Ignore intrinsic calls in cross-crate-inlining cost model
I noticed in a side project that a function which just compares to `[u64; 2]` for equality is not cross-crate-inlinable. That was surprising to me because I didn't think that code contained a function call, but of course our array comparisons are lowered to an intrinsic. Intrinsic calls don't make a function no longer a leaf, so it makes sense to add this as an exception to the "only leaves" cross-crate-inline heuristic.
This is the useful compare link: https://perf.rust-lang.org/compare.html?start=7cb1a81145a739c4fd858abe3c624ce8e6e5f9cd&end=c3f0a64dbf9fba4722dacf8e39d2fe00069c995e&stat=instructions%3Au because it disables CGU merging in both commits, so effects that cause changes in the sysroot to perturb partitioning downstream are excluded. Perturbations to what is and isn't cross-crate-inlinable in the sysroot has chaotic effects on what items are in which CGUs after merging. It looks like before this PR by sheer luck some of the CGUs dirtied by the patch in eza incr-unchanged happened to be merged together, and with this PR they are not.
The perf runs on this PR point to a nice runtime performance improvement.
Reimplement DestinationPropagation according to live ranges.
This PR reimplements DestinationPropagation as a problem of merging live-ranges of locals. We merge locals that have disjoint live-ranges. This allows merging several locals in the same round by updating live range information.
Live ranges are mainly computed using the `MaybeLiveLocals` analysis. The subtlety is that we split each statement and terminator in 2 positions. The first position is the regular statement. The second position is a shadow, which is always more live. It encodes partial writes and dead writes as a local being live for half a statement. This half statement ensures that writes conflict with another local's writes and regular liveness.
r? `@Amanieu`
Add amdgpu test for addrspacecasting global vars and the gpu-kernel calling convention
Add two tests that can now be added, as the amdgpu is merged.
- Global variables are casted to the default address space since rust-lang/rust#135026
- gpu-kernel calling convention, translatos to amdgpu_kernel rust-lang/rust#135047
Tracking issue: rust-lang/rust#135024
lint ImproperCTypes: refactor linting architecture (part 1)
This is the first PR in an effort to split rust-lang/rust#134697 into individually-mergeable parts.
This one focuses on properly packaging the lint and its tests, as well as properly separate the "linting" and "type-checking" code.
There is exactly one user-visible change: the safety of `Option<Box<FFISafePointee>>` is now the same in `extern` blocks and function definitions: it is safe.
r? `@tgross35` because you are already looking at the original
Rollup of 5 pull requests
Successful merges:
- rust-lang/rust#139524 (Add socket extensions for cygwin)
- rust-lang/rust#145940 (single buffer for exponent fmt of integers)
- rust-lang/rust#146206 (identity uses are ok, even if there are no defining uses)
- rust-lang/rust#146272 (Update comment for `-Werror` on LLVM builds)
- rust-lang/rust#146280 (Make `LetChainsPolicy` public for rustfmt usage)
r? `@ghost`
`@rustbot` modify labels: rollup
Mainly, we realise that the non-null assumption on a Box<_> argument
does not depend on what side of the FFI boundary the function is on.
And anyway, this is not the way to deal with this assumption being maybe violated.
identity uses are ok, even if there are no defining uses
fixrust-lang/rust#146191
I've tried moving the "is this an identity use" check to `fn clone_and_resolve_opaque_types` and this would allow the following code to compile as it now ignores `Opaque<'!a> = Opaque<'!a>` while they previously resulted in errors 71289c378d/tests/ui/type-alias-impl-trait/hkl_forbidden.rs (L42-L46)
The closure signature gets inferred to `for<'a> fn(&'a ()) -> Inner<'a>`. The closure then has a defining use `Inner<'a_latbound> = &'a_latebound ()` while the parent function has a non-defining `Inner<'!a> = Inner<'!a>`. By eagerly discarding identity uses we don't error on the non-defining use in the parent.
r? `@BoxyUwU`
Split `run-make` into two {`run-make`,`run-make-cargo`} test suites
## Summary
Split `tests/run-make` into two test suites, to make it faster and more convenient for contributors to run run-make tests that do not need in-tree `cargo`.
| New test suites | Explanation |
| ---------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| `tests/run-make` | The "fast path" test suite intended for run-make tests that do not need in-tree `cargo`. These tests may not use `cargo`. |
| `tests/run-make-cargo` | The "slow path" test suite that requires checking out `cargo` submodule and building in-tree `cargo`, and thus will have access to in-tree `cargo`. In practice, these constitute a very small portion of the original `run-make` tests. |
This PR carries out [MCP 847: Split run-make test suite into slower-building test suite with suitably-staged cargo and faster-building test suite without cargo](https://github.com/rust-lang/compiler-team/issues/847).
