We want to allow setting this on the CLI, override it only in MIR
passes, and disable it altogether in mir-opt tests.
The default value is "only for NLL MIR dumps", which is considered off
for all intents and purposes, except for `rustc_borrowck` when an NLL
MIR dump is requested.
Use a reduced recursion limit in the MIR inliner's cycle breaker
This probably papers over https://github.com/rust-lang/rust/issues/128887, but primarily I'm opening this PR because multiple compiler people have thought about making this change which probably means it's a good idea.
r? compiler-errors
Add `needs-unwind` compiletest directive to `libtest-thread-limit` and replace some `Path` with `path` in `run-make`
Part of #121876 and the associated [Google Summer of Code project](https://blog.rust-lang.org/2024/05/01/gsoc-2024-selected-projects.html).
This PR does two things:
1. Add this to `libtest-thread-limit` ([Why?](https://github.com/rust-lang/rust/pull/128507#issuecomment-2315158014))
```
//@ needs-unwind
// Reason: this should be ignored in cg_clif (Cranelift) CI and anywhere
// else that uses panic=abort.
```
2. Use `path` instead of `Path` to simplify multiple run-make tests.
debug-fmt-detail option
I'd like to propose a new option that makes `#[derive(Debug)]` generate no-op implementations that don't print anything, and makes `{:?}` in format strings a no-op.
There are a couple of motivations for this:
1. A more thorough stripping of debug symbols. Binaries stripped of debug symbols still retain some of them through `Debug` implementations. It's hard to avoid that without compiler's help, because debug formatting can be used in many places, including dependencies, and their loggers, asserts, panics, etc.
* In my testing it gives about 2% binary size reduction on top of all other binary-minimizing best practices (including `panic_immediate_abort`). There are targets like Web WASM or embedded where users pay attention to binary sizes.
* Users distributing closed-source binaries may not want to "leak" any symbol names as a matter of principle.
2. Adds ability to test whether code depends on specifics of the `Debug` format implementation in unwise ways (e.g. trying to get data unavailable via public interface, or using it as a serialization format). Because current Rust's debug implementation doesn't change, there's a risk of it becoming a fragile de-facto API that [won't be possible to change in the future](https://www.hyrumslaw.com/). An option that "breaks" it can act as a [grease](https://www.rfc-editor.org/rfc/rfc8701.html).
This implementation is a `-Z fmt-debug=opt` flag that takes:
* `full` — the default, current state.
* `none` — makes derived `Debug` and `{:?}` no-ops. Explicit `impl Debug for T` implementations are left unharmed, but `{:?}` format won't use them, so they may get dead-code eliminated if they aren't invoked directly.
* `shallow` — makes derived `Debug` print only the type's name, without recursing into fields. Fieldless enums print their variant names. `{:?}` works.
The `shallow` option is a compromise between minimizing the `Debug` code, and compatibility. There are popular proc-macro crates that use `Debug::fmt` as a way to convert enum values into their Rust source code.
There's a corresponding `cfg` flag: `#[cfg(fmt_debug = "none")]` that can be used in user code to react to this setting to minimize custom `Debug` implementations or remove unnecessary formatting helper functions.
* Use a lookup table for 8-bit integers and the Karatsuba square root
algorithm for larger integers.
* Include optimization hints that give the compiler the exact numeric
range of results.
* Choose test inputs more thoroughly and systematically.
* Check that `isqrt` and `checked_isqrt` have equivalent results for
signed types, either equivalent numerically or equivalent as a panic
and a `None`.
* Check that `isqrt` has numerically-equivalent results for unsigned
types and their `NonZero` counterparts.
* Reuse `ilog10` benchmarks, plus benchmarks that use a uniform
distribution.
Rollup of 14 pull requests
Successful merges:
- #128192 (rustc_target: Add various aarch64 features)
- #129170 (Add an ability to convert between `Span` and `visit::Location`)
- #129343 (Emit specific message for time<=0.3.35)
- #129378 (Clean up cfg-gating of ProcessPrng extern)
- #129401 (Partially stabilize `feature(new_uninit)`)
- #129467 (derive(SmartPointer): assume pointee from the single generic and better error messages)
- #129494 (format code in tests/ui/threads-sendsync)
- #129617 (Update books)
- #129673 (Add fmt::Debug to sync::Weak<T, A>)
- #129683 (copysign with sign being a NaN can have non-portable results)
- #129689 (Move `'tcx` lifetime off of impl and onto methods for `CrateMetadataRef`)
- #129695 (Fix path to run clippy on rustdoc)
- #129712 (Correct trusty targets to be tier 3)
- #129715 (Update `compiler_builtins` to `0.1.123`)
r? `@ghost`
`@rustbot` modify labels: rollup
Correct trusty targets to be tier 3
The Trusty targets were added in https://github.com/rust-lang/rust/pull/129490, but in that PR I accidentally marked them as tier 2. This PR corrects the target metadata to mark them as tier 3.
