Test and fix `size_hint` for slice’s [r]split* iterators
Adds extensive test (of `size_hint`) for all the _[r]split*_ iterators.
Fixes `size_hint` upper bound for _split_inclusive*_ iterators which was one higher than necessary for non-empty slices.
Fixes `size_hint` lower bound for _[r]splitn*_ iterators when _n == 0_, which was one too high.
**Lower bound being one too high was a logic error, violating the correctness condition of `size_hint`.**
_Edit:_ I’ve opened an issue for that bug, so this PR fixes#87978
Implement `black_box` using intrinsic
Introduce `black_box` intrinsic, as suggested in https://github.com/rust-lang/rust/pull/87590#discussion_r680468700.
This is still codegenned as empty inline assembly for LLVM. For MIR interpretation and cranelift it's treated as identity.
cc `@Amanieu` as this is related to inline assembly
cc `@bjorn3` for rustc_codegen_cranelift changes
cc `@RalfJung` as this affects MIRI
r? `@nagisa` I suppose
Adds extensive test for all the [r]split* iterators.
Fixes size_hint upper bound for split_inclusive* iterators which was one higher than necessary for non-empty slices.
Fixes size_hint lower bound for [r]splitn* iterators when n==0, which was one too high.
The new implementation allows some `memcpy`s to be optimized away,
so the uninit value in ui/sanitize/memory.rs is constructed directly
onto the return place. Therefore the sanitizer now says that the
value is allocated by `main` rather than `random`.
Implement Extend<(A, B)> for (Extend<A>, Extend<B>)
I oriented myself at the implementation of `Iterator::unzip` and also rewrote the impl in terms of `(A, B)::extend` after that.
Since (A, B) now also implements Extend we could also mention in the documentation of unzip that it can do "nested unzipping" (you could unzip `Iterator<Item=(A, (B, C))>` into `(Vec<A>, (Vec<B>, Vec<C>))` for example) but I'm not sure of that so I'm asking here 🙂
(P.S. I saw a couple of people asking if there is an unzip3 but there isn't. So this could be a way to get equivalent functionality)
Added the `Option::unzip()` method
* Adds the `Option::unzip()` method to turn an `Option<(T, U)>` into `(Option<T>, Option<U>)` under the `unzip_option` feature
* Adds tests for both `Option::unzip()` and `Option::zip()`, I noticed that `.zip()` didn't have any
* Adds `#[inline]` to a few of `Option`'s methods that were missing it
Constify implementations of `(Try)From` for int types
I believe this to be one of the (many?) things blocking const (Range) iterators.
~~If this is to be merged maybe that should wait until `#![feature(const_trait_impl)]` no longer needs `#![allow(incomplete_features)]`?~~ - Done
Avoid using the `copy_nonoverlapping` wrapper through `mem::replace`.
This is a much simpler way to achieve the pre-#86003 behavior of `mem::replace` not needing dynamically-sized `memcpy`s (at least before inlining), than re-doing #81238 (which needs #86699 or something similar).
I didn't notice it until recently, but `ptr::write` already explicitly avoided using the wrapper, while `ptr::read` just called the wrapper (and was the reason for us observing any behavior change from #86003 in Rust-GPU).
<hr/>
The codegen test I've added fails without the change to `core::ptr::read` like this (ignore the `v0` mangling, I was using a worktree with it turned on by default, for this):
```llvm
13: ; core::intrinsics::copy_nonoverlapping::<u8>
14: ; Function Attrs: inlinehint nonlazybind uwtable
15: define internal void `@_RINvNtCscK5tvALCJol_4core10intrinsics19copy_nonoverlappinghECsaS4X3EinRE8_25mem_replace_direct_memcpy(i8*` %src, i8* %dst, i64 %count) unnamed_addr #0 {
16: start:
17: %0 = mul i64 %count, 1
18: call void `@llvm.memcpy.p0i8.p0i8.i64(i8*` align 1 %dst, i8* align 1 %src, i64 %0, i1 false)
not:17 !~~~~~~~~~~~~~~~~~~~~~ error: no match expected
19: ret void
20: }
```
With the `core::ptr::read` change, `core::intrinsics::copy_nonoverlapping` doesn't get instantiated and the test passes.
