| Age | Commit message (Collapse) | Author | Lines |
|---|
|
linker: Fix staticlib naming for UEFI
And one minor refactoring in the second commit.
|
|
Group test diffs by stage in post-merge analysis
I think that this is clearer than including the stage in the test name.
To test e.g. on [this PR](https://github.com/rust-lang/rust/pull/138523):
```bash
$ curl https://ci-artifacts.rust-lang.org/rustc-builds/282865097d138c7f0f7a7566db5b761312dd145c/metrics-aarch64-gnu.json > metrics.json
$ cargo run --manifest-path src/ci/citool/Cargo.toml postprocess-metrics metrics.json --job-name aarch64-gnu --parent d9e5539a39192028a7b15ae596a8685017faecee > out.md
```
r? `@marcoieni`
|
|
r=oli-obk,traviscross
Stabilize `#![feature(precise_capturing_in_traits)]`
# Precise capturing (`+ use<>` bounds) in traits - Stabilization Report
Fixes https://github.com/rust-lang/rust/issues/130044.
## Stabilization summary
This report proposes the stabilization of `use<>` precise capturing bounds in return-position impl traits in traits (RPITITs). This completes a missing part of [RFC 3617 "Precise capturing"].
Precise capturing in traits was not ready for stabilization when the first subset was proposed for stabilization (namely, RPITs on free and inherent functions - https://github.com/rust-lang/rust/pull/127672) since this feature has a slightly different implementation, and it hadn't yet been implemented or tested at the time. It is now complete, and the type system implications of this stabilization are detailed below.
## Motivation
Currently, RPITITs capture all in-scope lifetimes, according to the decision made in the ["lifetime capture rules 2024" RFC](https://rust-lang.github.io/rfcs/3498-lifetime-capture-rules-2024.html#return-position-impl-trait-in-trait-rpitit). However, traits can be designed such that some lifetimes in arguments may not want to be captured. There is currently no way to express this.
## Major design decisions since the RFC
No major decisions were made. This is simply an extension to the RFC that was understood as a follow-up from the original stabilization.
## What is stabilized?
Users may write `+ use<'a, T>` bounds on their RPITITs. This conceptually modifies the desugaring of the RPITIT to omit the lifetimes that we would copy over from the method. For example,
```rust
trait Foo {
fn method<'a>(&'a self) -> impl Sized;
// ... desugars to something like:
type RPITIT_1<'a>: Sized;
fn method_desugared<'a>(&'a self) -> Self::RPITIT_1<'a>;
// ... whereas with precise capturing ...
fn precise<'a>(&'a self) -> impl Sized + use<Self>;
// ... desugars to something like:
type RPITIT_2: Sized;
fn precise_desugared<'a>(&'a self) -> Self::RPITIT_2;
}
```
And thus the GAT doesn't name `'a`. In the compiler internals, it's not implemented exactly like this, but not in a way that users should expect to be able to observe.
#### Limitations on what generics must be captured
Currently, we require that all generics from the trait (including the `Self`) type are captured. This is because the generics from the trait are required to be *invariant* in order to do associated type normalization.
And like regular precise capturing bounds, all type and const generics in scope must be captured.
Thus, only the in-scope method lifetimes may be relaxed with this syntax today.
## What isn't stabilized? (a.k.a. potential future work)
See section above. Relaxing the requirement to capture all type and const generics in scope may be relaxed when https://github.com/rust-lang/rust/issues/130043 is implemented, however it currently interacts with some underexplored corners of the type system (e.g. unconstrained type bivariance) so I don't expect it to come soon after.
## Implementation summary
This functionality is implemented analogously to the way that *opaque type* precise capturing works.