Fix path to run clippy on rustdoc
Took me a while to find out that the path clippy expected was `src/tools/rustdoc` and not `src/librustdoc`. I think it makes more sense this way as most commands rely on source paths.
r? ```@Kobzol```
Move `'tcx` lifetime off of impl and onto methods for `CrateMetadataRef`
Unconstrained type and const variables are not allowed, but unconstrained lifetimes are. This is not very good style, though, and it leads to unnecessary captures of a lifetime in edition 2024 (not that it matters, but it does trigger the edition migration lint).
copysign with sign being a NaN can have non-portable results
Follow-up to https://github.com/rust-lang/rust/pull/129559.
Cc ```@tgross35``` ```@beetrees```
There's no portable variant we can recommend instead here, is there? Something with a semantics like "if `sign` is a NaN, then return `self` unaltered, otherwise return `self` with the sign changed to that of `sign`"?
Add fmt::Debug to sync::Weak<T, A>
Currently, `sync::Weak<T>` implements `Debug`, but `sync::Weak<T, A>` does not. This appears to be an oversight, as `rc::Weak<T, A>` implements `Debug`. (Note: `sync::Weak` is the weak for `Arc`, and `rc::Weak` is the weak for `Rc`.)
This PR adds the Debug trait for `sync::Weak<T, A>`. The issue was initially brought up here: https://github.com/rust-lang/wg-allocators/issues/131
Update books
## rust-lang/book
4 commits in 04bc1396bb857f35b5dda1d773c9571e1f253304..e7d217be2a75ef1753f0988d6ccaba4d7e376259
2024-08-14 01:19:47 UTC to 2024-08-13 16:51:00 UTC
- Backport/forward port ch12 (rust-lang/book#4008)
- Found some more things to fix in ch7; I forgot to update the snapshot (rust-lang/book#4007)
- Remove redundant sentence. Send to nostarch (rust-lang/book#4006)
- Fix: typo (rust-lang/book#4003)
## rust-lang/edition-guide
5 commits in aeeb287d41a0332c210da122bea8e0e91844ab3e..eeba2cb9c37ab74118a4fb5e5233f7397e4a91f8
2024-08-19 23:28:06 UTC to 2024-08-15 15:12:33 UTC
- 2024: Add rustdoc combined doctests (rust-lang/edition-guide#320)
- Update for unsafe attributes stabilization (rust-lang/edition-guide#319)
- 2024: Add macro-fragment-specifiers. (rust-lang/edition-guide#312)
- Fix deprecated_safe_2024 link (rust-lang/edition-guide#317)
- Add 2024 unsafe functions (rust-lang/edition-guide#304)
## rust-embedded/book
1 commits in 019f3928d8b939ec71b63722dcc2e46330156441..ff5d61d56f11e1986bfa9652c6aff7731576c37d
2024-08-20 07:26:19 UTC to 2024-08-20 07:26:19 UTC
- Use aligned address to demonstrate HardFault (rust-embedded/book#374)
## rust-lang/nomicon
1 commits in 6ecf95c5f2bfa0e6314dfe282bf775fd1405f7e9..14649f15d232d509478206ee9ed5105641aa60d0
2024-08-14 14:49:09 UTC to 2024-08-14 14:49:09 UTC
- CI: Switch to merge queue (rust-lang/nomicon#459)
## rust-lang/reference
14 commits in 62cd0df95061ba0ac886333f5cd7f3012f149da1..0668397076da350c404dadcf07b6cbc433ad3743
2024-08-11 21:06:12 +0000 to 2024-08-27 21:47:20 +0000
- Update enum.md (rust-lang/reference#1354)
- Be consistent about how "Edition differences" is capitalized (rust-lang/reference#1586)
- Sync denied lints with upstream (rust-lang/reference#1589)
- const_eval: update for const-fn float stabilization (rust-lang/reference#1566)
- Add spec identifier syntax to destructors.md (rust-lang/reference#1571)
- Say that `pub(in path)` can't depend on `use` statements (rust-lang/reference#1559)
- bytes inside implicitly const-promoted expressions are immutable (rust-lang/reference#1554)
- Tweak `repr(transparent)` to mention requiring *at most* one non-1-ZST (rust-lang/reference#1568)
- operator expressions: add &raw (rust-lang/reference#1567)
- Rewrite the automatic std link translation, and switch to automatic links (rust-lang/reference#1578)
- Add some basic docs for unsafe attrs (rust-lang/reference#1539)
- don't capitalize Undefined Behavior (rust-lang/reference#1575)
