]> git.lizzy.rs Git - rust.git/log
rust.git
4 years agoRollup merge of #73359 - jonas-schievink:do-the-shimmy, r=matthewjasper
Manish Goregaokar [Sat, 20 Jun 2020 02:42:57 +0000 (19:42 -0700)]
Rollup merge of #73359 - jonas-schievink:do-the-shimmy, r=matthewjasper

shim.rs: avoid creating `Call` terminators calling `Self`

Also contains some cleanup and doc comment additions so I could make sense of the code.

Fixes https://github.com/rust-lang/rust/issues/73109
Closes https://github.com/rust-lang/rust/pull/73175

r? @matthewjasper

4 years agoRollup merge of #73347 - tmiasko:incompatible-sanitizers, r=nikic
Manish Goregaokar [Sat, 20 Jun 2020 02:42:55 +0000 (19:42 -0700)]
Rollup merge of #73347 - tmiasko:incompatible-sanitizers, r=nikic

Diagnose use of incompatible sanitizers

Emit an error when incompatible sanitizer are configured through command
line options. Previously the last one configured prevailed and others
were silently ignored.

Additionally use a set to represent configured sanitizers, making it
possible to enable multiple sanitizers at once. At least in principle,
since currently all of them are considered to be incompatible with
others.

4 years agoRollup merge of #73027 - doctorn:issue-72690, r=estebank
Manish Goregaokar [Sat, 20 Jun 2020 02:42:53 +0000 (19:42 -0700)]
Rollup merge of #73027 - doctorn:issue-72690, r=estebank

Make `need_type_info_err` more conservative

Makes sure arg patterns we are going to suggest on are actually contained within the span of the obligation that caused the inference error (credit to @lcnr for suggesting this fix).

There's a subtle trade-off regarding the handling of local patterns which I've left a comment about.

Resolves #72690

4 years agoRollup merge of #72934 - christianpoveda:mut-borrows-in-consts, r=oli-obk
Manish Goregaokar [Sat, 20 Jun 2020 02:42:51 +0000 (19:42 -0700)]
Rollup merge of #72934 - christianpoveda:mut-borrows-in-consts, r=oli-obk

forbid mutable references in all constant contexts except for const-fns

PR to address #71212

cc: @ecstatic-morse

4 years agoRollup merge of #72791 - lcnr:coerce-refactor, r=estebank
Manish Goregaokar [Sat, 20 Jun 2020 02:42:49 +0000 (19:42 -0700)]
Rollup merge of #72791 - lcnr:coerce-refactor, r=estebank

update coerce docs and unify relevant tests

Merges `test/ui/coerce` with `test/ui/coercion`.
Updates the documentation of `librustc_typeck/check/coercion.rs`.
Adds 2 new coercion tests.

4 years agoRollup merge of #72689 - lcnr:common_str, r=estebank
Manish Goregaokar [Sat, 20 Jun 2020 02:42:47 +0000 (19:42 -0700)]
Rollup merge of #72689 - lcnr:common_str, r=estebank

add str to common types

I already expected this to be the case and it may slightly improve perf.

Afaict if we ever want to change str into a lang item this would have to get reverted.
As that would be fairly simple I don't believe this to cause any problems in the future.

4 years agoRollup merge of #71899 - cuviper:try_find_map, r=dtolnay
Manish Goregaokar [Sat, 20 Jun 2020 02:42:45 +0000 (19:42 -0700)]
Rollup merge of #71899 - cuviper:try_find_map, r=dtolnay

Refactor `try_find` a little

~~This works like `find_map`, but mapping to a `Try` type. It stops when `Ok` is `Some(value)`, with an additional short-circuit on `Try::Error`. This is similar to the unstable `try_find`, but has the advantage of being able to directly return the user's `R: Try` type directly, rather than converting to `Result`.~~
(removed -- `try_find_map` was declined in review)

This PR also refactors `try_find` a little to match style. The `E` type parameter was unnecessary, so it's now removed. The folding closure now has reduced parametricity on just `T = Self::Item`, rather
than the whole `Self` iterator type. There's otherwise no functional change in this.

4 years agoRollup merge of #71420 - RalfJung:specialization-incomplete, r=matthewjasper
Manish Goregaokar [Sat, 20 Jun 2020 02:42:43 +0000 (19:42 -0700)]
Rollup merge of #71420 - RalfJung:specialization-incomplete, r=matthewjasper

Specialization is unsound

As discussed in https://github.com/rust-lang/rust/issues/31844#issuecomment-617013949, it might be a good idea to warn users of specialization that the feature they are using is unsound.

I also expanded the "incomplete feature" warning to link the user to the tracking issue.

4 years agoAuto merge of #73511 - Manishearth:rollup-3iffxd8, r=Manishearth
bors [Fri, 19 Jun 2020 23:04:41 +0000 (23:04 +0000)]
Auto merge of #73511 - Manishearth:rollup-3iffxd8, r=Manishearth

Rollup of 13 pull requests

Successful merges:

 - #71568 (Document unsafety in slice/sort.rs)
 - #72709 (`#[deny(unsafe_op_in_unsafe_fn)]` in liballoc)
 - #73214 (Add asm!() support for hexagon)
 - #73248 (save_analysis: improve handling of enum struct variant)
 - #73257 (ty: projections in `transparent_newtype_field`)
 - #73261 (Suggest `?Sized` when applicable for ADTs)
 - #73300 (Implement crate-level-only lints checking.)
 - #73334 (Note numeric literals that can never fit in an expected type)
 - #73357 (Use `LocalDefId` for import IDs in trait map)
 - #73364 (asm: Allow multiple template string arguments; interpret them as newline-separated)
 - #73382 (Only display other method receiver candidates if they actually apply)
 - #73465 (Add specialization of `ToString for char`)
 - #73489 (Refactor hir::Place)

Failed merges:

r? @ghost

4 years agoRefactor `try_find` a little
Josh Stone [Mon, 4 May 2020 18:25:54 +0000 (11:25 -0700)]
Refactor `try_find` a little

The `E` type parameter was unnecessary, so it's now removed. The folding
closure now has reduced parametricity on just `T = Self::Item`, rather
than the whole `Self` iterator type. There's otherwise no functional
change in this.

4 years agoadd new error code
Christian Poveda [Thu, 18 Jun 2020 19:52:37 +0000 (14:52 -0500)]
add new error code

4 years agoAuto merge of #73257 - davidtwco:issue-73249-improper-ctypes-projection, r=lcnr,varkor
bors [Fri, 19 Jun 2020 19:14:45 +0000 (19:14 +0000)]
Auto merge of #73257 - davidtwco:issue-73249-improper-ctypes-projection, r=lcnr,varkor

ty: projections in `transparent_newtype_field`

Fixes #73249.

This PR modifies `transparent_newtype_field` so that it handles
projections with generic parameters, where `normalize_erasing_regions`
would ICE.

4 years agoupdate diagnostics for &mut in constants
Christian Poveda [Mon, 15 Jun 2020 17:05:24 +0000 (12:05 -0500)]
update diagnostics for &mut in constants

4 years agoupdate tests
Christian Poveda [Mon, 8 Jun 2020 20:16:23 +0000 (15:16 -0500)]
update tests

4 years agoforbid mutable references in all constants except const fns
Christian Poveda [Sun, 7 Jun 2020 00:25:02 +0000 (19:25 -0500)]
forbid mutable references in all constants except const fns

4 years agoRollup merge of #73489 - sexxi-goose:init_place_refactor, r=nikomatsakis
Manish Goregaokar [Fri, 19 Jun 2020 16:15:20 +0000 (09:15 -0700)]
Rollup merge of #73489 - sexxi-goose:init_place_refactor, r=nikomatsakis

Refactor hir::Place

For the following code
```rust
let c = || bar(foo.x, foo.x)
```

We generate two different `hir::Place`s for both `foo.x`.
Handling this adds overhead for analysis we need to do for RFC 2229.

