]> git.lizzy.rs Git - rust.git/log
rust.git
20 months agoImplItemKind::TyAlias => ImplItemKind::Type
Michael Goulet [Sun, 9 Oct 2022 07:09:57 +0000 (07:09 +0000)]
ImplItemKind::TyAlias => ImplItemKind::Type

21 months agoSuggest `.into()` when all other coercion suggestions fail
Michael Goulet [Fri, 30 Sep 2022 00:06:19 +0000 (00:06 +0000)]
Suggest `.into()` when all other coercion suggestions fail

21 months agoAuto merge of #101768 - sunfishcode:sunfishcode/wasi-stdio-lock-asfd, r=joshtriplett
bors [Tue, 4 Oct 2022 23:22:16 +0000 (23:22 +0000)]
Auto merge of #101768 - sunfishcode:sunfishcode/wasi-stdio-lock-asfd, r=joshtriplett

Add `AsFd` implementations for stdio lock types on WASI.

This mirrors the implementations on Unix platforms, and also mirrors the existing `AsRawFd` impls.

This is similar to #100892, but is for the `*Lock` types.

21 months agoAuto merge of #102666 - matthiaskrgr:rollup-tuge18t, r=matthiaskrgr
bors [Tue, 4 Oct 2022 16:29:26 +0000 (16:29 +0000)]
Auto merge of #102666 - matthiaskrgr:rollup-tuge18t, r=matthiaskrgr

Rollup of 6 pull requests

Successful merges:

 - #102241 (Package `rust-docs-json` into nightly components (take 3))
 - #102488 (Check generic argument compatibility when projecting assoc ty)
 - #102647 (Only allow ~const bounds for traits with #[const_trait])
 - #102648 (Add test for #102605)
 - #102651 (It's not about types or consts, but the lack of regions)
 - #102653 (resolve instance: missing value to `delay_span_bug`)

Failed merges:

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

21 months agoRollup merge of #102653 - lcnr:delay_span_bug, r=fee1-dead
Matthias Krüger [Tue, 4 Oct 2022 16:26:41 +0000 (18:26 +0200)]
Rollup merge of #102653 - lcnr:delay_span_bug, r=fee1-dead

resolve instance: missing value to `delay_span_bug`

21 months agoRollup merge of #102651 - oli-obk:non_region_things, r=lcnr
Matthias Krüger [Tue, 4 Oct 2022 16:26:41 +0000 (18:26 +0200)]
Rollup merge of #102651 - oli-obk:non_region_things, r=lcnr

It's not about types or consts, but the lack of regions

pulled out of https://github.com/rust-lang/rust/pull/101900 which adds a fourth kind of non-lifetime generic parameter, and the naming of these methods would get ridiculous.

21 months agoRollup merge of #102648 - Rageking8:add-test-for-#102605, r=compiler-errors
Matthias Krüger [Tue, 4 Oct 2022 16:26:40 +0000 (18:26 +0200)]
Rollup merge of #102648 - Rageking8:add-test-for-#102605, r=compiler-errors

Add test for #102605

Fixes #102605

21 months agoRollup merge of #102647 - oli-obk:tilde_const_bounds, r=fee1-dead
Matthias Krüger [Tue, 4 Oct 2022 16:26:39 +0000 (18:26 +0200)]
Rollup merge of #102647 - oli-obk:tilde_const_bounds, r=fee1-dead

Only allow ~const bounds for traits with #[const_trait]

r? `@fee1-dead`

21 months agoRollup merge of #102488 - compiler-errors:gat-compatibility, r=oli-obk
Matthias Krüger [Tue, 4 Oct 2022 16:26:39 +0000 (18:26 +0200)]
Rollup merge of #102488 - compiler-errors:gat-compatibility, r=oli-obk

Check generic argument compatibility when projecting assoc ty

Fixes #102114

21 months agoRollup merge of #102241 - jyn514:manifest-json-docs, r=Mark-Simulacrum
Matthias Krüger [Tue, 4 Oct 2022 16:26:38 +0000 (18:26 +0200)]
Rollup merge of #102241 - jyn514:manifest-json-docs, r=Mark-Simulacrum

Package `rust-docs-json` into nightly components (take 3)

`dist` creates a `rust-docs-json.tar.xz` tarfile. But build-manifest expected it to be named `rust-docs-json-preview.tar.xz`. Change build-manifest to allow the name without the `-preview` suffix.

I haven't actually tested this :( build-manifest is a pain to run locally.

21 months agomissing value to delay_span_bug
lcnr [Tue, 4 Oct 2022 12:18:11 +0000 (14:18 +0200)]
missing value to delay_span_bug

21 months agoIt's not about types or consts, but the lack of regions
Oli Scherer [Tue, 4 Oct 2022 09:43:34 +0000 (09:43 +0000)]
It's not about types or consts, but the lack of regions

21 months agoAuto merge of #102652 - Dylan-DPC:rollup-6ff8ct8, r=Dylan-DPC
bors [Tue, 4 Oct 2022 13:04:57 +0000 (13:04 +0000)]
Auto merge of #102652 - Dylan-DPC:rollup-6ff8ct8, r=Dylan-DPC

Rollup of 6 pull requests

Successful merges:

 - #101189 (Implement `Ready::into_inner()`)
 - #101642 (Fix in-place collection leak when remaining element destructor panic)
 - #102489 (Normalize substs before resolving instance in `NoopMethodCall` lint)
 - #102559 (Don't ICE when trying to copy unsized value in const prop)
 - #102568 (Lint against nested opaque types that don't satisfy associated type bounds)
 - #102633 (Fix rustdoc ICE in invalid_rust_codeblocks lint)

Failed merges:

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

21 months agoRollup merge of #102633 - Nilstrieb:rustdoc-lint-🏳️‍⚧️late, r=davidtwco
Dylan DPC [Tue, 4 Oct 2022 10:41:03 +0000 (16:11 +0530)]
Rollup merge of #102633 - Nilstrieb:rustdoc-lint-🏳️‍⚧️late, r=davidtwco

