]> git.lizzy.rs Git - rust.git/log
rust.git
22 months agodon't normalize wf predicates
lcnr [Wed, 13 Jul 2022 17:42:08 +0000 (19:42 +0200)]
don't normalize wf predicates

this allows us to soundly use unnormalized projections for wf

22 months agoAuto merge of #100205 - cjgillot:noice-doc, r=camelid
bors [Tue, 9 Aug 2022 05:05:46 +0000 (05:05 +0000)]
Auto merge of #100205 - cjgillot:noice-doc, r=camelid

Avoid ICE in rustdoc when using `Fn` bounds

Fixes https://github.com/rust-lang/rust/issues/100143

22 months agoAdd regression test comment
Noah Lev [Mon, 8 Aug 2022 21:09:55 +0000 (14:09 -0700)]
Add regression test comment

22 months agoSynthetize a trait ref when none is available.
Camille GILLOT [Mon, 8 Aug 2022 19:09:09 +0000 (21:09 +0200)]
Synthetize a trait ref when none is available.

22 months agoAuto merge of #98863 - compiler-errors:projection-msg, r=estebank
bors [Mon, 8 Aug 2022 10:16:20 +0000 (10:16 +0000)]
Auto merge of #98863 - compiler-errors:projection-msg, r=estebank

Implement special-cased projection error message for some common traits

Not sure what the best phrasing is, but I feel like these are more clear than the plain `<Type as Iterator>::Output == Type` messages.

If this is actually a good idea, are there any other traits this could benefit?

22 months agoAuto merge of #98489 - cjgillot:naked-nohir, r=davidtwco,tmiasko
bors [Mon, 8 Aug 2022 07:31:12 +0000 (07:31 +0000)]
Auto merge of #98489 - cjgillot:naked-nohir, r=davidtwco,tmiasko

Only fetch HIR for naked functions that have the attribute.

22 months agoAuto merge of #100207 - notriddle:notriddle/skipped-inline-module, r=Manishearth
bors [Mon, 8 Aug 2022 04:39:29 +0000 (04:39 +0000)]
Auto merge of #100207 - notriddle:notriddle/skipped-inline-module, r=Manishearth

rustdoc: do not mark the contents of a skipped module as inlined

Fixes #100204

22 months agoAdjust wording
Michael Goulet [Mon, 8 Aug 2022 00:13:41 +0000 (00:13 +0000)]
Adjust wording

22 months agoImplement special-cased projection error message for some common traits
Michael Goulet [Sun, 3 Jul 2022 20:17:08 +0000 (20:17 +0000)]
Implement special-cased projection error message for some common traits

22 months agoAuto merge of #100245 - matthiaskrgr:rollup-tfoo650, r=matthiaskrgr
bors [Sun, 7 Aug 2022 21:25:29 +0000 (21:25 +0000)]
Auto merge of #100245 - matthiaskrgr:rollup-tfoo650, r=matthiaskrgr

Rollup of 6 pull requests

Successful merges:

 - #100019 (Revive suggestions for boxed trait objects instead of impl Trait)
 - #100038 (Document the `no-std` target option in config.toml.example)
 - #100194 (Remove even more box syntax uses from src/test)
 - #100206 (test: skip terminfo parsing in Miri)
 - #100230 (Use start_point instead of next_point to point to elided lifetime amp…)
 - #100244 (Add armv4t-none-eabi take2)

Failed merges:

r? `@ghost`
`@rustbot` modify labels: rollup

22 months agoRollup merge of #100244 - Lokathor:add-armv4t-none-eabi-take2, r=jackh726
Matthias Krüger [Sun, 7 Aug 2022 19:10:27 +0000 (21:10 +0200)]
Rollup merge of #100244 - Lokathor:add-armv4t-none-eabi-take2, r=jackh726

Add armv4t-none-eabi take2

This is the same as the previous PR (https://github.com/rust-lang/rust/pull/99226) but i just made a fresh branch without a merge commit in it.

---

### armv4t-none-eabi target quiz

> A tier 3 target must have a designated developer or developers (the "target maintainers") on record to be CCed when issues arise regarding the target.

That's me!

> Targets must use naming consistent with any existing targets

We're using the existing name as recognized by LLVM and GCC

> Tier 3 targets may have unusual requirements to build or use, but must not create legal issues or impose onerous legal terms for the Rust project or for Rust developers or users.

No legal issues here.

>> The target must not introduce license incompatibilities.

No license requirements here.

>> Anything added to the Rust repository must be under the standard Rust license (MIT OR Apache-2.0).

check

>> The target must not cause the Rust tools or libraries built for any other host (even when supporting cross-compilation to the target) to depend on any new dependency less permissive than the Rust licensing policy.

no new deps, we're just adding a rustc target description file for a target llvm already knows about.

>> Compiling, linking, and emitting functional binaries, libraries, or other code for the target (whether hosted on the target itself or cross-compiling from another target) must not depend on proprietary (non-FOSS) libraries.

bare-metal target, doesn't rely on any libs at all.

> Tier 3 targets should attempt to implement as much of the standard libraries as possible and appropriate

`core` only here. You could build `alloc` too, but you'd have to bring your own global allocator.

> The target must provide documentation for the Rust community explaining how to build for the target, using cross-compilation if possible.

LLVM knows how to do it, you just need the GNU Binutils linker because LLVM's linker doesn't work that far back. That's in the docs as part of this PR.

