Commit Graph

281737 Commits

Author SHA1 Message Date
dianne
20149629ba "classic2021" ruleset: experimentally add fallback-to-outer (eat both)
My reasoning: the ruleset implemented by the same feature gate in
Edition 2024 always tries to eat the inherited reference first. For
consistency, it makes sense to me to say across all editions that users
should consider the inherited reference's mutability when wondering if a
`&mut` pattern will type.
2025-02-18 18:00:17 -08:00
dianne
799e0f7690 add FIXMEs for diagnostic improvements 2025-02-18 18:00:17 -08:00
dianne
2c595d649a update unstable book 2025-02-18 18:00:17 -08:00
dianne
1ed74aaf0c add mixed-edition tests 2025-02-18 18:00:11 -08:00
dianne
443c51d5d6 "structural2021" ruleset: add fallback-to-outer (eat both) deref rule 2025-02-18 17:44:28 -08:00
dianne
8dc64a405d "classic2021" and "structural2021" rulesets: add eat-inherited-ref-alone deref rules 2025-02-18 17:44:28 -08:00
dianne
3e77657312 remove old edition-2021-specific tests
These are superseded by the old-edition revisions on the shared tests.
2025-02-18 17:44:28 -08:00
dianne
e24833a4a8 add test revisions for old-edition behavior of feature gates
This also adds `#[cfg]` attributes to tests for bindings' types,
to make it visually clearer which revisions type successfully.
2025-02-18 17:44:28 -08:00
bors
f44efbf9e1 Auto merge of #137235 - matthiaskrgr:rollup-2kjua2t, r=matthiaskrgr
Rollup of 10 pull requests

Successful merges:

 - #135711 (Do not ICE on default_field_value const with lifetimes)
 - #136599 (librustdoc: more usages of `Joined::joined`)
 - #136876 (Locking documentation updates)
 - #137000 (Deeply normalize item bounds in new solver)
 - #137126 (fix docs for inherent str constructors)
 - #137161 (Pattern Migration 2024: fix incorrect messages/suggestions when errors arise in macro expansions)
 - #137191 (Update mdbook and move error_index_generator)
 - #137203 (Improve MIR modification)
 - #137206 (Make E0599 a structured error)
 - #137218 (misc `layout_of` cleanup)

r? `@ghost`
`@rustbot` modify labels: rollup
2025-02-18 21:08:58 +00:00
bors
827a0d638d Auto merge of #137231 - Urgau:rollup-heiq934, r=Urgau
Rollup of 9 pull requests

Successful merges:

 - #136750 (Make ub_check message clear that it's not an assert)
 - #137151 (Install more signal stack trace handlers)
 - #137167 (tests: Also gate `f16::erfc()` doctest with `reliable_f16_math` cfg)
 - #137195 (cg_clif: use exclusively ABI alignment)
 - #137202 (Enforce T: Hash for Interned<...>)
 - #137205 (Remove `std::os::wasi::fs::FileExt::tell`)
 - #137211 (don't ICE for alias-relate goals with error term)
 - #137214 (add last std diagnostic items for clippy)
 - #137221 (Remove scrutinee_hir_id from ExprKind::Match)

r? `@ghost`
`@rustbot` modify labels: rollup
2025-02-18 18:04:32 +00:00
Matthias Krüger
bafff1b3af
Rollup merge of #137218 - lukas-code:layout_of_cleanup, r=compiler-errors
misc `layout_of` cleanup

See individual commits for details.

r? `@oli-obk` but feel free to reassign
2025-02-18 18:40:55 +01:00
Matthias Krüger
a66ef2f40e
Rollup merge of #137206 - estebank:e0599-structured, r=jieyouxu
Make E0599 a structured error
2025-02-18 18:40:54 +01:00
Matthias Krüger
10dd016a80
Rollup merge of #137203 - nnethercote:improve-MIR-modification, r=compiler-errors
Improve MIR modification

A few commits that simplify code that manipulates MIR bodies.

r? `@tmiasko`
2025-02-18 18:40:54 +01:00
Matthias Krüger
05dbe6ddff
Rollup merge of #137191 - ehuss:update-mdbook, r=jieyouxu
Update mdbook and move error_index_generator

This moves error_index_generator to the rustbook workspace so that it can share the dependency with mdbook. I had forgotten that error_index_generator is using mdbook.

This includes a corresponding update to mdbook which avoids a regression in error_index_generator.