We also want to store type information at each Projection to support
analysis as part of the RFC. This resembles what we have for
`mir::Place`

This commit modifies the Place as follows:
- Rename to `PlaceWithHirId`, where there `hir_id` is that of the
expressioin.
- Move any other information that describes the access out to another
struct now called `Place`.
- Removed `Span`, it can be accessed using the [hir
API](https://doc.rust-lang.org/nightly/nightly-rustc/rustc_middle/hir/map/struct.Map.html#method.span)
- Modify `Projection` to be a strucutre of its own, that currently only
contains the `ProjectionKind`.

Adding type information to projections wil be completed as part of https://github.com/rust-lang/project-rfc-2229/issues/5

Closes https://github.com/rust-lang/project-rfc-2229/issues/3

4 years agoRollup merge of #73465 - lzutao:spec-char-tostring, r=sfackler
Manish Goregaokar [Fri, 19 Jun 2020 16:15:18 +0000 (09:15 -0700)]
Rollup merge of #73465 - lzutao:spec-char-tostring, r=sfackler

Add specialization of `ToString for char`

Closes #73462

4 years agoRollup merge of #73382 - Aaron1011:fix/self-receiver-candidates, r=petrochenkov
Manish Goregaokar [Fri, 19 Jun 2020 16:15:16 +0000 (09:15 -0700)]
Rollup merge of #73382 - Aaron1011:fix/self-receiver-candidates, r=petrochenkov

Only display other method receiver candidates if they actually apply

Previously, we would suggest `Box<Self>` as a valid receiver, even if
method resolution only succeeded due to an autoderef (e.g. to `&self`)

4 years agoRollup merge of #73364 - joshtriplett:inline-asm, r=Amanieu
Manish Goregaokar [Fri, 19 Jun 2020 16:15:14 +0000 (09:15 -0700)]
Rollup merge of #73364 - joshtriplett:inline-asm, r=Amanieu

asm: Allow multiple template string arguments; interpret them as newline-separated

Allow the `asm!` macro to accept a series of template arguments, and interpret them as if they were concatenated with a '\n' between them. This allows writing an `asm!` where each line of assembly appears in a separate template string argument.

This syntax makes it possible for rustfmt to reliably format and indent each line of assembly, without risking changes to the inside of a template string. It also avoids the complexity of having the user carefully format and indent a multi-line string (including where to put the surrounding quotes), and avoids the extra indentation and lines of a call to `concat!`.

For example, rewriting the second example from the [blog post on the new inline assembly syntax](https://blog.rust-lang.org/inside-rust/2020/06/08/new-inline-asm.html) using multiple template strings:

```rust

fn main() {
    let mut bits = [0u8; 64];
    for value in 0..=1024u64 {
        let popcnt;
        unsafe {
            asm!(
                "    popcnt {popcnt}, {v}",
                "2:",
                "    blsi rax, {v}",
                "    jz 1f",
                "    xor {v}, rax",
                "    tzcnt rax, rax",
                "    stosb",
                "    jmp 2b",
                "1:",
                v = inout(reg) value => _,
                popcnt = out(reg) popcnt,
                out("rax") _, // scratch
                inout("rdi") bits.as_mut_ptr() => _,
            );
        }
        println!("bits of {}: {:?}", value, &bits[0..popcnt]);
    }
}
```

Note that all the template strings must appear before all other arguments; you cannot, for instance, provide a series of template strings intermixed with the corresponding operands.

4 years agoRollup merge of #73357 - petrochenkov:tmap, r=davidtwco
Manish Goregaokar [Fri, 19 Jun 2020 16:15:12 +0000 (09:15 -0700)]
Rollup merge of #73357 - petrochenkov:tmap, r=davidtwco

Use `LocalDefId` for import IDs in trait map

cc https://github.com/rust-lang/rust/pull/73291#discussion_r439734867

4 years agoRollup merge of #73334 - ayazhafiz:err/num-type-cannot-fit, r=estebank
Manish Goregaokar [Fri, 19 Jun 2020 16:15:10 +0000 (09:15 -0700)]
Rollup merge of #73334 - ayazhafiz:err/num-type-cannot-fit, r=estebank

Note numeric literals that can never fit in an expected type

re https://github.com/rust-lang/rust/pull/72380#discussion_r438289385

Given the toy code

```rust
fn is_positive(n: usize) {
  n > -1_isize;
}
```

We currently get a type mismatch error like the following:

```
error[E0308]: mismatched types
 --> src/main.rs:2:9
  |
2 |     n > -1_isize;
  |         ^^^^^^^^ expected `usize`, found `isize`
  |
help: you can convert an `isize` to `usize` and panic if the converted value wouldn't fit
  |
2 |     n > (-1_isize).try_into().unwrap();
  |         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
```

But clearly, `-1` can never fit into a `usize`, so the suggestion will
always panic. A more useful message would tell the user that the value
can never fit in the expected type:

```
error[E0308]: mismatched types
 --> test.rs:2:9
  |
2 |     n > -1_isize;
  |         ^^^^^^^^ expected `usize`, found `isize`
  |
note: `-1_isize` can never fit into `usize`
 --> test.rs:2:9
  |
2 |     n > -1_isize;
  |         ^^^^^^^^
```

Which is what this commit implements.

I only added this check for negative literals because

- Currently we can only perform such a check for literals (constant
  value propagation is outside the scope of the typechecker at this
  point)
- A lint error for out-of-range numeric literals is already emitted

IMO it makes more sense to put this check in librustc_lint, but as far
as I can tell the typecheck pass happens before the lint pass, so I've
added it here.

r? @estebank

4 years agoRollup merge of #73300 - crlf0710:crate_level_only_check, r=petrochenkov
Manish Goregaokar [Fri, 19 Jun 2020 16:15:08 +0000 (09:15 -0700)]
Rollup merge of #73300 - crlf0710:crate_level_only_check, r=petrochenkov

Implement crate-level-only lints checking.

This implements a crate_level_only flag on lints, and when it is true, it becomes an error when user tries  to specify this flag upon nodes other than crate node.

This also turns on this flag for all non_ascii_ident lints.

4 years agoRollup merge of #73261 - estebank:generics-sized, r=nikomatsakis
Manish Goregaokar [Fri, 19 Jun 2020 16:15:06 +0000 (09:15 -0700)]
Rollup merge of #73261 - estebank:generics-sized, r=nikomatsakis

Suggest `?Sized` when applicable for ADTs

Address #71790, fix #27964.

4 years agoRollup merge of #73257 - davidtwco:issue-73249-improper-ctypes-projection, r=lcnr...
Manish Goregaokar [Fri, 19 Jun 2020 16:15:04 +0000 (09:15 -0700)]
Rollup merge of #73257 - davidtwco:issue-73249-improper-ctypes-projection, r=lcnr,varkor

ty: projections in `transparent_newtype_field`

Fixes #73249.

This PR modifies `transparent_newtype_field` so that it handles
projections with generic parameters, where `normalize_erasing_regions`
would ICE.