> Tier 3 targets must not impose burden on the authors of pull requests, or other developers in the community, to maintain the target.

No burdens, LLVM already knows how to do this. Further, because this is a cpu-feature variant of an existing tier3 target the `compiler-builtins` crate has already been updated as necessary to fix any missing builtin function gaps.

> Patches adding or updating tier 3 targets must not break any existing tier 2 or tier 1 target, and must not knowingly break another tier 3 target without approval of either the compiler team or the maintainers of the other tier 3 target.

check.

22 months agoRollup merge of #100230 - cjgillot:noice-multibyte-amp, r=compiler-errors
Matthias Krüger [Sun, 7 Aug 2022 19:10:26 +0000 (21:10 +0200)]
Rollup merge of #100230 - cjgillot:noice-multibyte-amp, r=compiler-errors

Use start_point instead of next_point to point to elided lifetime amp…

Using `next_point` creates a span which points inside the multibyte token, ICEing.

Fixes https://github.com/rust-lang/rust/issues/100224

22 months agoRollup merge of #100206 - RalfJung:miri-terminfo, r=thomcc
Matthias Krüger [Sun, 7 Aug 2022 19:10:25 +0000 (21:10 +0200)]
Rollup merge of #100206 - RalfJung:miri-terminfo, r=thomcc

test: skip terminfo parsing in Miri

Terminfo parsing takes a significant amount of time in Miri, making libtest startup very slow. To work around that Miri in fact unsets the `TERM` variable. However, this means we don't get colors in `cargo miri test`.

So I propose we add some logic in libtest that skips parsing terminfo files under Miri, and just uses the regular basic coloring commands (taken from the `colored` crate).

As far as I can see, these two commands are all that libtest ever needs from terminfo, so Miri doesn't even lose any functionality through this. If you want I can entirely remove the terminfo parsing code and just use these commands instead.

Cc https://github.com/rust-lang/miri/issues/2292 `@saethlin`

22 months agoRollup merge of #100194 - est31:box_syntax_tests, r=Mark-Simulacrum
Matthias Krüger [Sun, 7 Aug 2022 19:10:24 +0000 (21:10 +0200)]
Rollup merge of #100194 - est31:box_syntax_tests, r=Mark-Simulacrum

Remove even more box syntax uses from src/test

Prior work, notably #88316 has removed box syntax from most of the testsuite.
However, some tests were left out.
This commit removes box_syntax uses from more locations in src/test.
This migrates the tests where `box` is mostly an "implementation detail" and not the primary thing being tested by the test.
Furthermore, some tests from the mir-opt test suite are not being migrated.

22 months agoRollup merge of #100038 - nicholasbishop:bishop-add-no-std, r=Mark-Simulacrum
Matthias Krüger [Sun, 7 Aug 2022 19:10:23 +0000 (21:10 +0200)]
Rollup merge of #100038 - nicholasbishop:bishop-add-no-std, r=Mark-Simulacrum

Document the `no-std` target option in config.toml.example

22 months agoRollup merge of #100019 - TaKO8Ki:suggest-boxed-trait-objects-instead-of-impl-trait...
Matthias Krüger [Sun, 7 Aug 2022 19:10:22 +0000 (21:10 +0200)]
Rollup merge of #100019 - TaKO8Ki:suggest-boxed-trait-objects-instead-of-impl-trait, r=compiler-errors

Revive suggestions for boxed trait objects instead of impl Trait

The suggestion implemented in #75608 was not working properly, so I fixed it.

22 months agoAuto merge of #100218 - nicholasbishop:bishop-update-cb, r=Mark-Simulacrum
bors [Sun, 7 Aug 2022 18:44:41 +0000 (18:44 +0000)]
Auto merge of #100218 - nicholasbishop:bishop-update-cb, r=Mark-Simulacrum

Update compiler_builtins to 0.1.78

Among other things, this pulls in https://github.com/rust-lang/compiler-builtins/pull/475, which fixes some i128/u128 arithmetic operations on the `x86_64-unknown-uefi` target.

22 months agoUpdate SUMMARY.md
Lokathor [Sun, 7 Aug 2022 18:42:25 +0000 (12:42 -0600)]
Update SUMMARY.md

22 months agoCreate armv4t_none_eabi.md
Lokathor [Sun, 7 Aug 2022 18:41:38 +0000 (12:41 -0600)]
Create armv4t_none_eabi.md

22 months agoCreate armv4t_none_eabi.rs
Lokathor [Sun, 7 Aug 2022 18:40:41 +0000 (12:40 -0600)]
Create armv4t_none_eabi.rs

22 months agoAuto merge of #99833 - andrewpollack:fuchsia-rust-ci-zircon-lib-improvement, r=Mark...
bors [Sun, 7 Aug 2022 15:58:30 +0000 (15:58 +0000)]
Auto merge of #99833 - andrewpollack:fuchsia-rust-ci-zircon-lib-improvement, r=Mark-Simulacrum

Fuchsia CI zircon lib improvement

Removing Zircon build process, instead pulling `sysroot` and related libs directly from Fuchsia SDK

cc. `@tmandry` `@djkoloski`

22 months agotest: skip terminfo parsing in Miri
Ralf Jung [Sat, 6 Aug 2022 18:17:40 +0000 (14:17 -0400)]
test: skip terminfo parsing in Miri

