Extend the `SizeSkeleton` evaluator to shortcut zero-sized arrays, thus
considering `[T; 0]` to have a compile-time fixed-size of 0.

The existing evaluator already deals with generic arrays under the
feature-guard `transmute_const_generics`. However, it merely allows
comparing fixed-size types with fixed-size types, and generic types with
generic types. For generic types, it merely compares whether their
arguments match (ordering them first). Even if their exact sizes are not
known at compile time, it can ensure that they will eventually be the
same.

This patch extends this by shortcutting the size-evaluation of zero
sized arrays and thus allowing size comparisons of `()` with `[T; 0]`,
where one contains generics and the other does not.

This code is guarded by `transmute_const_generics` (rust-lang#109929), even
though it is unclear whether it should be. However, this assumes that a
separate stabilization PR is required to move this out of the feature
guard.

Initially reported in rust-lang#98104.

Co-authored-by: Mara Bos <[email protected]>

Co-authored-by: Mara Bos <[email protected]>

rust-lang#91913

Code changes in raw_vec require blessing UI tests every time

ACP: rust-lang/libs-team#156

Signed-off-by: Andrew Wock <[email protected]>

…d increase contrast by setting text color to black

On modern Linux with Intel AMX and 1KiB matrices,
Arm SVE with potentially 2KiB vectors,
and RISCV Vectors with up to 16KiB vectors,
we must handle dynamic signal stack sizes.

We can do so unconditionally by using getauxval,
but assuming it may return 0 as an answer,
thus falling back to the old constant if needed.

…igstksz, r=m-ou-se

Dynamically size sigaltstk in std

On modern Linux with Intel AMX and 1KiB matrices,
Arm SVE with potentially 2KiB vectors,
and RISCV Vectors with up to 16KiB vectors,
we must handle dynamic signal stack sizes.

We can do so unconditionally by using getauxval,
but assuming it may return 0 as an answer,
thus falling back to the old constant if needed.

Fixes rust-lang#107795

compiler: allow transmute of ZST arrays with generics

Extend the `SizeSkeleton` evaluator to shortcut zero-sized arrays, thus considering `[T; 0]` to have a compile-time fixed-size of 0.

The existing evaluator already deals with generic arrays under the feature-guard `transmute_const_generics`. However, it merely allows comparing fixed-size types with fixed-size types, and generic types with generic types. For generic types, it merely compares whether their arguments match (ordering them first). Even if their exact sizes are not known at compile time, it can ensure that they will eventually be the same.

This patch extends this by shortcutting the size-evaluation of zero sized arrays and thus allowing size comparisons of `()` with `[T; 0]`, where one contains generics and the other does not.

This code is guarded by `transmute_const_generics` (rust-lang#109929), even though it is unclear whether it should be. However, this assumes that a separate stabilization PR is required to move this out of the feature guard.

Initially reported in rust-lang#98104.

…end, r=cjgillot

Allow targets to override default codegen backend

Implements rust-lang/compiler-team#670.

Improve std::fs::read_to_string example

Resolves  [rust-lang#118621](rust-lang#118621)

For the original code to succeed it requires address.txt to contain a socketaddress, however it is much easier to follow if this is just any strong - eg address could be a street address or just text.

Also changed the variable name from "foo" to something more meaningful as cargo clippy warns you against using foo as a placeholder.

```
$ cat main.rs
use std::fs;
use std::error::Error;

fn main() -> Result<(), Box<dyn Error>> {
    let addr: String = fs::read_to_string("address.txt")?.parse()?;
    println!("{}", addr);
    Ok(())
}

$ cat address.txt
123 rusty lane
san francisco 94999

$ cargo run
    Finished dev [unoptimized + debuginfo] target(s) in 0.00s
     Running `/home/haydon/workspace/rust-test-pr/tester/target/debug/tester`
123 rusty lane
san francisco 94999

```

…nieu

Vec::try_with_capacity

Related to rust-lang#91913

Implements try_with_capacity for `Vec`, `VecDeque`, and `String`. I can follow it up with more collections if desired.

`Vec::try_with_capacity()` is functionally equivalent to the current stable:

```rust
let mut v = Vec::new();
v.try_reserve_exact(n)?
```

However, `try_reserve` calls non-inlined `finish_grow`, which requires old and new `Layout`, and is designed to reallocate memory. There is benefit to using `try_with_capacity`, besides syntax convenience, because it generates much smaller code at the call site with a direct call to the allocator. There's codegen test included.

It's also a very desirable functionality for users of `no_global_oom_handling` (Rust-for-Linux), since it makes a very commonly used function available in that environment (`with_capacity` is used much more frequently than all `(try_)reserve(_exact)`).

…enkov

Remove `feed_local_def_id`

best reviewed commit by commit

Basically I returned `TyCtxtFeed` from `create_def` and then preserved that in the local caches

based on rust-lang#121084

r? ```@petrochenkov```

Implement MaybeUninit::fill{,_with,_from}

ACP: rust-lang/libs-team#156

…, r=notriddle

Add missing background color for top-level rust documentation page and increase contrast by setting text color to black

Fixes rust-lang#121954.

r? `@notriddle`

Rust is a proper name: rust → Rust

I only went over the library where it might be user-visible -- I noticed this in the `time` docs.

Make `x t miri` respect `MIRI_TEMP`

(I don't want to override `TMPDIR`, as that might affect other things)

r? `@RalfJung`

…affleLapkin

Make not finding core a fatal error

Similar to rust-lang#120472, this prevents terminal spam. In particular, it makes the good diagnostic visible when you try to use a target that's not installed.

…ler-errors

Cancel parsing ever made during recovery

Fixes rust-lang#122112

It would be nice if diagnostics from recovery were automatically cancelled... 🤔

…ld-belongs-to-unix, r=ChrisDenton

Fix `tidy --bless` on  ̶X̶e̶n̶i̶x̶ Windows

As reported in rust-lang#120628 (comment) the requested `tidy --bless` implementation didn't take into account the fact that earlier the linting code canonicalized things to use the OS path separator. This makes it so that the path separator is always rewritten back as '/', which should fix the variance there.

r? `@ChrisDenton`