Enable `f16` tests on x86 and x86-64
Since the `compiler_builtins` update [1], ABI bugs on x86 should be resolved. Enable tests for f16 on these platforms now.
`f16` math functions (`reliable_f16_math`) are still excluded because there is an LLVM crash for powi [2].
[1]: https://github.com/rust-lang/rust/pull/125016
[2]: https://github.com/llvm/llvm-project/issues/105747
try-job: dist-i586-gnu-i586-i686-musl
try-job: x86_64-apple-1
epoll: handle edge case for epoll_ctl
There is a test case that revealed that our implementation differs from the real system:
- Set up an epoll watching the FD
- Call epoll_wait
- Set up another epoll watching the same FD
- Call epoll_wait on the first epoll. Nothing should be reported!
This happened because, in ``epoll_ctl``, we used ``check_and_update_readiness``, which is a function that would return notification for all epoll file description that registered a particular file description. But we shouldn't do that because no notification should be returned if there is no I/O activity between two ``epoll_wait`` (every first ``epoll_wait`` that happens after ``epoll_ctl`` is an exception, we should return notification that reflects the readiness of file description).
r? `@oli-obk`
epoll: Add a EINVAL case
In ``epoll_ctl`` documentation, it is mentioned that:
> EINVAL epfd is not an epoll file descriptor, or fd is the same as epfd, or the requested operation op is not supported by this interface.
So I added this EINVAL case for ``epfd == fd`` in ``epoll_ctl``
fix: rust-analyzer should watch build files from rust-project.json
rust-analyzer always watches Cargo.toml for changes, but other build systems using rust-project.json have their own build files.
Ensure we also watch those for changes, so we know when to reconfigure rust-analyzer when dependencies change.
- `new_zeroed` variants move to `new_zeroed_alloc`
- the `write` fn moves to `box_uninit_write`
The remainder will be stabilized in upcoming patches, as
it was decided to only stabilize `uninit*` and `assume_init`.
fix: Wrong `Self: Sized` predicate for trait assoc items
Again while implementing object safety like #17939😅
If we call `generic_predicates_query` on `fn foo` in the following code;
```
trait Foo {
fn foo();
}
```
It returns implicit bound `Self: Sized`, even though `Self` is not appearing as a generic parameter inside angle brackets, but as a parent generic parameter, "trait self".
This PR prevent pushing "implicit" `Self: Sized` predicates in such cases
Documents that `BikeshedIntrinsicFrom` models transmute-via-union,
which is slightly more expressive than the transmute-via-cast
implemented by `transmute_copy`. Additionally, we provide an
implementation of transmute-via-union as a method on the
`BikeshedIntrinsicFrom` trait with additional documentation on
the boundary between trait invariants and caller obligations.
Whether or not transmute-via-union is the right kind of transmute
to model remains up for discussion [1]. Regardless, it seems wise
to document the present behavior.
[1] https://rust-lang.zulipchat.com/#narrow/stream/216762-project-safe-transmute/topic/What.20'kind'.20of.20transmute.20to.20model.3F/near/426331967
This change allows setting either `build.cargo` or `build.rustc` without requiring
both to be set simultaneously, which was not possible previously.
To try it, set `build.rustc` without setting `build.cargo`, and try to bootstrap on clean build.
Signed-off-by: onur-ozkan <work@onurozkan.dev>
Rollup of 9 pull requests
Successful merges:
- #128511 (Document WebAssembly target feature expectations)
- #129243 (do not build `cargo-miri` by default on stable channel)
- #129263 (Add a missing compatibility note in the 1.80.0 release notes)
- #129276 (Stabilize feature `char_indices_offset`)
- #129350 (adapt integer comparison tests for LLVM 20 IR changes)
- #129408 (Fix handling of macro arguments within the `dropping_copy_types` lint)
- #129426 (rustdoc-search: use tighter json for names and parents)
- #129437 (Fix typo in a help diagnostic)
- #129457 (kobzol vacation)
r? `@ghost`
`@rustbot` modify labels: rollup
Fix handling of macro arguments within the `dropping_copy_types` lint
This PR fixes the handling of spans with different context (aka macro arguments) than the primary expression within the different `{drop,forget}ing_copy_types` and `{drop,forget}ing_references` lints.
<details>
<summary>Before</summary>
```
warning: calls to `std::mem::drop` with a value that implements `Copy` does nothing
--> drop_writeln.rs:5:5
|
5 | drop(writeln!(&mut msg, "test"));
| ^^^^^--------------------------^
| |
| argument has type `Result<(), std::fmt::Error>`
|
= note: `#[warn(dropping_copy_types)]` on by default
help: use `let _ = ...` to ignore the expression or result
--> /home/[..]/.rustup/toolchains/nightly-x86_64-unknown-linux-gnu/lib/rustlib/src/rust/library/core/src/macros/mod.rs:688:9
|
68| let _ =
| ~~~~~~~
```
</details>
<details>
<summary>With this PR</summary>
```
warning: calls to `std::mem::drop` with a value that implements `Copy` does nothing
--> drop_writeln.rs:5:5
|
5 | drop(writeln!(&mut msg, "test"));
| ^^^^^--------------------------^
| |
| argument has type `Result<(), std::fmt::Error>`
|
= note: `#[warn(dropping_copy_types)]` on by default
help: use `let _ = ...` to ignore the expression or result
|
5 - drop(writeln!(&mut msg, "test"));
5 + let _ = writeln!(&mut msg, "test");
|
```
</details>
``````@rustbot`````` label +L-dropping_copy_types
Stabilize feature `char_indices_offset`
Stabilized API:
```rust
impl CharIndices<'_> {
pub fn offset(&self) -> usize;
}
```
Tracking issue: https://github.com/rust-lang/rust/issues/83871
Closes https://github.com/rust-lang/rust/issues/83871
I also attempted to improved the documentation to make it more clear that it returns the offset of the character that will be returned by the next call to `next()`.