22 months agoShifting CI to pull Zircon libraries directly from Fuchsia SDK
Andrew Pollack [Wed, 27 Jul 2022 23:47:19 +0000 (23:47 +0000)]
Shifting CI to pull Zircon libraries directly from Fuchsia SDK

PR feedback

PR Followups

Updating clang download

Updating clang download

Restructuring env used

Restructuring env used

Adding chmod

Adding chmod

Adding chmod

22 months agoAuto merge of #99983 - RalfJung:more-layout-checks, r=eddyb
bors [Sun, 7 Aug 2022 13:17:54 +0000 (13:17 +0000)]
Auto merge of #99983 - RalfJung:more-layout-checks, r=eddyb

More layout sanity checks

r? `@eddyb`

22 months agoUse start_point instead of next_point to point to elided lifetime ampersand.
Camille GILLOT [Sun, 7 Aug 2022 12:35:11 +0000 (14:35 +0200)]
Use start_point instead of next_point to point to elided lifetime ampersand.

22 months agofix wrong suggestions for boxed trait objects instead of impl trait
Takayuki Maeda [Mon, 1 Aug 2022 11:22:50 +0000 (20:22 +0900)]
fix wrong suggestions for boxed trait objects instead of impl trait

22 months agorevive suggestions for boxed trait objects instead of impl Trait
Takayuki Maeda [Mon, 1 Aug 2022 09:42:16 +0000 (18:42 +0900)]
revive suggestions for boxed trait objects instead of impl Trait

22 months agoAuto merge of #100091 - chenyukang:add-check-for-link-ordinal, r=michaelwoerister
bors [Sun, 7 Aug 2022 05:37:29 +0000 (05:37 +0000)]
Auto merge of #100091 - chenyukang:add-check-for-link-ordinal, r=michaelwoerister

Check link ordinal to make sure it is targetted  for foreign function

Fix #100009, when link ordinal is not target for foreign functions, emit an error.

cc `@dpaoliello`

22 months agoAuto merge of #100004 - jyn514:exclude-single-test, r=Mark-Simulacrum
bors [Sun, 7 Aug 2022 02:56:48 +0000 (02:56 +0000)]
Auto merge of #100004 - jyn514:exclude-single-test, r=Mark-Simulacrum

Move `x test --skip` to be part of `--exclude`

`--skip` is inconsistent with the rest of the interface and redundant with `--exclude`.
Fix --exclude to work properly for files and directories rather than having a separate flag.

Fixes https://github.com/rust-lang/rust/issues/96342. cc https://github.com/rust-lang/rust/pull/96493#issuecomment-1200521720

r? `@Mark-Simulacrum`

22 months agoRemove even more box syntax uses from src/test
est31 [Thu, 7 Jul 2022 02:36:10 +0000 (04:36 +0200)]
Remove even more box syntax uses from src/test

Prior work, notably 6550021124451628b1efc60c59284465b109e3aa from #88316
has removed box syntax from most of the testsuite. However,
some tests were left out.
This commit removes box_syntax uses from more locations in src/test.
Some tests that are very box syntax specific are not being migrated.

22 months agoUpdate compiler_builtins to 0.1.78
Nicholas Bishop [Sun, 7 Aug 2022 00:49:48 +0000 (20:49 -0400)]
Update compiler_builtins to 0.1.78

22 months agoAuto merge of #100213 - matthiaskrgr:rollup-mqe7t1n, r=matthiaskrgr
bors [Sat, 6 Aug 2022 23:38:28 +0000 (23:38 +0000)]
Auto merge of #100213 - matthiaskrgr:rollup-mqe7t1n, r=matthiaskrgr

Rollup of 5 pull requests

Successful merges:

 - #100071 (deps: dedupe `annotate-snippets` crate versions)
 - #100127 (Remove Windows function preloading)
 - #100130 (Avoid pointing out `return` span if it has nothing to do with type error)
 - #100169 (Optimize `pointer::as_aligned_to`)
 - #100175 (ascii -> ASCII in code comment)

Failed merges:

r? `@ghost`
`@rustbot` modify labels: rollup

22 months agoRollup merge of #100175 - fxn:patch-1, r=Mark-Simulacrum
Matthias Krüger [Sat, 6 Aug 2022 23:19:35 +0000 (01:19 +0200)]
Rollup merge of #100175 - fxn:patch-1, r=Mark-Simulacrum

ascii -> ASCII in code comment

Easy one I spotted while reading source code.

22 months agoRollup merge of #100169 - WaffleLapkin:optimize_is_aligned_to, r=workingjubilee
Matthias Krüger [Sat, 6 Aug 2022 23:19:34 +0000 (01:19 +0200)]
Rollup merge of #100169 - WaffleLapkin:optimize_is_aligned_to, r=workingjubilee

Optimize `pointer::as_aligned_to`

This PR replaces `addr % align` with `addr & align - 1`, which is correct due to `align` being a power of two.

Here is a proof that this makes things better: [[godbolt]](https://godbolt.org/z/Wbq3hx6YG).

This PR also removes `assume(align != 0)`, with the new impl it does not improve anything anymore ([[godbolt]](https://rust.godbolt.org/z/zcnrG4777), [[original concern]](https://github.com/rust-lang/rust/pull/95643#discussion_r843326903)).

