Based on @paolobarbolini's tip that the unsafe block was unnecessary in
this case.
Not much left of `hexify()` after this, so seemed clearer to just inline
it.
To make this work, the `#[rustc_on_unimplemented]` data needs to be used to
report method resolution errors, which is most of what this commit does.
Fixes#94581
The initial stdlib modifications for L4Re just used the linux specifics
directly because they were reasonably close to L4Re's behavior.
However, this breaks when Linux-specific code relies on code that is
only available for the linux target, such as in #81825.
Put L4Re into its own platform to avoid such breakage in the future.
This uses the Linux-specific code as a starting point, which seems to be
in line with other OSes with a unix-y interface such as Fuchsia.
L4Re provides limited POSIX support which includes support for
standard I/O streams, and a limited implementation of the standard file
handling API. However, because as a capability based OS it strives to
only make a local view available to each application, there are
currently no standardized special files like /dev/null that could serve
to sanitize closed standard FDs.
For now, skip any attempts to sanitize standard streams until a more
complete POSIX runtime is available.
unix: reduce the size of DirEntry
On platforms where we call `readdir` instead of `readdir_r`, we store
the name as an allocated `CString` for variable length. There's no point
carrying around a full `dirent64` with its fixed-length `d_name` too.
Reverted atomic_mut_ptr feature removal causing compilation break
Fixes a regression introduced as part of https://github.com/rust-lang/rust/pull/94546
Std no longer compiles on nightly while using the following commnd:
export RUSTFLAGS='-C target-feature=+atomics,+bulk-memory'
cargo build --target wasm32-unknown-unknown -Z build-std=panic_abort,std
I can help add tests to avoid future breaks but i couldn't understand the test framework
unix: Avoid name conversions in `remove_dir_all_recursive`
Each recursive call was creating an `OsString` for a `&Path`, only for
it to be turned into a `CString` right away. Instead we can directly
pass `.name_cstr()`, saving two allocations each time.
Add core::hint::must_use
The example code in this documentation is minimized from a real-world situation in the `anyhow` crate where this function would have been valuable.
Having this provided by the standard library is especially useful for proc macros, even more than for macro_rules. That's because proc macro crates aren't allowed to export anything other than macros, so they couldn't make their own `must_use` function for their macro-generated code to call.
<br>
## Rendered documentation
> An identity function that causes an `unused_must_use` warning to be triggered if the given value is not used (returned, stored in a variable, etc) by the caller.
>
> This is primarily intended for use in macro-generated code, in which a [`#[must_use]` attribute][must_use] either on a type or a function would not be convenient.
>
> [must_use]: https://doc.rust-lang.org/reference/attributes/diagnostics.html#the-must_use-attribute
>
> ### Example
>
> ```rust
> #![feature(hint_must_use)]
>
> use core::fmt;
>
> pub struct Error(/* ... */);
>
> #[macro_export]
> macro_rules! make_error {
> ($($args:expr),*) => {
> core::hint::must_use({
> let error = $crate::make_error(core::format_args!($($args),*));
> error
> })
> };
> }
>
> // Implementation detail of make_error! macro.
> #[doc(hidden)]
> pub fn make_error(args: fmt::Arguments<'_>) -> Error {
> Error(/* ... */)
> }
>
> fn demo() -> Option<Error> {
> if true {
> // Oops, meant to write `return Some(make_error!("..."));`
> Some(make_error!("..."));
> }
> None
> }
> ```
>
> In the above example, we'd like an `unused_must_use` lint to apply to the value created by `make_error!`. However, neither `#[must_use]` on a struct nor `#[must_use]` on a function is appropriate here, so the macro expands using `core::hint::must_use` instead.
>
> - We wouldn't want `#[must_use]` on the `struct Error` because that would make the following unproblematic code trigger a warning:
>
> ```rust
> fn f(arg: &str) -> Result<(), Error>
>
> #[test]
> fn t() {
> // Assert that `f` returns error if passed an empty string.
> // A value of type `Error` is unused here but that's not a problem.
> f("").unwrap_err();
> }
> ```
>
> - Using `#[must_use]` on `fn make_error` can't help because the return value *is* used, as the right-hand side of a `let` statement. The `let` statement looks useless but is in fact necessary for ensuring that temporaries within the `format_args` expansion are not kept alive past the creation of the `Error`, as keeping them alive past that point can cause autotrait issues in async code:
>
> ```rust
> async fn f() {
> // Using `let` inside the make_error expansion causes temporaries like
> // `unsync()` to drop at the semicolon of that `let` statement, which
> // is prior to the await point. They would otherwise stay around until
> // the semicolon on *this* statement, which is after the await point,
> // and the enclosing Future would not implement Send.
> log(make_error!("look: {:p}", unsync())).await;
> }
>
> async fn log(error: Error) {/* ... */}
>
> // Returns something without a Sync impl.
