]> git.lizzy.rs Git - rust.git/log
rust.git
2 years agoRollup merge of #97216 - RalfJung:null-ptr-check, r=oli-obk
Yuki Okushi [Wed, 1 Jun 2022 14:36:46 +0000 (23:36 +0900)]
Rollup merge of #97216 - RalfJung:null-ptr-check, r=oli-obk

Ensure we never consider the null pointer dereferencable

This replaces the checks that are being removed in https://github.com/rust-lang/rust/pull/97188. Those checks were too early and hence incorrect.

2 years agoRollup merge of #94647 - Urgau:hash-map-many-mut, r=Amanieu
Yuki Okushi [Wed, 1 Jun 2022 14:36:45 +0000 (23:36 +0900)]
Rollup merge of #94647 - Urgau:hash-map-many-mut, r=Amanieu

Expose `get_many_mut` and `get_many_unchecked_mut` to HashMap

This pull-request expose the function [`get_many_mut`](https://docs.rs/hashbrown/0.12.0/hashbrown/struct.HashMap.html#method.get_many_mut) and [`get_many_unchecked_mut`](https://docs.rs/hashbrown/0.12.0/hashbrown/struct.HashMap.html#method.get_many_unchecked_mut) from `hashbrown` to the standard library `HashMap` type. They obviously keep the same API and are added under the (new) `map_many_mut` feature.

- `get_many_mut`: Attempts to get mutable references to `N` values in the map at once.
- `get_many_unchecked_mut`: Attempts to get mutable references to `N` values in the map at once, without validating that the values are unique.

2 years agoupdate Miri
Ralf Jung [Wed, 1 Jun 2022 13:26:26 +0000 (09:26 -0400)]
update Miri

2 years agorename `sp` to `span`
Christian Poveda [Wed, 1 Jun 2022 12:14:33 +0000 (07:14 -0500)]
rename `sp` to `span`

2 years agoUpdate sync.rs
Dylan DPC [Wed, 1 Jun 2022 12:12:36 +0000 (14:12 +0200)]
Update sync.rs

2 years agoUpdate sync.rs
Dylan DPC [Wed, 1 Jun 2022 11:47:39 +0000 (13:47 +0200)]
Update sync.rs

2 years agoAdd regression test for #71546
Yuki Okushi [Tue, 31 May 2022 11:35:41 +0000 (20:35 +0900)]
Add regression test for #71546

2 years agoHide anonymous reexports in generated documentation
Guillaume Gomez [Wed, 1 Jun 2022 10:24:39 +0000 (12:24 +0200)]
Hide anonymous reexports in generated documentation

2 years agoremove an unnecessary `Option`
Takayuki Maeda [Wed, 1 Jun 2022 08:06:07 +0000 (17:06 +0900)]
remove an unnecessary `Option`

2 years agoUpdate cargo
Eric Huss [Wed, 1 Jun 2022 05:50:43 +0000 (22:50 -0700)]
Update cargo

2 years agoTweak insert docs
Ariel Davis [Wed, 1 Jun 2022 05:08:14 +0000 (22:08 -0700)]
Tweak insert docs

2 years agoAuto merge of #97553 - nbdd0121:lib, r=Mark-Simulacrum
bors [Wed, 1 Jun 2022 04:52:11 +0000 (04:52 +0000)]
Auto merge of #97553 - nbdd0121:lib, r=Mark-Simulacrum

Add `#[inline]` to `Vec`'s `Deref/DerefMut`

This should help #97552 (although I haven't verified).

2 years agouse 128 cache align for m1 mac
yifei [Thu, 26 May 2022 06:25:25 +0000 (14:25 +0800)]
use 128 cache align for m1 mac

2 years agoAuto merge of #97435 - Patryk27:bump-compiler-builtins, r=Dylan-DPC
bors [Wed, 1 Jun 2022 01:49:04 +0000 (01:49 +0000)]
Auto merge of #97435 - Patryk27:bump-compiler-builtins, r=Dylan-DPC

library/std: Bump compiler_builtins

Some neat changes include faster float conversions & fixes for AVR 🙂

(note that's it's my first time upgrading `compiler_builtins`, so I'm not 100% sure if bumping `library/std/Cargo.toml` is enough; certainly seems to be so, though.)

2 years agoRemove #[rustc_box] attr during lowering
est31 [Wed, 25 May 2022 00:30:55 +0000 (02:30 +0200)]
Remove #[rustc_box] attr during lowering

2 years agoLazify `SourceFile::lines`.
Nicholas Nethercote [Mon, 30 May 2022 05:59:45 +0000 (15:59 +1000)]
Lazify `SourceFile::lines`.

`SourceFile::lines` is a big part of metadata. It's stored in a compressed form
(a difference list) to save disk space. Decoding it is a big fraction of
compile time for very small crates/programs.

This commit introduces a new type `SourceFileLines` which has a `Lines`
form and a `Diffs` form. The latter is used when the metadata is first
read, and it is only decoded into the `Lines` form when line data is
actually needed. This avoids the decoding cost for many files,
especially in `std`. It's a performance win of up to 15% for tiny
crates/programs where metadata decoding is a high part of compilation
costs.