4 years agoRollup merge of #73248 - marmeladema:save-analysis-various-fixes, r=Xanewok
Manish Goregaokar [Fri, 19 Jun 2020 16:15:02 +0000 (09:15 -0700)]
Rollup merge of #73248 - marmeladema:save-analysis-various-fixes, r=Xanewok

save_analysis: improve handling of enum struct variant

Fixes #61385

4 years agoRollup merge of #73214 - androm3da:hex_inline_asm_00, r=Amanieu
Manish Goregaokar [Fri, 19 Jun 2020 16:15:00 +0000 (09:15 -0700)]
Rollup merge of #73214 - androm3da:hex_inline_asm_00, r=Amanieu

Add asm!() support for hexagon

4 years agoRollup merge of #72709 - LeSeulArtichaut:unsafe-liballoc, r=nikomatsakis
Manish Goregaokar [Fri, 19 Jun 2020 16:14:58 +0000 (09:14 -0700)]
Rollup merge of #72709 - LeSeulArtichaut:unsafe-liballoc, r=nikomatsakis

`#[deny(unsafe_op_in_unsafe_fn)]` in liballoc

This PR proposes to make use of the new `unsafe_op_in_unsafe_fn` lint, i.e. no longer consider the body of an unsafe function as an unsafe block and require explicit unsafe block to perform unsafe operations.

This has been first (partly) suggested by @Mark-Simulacrum in https://github.com/rust-lang/rust/pull/69245#issuecomment-587817065

Tracking issue for the feature: #71668.
~~Blocked on #71862.~~
r? @Mark-Simulacrum cc @nikomatsakis can you confirm that those changes are desirable? Should I restrict it to only BTree for the moment?

4 years agoRollup merge of #71568 - hbina:document_unsafety_slice_sort, r=joshtriplett
Manish Goregaokar [Fri, 19 Jun 2020 16:14:56 +0000 (09:14 -0700)]
Rollup merge of #71568 - hbina:document_unsafety_slice_sort, r=joshtriplett

Document unsafety in slice/sort.rs

Let me know if these documentations are accurate c:

I don't think I am capable enough to document the safety of `partition_blocks`, however.

Related issue #66219

4 years agoAuto merge of #73504 - RalfJung:rollup-iy8hsvl, r=RalfJung
bors [Fri, 19 Jun 2020 12:30:20 +0000 (12:30 +0000)]
Auto merge of #73504 - RalfJung:rollup-iy8hsvl, r=RalfJung

Rollup of 10 pull requests

Successful merges:

 - #72280 (Fix up autoderef when reborrowing)
 - #72785 (linker: MSVC supports linking static libraries as a whole archive)
 - #73011 (first stage of implementing LLVM code coverage)
 - #73044 (compiletest: Add directives to detect sanitizer support)
 - #73054 (memory access sanity checks: abort instead of panic)
 - #73136 (Change how compiler-builtins gets many CGUs)
 - #73280 (Add E0763)
 - #73317 (bootstrap: read config from $RUST_BOOTSTRAP_CONFIG)
 - #73350 (bootstrap/install.rs: support a nonexistent `prefix` in `x.py install`)
 - #73352 (Speed up bootstrap a little.)

Failed merges:

r? @ghost

4 years agoRollup merge of #73352 - ehuss:bootstrap-metadata, r=Mark-Simulacrum
Ralf Jung [Fri, 19 Jun 2020 12:29:35 +0000 (14:29 +0200)]
Rollup merge of #73352 - ehuss:bootstrap-metadata, r=Mark-Simulacrum

Speed up bootstrap a little.

The bootstrap script was calling `cargo metadata` 3 times (or 6 with `-v`). This is a very expensive operation, and this attempts to avoid the extra calls. On my system, a simple command like `./x.py test -h -v` goes from about 3 seconds to 0.4.

An overview of the changes:

- Call `cargo metadata` only once with `--no-deps`. Optional dependencies are filtered in `in_tree_crates` (handling `profiler_builtins` and `rustc_codegen_llvm` which are driven by the config).
- Remove a duplicate call to `metadata::build` when using `-v`. I'm not sure why it was there, it looks like a mistake or vestigial from previous behavior.
- Remove check for `_shim`, I believe all the `_shim` crates are now gone.
- Remove check for `rustc_` and `*san` for `test::Crate::should_run`, these are no longer dependencies in the `test` tree.
- Use relative paths in `./x.py test -h -v` output.
- Some code cleanup (remove unnecessary `find_compiler_crates`, etc.).
- Show suite paths (`src/test/ui/...`) in `./x.py test -h -v` output.
- Some doc comments.

4 years agoRollup merge of #73350 - nodakai:install-rs-support-nonexistent-prefix, r=Mark-Simulacrum
Ralf Jung [Fri, 19 Jun 2020 12:29:33 +0000 (14:29 +0200)]
Rollup merge of #73350 - nodakai:install-rs-support-nonexistent-prefix, r=Mark-Simulacrum

bootstrap/install.rs: support a nonexistent `prefix` in `x.py install`

PR #49778 introduced fs::canonicalize() which fails for a nonexistent path.
This is a surprise for someone used to GNU Autotools' configure which can create any necessary intermediate directories in prefix.

This change makes it run fs::create_dir_all() before canonicalize().

4 years agoRollup merge of #73317 - davidtwco:bootstrap-config-env-var, r=Mark-Simulacrum
Ralf Jung [Fri, 19 Jun 2020 12:29:31 +0000 (14:29 +0200)]
Rollup merge of #73317 - davidtwco:bootstrap-config-env-var, r=Mark-Simulacrum

bootstrap: read config from $RUST_BOOTSTRAP_CONFIG

This PR modifies bootstrap so that `config.toml` is read first from `RUST_BOOTSTRAP_CONFIG`, then `--config` and finally `config.toml` in the current directory.

This is a subjective change, intended to improve the ergnomics when using "development shells" for rustc development (for example, using tools such as Nix) which set environment variables to ensure a reproducible environment (these development shells can then be version controlled, e.g. [my rustc shell](https://github.com/davidtwco/veritas/blob/6b74a5c170b6efb2c7b094352932f9158f97eec0/nix/shells/rustc.nix)). By optionally reading `config.toml` from an environment variable, a `config.toml` can be defined in the development shell and a path to it exposed in the `RUST_BOOTSTRAP_CONFIG` environment variable - avoiding the need to manually symlink the contents of this file to `config.toml` in the working directory.

4 years agoRollup merge of #73280 - GuillaumeGomez:add-e0763, r=petrochenkov
Ralf Jung [Fri, 19 Jun 2020 12:29:29 +0000 (14:29 +0200)]
Rollup merge of #73280 - GuillaumeGomez:add-e0763, r=petrochenkov

Add E0763

4 years agoRollup merge of #73136 - alexcrichton:thinlto-compiler-builtins, r=Mark-Simulacrum
Ralf Jung [Fri, 19 Jun 2020 12:29:26 +0000 (14:29 +0200)]
Rollup merge of #73136 - alexcrichton:thinlto-compiler-builtins, r=Mark-Simulacrum

Change how compiler-builtins gets many CGUs

This commit intends to fix an accidental regression from #70846. The
goal of #70846 was to build compiler-builtins with a maximal number of
CGUs to ensure that each module in the source corresponds to an object
file. This high degree of control for compiler-builtins is desirable to
ensure that there's at most one exported symbol per CGU, ideally
enabling compiler-builtins to not conflict with the system libgcc as
often.