Fix rustdoc ICE in invalid_rust_codeblocks lint

The diagnostic message extraction code didn't handle translations yet.
Fixes #102603
Fixes #102631

r? `@davidtwco`

21 months agoRollup merge of #102568 - compiler-errors:lint-unsatisfied-opaques, r=oli-obk
Dylan DPC [Tue, 4 Oct 2022 10:41:02 +0000 (16:11 +0530)]
Rollup merge of #102568 - compiler-errors:lint-unsatisfied-opaques, r=oli-obk

Lint against nested opaque types that don't satisfy associated type bounds

See the test failures for examples of places where this lint would fire.

r? `@oli-obk`

21 months agoRollup merge of #102559 - compiler-errors:issue-102553, r=oli-obk
Dylan DPC [Tue, 4 Oct 2022 10:41:02 +0000 (16:11 +0530)]
Rollup merge of #102559 - compiler-errors:issue-102553, r=oli-obk

Don't ICE when trying to copy unsized value in const prop

When we have a trivially false where-clause predicate like `Self: Sized` where `Self = dyn Trait`, we sometimes don't throw an error during typeck for an illegal operation such as copying an unsized type.

This, unfortunately, cannot be made into an error (at least not without some migration -- see #95611 for example), but we should at least not ICE, since this function will never actually be reachable from main, for example.

r? `@RalfJung` since I think you added these assertions? but feel free to reassign.

Fixes #102553

21 months agoRollup merge of #102489 - compiler-errors:issue-102074, r=oli-obk
Dylan DPC [Tue, 4 Oct 2022 10:41:01 +0000 (16:11 +0530)]
Rollup merge of #102489 - compiler-errors:issue-102074, r=oli-obk

Normalize substs before resolving instance in `NoopMethodCall` lint

Fixes #102074

r? types

21 months agoRollup merge of #101642 - SkiFire13:fix-inplace-collection-leak, r=the8472
Dylan DPC [Tue, 4 Oct 2022 10:41:01 +0000 (16:11 +0530)]
Rollup merge of #101642 - SkiFire13:fix-inplace-collection-leak, r=the8472

Fix in-place collection leak when remaining element destructor panic

Fixes #101628

cc `@the8472`

I went for the drop guard route, placing it immediately before the `forget_allocation_drop_remaining` call and after the comment, as to signal they are closely related.

I also updated the test to check for the leak, though the only change really needed was removing the leak clean up for miri since now that's no longer leaked.

21 months agoRollup merge of #101189 - daxpedda:ready-into-inner, r=joshtriplett
Dylan DPC [Tue, 4 Oct 2022 10:41:00 +0000 (16:11 +0530)]
Rollup merge of #101189 - daxpedda:ready-into-inner, r=joshtriplett

Implement `Ready::into_inner()`

Tracking issue: #101196.

This implements a method to unwrap the value inside a `Ready` outside an async context.
See https://docs.rs/futures/0.3.24/futures/future/struct.Ready.html#method.into_inner for previous work.

This was discussed in [Zulip beforehand](https://rust-lang.zulipchat.com/#narrow/stream/219381-t-libs/topic/.60Ready.3A.3Ainto_inner.28.29.60):
> An example I'm hitting right now:
I have a cross-platform library that provides a functions that returns a `Future`. The only reason why it returns a `Future` is because the WASM platform requires it, but the native doesn't, to make a cross-platform API that is equal for all I just return a `Ready` on the native targets.
>
> Now I would like to expose native-only functions that aren't async, that users can use to avoid having to deal with async when they are targeting native. With `into_inner` that's easily solvable now.
>
> I want to point out that some internal restructuring could be used to solve that problem too, but in this case it's not that simple, the library uses internal traits that return the `Future` already and playing around with that would introduce unnecessary `cfg` in a lot more places. So it is really only a quality-of-life feature.

21 months agoAuto merge of #102622 - camsteffen:move-layout, r=fee1-dead
bors [Tue, 4 Oct 2022 09:29:07 +0000 (09:29 +0000)]
Auto merge of #102622 - camsteffen:move-layout, r=fee1-dead

Move layout_of and friends from rustc_middle to rustc_ty_utils

Breaks up the very large module that is `rustc_middle::ty::layout` by fork-lifting some queries into `rustc_ty_utils::{abi, layout}`.

This does set back `rustc_ty_utils` to having untranslatable diagnostics. I'd like to leave this as a separate task.

21 months agoMerge the `~const` and `impl const` checks and add some explanatory notes
Oli Scherer [Tue, 4 Oct 2022 08:59:20 +0000 (08:59 +0000)]
Merge the `~const` and `impl const` checks and add some explanatory notes

21 months agoadd test for #102605
Rageking8 [Tue, 4 Oct 2022 08:13:50 +0000 (16:13 +0800)]
add test for #102605

21 months agoOnly allow ~const bounds for traits with #[const_trait]
Oli Scherer [Mon, 3 Oct 2022 17:08:42 +0000 (17:08 +0000)]
Only allow ~const bounds for traits with #[const_trait]

21 months agoAuto merge of #102644 - matthiaskrgr:rollup-rg0sw41, r=matthiaskrgr
bors [Tue, 4 Oct 2022 06:47:21 +0000 (06:47 +0000)]
Auto merge of #102644 - matthiaskrgr:rollup-rg0sw41, r=matthiaskrgr

Rollup of 7 pull requests

Successful merges:

 - #102441 (Suggest unwrap_or_else when a closure is given)
 - #102547 (Migrate CSS theme for search results)
 - #102567 (Delay evaluating lint primary message until after it would be suppressed)
 - #102624 (rustdoc: remove font family CSS on `.rustdoc-toggle summary::before`)
 - #102628 (Change the parameter name of From::from to `value`)
 - #102637 (Ignore fuchsia on two compiler tests)
 - #102639 (Improve spans when splitting multi-char operator tokens for proc macros.)

Failed merges:

 - #102496 (Suggest `.into()` when all other coercion suggestions fail)

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

21 months agoRollup merge of #102639 - nnethercote:improve-spans-splitting, r=Aaron1011
Matthias Krüger [Tue, 4 Oct 2022 04:14:13 +0000 (06:14 +0200)]
Rollup merge of #102639 - nnethercote:improve-spans-splitting, r=Aaron1011

Improve spans when splitting multi-char operator tokens for proc macros.