<hr/>
r? `@m-ou-se` cc `@nagisa` (codegen test) `@oli-obk` / `@RalfJung` (miri diagnostic changes)
rustc: Fill out remaining parts of C-unwind ABI
This commit intends to fill out some of the remaining pieces of the
C-unwind ABI. This has a number of other changes with it though to move
this design space forward a bit. Notably contained within here is:
* On `panic=unwind`, the `extern "C"` ABI is now considered as "may
unwind". This fixes a longstanding soundness issue where if you
`panic!()` in an `extern "C"` function defined in Rust that's actually
UB because the LLVM representation for the function has the `nounwind`
attribute, but then you unwind.
* Whether or not a function unwinds now mainly considers the ABI of the
function instead of first checking the panic strategy. This fixes a
miscompile of `extern "C-unwind"` with `panic=abort` because that ABI
can still unwind.
* The aborting stub for non-unwinding ABIs with `panic=unwind` has been
reimplemented. Previously this was done as a small tweak during MIR
generation, but this has been moved to a separate and dedicated MIR
pass. This new pass will, for appropriate functions and function
calls, insert a `cleanup` landing pad for any function call that may
unwind within a function that is itself not allowed to unwind. Note
that this subtly changes some behavior from before where previously on
an unwind which was caught-to-abort it would run active destructors in
the function, and now it simply immediately aborts the process.
* The `#[unwind]` attribute has been removed and all users in tests and
such are now using `C-unwind` and `#![feature(c_unwind)]`.
I think this is largely the last piece of the RFC to implement.
Unfortunately I believe this is still not stabilizable as-is because
activating the feature gate changes the behavior of the existing `extern
"C"` ABI in a way that has no replacement. My thinking for how to enable
this is that we add support for the `C-unwind` ABI on stable Rust first,
and then after it hits stable we change the behavior of the `C` ABI.
That way anyone straddling stable/beta/nightly can switch to `C-unwind`
safely.
#[inline] slice::Iter::advance_by
https://github.com/rust-lang/rust/pull/87387#issuecomment-891942661 was marked as a regression. One of the methods in the PR was missing an inline annotation unlike all the other methods on slice iterators.
Let's see if that makes a difference.
Make wrapping_neg() use wrapping_sub(), #[inline(always)]
This is a follow-up change to the fix for #75598. It simplifies the implementation of wrapping_neg() for all integer types by just calling 0.wrapping_sub(self) and always inlines it. This leads to much less assembly code being emitted for opt-level≤1 and thus much better performance for debug-compiled code.
Background is [this discussion on the internals forum](https://internals.rust-lang.org/t/why-does-rust-generate-10x-as-much-unoptimized-assembly-as-gcc/14930).
Add `core::stream::from_iter`
_Tracking issue: https://github.com/rust-lang/rust/issues/81798_
This_ PR implements `std::stream::from_iter`, as outlined in the _"Converting an Iterator to a Stream"_ section of the [Stream RFC](https://github.com/nellshamrell/rfcs/blob/add-async-stream-rfc/text/0000-async-stream.md#converting-an-iterator-to-a-stream). This function enables converting an `Iterator` to a `Stream` by wrapping each item in the iterator with a `Poll::Ready` instance.
r? `@tmandry`
cc/ `@rust-lang/libs` `@rust-lang/wg-async-foundations`
## Example
Being able to convert from an iterator into a stream is useful when refactoring from iterative loops into a more functional adapter-based style. This is fairly common when using more complex `filter` / `map` / `find` chains. In its basic form this conversion looks like this:
**before**
```rust
let mut output = vec![];
for item in my_vec {
let out = do_io(item).await?;
output.push(out);
}
```
**after**
```rust
use std::stream;
let output = stream::from_iter(my_vec.iter())
.map(async |item| do_io(item).await)
.collect()?;
```
Having a way to convert an `Iterator` to a `Stream` is essential in enabling this flow.
## Implementation Notes
This PR makes use of `unsafe {}` to pin an item. Currently we're having conversations on the libs stream in Zulip how to bring `pin-project` in as a dependency to `core` so we can omit the `unsafe {}`.
This PR also includes a documentation block which references `Stream::next` which currently doesn't exist in the stdlib (originally included in the RFC and PR, but later omitted because of an unresolved issue). `stream::from_iter` can't stabilize before `Stream` does, and there's still a chance we may stabilize `Stream` with a `next` method. So this PR includes documentation referencing that method, which we can remove as part of stabilization if by any chance we don't have `Stream::next`.
