Split up the `unknown_or_malformed_diagnostic_attributes` lint
This splits up the lint into the following lint group:
- `unknown_diagnostic_attributes` - triggers if the attribute is unknown to the current compiler
- `misplaced_diagnostic_attributes` - triggers if the attribute exists but it is not placed on the item kind it's meant for
- `malformed_diagnostic_attributes` - triggers if the attribute's syntax or options are invalid
- `malformed_diagnostic_format_literals` - triggers if the format string literal is invalid, for example if it has unpaired curly braces or invalid parameters
- this pr doesn't create it, but future lints for things like deprecations can also go here.
This PR does not start emitting lints in places that previously did not.
## Motivation
I want to have finer control over what `unknown_or_malformed_diagnostic_attributes` does
I have a project with fairly low msrv that is/will have a lower msrv than future diagnostic attributes. So lints will be emitted when I or others compile it on a lower msrv.
At this time, there are two options to silence these lints:
- `#[allow(unknown_or_malformed_diagnostic_attributes)]` - this risks diagnostic regressions if I (or others) mess up using the attribute, or if the attribute's syntax ever changes.
- write a build script to detect the compiler version and emit cfgs, and then conditionally enable the attribute:
```rust
#[cfg_attr(rust_version_99, diagnostic::new_attr_in_rust_99(thing = ..))]`
struct Foo;
```
or conditionally `allow` the lint:
```rust
// lib.rs
#![cfg_attr(not(current_rust), allow(unknown_or_malformed_diagnostic_attributes))]
```
I like to avoid using build scripts if I can, so the following works much better for me. That is what this PR will let me do in the future:
```rust
#[allow(unknown_diagnostic_attribute, reason = "attribute came out in rust 1.99 but msrv is 1.70")]
#[diagnostic::new_attr_in_rust_99(thing = ..)]`
struct Foo;
Rollup of 9 pull requests
Successful merges:
- rust-lang/rust#143403 (Port several trait/coherence-related attributes the new attribute system)
- rust-lang/rust#143633 (fix: correct assertion to check for 'noinline' attribute presence before removal)
- rust-lang/rust#143647 (Clarify and expand documentation for std::sys_common dependency structure)
- rust-lang/rust#143716 (compiler: doc/comment some codegen-for-functions interfaces)
- rust-lang/rust#143747 (Add target maintainer information for aarch64-unknown-linux-musl)
- rust-lang/rust#143759 (Fix typos in function names in the `target_feature` test)
- rust-lang/rust#143767 (Bump `src/tools/x` to Edition 2024 and some cleanups)
- rust-lang/rust#143769 (Remove support for SwitchInt edge effects in backward dataflow)
- rust-lang/rust#143770 (build-helper: clippy fixes)
r? `@ghost`
`@rustbot` modify labels: rollup
Add target maintainer information for aarch64-unknown-linux-musl
Mentioning ``@famfo`` so that they can review the documentation. We're both very invested in this target; I originally promoted it to tier 2 with host tools in rust-lang/rust#76420 back in 2020.
fix: Include frontmatter in -Zunpretty output
In the implementation (rust-lang/rust#140035), this was left as an open question for
the tracking issue (rust-lang/rust#136889). My assumption is that this should be
carried over.
The test was carried over from rust-lang/rust#137193 which was superseded by rust-lang/rust#140035.
Thankfully, either way, `-Zunpretty` is unstable and we can always
change it even if we stabilize frontmatter.
Fix `--skip-std-check-if-no-download-rustc`
Since https://github.com/rust-lang/rust/pull/143048, we now explicitly set the build compiler for `check::Std`, which caused it to be built before we checked `--skip-std-check-if-no-download-rustc`. So I moved the check earlier to `make_run`, which resolves it.
I also added a regression test for this. Sadly we can't really test for the positive case easily (when download-ci-rustc is enabled), but we can test the negative cases, where it is disabled.