A `Lock` is needed because the methods that access lines data (which can
trigger decoding) take `&self` rather than `&mut self`. To allow for this,
`SourceFile::lines` now takes a `FnMut` that operates on the lines slice rather
than returning the lines slice.

2 years agoUse #[rustc_box] in alloc instead of box syntax
est31 [Sat, 28 May 2022 14:37:52 +0000 (16:37 +0200)]
Use #[rustc_box] in alloc instead of box syntax

2 years agoAdd #[rustc_box]
est31 [Sun, 22 May 2022 21:10:27 +0000 (23:10 +0200)]
Add #[rustc_box]

This commit adds an alternative content boxing syntax,
and uses it inside alloc.

The usage inside the very performance relevant code in
liballoc is the only remaining relevant usage of box syntax
in the compiler (outside of tests, which are comparatively
easy to port).

box syntax was originally designed to be used by all Rust
developers. This introduces a replacement syntax more tailored
to only being used inside the Rust compiler, and with it,
lays the groundwork for eventually removing box syntax.

2 years agoAuto merge of #97600 - matthiaskrgr:rollup-yivyeu5, r=matthiaskrgr
bors [Tue, 31 May 2022 23:24:39 +0000 (23:24 +0000)]
Auto merge of #97600 - matthiaskrgr:rollup-yivyeu5, r=matthiaskrgr

Rollup of 4 pull requests

Successful merges:

 - #97316 (Put a bound on collection misbehavior)
 - #97578 (alloc: remove repeated word in comment)
 - #97593 (:arrow_up: rust-analyzer)
 - #97596 (Fixup feature name to be more consistent with others)

Failed merges:

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

2 years agoSpelling correction.
Ximon Eighteen [Tue, 31 May 2022 22:23:23 +0000 (00:23 +0200)]
Spelling correction.

2 years agoExpose get_many_mut and get_many_unchecked_mut to HashMap
Loïc BRANSTETT [Sat, 5 Mar 2022 17:32:33 +0000 (18:32 +0100)]
Expose get_many_mut and get_many_unchecked_mut to HashMap

2 years agomigrate `check_for_for_in_in_typo` diagnostic
Christian Poveda [Tue, 31 May 2022 21:28:05 +0000 (16:28 -0500)]
migrate `check_for_for_in_in_typo` diagnostic

2 years agoRollup merge of #97596 - WaffleLapkin:fixup_feature_name, r=compiler-errors
Matthias Krüger [Tue, 31 May 2022 21:11:37 +0000 (23:11 +0200)]
Rollup merge of #97596 - WaffleLapkin:fixup_feature_name, r=compiler-errors

Fixup feature name to be more consistent with others

`slice_from_mut_ptr_range_const` -> `const_slice_from_mut_ptr_range`, we usually have `const` in the front.

I've made a typo in  #97419

2 years agoRollup merge of #97593 - lnicola:rust-analyzer-2022-05-31, r=lnicola
Matthias Krüger [Tue, 31 May 2022 21:11:36 +0000 (23:11 +0200)]
Rollup merge of #97593 - lnicola:rust-analyzer-2022-05-31, r=lnicola

:arrow_up: rust-analyzer

r? `@ghost`

2 years agoRollup merge of #97578 - ojeda:checkpatch, r=JohnTitor
Matthias Krüger [Tue, 31 May 2022 21:11:35 +0000 (23:11 +0200)]
Rollup merge of #97578 - ojeda:checkpatch, r=JohnTitor

alloc: remove repeated word in comment

Linux's `checkpatch.pl` reports:

```txt
#42544: FILE: rust/alloc/vec/mod.rs:2692:
WARNING: Possible repeated word: 'to'
+            // - Elements are :Copy so it's OK to to copy them, without doing
```

Signed-off-by: Miguel Ojeda <ojeda@kernel.org>
2 years agoRollup merge of #97316 - CAD97:bound-misbehavior, r=dtolnay
Matthias Krüger [Tue, 31 May 2022 21:11:34 +0000 (23:11 +0200)]
Rollup merge of #97316 - CAD97:bound-misbehavior, r=dtolnay

Put a bound on collection misbehavior

As currently written, when a logic error occurs in a collection's trait parameters, this allows *completely arbitrary* misbehavior, so long as it does not cause undefined behavior in std. However, because the extent of misbehavior is not specified, it is allowed for *any* code in std to start misbehaving in arbitrary ways which are not formally UB; consider the theoretical example of a global which gets set on an observed logic error. Because the misbehavior is only bound by not resulting in UB from safe APIs and the crate-level encapsulation boundary of all of std, this makes writing user unsafe code that utilizes std theoretically impossible, as it now relies on undocumented QOI (quality of implementation) that unrelated parts of std cannot be caused to misbehave by a misuse of std::collections APIs.

In practice, this is a nonconcern, because std has reasonable QOI and an implementation that takes advantage of this freedom is essentially a malicious implementation and only compliant by the most langauage-lawyer reading of the documentation.

To close this hole, we just add a small clause to the existing logic error paragraph that ensures that any misbehavior is limited to the collection which observed the logic error, making it more plausible to prove the soundness of user unsafe code.