- add the `const` operand to docs for inline assembly (rust-lang/reference#1556)
- Typo: 'a' to 'an' in type-coercions.md (rust-lang/reference#1572)
## rust-lang/rust-by-example
1 commits in 8f94061936e492159f4f6c09c0f917a7521893ff..859786c5bc99301bbc22fc631a5c2b341860da08
2024-08-26 10:30:48 UTC to 2024-08-26 10:30:48 UTC
- Update primitives.md with examples (rust-lang/rust-by-example#1878)
## rust-lang/rustc-dev-guide
7 commits in 43d83780db545a1ed6d45773312fc578987e3968..fa928a6d19e1666d8d811dfe3fd35cdad3b4e459
2024-08-26 14:46:50 UTC to 2024-08-12 21:07:49 UTC
- Fix x.py reference (rust-lang/rustc-dev-guide#2049)
- Update `stabilization_guide.md` (rust-lang/rustc-dev-guide#2034)
- Explain the internal `#[rustc_*]` TEST attributes used for debugging and inside tests (rust-lang/rustc-dev-guide#2046)
- missing char (rust-lang/rustc-dev-guide#2047)
- Replace direct http links to rustc-dev-guide.rust-lang.org (rust-lang/rustc-dev-guide#2044)
- Update index.html, 39. The MIR: fix typo (rust-lang/rustc-dev-guide#2043)
- Update LLVM docs (rust-lang/rustc-dev-guide#2039)
derive(SmartPointer): assume pointee from the single generic and better error messages
Fix#129465
Actually RFC says that `#[pointee]` can be inferred when there is no ambiguity, or there is only one generic type parameter so to say.
cc ```@Darksonn```
r? ```@compiler-errors```
Clean up cfg-gating of ProcessPrng extern
This removes a bit of duplication and is consistent with how `api-ms-win-core-synch-l1-2-0` externs are imported.
Emit specific message for time<=0.3.35
```
error[E0282]: type annotations needed for `Box<_>`
--> /home/gh-estebank/.cargo/registry/src/index.crates.io-6f17d22bba15001f/time-0.3.34/src/format_description/parse/mod.rs:83:9
|
83 | let items = format_items
| ^^^^^
...
86 | Ok(items.into())
| ---- type must be known at this point
|
= note: this is an inference error on `time` caused by a change in Rust 1.80.0; update `time` to version `>=0.3.36`
```
Partially mitigate the fallout from https://github.com/rust-lang/rust/issues/127343. Although the biggest benefit of this would have been if we had had this in 1.80 before it became stable, the long-tail of that change will be felt for a *long* time, so better late than never.
We can also emit an even more targeted error instead of this inference failure.
Add an ability to convert between `Span` and `visit::Location`
AFAIK, there is no way to create a `Location` from a `Span` because its only field is private. This makes it impossible to use visitor methods like `visit_statement` or `visit_terminator`.
This PR adds an implementation for`From<Span>` for `Location` to fix this.
r? ```@celinval```
rustc_target: Add various aarch64 features
Add various aarch64 features already supported by LLVM and Linux.
Additionally include some comment fixes to ensure consistency of feature names with the Arm ARM.
Compiler support for features added to stdarch by https://github.com/rust-lang/stdarch/pull/1614.
Tracking issue for unstable aarch64 features is https://github.com/rust-lang/rust/issues/127764.
List of added features:
- FEAT_CSSC
- FEAT_ECV
- FEAT_FAMINMAX
- FEAT_FLAGM2
- FEAT_FP8
- FEAT_FP8DOT2
- FEAT_FP8DOT4
- FEAT_FP8FMA
- FEAT_HBC
- FEAT_LSE128
- FEAT_LSE2
- FEAT_LUT
- FEAT_MOPS
- FEAT_LRCPC3
- FEAT_SVE_B16B16
- FEAT_SVE2p1
- FEAT_WFxT
- FEAT_SME
- FEAT_SME_F16F16
- FEAT_SME_F64F64
- FEAT_SME_F8F16
- FEAT_SME_F8F32
- FEAT_SME_FA64
- FEAT_SME_I16I64
- FEAT_SME_LUTv2
- FEAT_SME2
- FEAT_SME2p1
- FEAT_SSVE_FP8DOT2
- FEAT_SSVE_FP8DOT4
- FEAT_SSVE_FP8FMA
FEAT_FPMR is added in the first commit and then removed in a separate one to highlight it being removed from upstream LLVM 19. The intention is for it to be detectable at runtime through stdarch but not have a corresponding Rust compile-time feature.