Namely, we currently use *variance* to model the capturedness of lifetimes. However, since RPITITs are anonymous GATs instead of opaque types, we instead modify the type relation of GATs to consider variances for RPITITs (along with opaque types which it has done since https://github.com/rust-lang/rust/pull/103491).
https://github.com/rust-lang/rust/blob/30f168ef811aec63124eac677e14699baa9395bd/compiler/rustc_middle/src/ty/util.rs#L954-L976
https://github.com/rust-lang/rust/blob/30f168ef811aec63124eac677e14699baa9395bd/compiler/rustc_type_ir/src/relate.rs#L240-L244
Using variance to model capturedness is an implementation detail, and in the future it would be desirable if opaques and RPITITs simply did not include the uncaptured lifetimes in their generics. This can be changed in a forwards-compatible way, and almost certainly would not be observable by users (at least not negatively, since it may indeed fix some bugs along the way).
## Tests
* Test that the lifetime isn't actually captured: `tests/ui/impl-trait/precise-capturing/rpitit.rs` and `tests/ui/impl-trait/precise-capturing/rpitit-outlives.rs` and `tests/ui/impl-trait/precise-capturing/rpitit-outlives-2.rs`.
* Technical test for variance computation: `tests/ui/impl-trait/in-trait/variance.rs`.
* Test that you must capture all trait generics: `tests/ui/impl-trait/precise-capturing/forgot-to-capture-type.rs`.
* Test that you cannot capture more than what the trait specifies: `tests/ui/impl-trait/precise-capturing/rpitit-captures-more-method-lifetimes.rs` and `tests/ui/impl-trait/precise-capturing/rpitit-impl-captures-too-much.rs`.
* Undercapturing (refinement) lint: `tests/ui/impl-trait/in-trait/refine-captures.rs`.
### What other unstable features may be exposed by this feature?
I don't believe that this exposes any new unstable features indirectly.
## Remaining bugs and open issues
Not aware of any open issues or bugs.
## Tooling support
Rustfmt: :white_check_mark: Supports formatting `+ use<>` everywhere.
Clippy: :white_check_mark: No support needed, unless specific clippy lints are impl'd to care for precise capturing itself.
Rustdoc: :white_check_mark: Rendering `+ use<>` precise capturing bounds is supported.
Rust-analyzer: :white_check_mark: Parser support, and then lifetime support isn't needed https://github.com/rust-lang/rust/pull/138128#issuecomment-2705292494 (previous: ~~:question: There is parser support, but I am unsure of rust-analyzer's level of support for RPITITs in general.~~)
## History
Tracking issue: https://github.com/rust-lang/rust/issues/130044
* https://github.com/rust-lang/rust/pull/131033
* https://github.com/rust-lang/rust/pull/132795
* https://github.com/rust-lang/rust/pull/136554
|
|
Rollup of 8 pull requests
Successful merges:
- #135745 (Recognise new IPv6 non-global range from IETF RFC 9602)
- #137247 (cg_llvm: Reduce the visibility of types, modules and using declarations in `rustc_codegen_llvm`.)
- #138317 (privacy: Visit types and traits in impls in type privacy lints)
- #138581 (Abort in deadlock handler if we fail to get a query map)
- #138776 (coverage: Separate span-extraction from unexpansion)
- #138886 (Fix autofix for `self` and `self as …` in `unused_imports` lint)
- #138924 (Reduce `kw::Empty` usage, part 3)
- #138929 (Visitors track whether an assoc item is in a trait impl or an inherent impl)
r? `@ghost`
`@rustbot` modify labels: rollup
|
|
Visitors track whether an assoc item is in a trait impl or an inherent impl
`AssocCtxt::Impl` now contains an `of_trait` field. This allows ast lowering and nameres to not have to track whether we're in a trait impl or an inherent impl.
|
|
Reduce `kw::Empty` usage, part 3
Remove some more `kw::Empty` uses, in support of #137978.
r? `@davidtwco`
|
|
Fix autofix for `self` and `self as …` in `unused_imports` lint
This fixes two problems with the autofixes for the `unused_imports` lint:
- `use std::collections::{HashMap, self as coll};` would suggest, when `HashMap` is unused, the incorrect `use std::collections::self as coll;` which does not compile.
- `use std::borrow::{self, Cow};` would suggest, when `self` is unused, `use std::borrow::{Cow};`, which contains unnecessary brackets.
The first problem was reported in rust-lang/rust-clippy#14450, the second found while fixing the first one.