This is not meant to be formal; a formal refinement would likely need to mention that values derived from the collection can also misbehave after a logic error is observed, as well as define what it means to "observe" a logic error in the first place. This fix errs on the side of informality in order to close the hole without complicating a normal reading which can assume a reasonable nonmalicious QOI.

See also [discussion on IRLO][1].

[1]: https://internals.rust-lang.org/t/using-std-collections-and-unsafe-anything-can-happen/16640

r? rust-lang/libs-api ```@rustbot``` label +T-libs-api -T-libs

This technically adds a new guarantee to the documentation, though I argue as written it's one already implicitly provided.

2 years agoAuto merge of #97582 - tmiasko:pointer-address-cast, r=oli-obk
bors [Tue, 31 May 2022 21:06:11 +0000 (21:06 +0000)]
Auto merge of #97582 - tmiasko:pointer-address-cast, r=oli-obk

Add a pointer to address cast kind

A pointer to address cast are often special-cased. Introduce a dedicated cast kind to make them easy distinguishable.

2 years agoFix lower_generics rustdocs
Santiago Pastorino [Tue, 31 May 2022 20:40:50 +0000 (17:40 -0300)]
Fix lower_generics rustdocs

2 years agomerge diagnostics about incorrect uses of `.await`
Christian Poveda [Tue, 31 May 2022 20:07:44 +0000 (15:07 -0500)]
merge diagnostics about incorrect uses of `.await`

2 years agoInline lower_generics_mut and remove GenericsCtor
Santiago Pastorino [Tue, 31 May 2022 19:57:16 +0000 (16:57 -0300)]
Inline lower_generics_mut and remove GenericsCtor

2 years agoRemove ImplTraitContext::reborrow
Santiago Pastorino [Tue, 31 May 2022 19:46:15 +0000 (16:46 -0300)]
Remove ImplTraitContext::reborrow

2 years agoderive Copy, Clone, PartialEq, Eq for ImplTraitContext
Santiago Pastorino [Tue, 31 May 2022 19:47:40 +0000 (16:47 -0300)]
derive Copy, Clone, PartialEq, Eq for ImplTraitContext

2 years agomigrate `error_on_incorrect_await` diagnostic
Christian Poveda [Tue, 31 May 2022 19:32:07 +0000 (14:32 -0500)]
migrate `error_on_incorrect_await` diagnostic

2 years agoFixup feature name to be more consistent with others
Maybe Waffle [Tue, 31 May 2022 19:12:28 +0000 (23:12 +0400)]
Fixup feature name to be more consistent with others

`slice_from_mut_ptr_range_const` -> `const_slice_from_mut_ptr_range`,
we usually have `const` in the front.

2 years agoMove lower_generics definition to item.rs
Santiago Pastorino [Tue, 31 May 2022 18:33:51 +0000 (15:33 -0300)]
Move lower_generics definition to item.rs

2 years ago:arrow_up: rust-analyzer
Laurențiu Nicola [Tue, 31 May 2022 18:32:15 +0000 (21:32 +0300)]
:arrow_up: rust-analyzer

2 years agoRename add_implicit_generics to lower_generics
Santiago Pastorino [Tue, 31 May 2022 18:32:05 +0000 (15:32 -0300)]
Rename add_implicit_generics to lower_generics

2 years agoReplace calls to lower_generics with calls to add_implicit_generics
Santiago Pastorino [Tue, 31 May 2022 18:30:14 +0000 (15:30 -0300)]
Replace calls to lower_generics with calls to add_implicit_generics

2 years agorustdoc: also index raw pointers
Michael Howell [Tue, 31 May 2022 18:18:51 +0000 (11:18 -0700)]
rustdoc: also index raw pointers

Co-authored-by: Noah Lev <camelidcamel@gmail.com>
2 years agorustdoc: also index impl trait
Michael Howell [Tue, 31 May 2022 18:12:02 +0000 (11:12 -0700)]
rustdoc: also index impl trait

2 years agoAuto merge of #97570 - JakobDegen:dse-test, r=tmiasko
bors [Tue, 31 May 2022 17:59:41 +0000 (17:59 +0000)]
Auto merge of #97570 - JakobDegen:dse-test, r=tmiasko

Fix TLS access mir opt test and remove stale files

Thanks `@pietroalbini` for noticing that the TLS test was not doing what it was supposed to. Switched to `PreCodegen` because `SimplifyCfg` does not run on opt level 0.

Also addresses the easy part of #97564 .

r? rust-lang/mir-opt

2 years agouse `suggestion_short` for incorrect semicolon diagnostic
Christian Poveda [Tue, 31 May 2022 17:56:26 +0000 (12:56 -0500)]
use `suggestion_short` for incorrect semicolon diagnostic

2 years agoAdd itctx: ImplTraitContext arg to add_implicit_generics
Santiago Pastorino [Tue, 31 May 2022 17:54:30 +0000 (14:54 -0300)]
Add itctx: ImplTraitContext arg to add_implicit_generics

2 years agoTake and restore fields in with_hir_id_owner calls
Santiago Pastorino [Tue, 31 May 2022 14:48:12 +0000 (11:48 -0300)]
Take and restore fields in with_hir_id_owner calls

2 years agomigrate `recover_from_await_method_call` diagnostic
Christian Poveda [Tue, 31 May 2022 17:33:35 +0000 (12:33 -0500)]
migrate `recover_from_await_method_call` diagnostic

