]> git.lizzy.rs Git - rust.git/log
rust.git
2 years agoAdd debug asserts in thread local cell set methods.
Mara Bos [Mon, 20 Dec 2021 18:20:56 +0000 (19:20 +0100)]
Add debug asserts in thread local cell set methods.

2 years agoAdd tracking issue number for local_key_cell_methods.
Mara Bos [Mon, 20 Dec 2021 12:49:10 +0000 (13:49 +0100)]
Add tracking issue number for local_key_cell_methods.

2 years agoRename LocalKey's with_{ref,mut} to with_borrow{,_mut}.
Mara Bos [Mon, 20 Dec 2021 12:38:07 +0000 (13:38 +0100)]
Rename LocalKey's with_{ref,mut} to with_borrow{,_mut}.

2 years agoImplement RFC 3184 - thread local cell methods.
Mara Bos [Tue, 19 Oct 2021 12:37:27 +0000 (14:37 +0200)]
Implement RFC 3184 - thread local cell methods.

2 years agoAuto merge of #94546 - JmPotato:std-features-cleanup, r=m-ou-se
bors [Sat, 5 Mar 2022 07:26:54 +0000 (07:26 +0000)]
Auto merge of #94546 - JmPotato:std-features-cleanup, r=m-ou-se

Clean up the std library's #![feature]s

Signed-off-by: JmPotato <ghzpotato@gmail.com>
This is part of https://github.com/rust-lang/rust/issues/87766.

r? `@m-ou-se`

2 years agoAuto merge of #94634 - Dylan-DPC:rollup-8wx1yrj, r=Dylan-DPC
bors [Sat, 5 Mar 2022 04:56:35 +0000 (04:56 +0000)]
Auto merge of #94634 - Dylan-DPC:rollup-8wx1yrj, r=Dylan-DPC

Rollup of 6 pull requests

Successful merges:

 - #94446 (UNIX `remove_dir_all()`: Try recursing first on the slow path)
 - #94460 (Reenable generator drop tracking tests and fix mutation handling)
 - #94620 (Edit docs on consistency of `PartialOrd` and `PartialEq`)
 - #94624 (Downgrade `#[test]` on macro call to warning)
 - #94626 (Add known-bug directive to issue #47511 test case)
 - #94631 (Fix typo in c-variadic)

Failed merges:

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

2 years agoRollup merge of #94631 - nebulatgs:patch-1, r=Dylan-DPC
Dylan DPC [Sat, 5 Mar 2022 03:46:41 +0000 (04:46 +0100)]
Rollup merge of #94631 - nebulatgs:patch-1, r=Dylan-DPC

Fix typo in c-variadic

Fixes a typo in the Unstable Book

2 years agoRollup merge of #94626 - marmeladema:issue-47511-known-bug, r=jackh726
Dylan DPC [Sat, 5 Mar 2022 03:46:40 +0000 (04:46 +0100)]
Rollup merge of #94626 - marmeladema:issue-47511-known-bug, r=jackh726

Add known-bug directive to issue #47511 test case

2 years agoRollup merge of #94624 - estebank:regression-94508, r=Dylan-DPC
Dylan DPC [Sat, 5 Mar 2022 03:46:39 +0000 (04:46 +0100)]
Rollup merge of #94624 - estebank:regression-94508, r=Dylan-DPC

Downgrade `#[test]` on macro call to warning

Follow up to #92959. Address #94508.

2 years agoRollup merge of #94620 - pierwill:partialord-constistency, r=yaahc
Dylan DPC [Sat, 5 Mar 2022 03:46:38 +0000 (04:46 +0100)]
Rollup merge of #94620 - pierwill:partialord-constistency, r=yaahc

Edit docs on consistency of `PartialOrd` and `PartialEq`

Use ordered list to make the information about implementations more readable.

2 years agoRollup merge of #94460 - eholk:reenable-drop-tracking-tests, r=tmiasko
Dylan DPC [Sat, 5 Mar 2022 03:46:37 +0000 (04:46 +0100)]
Rollup merge of #94460 - eholk:reenable-drop-tracking-tests, r=tmiasko

Reenable generator drop tracking tests and fix mutation handling

The previous PR, #94068, was overly zealous in counting mutations as borrows, which effectively nullified drop tracking. We would have caught this except the drop tracking tests were still ignored, despite having the option of using the `-Zdrop-tracking` flag now.

This PR fixes the issue introduced by #94068 by only counting mutations as borrows the mutated place has a project. This is sufficient to distinguish `x.y = 42` (which should count as a borrow of `x`) from `x = 42` (which is not a borrow of `x` because the whole variable is overwritten).

This PR also re-enables the drop tracking regression tests using the `-Zdrop-tracking` flag so we will avoid introducing these sorts of issues in the future.

Thanks to ``@tmiasko`` for noticing this problem and pointing it out!

r? ``@tmiasko``

2 years agoRollup merge of #94446 - rusticstuff:remove_dir_all-illumos-fix, r=cuviper
Dylan DPC [Sat, 5 Mar 2022 03:46:37 +0000 (04:46 +0100)]
Rollup merge of #94446 - rusticstuff:remove_dir_all-illumos-fix, r=cuviper

UNIX `remove_dir_all()`: Try recursing first on the slow path

This only affects the _slow_ code path - if there is no `dirent.d_type` or if it is `DT_UNKNOWN`.

POSIX specifies that calling `unlink()` or `unlinkat(..., 0)` on a directory is allowed to succeed:
> The _path_ argument shall not name a directory unless the process has appropriate privileges and the implementation supports using _unlink()_ on directories.

This however can cause dangling inodes requiring an fsck e.g. on Illumos UFS, so we have to avoid that in the common case. We now just try to recurse into it first and unlink() if we can't open it as a directory.

The other two commits integrate the Macos x86-64 implementation reducing redundancy. Split into two commits for better reviewing.

Fixes #94335.

2 years agoClean up the std library's #![feature]s
JmPotato [Thu, 3 Mar 2022 06:24:47 +0000 (14:24 +0800)]
Clean up the std library's #![feature]s

Signed-off-by: JmPotato <ghzpotato@gmail.com>
2 years agoAuto merge of #93142 - estebank:missing-main, r=wesleywiser
bors [Sat, 5 Mar 2022 02:46:24 +0000 (02:46 +0000)]
Auto merge of #93142 - estebank:missing-main, r=wesleywiser

Do not point at whole file missing `fn main`

Only point at the end of the crate. We could try making it point at the
beginning of the crate, but that is confused with `DUMMY_SP`, causing
the output to be *worse*.

