]> git.lizzy.rs Git - rust.git/log
rust.git
4 years agoAuto merge of #64882 - ehuss:stabilize-bare-extern, r=eddyb
bors [Fri, 8 Nov 2019 01:15:50 +0000 (01:15 +0000)]
Auto merge of #64882 - ehuss:stabilize-bare-extern, r=eddyb

Stabilize --extern flag without a path.

This stabilizes the `--extern` flag without a path, implemented in #54116.

This flag is used to add a crate that may be found in the search path to the extern prelude. The intent of stabilizing this now is to change Cargo to emit this flag for `proc_macro` when building a proc-macro crate. This will allow the ability to elide `extern crate proc_macro;` for proc-macros, one of the few places where it is still necessary.

It is intended that Cargo may also use this flag for other cases in the future as part of the [std-aware work](https://github.com/rust-lang/wg-cargo-std-aware/). There will likely be some kind of syntax where users may declare dependencies on other crates (such as `alloc`), and Cargo will use this flag so that they may be used like any other crate. At this time there are no short-term plans to use it for anything other than proc-macro.

This will not help for non-proc-macro crates that use `proc_macro`, which I believe is not too common?

An alternate approach for proc-macro is to use the `meta` crate, but from my inquiries there doesn't appear to be anyone interested in pushing that forward. The `meta` crate also doesn't help with things like `alloc` or `test`.

cc #57288

4 years agoAuto merge of #66189 - Centril:rollup-3bsf45s, r=Centril
bors [Thu, 7 Nov 2019 22:02:41 +0000 (22:02 +0000)]
Auto merge of #66189 - Centril:rollup-3bsf45s, r=Centril

Rollup of 5 pull requests

Successful merges:

 - #63793 (Have tidy ensure that we document all `unsafe` blocks in libcore)
 - #64696 ([rustdoc] add sub settings)
 - #65916 (syntax: move stuff around)
 - #66087 (Update some build-pass ui tests to use check-pass where applicable)
 - #66182 (invalid_value lint: fix help text)

Failed merges:

r? @ghost

4 years agoUpdate for unstable option refactoring.
Eric Huss [Wed, 16 Oct 2019 16:55:40 +0000 (09:55 -0700)]
Update for unstable option refactoring.

4 years agoRemove docs on --extern metadata precedence.
Eric Huss [Wed, 16 Oct 2019 14:37:53 +0000 (07:37 -0700)]
Remove docs on --extern metadata precedence.

4 years agoAdd test for --extern alloc=librustc.rlib
Eric Huss [Tue, 8 Oct 2019 16:45:24 +0000 (09:45 -0700)]
Add test for --extern alloc=librustc.rlib

4 years agoUpdate src/test/ui-fulldeps/pathless-extern-unstable.rs
Eric Huss [Mon, 7 Oct 2019 14:58:28 +0000 (07:58 -0700)]
Update src/test/ui-fulldeps/pathless-extern-unstable.rs

Add ERROR

Co-Authored-By: Mazdak Farrokhzad <twingoow@gmail.com>
4 years agoUpdate extern linking documentation.
Eric Huss [Mon, 30 Sep 2019 01:20:03 +0000 (18:20 -0700)]
Update extern linking documentation.

4 years agoUpdate built-in help for --extern.
Eric Huss [Mon, 30 Sep 2019 01:17:48 +0000 (18:17 -0700)]
Update built-in help for --extern.

4 years agoAdd more --extern tests.
Eric Huss [Mon, 30 Sep 2019 01:17:27 +0000 (18:17 -0700)]
Add more --extern tests.

4 years agoStabilize --extern flag without a path.
Eric Huss [Sat, 28 Sep 2019 19:04:06 +0000 (12:04 -0700)]
Stabilize --extern flag without a path.

4 years agoRollup merge of #66182 - RalfJung:invalid-value, r=Centril
Mazdak Farrokhzad [Thu, 7 Nov 2019 13:27:26 +0000 (14:27 +0100)]
Rollup merge of #66182 - RalfJung:invalid-value, r=Centril

invalid_value lint: fix help text

Now that we also warn about `MaybUninit::uninit().assume_init()`, just telling people "use `MaybeUninit`" isn't always sufficient. And anyway this seems like an important enough point to mention it here.

4 years agoRollup merge of #66087 - tmiasko:ui-mode, r=Centril
Mazdak Farrokhzad [Thu, 7 Nov 2019 13:27:24 +0000 (14:27 +0100)]
Rollup merge of #66087 - tmiasko:ui-mode, r=Centril

Update some build-pass ui tests to use check-pass where applicable

Helps with issue https://github.com/rust-lang/rust/issues/62277.

4 years agoRollup merge of #65916 - Centril:split-syntax-3, r=davidtwco
Mazdak Farrokhzad [Thu, 7 Nov 2019 13:27:23 +0000 (14:27 +0100)]
Rollup merge of #65916 - Centril:split-syntax-3, r=davidtwco

syntax: move stuff around

Part of https://github.com/rust-lang/rust/pull/65324.

r? @davidtwco
cc @estebank @petrochenkov

4 years agoRollup merge of #64696 - GuillaumeGomez:rustdoc-sub-settings, r=kinnison
Mazdak Farrokhzad [Thu, 7 Nov 2019 13:27:21 +0000 (14:27 +0100)]
Rollup merge of #64696 - GuillaumeGomez:rustdoc-sub-settings, r=kinnison

[rustdoc] add sub settings

This PR is to give a finer control over what types are automatically expanded or not as well as the possibility to add sub-settings in the settings page.