When a two-char (or three-char) operator token is split into single-char operator tokens before being passed to a proc macro, the single-char tokens are given the original span of length two (or three). This PR gives them more accurate spans.

r? `@Aaron1011`
cc `@petrochenkov`

21 months agoRollup merge of #102637 - andrewpollack:ignore-fuchsia-two-tests, r=tmandry
Matthias Krüger [Tue, 4 Oct 2022 04:14:12 +0000 (06:14 +0200)]
Rollup merge of #102637 - andrewpollack:ignore-fuchsia-two-tests, r=tmandry

Ignore fuchsia on two compiler tests

Adding `ignore-fuchsia` to two irrelevant compiler tests

cc. `@djkoloski`

r? `@tmandry`

21 months agoRollup merge of #102628 - H4x5:master, r=scottmcm
Matthias Krüger [Tue, 4 Oct 2022 04:14:12 +0000 (06:14 +0200)]
Rollup merge of #102628 - H4x5:master, r=scottmcm

Change the parameter name of From::from to `value`

The `From` trait is currently defined as:
```rust
pub trait From<T>: Sized {
    fn from(_: T) -> Self;
}
```

The name of the argument is `_`. I am proposing to change it to `value`, ie.
```rust
pub trait From<T>: Sized {
    fn from(value: T) -> Self;
}
```

This would be more consistent with the `TryFrom`, which looks like this:
```rust
pub trait TryFrom<T>: Sized {
    type Error;
    fn try_from(value: T) -> Result<Self, Self::Error>;
}
```

The reason for this proposal is twofold:
1. Consistency with the rest of the standard library. The `TryFrom` trait uses `value`, and no `From` implementation uses the default name (as it is quite useless).
2. When generating trait implementations with rust-analyzer/IntelliJ, the parameter name is copied, and it always has to be changed.

Optionally, another name like `x` could be used. I only propose `value` for consistency with `TryFrom`.

Changing parameter names is not a breaking change.

Note: this was originally posted as an internals thread [here](https://internals.rust-lang.org/t/change-the-argument-name-of-from-from/17480)

21 months agoRollup merge of #102624 - notriddle:notriddle/summary-before, r=GuillaumeGomez
Matthias Krüger [Tue, 4 Oct 2022 04:14:11 +0000 (06:14 +0200)]
Rollup merge of #102624 - notriddle:notriddle/summary-before, r=GuillaumeGomez

rustdoc: remove font family CSS on `.rustdoc-toggle summary::before`

This rule became irrelevant since c58246efe47bea09d4f3e70f536e4c9bb7770749 made it so that the `summary::before` pseudo-element contains an SVG instead of text.

21 months agoRollup merge of #102567 - compiler-errors:issue-102561, r=davidtwco
Matthias Krüger [Tue, 4 Oct 2022 04:14:11 +0000 (06:14 +0200)]
Rollup merge of #102567 - compiler-errors:issue-102561, r=davidtwco

Delay evaluating lint primary message until after it would be suppressed

Fixes #102561
Fixes #102572

21 months agoRollup merge of #102547 - GuillaumeGomez:migrate-css-theme-search-result, r=notriddle...
Matthias Krüger [Tue, 4 Oct 2022 04:14:10 +0000 (06:14 +0200)]
Rollup merge of #102547 - GuillaumeGomez:migrate-css-theme-search-result, r=notriddle,jsha

Migrate CSS theme for search results

Part of https://github.com/rust-lang/rust/pull/98460.

Just like https://github.com/rust-lang/rust/pull/102237, I unified theme to how the `ayu` handles this one: only one color for the background when search results are focused or hovered.

You can test it [here](https://rustdoc.crud.net/imperio/migrate-css-theme-search-result/lib2/index.html?search=coo).

cc `@jsha`
r? `@notriddle`

PS: The repetition in GUI tests is getting out of hand so I opened https://github.com/GuillaumeGomez/browser-UI-test/issues/363 to think about adding possibility to declare functions so we can greatly improve this.

21 months agoRollup merge of #102441 - chenyukang:fix-102320-unwrap_or_else, r=compiler-errors
Matthias Krüger [Tue, 4 Oct 2022 04:14:10 +0000 (06:14 +0200)]
Rollup merge of #102441 - chenyukang:fix-102320-unwrap_or_else, r=compiler-errors

Suggest unwrap_or_else when a closure is given

Fixes #102320

r? `@compiler-errors`

21 months agoAuto merge of #102395 - davidtwco:translation-rename-typeck, r=compiler-errors
bors [Tue, 4 Oct 2022 03:57:50 +0000 (03:57 +0000)]
Auto merge of #102395 - davidtwco:translation-rename-typeck, r=compiler-errors

errors: rename `typeck.ftl` to `hir_analysis.ftl`

In #102306, `rustc_typeck` was renamed to `rustc_hir_analysis` but the diagnostic resources were not renamed - which is what this pull request changes.

21 months agoWe are able to resolve methods even if they need subst
Michael Goulet [Tue, 4 Oct 2022 03:22:45 +0000 (03:22 +0000)]
We are able to resolve methods even if they need subst

21 months agoNormalize substs before resolving instance in NoopMethodCall lint
Michael Goulet [Thu, 29 Sep 2022 22:03:03 +0000 (22:03 +0000)]
Normalize substs before resolving instance in NoopMethodCall lint

21 months agoAuto merge of #99099 - Stargateur:phantomdata_debug, r=joshtriplett
bors [Tue, 4 Oct 2022 00:56:14 +0000 (00:56 +0000)]
Auto merge of #99099 - Stargateur:phantomdata_debug, r=joshtriplett

Add T to PhantomData impl Debug

This add debug information for `PhantomData`, I believe it's make sense to add this to debug impl of `PhantomData` since `T` is what define what is the `PhantomData` just write `"PhantomData"` is not very useful for debugging.

Alternative:

* `PhantomData::<{}>`
* `PhantomData { t: "str_type" }`

`@rustbot` label +T-libs-api -T-libs

21 months agoImprove spans when splitting multi-char operator tokens for proc macros.
Nicholas Nethercote [Mon, 3 Oct 2022 02:57:47 +0000 (13:57 +1100)]
Improve spans when splitting multi-char operator tokens for proc macros.

