Tweak spans for self arg, fix borrow suggestion for signature mismatch
1. Adjust a suggestion message that was annoying me
2. Fix#112503 by recording the right spans for the `self` part of the `&self` 0th argument
3. Remove the suggestion for adjusting a trait signature on type mismatch, bc that's gonna probably break all the other impls of the trait even if it fixes its one usage 😅
Reuse the MIR validator for MIR inlining
Instead of having the inliner home-cook its own validation, we just check that the substituted MIR body passes the regular validation.
The MIR validation is first split in two: control flow validation (MIR syntax and CFG invariants) and type validation (subtyping relationship in assignments and projections). Only the latter can be affected by instantiating type parameters.
editor/code: Break down CI steps to know what is failing easily
This do the thing I mentioned in https://github.com/rust-lang/rust-analyzer/pull/15265#issuecomment-1634424385
This aims to improve CI status check more readable.
I tried to use [`jobs.<job_id>.if`](https://docs.github.com/en/actions/using-workflows/workflow-syntax-for-github-actions#jobsjob_idif) to make the configuration
more shortly once.
But it could not fire the `end-success` or `end-failure` status if some jobs in the workflow were skipped. This causes an integration problem with bors.
By their reasons, this patch still uses `jobs.<job_id>.steps[*].if`.
---
To do this change, we reorganize npm-script.
| previous | after |
|--------------------|----------------------------------------|
| `npm run lint` | `npm run lint && npm run format:check` |
| `npm run fix` | `npm run lint:fix && npm run format` |
The previous `npm run fix` sometimes does not complete fix automatically because ESLint's autofix doees not follow prettier's formatting. So we need to run `npm run lint:fix && npm run format` by this order.
To do this change, we reorganize npm-script.
| previous | after |
|--------------------|----------------------------------------|
| `npm run lint` | `npm run lint && npm run format:check` |
| `npm run fix` | `npm run lint:fix && npm run format` |
The previous `npm run fix` sometimes does not complete fix automatically
because ESLint's autofix doees not follow prettier's formatting.
So we need to run `npm run lint:fix && npm run format` by this order.
Sync rustc_codegen_cranelift
This time Cranelift has been updated to 0.98. A couple of bugs have been fixed and a decent amount of x86 vendor intrinsics have been implemented.
r? `@ghost`
`@rustbot` label +A-codegen +A-cranelift +T-compiler
limit `change_visibility` assist to applicable items
this pr limits the `change_visibility` assist to applicable items. top level items in this context means items that are not nested within `fn`s or `trait`s.
now
```rs
fn foo {
// assists on this `struct` keyword won't include `change_visibility`
struct Bar {}
}
trait Foo {
// same with the `fn` here
fn bar();
}
```
Multibyte character removal in String::pop and String::remove doctests
I think it would be useful to have the doctests for the `String::pop()` and `String::remove()` methods demonstrate that they work on multibyte UTF-8 sequences.
Fix size_hint for EncodeUtf16
More realistic upper and lower bounds, and handle the case where the iterator is located within a surrogate pair.
Resolves#113897
fix docs & example for `std::os::unix::prelude::FileExt::write_at`
Changelog:
* used `File::create` instead of `File::read` to get a writeable file
* explicity mentioned the bug with `pwrite64` in docs
Unfortunately, I don't think that there is really much we can do about this since the feature has already been stabilised.
We could potentially add a clippy lint warning people on Linux that using `write_at` with the `O_APPEND` flag does not exhibit the behaviour that they would have assumed.
fixes#113627
Allow limited access to `OsString` bytes
This extends #109698 to allow no-cost conversion between `Vec<u8>` and `OsString` as suggested in feedback from `os_str_bytes` crate in #111544.
Update the tracking issue for `const_cstr_from_ptr`
Tracking issue #101719 was for `const_cstr_methods`, #113219 is a new issue specific for `const_cstr_from_ptr`.
(I believe #101719 could also be closed)
```@rustbot``` label +T-libs-api +A-docs
Normalize expected ty in call arguments
fix#15321
I'm not sure if we should do this, or add a normalize in the beginning of `infer_expr_inner`, or somewhere else. r? `@lowr`
[RFC] Support `.comment` section like GCC/Clang (`!llvm.ident`)
Both GCC and Clang write by default a `.comment` section with compiler information:
```txt
$ gcc -c -xc /dev/null && readelf -p '.comment' null.o
String dump of section '.comment':
[ 1] GCC: (GNU) 11.2.0
$ clang -c -xc /dev/null && readelf -p '.comment' null.o
String dump of section '.comment':
[ 1] clang version 14.0.1 (https://github.com/llvm/llvm-project.git c62053979489ccb002efe411c3af059addcb5d7d)
```
They also implement the `-Qn` flag to avoid doing so:
```txt
$ gcc -Qn -c -xc /dev/null && readelf -p '.comment' null.o
readelf: Warning: Section '.comment' was not dumped because it does not exist!
$ clang -Qn -c -xc /dev/null && readelf -p '.comment' null.o
readelf: Warning: Section '.comment' was not dumped because it does not exist!
```
So far, `rustc` only does it for WebAssembly targets and only when debug info is enabled:
```txt
$ echo 'fn main(){}' | rustc --target=wasm32-unknown-unknown --emit=llvm-ir -Cdebuginfo=2 - && grep llvm.ident rust_out.ll
!llvm.ident = !{!27}
```
The RFC part of this PR is about which behavior should `rustc` follow:
- Always add it.
- Add it by default, i.e. have an opt-out flag (GCC, Clang).
- Have an opt-in flag.
- Never add it (current).
There is also the question of whether debug info being enabled matters for that decision, given the current behavior of WebAssembly targets.
For instance, adding it by default gets us closer to other popular compilers, but that may surprise some users with an information leak. The most conservative option is to only do so opt-in, even if debug info is enabled (some users may be stripping debug info and not expecting something else to be leaked elsewhere).
Implementation-wise, this covers both `ModuleLlvm::new()` and `ModuleLlvm::new_metadata()` cases by moving the addition to `context::create_module` and adds a few test cases.
ThinLTO also sees the `llvm.ident` named metadata duplicated (in temporary outputs), so this deduplicates it like it is done for `wasm.custom_sections`. The tests also check this duplication does not take place.
