Add some track_caller info to precondition panics
Currently, when you encounter a precondition check, you'll always get the caller location of the implementation of the precondition checks. But with this PR, you'll be told the location of the invalid call. Which is useful.
I thought of this while looking at https://github.com/rust-lang/rust/pull/129642#issuecomment-2311703898.
The changes to `tests/ui/const*` happen because the const-eval interpreter skips `#[track_caller]` frames in its backtraces.
The perf implications of this are:
* Increased debug binary sizes. The caller_location implementation requires that the additional data we want to display here be stored in const allocations, which are deduplicated but not across crates. There is no impact on optimized build sizes. The panic path and the caller location data get optimized out.
* The compile time hit to opt-incr-patched bitmaps happens because the patch changes the line number of some function calls with precondition checks, causing us to go from 0 dirty CGUs to 1 dirty CGU.
* The other compile time hits are marginal but real, and due to doing a handful of new queries. Adding more useful data isn't completely free.
speed up charsearcher for ascii chars
attempt at fixing rust-lang/rust#82471
this implementation should be valid because ascii characters are always one byte and there are no continuation bytes that overlap with ascii characters
im not completely sure that this is _always_ an improvement but it seems to be an improvement for this case and i dont think it can significantly regress any cases
Attempt to improve the `std::fs::create_dir_all` docs related to atomicity
The original paragraph was added in rust-lang/rust#124520. It doesn't match the actual code logic. It says "function returns an error" if "the parent components" _(which also implies directories)_ "have been created already". The code is as follows:
e88e854634/library/std/src/fs.rs (L3146)e88e854634/library/std/src/fs.rs (L3160)
These lines suppress all errors if any path component is a directory. I've updated the paragraph to mirror this.
Clarify WTF-8 safety docs
This PR is a follow-up to PR #140159, which clarifies ~~two things~~:
- the WTF-8 safety comment [was confusing](https://github.com/rust-lang/rust/pull/140159#discussion_r2082766965), either surrogate condition is actually sufficient for safety, both are not required
- ~~the private `os_str::Slice` type name is easily confused with `std::slice`~~
~~Happy to bikeshed the `OsSlice` name, other alternatives are `OsStrSlice` and `StrSlice`. Now it's got a distinct name from `std::slice`, it's easy to search and replace.~~
cc ``@thaliaarchi`` ``@workingjubilee``
Implement `normalize_lexically`
Implements #134694
This is, I think, the most straightforward implementation I could do, which will hopefully more easily allow experimentation if we decide to change the design here.
Implement `advance_by` via `try_fold` for `Sized` iterators
When `try_fold` is overriden, it is usually easier for compilers to optimize.
Example difference: https://iter.godbolt.org/z/z8cEfnKro
use `cfg_select!` to select the right `VaListImpl` definition
tracking issue: https://github.com/rust-lang/rust/issues/44930
Just a bit of cleanup really.
We could use `PhantomInvariantLifetime<'f>` (https://github.com/rust-lang/rust/issues/135806) to make it more precise what that `PhantomData<&'f mut &'f c_void>` marker is doing. I'm not sure how ready that feature is though, `@jhpratt` are these types good to use internally?
---
Some research into the lifetimes of `VaList` and `VaListImpl`:
It's easy to see why the lifetime of these types should not be extended, a `VaList` or `VaListImpl` escaping its function is a bad idea. I don't currently see why coercing the lifetime to a shorter lifetime is problematic though, but probably I just don't understand variance well enough to see it. The history does not provide much explanation:
- 08140878fe original implementation
- b9ea653aee adds `VaListImpl<'f>`, but it is only covariant in `'f`
- https://github.com/rust-lang/rust/pull/62639 makes `VaListImpl<'f>` invariant over `'f` (because `VaList<'a, 'f>` is already invariant over `'f`, but I think that is just an implementation detail?)
Beyond that I don't see how the lifetime situation can be simplified significantly, e.g. this function really needs `'copy` to be unconstrained.
```rust
/// Copies the `va_list` at the current location.
pub unsafe fn with_copy<F, R>(&self, f: F) -> R
where
F: for<'copy> FnOnce(VaList<'copy, 'f>) -> R,
{
let mut ap = self.clone();
let ret = f(ap.as_va_list());
// SAFETY: the caller must uphold the safety contract for `va_end`.
unsafe {
va_end(&mut ap);
}
ret
}
```
`@rustbot` label +F-c_variadic
r? `@workingjubilee`
Docs(lib): Fix `extract_if` docs
Various fixes to the documentation comments of the several `extract_if` collection methods available. It originally started with a small typo fix in `Vec`'s spotted when reading the 1.87 release notes, but then by looking at the others' for comparison in order to try determining what was the intended sentence, some inconsistencies were spotted. Therefore, some other changes are also proposed here to reduce these avoidable differences, going more and more nit-picky along the way. See the individual commits for more details about each change.
`@rustbot` label T-libs A-collections A-docs
rename internal panicking::try to catch_unwind
The public function is called `catch_unwind`, the intrinsic at some point got renamed to `catch_unwind` -- there's no reason to have the internal implementation of this still be called `try`, so let's rename it to match the rest.
additional edge cases tests for `path.rs` 🧪
This pull request adds a few new edge case tests to the `std::path` module. The new tests cover scenarios such as paths with only separators, non-ASCII and Unicode characters, embedded new lines, etc. Each new test is documented with some helpful in-line comments as well.
std: sys: net: uefi: Implement TCP4 connect
- Implement TCP4 connect using EFI_TCP4_PROTOCOL.