Fixesrust-lang/rust#135573 (for the tests that do not need in-tree `cargo`).
Fixesrust-lang/rust#134109.
## Remarks
- I considered if we want to split by in-tree tools previously. However, as discussed rust-lang/rust#134109, in practice `rustdoc` is not very slow to build, but `cargo` takes a good few minutes. So, the partition boundary was determined to be along in-tree `cargo` availability.
- The `run-make` tests previously that wanted to use `cargo` cannot just use the bootstrap `cargo`, otherwise they would run into situations where bootstrap `cargo` can significantly diverge from in-tree `cargo` (see https://github.com/rust-lang/rust/pull/130642).
---
try-job: aarch64-msvc-1
try-job: test-various
try-job: x86_64-gnu-debug
try-job: aarch64-gnu-debug
try-job: aarch64-apple
try-job: dist-various-1
Disallow shebang in `--cfg` and `--check-cfg` arguments
This PR is similar to https://github.com/rust-lang/rust/issues/146130, where we disallowed frontmatter in `--cfg` and `--check-cfg` arguments. While fixing the other one we also discovered that shebang `#!/usr/bin/shebang` are currently also allowed in `--cfg` and `--check-cfg` arguments.
Allowing shebang in them (which are just ignored) was never intended, this PR fixes that by not stripping shebang for `--cfg` and `--check-cfg` arguments.
This is technically a breaking-change, although I don't expect anyone to actually rely on this unintended behavior.
Fixes https://github.com/rust-lang/rust/issues/146130#issuecomment-3246299499
r? fmease
Suggest bounds in more cases, accounting for type parameters referenced in predicate
Use a `ty::Visitor` to see if the failed predicate references a type parameter. If it does, then we only suggest adding a bound to an (associated) item only if the referenced parameter is present in its generics.
Provide adding bound suggestion in trait and impl associated functions in cases we previously weren't:
```
error[E0277]: `?` couldn't convert the error to `ApplicationError`
--> $DIR/suggest-complex-bound-on-method.rs:18:16
|
LL | t.run()?;
| -----^ the trait `From<<T as Trait>::Error>` is not implemented for `ApplicationError`
| |
| this can't be annotated with `?` because it has type `Result<_, <T as Trait>::Error>`
|
note: `ApplicationError` needs to implement `From<<T as Trait>::Error>`
--> $DIR/suggest-complex-bound-on-method.rs:12:1
|
LL | enum ApplicationError {
| ^^^^^^^^^^^^^^^^^^^^^
= note: the question mark operation (`?`) implicitly performs a conversion on the error value using the `From` trait
help: consider introducing a `where` clause, but there might be an alternative better way to express this requirement
|
LL | fn thing<T: Trait>(&self, t: T) -> Result<(), ApplicationError> where ApplicationError: From<<T as Trait>::Error> {
| +++++++++++++++++++++++++++++++++++++++++++++++++
```
Fixrust-lang/rust#144734.
Ensure that `--html-after-content` option is used to check `scrape_examples_ice` rustdoc GUI test
Follow-up of https://github.com/rust-lang/rust/pull/146091.
This test ensures that the spans are correctly handled when a "local source file" is not the first in the file list. To ensure it's actually what's tested, this test checks that the option is actually used by adding an element.
r? `@lolbinarycat`
Add `__isPlatformVersionAtLeast` and `__isOSVersionAtLeast` symbols
## Motivation
When Objective-C code uses ```@available(...)`,`` Clang inserts a call to [`__isPlatformVersionAtLeast`](https://github.com/llvm/llvm-project/blob/llvmorg-20.1.0/compiler-rt/lib/builtins/os_version_check.c#L276) (`__isOSVersionAtLeast` in older Clang versions). These symbols not being available sometimes ends up causing linker errors. See the new test `tests/run-make/apple-c-available-links` for a minimal reproducer.
The workaround is to link `libclang_rt.osx.a`, see e.g. https://github.com/alexcrichton/curl-rust/issues/279. But that's very difficult for users to figure out (and the backreferences to that issue indicates that people are still running into this in their own projects every so often).
For another recent example, this is preventing `rustc` from using LLVM assertions on macOS, see https://github.com/rust-lang/rust/pull/62592#issuecomment-510670657 and https://github.com/rust-lang/rust/pull/134275#issuecomment-2543067830.