22 months agoRollup merge of #100130 - compiler-errors:erroneous-return-span, r=lcnr
Matthias Krüger [Sat, 6 Aug 2022 23:19:33 +0000 (01:19 +0200)]
Rollup merge of #100130 - compiler-errors:erroneous-return-span, r=lcnr

Avoid pointing out `return` span if it has nothing to do with type error

This code:

```rust
fn f(_: String) {}

fn main() {
    let x = || {
        if true {
            return ();
        }
        f("");
    };
}
```

Emits this:
```
   Compiling playground v0.0.1 (/playground)
error[E0308]: mismatched types
 --> src/main.rs:8:11
  |
8 |         f("");
  |           ^^- help: try using a conversion method: `.to_string()`
  |           |
  |           expected struct `String`, found `&str`
  |
note: return type inferred to be `String` here
 --> src/main.rs:6:20
  |
6 |             return ();
  |                    ^^
```

Specifically, that note has nothing to do with the type error in question. This is because the change implemented in #84244 tries to point out the `return` span on _any_ type coercion error within a closure that happens after a `return` statement, regardless of if the error has anything to do with it.

This is really easy to trigger -- just needs a closure (or an `async`) and an early return (or any other form, e.g. `?` operator suffices) -- and super distracting in production codebases. I'm letting #84128 regress because that issue is much harder to fix correctly, and I can re-open that issue after this lands.

As a drive-by, I added a `resolve_vars_if_possible` to the coercion error logic, which leads to some error improvements. Unrelated to the issue above, though.

22 months agoRollup merge of #100127 - ChrisDenton:remove-init, r=thomcc
Matthias Krüger [Sat, 6 Aug 2022 23:19:32 +0000 (01:19 +0200)]
Rollup merge of #100127 - ChrisDenton:remove-init, r=thomcc

Remove Windows function preloading

After `@Mark-Simulacrum` asked me to provide guidance for when optionally imported functions should be preloaded, I realised my justifications were now quite weak. I think the strongest argument that can be made is that it avoids some degree of nondeterminism when calling these functions (in as far as system API calls can be said to be deterministic). However, I don't think that's particularly convincing unless there's a real world use case where it matters. Further discussion with `@thomcc` has strengthened my feeling that preloading isn't really needed.

Note that `WaitOnAddress` needed some adjustment to work without preloading. I opted not to use a macro for this special case as it seemed silly to do so for just one thing (and I don't like macros tbh).

22 months agoRollup merge of #100071 - klensy:annotate-snippets-bump, r=Mark-Simulacrum
Matthias Krüger [Sat, 6 Aug 2022 23:19:32 +0000 (01:19 +0200)]
Rollup merge of #100071 - klensy:annotate-snippets-bump, r=Mark-Simulacrum

deps: dedupe `annotate-snippets` crate versions

Dedupes `annotate-snippets` crate versions (https://github.com/rust-lang/annotate-snippets-rs/blob/0.9.1/CHANGELOG.md). Should work, but there is not a lot of tests.

Looks like switching to that crate a bit stalled.

22 months agoAuto merge of #100117 - nicholasbishop:bishop-update-cc, r=Mark-Simulacrum
bors [Sat, 6 Aug 2022 21:10:52 +0000 (21:10 +0000)]
Auto merge of #100117 - nicholasbishop:bishop-update-cc, r=Mark-Simulacrum

Bump cc version in bootstrap

Among other changes, the newer cc release pulls in this fix:
https://github.com/rust-lang/cc-rs/commit/b2792e33ff91b92e2e920e54d582b0c334670c37

This fixes errors when building compiler_builtins for UEFI targets.

22 months agoDocument the `no-std` target option in config.toml.example
Nicholas Bishop [Mon, 1 Aug 2022 22:12:07 +0000 (18:12 -0400)]
Document the `no-std` target option in config.toml.example

22 months agorustdoc: do not mark the contents of a skipped module as inlined
Michael Howell [Sat, 6 Aug 2022 20:17:40 +0000 (13:17 -0700)]
rustdoc: do not mark the contents of a skipped module as inlined

22 months agoAuto merge of #99524 - cuviper:relnotes-1.63.0, r=Mark-Simulacrum
bors [Sat, 6 Aug 2022 17:50:41 +0000 (17:50 +0000)]
Auto merge of #99524 - cuviper:relnotes-1.63.0, r=Mark-Simulacrum

Add release notes for Rust 1.63.0

22 months agoAvoid ICE in rustdoc.
Camille GILLOT [Sat, 6 Aug 2022 17:42:52 +0000 (19:42 +0200)]
Avoid ICE in rustdoc.

22 months agoAuto merge of #100195 - matthiaskrgr:rollup-ovzyyb0, r=matthiaskrgr
bors [Sat, 6 Aug 2022 15:09:59 +0000 (15:09 +0000)]
Auto merge of #100195 - matthiaskrgr:rollup-ovzyyb0, r=matthiaskrgr

Rollup of 4 pull requests

Successful merges:

 - #100094 (Detect type mismatch due to loop that might never iterate)
 - #100132 (Use (actually) dummy place for let-else divergence)
 - #100167 (Recover `require`, `include` instead of `use` in item)
 - #100193 (Remove more Clean trait implementations)

Failed merges:

r? `@ghost`
`@rustbot` modify labels: rollup

22 months agoRollup merge of #100193 - GuillaumeGomez:rm-clean-impls, r=notriddle
Matthias Krüger [Sat, 6 Aug 2022 14:16:00 +0000 (16:16 +0200)]
Rollup merge of #100193 - GuillaumeGomez:rm-clean-impls, r=notriddle