2 years agoFix comment in poly_project_and_unify_type
Michael Goulet [Mon, 30 May 2022 21:36:12 +0000 (14:36 -0700)]
Fix comment in poly_project_and_unify_type

2 years agomigrate `maybe_consume_incorrect_semicolon` diagnostic
Christian Poveda [Tue, 31 May 2022 16:45:44 +0000 (11:45 -0500)]
migrate `maybe_consume_incorrect_semicolon` diagnostic

2 years agomigrate `maybe_recover_from_bad_qpath_stage_2` diagnostic
Christian Poveda [Tue, 31 May 2022 16:22:26 +0000 (11:22 -0500)]
migrate `maybe_recover_from_bad_qpath_stage_2` diagnostic

2 years agoMove defs and bounds from Universal to LoweringContext
Santiago Pastorino [Tue, 31 May 2022 14:31:57 +0000 (11:31 -0300)]
Move defs and bounds from Universal to LoweringContext

2 years agoApply track_caller to closure on `expect_non_local()`
Yuki Okushi [Tue, 10 May 2022 11:17:38 +0000 (20:17 +0900)]
Apply track_caller to closure on `expect_non_local()`

2 years agoAuto merge of #97419 - WaffleLapkin:const_from_ptr_range, r=oli-obk
bors [Tue, 31 May 2022 14:55:33 +0000 (14:55 +0000)]
Auto merge of #97419 - WaffleLapkin:const_from_ptr_range, r=oli-obk

Make `from{,_mut}_ptr_range` const

This PR makes the following APIs `const`:
```rust
// core::slice

pub const unsafe fn from_ptr_range<'a, T>(range: Range<*const T>) -> &'a [T];
pub const unsafe fn from_mut_ptr_range<'a, T>(range: Range<*mut T>) -> &'a mut [T];
```

Tracking issue: #89792.
Feature for `from_ptr_range` as a `const fn`: `slice_from_ptr_range_const`.
Feature for `from_mut_ptr_range` as a `const fn`: `slice_from_mut_ptr_range_const`.

r? `@oli-obk`

2 years agoAuto merge of #97521 - SkiFire13:clarify-vec-as-ptr, r=Dylan-DPC
bors [Tue, 31 May 2022 12:14:51 +0000 (12:14 +0000)]
Auto merge of #97521 - SkiFire13:clarify-vec-as-ptr, r=Dylan-DPC

Clarify the guarantees of Vec::as_ptr and Vec::as_mut_ptr when there's no allocation

Currently the documentation says they return a pointer to the vector's buffer, which has the implied precondition that the vector allocated some memory. However `Vec`'s documentation also specifies that it won't always allocate, so it's unclear whether the pointer returned is valid in that case. Of course you won't be able to read/write actual bytes to/from it since the capacity is 0, but there's an exception: zero sized read/writes. They are still valid as long as the pointer is not null and the memory it points to wasn't deallocated, but `Vec::as_ptr` and `Vec::as_mut_ptr` don't specify that's not the case. This PR thus specifies they are actually valid for zero sized reads since `Vec` is implemented to hold a dangling pointer in those cases, which is neither null nor was deallocated.

2 years agoalloc: remove repeated word in comment
Miguel Ojeda [Tue, 31 May 2022 10:28:57 +0000 (12:28 +0200)]
alloc: remove repeated word in comment

Linux's `checkpatch.pl` reports:

```txt
#42544: FILE: rust/alloc/vec/mod.rs:2692:
WARNING: Possible repeated word: 'to'
+            // - Elements are :Copy so it's OK to to copy them, without doing
```

Signed-off-by: Miguel Ojeda <ojeda@kernel.org>
2 years agoAuto merge of #97526 - Nilstrieb:unicode-is-printable-fastpath, r=joshtriplett
bors [Tue, 31 May 2022 09:34:00 +0000 (09:34 +0000)]
Auto merge of #97526 - Nilstrieb:unicode-is-printable-fastpath, r=joshtriplett

Add unicode fast path to `is_printable`

Before, it would enter the full expensive check even for normal ascii characters. Now, it skips the check for the ascii characters in `32..127`. This range was checked manually from the current behavior.

I ran the `tracing` test suite in miri, and it was really slow. I looked at a profile, and miri spent most of the time in `core::char::methods::escape_debug_ext`, where half of that was dominated by `core::unicode::printable::is_printable`. So I optimized it here.

The tracing profile:
![The tracing profile](https://user-images.githubusercontent.com/48135649/170883650-23876e7b-3fd1-4e8b-9001-47672e06d914.svg)

2 years agoAdd unicode fast path to `is_printable`
Nilstrieb [Sun, 29 May 2022 17:21:57 +0000 (19:21 +0200)]
Add unicode fast path to `is_printable`

Before, it would enter the full expensive check even for normal ascii
characters. Now, it skips the check for the ascii characters in
`32..127`. This range was checked manually from the current behavior.

