improve case conversion happy path
Someone shared the source code for [Go's string case conversion](19156a5474/src/strings/strings.go (L558-L616)).
It features a hot path for ascii-only strings (although I assume for reasons specific to go, they've opted for a read safe hot loop).
I've borrowed these ideas and also kept our existing code to provide a fast path + seamless utf-8 correct path fallback.
(Naive) Benchmarks can be found here https://github.com/conradludgate/case-conv
For the cases where non-ascii is found near the start, the performance of this algorithm does fall back to original speeds and has not had any measurable speed loss
omit `record_accesses` function when collecting `MonoItem`s
This PR fixes the FIXME in the impl of `record_accesses` function.
[Edit] We can call `instantiation_mode` when push the `MonoItem` into `neighbors`. This avoids extra local variables `accesses: SmallVec<[_; 128]>`
Only allow `compiletest` to use `feature(test)`, not any other feature
Using language features occasionally causes issues when using nightly to bootstrap, rather than beta.
See #59264 for additional context.
`imported_source_files` adjusts lots of file positions, and then calls
`new_imported_source_file`, which then adjust them all again. This
commit combines the two adjustments into one, for a small perf win.
- Remove download-rustc handling from bootstrap.py
- Allow a custom `pattern` in `builder.unpack()`
- Only download rustc once another part of bootstrap depends on it.
This is somewhat necessary since the download functions rely on having a full
`Builder`, which isn't available until after config parsing finishes.
Simplify implementation of `-Z gcc-ld`
- The logic is now unified for all targets (wasm targets should also be supported now)
- Additional "symlink" files like `ld64` are eliminated
- lld-wrapper is used for propagating the correct lld flavor
- Cleanup "unwrap or exit" logic in lld-wrapper
cc https://github.com/rust-lang/rust/issues/97352
r? `@bjorn3`
- The logic is now unified for all targets (wasm targets should also be supported now)
- Additional "symlink" files like `ld64` are eliminated
- lld-wrapper is used for propagating the correct lld flavor
- Cleanup "unwrap or exit" logic in lld-wrapper
Update cargo
10 commits in a4c1cd0eb6b18082a7e693f5a665548fe1534be4..39ad1039d9e3e1746177bf5d134af4c164f95528
2022-05-20 00:55:25 +0000 to 2022-05-25 00:50:02 +0000
* doc: discuss build script instruction order (rust-lang/cargo#10600)
* Require http-registry URLs to end with a '/' (rust-lang/cargo#10698)
* No printing executable names when running tests and benchmarks with json message format (rust-lang/cargo#10691)
* Restore proper error for crate not in local reg (rust-lang/cargo#10683)
* Update libcurl (rust-lang/cargo#10696)
* Fixed small typos (rust-lang/cargo#10693)
* fix bugs with `workspace` key and `update_toml` (rust-lang/cargo#10685)
* Bump to 0.64.0, update changelog (rust-lang/cargo#10687)
* List C compiler as a build dependency in README (rust-lang/cargo#10678)
* Add unstable `rustc-check-cfg` build script output (rust-lang/cargo#10539)
r? `@ehuss`
Rollup of 5 pull requests
Successful merges:
- #97302 (Do writeback of Closure params before visiting the parent expression)
- #97328 (rustc: Fix ICE in native library error reporting)
- #97351 (Output correct type responsible for structural match violation)
- #97398 (Add regression test for #82830)
- #97400 (Fix a typo on Struct `Substructure`)
Failed merges:
r? `@ghost`
`@rustbot` modify labels: rollup
Output correct type responsible for structural match violation
Previously we included the outermost type that caused a structural match violation in the error message and stated that that type must be annotated with `#[derive(Eq, PartialEq)]` even if it already had that annotation. This PR outputs the correct type in the error message.
Fixes https://github.com/rust-lang/rust/issues/97278
Do writeback of Closure params before visiting the parent expression
This means that given the expression:
```
let x = |a: Vec<_>| {};
```
We will visit the HIR node for `a` before `x`, and report the ambiguity on the former instead of the latter. This also moves writeback for struct field ids and const blocks before, but the ordering of this and walking the expr doesn't seem to matter.
This is a >= condition, so we need a maximum size of 7 to not
create a stack protector in basic mode.
The reason this still worked is that the alloca type was converted
into an integer (rather than an array). The way these heuristics
are implemented in LLVM is rather questionable and not resilient
to optimization.