![Screenshot from 2019-09-23 00-46-14](https://user-images.githubusercontent.com/3050060/65395521-15aff300-dd9c-11e9-9437-429ca347d455.png)

r? @Mark-Simulacrum

4 years agoRollup merge of #63793 - oli-obk:🧹, r=dtolnay
Mazdak Farrokhzad [Thu, 7 Nov 2019 13:27:20 +0000 (14:27 +0100)]
Rollup merge of #63793 - oli-obk:🧹, r=dtolnay

Have tidy ensure that we document all `unsafe` blocks in libcore

cc @rust-lang/libs

I documented a few and added ignore flags on the other files. We can incrementally document the files, but won't regress any files this way.

4 years agomove syntax::{parse::literal -> util::literal}
Mazdak Farrokhzad [Fri, 11 Oct 2019 16:47:14 +0000 (18:47 +0200)]
move syntax::{parse::literal -> util::literal}

4 years agomove syntax::parse::lexer::comments -> syntax::util::comments
Mazdak Farrokhzad [Fri, 11 Oct 2019 12:39:52 +0000 (14:39 +0200)]
move syntax::parse::lexer::comments -> syntax::util::comments

4 years agosyntax::attr: remove usage of lexer
Mazdak Farrokhzad [Fri, 11 Oct 2019 12:14:30 +0000 (14:14 +0200)]
syntax::attr: remove usage of lexer

4 years agomove parse::classify -> util::classify
Mazdak Farrokhzad [Fri, 11 Oct 2019 11:17:20 +0000 (13:17 +0200)]
move parse::classify -> util::classify

4 years agomove PResult to librustc_errors
Mazdak Farrokhzad [Fri, 11 Oct 2019 11:06:36 +0000 (13:06 +0200)]
move PResult to librustc_errors

4 years agosyntax::parser::token -> syntax::token
Mazdak Farrokhzad [Fri, 11 Oct 2019 10:46:32 +0000 (12:46 +0200)]
syntax::parser::token -> syntax::token

4 years agomove unescape_error_reporting to lexer/
Mazdak Farrokhzad [Fri, 11 Oct 2019 10:32:18 +0000 (12:32 +0200)]
move unescape_error_reporting to lexer/

4 years agosyntax: simplify imports
Mazdak Farrokhzad [Fri, 11 Oct 2019 10:29:23 +0000 (12:29 +0200)]
syntax: simplify imports

4 years agomove parse/parser.rs -> parse/parser/mod.rs
Mazdak Farrokhzad [Fri, 25 Oct 2019 02:57:40 +0000 (04:57 +0200)]
move parse/parser.rs -> parse/parser/mod.rs

4 years agoinvalid_value lint: fix help text
Ralf Jung [Thu, 7 Nov 2019 09:22:50 +0000 (10:22 +0100)]
invalid_value lint: fix help text

4 years agoAuto merge of #66180 - Centril:rollup-c1ji943, r=Centril
bors [Thu, 7 Nov 2019 07:55:39 +0000 (07:55 +0000)]
Auto merge of #66180 - Centril:rollup-c1ji943, r=Centril

Rollup of 5 pull requests

Successful merges:

 - #59789 (Revert two unapproved changes to rustc_typeck.)
 - #65752 (Use structured suggestions for missing associated items)
 - #65884 (syntax: ABI-oblivious grammar)
 - #65974 (A scheme for more macro-matcher friendly pre-expansion gating)
 - #66017 (Add future incompatibility lint for `array.into_iter()`)

Failed merges:

 - #66056 (rustc_metadata: Some reorganization of the module structure)

r? @ghost

4 years agoRollup merge of #66017 - LukasKalbertodt:array-into-iter-lint, r=matthewjasper
Mazdak Farrokhzad [Thu, 7 Nov 2019 07:51:58 +0000 (08:51 +0100)]
Rollup merge of #66017 - LukasKalbertodt:array-into-iter-lint, r=matthewjasper

Add future incompatibility lint for `array.into_iter()`

This is for #65819. This lint warns when calling `into_iter` on an array directly. That's because today the method call resolves to `<&[T] as IntoIterator>::into_iter` but that would change when adding `IntoIterator` impls for arrays. This problem is discussed in detail in #65819.

We still haven't decided how to proceed exactly, but it seems like adding a lint is a good idea regardless?

Also: this is the first time I implement a lint, so there are probably a lot of things I can improve. I used a different strategy than @scottmcm describes [here](https://github.com/rust-lang/rust/pull/65819#issuecomment-548667847) since I already started implementing this before they commented.