2 years agoAuto merge of #97574 - Dylan-DPC:rollup-jq850l6, r=Dylan-DPC
bors [Tue, 31 May 2022 06:53:02 +0000 (06:53 +0000)]
Auto merge of #97574 - Dylan-DPC:rollup-jq850l6, r=Dylan-DPC

Rollup of 6 pull requests

Successful merges:

 - #97089 (Improve settings theme display)
 - #97229 (Document the current aliasing rules for `Box<T>`.)
 - #97371 (Suggest adding a semicolon to a closure without block)
 - #97455 (Stabilize `toowned_clone_into`)
 - #97565 (Add doc alias `memset` to `write_bytes`)
 - #97569 (Remove `memset` alias from `fill_with`.)

Failed merges:

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

2 years agoRollup merge of #97569 - thomcc:fill_with_isnt_memset, r=Amanieu
Dylan DPC [Tue, 31 May 2022 05:57:37 +0000 (07:57 +0200)]
Rollup merge of #97569 - thomcc:fill_with_isnt_memset, r=Amanieu

Remove `memset` alias from `fill_with`.

In https://rust-lang.zulipchat.com/#narrow/stream/219381-t-libs/topic/Unsafe.20and.20Safe.20versions.20of.20APIs.20both.20getting.20the.20same.20alias/near/284413029 `@Amanieu` pointed out that we had this, which is not really right.

In our guidelines we say that we will "not add an alias for a function that's only somewhat similar or related", which applies here. Memset doesn't take a closure, not even conceptually.

2 years agoRollup merge of #97565 - lukas-code:patch-1, r=thomcc
Dylan DPC [Tue, 31 May 2022 05:57:36 +0000 (07:57 +0200)]
Rollup merge of #97565 - lukas-code:patch-1, r=thomcc

Add doc alias `memset` to `write_bytes`

I were looking for `memset` in rust, but the docs only pointed me to `slice::fill`.

With only the old aliases, one might write code like this, which is incorrect if the memory is uninitialized.
``` Rust
core::slice::from_raw_parts(ptr, len).fill(0)
```

2 years agoRollup merge of #97455 - JohnTitor:stabilize-toowned-clone-into, r=dtolnay
Dylan DPC [Tue, 31 May 2022 05:57:35 +0000 (07:57 +0200)]
Rollup merge of #97455 - JohnTitor:stabilize-toowned-clone-into, r=dtolnay

Stabilize `toowned_clone_into`

Closes #41263
FCP has been done: https://github.com/rust-lang/rust/issues/41263#issuecomment-1100760750

2 years agoRollup merge of #97371 - ChayimFriedman2:closure-non-block-add-semicolon, r=oli-obk
Dylan DPC [Tue, 31 May 2022 05:57:34 +0000 (07:57 +0200)]
Rollup merge of #97371 - ChayimFriedman2:closure-non-block-add-semicolon, r=oli-obk

Suggest adding a semicolon to a closure without block

This transforms `|| expr` into `|| { expr; }`.

Closes #97359.

2 years agoRollup merge of #97229 - Nilstrieb:doc-box-noalias, r=dtolnay
Dylan DPC [Tue, 31 May 2022 05:57:33 +0000 (07:57 +0200)]
Rollup merge of #97229 - Nilstrieb:doc-box-noalias, r=dtolnay

Document the current aliasing rules for `Box<T>`.

Currently, `Box<T>` gets `noalias`, meaning it has the same rules as `&mut T`. This is sparsely documented, even though it can have quite a big impact on unsafe code using box. Therefore, these rules are documented here, with a big warning that they are not normative and subject to change, since we have not yet committed to an aliasing model and the state of `Box<T>` is especially uncertain.

If you have any suggestions and improvements, make sure to leave them here. This is mostly intended to inform people about what is currently going on (to prevent misunderstandings such as [Jon Gjengset's Box aliasing](https://www.youtube.com/watch?v=EY7Wi9fV5bk)).

This is supposed to _only document current UB_ and not add any new guarantees or rules.

2 years agoRollup merge of #97089 - GuillaumeGomez:improve-settings-theme-display, r=jsha
Dylan DPC [Tue, 31 May 2022 05:57:32 +0000 (07:57 +0200)]
Rollup merge of #97089 - GuillaumeGomez:improve-settings-theme-display, r=jsha

Improve settings theme display

This is a follow-up of #96958. In this PR, I changed how the theme radio buttons are displayed and improved their look as well.

It now looks like this:

![Screenshot from 2022-05-17 20-46-20](https://user-images.githubusercontent.com/3050060/168887703-a01e3bd5-9644-4012-ac11-2ae7bacd6be6.png)
![Screenshot from 2022-05-17 20-46-12](https://user-images.githubusercontent.com/3050060/168887707-132f8b2d-1163-462f-b7dd-f861121bdee7.png)

You can test it [here](https://rustdoc.crud.net/imperio/improve-settings-theme-display/doc/foo/index.html).

r? `@jsha`

2 years agoAuto merge of #96881 - est31:join_osstr, r=dtolnay
bors [Tue, 31 May 2022 04:28:29 +0000 (04:28 +0000)]
Auto merge of #96881 - est31:join_osstr, r=dtolnay

Implement [OsStr]::join

Implements join for `OsStr` and `OsString` slices:

```Rust
    let strings = [OsStr::new("hello"), OsStr::new("dear"), OsStr::new("world")];
    assert_eq!("hello dear world", strings.join(OsStr::new(" ")));
````

This saves one from converting to strings and back, or from implementing it manually.