> fn unsync() -> *const () {
> 0 as *const ()
> }
> ```
Enable `close_read_wakes_up` test on Windows
I wonder if we could/should try enabling this again? It was closed by #38867 due to #31657. I've tried running this test (along with other tests) on my machine a number of times and haven't seen this fail yet,
Caveat: the worst that can happen is this succeeds initially but then causes random hangs in CI. This is not a great failure mode and would be a reason not to do this.
If this does work out, closes#39006
r? `@Mark-Simulacrum`
On platforms where we call `readdir` instead of `readdir_r`, we store
the name as an allocated `CString` for variable length. There's no point
carrying around a full `dirent64` with its fixed-length `d_name` too.
Remove argument from closure in thread::Scope::spawn.
This implements ```@danielhenrymantilla's``` [suggestion](https://github.com/rust-lang/rust/issues/93203#issuecomment-1040798286) for improving the scoped threads interface.
Summary:
The `Scope` type gets an extra lifetime argument, which represents basically its own lifetime that will be used in `&'scope Scope<'scope, 'env>`:
```diff
- pub struct Scope<'env> { .. };
+ pub struct Scope<'scope, 'env: 'scope> { .. }
pub fn scope<'env, F, T>(f: F) -> T
where
- F: FnOnce(&Scope<'env>) -> T;
+ F: for<'scope> FnOnce(&'scope Scope<'scope, 'env>) -> T;
```
This simplifies the `spawn` function, which now no longer passes an argument to the closure you give it, and now uses the `'scope` lifetime for everything:
```diff
- pub fn spawn<'scope, F, T>(&'scope self, f: F) -> ScopedJoinHandle<'scope, T>
+ pub fn spawn<F, T>(&'scope self, f: F) -> ScopedJoinHandle<'scope, T>
where
- F: FnOnce(&Scope<'env>) -> T + Send + 'env,
+ F: FnOnce() -> T + Send + 'scope,
- T: Send + 'env;
+ T: Send + 'scope;
```
The only difference the user will notice, is that their closure now takes no arguments anymore, even when spawning threads from spawned threads:
```diff
thread::scope(|s| {
- s.spawn(|_| {
+ s.spawn(|| {
...
});
- s.spawn(|s| {
+ s.spawn(|| {
...
- s.spawn(|_| ...);
+ s.spawn(|| ...);
});
});
```
<details><summary>And, as a bonus, errors get <em>slightly</em> better because now any lifetime issues point to the outermost <code>s</code> (since there is only one <code>s</code>), rather than the innermost <code>s</code>, making it clear that the lifetime lasts for the entire <code>thread::scope</code>.
</summary>
```diff
error[E0373]: closure may outlive the current function, but it borrows `a`, which is owned by the current function
--> src/main.rs:9:21
|
- 7 | s.spawn(|s| {
- | - has type `&Scope<'1>`
+ 6 | thread::scope(|s| {
+ | - lifetime `'1` appears in the type of `s`
9 | s.spawn(|| println!("{:?}", a)); // might run after `a` is dropped
| ^^ - `a` is borrowed here
| |
| may outlive borrowed value `a`
|
note: function requires argument type to outlive `'1`
--> src/main.rs:9:13
|
9 | s.spawn(|| println!("{:?}", a)); // might run after `a` is dropped
| ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
help: to force the closure to take ownership of `a` (and any other referenced variables), use the `move` keyword
|
9 | s.spawn(move || println!("{:?}", a)); // might run after `a` is dropped
| ++++
"
```
</details>
The downside is that the signature of `scope` and `Scope` gets slightly more complex, but in most cases the user wouldn't need to write those, as they just use the argument provided by `thread::scope` without having to name its type.
Another downside is that this does not work nicely in Rust 2015 and Rust 2018, since in those editions, `s` would be captured by reference and not by copy. In those editions, the user would need to use `move ||` to capture `s` by copy. (Which is what the compiler suggests in the error.)
Add Result::{ok, err, and, or, unwrap_or} as const
Already opened tracking issue #92384.
I don't think that this should actually cause any issues as long as the constness is unstable, but we may want to double-check that this doesn't get interpreted as a weird `Drop` bound even for non-const usages.
Each recursive call was creating an `OsString` for a `&Path`, only for
it to be turned into a `CString` right away. Instead we can directly
pass `.name_cstr()`, saving two allocations each time.
Stabilize const_fn_fn_ptr_basics, const_fn_trait_bound, and const_impl_trait
# Stabilization Report
This PR serves as a request for stabilization for three const evaluation features:
1. `const_fn_fn_ptr_basics`
2. `const_fn_trait_bound`
3. `const_impl_trait`
These are being stabilized together because they are relatively minor and related updates to existing functionality.
## `const_fn_fn_ptr_basics`
Allows creating, passing, and casting function pointers in a `const fn`.