### TODO

- [x] Decide if we want this lint -> apparently [we want](https://github.com/rust-lang/rust/pull/65819#issuecomment-548964818)
- [x] Open a lint-tracking-issue and add the correct issue number in the code -> https://github.com/rust-lang/rust/issues/66145

4 years agoRollup merge of #65974 - Centril:matcher-friendly-gating, r=petrochenkov
Mazdak Farrokhzad [Thu, 7 Nov 2019 07:51:57 +0000 (08:51 +0100)]
Rollup merge of #65974 - Centril:matcher-friendly-gating, r=petrochenkov

A scheme for more macro-matcher friendly pre-expansion gating

Pre-expansion gating will now avoid gating macro matchers that did not result in `Success(...)`. That is, the following is now OK despite `box 42` being a valid `expr` and that form being pre-expansion gated:

```rust
macro_rules! m {
    ($e:expr) => { 0 }; // This fails on the input below due to `, foo`.
    (box $e:expr, foo) => { 1 }; // Successful matcher, we should get `2`.
}

fn main() {
    assert_eq!(1, m!(box 42, foo));
}
```

Closes https://github.com/rust-lang/rust/issues/65846.

r? @petrochenkov
cc @Mark-Simulacrum

4 years agoRollup merge of #65884 - Centril:non-hardcoded-abis, r=petrochenkov
Mazdak Farrokhzad [Thu, 7 Nov 2019 07:51:55 +0000 (08:51 +0100)]
Rollup merge of #65884 - Centril:non-hardcoded-abis, r=petrochenkov

syntax: ABI-oblivious grammar

This PR has the following effects:

1. `extern $lit` is now legal where `$lit:literal` and `$lit` is substituted for a string literal.

2. `extern "abi_that_does_not_exist"` is now *syntactically* legal whereas before, the set of ABI strings was hard-coded into the grammar of the language. With this PR, the set of ABIs are instead validated and translated during lowering. That seems more appropriate.

3. `ast::FloatTy` is now distinct from `rustc_target::abi::FloatTy`. The former is used substantially more and the translation between them is only necessary in a single place.

4. As a result of 2-3, libsyntax no longer depends on librustc_target, which should improve pipe-lining somewhat.

cc @rust-lang/lang -- the points 1-2 slightly change the definition of the language but in a way which seems consistent with our general principles (in particular wrt. the discussions of turning things into semantic errors). I expect this to be uncontroversial but it's worth letting y'all know. :)

r? @varkor

4 years agoRollup merge of #65752 - estebank:sugg, r=Centril
Mazdak Farrokhzad [Thu, 7 Nov 2019 07:51:54 +0000 (08:51 +0100)]
Rollup merge of #65752 - estebank:sugg, r=Centril

Use structured suggestions for missing associated items

When encountering an `impl` that is missing associated items required by its `trait`, use structured suggestions at an appropriate place in the `impl`.

4 years agoRollup merge of #59789 - eddyb:typeck-reverts, r=nikomatsakis
Mazdak Farrokhzad [Thu, 7 Nov 2019 07:51:52 +0000 (08:51 +0100)]
Rollup merge of #59789 - eddyb:typeck-reverts, r=nikomatsakis

Revert two unapproved changes to rustc_typeck.

There was a breakdown in process (https://github.com/rust-lang/rust/pull/59004#issuecomment-477600735, https://github.com/rust-lang/rust/pull/58894#discussion_r272795560) and two changes were made to `rustc_typeck`'s "collect" queries, for rustdoc, that were neither needed *nor* correct.
I'm reverting them here, and will fix up rustdoc *somehow*, if necessary.

cc @rust-lang/compiler How do we ensure this doesn't happen again?

r? @nikomatsakis or @oli-obk

4 years agoUse -Zsave-analysis when compiling issue-37323.rs
Tomasz MiÄ…sko [Wed, 6 Nov 2019 00:00:00 +0000 (00:00 +0000)]
Use -Zsave-analysis when compiling issue-37323.rs

4 years agosyntax: use distinct FloatTy from rustc_target.
Mazdak Farrokhzad [Mon, 28 Oct 2019 02:46:22 +0000 (03:46 +0100)]
syntax: use distinct FloatTy from rustc_target.

We also sever syntax's dependency on rustc_target as a result.
This should slightly improve pipe-lining.

Moreover, some cleanup is done in related code.

4 years agotest that extern lits accept raw strings annd escapes
Mazdak Farrokhzad [Wed, 30 Oct 2019 12:53:53 +0000 (13:53 +0100)]
test that extern lits accept raw strings annd escapes

4 years agotest that parser is oblivious to set of ABIs
Mazdak Farrokhzad [Mon, 28 Oct 2019 03:13:38 +0000 (04:13 +0100)]
test that parser is oblivious to set of ABIs

4 years agoparser: allow ABIs from literal macro fragments
Mazdak Farrokhzad [Sun, 27 Oct 2019 23:29:23 +0000 (00:29 +0100)]
parser: allow ABIs from literal macro fragments

4 years agocleanup can_begin_const_arg
Mazdak Farrokhzad [Sun, 27 Oct 2019 22:22:59 +0000 (23:22 +0100)]
cleanup can_begin_const_arg

4 years agoparser: don't hardcode ABIs into grammar
Mazdak Farrokhzad [Sun, 27 Oct 2019 22:14:35 +0000 (23:14 +0100)]
parser: don't hardcode ABIs into grammar

4 years agoAuto merge of #66175 - JohnTitor:rollup-ihqk5vn, r=JohnTitor
bors [Thu, 7 Nov 2019 03:54:25 +0000 (03:54 +0000)]
Auto merge of #66175 - JohnTitor:rollup-ihqk5vn, r=JohnTitor

Rollup of 12 pull requests

Successful merges:

 - #65794 (gate rustc_on_unimplemented under rustc_attrs)
 - #65945 (Optimize long-linker-command-line test)
 - #66044 (Improve uninit/zeroed lint)
 - #66076 (HIR docs: mention how to resolve method paths)
 - #66084 (Do not require extra LLVM backends for `x.py test` to pass)
 - #66111 (improve from_raw_parts docs)
 - #66114 (Improve std::thread::Result documentation)
 - #66117 (Fixed PhantomData markers in Arc and Rc)
 - #66146 (Remove unused parameters in `__thread_local_inner`)
 - #66147 (Miri: Refactor to_scalar_ptr out of existence)
 - #66162 (Fix broken link in README)
 - #66171 (Update link on CONTRIBUTING.md)

Failed merges:

r? @ghost

4 years agoRollup merge of #66171 - JohnTitor:update-link, r=ehuss
Yuki Okushi [Thu, 7 Nov 2019 00:20:49 +0000 (09:20 +0900)]
Rollup merge of #66171 - JohnTitor:update-link, r=ehuss

Update link on CONTRIBUTING.md

This is a lost article of #66162
Follow-up for rust-lang/rustc-guide#491

CC: @mark-i-m
4 years agoRollup merge of #66162 - Darksonn:master, r=ehuss
Yuki Okushi [Thu, 7 Nov 2019 00:20:48 +0000 (09:20 +0900)]
Rollup merge of #66162 - Darksonn:master, r=ehuss

Fix broken link in README

The link to the rustc guide on building and running the compiler is broken. It was broken in [rustc-guide#491](https://github.com/rust-lang/rustc-guide/pull/491).

4 years agoRollup merge of #66147 - RalfJung:no-scalar-ptr, r=oli-obk
Yuki Okushi [Thu, 7 Nov 2019 00:20:46 +0000 (09:20 +0900)]
Rollup merge of #66147 - RalfJung:no-scalar-ptr, r=oli-obk

Miri: Refactor to_scalar_ptr out of existence

`to_scalar_ptr` is somewhat subtle as it just throws away the 2nd component of a `ScalarPair` if there is one -- without any check if this is truly a pointer or so. And indeed we used it wrong on two occasions!

So I fixed those two, and then refactored things such that everyone calls `ref_to_mplace` instead (which they did anyway, I just moved up the calls), which is the only place that should interpret a `ScalarPair` as a wide ptr -- and it checks the type first. Thus we can remove `to_scalar_ptr` and `to_meta`.

r? @oli-obk

4 years agoRollup merge of #66146 - 3442853561:patch-2, r=Mark-Simulacrum
Yuki Okushi [Thu, 7 Nov 2019 00:20:45 +0000 (09:20 +0900)]
Rollup merge of #66146 - 3442853561:patch-2, r=Mark-Simulacrum

Remove unused parameters in `__thread_local_inner`

Fixes #65993.

4 years agoRollup merge of #66117 - olegnn:fixed_linked_list_marker, r=RalfJung
Yuki Okushi [Thu, 7 Nov 2019 00:20:44 +0000 (09:20 +0900)]
Rollup merge of #66117 - olegnn:fixed_linked_list_marker, r=RalfJung

Fixed PhantomData markers in Arc and Rc

Include owned internal structs in `PhantomData` markers in `Arc` (`PhantomData<T>` => `PhantomData<ArcInner<T>>`) and `Rc` (`PhantomData<T>` => `PhantomData<RcBox<T>>`).

4 years agoRollup merge of #66114 - golddranks:improve_thread_docs, r=Centril
Yuki Okushi [Thu, 7 Nov 2019 00:20:42 +0000 (09:20 +0900)]
Rollup merge of #66114 - golddranks:improve_thread_docs, r=Centril

Improve std::thread::Result documentation

Thanks to @dtolnay for pointing out the different premise of the contents of the `Err` variant in `std::thread::Result` WRT normal error handling.

4 years agoRollup merge of #66111 - RalfJung:from_raw_parts, r=Centril
Yuki Okushi [Thu, 7 Nov 2019 00:20:41 +0000 (09:20 +0900)]
Rollup merge of #66111 - RalfJung:from_raw_parts, r=Centril

improve from_raw_parts docs

Triggered by https://github.com/rust-lang/rfcs/pull/2806. Hopefully this helps clarify that joining slices across allocations is not possible in Rust currently.

r? @Centril

4 years agoRollup merge of #66084 - petrochenkov:x86arm, r=alexcrichton
Yuki Okushi [Thu, 7 Nov 2019 00:20:39 +0000 (09:20 +0900)]
Rollup merge of #66084 - petrochenkov:x86arm, r=alexcrichton

Do not require extra LLVM backends for `x.py test` to pass

For long time our testing passed with a partially built LLVM
```
[llvm]
targets = "X86;ARM"
```
, a [recent PR](https://github.com/rust-lang/rust/pull/65809) changed that.

4 years agoRollup merge of #66076 - RalfJung:qpath, r=davidtwco,oli-obk
Yuki Okushi [Thu, 7 Nov 2019 00:20:38 +0000 (09:20 +0900)]
Rollup merge of #66076 - RalfJung:qpath, r=davidtwco,oli-obk

HIR docs: mention how to resolve method paths

4 years agoRollup merge of #66044 - RalfJung:uninit-lint, r=oli-obk
Yuki Okushi [Thu, 7 Nov 2019 00:20:36 +0000 (09:20 +0900)]
Rollup merge of #66044 - RalfJung:uninit-lint, r=oli-obk

Improve uninit/zeroed lint

* Also warn when creating a raw pointer with a NULL vtable.
* Also identify `MaybeUninit::uninit().assume_init()` and `MaybeUninit::zeroed().assume_init()` as dangerous.