This PR has been re-filed after #96744 was first accidentally merged and then reverted.

The change is instantly stable and thus:

r? rust-lang/libs-api `@rustbot` label +T-libs-api -T-libs

cc `@thomcc` `@m-ou-se` `@faptc`

2 years agoAuto merge of #97566 - compiler-errors:rollup-qfxw4j8, r=compiler-errors
bors [Tue, 31 May 2022 01:17:00 +0000 (01:17 +0000)]
Auto merge of #97566 - compiler-errors:rollup-qfxw4j8, r=compiler-errors

Rollup of 6 pull requests

Successful merges:

 - #89685 (refactor: VecDeques Iter fields to private)
 - #97172 (Optimize the diagnostic generation for `extern unsafe`)
 - #97395 (Miri call ABI check: ensure type size+align stay the same)
 - #97431 (don't do `Sized` and other return type checks on RPIT's real type)
 - #97555 (Source code page: line number click adds `NaN`)
 - #97558 (Fix typos in comment)

Failed merges:

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

2 years agoBTreeSet->BTreeMap (fix copy/paste mistake in documentation)
David Tolnay [Tue, 31 May 2022 00:56:35 +0000 (17:56 -0700)]
BTreeSet->BTreeMap (fix copy/paste mistake in documentation)

Co-authored-by: lcnr <rust@lcnr.de>
2 years agoAdd a pointer to address cast kind
Tomasz Miąsko [Tue, 31 May 2022 00:00:00 +0000 (00:00 +0000)]
Add a pointer to address cast kind

A pointer to address cast are often special-cased.
Introduce a dedicated cast kind to make them easy distinguishable.

2 years agoFix typo uniqeness -> uniqueness
David Tolnay [Mon, 30 May 2022 23:49:28 +0000 (16:49 -0700)]
Fix typo uniqeness -> uniqueness

2 years agonormalize harder
Maybe Waffle [Mon, 30 May 2022 23:46:04 +0000 (03:46 +0400)]
normalize harder

2 years agoFix TLS access mir opt test and remove stale files
Jakob Degen [Mon, 30 May 2022 23:26:02 +0000 (16:26 -0700)]
Fix TLS access mir opt test and remove stale files

2 years agoRemove `memset` alias from `fill_with`.
Thom Chiovoloni [Mon, 30 May 2022 23:26:00 +0000 (16:26 -0700)]
Remove `memset` alias from `fill_with`.

2 years agoRollup merge of #97558 - stoeckmann:typos, r=compiler-errors
Michael Goulet [Mon, 30 May 2022 22:57:31 +0000 (15:57 -0700)]
Rollup merge of #97558 - stoeckmann:typos, r=compiler-errors

Fix typos in comment

Just two typos in HashingControls' description.

2 years agoRollup merge of #97555 - GuillaumeGomez:line-number-click, r=notriddle
Michael Goulet [Mon, 30 May 2022 22:57:30 +0000 (15:57 -0700)]
Rollup merge of #97555 - GuillaumeGomez:line-number-click, r=notriddle

Source code page: line number click adds `NaN`

When you click on the parent element of the line numbers in the source code pages, it'll add `NaN` (like in https://doc.rust-lang.org/nightly/src/alloc/lib.rs.html#NaN). This PR fixes this bug.

cc ``@jsha``
r? ``@notriddle``

2 years agoRollup merge of #97431 - compiler-errors:issue-97413, r=oli-obk
Michael Goulet [Mon, 30 May 2022 22:57:29 +0000 (15:57 -0700)]
Rollup merge of #97431 - compiler-errors:issue-97413, r=oli-obk

don't do `Sized` and other return type checks on RPIT's real type

Fixes an ICE where we're doing `Sized` check against the RPIT's real type, instead of against the opaque type. This differs from what we're doing in MIR typeck, which causes ICE #97226.

This regressed in #96516 -- this adjusts that fix to be a bit more conservative. That PR was backported and thus the ICE is also present in stable. Not sure if it's worth to beta and/or stable backport, probably not the latter but I could believe the former.

r? `@oli-obk`

cc: another attempt to fix this ICE #97413. I believe this PR addresses the root cause.

2 years agoRollup merge of #97395 - RalfJung:call-abi, r=oli-obk
Michael Goulet [Mon, 30 May 2022 22:57:28 +0000 (15:57 -0700)]
Rollup merge of #97395 - RalfJung:call-abi, r=oli-obk

Miri call ABI check: ensure type size+align stay the same

We should almost certainly not accept calls where caller and callee disagree on the size or alignment of the type.

The checks we do *almost* imply that, except that `ScalarPair` types can have `repr(align)` and thus differ in size/align even when they are pairs of the same primitive type.

r? ``@oli-obk``

2 years agoRollup merge of #97172 - SparrowLii:unsafe_extern, r=compiler-errors
Michael Goulet [Mon, 30 May 2022 22:57:27 +0000 (15:57 -0700)]
Rollup merge of #97172 - SparrowLii:unsafe_extern, r=compiler-errors

Optimize the diagnostic generation for `extern unsafe`

This PR does the following about diagnostic generation when parsing foreign mod:
1. Fixes the FIXME about avoiding depending on the error message text.
2. Continue parsing when `unsafe` is followed by `{` (just like `unsafe extern {...}`).
3. Add test case.