The following is an example of what is now allowed:
```rust
const fn get_function() -> fn() {
fn foo() {
println!("Hello, World!");
}
foo
}
```
Casts between function pointer types are allowed, as well as transmuting from integers:
```rust
const fn get_function() -> fn() {
unsafe {
std::mem::transmute(0x1234usize)
}
}
```
However, casting from a function pointer to an integer is not allowed:
```rust
const fn fn_to_usize(f: fn()) -> usize {
f as usize //~ pointers cannot be cast to integers during const eval
}
```
Calling function pointers is also not allowed.
```rust
const fn call_fn_ptr(f: fn()) {
f() //~ function pointers are not allowed in const fn
}
```
### Test Coverage
The following tests include code that exercises this feature:
- `src/test/ui/consts/issue-37550.rs`
- `src/test/ui/consts/issue-46553.rs`
- `src/test/ui/consts/issue-56164.rs`
- `src/test/ui/consts/min_const_fn/allow_const_fn_ptr_run_pass.rs`
- `src/test/ui/consts/min_const_fn/cast_fn.rs`
- `src/test/ui/consts/min_const_fn/cmp_fn_pointers.rs`
## `const_fn_trait_bound`
Allows trait bounds in `const fn`. Additionally, this feature allows creating and passing `dyn Trait` objects.
Examples such as the following are allowed by this feature:
```rust
const fn do_thing<T: Foo>(_x: &T) {
// ...
}
```
Previously only `Sized` was allowed as a trait bound.
There is no way to call methods from the trait because trait methods cannot currently be marked as const. Allowing trait bounds in const functions does allow the const function to use the trait's associated types and constants.
This feature also allowes `dyn Trait` types. These work equivalently to non-const code. Similar to other pointers in const code, the value of a `dyn Trait` pointer cannot be observed.
Note that due to https://github.com/rust-lang/rust/issues/90912, it was already possible to do the example above as follows:
```rust
const fn do_thing<T>(_x: &T) where (T,): Foo {
// ...
}
```
### Test Coverage
The following tests include code that exercises `const_fn_trait_bound`:
- `src/test/ui/consts/const-fn.rs`
- `src/test/ui/consts/issue-88071.rs`
- `src/test/ui/consts/min_const_fn/min_const_fn.rs`
- `src/test/ui/consts/min_const_fn/min_const_fn_dyn.rs`
- `src/test/ui/nll/issue-55825-const-fn.rs`
- Many of the tests in `src/test/ui/rfc-2632-const-trait-impl/` also exercise this feature.
## `const_impl_trait`
Allows argument and return position `impl Trait` in a `const fn`, such as in the following example:
```rust
const fn do_thing(x: impl Foo) -> impl Foo {
x
}
```
Similar to generic parameters and function pointers, this allows the creation of such opaque types, but not doing anything with them beyond accessing associated types and constants.
### Test Coverage
The following tests exercise this feature:
- `src/test/ui/type-alias-impl-trait/issue-53096.rs`
- `src/test/ui/type-alias-impl-trait/issue-53678-generator-and-const-fn.rs`
## Documentation
These features are documented along with the other const evaluation features in the Rust Reference at https://doc.rust-lang.org/stable/reference/const_eval.html.
There is a PR that updates this documentation to reflect the capabilities enabled by these features at https://github.com/rust-lang/reference/pull/1166.
Tracking issues: #57563, #63997, #93706
libunwind: readd link attrs to _Unwind_Backtrace
It seems the removal of these in 1c07096a45 was unintended; readding them fixes the build.
fixesrust-lang/rust#93349
r? `@alexcrichton`
Unix path::absolute: Fix leading "." component
Testing leading `.` and `..` components were missing from the unix tests.
This PR adds them and fixes the leading `.` case. It also fixes the test cases so that they do an exact comparison.
This problem reported by ``@axetroy``
Windows 11 no longer turn paths ending with dos device names into device paths.
E.g. `C:\path\to\COM1.txt` used to get turned into `\\.\COM1`. Whereas now the path is left as is.
UNIX `remove_dir_all()`: Try recursing first on the slow path
This only affects the _slow_ code path - if there is no `dirent.d_type` or if it is `DT_UNKNOWN`.
POSIX specifies that calling `unlink()` or `unlinkat(..., 0)` on a directory is allowed to succeed:
> The _path_ argument shall not name a directory unless the process has appropriate privileges and the implementation supports using _unlink()_ on directories.
This however can cause dangling inodes requiring an fsck e.g. on Illumos UFS, so we have to avoid that in the common case. We now just try to recurse into it first and unlink() if we can't open it as a directory.
The other two commits integrate the Macos x86-64 implementation reducing redundancy. Split into two commits for better reviewing.
Fixes#94335.