Remove more Clean trait implementations

Follow-up of https://github.com/rust-lang/rust/pull/99638.

r? `@notriddle`

22 months agoRollup merge of #100167 - chenyukang:require-suggestion, r=estebank
Matthias Krüger [Sat, 6 Aug 2022 14:15:59 +0000 (16:15 +0200)]
Rollup merge of #100167 - chenyukang:require-suggestion, r=estebank

Recover `require`, `include` instead of `use` in item

Fix #100140

22 months agoRollup merge of #100132 - compiler-errors:issue-100103, r=tmiasko
Matthias Krüger [Sat, 6 Aug 2022 14:15:56 +0000 (16:15 +0200)]
Rollup merge of #100132 - compiler-errors:issue-100103, r=tmiasko

Use (actually) dummy place for let-else divergence

Fixes #100103

22 months agoRollup merge of #100094 - lyming2007:issue-98982, r=estebank
Matthias Krüger [Sat, 6 Aug 2022 14:15:56 +0000 (16:15 +0200)]
Rollup merge of #100094 - lyming2007:issue-98982, r=estebank

Detect type mismatch due to loop that might never iterate

When loop as tail expression causes a miss match type E0308 error, recursively get the return statement and add diagnostic information on it.

22 months agoAuto merge of #99893 - compiler-errors:issue-99387, r=davidtwco
bors [Sat, 6 Aug 2022 12:29:11 +0000 (12:29 +0000)]
Auto merge of #99893 - compiler-errors:issue-99387, r=davidtwco

Delay formatting trimmed path until lint/error is emitted

Fixes #99387

r? `@davidtwco`

22 months agoremove Clean trait implementation for hir::PolyTraitRef
Guillaume Gomez [Sat, 6 Aug 2022 09:54:54 +0000 (11:54 +0200)]
remove Clean trait implementation for hir::PolyTraitRef

22 months agoremove Clean trait implementation for hir::GenericBound
Guillaume Gomez [Sat, 6 Aug 2022 09:52:27 +0000 (11:52 +0200)]
remove Clean trait implementation for hir::GenericBound

22 months agoAuto merge of #100172 - Mark-Simulacrum:bump-165, r=Mark-Simulacrum
bors [Sat, 6 Aug 2022 09:37:32 +0000 (09:37 +0000)]
Auto merge of #100172 - Mark-Simulacrum:bump-165, r=Mark-Simulacrum

Bump to 1.65.0

r? `@Mark-Simulacrum`

22 months agoAuto merge of #99743 - compiler-errors:fulfillment-context-cleanups, r=jackh726
bors [Sat, 6 Aug 2022 06:48:15 +0000 (06:48 +0000)]
Auto merge of #99743 - compiler-errors:fulfillment-context-cleanups, r=jackh726

Some `FulfillmentContext`-related cleanups

Use `ObligationCtxt` in some places, remove some `FulfillmentContext`s in others...

r? types

22 months agoAuto merge of #100035 - workingjubilee:merge-functions, r=nikic
bors [Fri, 5 Aug 2022 23:11:49 +0000 (23:11 +0000)]
Auto merge of #100035 - workingjubilee:merge-functions, r=nikic

Enable function merging when opt is for size

It is, of course, natural to want to merge aliasing functions when
optimizing for code size, since that can eliminate several bytes.
And an exhaustive match helps make the code less brittle.

Closes #98215.

22 months agoEnable function merging when opt is for size
Jubilee Young [Mon, 1 Aug 2022 21:04:14 +0000 (14:04 -0700)]
Enable function merging when opt is for size

It is, of course, natural to want to merge aliasing functions when
optimizing for code size, since that can eliminate several bytes.
And an exhaustive match helps make the code less brittle.

22 months agoimplement #98982
Yiming Lei [Wed, 3 Aug 2022 04:46:07 +0000 (21:46 -0700)]
implement #98982
when loop as tail expression for miss match type E0308 error, recursively get
the return statement and add diagnostic information on it
use rustc_hir::intravisit to collect the return expression
modified:   compiler/rustc_typeck/src/check/coercion.rs
new file:   src/test/ui/typeck/issue-98982.rs
new file:   src/test/ui/typeck/issue-98982.stderr

22 months agoUpdate RELEASES.md
Josh Stone [Fri, 5 Aug 2022 17:02:46 +0000 (10:02 -0700)]
Update RELEASES.md

Co-authored-by: Trevor Spiteri <tspiteri@ieee.org>
22 months agoUpdate RELEASES.md
Josh Stone [Fri, 5 Aug 2022 17:02:38 +0000 (10:02 -0700)]
Update RELEASES.md

Co-authored-by: Trevor Spiteri <tspiteri@ieee.org>
22 months agoascii -> ASCII in code comment
Xavier Noria [Fri, 5 Aug 2022 16:45:42 +0000 (18:45 +0200)]
ascii -> ASCII in code comment

22 months agomove DiagnosticArgFromDisplay into rustc_errors
Michael Goulet [Fri, 5 Aug 2022 16:43:52 +0000 (16:43 +0000)]
move DiagnosticArgFromDisplay into rustc_errors

22 months agoDelay formatting trimmed path until lint/error is emitted
Michael Goulet [Fri, 29 Jul 2022 06:48:39 +0000 (06:48 +0000)]
Delay formatting trimmed path until lint/error is emitted