Fixes: https://github.com/rust-lang/rust/issues/143705
r? ```@RalfJung```
Add an aarch64-msvc build running on ARM64 Windows
Resurrecting rust-lang/rust#126341
Per <https://github.com/rust-lang/rfcs/pull/3817> we intend to promote `aarch64-pc-windows-msvc` to Tier 1. As part of that work, we are adding a pre-merge CI job to validate that changes do not break this target.
Additionally, for consistency, the `dist-aarch64-msvc` job will also be run on Arm64 Windows runners.
r? ``@Kobzol``
try-job: `*aarch64-msvc*`
Rollup of 9 pull requests
Successful merges:
- rust-lang/rust#143446 (use `--dynamic-list` for exporting executable symbols)
- rust-lang/rust#143590 (Fix weird rustdoc output when single and glob reexport conflict on a name)
- rust-lang/rust#143599 (emit `.att_syntax` when global/naked asm use that option)
- rust-lang/rust#143615 (Fix handling of no_std targets in `doc::Std` step)
- rust-lang/rust#143632 (fix: correct parameter names in LLVMRustBuildMinNum and LLVMRustBuildMaxNum FFI declarations)
- rust-lang/rust#143640 (Constify `Fn*` traits)
- rust-lang/rust#143651 (Win: Use exceptions with empty data for SEH panic exception copies instead of a new panic)
- rust-lang/rust#143660 (Disable docs for `compiler-builtins` and `sysroot`)
- rust-lang/rust#143665 ([rustdoc-json] Add tests for `#[doc(hidden)]` handling of items.)
r? `@ghost`
`@rustbot` modify labels: rollup
Disable docs for `compiler-builtins` and `sysroot`
Bootstrap already had a manual doc filter for the `sysroot` crate, but
other library crates keep themselves out of the public docs by setting
`[lib] doc = false` in their manifest. This seems like a better solution
to hide `compiler-builtins` docs, and removes the `sysroot` hack too.
Fixesrust-lang/rust#143215 (after backport)
```@rustbot``` label beta-nominated
Fix handling of no_std targets in `doc::Std` step
The previous logic was wrong for no_std targets, it just didn't do anything. The logic was added there because by default, the `Std` step would otherwise have a list of all std crates to check, but these would fail for no_std targets. What has to happen instead is to select the default set of packages to check/doc/build, which currently happens in the `std_cargo` function, but the `self.crates` list was overriding that.
In general, using `crates: Vec<String>` in the `Std` steps is quite fishy, because it's difficult to distinguish between all crates (either they are all enumerated or `crates` is empty) and the default (e.g. `x <kind> [library]`) vs a subset (e.g. `x <kind> core`). I wanted to improve that using an enum that would distinguish these situations, avoid passing `-p` for all of the crates explicitly, and unify the selection of packages to compile/check/... in `std_cargo`, based on this enum.
However, I found out from some other bootstrap comments that when you pass `-p` explicitly for all crates, cargo behaves differently (apparently for check it will also check targets/examples etc. with `-p`, but not without it). Furthermore, the doc step has a special case where it does not document the `sysroot` package. So as usually, unifying this logic would get into some edge cases... So instead I opted for a seemingly simpler solution, where I try to prefilter only two allowed crates (core and alloc) for no_std targets in the `std_crates_for_run_make` function.
It's not perfect, but I think it's better than the status quo (words to live by when working on bootstrap...).
Fixes [this Zulip topic](https://rust-lang.zulipchat.com/#narrow/channel/326414-t-infra.2Fbootstrap/topic/docs.20for.20non-host.20targets.3F).
r? `@jieyouxu`
Fix weird rustdoc output when single and glob reexport conflict on a name
Fixesrust-lang/rust#143107.
