For the given code paths, the amount of space used in the previous chunk
is irrelevant.

(This will almost never make a difference to behaviour, but it makes the
code clearer.)

`DroplessArena` and `TypedArena` use an aggressive growth strategy: the
first chunk is 4 KiB, the second is 8 KiB, and it keeps on doubling
indefinitely. DHAT profiles show that sometimes this results in large
chunks (e.g. 16-128 MiB) that are barely filled. Although these don't
contribute to RSS, they clog up the DHAT profiles.

This commit changes things so that the doubling stops at 2 MiB. This is
large enough that chunk allocations are still rare (you might get 100s
instead of 10s of them) but avoids lots of unused space in the worst
case. It gives a slight speed-up to cycle counts in some cases.

Add explicit references to the BuildHasher trait

Fixes rust-lang#71652

…rowth, r=oli-obk

Be less aggressive with `DroplessArena`/`TypedArena` growth.

`DroplessArena` and `TypedArena` use an aggressive growth strategy: the first chunk is 4 KiB, the second is 8 KiB, and it keeps on doubling indefinitely. DHAT profiles show that sometimes this results in large chunks (e.g. 16-128 MiB) that are barely filled.

This commit changes things so that the doubling stops at 2 MiB. This is large enough that chunk allocations are still rare (you might get 100s instead of 10s of them) but avoids lots of unused space in the worst case. It makes the same change to `TypedArena`, too.

…oli-obk

rustc_driver: factor out computing the exit code

In a recent Miri PR I [added a convenience wrapper](https://github.com/rust-lang/miri/pull/1405/files#diff-c3d602c5c8035a16699ce9c015bfeceaR125) around `catch_fatal_errors` and `run_compiler` that @oli-obk suggested I could upstream. However, after seeing what could be shared between `rustc_driver::main`, clippy and Miri, really the only thing I found is computing the exit code -- so that's what this PR does.

What prevents using the Miri convenience function in `rustc_driver::main` and clippy is that they do extra work inside `catch_fatal_errors`, and while I could abstract that away, clippy actually *computes the callbacks* inside there, and I fond no good way to abstract that and thus gave up. Maybe the clippy thing could be moved out, I am not sure if it ever can actually raise a `FatalErrorMarker` -- someone more knowledgeable in clippy would have to do that.

…Dylan-DPC

remove extra space from crate-level doctest names

Before:

```
running 2 tests
test src/test/rustdoc-ui/doctest-output.rs - foo::bar (line 11) ... ok
test src/test/rustdoc-ui/doctest-output.rs -  (line 5) ... ok

test result: ok. 2 passed; 0 failed; 0 ignored; 0 measured; 0 filtered out
```

After:

```
running 2 tests
test src/test/rustdoc-ui/doctest-output.rs - foo::bar (line 11) ... ok
test src/test/rustdoc-ui/doctest-output.rs - (line 5) ... ok

test result: ok. 2 passed; 0 failed; 0 ignored; 0 measured; 0 filtered out
```

Remove `lang_items\(\).*\.unwrap\(\)`

Follows up rust-lang#72170 to remove the remaining uses of `lang_items\(\).*\.unwrap\(\)` (avoids a bunch of potential ICEs when working in `#![no_core]`).

Resolves rust-lang#72195

doc: add links to rotate_(left|right)