Closes https://github.com/rust-lang/rust/issues/137052
2025-02-18 18:40:53 +01:00
Matthias Krüger
54db888355
Rollup merge of #137161 - dianne:pat-migration-bookkeeping-for-macros, r=Nadrieril
Pattern Migration 2024: fix incorrect messages/suggestions when errors arise in macro expansions

See the diff between the two commits for how this affected the error message and suggestion. In order to decide how to format those, the pattern migration diagnostic keeps track of which parts of the user's pattern cause problems in Edition 2024. However, it neglected to do some of this bookkeeping when pointing to macro expansion sites. This fixes that.
2025-02-18 18:40:52 +01:00
Matthias Krüger
ac3b179c3b
Rollup merge of #137126 - m4rch3n1ng:fix-inherent-str-docs, r=Amanieu
fix docs for inherent str constructors

related to #131114

when implementing inherent str constructors in #136517, i forgot to change the docs, so the code examples still imported the `std::str` module and used the constructor from there, instead of using "itself" (the inherent constructor).
2025-02-18 18:40:52 +01:00
Matthias Krüger
c8d904125e
Rollup merge of #137000 - compiler-errors:deeply-normalize-item-bounds, r=lcnr
Deeply normalize item bounds in new solver

Built on #136863.

Fixes https://github.com/rust-lang/trait-system-refactor-initiative/issues/142.
Fixes https://github.com/rust-lang/trait-system-refactor-initiative/issues/151.

cc https://github.com/rust-lang/trait-system-refactor-initiative/issues/116

First commit reworks candidate preference for projection bounds to prefer param-env projection clauses even if the corresponding trait ref doesn't come from the param-env.

Second commit adjusts the associated type item bounds check to deeply normalize in the new solver. This causes some test fallout which I will point out.

r? lcnr
2025-02-18 18:40:51 +01:00
Matthias Krüger
5a942d67a6
Rollup merge of #136876 - joshtriplett:locking-might-not-be-advisory, r=Amanieu
Locking documentation updates

- Reword file lock documentation to clarify advisory vs mandatory. Remove the
  word "advisory", and make it more explicit that the lock may be advisory or
  mandatory depending on platform.

- Document that locking a file fails on Windows if the file is opened only for append
2025-02-18 18:40:50 +01:00
Matthias Krüger
58a622b019
Rollup merge of #136599 - yotamofek:pr/rustdoc-more-joined, r=GuillaumeGomez
librustdoc: more usages of `Joined::joined`

Some missed opportunities from #136244
r? ```@GuillaumeGomez``` since you reviewed the last one (feel free to re-assign, of course 😊)
First two commits are just drive-by cleanups
2025-02-18 18:40:50 +01:00
Matthias Krüger
7edd17cfbb
Rollup merge of #135711 - estebank:issue-135649, r=davidtwco
Do not ICE on default_field_value const with lifetimes

`#![feature(default_field_values)]` uses a `const` body that should be treated as inline `const`s, but is actually being detected otherwise. This is similar to the situation in #78174, so we take the same solution: we check if the const actually comes from a field, and if it does, we use that logic to get the appropriate lifetimes and not ICE during borrowck.

Fix #135649.
2025-02-18 18:40:49 +01:00
Urgau
3ebe11b2e8
Rollup merge of #137221 - bjorn3:remove_unused_field, r=compiler-errors
Remove scrutinee_hir_id from ExprKind::Match

It is unused.
2025-02-18 18:34:19 +01:00
Urgau
b7c2da2231
Rollup merge of #137214 - cyrgani:clippy_diagnostic_items, r=compiler-errors
add last std diagnostic items for clippy

Part of https://github.com/rust-lang/rust-clippy/issues/5393.
Add diagnostic item attributes to the items in `std` and `core` where clippy currently uses hardcoded paths (https://github.com/rust-lang/rust-clippy/blob/master/clippy_utils/src/paths.rs).
2025-02-18 18:34:18 +01:00
Urgau
61f04328db
Rollup merge of #137211 - lcnr:alias-relate-accept-error, r=compiler-errors
don't ICE for alias-relate goals with error term

see comment, fixes https://github.com/rust-lang/trait-system-refactor-initiative/issues/165

r? ``@compiler-errors``
2025-02-18 18:34:17 +01:00
Urgau
73e5abd175
Rollup merge of #137205 - thaliaarchi:remove-wasi-fileext-tell, r=alexcrichton
Remove `std::os::wasi::fs::FileExt::tell`

Following #137165 (Use `tell` for `<File as Seek>::stream_position`), `tell` is now directly exposed via `stream_position`, making `<File as FileExt>::tell` redundant. Remove it.