4 years agoRollup merge of #65945 - tmiasko:long-linker-command-line, r=alexcrichton
Yuki Okushi [Thu, 7 Nov 2019 00:20:34 +0000 (09:20 +0900)]
Rollup merge of #65945 - tmiasko:long-linker-command-line, r=alexcrichton

Optimize long-linker-command-line test

Replace O(n^3) text matching with inexpensive hash set lookups.

On my machine this reduces the total runtime of complete
run-make-fulldeps suite from roughly 75 seconds to 45 seconds.

4 years agoRollup merge of #65794 - Centril:unimpl-internal, r=varkor
Yuki Okushi [Thu, 7 Nov 2019 00:20:33 +0000 (09:20 +0900)]
Rollup merge of #65794 - Centril:unimpl-internal, r=varkor

gate rustc_on_unimplemented under rustc_attrs

Move `rustc_on_implemented` from the `on_implemented` gate to `rustc_attrs` as it is internal.

Closes #29628

r? @varkor

4 years agoAuto merge of #65750 - nnethercote:cheaper-doc-comments, r=petrochenkov
bors [Thu, 7 Nov 2019 00:10:52 +0000 (00:10 +0000)]
Auto merge of #65750 - nnethercote:cheaper-doc-comments, r=petrochenkov

Cheaper doc comments

This PR implements the idea from #60935: represent doc comments more cheaply, rather than converting them into `#[doc="..."]` attribute form. Unlike #60936 (which is about coalescing doc comments to reduce their number), this approach does not have any backwards compatibility concerns, and it eliminates about 80-90% of the current cost of doc comments (as estimated using the numbers in #60930, which eliminated the cost of doc comments entirely by treating them as normal comments).

r? @petrochenkov

4 years agoFix other broken link
Yuki Okushi [Wed, 6 Nov 2019 23:30:44 +0000 (08:30 +0900)]
Fix other broken link

4 years agoUpdate link on CONTRIBUTING.md
Yuki Okushi [Wed, 6 Nov 2019 21:48:28 +0000 (06:48 +0900)]
Update link on CONTRIBUTING.md

4 years agoAuto merge of #65728 - ecstatic-morse:promotion-const-proj, r=eddyb
bors [Wed, 6 Nov 2019 18:12:57 +0000 (18:12 +0000)]
Auto merge of #65728 - ecstatic-morse:promotion-const-proj, r=eddyb

Fix promotion in a `const` when projections are present

Resolves #65727.

This marks the entire local as "needs promotion" when only a projection of that local appears in a promotable context. This should only affect promotion in a `const` or `static`, not in a `fn` or `const fn`, which is handled in `promote_consts.rs`.

r? @eddyb

4 years agoreview comments
Esteban Küber [Wed, 30 Oct 2019 19:55:38 +0000 (12:55 -0700)]
review comments

4 years agoSuggest missing item from `trait` in `impl`
Esteban Küber [Thu, 24 Oct 2019 05:45:15 +0000 (22:45 -0700)]
Suggest missing item from `trait` in `impl`

4 years agoTry with crate::error::Error
Pyry Kontio [Wed, 6 Nov 2019 16:45:30 +0000 (01:45 +0900)]
Try with crate::error::Error

4 years agoFix broken link in README
Alice Ryhl [Wed, 6 Nov 2019 15:41:24 +0000 (16:41 +0100)]
Fix broken link in README

4 years agoReplace `array.into_iter()` with `iter()` in `libtest/tests.rs`
Lukas Kalbertodt [Fri, 1 Nov 2019 12:43:44 +0000 (13:43 +0100)]
Replace `array.into_iter()` with `iter()` in `libtest/tests.rs`

4 years agoAdd UI test for `array.into_iter()` lint
Lukas Kalbertodt [Fri, 1 Nov 2019 11:04:31 +0000 (12:04 +0100)]
Add UI test for `array.into_iter()` lint

4 years agoAdd future incompatibility lint for `array.into_iter()`
Lukas Kalbertodt [Fri, 1 Nov 2019 11:04:18 +0000 (12:04 +0100)]
Add future incompatibility lint for `array.into_iter()`

As we might want to add `IntoIterator` impls for arrays in the future,
and since that introduces a breaking change, this lint warns and
suggests using `iter()` instead (which is shorter and more explicit).

4 years agoAuto merge of #65134 - davidtwco:issue-19834-improper-ctypes-in-extern-C-fn, r=rkruppe
bors [Wed, 6 Nov 2019 12:45:35 +0000 (12:45 +0000)]
Auto merge of #65134 - davidtwco:issue-19834-improper-ctypes-in-extern-C-fn, r=rkruppe

improper_ctypes: `extern "C"` fns

cc #19834. Fixes #65867.

This pull request implements the change [described in this comment](https://github.com/rust-lang/rust/issues/19834#issuecomment-466671572).

cc @rkruppe @varkor @shepmaster

4 years agoMake doc comments cheaper with `AttrKind`.
Nicholas Nethercote [Wed, 23 Oct 2019 19:33:12 +0000 (06:33 +1100)]
Make doc comments cheaper with `AttrKind`.