Rollup of 5 pull requests
Successful merges:
- #94362 (Add well known values to `--check-cfg` implementation)
- #94577 (only disable SIMD for doctests in Miri (not for the stdlib build itself))
- #94595 (Fix invalid `unresolved imports` errors for a single-segment import)
- #94596 (Delay bug in expr adjustment when check_expr is called multiple times)
- #94618 (Don't round stack size up for created threads in Windows)
Failed merges:
r? `@ghost`
`@rustbot` modify labels: rollup
only disable SIMD for doctests in Miri (not for the stdlib build itself)
Also we can enable library/core/tests/simd.rs now, Miri supports enough SIMD for that.
Enable conditional compilation checking on the Rust codebase
This pull-request enable conditional compilation checking on every rust project build by the `bootstrap` tool.
To be more specific, this PR only enable well known names checking + extra names (bootstrap, parallel_compiler, ...).
r? `@Mark-Simulacrum`
Add #[track_caller] to track callers when initializing poisoned Once
This PR is for this Issue.
https://github.com/rust-lang/rust/issues/87707
With this fix, we expect to be able to track the caller when poisoned Once is initialized.
Before calling `CreateProcessW`, stdio handles are passed through
`stdio::get_handle`, which already converts NULL to
`INVALID_HANDLE_VALUE`, so we don't need extra checks for NULL after
that point.
This addresses #90964 by making the std API consistent about presenting
absent stdio handles on Windows as NULL handles. Stdio handles may be
absent due to `#![windows_subsystem = "windows"]`, due to the console
being detached, or due to a child process having been launched from a
parent where stdio handles are absent.
Specifically, this fixes the case of child processes of parents with absent
stdio, which previously ended up with `stdin().as_raw_handle()` returning
`INVALID_HANDLE_VALUE`, which was surprising, and which overlapped with an
unrelated valid handle value. With this patch, `stdin().as_raw_handle()`
now returns null in these situation, which is consistent with what it
does in the parent process.
And, document this in the "Windows Portability Considerations" sections of
the relevant documentation.
This only affects the `slow` code path, if there is no `dirent.d_type` or if
the type is `DT_UNKNOWN`.
POSIX specifies that calling `unlink()` or `unlinkat(..., 0)` on a directory can
succeed:
> "The _path_ argument shall not name a directory unless the process has
> appropriate privileges and the implementation supports using _unlink()_ on
> directories."
This however can cause orphaned directories requiring an fsck e.g. on Illumos
UFS, so we have to avoid that in the common case. We now just try to recurse
into it first and unlink() if we can't open it as a directory.
Use `HandleOrNull` and `HandleOrInvalid` in the Windows FFI bindings.
Use the new `HandleOrNull` and `HandleOrInvalid` types that were introduced
as part of [I/O safety] in a few functions in the Windows FFI bindings.
This factors out an `unsafe` block and two `unsafe` function calls in the
Windows implementation code.
And, it helps test `HandleOrNull` and `HandleOrInvalid`, and indeed, it
turned up a bug: `OwnedHandle` also needs to be `#[repr(transparent)]`,
as it's used inside of `HandleOrNull` and `HandleOrInvalid` which are also
`#[repr(transparent)]`.
r? ```@joshtriplett```
[I/O safety]: https://github.com/rust-lang/rust/issues/87074
Make regular stdio lock() return 'static handles
This also deletes the unstable API surface area previously added to expose this
functionality on new methods rather than built into the current set.
Closes#86845 (tracking issue for unstable API needed without this)
r? ``````@dtolnay`````` to kick off T-libs-api FCP
Clarification of default socket flags
This PR outlines the decision to disable inheritance of socket objects when possible to child processes in the documentation.
Use the new `HandleOrNull` and `HandleOrInvalid` types that were introduced
as part of [I/O safety] in a few functions in the Windows FFI bindings.
This factors out an `unsafe` block and two `unsafe` function calls in the
Windows implementation code.
And, it helps test `HandleOrNull` and `HandleOrInvalid`, which indeed turned
up a bug: `OwnedHandle` also needs to be `#[repr(transparent)]`, as it's
used inside of `HandleOrNull` and `HandleOrInvalid` which are also
`#[repr(transparent)]`.
[I/O safety]: https://github.com/rust-lang/rust/issues/87074
Miri/CTFE: properly treat overflow in (signed) division/rem as UB
To my surprise, it looks like LLVM treats overflow of signed div/rem as UB. From what I can tell, MIR `Div`/`Rem` directly lowers to the corresponding LLVM operation, so to make that correct we also have to consider these overflows UB in the CTFE/Miri interpreter engine.
r? `@oli-obk`
Update the documentation for `{As,Into,From}Raw{Fd,Handle,Socket}`.
This change weakens the descriptions of the
`{as,into,from}_raw_{fd,handle,socket}` descriptions from saying that
they *do* express ownership relations to say that they are *typically used*
in ways that express ownership relations. This is needed since, for
example, std's own [`RawFd`] implements `{As,From,Into}Fd` without any of
the ownership relationships.
