Do not accidentally treat multi-segment meta-items as single-segment
Fixes https://github.com/rust-lang/rust/issues/55168 and many other regressions from https://github.com/rust-lang/rust/pull/50030
Basically, attributes like `#[any::prefix::foo]` were commonly interpreted as `#[foo]` due to `name()` successfully returning the last segment (this applies to nested things as well `#[attr(any::prefix::foo)]`).
Rollup of 37 pull requests
Successful merges:
- #58854 (appveyor: Use VS2017 for all our images)
- #58855 (std: Spin for a global malloc lock on wasm32)
- #58873 (Fix "Auto-hide item methods documentation" setting)
- #58901 (Change `std::fs::copy` to use `copyfile` on MacOS and iOS)
- #58933 (Move alloc::prelude::* to alloc::prelude::v1, make alloc a subset of std)
- #58938 (core: ensure VaList passes improper_ctypes lint)
- #58941 (MIPS: add r6 support)
- #58949 (SGX target: Expose thread id function in os module)
- #58959 (Add release notes for PR #56243)
- #58976 (Default to integrated `rust-lld` linker for UEFI targets)
- #59009 (Fix SGX implementations of read/write_vectored.)
- #59025 (Fix generic argument lookup for Self)
- #59036 (Fix ICE in MIR pretty printing)
- #59037 (Avoid some common false positives in intra doc link checking)
- #59072 (we can now skip should_panic tests with the libtest harness)
- #59079 (add suggestions to invalid macro item error)
- #59082 (A few improvements to comments in user-facing crates)
- #59102 (Consistent naming for duration_float methods and additional f32 methods)
- #59118 (rustc: fix ICE when trait alias has bare Self)
- #59139 (Unregress using scalar unions in constants.)
- #59146 (Suggest return lifetime when there's only one named lifetime)
- #59147 (Make std time tests more robust for platform differences)
- #59152 (Stabilize Range*::contains.)
- #59156 ([wg-async-await] Add regression test for #55809.)
- #59158 (Revert "Don't generate minification variable if minification disabled")
- #59169 (Add `-Z allow_features=...` flag)
- #59173 (bootstrap: Default to a sensible llvm-suffix.)
- #59175 (Don't run test launching `echo` since that doesn't exist on Windows)
- #59180 (Use try blocks in rustc_codegen_ssa)
- #59185 (No old chestnuts in iter::repeat docs)
- #59201 (Remove restriction on isize/usize in repr(simd))
- #59204 (Output diagnostic information for rustdoc)
- #59206 (Improved test output)
- #59208 (Reduce a Code Repetition Related to Bit Operation)
- #59212 (Add x86_64 musl host to the manifest)
- #59221 (Option and Result: Add references to documentation of as_ref and as_mut)
- #59231 (Stabilize Option::copied)
Option and Result: Add references to documentation of as_ref and as_mut
This makes the documentation more consistent with that of `Pin::as_ref` which converts "from `&Pin<Pointer<T>>` to `Pin<&t>`".
This generally makes it clearer that the reference is going inside the option.
Remove restriction on isize/usize in repr(simd)
As discussed in #55078, there's no known reason for this restriction.
It's unlikely that repr(simd) will be stabilized in its current form, but
might as well remove some restrictions on it.
This removes the branch in `is_machine` which returns false for these types.
`is_machine` is only used for the repr(simd) type validation check.
No old chestnuts in iter::repeat docs
The current language may be amusing, yet is just imprecise and most especially difficult to understand for someone who speaks English as a foreign language.
bootstrap: Default to a sensible llvm-suffix.
I used version-channel-sha, hopefully that should work.
I checked that bootstrap builds, but I cannot check anything else since the llvm
build process is started from cargo, and thus calls clang, and thus I hit the
same bug I hope to fix with this change.
Hopefully fixes#59034.
Add `-Z allow_features=...` flag
Adds a compiler option to allow only whitelisted features.
For projects on nightly that want to prevent feature-creep (and maybe, someday, move off of nightly). Not being able to enforce this has been a problem on Fuchsia and at other big companies.
This doesn't support filtering edition feature flags, but someone is welcome to add that if they need it.
Revert "Don't generate minification variable if minification disabled"
Reverts #58643Fixes#59157https://github.com/rust-lang/rust/pull/58643 made us stop generating minification variables when minification is disabled, however they may still be needed for parent crates that were generated with minification (this will always be the case for libstd and libcore)
r? @QuietMisdreavus @GuillaumeGomez
[wg-async-await] Add regression test for #55809.
Fixes#55809.
This PR adds a regression test for #55809 which checks that a
overflow does not occur when evaluating a requirement for async
functions and `&mut` arguments in some specific circumstances.
Stabilize Range*::contains.
Closes https://github.com/rust-lang/rust/issues/32311. There's also a bit of rustfmt on range.rs thrown in for good measure (I forgot to turn off format-on-save in VSCode).