This change will make it so that VSCode will *not* underline the whole
file when `main` is missing, so other errors will be visible.

2 years agoDo not point at whole file missing `fn main`
Esteban Kuber [Fri, 21 Jan 2022 00:15:39 +0000 (00:15 +0000)]
Do not point at whole file missing `fn main`

Only point at the end of the crate. We could try making it point at the
beginning of the crate, but that is confused with `DUMMY_SP`, causing
the output to be *worse*.

This change will make it so that VSCode will *not* underline the whole
file when `main` is missing, so other errors will be visible.

2 years agoFix typo in c-variadic
Nebula [Sat, 5 Mar 2022 02:34:38 +0000 (21:34 -0500)]
Fix typo in c-variadic

2 years agoAuto merge of #94628 - Dylan-DPC:rollup-v2slupe, r=Dylan-DPC
bors [Sat, 5 Mar 2022 00:15:54 +0000 (00:15 +0000)]
Auto merge of #94628 - Dylan-DPC:rollup-v2slupe, r=Dylan-DPC

Rollup of 5 pull requests

Successful merges:

 - #94362 (Add well known values to `--check-cfg` implementation)
 - #94577 (only disable SIMD for doctests in Miri (not for the stdlib build itself))
 - #94595 (Fix invalid `unresolved imports` errors for a single-segment import)
 - #94596 (Delay bug in expr adjustment when check_expr is called multiple times)
 - #94618 (Don't round stack size up for created threads in Windows)

Failed merges:

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

2 years agoRollup merge of #94618 - lewisclark:remove-stack-size-rounding, r=yaahc
Dylan DPC [Fri, 4 Mar 2022 21:58:37 +0000 (22:58 +0100)]
Rollup merge of #94618 - lewisclark:remove-stack-size-rounding, r=yaahc

Don't round stack size up for created threads in Windows

Fixes #94454

Windows does the rounding itself, so there isn't a need to explicity do the rounding beforehand, as mentioned by ```@ChrisDenton``` in #94454

> The operating system rounds up the specified size to the nearest multiple of the system's allocation granularity (typically 64 KB). To retrieve the allocation granularity of the current system, use the [GetSystemInfo](https://docs.microsoft.com/en-us/windows/win32/api/sysinfoapi/nf-sysinfoapi-getsysteminfo) function.

https://docs.microsoft.com/en-us/windows/win32/procthread/thread-stack-size

2 years agoRollup merge of #94596 - compiler-errors:delay-adjustment-duplicate, r=estebank
Dylan DPC [Fri, 4 Mar 2022 21:58:36 +0000 (22:58 +0100)]
Rollup merge of #94596 - compiler-errors:delay-adjustment-duplicate, r=estebank

Delay bug in expr adjustment when check_expr is called multiple times

Instead of including slightly more complicated logic in `check_argument_types` to fix the bug (#94516) I introduced in #94438, and inevitably have this bug appear once again when some other diagnostic is written that causes `check_expr` to be called an expression during a (bad) code path, just delay the bug in adjustment logic.

I am open to other implementations that don't delay the bug here.

Fixes #94516

2 years agoRollup merge of #94595 - TaKO8Ki:fix-invalid-unresolved-imports-errors-for-asterisk...
Dylan DPC [Fri, 4 Mar 2022 21:58:35 +0000 (22:58 +0100)]
Rollup merge of #94595 - TaKO8Ki:fix-invalid-unresolved-imports-errors-for-asterisk-wildcard-syntax, r=estebank

Fix invalid `unresolved imports` errors for a single-segment import

closes #90248

2 years agoRollup merge of #94577 - RalfJung:simd-miri, r=scottmcm
Dylan DPC [Fri, 4 Mar 2022 21:58:34 +0000 (22:58 +0100)]
Rollup merge of #94577 - RalfJung:simd-miri, r=scottmcm

only disable SIMD for doctests in Miri (not for the stdlib build itself)

Also we can enable library/core/tests/simd.rs now, Miri supports enough SIMD for that.

2 years agoRollup merge of #94362 - Urgau:check-cfg-values, r=petrochenkov
Dylan DPC [Fri, 4 Mar 2022 21:58:34 +0000 (22:58 +0100)]
Rollup merge of #94362 - Urgau:check-cfg-values, r=petrochenkov

Add well known values to `--check-cfg` implementation

This pull-request adds well known values for the well known names via `--check-cfg=values()`.

[RFC 3013: Checking conditional compilation at compile time](https://rust-lang.github.io/rfcs/3013-conditional-compilation-checking.html#checking-conditional-compilation-at-compile-time) doesn't define this at all, but this seems a nice improvement.
The activation is done by a empty `values()` (new syntax) similar to `names()` except that `names(foo)` also activate well known names while `values(aa, "aa", "kk")` would not.

As stated this use a different activation logic because well known values for the well known names are not always sufficient.
In fact this is problematic for every `target_*` cfg because of non builtin targets, as the current implementation use those built-ins targets to create the list the well known values.

The implementation is straight forward, first we gather (if necessary) all the values (lazily or not) and then we apply them.

r? ```@petrochenkov```

2 years agoAuto merge of #94298 - Urgau:rustbuild-check-cfg, r=Mark-Simulacrum
bors [Fri, 4 Mar 2022 21:52:34 +0000 (21:52 +0000)]
Auto merge of #94298 - Urgau:rustbuild-check-cfg, r=Mark-Simulacrum

Enable conditional compilation checking on the Rust codebase

This pull-request enable conditional compilation checking on every rust project build by the `bootstrap` tool.

To be more specific, this PR only enable well known names checking + extra names (bootstrap, parallel_compiler, ...).

r? `@Mark-Simulacrum`

2 years agoAdd known-bug directive to issue #47511 test case
marmeladema [Fri, 4 Mar 2022 21:47:56 +0000 (22:47 +0100)]
Add known-bug directive to issue #47511 test case

2 years agoDowngrade `#[test]` on macro call to warning
Esteban Kuber [Fri, 4 Mar 2022 20:34:10 +0000 (20:34 +0000)]
Downgrade `#[test]` on macro call to warning

Follow up to #92959. Address #94508.