In #70846, however, only part of the compiler understands that
compiler-builtins is built with many CGUs. The rest of the compiler
thinks it's building with `sess.codegen_units()`. Notably the
calculation of `sess.lto()` consults `sess.codegen_units()`, which when
there's only one CGU it disables ThinLTO. This means that
compiler-builtins is built without ThinLTO, which is quite harmful to
performance! This is the root of the cause from #73135 where intrinsics
were found to not be inlining trivial functions.

The fix applied in this commit is to remove the special-casing of
compiler-builtins in the compiler. Instead the build system is now
responsible for special-casing compiler-builtins. It doesn't know
exactly how many CGUs will be needed but it passes a large number that
is assumed to be much greater than the number of source-level modules
needed. After reading the various locations in the compiler source, this
seemed like the best solution rather than adding more and more special
casing in the compiler for compiler-builtins.

Closes #73135

4 years agoRollup merge of #73054 - RalfJung:dont-panic, r=Mark-Simulacrum
Ralf Jung [Fri, 19 Jun 2020 12:29:24 +0000 (14:29 +0200)]
Rollup merge of #73054 - RalfJung:dont-panic, r=Mark-Simulacrum

memory access sanity checks: abort instead of panic

Suggested by @Mark-Simulacrum, this should help reduce the performance impact of these checks.

4 years agoRollup merge of #73044 - tmiasko:compiletest-san, r=nikomatsakis
Ralf Jung [Fri, 19 Jun 2020 12:29:22 +0000 (14:29 +0200)]
Rollup merge of #73044 - tmiasko:compiletest-san, r=nikomatsakis

compiletest: Add directives to detect sanitizer support

Add needs-sanitizer-{address,leak,memory,thread} directive indicating
that test requires target with support for specific sanitizer.

This is an addition to the existing needs-sanitizer-support directive
indicating that test requires a sanitizer runtime library.

The existing needs-sanitizer-support directive could be incorporated into the
new ones, but I decided to retain it, since it enables running sanitizer
codegen tests even when building of sanitizer runtime libraries is disabled.

4 years agoRollup merge of #73011 - richkadel:llvm-count-from-mir-pass, r=tmandry
Ralf Jung [Fri, 19 Jun 2020 12:29:20 +0000 (14:29 +0200)]
Rollup merge of #73011 - richkadel:llvm-count-from-mir-pass, r=tmandry

first stage of implementing LLVM code coverage

This PR replaces #70680 (WIP toward LLVM Code Coverage for Rust) since I am re-implementing the Rust LLVM code coverage feature in a different part of the compiler (in MIR pass(es) vs AST).

This PR updates rustc with `-Zinstrument-coverage` option that injects the llvm intrinsic `instrprof.increment()` for code generation.

This initial version only injects counters at the top of each function, and does not yet implement the required coverage map.

Upcoming PRs will add the coverage map, and add more counters and/or counter expressions for each conditional code branch.

Rust compiler MCP https://github.com/rust-lang/compiler-team/issues/278
Relevant issue: #34701 - Implement support for LLVMs code coverage instrumentation

***[I put together some development notes here, under a separate branch.](https://github.com/richkadel/rust/blob/cfa0b21d34ee64e4ebee226101bd2ef0c6757865/src/test/codegen/coverage-experiments/README-THIS-IS-TEMPORARY.md)***

4 years agoRollup merge of #72785 - petrochenkov:wholemsvc, r=matthewjasper
Ralf Jung [Fri, 19 Jun 2020 12:29:18 +0000 (14:29 +0200)]
Rollup merge of #72785 - petrochenkov:wholemsvc, r=matthewjasper

linker: MSVC supports linking static libraries as a whole archive

4 years agoRollup merge of #72280 - nbdd0121:typeck, r=nikomatsakis
Ralf Jung [Fri, 19 Jun 2020 12:29:16 +0000 (14:29 +0200)]
Rollup merge of #72280 - nbdd0121:typeck, r=nikomatsakis

Fix up autoderef when reborrowing

Currently `(f)()` and `f.call_mut()` behaves differently if expression `f` contains autoderef in it. This causes a weird error in #72225.

When `f` is type checked, `Deref` is used (this is expected as we can't yet determine if we should use `Fn` or `FnMut`). When subsequently we determine the actual trait to be used, when using the `f.call_mut()` syntax the `Deref` is patched to `DerefMut`, while for the `(f)()` syntax case it is not.

This PR replicates the fixup for the first case.

Fixes #72225
Fixes #68590

4 years agoApply suggestions from code review
LeSeulArtichaut [Sun, 31 May 2020 13:06:51 +0000 (15:06 +0200)]
Apply suggestions from code review

Co-authored-by: nikomatsakis <niko@alum.mit.edu>
4 years ago`#[deny(unsafe_op_in_unsafe_fn)]` in liballoc
LeSeulArtichaut [Thu, 28 May 2020 21:27:00 +0000 (23:27 +0200)]
`#[deny(unsafe_op_in_unsafe_fn)]` in liballoc

4 years agoty: simplify `transparent_newtype_field`
David Wood [Fri, 12 Jun 2020 15:14:05 +0000 (16:14 +0100)]
ty: simplify `transparent_newtype_field`

This commit removes the normalization from `transparent_newtype_field` -
turns out it wasn't necessary and that makes it a bunch simpler -
particularly when handling projections.

Signed-off-by: David Wood <david@davidtw.co>
4 years agolint: unify enum variant, union and struct logic
David Wood [Fri, 12 Jun 2020 13:11:45 +0000 (14:11 +0100)]
lint: unify enum variant, union and struct logic

This commit applies the changes introduced in #72890 to both enum
variants and unions - where the logic prior to #72890 was duplicated.

Signed-off-by: David Wood <david@davidtw.co>
4 years agolint: prohibit fields with opaque types
David Wood [Fri, 12 Jun 2020 12:30:39 +0000 (13:30 +0100)]
lint: prohibit fields with opaque types

Opaque types cannot be used in extern declarations, and normally cannot
exist in fields - except with type aliases to `impl Trait` and
projections which normalize to them.

Signed-off-by: David Wood <david@davidtw.co>
4 years agosave_analysis: improve handling of enum struct variant
marmeladema [Thu, 11 Jun 2020 18:08:33 +0000 (19:08 +0100)]
save_analysis: improve handling of enum struct variant

Fixes #61385

4 years agoty: projections in `transparent_newtype_field`
David Wood [Thu, 11 Jun 2020 21:53:32 +0000 (22:53 +0100)]
ty: projections in `transparent_newtype_field`

This commit modifies `transparent_newtype_field` so that it handles
projections with generic parameters, where `normalize_erasing_regions`
would ICE.