It is also a blocker for [setting the correct minimum OS version in `cc-rs`](https://github.com/rust-lang/rust/issues/136113), since fixing this in `cc-rs` might end up introducing linker errors in places where we weren't before (by default, if using e.g. ```@available(macos`` 10.15, *)`, the symbol usually happens to be left out, since `clang` defaults to compiling for the host macOS version, and thus things _seem_ to work - but the availability check actually compiles down to nothing, which is a huge correctness footgun for running on older OSes).
(My super secret evil agenda is also to expose some variant of ```@available``` in Rust's `std` after https://github.com/rust-lang/rfcs/pull/3750 progresses further, will probably file an ACP for this later. But I believe this PR has value regardless of those future plans, since we'd be making C/Objective-C/Swift interop easier).
## Solution
Implement `__isPlatformVersionAtLeast` and `__isOSVersionAtLeast` as part of the "public ABI" that `std` exposes.
**This is insta-stable**, in the same sense that additions to `compiler-builtins` are insta-stable, though the availability of these symbols can probably be considered a "quality of implementation" detail rather than a stable promise.
I originally proposed to implement this in `compiler-builtins`, see https://github.com/rust-lang/compiler-builtins/pull/794, but we discussed moving it to `std` instead ([Zulip thread](https://rust-lang.zulipchat.com/#narrow/channel/219381-t-libs/topic/Provide.20.60__isPlatformVersionAtLeast.60.20in.20.60std.60.3F/with/507880717)), which makes the implementation substantially simpler, and we avoid gnarly issues with requiring the user to link `libSystem.dylib` (since `std` unconditionally does that).
Note that this does not solve the linker errors for (pure) `#![no_std]` users, but that's _probably_ fine, if you are using ```@available``` to test the OS version on Apple platforms, you're likely also using `std` (and it is still possible to work around by linking `libclang_rt.*.a`).
A thing to note about the implementation, I've choosen to stray a bit from LLVM's upstream implementation, and not use `_availability_version_check` since [it has problems when compiling with an older SDK](https://github.com/llvm/llvm-project/issues/64227). Instead, we use `sysctl kern.osproductversion` when available to still avoid the costly PList lookup in most cases, but still with a fall back to the PList lookup when that is not available (with the PList fallback being is similar to LLVM's implementation).
## Testing
Apple has a lot of different "modes" that they can run binaries in, which can be a bit difficult to find your bearings in, but I've tried to be as thorough as I could in testing them all.
Tested using roughly the equivalent of `./x test library/std -- platform_version` on the following configurations:
- macOS 14.7.3 on a Macbook Pro M2
- `aarch64-apple-darwin`
- `x86_64-apple-darwin` (under Rosetta)
- `aarch64-apple-ios-macabi`
- `x86_64-apple-ios-macabi` (under Rosetta)
- `aarch64-apple-ios` (using Xcode's "Designed for iPad" setting)
- `aarch64-apple-ios-sim` (in iOS Simulator, as iPhone with iOS 17.5)
- `aarch64-apple-ios-sim` (in iOS Simulator, as iPad with iOS 18.2)
- `aarch64-apple-tvos-sim` (in tvOS Simulator)
- `aarch64-apple-watchos-sim` (in watchOS Simulator)
- `aarch64-apple-ios-sim` (in visionOS simulator, using Xcode's "Designed for iPad" setting)
- `aarch64-apple-visionos-sim` (in visionOS Simulator)
- macOS 15.3.1 VM
- `aarch64-apple-darwin`
- `aarch64-apple-ios-macabi`
- macOS 10.12.6 on an Intel Macbook from 2013
- `x86_64-apple-darwin`
- `i686-apple-darwin`
- `x86_64-apple-ios` (in iOS Simulator)
- iOS 9.3.6 on a 1st generation iPad Mini
- `armv7-apple-ios` with an older compiler
Along with manually inspecting the output of `version_from_sysctl()` and `version_from_plist()`, and verifying that they actually match what's expected.
I believe the only real omissions here would be:
- `aarch64-apple-ios` on a newer iPhone that has `sysctl` available (iOS 11.4 or above).
- `aarch64-apple-ios` on a Vision Pro using Xcode's "Designed for iPad" setting.
But I don't have the hardware available to test those.
