optimize str.replace
Adds a fast path for str.replace for the ascii to ascii case. This allows for autovectorizing the code. Also should this instead be done with specialization? This way we could remove one branch. I think it is the kind of branch that is easy to predict though.
Benchmark for the fast path (replace all "a" with "b" in the rust wikipedia article, using criterion) :
| N | Speedup | Time New (ns) | Time Old (ns) |
|----------|---------|---------------|---------------|
| 2 | 2.03 | 13.567 | 27.576 |
| 8 | 1.73 | 17.478 | 30.259 |
| 11 | 2.46 | 18.296 | 45.055 |
| 16 | 2.71 | 17.181 | 46.526 |
| 37 | 4.43 | 18.526 | 81.997 |
| 64 | 8.54 | 18.670 | 159.470 |
| 200 | 9.82 | 29.634 | 291.010 |
| 2000 | 24.34 | 81.114 | 1974.300 |
| 20000 | 30.61 | 598.520 | 18318.000 |
| 1000000 | 29.31 | 33458.000 | 980540.000 |
Rollup of 7 pull requests
Successful merges:
- #128391 (Change orphan hint from "only" to "any uncovered type inside...")
- #131583 (Setting up indirect access to external data for loongarch64-linux-{musl,ohos})
- #131595 (rustdoc-JSON: Rename "object safe" to "dyn compatible")
- #131748 (cleanup canonical queries)
- #131798 (llvm: Tolerate propagated range metadata)
- #131815 (compiler: use `is_none_or` where it is clearly better)
- #131822 (extract `expr_assign_expected_bool_error`)
r? `@ghost`
`@rustbot` modify labels: rollup
compiler: use `is_none_or` where it is clearly better
heuristic was: if it easily allows removing bangs entirely? worth it. if it requires more effort or just moves the bang? not.
llvm: Tolerate propagated range metadata
llvm/llvm-project#91101 propagates range information across inlining, resulting in more metadata in this test. Tolerate the range metadata if it appears.
``@rustbot:`` label +llvm-main
r? ``@durin42``
Please wait a moment before approving, putting the llvm-main tag on it to make sure it fixes the integration test.
cleanup canonical queries
best reviewed commit by commit. adding `CanonicalQueryInput` to stop returning `defining_opaque_types` in query responses is the most involved change here.
r? ``@compiler-errors``
rustdoc-JSON: Rename "object safe" to "dyn compatible"
~~Blocked: Sits atop #131594. Only the last commit is relevant.~~ (rebased)
Part of #130852.
r? aDotInTheVoid or rustdoc
Setting up indirect access to external data for loongarch64-linux-{musl,ohos}
In issue #118053, the `loongarch64-unknown-linux-gnu` target needs indirection to access external data, and so do the `loongarch64-unknown-linux-musl` and `loongarch64-unknown-linux-ohos` targets.
Make destructors on `extern "C"` frames to be executed
This would make the example in #123231 print "Noisy Drop". I didn't mark this as fixing the issue because the behaviour is yet to be spec'ed.
Tracking:
- https://github.com/rust-lang/rust/issues/74990
internal: Use local time when formatting logs
When debugging rust-analyzer and looking at logs, it's much easier to read when the timestamp is in the local timezone.
Before:
2024-08-28T20:55:38.792321Z INFO ParseQuery: invoked at R18460
After:
2024-08-28T13:55:38.792321-07:00 INFO ParseQuery: invoked at R18460
When debugging rust-analyzer and looking at logs, it's much easier to read
when the timestamp is in the local timezone.
Before:
2024-08-28T20:55:38.792321Z INFO ParseQuery: invoked at R18460
After:
2024-08-28T13:55:38.792321-07:00 INFO ParseQuery: invoked at R18460
llvm/llvm-project#91101 propagates range information across inlining,
resulting in more metadata in this test. Tolerate the range metadata if
it appears.