2 years agoRollup merge of #89685 - DeveloperC286:iter_fields_to_private, r=oli-obk
Michael Goulet [Mon, 30 May 2022 22:57:27 +0000 (15:57 -0700)]
Rollup merge of #89685 - DeveloperC286:iter_fields_to_private, r=oli-obk

refactor: VecDeques Iter fields to private

Made the fields of VecDeque's Iter private by creating a Iter::new(...) function to create a new instance of Iter and migrating usage to use Iter::new(...).

2 years agoAuto merge of #97357 - davidtwco:diagnostic-translation-typed-subdiagnostic-simplific...
bors [Mon, 30 May 2022 22:49:36 +0000 (22:49 +0000)]
Auto merge of #97357 - davidtwco:diagnostic-translation-typed-subdiagnostic-simplification, r=oli-obk

errors: simplify referring to fluent attributes

To render the message of a Fluent attribute, the identifier of the Fluent message must be known. `DiagnosticMessage::FluentIdentifier` contains both the message's identifier and optionally the identifier of an attribute. Generated constants for each attribute would therefore need to be named uniquely (amongst all error messages) or be able to refer to only the attribute identifier which will be combined with a message identifier later. In this commit, the latter strategy is implemented as part of the `Diagnostic` type's functions for adding subdiagnostics of various kinds.

r? `@oli-obk`

2 years agoUpdate mut_ptr.rs
Lukas [Mon, 30 May 2022 22:40:34 +0000 (22:40 +0000)]
Update mut_ptr.rs

2 years agoUpdate intrinsics.rs
Lukas [Mon, 30 May 2022 22:38:29 +0000 (22:38 +0000)]
Update intrinsics.rs

2 years agoaddress comments
Michael Goulet [Mon, 30 May 2022 21:18:03 +0000 (14:18 -0700)]
address comments

2 years agobetter enum suggestions
Michael Goulet [Wed, 27 Apr 2022 02:59:50 +0000 (19:59 -0700)]
better enum suggestions

2 years agoCheck for `can_have_side_effects()` and `in_external_macro()` inside `suggest_missing...
Chayim Refael Friedman [Mon, 30 May 2022 20:35:51 +0000 (20:35 +0000)]
Check for `can_have_side_effects()` and `in_external_macro()` inside `suggest_missing_semicolon()`

2 years agoAuto merge of #97025 - ouz-a:mini-derefer-generator, r=davidtwco
bors [Mon, 30 May 2022 20:06:25 +0000 (20:06 +0000)]
Auto merge of #97025 - ouz-a:mini-derefer-generator, r=davidtwco

Add validation layer for Derefer

_Follow up work to #96549 #96116 #95857 #95649_

This adds validation for Derefer making sure it is always the first projection.

r? rust-lang/mir-opt

2 years agoFix typos in comment
Tobias Stoeckmann [Mon, 30 May 2022 19:21:32 +0000 (21:21 +0200)]
Fix typos in comment

2 years agoAuto merge of #97480 - conradludgate:faster-format-literals, r=joshtriplett
bors [Mon, 30 May 2022 17:39:58 +0000 (17:39 +0000)]
Auto merge of #97480 - conradludgate:faster-format-literals, r=joshtriplett

improve format impl for literals

The basic idea of this change can be seen here https://godbolt.org/z/MT37cWoe1.

Updates the format impl to have a fast path for string literals and the default path for regular format args.

This change will allow `format!("string literal")` to be used interchangably with `"string literal".to_owned()`.

This would be relevant in the case of `f!"string literal"` being legal (https://github.com/rust-lang/rfcs/pull/3267) in which case it would be the easiest way to create owned strings from literals, while also being just as efficient as any other impl

2 years agovalidate derefer, run derefer inside generator
ouz-a [Wed, 11 May 2022 21:27:06 +0000 (00:27 +0300)]
validate derefer, run derefer inside generator

2 years agoAdd line number click GUI test
Guillaume Gomez [Mon, 30 May 2022 15:14:46 +0000 (17:14 +0200)]
Add line number click GUI test

2 years agoImprove source-code-page.goml GUI test code
Guillaume Gomez [Mon, 30 May 2022 15:07:21 +0000 (17:07 +0200)]
Improve source-code-page.goml GUI test code

2 years agoAuto merge of #97548 - Dylan-DPC:rollup-9x0va1d, r=Dylan-DPC
bors [Mon, 30 May 2022 14:59:12 +0000 (14:59 +0000)]
Auto merge of #97548 - Dylan-DPC:rollup-9x0va1d, r=Dylan-DPC

Rollup of 6 pull requests

Successful merges:

 - #97494 (Use Box::new() instead of box syntax in library tests)
 - #97499 (Remove "sys isn't exported yet" phrase)
 - #97504 (Ensure source file present when calculating max line number)
 - #97519 (Re-add help_on_error for download-ci-llvm)
 - #97531 (Note pattern mismatch coming from `for` loop desugaring)
 - #97545 (Reword safety comments in core/hash/sip.rs)