2 years agoEdit docs on consistency of `PartialOrd` and `PartialEq`
pierwill [Fri, 4 Mar 2022 19:31:23 +0000 (13:31 -0600)]
Edit docs on consistency of `PartialOrd` and `PartialEq`

Use ordered list to make the information about implementations more readable.

2 years agoDistinguish binding assignments, use Ty::needs_drop
Eric Holk [Fri, 4 Mar 2022 19:03:24 +0000 (11:03 -0800)]
Distinguish binding assignments, use Ty::needs_drop

This better captures the actual behavior, rather than using hacks around
whether the assignment has any projections.

2 years agoAuto merge of #94494 - jonhoo:bump-autocfg, r=Mark-Simulacrum
bors [Fri, 4 Mar 2022 19:01:44 +0000 (19:01 +0000)]
Auto merge of #94494 - jonhoo:bump-autocfg, r=Mark-Simulacrum

Bump autocfg to 1.1.0

autocfg 1.1.0 makes it so that rustflags from the build are correctly
passed to the compiler probes, which in turn means those probes more
accurately reflect the outer build conditions. This is particularly
important if rustflags includes _required_ `-Clink-arg=` flags without
which builds will fail, as older versions of `autocfg` will then fail
the probe and erroneously report the probed feature as unavailable.

See also
https://github.com/rust-lang/rust/issues/94007#issuecomment-1040668261

2 years agoDon't round stack size up for created threads
Lewis Clark [Fri, 4 Mar 2022 18:04:43 +0000 (18:04 +0000)]
Don't round stack size up for created threads

2 years agoAuto merge of #94612 - matthiaskrgr:rollup-2jm5wkr, r=matthiaskrgr
bors [Fri, 4 Mar 2022 16:37:46 +0000 (16:37 +0000)]
Auto merge of #94612 - matthiaskrgr:rollup-2jm5wkr, r=matthiaskrgr

Rollup of 6 pull requests

r? `@ghost`

2 years agoRollup merge of #94600 - est31:master, r=notriddle
Matthias Krüger [Fri, 4 Mar 2022 16:31:09 +0000 (17:31 +0100)]
Rollup merge of #94600 - est31:master, r=notriddle

Use if let instead of manual match

Factored out of #94139 . `if let` is better here than both `let ... else` and `let ... = match`.

2 years agoRollup merge of #94593 - estebank:issue-94510, r=davidtwco
Matthias Krüger [Fri, 4 Mar 2022 16:31:08 +0000 (17:31 +0100)]
Rollup merge of #94593 - estebank:issue-94510, r=davidtwco

Do not recover from `Ty?` in macro parsing

Follow up to #92746. Address #94510.

2 years agoRollup merge of #94568 - bjorn3:rustbuild_remove_dead_code, r=Mark-Simulacrum
Matthias Krüger [Fri, 4 Mar 2022 16:31:07 +0000 (17:31 +0100)]
Rollup merge of #94568 - bjorn3:rustbuild_remove_dead_code, r=Mark-Simulacrum

Remove some dead code from rustbuild

This should reduce build time a tiny bit.

2 years agoRollup merge of #94549 - m-ou-se:thread-is-finished, r=yaahc
Matthias Krüger [Fri, 4 Mar 2022 16:31:06 +0000 (17:31 +0100)]
Rollup merge of #94549 - m-ou-se:thread-is-finished, r=yaahc

Rename JoinHandle::is_running to is_finished.

This is renaming `is_running` to `is_finished` as discussed on the tracking issue here: https://github.com/rust-lang/rust/issues/90470#issuecomment-1050188499

Taking some of the docs suggestions from https://github.com/rust-lang/rust/pull/94033

2 years agoRollup merge of #94524 - bjorn3:remove_num_cpus, r=Mark-Simulacrum
Matthias Krüger [Fri, 4 Mar 2022 16:31:05 +0000 (17:31 +0100)]
Rollup merge of #94524 - bjorn3:remove_num_cpus, r=Mark-Simulacrum

Remove num_cpus dependency from bootstrap, build-manifest and rustc_s…

…ession

`std::threads::available_parallelism` was stabilized in rust 1.59.

r? ```````````````````````````@Mark-Simulacrum```````````````````````````

2 years agoRollup merge of #94236 - reez12g:add_track_caller_87707, r=yaahc
Matthias Krüger [Fri, 4 Mar 2022 16:31:04 +0000 (17:31 +0100)]
Rollup merge of #94236 - reez12g:add_track_caller_87707, r=yaahc

Add #[track_caller] to track callers when initializing poisoned Once

This PR is for this Issue.
https://github.com/rust-lang/rust/issues/87707

With this fix, we expect to be able to track the caller when poisoned Once is initialized.

2 years agoAuto merge of #94570 - shampoofactory:reopen-91719, r=workingjubilee
bors [Fri, 4 Mar 2022 13:06:14 +0000 (13:06 +0000)]
Auto merge of #94570 - shampoofactory:reopen-91719, r=workingjubilee

Reopen 91719

Reopened #91719, which was closed inadvertently due to technical difficulties.

2 years agoIntegrate macos x86-64 remove_dir_all() impl. Step 2: readd
Hans Kratz [Mon, 28 Feb 2022 11:30:23 +0000 (12:30 +0100)]
Integrate macos x86-64 remove_dir_all() impl. Step 2: readd

2 years agoIntegrate macos x86-64 remove_dir_all() impl. Step 1: remove
Hans Kratz [Mon, 28 Feb 2022 11:53:12 +0000 (12:53 +0100)]
Integrate macos x86-64 remove_dir_all() impl. Step 1: remove

2 years agoremove_dir_all(): try recursing first instead of trying to unlink()
Hans Kratz [Mon, 28 Feb 2022 09:34:06 +0000 (10:34 +0100)]
remove_dir_all(): try recursing first instead of trying to unlink()

This only affects the `slow` code path, if there is no `dirent.d_type` or if
the type is `DT_UNKNOWN`.

POSIX specifies that calling `unlink()` or `unlinkat(..., 0)` on a directory can
succeed:
> "The _path_ argument shall not name a directory unless the process has
> appropriate privileges and the implementation supports using _unlink()_ on
> directories."
This however can cause orphaned directories requiring an fsck e.g. on Illumos
UFS, so we have to avoid that in the common case. We now just try to recurse
into it first and unlink() if we can't open it as a directory.