``@rustbot`` label O-apple A-linkage -T-compiler -A-meta -A-run-make
try-job: aarch64-apple
Suggest parentheses when `match` or `if` expression in binop is parsed as statement
```
error[E0308]: mismatched types
--> $DIR/expr-as-stmt.rs:81:5
|
LL | match () { _ => true } && match () { _ => true };
| ^^^^^^^^^^^^^^^^^^^^^^ expected `()`, found `bool`
|
help: parentheses are required to parse this as an expression
|
LL | (match () { _ => true }) && match () { _ => true };
| + +
```
Address the common case from rust-lang/rust#88727. The original parse error is still outstanding, but the cases brought up in the thread are resolved.
fix ICE when suggesting `::new`
fixes https://github.com/rust-lang/rust/issues/146174
This code suggests to write `Foo::new(...)` when the user writes `Foo(...)` or `Foo { ... }` and the constructor is private, where `new` is some associated function that returns `Self`.
When checking that the return type of `new` is `Self`, we need to instantiate the parameters of `new` with infer vars, so we don't end up with a type like `Box<$param(0)>` in a context that doesn't have any parameters. But then we can't use `normalize_erasing_late_bound_regions` anymore because that goes though a query that can't deal with infer vars.
Since this is diagnostic-only code that is supposed to check for exactly `-> Self`, I think it's fine to just skip normalizing here, especially since The Correct Way<sup>TM</sup> would involve a probe and make this code even more complicated.
Also, the code here does almost the same thing, and these suggestions can probably be unified in the future: 4ca8078d37/compiler/rustc_hir_typeck/src/method/suggest.rs (L2123-L2129)
r? ````@compiler-errors````
cc ````@Qelxiros```` -- this should unblock https://github.com/rust-lang/rust/pull/144420
Fix LoongArch C function ABI when passing/returning structs containing floats
Similar to RISC-V, LoongArch passes structs containing only one or two floats (or a float–integer pair) in registers, as long as each element fits into a single corresponding register. Before this PR, Rust did not check the actual offset of the second float or integer; instead, it assumed the standard offset based on the default alignment. However, since the offset can be affected by `#[repr(align(N))]` and `#[repr(packed)]`, this led to miscompilations (see rust-lang/rust#145692). This PR fixes the issue by explicitly specifying the offset for the remainder of the cast.
Sanitizers target modificators
Depends on bool flag fix: https://github.com/rust-lang/rust/pull/138483.
Some sanitizers need to be target modifiers, and some do not. For now, we should mark all sanitizers as target modifiers except for these: AddressSanitizer, LeakSanitizer
For kCFI, the helper flag -Zsanitizer-cfi-normalize-integers should also be a target modifier.
Many test errors was with sanizer flags inconsistent with std deps. Tests are fixed with `-C unsafe-allow-abi-mismatch`.
fix(rustdoc): match rustc `--emit` precedence
Resolvesrust-lang/rust#141664
This changes rustdoc's `--emit` to allow only one instance of each type, regardless of the actual data that `--emit` carries. This matches rustc's `--emit` behavior.
As of the writing, only `dep-info` emit type carries extra data.
Disallow frontmatter in `--cfg` and `--check-cfg` arguments
This PR disallows the frontmatter syntax in `--cfg` and `--check-cfg` arguments.
Fixes https://github.com/rust-lang/rust/issues/146130
r? fmease
Allow `inline(always)` with a target feature behind a unstable feature `target_feature_inline_always`.
Rather than adding the inline always attribute to the function definition, we add it to the callsite. We can then check that the target features match and that the call would be safe to inline. If the function isn't inlined due to a mismatch, we emit a warning informing the user that the function can't be inlined due to the target feature mismatch.
See tracking issue rust-lang/rust#145574
On unused binding or binding not present in all patterns, suggest potential typo of unit struct/variant or const
When encountering an or-pattern with a binding not available in all patterns, look for consts and unit struct/variants that have similar names as the binding to detect typos.
```
error[E0408]: variable `Ban` is not bound in all patterns
--> $DIR/binding-typo.rs:22:9
|
LL | (Foo, _) | (Ban, Foo) => {}
| ^^^^^^^^ --- variable not in all patterns
| |
| pattern doesn't bind `Ban`
|
help: you might have meant to use the similarly named unit variant `Bar`
|
LL - (Foo, _) | (Ban, Foo) => {}
LL + (Foo, _) | (Bar, Foo) => {}
|
```
For items that are not in the immedate scope, suggest the full path for them:
```
error[E0408]: variable `Non` is not bound in all patterns
--> $DIR/binding-typo-2.rs:51:16
|
LL | (Non | Some(_))=> {}
| --- ^^^^^^^ pattern doesn't bind `Non`
| |
| variable not in all patterns
|
help: you might have meant to use the similarly named unit variant `None`
|
LL - (Non | Some(_))=> {}
LL + (core::option::Option::None | Some(_))=> {}
|
```
When encountering a typo in a pattern that gets interpreted as an unused binding, look for unit struct/variant of the same type as the binding:
```
error: unused variable: `Non`
--> $DIR/binding-typo-2.rs:36:9
|
LL | Non => {}
| ^^^
|
help: if this is intentional, prefix it with an underscore
|
LL | _Non => {}
| +
help: you might have meant to pattern match on the similarly named variant `None`
|
LL - Non => {}
LL + std::prelude::v1::None => {}
|
```
Suggest constant on unused binding in a pattern
```
error: unused variable: `Batery`
--> $DIR/binding-typo-2.rs:110:9
|
LL | Batery => {}
| ^^^^^^
|
help: if this is intentional, prefix it with an underscore
|
LL | _Batery => {}
| +
help: you might have meant to pattern match on the similarly named constant `Battery`
|
LL | Battery => {}
| +
```
Fixrust-lang/rust#51976.