This adds proper `# Safety` comments to `from_raw_{fd,handle,socket}`,
adds the requirement that raw handles be not opened with the
`FILE_FLAG_OVERLAPPED` flag, and merges the `OwnedHandle::from_raw_handle`
comment into the main `FromRawHandle::from_raw_handle` comment.
And, this changes `HandleOrNull` and `HandleOrInvalid` to not implement
`FromRawHandle`, since they are intended for limited use in FFI situations,
and not for generic use, and they have constraints that are stronger than
the those of `FromRawHandle`.
[`RawFd`]: https://doc.rust-lang.org/stable/std/os/unix/io/type.RawFd.html
There may eventually be something to say about `FILE_FLAG_OVERLAPPED` here,
however this appears to be independent of the other changes in this PR,
so remove them from this PR so that it can be discussed separately.
Rename `BorrowedFd::borrow_raw_fd` to `BorrowedFd::borrow_raw`.
Also, rename `BorrowedHandle::borrow_raw_handle` and
`BorrowedSocket::borrow_raw_socket` to `BorrowedHandle::borrow_raw` and
`BorrowedSocket::borrow_raw`.
This is just a minor rename to reduce redundancy in the user code calling
these functions, and to eliminate an inessential difference between
`BorrowedFd` code and `BorrowedHandle`/`BorrowedSocket` code.
While here, add a simple test exercising `BorrowedFd::borrow_raw_fd`.
r? ``````@joshtriplett``````
Add documentation about `BorrowedFd::to_owned`.
Following up on #88564, this adds documentation explaining why
`BorrowedFd::to_owned` returns another `BorrowedFd` rather than an
`OwnedFd`. And similar for `BorrowedHandle` and `BorrowedSocket`.
r? `````@joshtriplett`````
this avoids parsing mountinfo which can be huge on some systems and
something might be emulating cgroup fs for sandboxing reasons which means
it wouldn't show up as mountpoint
additionally the new implementation operates on a single pathbuffer, reducing allocations
Manually tested via
```
// spawn a new cgroup scope for the current user
$ sudo systemd-run -p CPUQuota="300%" --uid=$(id -u) -tdS
// quota.rs
#![feature(available_parallelism)]
fn main() {
println!("{:?}", std:🧵:available_parallelism()); // prints Ok(3)
}
```
Caveats
* cgroup v1 is ignored
* funky mountpoints (containing spaces, newlines or control chars) for cgroupfs will not be handled correctly since that would require unescaping /proc/self/mountinfo
The escaping behavior of procfs seems to be undocumented. systemd and docker default to `/sys/fs/cgroup` so it should be fine for most systems.
* quota will be ignored when `sched_getaffinity` doesn't work
* assumes procfs is mounted under `/proc` and cgroupfs mounted and readable somewhere in the directory tree
When CStr moves to core with an alias in std, this can link to
`crate::ffi::CStr`. However, linking in the reverse direction (from core
to std) requires a relative path, and that path can't work from both
core::ffi and std::os::raw (different number of `../` traversals
required).
The ability to interoperate with C code via FFI is not limited to crates
using std; this allows using these types without std.
The existing types in `std::os::raw` become type aliases for the ones in
`core::ffi`. This uses type aliases rather than re-exports, to allow the
std types to remain stable while the core types are unstable.
This also moves the currently unstable `NonZero_` variants and
`c_size_t`/`c_ssize_t`/`c_ptrdiff_t` types to `core::ffi`, while leaving
them unstable.
core can't depend on external crates the way std can. Rather than revert
usage of cfg_if, add a copy of it to core. This does not export our
copy, even unstably; such a change could occur in a later commit.
Sync portable-simd for bitmasks &c.
In the ideal case, where everything works easily and nothing has to be rearranged, it is as simple as:
- `git subtree pull -P library/portable-simd https://github.com/rust-lang/portable-simd - ${branch}`
- write the commit message
- `python x.py test --stage 1` to make sure it runs
- `git push` to your PR-to-rustc branch
If anything borks up this flow, you can fix it with sufficient git wizardry but you are usually better off going back to the source, fixing it, and starting over, before you open the PR.
r? `@calebzulawski`
Add Atomic*::from_mut_slice
Tracking issue #76314 for `from_mut` has a question about the possibility of `from_mut_slice`, and I found a real case for it. A user in the forum had a parallelism problem that could be solved by open-indexing updates to a vector of atomics, but they didn't want to affect the other code using that vector. Using `from_mut_slice`, they could borrow that data as atomics just long enough for their parallel loop.
ref: https://users.rust-lang.org/t/sharing-vector-with-rayon-par-iter-correctly/72022
use BOOL for TCP_NODELAY setsockopt value on Windows
This issue was found by the Wine project and mitigated there [^1].