2 years agoEnable conditional compilation checking on the Rust codebase
Loïc BRANSTETT [Wed, 23 Feb 2022 17:27:36 +0000 (18:27 +0100)]
Enable conditional compilation checking on the Rust codebase

2 years agoAuto merge of #94539 - tmiasko:string-attributes, r=nikic
bors [Fri, 4 Mar 2022 10:38:11 +0000 (10:38 +0000)]
Auto merge of #94539 - tmiasko:string-attributes, r=nikic

Pass LLVM string attributes as string slices

2 years agoAdd #![allow(unexpected_cfgs)] in preparation of global --check-cfg
Loïc BRANSTETT [Sat, 26 Feb 2022 00:31:36 +0000 (01:31 +0100)]
Add #![allow(unexpected_cfgs)] in preparation of global --check-cfg

2 years agoRemove invalid #[cfg(tests)] in index_map
Loïc BRANSTETT [Wed, 23 Feb 2022 17:22:28 +0000 (18:22 +0100)]
Remove invalid #[cfg(tests)] in index_map

2 years agoAdd well known values to --check-cfg implementation
Loïc BRANSTETT [Fri, 25 Feb 2022 15:10:26 +0000 (16:10 +0100)]
Add well known values to --check-cfg implementation

2 years agoUse if let instead of manual match
est31 [Fri, 4 Mar 2022 09:35:00 +0000 (10:35 +0100)]
Use if let instead of manual match

2 years agoAuto merge of #94159 - erikdesjardins:align-load, r=nikic
bors [Fri, 4 Mar 2022 08:14:31 +0000 (08:14 +0000)]
Auto merge of #94159 - erikdesjardins:align-load, r=nikic

Add !align metadata on loads of &/&mut/Box

Note that this refers to the alignment of what the loaded value points
to, _not_ the alignment of the loaded value itself.

r? `@ghost` (blocked on #94158)

2 years agoAuto merge of #94096 - cjgillot:ensure-stability, r=lcnr
bors [Fri, 4 Mar 2022 05:49:14 +0000 (05:49 +0000)]
Auto merge of #94096 - cjgillot:ensure-stability, r=lcnr

Ensure stability directives are checked in all cases

Split off  #93017

Stability and deprecation were not checked in all cases, for instance if a type error happened.
This PR moves the check earlier in the pipeline to ensure the errors are emitted in all cases.

r? `@lcnr`

2 years agofix invalid `unresolved imports` errors the asterisk wildcard syntax causes
Takayuki Maeda [Fri, 4 Mar 2022 04:07:39 +0000 (13:07 +0900)]
fix invalid `unresolved imports` errors the asterisk wildcard syntax causes

use a path variabale

2 years agodelay bug instead of skipping check_expr
Michael Goulet [Fri, 4 Mar 2022 04:39:50 +0000 (20:39 -0800)]
delay bug instead of skipping check_expr

2 years agoAdd #[track_caller] to track callers when initializing poisoned Once
reez12g [Mon, 21 Feb 2022 23:37:52 +0000 (08:37 +0900)]
Add #[track_caller] to track callers when initializing poisoned Once

2 years agoAuto merge of #94588 - Dylan-DPC:rollup-7pxd0i3, r=Dylan-DPC
bors [Fri, 4 Mar 2022 02:53:40 +0000 (02:53 +0000)]
Auto merge of #94588 - Dylan-DPC:rollup-7pxd0i3, r=Dylan-DPC

Rollup of 10 pull requests

Successful merges:

 - #88805 (Clarification of default socket flags)
 - #93418 (rustdoc & doc: no `shortcut` for `rel="icon"`)
 - #93913 (Remove the everybody loops pass)
 - #93965 (Make regular stdio lock() return 'static handles)
 - #94339 (ARM: Only allow using d16-d31 with asm! when supported by the target)
 - #94404 (Make Ord and PartialOrd opt-out in `newtype_index`)
 - #94466 (bootstrap: correct reading of flags for llvm)
 - #94572 (Use `HandleOrNull` and `HandleOrInvalid` in the Windows FFI bindings.)
 - #94575 (CTFE SwitchInt: update comment)
 - #94582 (Fix a bug in `x.py fmt` that prevents some files being formatted.)

Failed merges:

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

2 years agoDo not recover from `Ty?` in macro parsing
Esteban Kuber [Fri, 4 Mar 2022 02:03:55 +0000 (02:03 +0000)]
Do not recover from `Ty?` in macro parsing

Follow up to #92746. Address #94510.

2 years agoRollup merge of #94582 - nnethercote:fix-x-fmt-bug, r=Mark-Simulacrum
Dylan DPC [Fri, 4 Mar 2022 01:06:44 +0000 (02:06 +0100)]
Rollup merge of #94582 - nnethercote:fix-x-fmt-bug, r=Mark-Simulacrum

Fix a bug in `x.py fmt` that prevents some files being formatted.

If you have a file in the repository root with the same name as a file
somewhere within a directory, the latter currently won't get formatted.

I have experienced this multiple times and not understood what was
happening; I finally figured out the problem today. This commit fixes
the problem.

r? ```@Mark-Simulacrum```

2 years agoRollup merge of #94575 - RalfJung:switchint-comment, r=oli-obk
Dylan DPC [Fri, 4 Mar 2022 01:06:43 +0000 (02:06 +0100)]
Rollup merge of #94575 - RalfJung:switchint-comment, r=oli-obk

CTFE SwitchInt: update comment

I just wondered why this doesn't use `binary_op`.

r? ```@oli-obk```

2 years agoRollup merge of #94572 - sunfishcode:sunfishcode/handle-or, r=joshtriplett
Dylan DPC [Fri, 4 Mar 2022 01:06:42 +0000 (02:06 +0100)]
Rollup merge of #94572 - sunfishcode:sunfishcode/handle-or, r=joshtriplett

Use `HandleOrNull` and `HandleOrInvalid` in the Windows FFI bindings.

Use the new `HandleOrNull` and `HandleOrInvalid` types that were introduced
as part of [I/O safety] in a few functions in the Windows FFI bindings.

This factors out an `unsafe` block and two `unsafe` function calls in the
Windows implementation code.