`AttrKind` is a new type with two variants, `Normal` and `DocComment`. It's a
big performance win (over 10% in some cases) because `DocComment` lets doc
comments (which are common) be represented very cheaply.

`Attribute` gets some new helper methods to ease the transition:
- `has_name()`: check if the attribute name matches a single `Symbol`; for
  `DocComment` variants it succeeds if the symbol is `sym::doc`.
- `is_doc_comment()`: check if it has a `DocComment` kind.
- `{get,unwrap}_normal_item()`: extract the item from a `Normal` variant;
  panic otherwise.

Fixes #60935.

4 years agoRemove unnecessary `Deref` impl for `Attribute`.
Nicholas Nethercote [Wed, 23 Oct 2019 19:40:35 +0000 (06:40 +1100)]
Remove unnecessary `Deref` impl for `Attribute`.

This kind of thing just makes the code harder to read.

4 years agorollback gating for failing macro matchers
Mazdak Farrokhzad [Wed, 30 Oct 2019 16:34:00 +0000 (17:34 +0100)]
rollback gating for failing macro matchers

4 years agorevamp pre-expansion gating infra
Mazdak Farrokhzad [Wed, 30 Oct 2019 15:38:16 +0000 (16:38 +0100)]
revamp pre-expansion gating infra

4 years agoFix markdown link
Ralf Jung [Wed, 6 Nov 2019 10:15:30 +0000 (11:15 +0100)]
Fix markdown link

Co-Authored-By: Oliver Scherer <github35764891676564198441@oli-obk.de>
4 years agoThe unsafety in `iter.rs` is already documented wonderfully
Oliver Scherer [Fri, 25 Oct 2019 16:35:51 +0000 (18:35 +0200)]
The unsafety in `iter.rs` is already documented wonderfully

4 years agoRebase fallout
Oliver Scherer [Fri, 25 Oct 2019 10:03:34 +0000 (12:03 +0200)]
Rebase fallout

4 years agoSilence a deprecation warning
Oliver Scherer [Sat, 19 Oct 2019 11:02:17 +0000 (13:02 +0200)]
Silence a deprecation warning

4 years agoHalloween... time to get rid of 👻
Oliver Scherer [Sat, 19 Oct 2019 11:00:23 +0000 (13:00 +0200)]
Halloween... time to get rid of ðŸ‘»

4 years agoHave tidy ensure that we document all `unsafe` blocks in libcore
Oliver Scherer [Wed, 21 Aug 2019 17:56:46 +0000 (19:56 +0200)]
Have tidy ensure that we document all `unsafe` blocks in libcore

4 years agoAuto merge of #65830 - Quantumplation:master, r=davidtwco,estebank
bors [Wed, 6 Nov 2019 09:35:27 +0000 (09:35 +0000)]
Auto merge of #65830 - Quantumplation:master, r=davidtwco,estebank

Use ident.span instead of def_span in dead-code pass

Hello! First time contributor! :)

This should fix #58729.

According to @estebank in the duplicate #63064, def_span scans forward on the line until it finds a {,
and if it can't find one, falls back to the span for the whole item. This
was apparently written before the identifier span was explicitly tracked on
each node.

This means that if an unused function signature spans multiple lines, the
entire function (potentially hundreds of lines) gets flagged as dead code.
This could, for example, cause IDEs to add error squiggly's to the whole
function.

By using the span from the ident instead, we narrow the scope of this in
most cases. In a wider sense, it's probably safe to use ident.span
instead of def_span in most locations throughout the whole code base,
but since this is my first contribution, I kept it small.

Some interesting points that came up while I was working on this:
- I reorganized the tests a bit to bring some of the dead code ones all
into the same location
- A few tests were for things unrelated to dead code (like the
path-lookahead for parens), so I added #![allow(dead_code)] and
cleaned up the stderr file to reduce noise in the future
- The same fix doesn't apply to const and static declarations. I tried
adding these cases to the match expression, but that created a much
wider change to tests and error messages, so I left it off until I
could get some code review to validate the approach.

4 years ago--bless
Ralf Jung [Wed, 6 Nov 2019 08:52:11 +0000 (09:52 +0100)]
--bless

4 years agoimprove a comment
Ralf Jung [Wed, 6 Nov 2019 07:45:50 +0000 (08:45 +0100)]
improve a comment

4 years agoremvoe to_scalar_ptr and use ref_to_mplace everywhere
Ralf Jung [Wed, 6 Nov 2019 07:44:15 +0000 (08:44 +0100)]
remvoe to_scalar_ptr and use ref_to_mplace everywhere

4 years agomiri cast: avoid unnecessary to_scalar_ptr
Ralf Jung [Tue, 5 Nov 2019 22:04:17 +0000 (23:04 +0100)]
miri cast: avoid unnecessary to_scalar_ptr

4 years agoUpdate local.rs
3442853561 [Wed, 6 Nov 2019 08:39:48 +0000 (16:39 +0800)]
Update local.rs

Removed parameters not used in the macro

4 years agoFix the Error linking.
Pyry Kontio [Wed, 6 Nov 2019 07:57:59 +0000 (16:57 +0900)]
Fix the Error linking.