21 months agoExtend GUI test to check hover and focus on search results
Guillaume Gomez [Sat, 1 Oct 2022 15:11:33 +0000 (17:11 +0200)]
Extend GUI test to check hover and focus on search results

21 months agoMigrate search results theme style to CSS variables
Guillaume Gomez [Sat, 1 Oct 2022 13:50:47 +0000 (15:50 +0200)]
Migrate search results theme style to CSS variables

21 months agoIgnore fuchsia on two compiler tests
Andrew Pollack [Mon, 3 Oct 2022 21:11:47 +0000 (21:11 +0000)]
Ignore fuchsia on two compiler tests

21 months agoAuto merge of #102632 - matthiaskrgr:rollup-h8s3zmo, r=matthiaskrgr
bors [Mon, 3 Oct 2022 20:22:18 +0000 (20:22 +0000)]
Auto merge of #102632 - matthiaskrgr:rollup-h8s3zmo, r=matthiaskrgr

Rollup of 7 pull requests

Successful merges:

 - #98218 (Document the conditional existence of `alloc::sync` and `alloc::task`.)
 - #99216 (docs: be less harsh in wording for Vec::from_raw_parts)
 - #99460 (docs: Improve AsRef / AsMut docs on blanket impls)
 - #100470 (Tweak `FpCategory` example order.)
 - #101040 (Fix `#[derive(Default)]` on a generic `#[default]` enum adding unnecessary `Default` bounds)
 - #101308 (introduce `{char, u8}::is_ascii_octdigit`)
 - #102486 (Add diagnostic struct for const eval error in `rustc_middle`)

Failed merges:

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

21 months agoFix rustdoc ICE in invalid_rust_codeblocks lint
Nilstrieb [Mon, 3 Oct 2022 19:01:19 +0000 (21:01 +0200)]
Fix rustdoc ICE in invalid_rust_codeblocks lint

The diagnostic message extraction code didn't handle translations yet.

21 months agoRollup merge of #102486 - pierwill:middle-const-eval-err, r=compiler-errors
Matthias Krüger [Mon, 3 Oct 2022 18:58:56 +0000 (20:58 +0200)]
Rollup merge of #102486 - pierwill:middle-const-eval-err, r=compiler-errors

Add diagnostic struct for const eval error in `rustc_middle`

Part of #100717.

r? `@ghost`

21 months agoRollup merge of #101308 - nerdypepper:feature/is-ascii-octdigit, r=joshtriplett
Matthias Krüger [Mon, 3 Oct 2022 18:58:56 +0000 (20:58 +0200)]
Rollup merge of #101308 - nerdypepper:feature/is-ascii-octdigit, r=joshtriplett

introduce `{char, u8}::is_ascii_octdigit`

This feature adds two new APIs: `char::is_ascii_octdigit` and `u8::is_ascii_octdigit`, under the feature gate `is_ascii_octdigit`. These methods are shorthands for `char::is_digit(self, 8)` and `u8::is_digit(self, 8)`:

```rust
// core::char

impl char {
    pub fn is_ascii_octdigit(self) -> bool;
}

// core::num

impl u8 {
    pub fn is_ascii_octdigit(self) -> bool;
}
```

---

Couple of things I need help understanding:

- `const`ness: have I used the right attribute in this case?
- is there a way to run the tests for `core::char` alone, instead of `./x.py test library/core`?

21 months agoRollup merge of #101040 - danielhenrymantilla:no-bounds-for-default-annotated-derive...
Matthias Krüger [Mon, 3 Oct 2022 18:58:55 +0000 (20:58 +0200)]
Rollup merge of #101040 - danielhenrymantilla:no-bounds-for-default-annotated-derive, r=joshtriplett

Fix `#[derive(Default)]` on a generic `#[default]` enum adding unnecessary `Default` bounds

That is, given something like:

```rs
// #[default] on a generic enum does not add `Default` bounds to the type params.
#[derive(Default)]
enum MyOption<T> {
    #[default]
    None,
    Some(T),
}
```

then `MyOption<T> : Default`_as currently implemented_ only holds when `T : Default`, as reported by ```@5225225``` [over Zulip](https://rust-lang.zulipchat.com/#narrow/stream/122651-general/topic/.23.5Bderive.28Default.29.5D.20for.20enums.20with.20fields).

This is contrary to [what the accepted RFC proposes](https://rust-lang.github.io/rfcs/3107-derive-default-enum.html#generated-bounds) (_i.e._, that `T` be allowed not to be itself `Default`), and indeed seems to be a rather unnecessary limitation.

21 months agoRollup merge of #100470 - reitermarkus:patch-1, r=joshtriplett
Matthias Krüger [Mon, 3 Oct 2022 18:58:55 +0000 (20:58 +0200)]
Rollup merge of #100470 - reitermarkus:patch-1, r=joshtriplett

Tweak `FpCategory` example order.

Follow same order for variable declarations and assertions.

21 months agoRollup merge of #99460 - JanBeh:PR_asref_asmut_docs, r=joshtriplett
Matthias Krüger [Mon, 3 Oct 2022 18:58:54 +0000 (20:58 +0200)]
Rollup merge of #99460 - JanBeh:PR_asref_asmut_docs, r=joshtriplett

docs: Improve AsRef / AsMut docs on blanket impls

There are several issues with the current state of `AsRef` and `AsMut` as [discussed here on IRLO](https://internals.rust-lang.org/t/semantics-of-asref/17016). See also #39397, #45742, #73390, #98905, and the FIXMEs [here](https://github.com/rust-lang/rust/blob/1.62.0/library/core/src/convert/mod.rs#L509-L515) and [here](https://github.com/rust-lang/rust/blob/1.62.0/library/core/src/convert/mod.rs#L530-L536). These issues are difficult to fix. This PR aims to update the documentation to better reflect the status-quo and to give advice on how `AsRef` and `AsMut` should be used.