Fix #133750
(thanks to `@richardsamuels` for spotting the duplicate)
|
|
coverage: Separate span-extraction from unexpansion
Historically, coverage instrumentation has relied on eagerly “unexpanding” MIR spans back to ancestor spans that have the same context as the function body, and lie within that body. Doing so makes several subsequent operations more straightforward.
In order to support expansion regions, we need to stop doing that, and handle layers of macro-expansion more explicitly. This PR takes a step in that direction, by deferring some of the unexpansion steps, and concentrating them in one place (`spans::extract_refined_covspans`).
Unexpansion still takes place as before, but these changes will make it easier to experiment with expansion-aware coverage instrumentation.
|
|
Abort in deadlock handler if we fail to get a query map
Resolving query cycles requires the complete active query map, or it may miss query cycles. We did not check that the map is completely constructed before. If there is some error collecting the map, something has gone wrong already. This adds a check to abort/panic if we fail to construct the complete map.
This can help differentiate errors from the `deadlock detected` case if constructing query map has errors in practice.
An `Option` is not used for `collect_active_jobs` as the panic handler can still make use of a partial map.
|
|
privacy: Visit types and traits in impls in type privacy lints
With one exception to avoid false positives.
Fixes the same issue as https://github.com/rust-lang/rust/pull/134176.
|
|
cg_llvm: Reduce the visibility of types, modules and using declarations in `rustc_codegen_llvm`.
Final part of #135502
Reduces the visibility of types, modules and using declarations in the `rustc_codegen_llvm` to private or `pub(crate)` where possible, and marks unused fields and enum entries with `#[expect(dead_code)]`.
r? Zalathar
|
|
Recognise new IPv6 non-global range from IETF RFC 9602
This PR adds the `5f00::/16` range defined by [IETF RFC 9602](https://datatracker.ietf.org/doc/rfc9602/) to those ranges which `Ipv6Addr::is_global` recognises as a non-global IP. This range is used for Segment Routing (SRv6) SIDs.
See also: https://www.iana.org/assignments/iana-ipv6-special-registry/iana-ipv6-special-registry.xhtml
Unstable tracking issue: #27709
|
|
compiletest: Support matching on diagnostics without a span
Using `//~? ERROR my message` on any line of the test.
The new checks are exhaustive, like all other `//~` checks, and unlike the `error-pattern` directive that is sometimes used now to check for span-less diagnostics.
This will allow to eliminate most on `error-pattern` directives in compile-fail tests (except those that are intentionally imprecise due to platform-specific diagnostics).
I didn't migrate any of `error-pattern`s in this PR though, except those where the migration was necessary for the tests to pass.
|
|
|
|
Rollup of 9 pull requests
Successful merges:
- #138385 (Keyword tweaks)
- #138580 (resolve: Avoid some unstable iteration 2)
- #138652 (Reintroduce remote-test support in run-make tests)
- #138701 (Make default_codegen_backend serializable)
- #138755 ([rustdoc] Remove duplicated loop when computing doc cfgs)
- #138829 (Slightly reword triagebot ping message for `relnotes-interest-group`)
- #138837 (resolve: Avoid remaining unstable iteration)
- #138838 (Fix/tweak some tests in new solver)
- #138895 (Add a helper for building an owner id in ast lowering)
r? `@ghost`
`@rustbot` modify labels: rollup
|
|
|
|
|
|
|
|
|
|
|
|
Remove InstanceKind::generates_cgu_internal_copy
This PR should not contain any behavior changes. Before this PR, the logic for selecting instantiation mode is spread across all of
* `instantiation_mode`
* `cross_crate_inlinable`
* `generates_cgu_internal_copy`
* `requires_inline`
The last two of those functions are not well-designed. The function that actually decides if we generate a CGU-internal copy is `instantiation_mode`, _not_ `generates_cgu_internal_copy`. The function `requires_inline` documents that it is about the LLVM `inline` attribute and that it is a hint. The LLVM attribute is called `inlinehint`, this function is also used by other codegen backends, and since it is part of instantiation mode selection it is *not* a hint.