And, it helps test `HandleOrNull` and `HandleOrInvalid`, and indeed, it
turned up a bug: `OwnedHandle` also needs to be `#[repr(transparent)]`,
as it's used inside of `HandleOrNull` and `HandleOrInvalid` which are also
`#[repr(transparent)]`.

r? ```@joshtriplett```

[I/O safety]: https://github.com/rust-lang/rust/issues/87074

2 years agoRollup merge of #94466 - jonhoo:bootstrap-proper-env-flags, r=Mark-Simulacrum
Dylan DPC [Fri, 4 Mar 2022 01:06:41 +0000 (02:06 +0100)]
Rollup merge of #94466 - jonhoo:bootstrap-proper-env-flags, r=Mark-Simulacrum

bootstrap: correct reading of flags for llvm

First, this reverts the `CFLAGS`/`CXXFLAGS` of #93918. Those flags are
already read by `cc` and populated into `Build` earlier on in the
process. We shouldn't be overriding that based on `CFLAGS`, since `cc`
also respects overrides like `CFLAGS_{TARGET}` and `HOST_CFLAGS`, which
we want to take into account.

Second, this adds the same capability to specify target-specific
versions of `LDFLAGS` as we have through `cc` for the `C*` flags:
https://github.com/alexcrichton/cc-rs#external-configuration-via-environment-variables

2 years agoRollup merge of #94404 - pierwill:newtype-index-noord-2, r=Aaron1011
Dylan DPC [Fri, 4 Mar 2022 01:06:41 +0000 (02:06 +0100)]
Rollup merge of #94404 - pierwill:newtype-index-noord-2, r=Aaron1011

Make Ord and PartialOrd opt-out in `newtype_index`

Part of work on #90317. This will allow us to do

```diff
rustc_index::newtype_index! {
    /// A unique ID associated with a macro invocation and expansion.
    pub struct LocalExpnId {
        ENCODABLE = custom
        DEBUG_FORMAT = "expn{}"
+       ORD_IMPL = off
    }
}
```

2 years agoRollup merge of #94339 - Amanieu:arm-d32, r=nagisa
Dylan DPC [Fri, 4 Mar 2022 01:06:40 +0000 (02:06 +0100)]
Rollup merge of #94339 - Amanieu:arm-d32, r=nagisa

ARM: Only allow using d16-d31 with asm! when supported by the target

Support can be determined by checking for the "d32" LLVM feature.

r? ```````````````@nagisa```````````````

2 years agoRollup merge of #93965 - Mark-Simulacrum:owned-stdio, r=dtolnay
Dylan DPC [Fri, 4 Mar 2022 01:06:39 +0000 (02:06 +0100)]
Rollup merge of #93965 - Mark-Simulacrum:owned-stdio, r=dtolnay

Make regular stdio lock() return 'static handles

This also deletes the unstable API surface area previously added to expose this
functionality on new methods rather than built into the current set.

Closes #86845 (tracking issue for unstable API needed without this)

r? ``````@dtolnay`````` to kick off T-libs-api FCP

2 years agoRollup merge of #93913 - bjorn3:remove_everybody_loops, r=jackh726
Dylan DPC [Fri, 4 Mar 2022 01:06:38 +0000 (02:06 +0100)]
Rollup merge of #93913 - bjorn3:remove_everybody_loops, r=jackh726

Remove the everybody loops pass

It isn't used anymore by rustdoc.

Split out of https://github.com/rust-lang/rust/pull/92895. There has been some previous discussion there.

2 years agoRollup merge of #93418 - ojeda:no-shortcut, r=camelid
Dylan DPC [Fri, 4 Mar 2022 01:06:38 +0000 (02:06 +0100)]
Rollup merge of #93418 - ojeda:no-shortcut, r=camelid

rustdoc & doc: no `shortcut` for `rel="icon"`

According to https://html.spec.whatwg.org/multipage/links.html#rel-icon:

> For historical reasons, the `icon` keyword may be preceded by the keyword "`shortcut`".

And to https://developer.mozilla.org/en-US/docs/Web/HTML/Link_types:

> **Warning:** The `shortcut` link type is often seen before `icon`, but this link type is non-conforming, ignored and **web authors must not use it anymore.**

While it was removed from the Rust logo case a while ago in commit 085679c ("Use theme-adaptive SVG favicon from other Rust sites"), it is still there for the custom logo case.

Also updated a few other instances.

Signed-off-by: Miguel Ojeda <ojeda@kernel.org>
2 years agoRollup merge of #88805 - krhancoc:master, r=dtolnay
Dylan DPC [Fri, 4 Mar 2022 01:06:37 +0000 (02:06 +0100)]
Rollup merge of #88805 - krhancoc:master, r=dtolnay

Clarification of default socket flags

This PR outlines the decision to disable inheritance of socket objects when possible to child processes in the documentation.

2 years agoAuto merge of #94009 - compiler-errors:gat-rustdoc, r=GuillaumeGomez
bors [Fri, 4 Mar 2022 00:27:23 +0000 (00:27 +0000)]
Auto merge of #94009 - compiler-errors:gat-rustdoc, r=GuillaumeGomez

Support GATs in Rustdoc

Implements:
1. Rendering GATs in trait definitions and impl blocks
2. Rendering GATs in types (e.g. in the return type of a function)

Fixes #92341

This is my first rustdoc PR, so I have absolutely no idea how to produce tests for this. Advice from the rustdoc team would be wonderful!

I tested locally and things looked correct:
![image](https://user-images.githubusercontent.com/3674314/153988325-9732cbf3-0645-4e1a-9e64-ddfd93877b55.png)

2 years agoFix a bug in `x.py fmt` that prevents some files being formatted.
Nicholas Nethercote [Thu, 3 Mar 2022 22:26:31 +0000 (09:26 +1100)]
Fix a bug in `x.py fmt` that prevents some files being formatted.

If you have a file in the repository root with the same name as a file
somewhere within a directory, the latter currently won't get formatted.

I have experienced this multiple times and not understood what was
happening; I finally figured out the problem today. This commit fixes
the problem.