In particular:

- Explicitly mention that `AsRef` and `AsMut` do not auto-dereference generally for all dereferencable types (but only if inner type is a shared and/or mutable reference)
- Give advice to not use `AsRef` or `AsMut` for the sole purpose of dereferencing
- Suggest providing a transitive `AsRef` or `AsMut` implementation for types which implement `Deref`
- Add new section "Reflexivity" in documentation comments for `AsRef` and `AsMut`
- Provide better example for `AsMut`
- Added heading "Relation to `Borrow`" in `AsRef`'s docs to improve structure

21 months agoRollup merge of #99216 - duarten:master, r=joshtriplett
Matthias Krüger [Mon, 3 Oct 2022 18:58:53 +0000 (20:58 +0200)]
Rollup merge of #99216 - duarten:master, r=joshtriplett

docs: be less harsh in wording for Vec::from_raw_parts

In particular, be clear that it is sound to specify memory not
originating from a previous `Vec` allocation. That is already suggested
in other parts of the documentation about zero-alloc conversions to Box<[T]>.

Incorporate a constraint from `slice::from_raw_parts` that was missing
but needs to be fulfilled, since a `Vec` can be converted into a slice.

Fixes https://github.com/rust-lang/rust/issues/98780.

21 months agoRollup merge of #98218 - kpreid:nostdarc, r=joshtriplett
Matthias Krüger [Mon, 3 Oct 2022 18:58:53 +0000 (20:58 +0200)]
Rollup merge of #98218 - kpreid:nostdarc, r=joshtriplett

Document the conditional existence of `alloc::sync` and `alloc::task`.

`alloc` declares

```rust
#[cfg(target_has_atomic = "ptr")]
pub mod sync;
```

but there is no public documentation of this condition. This PR fixes that, so that users of `alloc` can understand how to make their code compile everywhere `alloc` does, if they are writing a library with impls for `Arc`.

The wording is copied from `std::sync::atomic::AtomicPtr`, with additional advice on how to `#[cfg]` for it.

I feel quite uncertain about whether the paragraph I added to `Arc`'s documentation should actually be there, as it is a distraction for anyone using `std`. On the other hand, maybe more reminders that no_std exists would benefit the ecosystem.

Note: `target_has_atomic` is [stabilized](https://github.com/rust-lang/rust/issues/32976) but [not yet documented in the reference](https://github.com/rust-lang/reference/pull/1171).

21 months agoClarify documentation
Giacomo Stevanato [Sat, 10 Sep 2022 13:00:37 +0000 (15:00 +0200)]
Clarify documentation

21 months agoChange the parameter name of From::from to `value`
H4x5 [Mon, 3 Oct 2022 17:36:57 +0000 (13:36 -0400)]
Change the parameter name of From::from to `value`

21 months agoAuto merge of #102627 - matthiaskrgr:rollup-2xtrqkw, r=matthiaskrgr
bors [Mon, 3 Oct 2022 17:31:45 +0000 (17:31 +0000)]
Auto merge of #102627 - matthiaskrgr:rollup-2xtrqkw, r=matthiaskrgr

Rollup of 8 pull requests

Successful merges:

 - #102439 (rustdoc: re-sugar more cross-crate trait bounds)
 - #102569 (Improve `FromStr` example)
 - #102597 (Avoid ICE in printing RPITIT type)
 - #102607 (Improve documentation of `slice::{from_ptr_range, from_ptr_range_mut}`)
 - #102613 (Fix ICE #101739)
 - #102615 (Cleanup some error code explanations)
 - #102617 (`HirId` for `deferred_transmute_checks`)
 - #102620 (Migrate `.stab` elements style to CSS variables)

Failed merges:

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

21 months agoAdd diagnostic struct for const eval error in `rustc_middle`
pierwill [Sun, 2 Oct 2022 21:02:57 +0000 (16:02 -0500)]
Add diagnostic struct for const eval error in `rustc_middle`

Co-authored-by: Michael Goulet <michael@errs.io>
21 months agoRollup merge of #102620 - GuillaumeGomez:css-stab-migration, r=notriddle
Matthias Krüger [Mon, 3 Oct 2022 17:12:20 +0000 (19:12 +0200)]
Rollup merge of #102620 - GuillaumeGomez:css-stab-migration, r=notriddle

Migrate `.stab` elements style to CSS variables

Part of https://github.com/rust-lang/rust/pull/98460.

There should be no UI changes.

r? `@notriddle`

21 months agoRollup merge of #102617 - lcnr:deferred_transmute_checks, r=compiler-errors
Matthias Krüger [Mon, 3 Oct 2022 17:12:19 +0000 (19:12 +0200)]
Rollup merge of #102617 - lcnr:deferred_transmute_checks, r=compiler-errors

`HirId` for `deferred_transmute_checks`

directly interacting with spans is annoying :sparkles:

21 months agoRollup merge of #102615 - Nilstrieb:there-are-many-error-codes, r=compiler-errors
Matthias Krüger [Mon, 3 Oct 2022 17:12:19 +0000 (19:12 +0200)]
Rollup merge of #102615 - Nilstrieb:there-are-many-error-codes, r=compiler-errors

Cleanup some error code explanations

E0045: Use a stable non-C ABI instead
E0092: Use an atomic intrinsic that actually exists
E0161: Don't use box_syntax
E0579: Format ranges in the rustfmt style
E0622: Use the rustfmt style
E0743: Remove feature gate as it's not needed

21 months agoRollup merge of #102613 - TaKO8Ki:fix-part-of-101739, r=compiler-errors
Matthias Krüger [Mon, 3 Oct 2022 17:12:19 +0000 (19:12 +0200)]
Rollup merge of #102613 - TaKO8Ki:fix-part-of-101739, r=compiler-errors

Fix ICE #101739

Fixes a part of #101739

This cannot cover the following case. It causes `too many args provided` error and obligation does not have references error. I want your advice to solve the following cases as well in this pull request or a follow-up.

```rust
#![crate_type = "lib"]
#![feature(transmutability)]
#![allow(dead_code, incomplete_features, non_camel_case_types)]

mod assert {
    use std::mem::BikeshedIntrinsicFrom;

    pub fn is_transmutable<
        Src,
        Dst,
        Context,
        const ASSUME_ALIGNMENT: bool,
        const ASSUME_LIFETIMES: bool,
        const ASSUME_VALIDITY: bool,
        const ASSUME_VISIBILITY: bool,
    >()
    where
        Dst: BikeshedIntrinsicFrom<
            Src,
            Context,
            ASSUME_ALIGNMENT,
            ASSUME_LIFETIMES,
            ASSUME_VALIDITY,
            ASSUME_VISIBILITY,
        >,
    {}
}

fn via_const() {
    struct Context;
    #[repr(C)] struct Src;
    #[repr(C)] struct Dst;

    const FALSE: bool = false;

    assert::is_transmutable::<Src, Dst, Context, FALSE, FALSE, FALSE, FALSE>();
}
```