Signed-off-by: David Wood <david@davidtw.co>
4 years agocoerce reborrow multi arg test
Bastian Kauschke [Sat, 30 May 2020 18:54:14 +0000 (20:54 +0200)]
coerce reborrow multi arg test

4 years agoupdate coercion docs
Bastian Kauschke [Sat, 30 May 2020 17:43:04 +0000 (19:43 +0200)]
update coercion docs

4 years agomerge coercion test folders
Bastian Kauschke [Sat, 30 May 2020 17:42:22 +0000 (19:42 +0200)]
merge coercion test folders

4 years agoAuto merge of #73498 - RalfJung:rollup-1mfjcju, r=RalfJung
bors [Fri, 19 Jun 2020 08:34:50 +0000 (08:34 +0000)]
Auto merge of #73498 - RalfJung:rollup-1mfjcju, r=RalfJung

Rollup of 13 pull requests

Successful merges:

 - #70740 (Enabling static-pie for musl)
 - #72331 (Report error when casting an C-like enum implementing Drop)
 - #72486 (Fix asinh of negative values)
 - #72497 (tag/niche terminology cleanup)
 - #72999 (Create self-contained directory and move there some of external binaries/libs)
 - #73130 (Remove const prop for indirects)
 - #73142 (Ensure std benchmarks get tested.)
 - #73305 (Disallow loading crates with non-ascii identifier name.)
 - #73346 (Add rust specific features to print target features)
 - #73362 (Test that bounds checks are elided when slice len is checked up-front)
 - #73459 (Reduce pointer casts in Box::into_boxed_slice)
 - #73464 (Document format correction)
 - #73479 (Minor tweaks to liballoc)

Failed merges:

r? @ghost

4 years agoRollup merge of #73479 - pickfire:liballoc-spell, r=dtolnay
Ralf Jung [Fri, 19 Jun 2020 06:56:19 +0000 (08:56 +0200)]
Rollup merge of #73479 - pickfire:liballoc-spell, r=dtolnay

Minor tweaks to liballoc

4 years agoRollup merge of #73464 - qy3u:fs-document-format-correction, r=jonas-schievink
Ralf Jung [Fri, 19 Jun 2020 06:56:17 +0000 (08:56 +0200)]
Rollup merge of #73464 - qy3u:fs-document-format-correction, r=jonas-schievink

Document format correction

Minor amendments to the document.

r? @steveklabnik

4 years agoRollup merge of #73459 - cuviper:into_boxed_slice-unicast, r=dtolnay
Ralf Jung [Fri, 19 Jun 2020 06:56:15 +0000 (08:56 +0200)]
Rollup merge of #73459 - cuviper:into_boxed_slice-unicast, r=dtolnay

Reduce pointer casts in Box::into_boxed_slice

We only need to cast the pointer once to change `Box<T>` to an array
`Box<[T; 1]>`, then we can let unsized coercion return `Box<[T]>`.

4 years agoRollup merge of #73362 - erikdesjardins:bounds, r=nikomatsakis
Ralf Jung [Fri, 19 Jun 2020 06:56:13 +0000 (08:56 +0200)]
Rollup merge of #73362 - erikdesjardins:bounds, r=nikomatsakis

Test that bounds checks are elided when slice len is checked up-front

Closes #69101

4 years agoRollup merge of #73346 - aszenz:patch-1, r=cuviper
Ralf Jung [Fri, 19 Jun 2020 06:56:12 +0000 (08:56 +0200)]
Rollup merge of #73346 - aszenz:patch-1, r=cuviper

Add rust specific features to print target features

Fixes #71583

`crt-static` is a rust specific target feature that's absent from llvm feature table, adding it there so that it shows under `rustc --print target-features`.

Probably the most native implementation I could think of, would love to get feedback.

4 years agoRollup merge of #73305 - crlf0710:disallow_loading_monsters, r=petrochenkov
Ralf Jung [Fri, 19 Jun 2020 06:56:10 +0000 (08:56 +0200)]
Rollup merge of #73305 - crlf0710:disallow_loading_monsters, r=petrochenkov

Disallow loading crates with non-ascii identifier name.

This turns off external crate loading with non-ascii identifier names.

cc #55467.

4 years agoRollup merge of #73142 - ehuss:std-benches, r=dtolnay
Ralf Jung [Fri, 19 Jun 2020 06:56:08 +0000 (08:56 +0200)]
Rollup merge of #73142 - ehuss:std-benches, r=dtolnay

Ensure std benchmarks get tested.

This ensures that the std benchmarks don't break in the future. Currently they aren't compiled or tested on CI, so they can easily bitrot.  Testing a benchmark runs it with one iteration. Adding these should only add a few seconds to CI.

Closes #54176
Closes #61913

4 years agoRollup merge of #73130 - wesleywiser:remove_const_prop_for_indirects, r=oli-obk
Ralf Jung [Fri, 19 Jun 2020 06:56:06 +0000 (08:56 +0200)]
Rollup merge of #73130 - wesleywiser:remove_const_prop_for_indirects, r=oli-obk

Remove const prop for indirects

This was only used by one mir-opt test and since it causes buggy behavior under `-Zmir-opt-level=2`, it seems like we should remove it.

This was split out from #71946.

Closes #72679
Closes #72372
Closes #72285

4 years agoRollup merge of #72999 - mati865:separate-self-contained-dir, r=Mark-Simulacrum
Ralf Jung [Fri, 19 Jun 2020 06:56:04 +0000 (08:56 +0200)]
Rollup merge of #72999 - mati865:separate-self-contained-dir, r=Mark-Simulacrum

Create self-contained directory and move there some of external binaries/libs

One of the steps to reach design described in https://github.com/rust-lang/rust/issues/68887#issuecomment-633048380
This PR moves things around and allows link code to handle the new directory structure.

4 years agoRollup merge of #72497 - RalfJung:tag-term, r=oli-obk
Ralf Jung [Fri, 19 Jun 2020 06:56:02 +0000 (08:56 +0200)]
Rollup merge of #72497 - RalfJung:tag-term, r=oli-obk

tag/niche terminology cleanup

The term "discriminant" was used in two ways throughout the compiler:
* every enum variant has a corresponding discriminant, that can be given explicitly with `Variant = N`.
* that discriminant is then encoded in memory to store which variant is active -- but this encoded form of the discriminant was also often called "discriminant", even though it is conceptually quite different (e.g., it can be smaller in size, or even use niche-filling).

After discussion with @eddyb, this renames the second term to "tag". The way the tag is encoded can be either `TagEncoding::Direct` (formerly `DiscriminantKind::Tag`) or `TagEncoding::Niche` (formerly `DiscrimianntKind::Niche`).

This finally resolves some long-standing confusion I had about the handling of variant indices and discriminants, which surfaced in https://github.com/rust-lang/rust/pull/72419.

(There is also a `DiscriminantKind` type in libcore, it remains unaffected. I think this corresponds to the discriminant, not the tag, so that seems all right.)

r? @eddyb

4 years agoRollup merge of #72486 - Ralith:asinh-fix, r=dtolnay
Ralf Jung [Fri, 19 Jun 2020 06:55:59 +0000 (08:55 +0200)]
Rollup merge of #72486 - Ralith:asinh-fix, r=dtolnay

Fix asinh of negative values

Rust's current implementation of asinh has [large errors](https://www.wolframalpha.com/input/?i=arcsinh%28x%29%2C+ln%28x%2B%28x%5E2%2B1%29%5E0.5%29%2C+x+from+-67452095.07139316+to+0) in its negative range. ~These are (mostly) not numerical, but rather seem due to an incorrect implementation.~ This appears to be due to avoidable catastrophic cancellation.
[Playground before/after](https://play.rust-lang.org/?version=stable&mode=debug&edition=2018&gist=bd04ae6d86d06612e4e389a8b95d19ab).
[glibc uses](https://github.com/bminor/glibc/blob/81dca813cc35f91414731fdd0ff6b756d5e1827f/sysdeps/ieee754/dbl-64/s_asinh.c#L56) abs here.

Many thanks to @danieldeankon for finding this weird behavior, @jebrosen for diagnosing it, and @toasteater for identifying the probable implementation error!