4 years agogate rustc_on_unimplemented under rustc_attrs
Mazdak Farrokhzad [Fri, 25 Oct 2019 05:19:07 +0000 (07:19 +0200)]
gate rustc_on_unimplemented under rustc_attrs

4 years agoAuto merge of #66143 - Centril:rollup-qmzuex0, r=Centril
bors [Wed, 6 Nov 2019 06:14:03 +0000 (06:14 +0000)]
Auto merge of #66143 - Centril:rollup-qmzuex0, r=Centril

Rollup of 9 pull requests

Successful merges:

 - #65776 (Rename `LocalInternedString` and more)
 - #65973 (caller_location: point to macro invocation sites, like file!/line!, and use in core::panic!.)
 - #66015 (librustc_lexer: Refactor the module)
 - #66062 (Configure LLVM module PIC level)
 - #66086 (bump smallvec to 1.0)
 - #66092 (Use KERN_ARND syscall for random numbers on NetBSD, same as FreeBSD.)
 - #66103 (Add target thumbv7neon-unknown-linux-musleabihf)
 - #66133 (Update the bundled `wasi-libc` repository)
 - #66139 (use American spelling for `pluralize!`)

Failed merges:

r? @ghost

4 years agoRollup merge of #66139 - euclio:pluralize, r=nagisa
Mazdak Farrokhzad [Wed, 6 Nov 2019 06:03:14 +0000 (07:03 +0100)]
Rollup merge of #66139 - euclio:pluralize, r=nagisa

use American spelling for `pluralize!`

4 years agoRollup merge of #66133 - alexcrichton:update-wasi-libc, r=Mark-Simulacrum
Mazdak Farrokhzad [Wed, 6 Nov 2019 06:03:12 +0000 (07:03 +0100)]
Rollup merge of #66133 - alexcrichton:update-wasi-libc, r=Mark-Simulacrum

Update the bundled `wasi-libc` repository

This updates the libc that the `wasm32-wasi` target links against to the
latest revision, mostly just bringing in minor bug fixes and minor wasm
size improvements.

4 years agoRollup merge of #66103 - smaeul:patch/thumb-musl, r=nagisa
Mazdak Farrokhzad [Wed, 6 Nov 2019 06:03:11 +0000 (07:03 +0100)]
Rollup merge of #66103 - smaeul:patch/thumb-musl, r=nagisa

Add target thumbv7neon-unknown-linux-musleabihf

This is a copy of thumbv7neon-unknown-linux-gnueabihf with musl changes
merged from armv7-unknown-linux-musleabihf. This appears to have been
missed when adding the other ARMv7-A thumb targets.

4 years agoRollup merge of #66092 - niacat:master, r=nagisa
Mazdak Farrokhzad [Wed, 6 Nov 2019 06:03:09 +0000 (07:03 +0100)]
Rollup merge of #66092 - niacat:master, r=nagisa

Use KERN_ARND syscall for random numbers on NetBSD, same as FreeBSD.

This system call is present on all supported NetBSD versions and provides an endless stream of non-blocking random data from the kernel's ChaCha20-based CSPRNG. It doesn't require a file like `/dev/urandom` to be opened.

The system call is documented here (under kern.arandom):
https://netbsd.gw.com/cgi-bin/man-cgi?sysctl+7+NetBSD-7.0

And defined here:
https://nxr.netbsd.org/xref/src/sys/sys/sysctl.h#273

The semantics are the same as FreeBSD so reading 256 bytes per call is fine.

Similar change for getrandom crate: rust-random/getrandom#115

4 years agoRollup merge of #66086 - RalfJung:smallvec, r=nagisa
Mazdak Farrokhzad [Wed, 6 Nov 2019 06:03:08 +0000 (07:03 +0100)]
Rollup merge of #66086 - RalfJung:smallvec, r=nagisa

bump smallvec to 1.0

This includes https://github.com/servo/rust-smallvec/pull/162, fixing an unsoundness in smallvec.

See https://github.com/servo/rust-smallvec/pull/175 for the 1.0 release announcement.

Cc @mbrubeck @emilio

4 years agoRollup merge of #66062 - smaeul:patch/pic-level, r=estebank
Mazdak Farrokhzad [Wed, 6 Nov 2019 06:03:06 +0000 (07:03 +0100)]
Rollup merge of #66062 - smaeul:patch/pic-level, r=estebank

Configure LLVM module PIC level

As of LLVM 9, this is required for 32-bit PowerPC to properly generate
PLT references. Previously, only BigPIC was supported; now LLVM supports
both BigPIC and SmallPIC, and there is no default value provided.

4 years agoRollup merge of #66015 - popzxc:refactor-librustc_parser, r=matklad
Mazdak Farrokhzad [Wed, 6 Nov 2019 06:03:05 +0000 (07:03 +0100)]
Rollup merge of #66015 - popzxc:refactor-librustc_parser, r=matklad

librustc_lexer: Refactor the module

This PR introduces a refactoring of the `librustc_lexer` in order to improve readability.

All the changes performed are only cosmetic and do not introduce any changes the lexer logic or performance.

Newly introduced modules `literal`, `token` and `utils` are just copy-pasted from the `lib.rs` and do not contain even cosmetic changes (I decided to do so so it'll be easier to review changes looking only on diff).

r? @petrochenkov

cc @Centril @matklad

4 years agoRollup merge of #65973 - eddyb:caller-location-panic, r=petrochenkov
Mazdak Farrokhzad [Wed, 6 Nov 2019 06:03:03 +0000 (07:03 +0100)]
Rollup merge of #65973 - eddyb:caller-location-panic, r=petrochenkov

caller_location: point to macro invocation sites, like file!/line!, and use in core::panic!.