Fix needless_lifetimes in stable_mir
Hi,
This PR fixes the following clippy warning
```
warning: the following explicit lifetimes could be elided: 'a
--> compiler/stable_mir/src/mir/visit.rs:490:6
|
490 | impl<'a> PlaceRef<'a> {
| ^^ ^^
|
= help: for further information visit https://rust-lang.github.io/rust-clippy/master/index.html#needless_lifetimes
= note: `#[warn(clippy::needless_lifetimes)]` on by default
help: elide the lifetimes
|
490 - impl<'a> PlaceRef<'a> {
490 + impl PlaceRef<'_> {
|
```
Best regards,
Michal
Fix trivially_copy_pass_by_ref in stable_mir
Hi,
This PR fixes the following clippy warnings
```
warning: this argument (8 byte) is passed by reference, but would be more efficient if passed by value (limit: 8 byte)
--> compiler/stable_mir/src/mir/body.rs:1042:34
|
1042 | fn subslice_ty(ty: Ty, from: &u64, to: &u64, from_end: &bool) -> Result<Ty, Error> {
| ^^^^ help: consider passing by value instead: `u64`
|
= help: for further information visit https://rust-lang.github.io/rust-clippy/master/index.html#trivially_copy_pass_by_ref
= note: requested on the command line with `-W clippy::trivially-copy-pass-by-ref`
warning: this argument (8 byte) is passed by reference, but would be more efficient if passed by value (limit: 8 byte)
--> compiler/stable_mir/src/mir/body.rs:1042:44
|
1042 | fn subslice_ty(ty: Ty, from: &u64, to: &u64, from_end: &bool) -> Result<Ty, Error> {
| ^^^^ help: consider passing by value instead: `u64`
|
= help: for further information visit https://rust-lang.github.io/rust-clippy/master/index.html#trivially_copy_pass_by_ref
warning: this argument (1 byte) is passed by reference, but would be more efficient if passed by value (limit: 8 byte)
--> compiler/stable_mir/src/mir/body.rs:1042:60
|
1042 | fn subslice_ty(ty: Ty, from: &u64, to: &u64, from_end: &bool) -> Result<Ty, Error> {
| ^^^^^ help: consider passing by value instead: `bool`
|
= help: for further information visit https://rust-lang.github.io/rust-clippy/master/index.html#trivially_copy_pass_by_ref
```
Best regards,
Michal
Emscripten: Xfail backtrace ui tests
It is possible to link libunwind and use the normal backtrace code, but it fails to symbolize stack traces. I investigated and could get the list of instruction pointers and symbol names, but I'm not sure how to use the dwarf info to map from instruction pointer to source location. In any case, fixing this is not a high priority.
See https://github.com/rust-lang/rust/issues/131738
r?jieyouxu
Rename `can_coerce` to `may_coerce`, and then structurally resolve correctly in the probe
We need to structurally resolve the lhs and rhs of the coercion. Also, renaming the method so it's less ambiguous about what it's doing... the word "may" gives more clear signal that it has false positives imo.
r? lcnr
Refactor some `core::fmt` macros
While looking at the macros in `core::fmt`, find that the macros are not well organized. So I created a patch to fix it.
[`core/src/fmt/num.rs`](https://github.com/rust-lang/rust/blob/master/library/core/src/fmt/num.rs)
* `impl_int!` and `impl_uint!` macro are **completly** same. It would be better to combine for readability
* `impl_int!` has a problem that the indenting is not uniform. It has unified into 4 spaces
* `debug` macro in `num` renamed to `impl_Debug`, And it was moved to a position close to the `impl_Display`.
[`core/src/fmt/float.rs`](https://github.com/rust-lang/rust/blob/master/library/core/src/fmt/float.rs)
[`core/src/fmt/nofloat.rs`](https://github.com/rust-lang/rust/blob/master/library/core/src/fmt/nofloat.rs)
* `floating` macro now receive multiple idents at once. It makes the code cleaner.