4 years agoRollup merge of #72331 - oddg:forbid-cast-of-cenum-implementing-drop, r=matthewjasper...
Ralf Jung [Fri, 19 Jun 2020 06:55:57 +0000 (08:55 +0200)]
Rollup merge of #72331 - oddg:forbid-cast-of-cenum-implementing-drop, r=matthewjasper,nikomatsakis

Report error when casting an C-like enum implementing Drop

Following approach described in https://github.com/rust-lang/rust/issues/35941

4 years agoRollup merge of #70740 - haraldh:static-pie, r=petrochenkov
Ralf Jung [Fri, 19 Jun 2020 06:55:55 +0000 (08:55 +0200)]
Rollup merge of #70740 - haraldh:static-pie, r=petrochenkov

Enabling static-pie for musl

and make it the default for the x86_64-unknown-linux-musl target

This is a quick implementation for https://github.com/rust-lang/rust/issues/70693

Opening it as a draft PR to gather some feedback, before I put more work in it.

```console
❯ cat hello.rs
fn main() {
    println!("main = {:#x}", &main as *const _ as usize);
}

❯  /tmp/rust-musl/bin/rustc  --target x86_64-unknown-linux-musl  ~/hello.rs

❯ ldd hello
statically linked

❯ file hello
hello: ELF 64-bit LSB shared object, x86-64, version 1 (GNU/Linux), statically linked, BuildID[sha1]=fec5cdc170f503a712a63a6958691ce5ce433654, with debug_info, not stripped

❯ ./hello
main = 0x7f233ca30008

❯ ./hello
main = 0x7f9ddc529008

❯ ./hello
main = 0x7f1e5a224008

❯ ./hello
main = 0x7f4485c7c008

❯ /tmp/rust-musl/bin/rustc  --target x86_64-unknown-linux-musl  -Z print-link-args  ~/hello.rs
"cc" "-Wl,--as-needed" "-Wl,-z,noexecstack" "-Wl,--eh-frame-hdr" "-m64" "-nostdlib" "/tmp/rust-musl/lib/rustlib/x86_64-unknown-linux-musl/lib/rcrt1.o" "/tmp/rust-musl/lib/rustlib/x86_64-unknown-linux-musl/lib/crti.o" "-L" "/tmp/rust-musl/lib/rustlib/x86_64-unknown-linux-musl/lib" "hello.hello.7rcbfp3g-cgu.0.rcgu.o" "hello.hello.7rcbfp3g-cgu.1.rcgu.o" "hello.hello.7rcbfp3g-cgu.2.rcgu.o" "hello.hello.7rcbfp3g-cgu.3.rcgu.o" "hello.hello.7rcbfp3g-cgu.4.rcgu.o" "hello.hello.7rcbfp3g-cgu.5.rcgu.o" "-o" "hello" "hello.1nxjf9so94czdgcz.rcgu.o" "-Wl,--gc-sections" "-static-pie" "-Wl,-zrelro" "-Wl,-znow" "-nodefaultlibs" "-L" "/tmp/rust-musl/lib/rustlib/x86_64-unknown-linux-musl/lib" "-Wl,--start-group" "-Wl,-Bstatic" "/tmp/rust-musl/lib/rustlib/x86_64-unknown-linux-musl/lib/libstd-0f9cb7646f9e2c34.rlib" "/tmp/rust-musl/lib/rustlib/x86_64-unknown-linux-musl/lib/libpanic_unwind-ba857f2f2e4e7187.rlib" "/tmp/rust-musl/lib/rustlib/x86_64-unknown-linux-musl/lib/libhashbrown-58ba5e25bbdf9d29.rlib" "/tmp/rust-musl/lib/rustlib/x86_64-unknown-linux-musl/lib/librustc_std_workspace_alloc-886bfe43afa847dc.rlib" "/tmp/rust-musl/lib/rustlib/x86_64-unknown-linux-musl/lib/libbacktrace-fbfb8fe99f19a67b.rlib" "/tmp/rust-musl/lib/rustlib/x86_64-unknown-linux-musl/lib/libbacktrace_sys-85fa859e7d364cc9.rlib" "/tmp/rust-musl/lib/rustlib/x86_64-unknown-linux-musl/lib/librustc_demangle-07ab026cd3ec0d82.rlib" "/tmp/rust-musl/lib/rustlib/x86_64-unknown-linux-musl/lib/libunwind-a8ec5932d92ea864.rlib" "/tmp/rust-musl/lib/rustlib/x86_64-unknown-linux-musl/lib/libcfg_if-0ba4cc2f38a198d5.rlib" "/tmp/rust-musl/lib/rustlib/x86_64-unknown-linux-musl/lib/liblibc-c1bb2b3ce4f78b7c.rlib" "/tmp/rust-musl/lib/rustlib/x86_64-unknown-linux-musl/lib/liballoc-0ff673c1cf0d451a.rlib" "/tmp/rust-musl/lib/rustlib/x86_64-unknown-linux-musl/lib/librustc_std_workspace_core-c8ff2001db856926.rlib" "/tmp/rust-musl/lib/rustlib/x86_64-unknown-linux-musl/lib/libcore-2ae14177140eeca2.rlib" "-Wl,--end-group" "/tmp/rust-musl/lib/rustlib/x86_64-unknown-linux-musl/lib/libcompiler_builtins-4fd81b5ce1b08a9c.rlib" "-static" "-Wl,-Bdynamic" "/tmp/rust-musl/lib/rustlib/x86_64-unknown-linux-musl/lib/crtn.o"
```

Closes https://github.com/rust-lang/rust/issues/70693
Closes https://github.com/rust-lang/rust/issues/53968

4 years agoRemove now-redundant branch
Benjamin Saunders [Fri, 19 Jun 2020 01:52:26 +0000 (18:52 -0700)]
Remove now-redundant branch

4 years agoAuto merge of #73486 - Manishearth:rollup-11iyqpc, r=Manishearth
bors [Fri, 19 Jun 2020 01:20:49 +0000 (01:20 +0000)]
Auto merge of #73486 - Manishearth:rollup-11iyqpc, r=Manishearth

Rollup of 17 pull requests

Successful merges:

 - #70551 (Make all uses of ty::Error delay a span bug)
 - #71338 (Expand "recursive opaque type" diagnostic)
 - #71976 (Improve diagnostics for `let x += 1`)
 - #72279 (add raw_ref macros)
 - #72628 (Add tests for 'impl Default for [T; N]')
 - #72804 (Further tweak lifetime errors involving `dyn Trait` and `impl Trait` in return position)
 - #72814 (remove visit_terminator_kind from MIR visitor)
 - #72836 (Complete the std::time documentation to warn about the inconsistencies between OS)
 - #72968 (Only highlight doc search results via mouseover if mouse has moved)
 - #73034 (Export `#[inline]` fns with extern indicators)
 - #73315 (Clean up some weird command strings)
 - #73320 (Make new type param suggestion more targetted)
 - #73361 (Tweak "non-primitive cast" error)
 - #73425 (Mention functions pointers in the documentation)
 - #73428 (Fix typo in librustc_ast docs)
 - #73447 (Improve document for `Result::as_deref(_mut)` methods)
 - #73476 (Added tooltip for should_panic code examples)

Failed merges:

r? @ghost

4 years agoRefactor hir::Place
Aman Arora [Wed, 17 Jun 2020 22:13:05 +0000 (18:13 -0400)]
Refactor hir::Place

For the following code
```rust
let c = || bar(foo.x, foo.x)
```

We generate two different `hir::Place`s for both `foo.x`.
Handling this adds overhead for analysis we need to do for RFC 2229.