21 months agoRollup merge of #102607 - WaffleLapkin:docky_docky_slice_from_ptr_range, r=joshtriplett
Matthias Krüger [Mon, 3 Oct 2022 17:12:18 +0000 (19:12 +0200)]
Rollup merge of #102607 - WaffleLapkin:docky_docky_slice_from_ptr_range, r=joshtriplett

Improve documentation of `slice::{from_ptr_range, from_ptr_range_mut}`

Document panic conditions (`T` is a ZST) and sync docs of shared/unique version.

cc `@wx-csy`

21 months agoRollup merge of #102597 - compiler-errors:issue-102571, r=davidtwco
Matthias Krüger [Mon, 3 Oct 2022 17:12:18 +0000 (19:12 +0200)]
Rollup merge of #102597 - compiler-errors:issue-102571, r=davidtwco

Avoid ICE in printing RPITIT type

Fixes #102571

21 months agoRollup merge of #102569 - eduardosm:from_str-example, r=joshtriplett
Matthias Krüger [Mon, 3 Oct 2022 17:12:17 +0000 (19:12 +0200)]
Rollup merge of #102569 - eduardosm:from_str-example, r=joshtriplett

Improve `FromStr` example

The `from_str` implementation from the example had an `unwrap` that would make it panic on invalid input strings. Instead of panicking, it nows returns an error to better reflect the intented behavior of the `FromStr` trait.

21 months agoRollup merge of #102439 - fmease:rustdoc-simplify-cross-crate-trait-bounds, r=Guillau...
Matthias Krüger [Mon, 3 Oct 2022 17:12:17 +0000 (19:12 +0200)]
Rollup merge of #102439 - fmease:rustdoc-simplify-cross-crate-trait-bounds, r=GuillaumeGomez

rustdoc: re-sugar more cross-crate trait bounds

Previously, we would only ever re-sugar cross-crate predicates like `Type: Trait, <Type as Trait>::Name == Rhs` to `Type: Trait<Name = Rhs>` if the `Type` was a generic parameter like `Self` or `T`. With this PR, `Type` can be any type.

Most notably, this means that we now re-sugar predicates involving associated types (where `Type` is of the form `Self::Name`) which are then picked up by the pre-existing logic that re-sugars them into bounds. As a result of that, the associated type `IntoIter` of `std`'s `IntoIterator` trait (re-exported from `core`) is no longer rendered as:

```rust
type IntoIter: Iterator
where
    <Self::IntoIter as Iterator>::Item == Self::Item;
```

but as one would expect: `type IntoIter: Iterator<Item = Self::Item>;`.

Cross-crate closure bounds like `F: Fn(i32) -> bool` are now also rendered properly (previously, the return type (`Self::Output`) would not be rendered and we would show the underlying equality predicate).

Fixes #77763.
Fixes #84579.
Fixes #102142.

`@rustbot` label T-rustdoc A-cross-crate-reexports
r? rustdoc

21 months agorustdoc: remove font family CSS on `.rustdoc-toggle summary::before`
Michael Howell [Mon, 3 Oct 2022 16:29:26 +0000 (09:29 -0700)]
rustdoc: remove font family CSS on `.rustdoc-toggle summary::before`

This rule became irrelevant since c58246efe47bea09d4f3e70f536e4c9bb7770749
made it so that the `summary::before` pseudo-element contains an SVG instead
of text.

21 months agoAdd SAFETY comments for AsFd implementations on stdin/stdout/stderr
Josh Triplett [Mon, 3 Oct 2022 11:58:00 +0000 (12:58 +0100)]
Add SAFETY comments for AsFd implementations on stdin/stdout/stderr

21 months agoAdd stability attributes.
Dan Gohman [Thu, 22 Sep 2022 23:14:14 +0000 (16:14 -0700)]
Add stability attributes.

21 months agoAuto merge of #102610 - dawnofmidnight:git-commit-hash, r=Mark-Simulacrum
bors [Mon, 3 Oct 2022 14:20:32 +0000 (14:20 +0000)]
Auto merge of #102610 - dawnofmidnight:git-commit-hash, r=Mark-Simulacrum

re-add git-commit-hash file to tarballs

rust-lang/rust#100557 removed the `git-commit-hash` file and replaced it with `git-commit-info`. However, build-manifest relies on the `git-commit-hash` file being present, so this adds it back.

r? `@Mark-Simulacrum`

21 months agoMove utils from rustc_middle to rustc_ty_utils
Cameron Steffen [Mon, 3 Oct 2022 00:06:14 +0000 (19:06 -0500)]
Move utils from rustc_middle to rustc_ty_utils

21 months agoMigrate stab elements style to CSS variables
Guillaume Gomez [Mon, 3 Oct 2022 13:31:59 +0000 (15:31 +0200)]
Migrate stab elements style to CSS variables

21 months agoerrors: rename `typeck.ftl` to `hir_analysis.ftl`
David Wood [Wed, 28 Sep 2022 09:21:33 +0000 (10:21 +0100)]
errors: rename `typeck.ftl` to `hir_analysis.ftl`

In #102306, `rustc_typeck` was renamed to `rustc_hir_analysis` but the
diagnostic resources were not renamed - which is what this commit
changes.

Signed-off-by: David Wood <david.wood@huawei.com>
21 months ago`HirId` for `deferred_transmute_checks`
lcnr [Mon, 3 Oct 2022 11:49:31 +0000 (13:49 +0200)]
`HirId` for `deferred_transmute_checks`

21 months agoAuto merge of #102551 - bjorn3:cg_ssa_cleanup, r=davidtwco
bors [Mon, 3 Oct 2022 11:02:58 +0000 (11:02 +0000)]
Auto merge of #102551 - bjorn3:cg_ssa_cleanup, r=davidtwco

Some more cleanup for rustc_codegen_ssa

With the aim to make non-LLVM like backends, like Cranelift, easier to support using cg_ssa.