- Tested on QEMU setup with connecting to TCP server on host.
Guarantee behavior of transmuting `Option::<T>::None` subject to NPO
In https://github.com/rust-lang/rust/pull/115333, we added a guarantee that transmuting from `[0u8; N]` to `Option<P>` is sound where `P` is a pointer type subject to the null pointer optimization (NPO). It would be useful to be able to guarantee the inverse - that a `None::<P>` value can be transmutes to an array and that will yield `[0u8; N]`.
Closes#117591
Rollup of 7 pull requests
Successful merges:
- #141405 (GetUserProfileDirectoryW is now documented to always store the size)
- #141427 (Disable `triagebot`'s `glacier` handler)
- #141429 (Dont walk into unsafe binders when emiting error for non-structural type match)
- #141438 (Do not try to confirm non-dyn compatible method)
- #141444 (Improve CONTRIBUTING.md grammar and clarity)
- #141446 (Add 2nd Solaris target maintainer)
- #141456 (Suggest correct `version("..")` predicate syntax in check-cfg)
r? `@ghost`
`@rustbot` modify labels: rollup
GetUserProfileDirectoryW is now documented to always store the size
Update to match https://github.com/MicrosoftDocs/sdk-api/pull/1810
Also fix a bug in the Miri implementation while I am starting at that code...
r? ```@ChrisDenton```
Fixes#141254
chore: fix typos in comment
## Fix Typos in Comments
This PR addresses several typos in the Rust standard library's documentation comments:
- In `library/std/src/sync/mpmc/list.rs`: Corrected "attemped" to "attempted"
- In `library/std/src/sys/thread_local/guard/key.rs`: Fixed "defering" to "deferring"
- In `library/std/src/sys/thread_local/guard/key.rs`: Fixed "futher" to "further"
These changes improve documentation readability and consistency without affecting any functional code.
Updated std doctests for wasm
This updates some doctests that fail to run on wasm. We will soon be supporting cross-compiled doctests, and the test-various job fails to run these tests. These tests fail because wasm32-wasip1 does not support threads.
std: fix aliasing bug in UNIX process implementation
`CStringArray` contained both `CString`s and their pointers. Unfortunately, since `CString` uses `Box`, moving the `CString`s into the `Vec` can (under stacked borrows) invalidate the pointer to the string, meaning the resulting `Vec<*const c_char>` was, from an opsem perspective, unusable. This PR removes removes the `Vec<CString>` from `CStringArray`, instead recreating the `CString`/`CStr` from the pointers when necessary. Also,`CStringArray` is now used for the process args as well, the old implementation was suffering from the same kind of bug.
- Implement TCP4 connect using EFI_TCP4_PROTOCOL.
- Tested on QEMU setup with connecting to TCP server on host.
Signed-off-by: Ayush Singh <ayush@beagleboard.org>
`CStringArray` contained both `CString`s and their pointers. Unfortunately, since `CString` uses `Box`, moving the `CString`s into the `Vec` can (under stacked borrows) invalidate the pointer to the string, meaning the resulting `Vec<*const c_char>` was, from an opsem perspective, unusable. This PR removes removes the `Vec<CString>` from `CStringArray`, instead recreating the `CString`/`CStr` from the pointers when necessary. Also,`CStringArray` is now used for the process args as well, the old implementation was suffering from the same kind of bug.
discuss deadlocks in the std::io::pipe() example
I think it's important to discuss deadlocks in examples of how to use pipes. The current example does include an explicit `drop()`, but it also implicitly relies on the fact that the `Command` object is temporary, so that it drops its copy of `pong_tx`. This sort of thing tends to trip people up when they use pipes for the first time. I might've gone overboard with the comments in this version, but I'm curious what folks think.
Update std doctests for android
This updates some doctests that fail to run on android. We will soon be supporting cross-compiled doctests, and the `arm-android` job fails to run these tests.
In summary:
- Android re-exports some traits from linux under a different path.
- Android doesn't seem to have common unix utilities like `true`, `false`, or `whoami`, so these are disabled.
Rollup of 7 pull requests
Successful merges:
- #135562 (Add ignore value suggestion in closure body)
- #139635 (Finalize repeat expr inference behaviour with inferred repeat counts)
- #139668 (Handle regions equivalent to 'static in non_local_bounds)
- #140218 (HIR ty lowering: Clean up & refactor the lowering of type-relative paths)
- #140435 (use uX::from instead of _ as uX in non - const contexts)
- #141130 (rustc_on_unimplemented cleanups)
- #141286 (Querify `coroutine_hidden_types`)
Failed merges:
- #140247 (Don't build `ParamEnv` and do trait solving in `ItemCtxt`s when lowering IATs)
r? `@ghost`
`@rustbot` modify labels: rollup
add doc alias `replace_first` for `str::replacen`
`replace_first` is a sensible name for a function, analogous to actually existing `<[_]>::split_first`, for example. (I just saw someone try to search for it)
I think it's reasonable to add such an alias for `replacen`, which replaces the first occurrence of passed a 1.