We also want to store type information at each Projection to support
analysis as part of the RFC. This resembles what we have for
`mir::Place`

This commit modifies the Place as follows:
- Rename to `PlaceWithHirId`, where there `hir_id` is that of the
expressioin.
- Move any other information that describes the access out to another
struct now called `Place`.
- Removed `Span`, it can be accessed using the [hir
API](https://doc.rust-lang.org/nightly/nightly-rustc/rustc_middle/hir/map/struct.Map.html#method.span)
- Modify `Projection` to be a strucutre of its own, that currently only
contains the `ProjectionKind`.

Adding type information to projections wil be completed as part of https://github.com/rust-lang/project-rfc-2229/issues/5

Closes https://github.com/rust-lang/project-rfc-2229/issues/3

Co-authored-by: Aman Arora <me@aman-arora.com>
Co-authored-by: Roxane Fruytier <roxane.fruytier@hotmail.com>
4 years agoRollup merge of #73476 - JakobDegen:should_panic_rustdoc, r=GuillaumeGomez
Manish Goregaokar [Thu, 18 Jun 2020 22:21:04 +0000 (15:21 -0700)]
Rollup merge of #73476 - JakobDegen:should_panic_rustdoc, r=GuillaumeGomez

Added tooltip for should_panic code examples

This change adds a tooltip to the documentation for `should_panic` examples. It currently displays identically to `compile_fail` examples, save for the changed text. It may be helpful to change the color that this displays in to make it visually more clear what is going on, but I'm unsure if additional colors wouldn't just be distracting.

I brought this [up on internals](https://internals.rust-lang.org/t/indicating-that-an-example-is-should-panic-in-docs/12544) a few days ago, and there seemed to be a mild positive response to it.

4 years agoRollup merge of #73447 - lzutao:stabilize-result_as_deref, r=dtolnay
Manish Goregaokar [Thu, 18 Jun 2020 22:21:03 +0000 (15:21 -0700)]
Rollup merge of #73447 - lzutao:stabilize-result_as_deref, r=dtolnay

Improve document for `Result::as_deref(_mut)` methods

cc #50264

4 years agoRollup merge of #73428 - pierwill:patch-1, r=jonas-schievink
Manish Goregaokar [Thu, 18 Jun 2020 22:21:01 +0000 (15:21 -0700)]
Rollup merge of #73428 - pierwill:patch-1, r=jonas-schievink

Fix typo in librustc_ast docs

Fixed sentence by removing a word.

4 years agoRollup merge of #73425 - poliorcetics:zeroed-functions-pointers, r=dtolnay
Manish Goregaokar [Thu, 18 Jun 2020 22:20:59 +0000 (15:20 -0700)]
Rollup merge of #73425 - poliorcetics:zeroed-functions-pointers, r=dtolnay

Mention functions pointers in the documentation

Fixes #51615.

This mentions function pointers in the documentation for `core::mem::zeroed`, adding them to the list of types that are **always** wrong when zeroed, with `&T` and `&mut T`.

@rustbot modify labels: T-doc, C-enhancement, T-libs

4 years agoRollup merge of #73361 - estebank:non-primitive-cast, r=davidtwco
Manish Goregaokar [Thu, 18 Jun 2020 22:20:57 +0000 (15:20 -0700)]
Rollup merge of #73361 - estebank:non-primitive-cast, r=davidtwco

Tweak "non-primitive cast" error

- Suggest borrowing expression if it would allow cast to work.
- Suggest using `<Type>::from(<expr>)` when appropriate.
- Minor tweak to `;` typo suggestion.

Partily address #47136.

4 years agoRollup merge of #73320 - estebank:type-param-sugg-more, r=davidtwco
Manish Goregaokar [Thu, 18 Jun 2020 22:20:55 +0000 (15:20 -0700)]
Rollup merge of #73320 - estebank:type-param-sugg-more, r=davidtwco

Make new type param suggestion more targetted

Do not suggest new type param when encountering a missing type in an ADT
field with generic parameters.

Fix #72640.

4 years agoRollup merge of #73315 - GuillaumeGomez:clean-up-config-strs, r=kinnison
Manish Goregaokar [Thu, 18 Jun 2020 22:20:53 +0000 (15:20 -0700)]
Rollup merge of #73315 - GuillaumeGomez:clean-up-config-strs, r=kinnison

Clean up some weird command strings

r? @kinnison

4 years agoRollup merge of #73034 - doctorn:nomangle-inline-linkage, r=matthewjasper
Manish Goregaokar [Thu, 18 Jun 2020 22:20:51 +0000 (15:20 -0700)]
Rollup merge of #73034 - doctorn:nomangle-inline-linkage, r=matthewjasper

Export `#[inline]` fns with extern indicators

In ancient history (#36280) we stopped `#[inline]` fns being codegened if they weren't used. However,

- #72944
- #72463

observe that when writing something like

```rust
#![crate_type = "cdylib"]

#[no_mangle]
#[inline]
pub extern "C" fn foo() {
    // ...
}
```

we really _do_ want `foo` to be codegened. This change makes this the case.

Resolves #72944, resolves #72463 (and maybe some more)

4 years agoRollup merge of #72968 - integer32llc:docs-arrow-keys, r=GuillaumeGomez
Manish Goregaokar [Thu, 18 Jun 2020 22:20:49 +0000 (15:20 -0700)]
Rollup merge of #72968 - integer32llc:docs-arrow-keys, r=GuillaumeGomez

Only highlight doc search results via mouseover if mouse has moved

## What happens

- Go to https://doc.rust-lang.org/stable/std/index.html
- Put your mouse cursor somewhere in the middle where search results will appear and then don't move the mouse
- Press 's' to focus the search box
- Type a query that brings up enough search results to go under where your mouse cursor is
- Press the down arrow
- The search result that is one below where your mouse cursor is will be highlighted.

## What I expected

When not currently using the mouse, I expect doing a search and then pressing the down arrow to always highlight the first search result immediately below the search box.

## The fix

This feels a bit hacky to me; I'm open to other solutions. This introduces a global JS var that keeps track of whether the person searching has moved their mouse after doing a search or not, and only uses the mouse position to highlight search results if the person HAS moved the mouse AFTER doing a search.

4 years agoRollup merge of #72836 - poliorcetics:std-time-os-specificities, r=shepmaster
Manish Goregaokar [Thu, 18 Jun 2020 22:20:47 +0000 (15:20 -0700)]
Rollup merge of #72836 - poliorcetics:std-time-os-specificities, r=shepmaster

Complete the std::time documentation to warn about the inconsistencies between OS

Fixes #48980.

I put the new documentation in `src/libstd/time.rs` at the module-level because it affects all types, even the one that are not directly system dependents if they are used with affected types, but there may be a better place for it.

4 years agoRollup merge of #72814 - RalfJung:mir-visir-terminator, r=oli-obk
Manish Goregaokar [Thu, 18 Jun 2020 22:20:45 +0000 (15:20 -0700)]
Rollup merge of #72814 - RalfJung:mir-visir-terminator, r=oli-obk

remove visit_terminator_kind from MIR visitor

For some reason, we had both `visit_terminator` and `visit_terminator_kind`. In contrast, for `Statement` we just have `visit_statement`. So this cleans things up by removing `visit_terminator_kind` and porting its users to `visit_terminator`.