21 months agocheck if const is ADT or not
Takayuki Maeda [Mon, 3 Oct 2022 08:51:18 +0000 (17:51 +0900)]
check if const is ADT or not

21 months agoAuto merge of #102614 - matthiaskrgr:rollup-fjyu7oo, r=matthiaskrgr
bors [Mon, 3 Oct 2022 07:30:34 +0000 (07:30 +0000)]
Auto merge of #102614 - matthiaskrgr:rollup-fjyu7oo, r=matthiaskrgr

Rollup of 3 pull requests

Successful merges:

 - #102550 (rustdoc: remove no-op CSS on `.impl, .method` etc)
 - #102591 (Fix duplicate usage of `a` article.)
 - #102592 (Remove a couple lifetimes that can be infered)

Failed merges:

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

21 months agoCleanup some error code explanations
nils [Mon, 3 Oct 2022 06:06:07 +0000 (08:06 +0200)]
Cleanup some error code explanations

E0045: Use a stable non-C ABI instead
E0092: Use an atomic intrinsic that actually exists
E0161: Don't use box_syntax
E0579: Format ranges in the rustfmt style
E0622: Use the rustfmt style
E0743: Remove feature gate as it's not needed

21 months agoadd a ui test for #101739
Takayuki Maeda [Mon, 3 Oct 2022 06:02:38 +0000 (15:02 +0900)]
add a ui test for #101739

21 months agoRollup merge of #102592 - WaffleLapkin:less_lifetimes, r=cjgillot
Matthias Krüger [Mon, 3 Oct 2022 06:00:48 +0000 (08:00 +0200)]
Rollup merge of #102592 - WaffleLapkin:less_lifetimes, r=cjgillot

Remove a couple lifetimes that can be infered

From the review: https://github.com/rust-lang/rust/pull/101986#discussion_r974497497

r? `@cjgillot`

21 months agoRollup merge of #102591 - JarvisCraft:fix-double-a-article, r=compiler-errors
Matthias Krüger [Mon, 3 Oct 2022 06:00:47 +0000 (08:00 +0200)]
Rollup merge of #102591 - JarvisCraft:fix-double-a-article, r=compiler-errors

Fix duplicate usage of `a` article.

This fixes a typo first appearing in #94624 in which test-macro diagnostic uses "a" article twice.

Since I searched the sources for " a a " sequences, I also fixed the same issue in a few files where I found it.

21 months agoRollup merge of #102550 - notriddle:notriddle/impl, r=GuillaumeGomez
Matthias Krüger [Mon, 3 Oct 2022 06:00:47 +0000 (08:00 +0200)]
Rollup merge of #102550 - notriddle:notriddle/impl, r=GuillaumeGomez

rustdoc: remove no-op CSS on `.impl, .method` etc

Preview: http://notriddle.com/notriddle-rustdoc-demos/impl/index.html

# `flex-basis: 100%`

When `.impl-items { flex-basis: 100% }` and `h3.impl, h3.method, h4.method, h3.type, h4.type, h4.associatedconstant` were added in https://github.com/rust-lang/rust/commit/34bd2b845b3acd84c5a9bddae3ff8081c19ec5e9, it seems like it was a mistake even then. According to MDN, [flex-basis] does nothing unless the box it's applied to is a flex *item*, a child of a flex container. However, when this was added, these elements were flex containers themselves.

[flex-basis]: https://developer.mozilla.org/en-US/docs/Web/CSS/flex-basis

# `position: relative`

This property was added to help with positioning the `[+]/[-]` toggle. It is no longer necessary, because `details.rustdoc-toggle` already has `position:relative` set on it.

21 months agoreturn when obligation has references_error
Takayuki Maeda [Mon, 3 Oct 2022 05:17:42 +0000 (14:17 +0900)]
return when obligation has references_error

21 months agoAdd some more operator cases to `dump-debug-span-debug.rs`.
Nicholas Nethercote [Mon, 3 Oct 2022 03:20:03 +0000 (14:20 +1100)]
Add some more operator cases to `dump-debug-span-debug.rs`.

21 months agoAuto merge of #102508 - nnethercote:even-more-lexer-improvements, r=matklad
bors [Mon, 3 Oct 2022 04:49:46 +0000 (04:49 +0000)]
Auto merge of #102508 - nnethercote:even-more-lexer-improvements, r=matklad

Even more lexer improvements

These are just about code clarity, rather than performance.

r? `@matklad`

21 months agore-add git-commit-hash file to tarballs
dawnofmidnight [Mon, 3 Oct 2022 03:21:26 +0000 (23:21 -0400)]
re-add git-commit-hash file to tarballs

rust-lang/rust#100557 removed the `git-commit-hash` file and replaced it
with `git-commit-info`. However, build-manifest relies on the
`git-commit-hash` file being present, so this adds it back.

21 months agorustdoc: re-sugar more cross-crate trait bounds
León Orell Valerian Liehr [Wed, 28 Sep 2022 20:54:39 +0000 (22:54 +0200)]
rustdoc: re-sugar more cross-crate trait bounds

21 months agoAuto merge of #102503 - cuviper:x86-stack-probes, r=nagisa
bors [Mon, 3 Oct 2022 02:09:05 +0000 (02:09 +0000)]
Auto merge of #102503 - cuviper:x86-stack-probes, r=nagisa

Enable inline stack probes on X86 with LLVM 16

The known problems with x86 inline-asm stack probes have been solved on LLVM main (16), so this flips the switch. Anyone using bleeding-edge LLVM with rustc can start testing this, as I have done locally. We'll get more direct rust-ci when LLVM 16 branches and we start our upgrade, and we can always patch or disable it then if we find new problems.

The previous attempt was #77885, reverted in #84708.