The problem was that the second reexport would overwrite the first, leading to having unexpected results. To fix it, I now group items by their original `DefId` and their name and keep tracks of all imports for this item (should very rarely be more than one though, and even less often more than 2).
cc `@lolbinarycat`
Add profiler to bootstrap command
This PR adds command profiling to the bootstrap command. It tracks the total execution time and records cache hits for each command. It also provides the ability to export execution result to a JSON file. Integrating this with Chrome tracing could further enhance observability.
r? `@Kobzol`
Assorted `run-make-support` maintenance
This PR should contain no functional changes.
- Commit 1: Removes the support library's CHANGELOG. In the very beginning, I thought maybe we would try to version this library. But this is a purely internal test support library, and it's just extra busywork trying to maintain changelog/versions. It's also hopelessly outdated.
- Commit 2: Resets version number to `0.0.0`. Ditto on busywork.
- Commit 3: Bump `run-make-support` to Edition 2024. The support library was already "compliant" with Edition 2024.
- Commit 4: Slightly organizes the support library dependencies.
- Commit 5: Previously, I tried hopelessly to maintain some manual formatting, but that was annoying because it required skipping rustfmt (so export ordering etc. could not be extra formatted). Give up, and do some rearrangements / module prefix tricks to get the `lib.rs` looking at least *reasonable*. IMO this is not a strict improvement, but I rather regain the ability to auto-format it with rustfmt.
- Commit {6,7}: Noticed in rust-lang/rust#143669 that we apparently had *both* {`is_msvc`, `is_windows_msvc`}. This PR removes `is_msvc` in favor of `is_windows_msvc` to make it unambiguous (and only retain one way of gating) as there are some UEFI targets which are MSVC but not Windows.
Best reviewed commit-by-commit.
r? `@Kobzol`
tidy: add support for `--extra-checks=auto:` feature
in preparation for rust-lang/rust#142924
also heavily refactored the parsing of the `--extra-checks` argument to warn about improper usage.
cc ```@GuillaumeGomez```
r? ```@Kobzol```
Since https://github.com/rust-lang/rust/pull/143048, we now explicitly set the build compiler for `check::Std`, which caused it to be built before we checked `--skip-std-check-if-no-download-rustc`. So I moved the check earlier to `make_run`, which resolves it.
I also added a regression test for this. Sadly we can't really test for the positive case easily (when download-ci-rustc is enabled), but we can test the negative cases, where it is disabled.
`rustc_pattern_analysis`: always check that deref patterns don't match on the same place as normal constructors
In rust-lang/rust#140106, deref pattern validation was tied to the `deref_patterns` feature to temporarily avoid affecting perf. However:
- As of rust-lang/rust#143414, box patterns are represented as deref patterns in `rustc_pattern_analysis`. Since they can be used by enabling `box_patterns` instead of `deref_patterns`, it was possible for them to skip validation, resulting in an ICE. This fixes that and adds a regression test.
- External tooling (e.g. rust-analyzer) will also need to validate matches containing deref patterns, which was not possible. This fixes that by making `compute_match_usefulness` validate deref patterns by default.
In order to avoid doing an extra pass for anything with patterns, the second commit makes `RustcPatCtxt` keep track of whether it encounters a deref pattern, so that it only does the check if so. This is purely for performance. If the perf impact of the first commit is negligible and the complexity cost introduced by the second commit is significant, it may be worth dropping the latter.
r? `@Nadrieril`
configure.py: Write last key in each section
The loop that writes the keys in each section of bootstrap.toml accumulates all the commented lines before a given key and emits them when it reaches the next key in the section. This ends up dropping lines accumulated for the last key
Fixesrust-lang/rust#143605
Simplify LLVM bitcode linker in bootstrap and add tests for it
This PR tries to simplify the `LlvmBitcodeLinker` step a little bit, and add tests for it. It also adds tests for `LldWrapper`.
r? `@jieyouxu`
Bootstrap already had a manual doc filter for the `sysroot` crate, but
other library crates keep themselves out of the public docs by setting
`[lib] doc = false` in their manifest. This seems like a better solution
to hide `compiler-builtins` docs, and removes the `sysroot` hack too.