fix drop scope for `super let` bindings within `if let`
Fixesrust-lang/rust#145328 by making non-lifetime-extended `super let` reuse the logic used to compute drop scopes for non-lifetime-extended temporaries.
Also fixesrust-lang/rust#145374, which regressed due to rust-lang/rust#143376 introducing `if let`-like scopes for match arms with guards.
Tracking issue for `super let`: rust-lang/rust#139076
This is a regression fix / breaking change for macros stably exposing `super let`, including `pin!` and `format_args!`.
Nominating to be discussed alongside rust-lang/rust#145328: ```@rustbot``` label +I-lang-nominated +I-libs-api-nominated
a more general version of https://github.com/rust-lang/rust/pull/146080.
after a bit of hacking in [`fluent.rs`](https://github.com/rust-lang/rust/blob/master/compiler/rustc_fluent_macro/src/fluent.rs), i discovered that i'm not the only one that is bad at following guidelines 😅. this pr lowercases the first letter of all the error messages in the codebase.
(i did not change things that are traditionally uppercased such as _MIR_, _ABI_ or _C_)
i think it's reasonable to run a `@bors try` so all the test suite is checked, as i cannot run some of the tests on my machine. i double checked (and replaced manually) all the old error messages, but better be safe than sorry.
in the future i will try to add a check in `x test tidy` that errors if an error message starts with an uppercase letter.
fix(lexer): Only allow horizontal whitespace in frontmatter
In writing up the reference for frontmatter, I realized that we probably
shouldn't be accepting Unicode Line Ending characters between the code
fence and infostring or trailing after the infostring or a code fence.
In digging into the unicode specification we use for Whitespace, it
divides it up into categories, so I'm deferring to what it says for
horizontal whitespace for what should be used within a line.
Note, I am leaving out support for Unicode Default Ignorable characters.
I figure that can be discussed outside of this change within the
reference and tracking issue.
Fixesrust-lang/rust#145971
Frontmatter tracking issue: rust-lang/rust#136889
Constify conversion traits (part 1)
This is the first part of rust-lang/rust#144289 being split into smaller pieces. It adds/moves constness of several traits under the `const_convert` feature:
* `From`
* `Into`
* `TryFrom`
* `TryInto`
* `FromStr`
* `AsRef`
* `AsMut`
* `Borrow`
* `BorrowMut`
* `Deref`
* `DerefMut`
There are a few methods that are intrinsically tied to these traits which I've included in the feature. Particularly, those which are wrappers over `AsRef`:
* `ByteStr::new` (unstable under `bstr` feature)
* `OsStr::new`
* `Path::new`
Those which directly use `Into`:
* `Result::into_ok`
* `Result::into_err`
And those which use `Deref` and `DerefMut`:
* `Pin::as_ref`
* `Pin::as_mut`
* `Pin::as_deref_mut`
* `Option::as_deref`
* `Option::as_deref_mut`
* `Result::as_deref`
* `Result::as_deref_mut`
(note: the `Option` and `Result` methods were suggested by ``@npmccallum`` initially as rust-lang/rust#146101)
The parts which are missing from this PR are:
* Anything that involves heap-allocated types
* Making any method const than the ones listed above
* Anything that could rely on the above, *or* could rely on system-specific code for `OsStr` or `Path` (note: this mostly makes these methods useless since `str` doesn't implement `AsRef<OsStr>` yet, but it's better to track the method for now and add impls later, IMHO)
r? ``@tgross35`` (who mostly already reviewed this)
Change rustdoc's `--emit` to allow only one instance of each type,
regardless of the actual data that `--emit` carries.
This matches rustc's `--emit` behavior.
As of the writing, only `dep-info` emit type carries extra data.
See <https://github.com/rust-lang/rust/issues/141664>