Windows' setsockopt expects a BOOL (a typedef for int) for TCP_NODELAY
[^2]. Windows itself is forgiving and will accept any positive optlen and
interpret the first byte of *optval as the value, so this bug does not
affect Windows itself, but does affect systems implementing Windows'
interface more strictly, such as Wine. Wine was previously passing this
through to the host's setsockopt, where, e.g., Linux requires that
optlen be correct for the chosen option, and TCP_NODELAY expects an int.
[^1]: d6ea38f32d
[^2]: https://docs.microsoft.com/en-us/windows/win32/api/winsock/nf-winsock-setsockopt
fix typo in btree/vec doc: Self -> self
this pr fixes#92345
the documentation refers to the object the method is called for, not the type, so it should be using the lower case self.
For MIRI, cfg out the swap vectorization logic from 94212
Because of #69488 the swap logic from #94212 doesn't currently work in MIRI.
Copying in smaller pieces is probably much worse for its performance anyway, so it'd probably rather just use the simple path regardless.
Part of #94371, though another PR will be needed for the CTFE aspect.
r? `@oli-obk`
cc `@RalfJung`
Rollup of 5 pull requests
Successful merges:
- #93603 (Populate liveness facts when calling `get_body_with_borrowck_facts` without `-Z polonius`)
- #93870 (Fix switch on discriminant detection in a presence of coverage counters)
- #94355 (Add one more case to avoid ICE)
- #94363 (Remove needless borrows from core::fmt)
- #94377 (`check_used` should only look at actual `used` attributes)
Failed merges:
r? `@ghost`
`@rustbot` modify labels: rollup
This function was updated in a recent PR (92911) to be called without the caller
information passed in, but the function signature itself was not altered with
cfg_attr at the time.
As an example:
#[test]
#[ignore = "not yet implemented"]
fn test_ignored() {
...
}
Will now render as:
running 2 tests
test tests::test_ignored ... ignored, not yet implemented
test result: ok. 1 passed; 0 failed; 1 ignored; 0 measured; 0 filtered out; finished in 0.00s
Stop manually SIMDing in `swap_nonoverlapping`
Like I previously did for `reverse` (#90821), this leaves it to LLVM to pick how to vectorize it, since it can know better the chunk size to use, compared to the "32 bytes always" approach we currently have.
A variety of codegen tests are included to confirm that the various cases are still being vectorized.
It does still need logic to type-erase in some cases, though, as while LLVM is now smart enough to vectorize over slices of things like `[u8; 4]`, it fails to do so over slices of `[u8; 3]`.
As a bonus, this change also means one no longer gets the spurious `memcpy`(s?) at the end up swapping a slice of `__m256`s: <https://rust.godbolt.org/z/joofr4v8Y>
<details>
<summary>ASM for this example</summary>
## Before (from godbolt)
note the `push`/`pop`s and `memcpy`
```x86
swap_m256_slice:
push r15
push r14
push r13
push r12
push rbx
sub rsp, 32
cmp rsi, rcx
jne .LBB0_6
mov r14, rsi
shl r14, 5
je .LBB0_6
mov r15, rdx
mov rbx, rdi
xor eax, eax
.LBB0_3:
mov rcx, rax
vmovaps ymm0, ymmword ptr [rbx + rax]
vmovaps ymm1, ymmword ptr [r15 + rax]
vmovaps ymmword ptr [rbx + rax], ymm1
vmovaps ymmword ptr [r15 + rax], ymm0
add rax, 32
add rcx, 64
cmp rcx, r14
jbe .LBB0_3
sub r14, rax
jbe .LBB0_6
add rbx, rax
add r15, rax
mov r12, rsp
mov r13, qword ptr [rip + memcpy@GOTPCREL]
mov rdi, r12
mov rsi, rbx
mov rdx, r14
vzeroupper
call r13
mov rdi, rbx
mov rsi, r15
mov rdx, r14
call r13
mov rdi, r15
mov rsi, r12
mov rdx, r14
call r13
.LBB0_6:
add rsp, 32
pop rbx
pop r12
pop r13
pop r14
pop r15
vzeroupper
ret
```
## After (from my machine)
Note no `rsp` manipulation, sorry for different ASM syntax
```x86
swap_m256_slice:
cmpq %r9, %rdx
jne .LBB1_6
testq %rdx, %rdx
je .LBB1_6
cmpq $1, %rdx
jne .LBB1_7
xorl %r10d, %r10d
jmp .LBB1_4
.LBB1_7:
movq %rdx, %r9
andq $-2, %r9
movl $32, %eax
xorl %r10d, %r10d
.p2align 4, 0x90
.LBB1_8:
vmovaps -32(%rcx,%rax), %ymm0
vmovaps -32(%r8,%rax), %ymm1
vmovaps %ymm1, -32(%rcx,%rax)
vmovaps %ymm0, -32(%r8,%rax)
vmovaps (%rcx,%rax), %ymm0
vmovaps (%r8,%rax), %ymm1
vmovaps %ymm1, (%rcx,%rax)
vmovaps %ymm0, (%r8,%rax)
addq $2, %r10
addq $64, %rax
cmpq %r10, %r9
jne .LBB1_8
.LBB1_4:
testb $1, %dl
je .LBB1_6
shlq $5, %r10
vmovaps (%rcx,%r10), %ymm0
vmovaps (%r8,%r10), %ymm1
vmovaps %ymm1, (%rcx,%r10)
vmovaps %ymm0, (%r8,%r10)
.LBB1_6:
vzeroupper
retq
```
</details>
This does all its copying operations as either the original type or as `MaybeUninit`s, so as far as I know there should be no potential abstract machine issues with reading padding bytes as integers.