The goal of this PR is to start cleaning up the logic into a sequence of checks that have a more logical flow and are easier to customize in the future (to do things like improve incrementality or improve optimizations without causing obscure linker errors because you forgot to update another part of the compiler).
|
|
Add a helper for building an owner id in ast lowering
Just some deduplication of owner-id creations. Will also help me later split up ast lowering into per-owner queries, as it won't be possible anymore to go from a NodeId to a DefId of an owner without doing extra work to check whether we have an owner id. So I'd just do that in the new `owner_id` function and keep the `local_def_id` function free of that logic
|
|
Fix/tweak some tests in new solver
Bunch of miscellaneous new solver tweaks that I found from the failing tests. Can split these out, but they all seemed small enough to not warrant separate PRs.
r? lcnr
|
|
resolve: Avoid remaining unstable iteration
Continuation of #138580.
This should be the performance sensitive part.
|
|
r=cuviper
Slightly reword triagebot ping message for `relnotes-interest-group`
Now that there's also a meta relnotes tracking issue.
r? ```@cuviper``` (or release)
|
|
[rustdoc] Remove duplicated loop when computing doc cfgs
Working on implementing https://github.com/rust-lang/rfcs/blob/master/text/3631-rustdoc-cfgs-handling.md and found this weird case where the first loop was actually not doing anything since we were passing `cfg(...)` to `Cfg::parse` instead of `cfg(...)` items.
Well, that should be a first nice cleanup before the rest comes in.
cc ```@notriddle```
r? ```@camelid```
|
|
r=workingjubilee
Make default_codegen_backend serializable
This PR makes default_codegen_backend serializable.
|
|
Reintroduce remote-test support in run-make tests
The old Makefile-based infrastructure included support for executing binaries with remote-test-client if configured, but that didn't get ported to run_make_support as part of the rmake migration.
This PR re-introduces back that support, with the same implementation (and limitations) of the original Makefile-based support.
[Old Makefile-based implementation of this](https://github.com/rust-lang/rust/blob/9b8accbeb6336fa24d02b2a8bcaecaf44fe2bb65/tests/run-make/tools.mk#L65-L74)
try-job: armhf-gnu
|
|
resolve: Avoid some unstable iteration 2
Continuation of https://github.com/rust-lang/rust/pull/138502.
|
|
Keyword tweaks
r? ```@Noratrieb```
|
|
The existing code produces `Some(kw::Empty)` for these invalid forms:
- a non-name-value, e.g. `#[rustc_allowed_through_unstable_modules]`
- a non-string arg, e.g. `#[rustc_allowed_through_unstable_modules = 3]`
The new code avoids the `kw::Empty` and is a little shorter. It will
produce `None` in those cases, which means E0789 won't be produced if
the `stable` attribute is missing for these invalid forms. This doesn't
matter, because these invalid forms will trigger an "malformed
`rustc_allowed_through_unstable_modules` attribute" anyway.
|
|
|
|
For the the symbols that might not be present, instead of `kw::Empty`.
|
|
This way, `None` represents "crate root without a name" instead of
`kw::Empty`. This changes makes it impossible to forget to handle the
exceptional case.
|
|
Instead of `kw::Empty`. It makes it clearer that this is a name that is
searched for and might not be found.
|
|
This is exactly the kind of case `Ident::dummy()` is for.
|
|
This error was untested.
|
|
Lower to a memset(undef) when Rvalue::Repeat repeats uninit
Fixes https://github.com/rust-lang/rust/issues/138625.
It is technically correct to just do nothing. But if we actually do nothing, we may miss that this is de-initializing something, so instead we just lower to a single memset that writes undef. This is still superior to the memcpy loop, in both quality of code we hand to the backend and LLVM's final output.
|
|
|
|
Lower BinOp::Cmp to llvm.{s,u}cmp.* intrinsics
Lowers `mir::BinOp::Cmp` (`three_way_compare` intrinsic) to the corresponding LLVM `llvm.{s,u}cmp.i8.*` intrinsics.
These are the intrinsics mentioned in https://github.com/rust-lang/rust/pull/118310, which are now available in LLVM 19.