2 years agoAuto merge of #94571 - matthiaskrgr:rollup-4ul5ydb, r=matthiaskrgr
bors [Thu, 3 Mar 2022 21:40:28 +0000 (21:40 +0000)]
Auto merge of #94571 - matthiaskrgr:rollup-4ul5ydb, r=matthiaskrgr

Rollup of 8 pull requests

Successful merges:

 - #92697 (Use cgroup quotas for calculating `available_parallelism`)
 - #94057 (improve comments for `simplify_type`)
 - #94547 (`parse_tt` cleanups)
 - #94550 (rustdoc: Add test for higher kinded functions generated by macros)
 - #94551 (Doc: Fix use of quote instead of backstick in Adapter::map.)
 - #94554 (Fix invalid lint_node_id being put on a removed stmt)
 - #94555 (all: fix some typos)
 - #94563 (Remove a unnecessary `..` pattern)

Failed merges:

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

2 years agoadd tests
Michael Goulet [Sun, 20 Feb 2022 03:34:05 +0000 (19:34 -0800)]
add tests

2 years agodon't forget generics for GATs in impls
Michael Goulet [Tue, 15 Feb 2022 03:06:08 +0000 (19:06 -0800)]
don't forget generics for GATs in impls

2 years agomake generic projection types print correctly
Michael Goulet [Tue, 15 Feb 2022 03:01:56 +0000 (19:01 -0800)]
make generic projection types print correctly

2 years agoonly disable SIMD for doctests in Miri (not for the stdlib build itself)
Ralf Jung [Thu, 3 Mar 2022 20:11:06 +0000 (15:11 -0500)]
only disable SIMD for doctests in Miri (not for the stdlib build itself)

2 years agoMiri can run this test now
Ralf Jung [Thu, 3 Mar 2022 19:52:14 +0000 (14:52 -0500)]
Miri can run this test now

2 years agofix a warning when building core tests with cfg(miri)
Ralf Jung [Thu, 3 Mar 2022 19:52:01 +0000 (14:52 -0500)]
fix a warning when building core tests with cfg(miri)

2 years agoAdd a missing `#[no_mangle]`
scottmcm [Thu, 3 Mar 2022 19:52:45 +0000 (19:52 +0000)]
Add a missing `#[no_mangle]`

2 years agoCTFE SwitchInt: update comment
Ralf Jung [Thu, 3 Mar 2022 19:46:29 +0000 (14:46 -0500)]
CTFE SwitchInt: update comment

2 years agoUse `HandleOrNull` and `HandleOrInvalid` in the Windows FFI bindings.
Dan Gohman [Thu, 3 Mar 2022 19:07:23 +0000 (11:07 -0800)]
Use `HandleOrNull` and `HandleOrInvalid` in the Windows FFI bindings.

Use the new `HandleOrNull` and `HandleOrInvalid` types that were introduced
as part of [I/O safety] in a few functions in the Windows FFI bindings.

This factors out an `unsafe` block and two `unsafe` function calls in the
Windows implementation code.

And, it helps test `HandleOrNull` and `HandleOrInvalid`, which indeed turned
up a bug: `OwnedHandle` also needs to be `#[repr(transparent)]`, as it's
used inside of `HandleOrNull` and `HandleOrInvalid` which are also
`#[repr(transparent)]`.

[I/O safety]: https://github.com/rust-lang/rust/issues/87074

2 years agoRollup merge of #94563 - TaKO8Ki:remove-unnecessary-patten-for-ignoring-remaining...
Matthias Krüger [Thu, 3 Mar 2022 19:01:49 +0000 (20:01 +0100)]
Rollup merge of #94563 - TaKO8Ki:remove-unnecessary-patten-for-ignoring-remaining-parts, r=Dylan-DPC

Remove a unnecessary `..` pattern

2 years agoRollup merge of #94555 - cuishuang:master, r=oli-obk
Matthias Krüger [Thu, 3 Mar 2022 19:01:48 +0000 (20:01 +0100)]
Rollup merge of #94555 - cuishuang:master, r=oli-obk

all: fix some typos

Signed-off-by: cuishuang <imcusg@gmail.com>
2 years agoRollup merge of #94554 - Urgau:stmt-node-id-ice, r=petrochenkov
Matthias Krüger [Thu, 3 Mar 2022 19:01:48 +0000 (20:01 +0100)]
Rollup merge of #94554 - Urgau:stmt-node-id-ice, r=petrochenkov

Fix invalid lint_node_id being put on a removed stmt

This pull-request remove a invalid `assign_id!` being put on an stmt node.

The problem is that this node is being removed away by a cfg making it unreachable when triggering a buffered lint.
The comment in the other match arm already tell to not assign a id because it could have a `#[cfg()]` so this is just respecting the comment.

Fixes https://github.com/rust-lang/rust/issues/94523
r? ```````@petrochenkov```````

2 years agoRollup merge of #94551 - darnuria:doc-map-backstick, r=dtolnay
Matthias Krüger [Thu, 3 Mar 2022 19:01:47 +0000 (20:01 +0100)]
Rollup merge of #94551 - darnuria:doc-map-backstick, r=dtolnay

Doc: Fix use of quote instead of backstick in Adapter::map.

A little commit to fix documentation rendering and semantics in https://doc.rust-lang.org/std/iter/struct.Map.html#notes-about-side-effects `"` where used around an expression instead \`.

Screenshot on doc.rust-lang.org:
![2022-03-03 11-21-43_backstick](https://user-images.githubusercontent.com/2827553/156546536-569b7692-7ac4-4388-8e93-c1628ddc6a0f.png)

Looking forward: Maybe reworking the doc to use assert_eq like the upper paragraph:
```
let v: Vec<i32> = vec![1, 2, 3].into_iter().map(|x| x + 1).rev().collect();

assert_eq!(v, [4, 3, 2]);
```

2 years agoRollup merge of #94550 - GuillaumeGomez:HKF-macros, r=notriddle
Matthias Krüger [Thu, 3 Mar 2022 19:01:46 +0000 (20:01 +0100)]
Rollup merge of #94550 - GuillaumeGomez:HKF-macros, r=notriddle

rustdoc: Add test for higher kinded functions generated by macros

Fixes #75564.