The main change here is to `core::panic!`, trying to fix this remaining regression: https://github.com/rust-lang/rust/pull/65927#issuecomment-547625147

However, in order for `caller_location` to be usable from macros the same way `file!()`/`line!()` are, it needs to have the same behavior (of extracting the macro invocation site `Span` and using that).

Arguably we would've had to do this at some point anyway, if we want to use `#[track_caller]` to replace the `file!()`/`line!()` uses from macros, but I'm not sure the RFC mentions this at all.

r? @petrochenkov cc @anp @nnethercote

4 years agoRollup merge of #65776 - nnethercote:rename-LocalInternedString-and-more, r=estebank
Mazdak Farrokhzad [Wed, 6 Nov 2019 06:03:01 +0000 (07:03 +0100)]
Rollup merge of #65776 - nnethercote:rename-LocalInternedString-and-more, r=estebank

Rename `LocalInternedString` and more

This PR renames `LocalInternedString` as `SymbolStr`, removes an unnecessary `impl` from it, improves comments, and cleans up some `SymbolStr` uses.

r? @estebank

4 years agoAddressed review comments.
Pyry Kontio [Wed, 6 Nov 2019 05:47:52 +0000 (14:47 +0900)]
Addressed review comments.

4 years agoAuto merge of #66141 - Centril:rollup-n2fcvp9, r=Centril
bors [Wed, 6 Nov 2019 02:29:21 +0000 (02:29 +0000)]
Auto merge of #66141 - Centril:rollup-n2fcvp9, r=Centril

Rollup of 11 pull requests

Successful merges:

 - #65892 (Remove `PartialEq` and `Eq` from the `SpecialDerives`.)
 - #66014 (Show type parameter name and definition in type mismatch error messages )
 - #66027 (Move has_panic_handler to query)
 - #66054 (syntax: Avoid span arithmetic for delimiter tokens)
 - #66068 (use silent emitter for rustdoc highlighting pass)
 - #66081 (let caller of check_ptr_access_align control the error message)
 - #66093 (Do not ICE with a precision flag in formatting str and no format arguments)
 - #66098 (Detect `::` -> `:` typo when involving turbofish)
 - #66101 (Tweak type mismatch caused by break on tail expr)
 - #66106 (Fix typo in explanation of `E0080`)
 - #66115 (rustc: remove "GlobalMetaData" dead code from hir::map::definitions.)

Failed merges:

r? @ghost

4 years agoRollup merge of #66115 - eddyb:global-meta-what, r=michaelwoerister
Mazdak Farrokhzad [Wed, 6 Nov 2019 02:28:20 +0000 (03:28 +0100)]
Rollup merge of #66115 - eddyb:global-meta-what, r=michaelwoerister

rustc: remove "GlobalMetaData" dead code from hir::map::definitions.

Spotted while refactoring uses of `DefIndex` and/or `LocalDefId`.

r? @michaelwoerister

4 years agoRollup merge of #66106 - JOE1994:master, r=GuillaumeGomez
Mazdak Farrokhzad [Wed, 6 Nov 2019 02:28:19 +0000 (03:28 +0100)]
Rollup merge of #66106 - JOE1994:master, r=GuillaumeGomez

Fix typo in explanation of `E0080`

Handling issue #66105 in Rust repo.
`evaluate an constant expression` to `evaluate a constant expression`

4 years agoRollup merge of #66101 - estebank:break-tail-e0308, r=Centril
Mazdak Farrokhzad [Wed, 6 Nov 2019 02:28:17 +0000 (03:28 +0100)]
Rollup merge of #66101 - estebank:break-tail-e0308, r=Centril

Tweak type mismatch caused by break on tail expr

When `break;` has a type mismatch because the `Destination` points at a tail
expression with an obligation flowing from a return type, point at the
return type.

Fix #39968.

4 years agoRollup merge of #66098 - estebank:path-asciption-typo, r=Centril
Mazdak Farrokhzad [Wed, 6 Nov 2019 02:28:15 +0000 (03:28 +0100)]
Rollup merge of #66098 - estebank:path-asciption-typo, r=Centril

Detect `::` -> `:` typo when involving turbofish

Fix #65569.

4 years agoRollup merge of #66093 - estebank:fmt-ice, r=Centril
Mazdak Farrokhzad [Wed, 6 Nov 2019 02:28:14 +0000 (03:28 +0100)]
Rollup merge of #66093 - estebank:fmt-ice, r=Centril

Do not ICE with a precision flag in formatting str and no format arguments

Fix #66065.

4 years agoRollup merge of #66081 - RalfJung:ptr-offset, r=zackmdavis
Mazdak Farrokhzad [Wed, 6 Nov 2019 02:28:12 +0000 (03:28 +0100)]
Rollup merge of #66081 - RalfJung:ptr-offset, r=zackmdavis

let caller of check_ptr_access_align control the error message

This is needed for https://github.com/rust-lang/miri/pull/1031

4 years agoRollup merge of #66068 - euclio:null-emitter, r=estebank
Mazdak Farrokhzad [Wed, 6 Nov 2019 02:28:11 +0000 (03:28 +0100)]
Rollup merge of #66068 - euclio:null-emitter, r=estebank

use silent emitter for rustdoc highlighting pass

Partially addresses #63284.