I couldn't find any follow-up PRs/discussions about this, please let me know if I missed something.
r? `@scottmcm`
|
|
|
|
|
|
It uses `libname.a` instead of the standard MSVC naming `name.lib`.
Naming for import libraries isn't touched.
|
|
Now that there's also a meta relnotes tracking issue.
|
|
|
|
Rollup of 7 pull requests
Successful merges:
- #138662 (Implement some basics in UEFI fs)
- #138800 (remove remnants of const_box feature)
- #138821 (match lowering cleanup: remove unused unsizing logic from `non_scalar_compare`)
- #138864 (Rework `--print` options documentation)
- #138868 (Add do_not_recommend typo help)
- #138882 (`with_scope` is only ever used for ast modules)
- #138894 (Update books)
r? `@ghost`
`@rustbot` modify labels: rollup
|
|
Update books
## rust-lang/book
23 commits in 81a976a237f84b8392c4ce1bd5fd076eb757a2eb..45f05367360f033f89235eacbbb54e8d73ce6b70
2025-03-21 23:23:52 UTC to 2025-03-13 14:14:37 UTC
- Ch. 21: call out Chrome multiple-connections issue (rust-lang/book#4297)
- Ch. 16: refactor 16-6 to using listing component (rust-lang/book#4295)
- Ch. 01: Show how to work offline (rust-lang/book#4294)
- Ch. 07: Clarify sentences about `pub use` (rust-lang/book#4293)
- Ch. 02: Consistent ordering of `use` statements (rust-lang/book#4292)
- Anchors on listings (rust-lang/book#4271)
- Ch. 17: another tweak to how we phrase things about sections (rust-lang/book#4288)
- Ch. 20: correct listing number (rust-lang/book#4287)
- Ch. 10.3: clarify language detail (rust-lang/book#4284)
- Ch. 17: minor typos and link reference (rust-lang/book#4286)
- Ch. 9: correctly demonstrate privacy with module (rust-lang/book#4282)
- Ch. 18: correct discussion of delegation in `Post` methods (rust-lang/book#4281)
- Ch. 20: tell folks to see the Reference for more ABI info (rust-lang/book#4165)
- Ch 10.1 minor clarifications (rust-lang/book#4256)
- Clarified the misunderstanding b/w crates, module, items (rust-lang/book#4232)
- Ferris: always show, even when it’s small (rust-lang/book#4280)
- Ch. 17: mention `use std::pin::{Pin, pin};` on introduction (rust-lang/book#4279)
- Persist printing error, NOT ErrorKind (rust-lang/book#4259)
- Typo: "2" should be "2 seconds" (rust-lang/book#4263)
- Ch. 17: fix tiny example consistency issue (rust-lang/book#4270)
- Bump ring from 0.17.8 to 0.17.13 in /listings/ch17-async-await/listing-17-02 (rust-lang/book#4261)
- Bump ring from 0.17.8 to 0.17.14 in /packages/trpl (rust-lang/book#4273)
- 2024 Print Edition: updates to Word docs and more fixes to Markdown text (rust-lang/book#4272)
## rust-lang/reference
5 commits in dda31c85f2ef2e5d2f0f2f643c9231690a30a626..e95ebdfee02514d93f79ec92ae310a804e87f01f
2025-03-24 15:56:46 UTC to 2025-03-18 02:25:06 UTC
- Fix diagnostic attribute typo (rust-lang/reference#1767)
- Mention that “every address” ≠ “every pointer” (rust-lang/reference#1761)
- Rework range pattern rules (rust-lang/reference#1756)
- Use warning block in behavior-considered-undefined (rust-lang/reference#1759)
- Add reference for asm-goto (rust-lang/reference#1693)
|
|
`with_scope` is only ever used for ast modules
Thus I renamed it to match other similar functions (`with_mod_rib`) and made it panic if used on non-modules
|
|
Add do_not_recommend typo help
|
|
Rework `--print` options documentation
This PR reworks the `--print` options documentation, by making it more like codegen options with a dedicated page.
I also added some examples and split some paragraph into multiple paragraph since we now have more place.
r? ```@jieyouxu```
|