* Modified the panic message more clearly in fallback function of `cfg(no_fp_fmt_parse)`
Delay ambiguous intra-doc link resolution after `Cache` has been populated
Fixes https://github.com/rust-lang/rust/issues/130233.
I was getting nowhere with #130278. I took a wrong turn at some point and ended making way too many changes so instead I started again back from 0 and this time it worked out as expected.
r? ```@notriddle```
Rustfmt `for<'a> async` correctly
In #127054, we decided to move the trait bound modifier for `async for<'a> Fn()` to `for<'a> async Fn()`. This wasn't adjusted in rustfmt, so this PR implements that. It also requires consolidating the bound formatting into the `Rewrite` impl for `PolyTraitRef`.
Fixes#131649
Don't check unsize goal in MIR validation when opaques remain
Similarly to `mir_assign_valid_types`, let's just skip when there are opaques. Fixes#130921.
Fix explicit_iter_loop in rustc_serialize
Hi,
This PR fixes some clippy warnings
```
warning: it is more concise to loop over references to containers instead of using explicit iteration methods
--> compiler/rustc_serialize/src/serialize.rs:675:18
|
675 | for e in self.iter() {
| ^^^^^^^^^^^ help: to write this more concisely, try: `self`
|
= help: for further information visit https://rust-lang.github.io/rust-clippy/master/index.html#explicit_iter_loop
```
Best regards,
Michal
Ignore lint-non-snake-case-crate#proc_macro_ on targets without unwind
The lint-non-snake-case-crate test may emit a warning in stderr if the target does not support unwinding
```
warning: building proc macro crate with `panic=abort` may crash the compiler should the proc-macro panic
```
Consequently, the test will fail on targets that don't support unwinding as written.
This change modifies the expected stderr for lint-non-snake-case-crate in the proc_macro_ to ignore lines that indicate a warning was emitted.
Try to improve error messages involving aliases in the solver
1. Treat aliases as rigid only if it may not be defined and it's well formed (i.e. for projections, its trait goal is satisfied).
2. Record goals that are related to alias normalization under a new `GoalKind`, so we can look into them in the `BestObligation` visitor.
3. Try to deduplicate errors due to self types of goals that are un-normalizable aliases.
r? lcnr
Add fast-path when computing the default visibility
This PR adds (or more correctly re-adds the) fast-path when computing the default visibility, by taking advantage of the fact that the "interposable" requested visibility always return the "default" codegen visibility.
Should address the small regression observed in https://github.com/rust-lang/rust/pull/131111#issuecomment-2402273967.
r? `@lqd`
Implement edition 2024 match ergonomics restrictions
This implements the minimalest version of [match ergonomics for edition 2024](https://rust-lang.github.io/rfcs/3627-match-ergonomics-2024.html). This minimal version makes it an error to ever reset the default binding mode. The implemented proposal is described precisely [here](https://hackmd.io/zUqs2ISNQ0Wrnxsa9nhD0Q#RFC-3627-nano), where it is called "RFC 3627-nano".
Rules:
- Rule 1C: When the DBM (default binding mode) is not `move` (whether or not behind a reference), writing `mut`, `ref`, or `ref mut` on a binding is an error.
- Rule 2C: Reference patterns can only match against references in the scrutinee when the DBM is `move`.
This minimal version is forward-compatible with the main proposals for match ergonomics 2024: [RFC3627](https://rust-lang.github.io/rfcs/3627-match-ergonomics-2024.html) itself, the alternative [rule 4-early variant](https://rust-lang.github.io/rfcs/3627-match-ergonomics-2024.html), and [others](https://hackmd.io/zUqs2ISNQ0Wrnxsa9nhD0Q). The idea is to give us more time to iron out a final proposal.
This includes a migration lint that desugars any offending pattern into one that doesn't make use of match ergonomics. Such patterns have identical meaning across editions.
This PR insta-stabilizes the proposed behavior onto edition 2024.
r? `@ghost`
Tracking:
- https://github.com/rust-lang/rust/issues/123076