The problem has been solved apparently so adding a test to prevent a regression.

r? ```@notriddle```

2 years agoRollup merge of #94547 - nnethercote:parse_tt-cleanups, r=petrochenkov
Matthias Krüger [Thu, 3 Mar 2022 19:01:45 +0000 (20:01 +0100)]
Rollup merge of #94547 - nnethercote:parse_tt-cleanups, r=petrochenkov

`parse_tt` cleanups

I've been looking closely at this code, and saw some opportunities to improve its readability.

r? ```````@petrochenkov```````

2 years agoRollup merge of #94057 - lcnr:simplify_type-uwu, r=nikomatsakis
Matthias Krüger [Thu, 3 Mar 2022 19:01:44 +0000 (20:01 +0100)]
Rollup merge of #94057 - lcnr:simplify_type-uwu, r=nikomatsakis

improve comments for `simplify_type`

Should now correctly describe what's going on. Experimented with checking the invariant for projections
but that ended up requiring fairly involved changes. I assume that it is not possible to get unsoundness here,
at least for now and I can pretty much guarantee that it's impossible to trigger it by accident.

r? `````@nikomatsakis````` cc #92721

2 years agoRollup merge of #92697 - the8472:cgroups, r=joshtriplett
Matthias Krüger [Thu, 3 Mar 2022 19:01:43 +0000 (20:01 +0100)]
Rollup merge of #92697 - the8472:cgroups, r=joshtriplett

Use cgroup quotas for calculating `available_parallelism`

Automated tests for this are possible but would require a bunch of assumptions. It requires root + a recent kernel, systemd and maybe docker. And even then it would need a helper binary since the test has to run in a separate process.

Limitations

* only supports cgroup v2 and assumes it's mounted under `/sys/fs/cgroup`
* procfs must be available
* the quota gets mixed into `sched_getaffinity`, so if the latter doesn't work then quota information gets ignored too

Manually tested via

```
// spawn a new cgroup scope for the current user
$ sudo systemd-run -p CPUQuota="300%" --uid=$(id -u) -tdS

// quota.rs
#![feature(available_parallelism)]
fn main() {
    println!("{:?}", std::thread::available_parallelism()); // prints Ok(3)
}
```

strace:

```
sched_getaffinity(3041643, 32, [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47]) = 32
openat(AT_FDCWD, "/proc/self/cgroup", O_RDONLY|O_CLOEXEC) = 3
statx(0, NULL, AT_STATX_SYNC_AS_STAT, STATX_ALL, NULL) = -1 EFAULT (Bad address)
statx(3, "", AT_STATX_SYNC_AS_STAT|AT_EMPTY_PATH, STATX_ALL, {stx_mask=STATX_BASIC_STATS|STATX_MNT_ID, stx_attributes=0, stx_mode=S_IFREG|0444, stx_size=0, ...}) = 0
lseek(3, 0, SEEK_CUR)                   = 0
read(3, "0::/system.slice/run-u31477.serv"..., 128) = 36
read(3, "", 92)                         = 0
close(3)                                = 0
statx(AT_FDCWD, "/sys/fs/cgroup/system.slice/run-u31477.service/cgroup.controllers", AT_STATX_SYNC_AS_STAT, STATX_ALL, {stx_mask=STATX_BASIC_STATS|STATX_MNT_ID, stx_attributes=0, stx_mode=S_IFREG|0444, stx_size=0, ...}) = 0
openat(AT_FDCWD, "/sys/fs/cgroup/system.slice/run-u31477.service/cpu.max", O_RDONLY|O_CLOEXEC) = 3
statx(3, "", AT_STATX_SYNC_AS_STAT|AT_EMPTY_PATH, STATX_ALL, {stx_mask=STATX_BASIC_STATS|STATX_MNT_ID, stx_attributes=0, stx_mode=S_IFREG|0644, stx_size=0, ...}) = 0
lseek(3, 0, SEEK_CUR)                   = 0
read(3, "300000 100000\n", 20)          = 14
read(3, "", 6)                          = 0
close(3)                                = 0
openat(AT_FDCWD, "/sys/fs/cgroup/system.slice/cpu.max", O_RDONLY|O_CLOEXEC) = 3
statx(3, "", AT_STATX_SYNC_AS_STAT|AT_EMPTY_PATH, STATX_ALL, {stx_mask=STATX_BASIC_STATS|STATX_MNT_ID, stx_attributes=0, stx_mode=S_IFREG|0644, stx_size=0, ...}) = 0
lseek(3, 0, SEEK_CUR)                   = 0
read(3, "max 100000\n", 20)             = 11
read(3, "", 9)                          = 0
close(3)                                = 0
openat(AT_FDCWD, "/sys/fs/cgroup/cpu.max", O_RDONLY|O_CLOEXEC) = -1 ENOENT (No such file or directory)
sched_getaffinity(0, 128, [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47]) = 40
```

r? ```````@joshtriplett```````
cc ```````@yoshuawuyts```````

Tracking issue and previous discussion: #74479

2 years agoBless nll tests.
Camille GILLOT [Thu, 3 Mar 2022 18:59:52 +0000 (19:59 +0100)]
Bless nll tests.

2 years agoAuto merge of #87835 - xFrednet:rfc-2383-expect-attribute-with-ids, r=wesleywiser
bors [Thu, 3 Mar 2022 18:59:32 +0000 (18:59 +0000)]
Auto merge of #87835 - xFrednet:rfc-2383-expect-attribute-with-ids, r=wesleywiser

Implementation of the `expect` attribute (RFC 2383)

This is an implementation of the `expect` attribute as described in [RFC-2383](https://rust-lang.github.io/rfcs/2383-lint-reasons.html). The attribute allows the suppression of lint message by expecting them. Unfulfilled lint expectations (meaning no expected lint was caught) will emit the `unfulfilled_lint_expectations` lint at the `expect` attribute.

### Example
#### input
```rs
// required feature flag
#![feature(lint_reasons)]

#[expect(unused_mut)] // Will warn about an unfulfilled expectation
#[expect(unused_variables)] // Will be fulfilled by x
fn main() {
    let x = 0;
}
```

#### output

```txt
warning: this lint expectation is unfulfilled
  --> $DIR/trigger_lint.rs:3:1
   |
LL | #[expect(unused_mut)] // Will warn about an unfulfilled expectation
   |          ^^^^^^^^^^
   |
   = note: `#[warn(unfulfilled_lint_expectations)]` on by default
```

### Implementation

This implementation introduces `Expect` as a new lint level for diagnostics, which have been expected. All lint expectations marked via the `expect` attribute are collected in the [`LintLevelsBuilder`] and assigned an ID that is stored in the new lint level. The `LintLevelsBuilder` stores all found expectations and the data needed to emit the `unfulfilled_lint_expectations` in the [`LintLevelsMap`] which is the result of the [`lint_levels()`] query.