<details>
<summary>Perf is essentially unchanged</summary>
Though perhaps with more target features this would help more, if it could pick bigger chunks
## Before
```
running 10 tests
test slice::swap_with_slice_4x_usize_30 ... bench: 894 ns/iter (+/- 11)
test slice::swap_with_slice_4x_usize_3000 ... bench: 99,476 ns/iter (+/- 2,784)
test slice::swap_with_slice_5x_usize_30 ... bench: 1,257 ns/iter (+/- 7)
test slice::swap_with_slice_5x_usize_3000 ... bench: 139,922 ns/iter (+/- 959)
test slice::swap_with_slice_rgb_30 ... bench: 328 ns/iter (+/- 27)
test slice::swap_with_slice_rgb_3000 ... bench: 16,215 ns/iter (+/- 176)
test slice::swap_with_slice_u8_30 ... bench: 312 ns/iter (+/- 9)
test slice::swap_with_slice_u8_3000 ... bench: 5,401 ns/iter (+/- 123)
test slice::swap_with_slice_usize_30 ... bench: 368 ns/iter (+/- 3)
test slice::swap_with_slice_usize_3000 ... bench: 28,472 ns/iter (+/- 3,913)
```
## After
```
running 10 tests
test slice::swap_with_slice_4x_usize_30 ... bench: 868 ns/iter (+/- 36)
test slice::swap_with_slice_4x_usize_3000 ... bench: 99,642 ns/iter (+/- 1,507)
test slice::swap_with_slice_5x_usize_30 ... bench: 1,194 ns/iter (+/- 11)
test slice::swap_with_slice_5x_usize_3000 ... bench: 139,761 ns/iter (+/- 5,018)
test slice::swap_with_slice_rgb_30 ... bench: 324 ns/iter (+/- 6)
test slice::swap_with_slice_rgb_3000 ... bench: 15,962 ns/iter (+/- 287)
test slice::swap_with_slice_u8_30 ... bench: 281 ns/iter (+/- 5)
test slice::swap_with_slice_u8_3000 ... bench: 5,324 ns/iter (+/- 40)
test slice::swap_with_slice_usize_30 ... bench: 275 ns/iter (+/- 5)
test slice::swap_with_slice_usize_3000 ... bench: 28,277 ns/iter (+/- 277)
```
</detail>
remove feature gate in control_flow examples
Stabilization was done in https://github.com/rust-lang/rust/pull/91091, but the two examples weren't updated accordingly.
Probably too late to put it into stable, but it should be in the next release :)
Fix a layout possible miscalculation in `alloc::RawVec`
A layout miscalculation could happen in `RawVec` when used with a type whose size isn't a multiple of its alignment. I don't know if such type can exist in Rust, but the Layout API provides ways to manipulate such types. Anyway, it is better to calculate memory size in a consistent way.
Some improvements to the async docs
The goal here is to make the docs overall a little bit more comprehensive and add more links between the things.
One thing that's not working yet is the links to the keywords. Somehow I couldn't get them to work.
r? ````@GuillaumeGomez```` do you know how I could get the keyword links to work?
Like I previously did for `reverse`, this leaves it to LLVM to pick how to vectorize it, since it can know better the chunk size to use, compared to the "32 bytes always" approach we currently have.
It does still need logic to type-erase where appropriate, though, as while LLVM is now smart enough to vectorize over slices of things like `[u8; 4]`, it fails to do so over slices of `[u8; 3]`.
As a bonus, this also means one no longer gets the spurious `memcpy`(s?) at the end up swapping a slice of `__m256`s: <https://rust.godbolt.org/z/joofr4v8Y>
This issue was found by the Wine project and mitigated there [1].
Windows' documented interface for `setsockopt` expects a `BOOL` (a
`typedef` for `int`) for `TCP_NODELAY` [2]. Windows is forgiving and
will accept any positive length and interpret the first byte of
`*option_value` as the value, so this bug does not affect Windows
itself, but does affect systems implementing Windows' interface more
strictly, such as Wine. Wine was previously passing this through to the
host's `setsockopt`, where, e.g., Linux requires that `option_len` be
correct for the chosen option, and `TCP_NODELAY` expects an `int`.