22 months agoAuto merge of #100174 - Dylan-DPC:rollup-wnskbk6, r=Dylan-DPC
bors [Fri, 5 Aug 2022 16:35:15 +0000 (16:35 +0000)]
Auto merge of #100174 - Dylan-DPC:rollup-wnskbk6, r=Dylan-DPC

Rollup of 6 pull requests

Successful merges:

 - #99835 (Suggest adding/removing `ref` for binding patterns)
 - #100155 (Use `node_type_opt` to skip over generics that were not expected)
 - #100157 (rustdoc: use `collect()` instead of repeatedly pushing)
 - #100158 (kmc-solid: Add a stub implementation of #98246 (`File::set_times`))
 - #100166 (Remove more Clean trait implementations)
 - #100168 (Improve diagnostics for `const a: = expr;`)

Failed merges:

r? `@ghost`
`@rustbot` modify labels: rollup

22 months agoRollup merge of #100168 - WaffleLapkin:improve_diagnostics_for_missing_type_in_a_cons...
Dylan DPC [Fri, 5 Aug 2022 16:24:37 +0000 (21:54 +0530)]
Rollup merge of #100168 - WaffleLapkin:improve_diagnostics_for_missing_type_in_a_const_item, r=compiler-errors

Improve diagnostics for `const a: = expr;`

Adds a suggestion to write a type when there is a colon, but the type is not present.
I've also shrunk spans a little, so the suggestions are a little nicer.

Resolves #100146

r? `@compiler-errors`

22 months agoRollup merge of #100166 - GuillaumeGomez:rm-clean-impls, r=Dylan-DPC
Dylan DPC [Fri, 5 Aug 2022 16:24:36 +0000 (21:54 +0530)]
Rollup merge of #100166 - GuillaumeGomez:rm-clean-impls, r=Dylan-DPC

Remove more Clean trait implementations

Follow-up of https://github.com/rust-lang/rust/pull/99638.

r? `@notriddle`

22 months agoRollup merge of #100158 - solid-rs:patch/kmc-solid/follow-up-98246, r=joshtriplett
Dylan DPC [Fri, 5 Aug 2022 16:24:35 +0000 (21:54 +0530)]
Rollup merge of #100158 - solid-rs:patch/kmc-solid/follow-up-98246, r=joshtriplett

kmc-solid: Add a stub implementation of #98246 (`File::set_times`)

 Fixes the build failure of the [`*-kmc-solid_*`](https://doc.rust-lang.org/nightly/rustc/platform-support/kmc-solid.html) Tier 3 targets after #98246.

This target does not support setting a modification time and access time separately, hence stubbing out the implementation.

22 months agoRollup merge of #100157 - rust-lang:notriddle/use-map-instead-of-repeated-push, r...
Dylan DPC [Fri, 5 Aug 2022 16:24:34 +0000 (21:54 +0530)]
Rollup merge of #100157 - rust-lang:notriddle/use-map-instead-of-repeated-push, r=Dylan-DPC

rustdoc: use `collect()` instead of repeatedly pushing

22 months agoRollup merge of #100155 - compiler-errors:issue-100154, r=jackh726
Dylan DPC [Fri, 5 Aug 2022 16:24:33 +0000 (21:54 +0530)]
Rollup merge of #100155 - compiler-errors:issue-100154, r=jackh726

Use `node_type_opt` to skip over generics that were not expected

Fixes #100154

22 months agoRollup merge of #99835 - TaKO8Ki:suggest-adding-or-removing-ref-for-binding-pattern...
Dylan DPC [Fri, 5 Aug 2022 16:24:32 +0000 (21:54 +0530)]
Rollup merge of #99835 - TaKO8Ki:suggest-adding-or-removing-ref-for-binding-pattern, r=estebank

Suggest adding/removing `ref` for binding patterns

This fixes what a fixme comment says.

r? `@estebank`

22 months agoBump to 1.65.0
Mark Rousskov [Fri, 5 Aug 2022 15:32:46 +0000 (11:32 -0400)]
Bump to 1.65.0

22 months agoOptimize `pointer::as_aligned_to`
Maybe Waffle [Fri, 5 Aug 2022 13:11:47 +0000 (17:11 +0400)]
Optimize `pointer::as_aligned_to`

22 months agoAuto merge of #100073 - dpaoliello:externvar, r=michaelwoerister
bors [Fri, 5 Aug 2022 13:05:34 +0000 (13:05 +0000)]
Auto merge of #100073 - dpaoliello:externvar, r=michaelwoerister

Add test for raw-dylib with an external variable

All existing tests of link kind `raw-dylib` only validate the ability to link against functions, but it is also possible to link against variables.

This adds tests for linking against a variable using `raw-dylib` both by-name and by-ordinal.

22 months agoImprove diagnostics for `const a: = expr;`
Maybe Waffle [Fri, 5 Aug 2022 11:41:42 +0000 (15:41 +0400)]
Improve diagnostics for `const a: = expr;`

22 months agorecover require,include instead of use in item
yukang [Fri, 5 Aug 2022 11:20:03 +0000 (19:20 +0800)]
recover require,include instead of use in item

22 months agoremove Clean trait implementation for hir::TraitItem
Guillaume Gomez [Fri, 5 Aug 2022 10:08:32 +0000 (12:08 +0200)]
remove Clean trait implementation for hir::TraitItem

22 months agoremove Clean trait implementation for hir::PolyTraitRef
Guillaume Gomez [Fri, 5 Aug 2022 10:01:30 +0000 (12:01 +0200)]
remove Clean trait implementation for hir::PolyTraitRef

22 months agoAuto merge of #95977 - FabianWolff:issue-92790-dead-tuple, r=estebank
bors [Fri, 5 Aug 2022 09:32:26 +0000 (09:32 +0000)]
Auto merge of #95977 - FabianWolff:issue-92790-dead-tuple, r=estebank

Warn about dead tuple struct fields

Continuation of #92972. Fixes #92790.