The [`rustc_errors::HandlerInner`] is the central error handler in rustc and handles the emission of all diagnostics. Lint message with the level `Expect` are suppressed during this emission, while the expectation ID is stored in a set which marks them as fulfilled. The last step is then so simply check if all expectations collected by the [`LintLevelsBuilder`] in the [`LintLevelsMap`] have been marked as fulfilled in the [`rustc_errors::HandlerInner`]. Otherwise, a new lint message will be emitted.

The implementation of the `LintExpectationId` required some special handling to make it stable between sessions. Lints can be emitted during [`EarlyLintPass`]es. At this stage, it's not possible to create a stable identifier. The level instead stores an unstable identifier, which is later converted to a stable `LintExpectationId`.

### Followup TO-DOs
All open TO-DOs have been marked with `FIXME` comments in the code. This is the combined list of them:

* [ ] The current implementation doesn't cover cases where the `unfulfilled_lint_expectations` lint is actually expected by another `expect` attribute.
   * This should be easily possible, but I wanted to get some feedback before putting more work into this.
   * This could also be done in a new PR to not add to much more code to this one
* [ ] Update unstable documentation to reflect this change.
* [ ] Update unstable expectation ids in [`HandlerInner::stashed_diagnostics`](https://doc.rust-lang.org/nightly/nightly-rustc/rustc_errors/struct.HandlerInner.html#structfield.stashed_diagnostics)

### Open questions
I also have a few open questions where I would like to get feedback on:
1. The RFC discussion included a suggestion to change the `expect` attribute to something else. (Initiated by `@Ixrec` [here](https://github.com/rust-lang/rfcs/pull/2383#issuecomment-378424091), suggestion from `@scottmcm` to use `#[should_lint(...)]` [here](https://github.com/rust-lang/rfcs/pull/2383#issuecomment-378648877)). No real conclusion was drawn on that point from my understanding. Is this still open for discussion, or was this discarded with the merge of the RFC?
2. How should the expect attribute deal with the new `force-warn` lint level?

---

This approach was inspired by a discussion with `@LeSeulArtichaut.`

RFC tracking issue: #54503

Mentoring/Implementation issue: #85549

[`LintLevelsBuilder`]: https://doc.rust-lang.org/nightly/nightly-rustc/rustc_lint/levels/struct.LintLevelsBuilder.html
[`LintLevelsMap`]: https://doc.rust-lang.org/nightly/nightly-rustc/rustc_middle/lint/struct.LintLevelMap.html
[`lint_levels()`]: https://doc.rust-lang.org/nightly/nightly-rustc/rustc_middle/ty/context/struct.TyCtxt.html#method.lint_levels
[`rustc_errors::HandlerInner`]: https://doc.rust-lang.org/nightly/nightly-rustc/rustc_errors/struct.HandlerInner.html
[`EarlyLintPass`]: https://doc.rust-lang.org/nightly/nightly-rustc/rustc_lint/trait.EarlyLintPass.html

2 years agoAdd autovectorization codegen test
Jubilee Young [Thu, 3 Mar 2022 07:02:03 +0000 (23:02 -0800)]
Add autovectorization codegen test

Co-authored-by: Scott McMurray <scottmcm@users.noreply.github.com>
2 years agoRedo the array-equality codegen tests for the different threshold
Scott McMurray [Fri, 10 Dec 2021 09:53:13 +0000 (01:53 -0800)]
Redo the array-equality codegen tests for the different threshold

2 years agoRevert #26494 regression
Vin Singh [Wed, 8 Dec 2021 12:01:13 +0000 (12:01 +0000)]
Revert #26494 regression

2 years agoRemove unused Default and Clone derives in config.rs
bjorn3 [Wed, 2 Mar 2022 17:29:00 +0000 (18:29 +0100)]
Remove unused Default and Clone derives in config.rs

2 years agoRemove some dead code from toolstate.rs
bjorn3 [Wed, 2 Mar 2022 17:23:08 +0000 (18:23 +0100)]
Remove some dead code from toolstate.rs

2 years agoMake `Ord`, `PartialOrd` opt-out in `newtype_index`
pierwill [Sat, 26 Feb 2022 20:44:09 +0000 (14:44 -0600)]
Make `Ord`, `PartialOrd` opt-out in `newtype_index`

Also remove `step` impl if `ORD_IMPL = off`

2 years agoMake unknown feature error deterministic.
Camille GILLOT [Mon, 21 Feb 2022 17:59:24 +0000 (18:59 +0100)]
Make unknown feature error deterministic.

2 years agoCleanup feature gates.
Camille GILLOT [Wed, 19 Jan 2022 15:24:49 +0000 (16:24 +0100)]
Cleanup feature gates.

2 years agoGate stability attrs with other attributes.
Camille GILLOT [Tue, 18 Jan 2022 19:37:14 +0000 (20:37 +0100)]
Gate stability attrs with other attributes.

2 years agobootstrap: correct reading of flags for llvm
Jon Gjengset [Tue, 1 Mar 2022 00:40:08 +0000 (16:40 -0800)]
bootstrap: correct reading of flags for llvm

First, this reverts the `CFLAGS`/`CXXFLAGS` of #93918. Those flags are
already read by `cc` and populated into `Build` earlier on in the
process. We shouldn't be overriding that based on `CFLAGS`, since `cc`
also respects overrides like `CFLAGS_{TARGET}` and `HOST_CFLAGS`, which
we want to take into account.

Second, this adds the same capability to specify target-specific
versions of `LDFLAGS` as we have through `cc` for the `C*` flags:
https://github.com/alexcrichton/cc-rs#external-configuration-via-environment-variables

Note that this also necessitated an update to compiletest to treat
CXXFLAGS separately from CFLAGS.

2 years agoRemove the everybody loops pass
bjorn3 [Fri, 25 Jun 2021 09:56:14 +0000 (11:56 +0200)]
Remove the everybody loops pass

It isn't used anymore by rustdoc

2 years agoBless test.
Camille GILLOT [Sun, 16 Jan 2022 18:50:46 +0000 (19:50 +0100)]
Bless test.

2 years agoForce ensure stability_index.
Camille GILLOT [Sun, 16 Jan 2022 18:36:22 +0000 (19:36 +0100)]
Force ensure stability_index.

2 years agoMove query providers.
Camille GILLOT [Sun, 16 Jan 2022 21:26:46 +0000 (22:26 +0100)]
Move query providers.

2 years agoRemove useless map.
Camille GILLOT [Sun, 16 Jan 2022 15:30:33 +0000 (16:30 +0100)]
Remove useless map.