prefix regardless of page depth. This works at depth 1 (`crate/index.html`) but breaks at depth 2+ (`crate/module/struct.Foo.html`) because the relative path doesn't account for the extra nesting.
This patch detects relative vs absolute URLs in the `Remote` branch of `url_parts` and prepends the necessary `../` segments based on the current page depth. Absolute URLs (`https://...`) and server-absolute paths (`/docs/...`) are unchanged.
This makes relative paths a viable option for `--extern-html-root-url`, which they weren't before.
Related to rust-lang/rust#152917, which exposed that relative paths weren't handled correctly.
cc @eggyal
Rémy Rakic