The language team has already commented on this in https://github.com/rust-lang/rust/pull/92972#issuecomment-1021511970; I have incorporated their requests here. Specifically, there is now a new allow-by-default `unused_tuple_struct_fields` lint (name bikesheddable), and fields of unit type are ignored (https://github.com/rust-lang/rust/pull/92972#issuecomment-1021815408), so error messages look like this:
```
error: field is never read: `1`
  --> $DIR/tuple-struct-field.rs:6:21
   |
LL | struct Wrapper(i32, [u8; LEN], String);
   |                     ^^^^^^^^^
   |
help: change the field to unit type to suppress this warning while preserving the field numbering
   |
LL | struct Wrapper(i32, (), String);
   |                     ~~
```
r? `@joshtriplett`

22 months agoAuto merge of #99867 - spastorino:refactor-remap-lifetimes, r=nikomatsakis
bors [Fri, 5 Aug 2022 06:35:12 +0000 (06:35 +0000)]
Auto merge of #99867 - spastorino:refactor-remap-lifetimes, r=nikomatsakis

Split create_def and lowering of lifetimes for opaque types and bare async fns

r? `@cjgillot`

This work is kind of half-way, but I think it could be merged anyway.
I think we should be able to remove all the vacant arms in `new_named_lifetime_with_res`, if I'm not wrong that requires visiting more nodes. We can do that as a follow up.
In follow-up PRs, besides the thing mentioned previously, I'll be trying to remove `LifetimeCaptureContext`, `captured_lifetimes` as a global data structure, global `binders_to_ignore` and all their friends :).

Also try to remap in a more general way based on def-ids.

22 months agoAuto merge of #97085 - rylev:test-issue-33172, r=wesleywiser
bors [Fri, 5 Aug 2022 03:26:47 +0000 (03:26 +0000)]
Auto merge of #97085 - rylev:test-issue-33172, r=wesleywiser

Add a test for issue #33172

Adds a test confirming that #33172 has been fixed.

CDB has some surprising results as it looks like the supposedly unmangled static's symbol name is prefixed when it shouldn't be.

r? `@wesleywiser`

Closes #33172

22 months agorustdoc: use `collect()` instead of repeatedly pushing to bounds
Michael Howell [Fri, 5 Aug 2022 01:13:53 +0000 (18:13 -0700)]
rustdoc: use `collect()` instead of repeatedly pushing to bounds

22 months agorustdoc: use `collect()` instead of repeatedly pushing to bindings
Michael Howell [Fri, 5 Aug 2022 00:19:57 +0000 (17:19 -0700)]
rustdoc: use `collect()` instead of repeatedly pushing to bindings

22 months agoAuto merge of #95026 - cuviper:bump-linux-min, r=Mark-Simulacrum
bors [Thu, 4 Aug 2022 23:56:07 +0000 (23:56 +0000)]
Auto merge of #95026 - cuviper:bump-linux-min, r=Mark-Simulacrum

Increase the minimum linux-gnu versions

This is implementing the MCP from rust-lang/compiler-team#493. It is
increasing the minimum requirements of a couple Tier 1 targets, and
others at lower tiers, so this should go through FCP sign-offs for both
`T-compiler` and `T-release`.

The new `linux-gnu` baseline is kernel 3.2 and glibc 2.17. We will also
take that kernel as the minimum floor for _all_ `*-linux-*` targets, so
it may be broadly assumed in the implementation of the standard library.
That does not preclude specific targets from having greater requirements
where it makes sense, like a new arch needing something newer, or a
platform like `linux-android` choosing a newer baseline.

22 months agoLink /rustroot/lib stuff for clang -m32 to see
Josh Stone [Thu, 4 Aug 2022 23:25:01 +0000 (16:25 -0700)]
Link /rustroot/lib stuff for clang -m32 to see

22 months agoopt node type
Michael Goulet [Thu, 4 Aug 2022 22:43:39 +0000 (22:43 +0000)]
opt node type

22 months agoAuto merge of #100151 - matthiaskrgr:rollup-irqwvj2, r=matthiaskrgr
bors [Thu, 4 Aug 2022 21:03:48 +0000 (21:03 +0000)]
Auto merge of #100151 - matthiaskrgr:rollup-irqwvj2, r=matthiaskrgr

Rollup of 7 pull requests

Successful merges:

 - #98796 (Do not exclusively suggest `;` when `,` is also a choice)
 - #99772 (Re-enable submodule archive downloads.)
 - #100058 (Suggest a positional formatting argument instead of a captured argument)
 - #100093 (Enable unused_parens for match arms)
 - #100095 (More EarlyBinder cleanups)
 - #100138 (Remove more Clean trait implementations)
 - #100148 (RustWrapper: update for TypedPointerType in LLVM)

Failed merges:

r? `@ghost`
`@rustbot` modify labels: rollup

22 months agoRollup merge of #100148 - durin42:llvm-16-pointertype, r=nikic
Matthias Krüger [Thu, 4 Aug 2022 20:25:06 +0000 (22:25 +0200)]
Rollup merge of #100148 - durin42:llvm-16-pointertype, r=nikic

