Don't ICE when no bound vars found while doing closure hir type check
The problem was that we were not visiting the const generic default argument in a bound where predicate when the HIR gets traversed in hir_analysis -> collect -> resolve_bound_vars.
Fixes [112574](https://github.com/rust-lang/rust/issues/112574)
Update crates for better MIPS R6 support
Update crates to remove dependency on old versions of rustix and linux-raw-sys. Update libc, rustix, and linux-raw-sys to enhance support for MIPS R6 introduced by #112374
Commands that do the update:
```shell
cargo +nightly update tempfile clap
cargo +nightly update linux-raw-sys rustix
```
Move `llvm.x86.*` shims into `shims::x86` and implement `_addcarry_u32` and `_subborrow_u{32,64}`
This PR moves all `llvm.x86.*` shims into `shims::x86` and adds `llvm.x86.addcarry.32`, `llvm.x86.subborrow.32` and `llvm.x86.subborrow.64`.
Additionally, it fixes the input carry semantics of `llvm.x86.addcarry.32`. The input carry is an 8-bit value that is interpreted as 1 when it is non-zero.
https://www.intel.com/content/www/us/en/docs/cpp-compiler/developer-guide-reference/2021-8/addcarry-u32-addcarry-u64.html
Add assembly test to make sure that inlining works as expected when closures inherit target features
Closes https://github.com/rust-lang/rust/issues/108338 (the added test proves that it is working correctly)
Clarify example in `Pin::new_unchecked` docs
This example in the docs of `Pin::new_unchecked` puzzled me for a relatively long time. Now I understand that it comes down to the difference between dropping the `Pin` vs dropping the pinned value.
I have extended the explanation to highlight this difference. In my opinion it is clearer now, and I hope it helps others understand `Pin` better.
fix OS-specific I/O safety docs since the io_safety feature is stable
Looks like this text was forgotten to be updated when `io_safety` got stabilized: it still says "once the io_safety feature is stable".
Also adjust the wording a bit for how these docs relate to the general concept of I/O safety.
prevent ice when threshold is 0 and enum has no variants
changelog: [`enum_variant_names`]: prevent ice when threshold is 0 and enum has no variants
r? `@y21`
Fixes the same ice issue raised during review of https://github.com/rust-lang/rust-clippy/pull/11496
Add OwnedTargetMachine to manage llvm:TargetMachine
LLVMRustDisposeTargetMachine taking a &mut could be undefined behaviour.
Wrapping it with a struct and using pointers instead avoids this problem.
In addition the TargetMachine is now automatically freed via the Wrappers drop impl. This should fix some memory leaks when
create_informational_target_machine was used, e.g. 327e6cf55c/compiler/rustc_codegen_llvm/src/llvm_util.rs (L291-L314)
r? `@Nilstrieb`
Rename the legacy feature gating macro
It had a really confusing name by shadowing the previous name, which has
caused issues in the past where people added their new syntax in the
legacy location.
This makes it clear.
Also adds a comment about the return type notation gating, which confused me why it was here at first before `@compiler-errors` told me why.
Fix debug printing of tuple
Self-explanatory. Didn't create a UI test, but I guess I could -- not sure where debug output shows up in rustc_attrs to make a sufficient test, tho.