Make std time tests more robust for platform differences
Previously, `time::tests::since_epoch` and `time::tests::system_time_math` would fail if the platform represents a SystemTime as unix epoch + `u64` nanoseconds.
r? @sfackler
A few improvements to comments in user-facing crates
Not too many this time, and all concern comments (almost all doc comments) in user-facing crates (libstd, libcore, liballoc).
r? @steveklabnik
Avoid some common false positives in intra doc link checking
The empty string case is never going to be a link. The numeric case may be a link, but if it were it would have resolved locally. It's more likely the makeshift markdown footnote notation (`[0]`, etc)
r? @QuietMisdreavus
Fix ICE in MIR pretty printing
A `Def::Variant` should be considered as a function in mir pretty
printing. Each variant has a constructor that we must print.
Given the following enum definition:
```rust
pub enum TestMe {
X(usize),
}
```
We will need to generate a constructor for the variant `X` with a
signature that looks something like the following:
```
fn TestMe::X(_1: usize) -> TestMe;
```
Fixes: #59021
Fix generic argument lookup for Self
Rewrite the SelfCtor early and use the replacement Def when
calculating the path_segs.
Note that this also changes which def is seen by the code that
computes user_self_ty and is_alias_variant_ctor; I don't see a
immediate issue with that, but I'm not 100% clear on the
implications.
Fixes#57924
r? @eddyb
Default to integrated `rust-lld` linker for UEFI targets
The `x86_64-unknown-uefi` target was added in https://github.com/rust-lang/rust/pull/56769 with the linker defaulting to `lld-link`. This means that a system linker with that name is required for linking.
I think defaulting to `rust-lld`, which is shipped with Rust, is a better default for the following reasons:
- Most systems don't have `lld-link` installed, so it forces users to install it first.
- The naming of LLD executables is not standarized, so users often need to create an additional symlink before things work. For example, on Ubuntu `apt install lld` leads to an executable named `lld-link-6.0`.
- We already default to `rust-lld` for [many targets](https://github.com/rust-lang/rust/search?utf8=%E2%9C%93&q=rust-lld&type=), including embedded and WASM targets, so doing the same for UEFI crates seems consistent to me. (It even seems like `x86_64-unknown-uefi` is the [only target](https://github.com/rust-lang/rust/search?q=lld-link&unscoped_q=lld-link) that uses `lld-link`.)
cc @dvdhrm who added the target and @kkk669 who [proposed to use `rust-lld`](https://github.com/rust-lang/rust/pull/56769#issuecomment-461119648).
SGX target: Expose thread id function in os module
In order to call `std::os::fortanix_sgx::usercalls::send`, you need the thread id. This exposes it through another function in `std::os::fortanix_sgx`.
I looked at how other platforms do this. On Windows and `cfg(unix)` you can get the OS handle from a `thread::JoinHandle`, but that's not sufficient, I need it for a `thread::Thread`. In the future, this functionality could be added to `thread::Thread` and this platform can follow suit.
r? @joshtriplett
MIPS: add r6 support
MIPS r6 is quite different with the previous version.
It use some new target triples:
mipsisa32r6-unknown-linux-gnu
mipsisa32r6el-unknown-linux-gnu
mipsisa64r6-unknown-linux-gnuabi64
mipsisa64r6el-unknown-linux-gnuabi64
This patch has been tested with Debian Port for mips64r6el,
and the support of these triples also is included in llvm:
https://reviews.llvm.org/rGe58c45a695f39004710b6ce940d489fee800dbd3
Move alloc::prelude::* to alloc::prelude::v1, make alloc a subset of std
This was one of the unresolved questions of https://github.com/rust-lang/rfcs/pull/2480. As the RFC says this is maybe not useful in the sense that we are unlikely to ever have a second version, but making the crate a true subset makes one less issue to think about if we stabilize it and later want to merge standard library crates and have Cargo feature flags to enable or disable parts of the `std` crate.
See also discussion in https://github.com/rust-lang/rust/pull/58175.
Also rename the feature gate and point to a dedicated tracking issue: https://github.com/rust-lang/rust/issues/58935
Change `std::fs::copy` to use `copyfile` on MacOS and iOS
`copyfile` on MacOS is similar to `CopyFileEx` on Windows. It supports copying resource forks, extended attributes, and file ACLs, none of which are copied by the current generic unix implementation.
The API is available from MacOS 10.7 and iOS 4.3 (and possibly earlier but I haven't checked).
Closes#58895.
std: Spin for a global malloc lock on wasm32
There's lots of comments in the code, but the main gist of this commit
is that the acquisition of the global malloc lock on the
`wasm32-unknown-unknown` target when threads are enabled will not spin
on contention rather than block.
appveyor: Use VS2017 for all our images
Originally added in #55935 to test build times, this was reverted
in #56201 due to a belief that it caused the exit code 259 spurious
errors. We've since learned, however, that the 259 exit code is likely
not related to this image update as we're getting it in a number of
locations now.
VS2017 looks like it may be required to compile LLVm in the near future,
notably discovered by #58408 where we attempted to update LLVM.