RustWrapper: update for TypedPointerType in LLVM

This is a result of https://reviews.llvm.org/D130592.

r? `@nikic`

22 months agoRollup merge of #100138 - GuillaumeGomez:rm-clean-impls, r=notriddle
Matthias Krüger [Thu, 4 Aug 2022 20:25:05 +0000 (22:25 +0200)]
Rollup merge of #100138 - GuillaumeGomez:rm-clean-impls, r=notriddle

Remove more Clean trait implementations

Follow-up of https://github.com/rust-lang/rust/pull/99638.

r? `@notriddle`

22 months agoRollup merge of #100095 - jackh726:early-binder, r=lcnr
Matthias Krüger [Thu, 4 Aug 2022 20:25:04 +0000 (22:25 +0200)]
Rollup merge of #100095 - jackh726:early-binder, r=lcnr

More EarlyBinder cleanups

Each commit is independent

r? types

22 months agoRollup merge of #100093 - wcampbell0x2a:unused-parens-for-match-arms, r=petrochenkov
Matthias Krüger [Thu, 4 Aug 2022 20:25:02 +0000 (22:25 +0200)]
Rollup merge of #100093 - wcampbell0x2a:unused-parens-for-match-arms, r=petrochenkov

Enable unused_parens for match arms

Fixes: https://github.com/rust-lang/rust/issues/92751
Currently I can't get the `stderr` to work with `./x.py test`, but this should fix the issue. Help would be appreciated!

22 months agoRollup merge of #100058 - TaKO8Ki:suggest-positional-formatting-argument-instead...
Matthias Krüger [Thu, 4 Aug 2022 20:25:01 +0000 (22:25 +0200)]
Rollup merge of #100058 - TaKO8Ki:suggest-positional-formatting-argument-instead-of-format-args-capture, r=estebank

Suggest a positional formatting argument instead of a captured argument

This patch fixes a part of #96999.

fixes #98241
fixes #97311

r? `@estebank`

22 months agoRollup merge of #99772 - ehuss:reenable-submodule-archive, r=Mark-Simulacrum
Matthias Krüger [Thu, 4 Aug 2022 20:25:00 +0000 (22:25 +0200)]
Rollup merge of #99772 - ehuss:reenable-submodule-archive, r=Mark-Simulacrum

Re-enable submodule archive downloads.

This is effectively a revert of #98423 (though it keeps the `--depth 1` flag since that is still helpful).

GitHub has indicated that they have been working on the original issue, and my testing shows that the llvm-project archive download now succeeds 100% of the time.

This should save about a minute on every job.

22 months agoRollup merge of #98796 - compiler-errors:no-semi-if-comma, r=estebank
Matthias Krüger [Thu, 4 Aug 2022 20:24:59 +0000 (22:24 +0200)]
Rollup merge of #98796 - compiler-errors:no-semi-if-comma, r=estebank

Do not exclusively suggest `;` when `,` is also a choice

Fixes #96791

22 months agoAdd test for raw-dylib with an external variable
Daniel Paoliello [Tue, 2 Aug 2022 18:36:41 +0000 (11:36 -0700)]
Add test for raw-dylib with an external variable

22 months agoReword some 1.63.0 release notes
Josh Stone [Thu, 4 Aug 2022 18:48:16 +0000 (11:48 -0700)]
Reword some 1.63.0 release notes

22 months agoFix typo
Santiago Pastorino [Thu, 4 Aug 2022 15:48:20 +0000 (12:48 -0300)]
Fix typo

22 months agoImprove opt_local_def_id docs
Santiago Pastorino [Thu, 4 Aug 2022 15:47:58 +0000 (12:47 -0300)]
Improve opt_local_def_id docs

22 months agoComplete the links for 1.63.0 stabilizations
Josh Stone [Thu, 4 Aug 2022 18:08:15 +0000 (11:08 -0700)]
Complete the links for 1.63.0 stabilizations

22 months agoApply suggested updates
Josh Stone [Wed, 20 Jul 2022 20:16:27 +0000 (13:16 -0700)]
Apply suggested updates

22 months agoAdd release notes for Rust 1.63.0
Josh Stone [Wed, 20 Jul 2022 19:22:34 +0000 (12:22 -0700)]
Add release notes for Rust 1.63.0

22 months agoRelease 1.62.1
Mark Rousskov [Fri, 15 Jul 2022 21:51:06 +0000 (17:51 -0400)]
Release 1.62.1

22 months agoDowngrade dist-powerpc-linux binutils to 2.30
Josh Stone [Thu, 4 Aug 2022 17:16:19 +0000 (10:16 -0700)]
Downgrade dist-powerpc-linux binutils to 2.30

With binutils 2.32, we were getting errors like this:

    relocation truncated to fit: R_PPC_PLTREL24 against symbol
        `__cxa_atexit@@GLIBC_2.1.3' defined in .plt section in
        /x-tools/powerpc-unknown-linux-gnu/powerpc-unknown-linux-gnu/sysroot/usr/lib/crt1.o

but it builds okay with binutils 2.30.

22 months agoImprove record_def_id_remap docs
Santiago Pastorino [Thu, 4 Aug 2022 15:47:19 +0000 (12:47 -0300)]
Improve record_def_id_remap docs