## Alternatives Considered
### `impl IntoStream for T: IntoIterator`
An obvious question would be whether we could make it so every iterator can automatically be converted into a stream by calling `into_stream` on it. The answer is: "perhaps, but it could cause type issues". Types like `std::collections` may want to opt to create manual implementations for `IntoStream` and `IntoIter`, which wouldn't be possible if it was implemented through a catch-all trait.
Possibly an alternative such as `impl IntoStream for T: Iterator` could work, but it feels somewhat restrictive. In the end, converting an iterator to a stream is likely to be a bit of a niche case. And even then, **adding a standalone function to convert an `Iterator` into a `Stream` would not be mutually exclusive with a blanket implementation**.
### Naming
The exact name can be debated in the period before stabilization. But I've chosen `stream::from_iter` rather than `stream::iter` because we are _creating a stream from an iterator_ rather than _iterating a stream_. We also expect to add a stream counterpart to `iter::from_fn` later on (blocked on async closures), and having `stream::from_fn` and `stream::from_iter` would feel like a consistent pair. It also has prior art in `async_std::stream::from_iter`.
## Future Directions
### Stream conversions for collections
This is a building block towards implementing `stream/stream_mut/into_stream` methods for `std::collections`, `std::vec`, and more. This would allow even quicker refactorings from using loops to using iterator adapters by omitting the import altogether:
**before**
```rust
use std::stream;
let output = stream::from_iter(my_vec.iter())
.map(async |item| do_io(item).await)
.collect()?;
```
**after**
```rust
let output = my_vec
.stream()
.map(async |item| do_io(item).await)
.collect()?;
```
This commit intends to fill out some of the remaining pieces of the
C-unwind ABI. This has a number of other changes with it though to move
this design space forward a bit. Notably contained within here is:
* On `panic=unwind`, the `extern "C"` ABI is now considered as "may
unwind". This fixes a longstanding soundness issue where if you
`panic!()` in an `extern "C"` function defined in Rust that's actually
UB because the LLVM representation for the function has the `nounwind`
attribute, but then you unwind.
* Whether or not a function unwinds now mainly considers the ABI of the
function instead of first checking the panic strategy. This fixes a
miscompile of `extern "C-unwind"` with `panic=abort` because that ABI
can still unwind.
* The aborting stub for non-unwinding ABIs with `panic=unwind` has been
reimplemented. Previously this was done as a small tweak during MIR
generation, but this has been moved to a separate and dedicated MIR
pass. This new pass will, for appropriate functions and function
calls, insert a `cleanup` landing pad for any function call that may
unwind within a function that is itself not allowed to unwind. Note
that this subtly changes some behavior from before where previously on
an unwind which was caught-to-abort it would run active destructors in
the function, and now it simply immediately aborts the process.
* The `#[unwind]` attribute has been removed and all users in tests and
such are now using `C-unwind` and `#![feature(c_unwind)]`.
I think this is largely the last piece of the RFC to implement.
Unfortunately I believe this is still not stabilizable as-is because
activating the feature gate changes the behavior of the existing `extern
"C"` ABI in a way that has no replacement. My thinking for how to enable
this is that we add support for the `C-unwind` ABI on stable Rust first,
and then after it hits stable we change the behavior of the `C` ABI.
That way anyone straddling stable/beta/nightly can switch to `C-unwind`
safely.
Add missing "allocated object" doc link to `<*mut T>::add`
The portion of the documentation expecting the link was already there, but it was rendered as "[allocated object]". The added reference is just copied from the documentation for `<*const T>::add`.
rfc3052 followup: Remove authors field from Cargo manifests
Since RFC 3052 soft deprecated the authors field, hiding it from
crates.io, docs.rs, and making Cargo not add it by default, and it is
not generally up to date/useful information for contributors, we may as well
remove it from crates in this repo.
Partially stabilize `const_slice_first_last`
This stabilizes the non-`mut` methods of `const_slice_first_last` as `const`. These methods are trivial to implement and have no blockers that I am aware of.
`@rustbot` label +A-const-fn +S-waiting-on-review +T-libs-api