21 months agoSync docs of `slice::{from_ptr_range, from_ptr_range_mut}`
Maybe Waffle [Mon, 3 Oct 2022 00:44:50 +0000 (00:44 +0000)]
Sync docs of `slice::{from_ptr_range, from_ptr_range_mut}`

21 months agoInvert `is_top_level` to avoid negation.
Nicholas Nethercote [Fri, 30 Sep 2022 21:57:22 +0000 (07:57 +1000)]
Invert `is_top_level` to avoid negation.

21 months agoRemove `TokenStreamBuilder`.
Nicholas Nethercote [Fri, 30 Sep 2022 06:51:35 +0000 (16:51 +1000)]
Remove `TokenStreamBuilder`.

It's now only used in one function. Also, the "should we glue the
tokens?" check is only necessary when pushing a `TokenTree::Token`, not
when pushing a `TokenTree::Delimited`.

As part of this, we now do the "should we glue the tokens?" check
immediately, which avoids having look back at the previous token. It
also puts all the logic dealing with token gluing in a single place.

21 months agoInline and remove `parse_token_tree_non_delim_non_eof`.
Nicholas Nethercote [Fri, 30 Sep 2022 06:50:02 +0000 (16:50 +1000)]
Inline and remove `parse_token_tree_non_delim_non_eof`.

It has a single call site.

21 months agoMerge `parse_token_trees_until_close_delim` and `parse_all_token_trees`.
Nicholas Nethercote [Fri, 30 Sep 2022 06:43:11 +0000 (16:43 +1000)]
Merge `parse_token_trees_until_close_delim` and `parse_all_token_trees`.

Because they're very similar, and this will allow some follow-up
changes.

21 months agoAdd comments to `TokenCursor::desugar`.
Nicholas Nethercote [Thu, 29 Sep 2022 21:39:54 +0000 (07:39 +1000)]
Add comments to `TokenCursor::desugar`.

It took me some time to work out what this code was doing.

21 months agoRewrite `Token::is_op`.
Nicholas Nethercote [Wed, 28 Sep 2022 01:29:47 +0000 (11:29 +1000)]
Rewrite `Token::is_op`.

An exhaustive match is more readable and more future-proof.

21 months agoAdd comments to `Spacing`.
Nicholas Nethercote [Wed, 28 Sep 2022 01:20:42 +0000 (11:20 +1000)]
Add comments to `Spacing`.

21 months agoDocument when `slice::from_ptr_range[_mut]` panic
Maybe Waffle [Mon, 3 Oct 2022 00:41:54 +0000 (00:41 +0000)]
Document when `slice::from_ptr_range[_mut]` panic

21 months agoClarify operator splitting.
Nicholas Nethercote [Wed, 28 Sep 2022 00:07:01 +0000 (10:07 +1000)]
Clarify operator splitting.

I found this code hard to read.

21 months agoAuto merge of #102305 - flba-eb:remove_exclude_list, r=Mark-Simulacrum
bors [Sun, 2 Oct 2022 23:47:48 +0000 (23:47 +0000)]
Auto merge of #102305 - flba-eb:remove_exclude_list, r=Mark-Simulacrum

Get rid of exclude-list for Windows-only tests

Main purpose of this change is to get rid of a quite long (and growing) list of excluded targets, while this test should only be useful on Windows (as far as I understand it). The `// only-windows` header seams to implement exactly what we need here.

I don't know why there are some whitespace changes, but `x.py fmt` and `.git/hooks/pre-push` are happy.

21 months agoAuto merge of #100557 - dawnofmidnight:tarball-commit-info, r=Mark-Simulacrum
bors [Sun, 2 Oct 2022 20:55:17 +0000 (20:55 +0000)]
Auto merge of #100557 - dawnofmidnight:tarball-commit-info, r=Mark-Simulacrum

fix: use git-commit-info for version information

Fixes #33286.
Fixes #86587.

This PR changes the current `git-commit-hash` file that `./x.py` dist puts in the `rustc-{version}-src.tar.{x,g}z` to contain the hash, the short hash, and the commit date from which the tarball was created, assuming git was available when it was. It uses this for reading the version so that rustc has all the appropriate metadata.

# Testing

Testing this is kind of a pain. I did it with something like
```sh
./x.py dist # ensure that `ignore-git` is `false` in config.toml
cp ./build/dist/rustc-1.65.0-dev-src.tar.gz ../rustc-1.65.0-dev-src.tar.gz
cd .. && tar -xzf rustc-1.65.0-dev-src && cd rustc-1.65.0-dev-src
./x.py build
```

Then, the output of  `rustc -vV` with the stage1 compiler should have the `commit-hash` and `commit-date` fields filled, rather than be `unknown`. To be completely sure, you can use `rustc --sysroot` with the stdlib that the original `./x.py dist` made, which will require that the metadata matches.

21 months agoAvoid ICE in printing RPITIT type
Michael Goulet [Sun, 2 Oct 2022 20:25:48 +0000 (20:25 +0000)]
Avoid ICE in printing RPITIT type

21 months agoAdd example to opaque_hidden_inferred_bound lint
Michael Goulet [Sun, 2 Oct 2022 19:07:31 +0000 (19:07 +0000)]
Add example to opaque_hidden_inferred_bound lint

21 months agoMake it a lint for all opaque types
Michael Goulet [Sun, 2 Oct 2022 05:45:15 +0000 (05:45 +0000)]
Make it a lint for all opaque types

21 months agoLint for unsatisfied nested opaques
Michael Goulet [Sun, 2 Oct 2022 04:45:54 +0000 (04:45 +0000)]
Lint for unsatisfied nested opaques

21 months agoDon't ICE when trying to copy unsized value in const prop
Michael Goulet [Sun, 2 Oct 2022 19:20:49 +0000 (19:20 +0000)]
Don't ICE when trying to copy unsized value in const prop

21 months agoFix duplicate usage of `a` article.
Petr Portnov [Sun, 2 Oct 2022 18:40:39 +0000 (21:40 +0300)]
Fix duplicate usage of `a` article.

This fixes a typo first appearing in #94624
in which test-macro diagnostic uses "a" article twice.

Since I searched sources for " a a " sequences,
I also fixed the same issue in a few source files where I found it.

Signed-off-by: Petr Portnov <gh@progrm-jarvis.ru>