4 years agoRollup merge of #72804 - estebank:opaque-missing-lts-in-fn-2, r=nikomatsakis
Manish Goregaokar [Thu, 18 Jun 2020 22:20:43 +0000 (15:20 -0700)]
Rollup merge of #72804 - estebank:opaque-missing-lts-in-fn-2, r=nikomatsakis

Further tweak lifetime errors involving `dyn Trait` and `impl Trait` in return position

* Suggest substituting `'static` lifetime in impl/dyn `Trait + 'static` instead of `Trait + 'static + '_`
* When `'static` is explicit, also suggest constraining argument with it
* Reduce verbosity of suggestion message and mention lifetime in label
* Tweak output for overlapping required/captured spans
* Give these errors an error code

Follow up to #72543.

r? @nikomatsakis

4 years agoRollup merge of #72628 - MikailBag:array-default-tests, r=shepmaster
Manish Goregaokar [Thu, 18 Jun 2020 22:20:41 +0000 (15:20 -0700)]
Rollup merge of #72628 - MikailBag:array-default-tests, r=shepmaster

Add tests for 'impl Default for [T; N]'

Related: #71690.
This pull request adds two tests:
- Even it T::default() panics, no leaks occur.
- [T; 0] is Default even if T is not.

I believe at some moment `Default` impl for arrays will be rewritten to use const generics instead of macros, and these tests will help to prevent behavior changes.

4 years agoRollup merge of #72279 - RalfJung:raw-ref-macros, r=nikomatsakis
Manish Goregaokar [Thu, 18 Jun 2020 22:20:39 +0000 (15:20 -0700)]
Rollup merge of #72279 - RalfJung:raw-ref-macros, r=nikomatsakis

add raw_ref macros

In https://github.com/rust-lang/rust/issues/64490, various people were in favor of exposing `&raw` as a macro first before making the actual syntax stable. So this PR (unstably) introduces those macros.

I'll create the tracking issue if we're okay moving forward with this.

4 years agoRollup merge of #71976 - mibac138:let-recovery, r=estebank
Manish Goregaokar [Thu, 18 Jun 2020 22:20:36 +0000 (15:20 -0700)]
Rollup merge of #71976 - mibac138:let-recovery, r=estebank

Improve diagnostics for `let x += 1`

Fixes(?) #66736

The code responsible for the `E0404` errors is [here](https://github.com/rust-lang/rust/blob/master/src/librustc_parse/parser/ty.rs#L399-L424) which I don't think can be easily modified to prevent emitting an error in one specific case. Because of this I couldn't get rid of `E0404` and instead added `E0067` along with a help message which will fix the problem.

r? @estebank

4 years agoRollup merge of #71338 - estebank:recursive-impl-trait, r=nikomatsakis
Manish Goregaokar [Thu, 18 Jun 2020 22:20:33 +0000 (15:20 -0700)]
Rollup merge of #71338 - estebank:recursive-impl-trait, r=nikomatsakis

Expand "recursive opaque type" diagnostic

Fix #70968, partially address #66523.

4 years agoRollup merge of #70551 - mark-i-m:ty-err-2, r=varkor
Manish Goregaokar [Thu, 18 Jun 2020 22:20:30 +0000 (15:20 -0700)]
Rollup merge of #70551 - mark-i-m:ty-err-2, r=varkor

Make all uses of ty::Error delay a span bug

r? @eddyb

A second attempt at https://github.com/rust-lang/rust/pull/70245

resolves https://github.com/rust-lang/rust/issues/70866

4 years agoAuto merge of #73446 - ecstatic-morse:issue-73431, r=pnkfelix
bors [Thu, 18 Jun 2020 21:50:45 +0000 (21:50 +0000)]
Auto merge of #73446 - ecstatic-morse:issue-73431, r=pnkfelix

Make novel structural match violations not a `bug`

Fixes (on master) #73431.

Ideally, `CustomEq` would emit a strict subset of the structural match errors that are found by `search_for_structural_match_violation`, since it allows more cases due to value-based reasoning. However, const qualification is more conservative than `search_for_structural_match_violation` around associated constants, since qualification does not try to substitute type parameters.

In the long term, we should probably make const qualification work for generic associated constants, but I don't like extending its capabilities even further.

r? @pnkfelix

4 years agoDisable core benches on wasm (benches not supported).
Eric Huss [Thu, 18 Jun 2020 21:45:19 +0000 (14:45 -0700)]
Disable core benches on wasm (benches not supported).

4 years agoreview comments: add comment
Esteban Küber [Thu, 18 Jun 2020 21:31:15 +0000 (14:31 -0700)]
review comments: add comment

4 years agoRearrange liballoc __impl_slice_eq1
Ivan Tham [Thu, 18 Jun 2020 16:34:49 +0000 (00:34 +0800)]
Rearrange liballoc __impl_slice_eq1

4 years agoLiballoc clean up macro_rules style
Ivan Tham [Thu, 18 Jun 2020 16:34:20 +0000 (00:34 +0800)]
Liballoc clean up macro_rules style

4 years agoFix liballoc doc spelling
Ivan Tham [Thu, 18 Jun 2020 16:33:24 +0000 (00:33 +0800)]
Fix liballoc doc spelling

4 years agoEnsure std benchmarks get tested.
Eric Huss [Mon, 8 Jun 2020 16:09:21 +0000 (09:09 -0700)]
Ensure std benchmarks get tested.

4 years agoRemoved trailing whitespace
Jake Degen [Thu, 18 Jun 2020 15:45:52 +0000 (11:45 -0400)]
Removed trailing whitespace

4 years agoAdded tooltip for should_panic code examples.
Jake Degen [Thu, 18 Jun 2020 12:48:37 +0000 (08:48 -0400)]
Added tooltip for should_panic code examples.

Previously, compile_fail and ignore code examples displayed a tooltip
indicating this in the documentation. This tooltip has now also been
added to should_panic examples.

4 years agoDocument format correction
qy3u [Thu, 18 Jun 2020 02:13:35 +0000 (10:13 +0800)]
Document format correction

4 years agoAuto merge of #73384 - petrochenkov:gnulink, r=cuviper
bors [Thu, 18 Jun 2020 04:29:23 +0000 (04:29 +0000)]
Auto merge of #73384 - petrochenkov:gnulink, r=cuviper

linker: Never pass `-no-pie` to non-gnu linkers

Fixes https://github.com/rust-lang/rust/issues/73370

4 years agoAdd specialization of `ToString for char`
Lzu Tao [Thu, 18 Jun 2020 02:54:35 +0000 (02:54 +0000)]
Add specialization of `ToString for char`

4 years agoImprove document for `Result::as_deref(_mut)`
Lzu Tao [Thu, 18 Jun 2020 02:09:53 +0000 (02:09 +0000)]
Improve document for `Result::as_deref(_mut)`

4 years agoImplement crate level only lints checking.
Charles Lew [Sat, 13 Jun 2020 01:58:24 +0000 (09:58 +0800)]
Implement crate level only lints checking.

4 years agoAuto merge of #73065 - Amanieu:tls-fix, r=oli-obk
bors [Thu, 18 Jun 2020 00:29:10 +0000 (00:29 +0000)]
Auto merge of #73065 - Amanieu:tls-fix, r=oli-obk

Fix link error with #[thread_local] introduced by #71192

r? @oli-obk

4 years agoRemove duplicate sentence fragment from mem::zeroed doc
David Tolnay [Thu, 18 Jun 2020 00:17:07 +0000 (17:17 -0700)]
Remove duplicate sentence fragment from mem::zeroed doc

4 years agoReduce pointer casts in Box::into_boxed_slice
Josh Stone [Wed, 17 Jun 2020 23:30:27 +0000 (16:30 -0700)]
Reduce pointer casts in Box::into_boxed_slice

We only need to cast the pointer once to change `Box<T>` to an array
`Box<[T; 1]>`, then we can let unsized coercion return `Box<[T]>`.