[1]: d6ea38f32d
[2]: https://docs.microsoft.com/en-us/windows/win32/api/winsock/nf-winsock-setsockopt
POSIX allows `getsockopt` to set `*option_len` to a smaller value if
necessary. Windows will set `*option_len` to 1 for boolean options even
when the caller passes a `BOOL` (`int`) with `*option_len` as 4.
Previously `level` was named `opt` and `option_name` was named `val`,
then extra names of `payload` or `slot` were used for the option value.
This change aligns the wrapper parameters with their names in POSIX.
Winsock uses similar but more abbreviated names: `level`, `optname`,
`optval`, `optlen`.
Fix miniz_oxide types showing up in std docs
Fixes#90526.
Thanks to ```````@camelid,``````` I rediscovered `doc(masked)`, allowing us to prevent `miniz_oxide` type to show up in std docs.
r? ```````@notriddle```````
removing architecture requirements for RustyHermit
RustHermit and HermitCore is able to run on aarch64 and x86_64. In the future these operating systems will also support RISC-V. Consequently, the dependency to a specific target should be removed.
The build process of `hermit-abi` fails if the architecture isn't supported.
core: Implement ASCII trim functions on byte slices
Hi ````````@rust-lang/libs!```````` This is a feature that I wished for when implementing serial protocols with microcontrollers. Often these protocols may contain leading or trailing whitespace, which needs to be removed. Because oftentimes drivers will operate on the byte level, decoding to unicode and checking for unicode whitespace is unnecessary overhead.
This PR adds three new methods to byte slices:
- `trim_ascii_start`
- `trim_ascii_end`
- `trim_ascii`
I did not find any pre-existing discussions about this, which surprises me a bit. Maybe I'm missing something, and this functionality is already possible through other means? There's https://github.com/rust-lang/rfcs/issues/2547 ("Trim methods on slices"), but that has a different purpose.
As per the [std dev guide](https://std-dev-guide.rust-lang.org/feature-lifecycle/new-unstable-features.html), this is a proposed implementation without any issue / RFC. If this is the wrong process, please let me know. However, I thought discussing code is easier than discussing a mere idea, and hacking on the stdlib was fun.
Tracking issue: https://github.com/rust-lang/rust/issues/94035
Guard against unwinding in cleanup code
Currently the only safe guard we have against double unwind is the panic count (which is local to Rust). When double unwinds indeed happen (e.g. C++ exception + Rust panic, or two C++ exceptions), then the second unwind actually goes through and the first unwind is leaked. This can cause UB. cc rust-lang/project-ffi-unwind#6
E.g. given the following C++ code:
```c++
extern "C" void foo() {
throw "A";
}
extern "C" void execute(void (*fn)()) {
try {
fn();
} catch(...) {
}
}
```
This program is well-defined to terminate:
```c++
struct dtor {
~dtor() noexcept(false) {
foo();
}
};
void a() {
dtor a;
dtor b;
}
int main() {
execute(a);
return 0;
}
```
But this Rust code doesn't catch the double unwind:
```rust
extern "C-unwind" {
fn foo();
fn execute(f: unsafe extern "C-unwind" fn());
}
struct Dtor;
impl Drop for Dtor {
fn drop(&mut self) {
unsafe { foo(); }
}
}
extern "C-unwind" fn a() {
let _a = Dtor;
let _b = Dtor;
}
fn main() {
unsafe { execute(a) };
}
```
To address this issue, this PR adds an unwind edge to an abort block, so that the Rust example aborts. This is similar to how clang guards against double unwind (except clang calls terminate per C++ spec and we abort).
The cost should be very small; it's an additional trap instruction (well, two for now, since we use TrapUnreachable, but that's a different issue) for each function with landing pads; if LLVM gains support to encode "abort/terminate" info directly in LSDA like GCC does, then it'll be free. It's an additional basic block though so compile time may be worse, so I'd like a perf run.
r? `@ghost`
`@rustbot` label: F-c_unwind
Add debug assertions to validate NUL terminator in c strings
The `unchecked` variants from the stdlib usually perform the check anyway if debug assertions are on (for example, `unwrap_unchecked`). This PR does the same thing for `CStr` and `CString`, validating the correctness for the NUL byte in debug mode.
Destabilise entry_insert
See: https://github.com/rust-lang/rust/pull/90345
I didn't revert the rename that was done in that PR, I left it as `entry_insert`.
Additionally, before that PR, `VacantEntry::insert_entry` seemingly had no stability attribute on it? I kept the attribute, just made it an unstable one, same as the one on `Entry`.
There didn't seem to be any mention of this in the RELEASES.md, so I don't think there's anything for me to do other than this?