`std::os::wasi::fs::FileExt::tell` is currently unstable and tracked in https://github.com/rust-lang/rust/issues/71213.

``@rustbot`` ping wasi
2025-02-18 18:34:17 +01:00
Urgau
5b00e0b064
Rollup merge of #137202 - Mark-Simulacrum:interned-is-hash, r=saethlin
Enforce T: Hash for Interned<...>

This adds panicking Hash impls for several resolver types that don't actually satisfy this condition. It's not obvious to me that rustc_resolve actually upholds the Interned guarantees but fixing that seems pretty hard (the structures have at minimum some interior mutability, so it's not really recursively hashable in place...). FIXME comments as such on those impls.

cc https://github.com/rust-lang/rust/pull/137196#issuecomment-2664350287

r? ``@saethlin``
2025-02-18 18:34:16 +01:00
Urgau
626fcdf48f
Rollup merge of #137195 - workingjubilee:remove-pref-align-from-cg-clif, r=bjorn3
cg_clif: use exclusively ABI alignment

This will minimize possible conflict with future updates to AbiAndPrefAlign that may remove some fields. It is also almost a bug to consider them.

r? ``@bjorn3``
2025-02-18 18:34:15 +01:00
Urgau
14fb84a5bb
Rollup merge of #137167 - martn3:reliable_f16_math-f16-erfc, r=tgross35
tests: Also gate `f16::erfc()` doctest with `reliable_f16_math` cfg

In #136324 the doctest for `f16::erf()` was gated with `reliable_f16_math`. Add the same gate on `f16::erfc()` to avoid:

    rust_out.71e2e529d20ea47d-cgu.0:\
    (.text._ZN8rust_out4main43_doctest_main_library_std_src_f16_rs_1321_017h485f3ffe6bf2a981E+0x38): \
    undefined reference to `__gnu_h2f_ieee'

on MIPS (and maybe other architectures).

r? tgross35
2025-02-18 18:34:15 +01:00
Urgau
1391f75512
Rollup merge of #137151 - Urgau:register-more-signals, r=workingjubilee
Install more signal stack trace handlers

This PR install the signal stack handler to more signals (`SIGILL`, ~~`SIGTRAP`~~, ~~`SIGABRT`~~, ~~`SIGFPE`~~, `SIGBUS`, ~~`SIGQUIT`~~).

Noticed in https://github.com/rust-lang/rust/issues/137138 that we didn't print anything for `SIGILL`, so I though we could just handle more signals.

r? `````@workingjubilee````` since you last touched it
2025-02-18 18:34:14 +01:00
Urgau
d7fe4c0e92
Rollup merge of #136750 - kornelski:ub-bug, r=saethlin
Make ub_check message clear that it's not an assert

I've seen a user assume that their unsound code was *safe*, because ub_check prevented the program from performing the unsafe operation.

This PR makes the panic message clearer that ub_check is a bug detector, not run-time safety protection.
2025-02-18 18:34:13 +01:00
Josh Triplett
ec2034d53d Reorder "This lock may be advisory or mandatory." earlier in the lock docs 2025-02-18 17:31:10 +01:00
Josh Triplett
35674eff6f Clarify that locking on Windows also works for files opened with .read(true) 2025-02-18 17:26:33 +01:00
bjorn3
768a5bd470
Remove scrutinee_hir_id from ExprKind::Match
It is unused
2025-02-18 13:51:32 +01:00
Lukas Markeffsky
2fbc413d83 cosmetic changes
- change function parameter order to `cx, ty, ...` to match the other
  functions in this file
- use `ct` identifier for `ty::Const` to match the majority of the
  compiler codebase
- remove useless return
- bring match arms in a more natural order
2025-02-18 13:22:46 +01:00
Lukas Markeffsky
67345f9203 remove useless parameter
Remove the `repr` parameter from the wrappers around `calc.univariant`,
because it's always defaulted. Only ADTs can have a repr and those call
`calc.layout_of_struct_or_enum` and not `calc.univariant`.
2025-02-18 13:22:46 +01:00
Lukas Markeffsky
1d1ac3d310 remove redundant code
- we normalize before calling `layout_of_uncached`, so we don't need to
  normalize again later
- we check for type/const errors at the top of `layout_of_uncached`, so
  we don't need to check again later
2025-02-18 13:22:45 +01:00
Lukas Markeffsky
7a667d206c remove unreachable cases
`ty::Placeholder` is used by the trait solver and computing its layout
was necessary, because the `PointerLike` trait used to be automatically
implemented for all types with pointer-like layout.
Nowadays, `PointerLike` requires user-written impls and the trait solver
no longer computes any layouts, so this can be removed.