Failed merges:

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

2 years agoFix invalid line number computation when clicking on something else than a line number
Guillaume Gomez [Mon, 30 May 2022 14:53:24 +0000 (16:53 +0200)]
Fix invalid line number computation when clicking on something else than a line number

2 years agonormalize forbidden slices
Maybe Waffle [Mon, 30 May 2022 14:27:07 +0000 (18:27 +0400)]
normalize forbidden slices

2 years agoAdd `#[inline]` to `Vec`'s `Deref/DerefMut`
Gary Guo [Mon, 30 May 2022 14:11:53 +0000 (15:11 +0100)]
Add `#[inline]` to `Vec`'s `Deref/DerefMut`

2 years agoerrors: simplify referring to fluent attributes
David Wood [Tue, 24 May 2022 14:09:47 +0000 (15:09 +0100)]
errors: simplify referring to fluent attributes

To render the message of a Fluent attribute, the identifier of the
Fluent message must be known. `DiagnosticMessage::FluentIdentifier`
contains both the message's identifier and optionally the identifier of
an attribute. Generated constants for each attribute would therefore
need to be named uniquely (amongst all error messages) or be able to
refer to only the attribute identifier which will be combined with a
message identifier later. In this commit, the latter strategy is
implemented as part of the `Diagnostic` type's functions for adding
subdiagnostics of various kinds.

Signed-off-by: David Wood <david.wood@huawei.com>
2 years agoRollup merge of #97545 - thomcc:sip-comment-safety, r=Dylan-DPC
Dylan DPC [Mon, 30 May 2022 12:33:53 +0000 (14:33 +0200)]
Rollup merge of #97545 - thomcc:sip-comment-safety, r=Dylan-DPC

Reword safety comments in core/hash/sip.rs

In https://rust-lang.zulipchat.com/#narrow/stream/136281-t-lang.2Fwg-unsafe-code-guidelines/topic/Is.20there.20any.20way.20to.20soundly.20do.20a.20masked.20out-of-bounds.20read.3F/near/284329248 it came up that this is using an atypical (and somewhat vague) phrasing of the safety requirement, so this slightly rewords it.

2 years agoRollup merge of #97531 - compiler-errors:for-loop-pat-mismatch, r=davidtwco
Dylan DPC [Mon, 30 May 2022 12:33:52 +0000 (14:33 +0200)]
Rollup merge of #97531 - compiler-errors:for-loop-pat-mismatch, r=davidtwco

Note pattern mismatch coming from `for` loop desugaring

Fixes #97163

2 years agoRollup merge of #97519 - binggh:readd-help-on-error, r=jyn514
Dylan DPC [Mon, 30 May 2022 12:33:51 +0000 (14:33 +0200)]
Rollup merge of #97519 - binggh:readd-help-on-error, r=jyn514

Re-add help_on_error for download-ci-llvm

Closes #97503

- Re-added `help_on_error` for `download_component()` and the downstream functions
- Removed dead code in `bootstrap.py`

Thanks `@jyn514` for the helpful tips!

(first contribution here, please let me know if I missed anything out!)

2 years agoRollup merge of #97504 - JohnTitor:cleanup-deps, r=davidtwco
Dylan DPC [Mon, 30 May 2022 12:33:50 +0000 (14:33 +0200)]
Rollup merge of #97504 - JohnTitor:cleanup-deps, r=davidtwco

Ensure source file present when calculating max line number

Resubmission of #89268, fixes #71363
The behavior difference of `simulate-remapped-rust-src-base` is not something we should take into account here, so limiting targets to run the test makes sense, I think.

r? `@davidtwco,` and `@estebank,` you might be interested in this change

2 years agoRollup merge of #97499 - est31:master, r=Dylan-DPC
Dylan DPC [Mon, 30 May 2022 12:33:49 +0000 (14:33 +0200)]
Rollup merge of #97499 - est31:master, r=Dylan-DPC

Remove "sys isn't exported yet" phrase

The oldest occurence is from 9e224c2bf18ebf8f871efb2e1aba43ed7970ebb7,
which is from the pre-1.0 days. In the years since then, std::sys still
hasn't been exported, and the last attempt was met with strong criticism:
https://github.com/rust-lang/rust/pull/97151

Thus, removing the "yet" part makes a lot of sense.

2 years agoRollup merge of #97494 - est31:remove_box_alloc_tests, r=Dylan-DPC
Dylan DPC [Mon, 30 May 2022 12:33:48 +0000 (14:33 +0200)]
Rollup merge of #97494 - est31:remove_box_alloc_tests, r=Dylan-DPC

Use Box::new() instead of box syntax in library tests

The tests inside `library/*` have no reason to use `box` syntax as they have 0 performance relevance. Therefore, we can safely remove them (instead of having to use alternatives like the one in #97293).

2 years agoAuto merge of #97546 - RalfJung:miri, r=oli-obk
bors [Mon, 30 May 2022 11:59:55 +0000 (11:59 +0000)]
Auto merge of #97546 - RalfJung:miri, r=oli-obk

update Miri

First update with the new ui test suite, let's hope this all works. :)
r? `@oli-obk`

Fixes https://github.com/rust-lang/rust/issues/97486