Unevaluated constants that aren't generic should have caused a const eval
error earlier during normalization.
2025-02-18 13:22:45 +01:00
Lukas Markeffsky
802b7abab7 clean up layout error diagnostics
- group the fluent slugs together
- reword (internal-only) "too generic" error to be more in line with
  the other errors
2025-02-18 13:22:45 +01:00
Lukas Markeffsky
d0a5bbbb8e document and test all LayoutError variants 2025-02-18 13:22:45 +01:00
cyrgani
a72402a0f9 add last std diagnostic items for clippy 2025-02-18 10:54:37 +01:00
lcnr
f910684616 don't ICE for alias-relate goals with error term 2025-02-18 10:30:37 +01:00
Jubilee
fa53181f42
cg_clif: Tweak formatting of global comments
Co-authored-by: bjorn3 <17426603+bjorn3@users.noreply.github.com>
2025-02-18 01:29:23 -08:00
bors
3b022d8cee Auto merge of #133852 - x17jiri:cold_path, r=saethlin
improve cold_path()

#120370 added a new instrinsic `cold_path()` and used it to fix `likely` and `unlikely`

However, in order to limit scope, the information about cold code paths is only used in 2-target switch instructions. This is sufficient for `likely` and `unlikely`, but limits usefulness of `cold_path` for idiomatic rust. For example, code like this:

```
if let Some(x) = y { ... }
```

may generate 3-target switch:

```
switch y.discriminator:
0 => true branch
1 = > false branch
_ => unreachable
```

and therefore marking a branch as cold will have no effect.

This PR improves `cold_path()` to work with arbitrary switch instructions.

Note that for 2-target switches, we can use `llvm.expect`, but for multiple targets we need to manually emit branch weights. I checked Clang and it also emits weights in this situation. The Clang's weight calculation is more complex that this PR, which I believe is mainly because `switch` in `C/C++` can have multiple cases going to the same target.
2025-02-18 07:49:09 +00:00
Esteban Küber
693f7035f1 Make E0599 a structured error 2025-02-18 04:50:33 +00:00
bors
aaa8614934 Auto merge of #137162 - nnethercote:remove-Map-2, r=Zalathar
Move methods from `Map` to `TyCtxt`, part 2.

Continuing the work started in #136466.

Every method gains a `hir_` prefix, though for the ones that already have a `par_` or `try_par_` prefix I added the `hir_` after that.

r? Zalathar
2025-02-18 04:26:30 +00:00
Thalia Archibald
d2f15971de Remove std::os::wasi::fs::FileExt::tell
Following #137165 (Use `tell` for `<File as Seek>::stream_position`),
`tell` is now directly exposed via `stream_position`, making
`<File as FileExt>::tell` redundant. Remove it.
2025-02-17 20:21:05 -08:00
Nicholas Nethercote
04eeda47ab Inline and replace Statement::replace_nop.
It has a single call site, and doesn't seem worth having as an API
function.
2025-02-18 13:43:43 +11:00
Mark Rousskov
9fc759099b Enforce T: Hash for Interned<...>
This adds panicking Hash impls for several resolver types that don't
actually satisfy this condition. It's not obvious to me that
rustc_resolve actually upholds the Interned guarantees but fixing that
seems pretty hard (the structures have at minimum some interior
mutability, so it's not really recursively hashable in place...).
2025-02-17 21:35:52 -05:00
Nicholas Nethercote
69f5e342bf Inline and remove BasicBlockData::retain_statements.
It has a single call site, and the code is clearer this way.
2025-02-18 13:31:08 +11:00
Nicholas Nethercote
627e08c909 Remove BasicBlockData::expand_statements.
The previous commit removed its single use. `MirPatch` is a more
flexible alternative.
2025-02-18 13:13:32 +11:00
Nicholas Nethercote
e3316ae453 Improve MirPatch documentation and naming.
It's currently lacking comments. This commit adds some, which is useful
because there are some methods with non-obvious behaviour.

The commit also renames two things:
- `patch_map` becomes `term_patch_map`, because it's only about
  terminators.
- `is_patched` becomes `is_term_patched`, for the same reason.

(I would guess that originally `MirPatch` only handled terminators, and
then over time it expanded to allow other modifications, but these names
weren't updated.)
2025-02-18 13:12:50 +11:00