support non-defining uses of opaques in borrowck
Reimplements the first part of rust-lang/rust#139587, but limited to only the new solver. To do so I also entirely rewrite the way we handle opaque types in borrowck even on stable. This should not impact behavior however.
We now support revealing uses during MIR borrowck with the new solver:
```rust
fn foo<'a>(x: &'a u32) -> impl Sized + use<'a> {
let local = 1;
foo::<'_>(&local);
x
}
```
### How do opaque types work right now
Whenever we use an opaque type during type checking, we remember this use in the `opaque_type_storage` of the `InferCtxt`.
Right now, we collect all *member constraints* at the end of MIR type check by looking at all uses from the `opaque_type_storage`. We then apply these constraints while computing the region values for each SCC. This does not add actual region constraints but directly updates the final region values.
This means we need to manually handle any constraints from member constraints for diagnostics. We do this by separately tracking `applied_member_constraints` in the `RegionInferenceContext`.
After we've finished computing the region values, it is now immutable and we check whether all member constraints hold. If not, we error.
We now map the hidden types of our defining uses to the defining scope. This assumes that all member constraints apply. To handle non-member regions, we simply map any region in the hidden type we fail to map to a choice region to `'erased` b1b26b834d/compiler/rustc_borrowck/src/region_infer/opaque_types.rs (L126-L132)
### How do we handle opaque types with this PR
MIR type check still works the same by populating the `opaque_type_storage` whenever we use an opaque type.
We now have a new step `fn handle_opaque_type_uses` which happens between MIR type check and `compute_regions`.
This step looks at all opaque type uses in the storage and first checks whether they are defining: are the arguments of the `opaque_type_key` unique region parameters. *With the new solver we silently ignore any *non-defining* uses here. The old solver emits an errors.*
`fn compute_concrete_opaque_types`: We then collect all member constraints for the defining uses and apply them just like we did before. However, we do it on a temporary region graph which is only used while computing the concrete opaque types. We then use this region graph to compute the concrete type which we then store in the `root_cx`.
`fn apply_computed_concrete_opaque_types`: Now that we know the final concrete type of each opaque type and have mapped them to the definition of the opaque. We iterate over all opaque type uses and equate their hidden type with the instantiated final concrete type. This is the step which actually mutates the region graph.
The actual region checking can now entirely ignores opaque types (outside of the `ConstraintCategory` from checking the opaque type uses).
### Diagnostics issue (chill)
Because we now simply use type equality to "apply member constraints" we get ordinary `OutlivesConstraint`s, even if the regions were already related to another.
This is generally not an issue, expect that it can *hide* the actual region constraints which resulted in the final value of the opaque. The constraints we get from checking against the final opaque type definition relies on constraints we used to compute that definition.
I mostly handle this by updating `find_constraint_path_between_regions` to first ignore member constraints in its search and only if that does not find a path, retry while considering member constraints.
### Diagnostics issue (not chill)
A separate issue is that `find_constraint_paths_between_regions` currently looks up member constraints by their **scc**, not by region value:
2c1ac85679/compiler/rustc_borrowck/src/region_infer/mod.rs (L1768-L1775)
This means that in the `borrowck-4` test, the resulting constraint path is currently
```
('?2: '?5) due to Single(bb0[5]) (None) (Boring) (ConstraintSccIndex(1): ConstraintSccIndex(1)),
('?5: '?3) due to Single(bb0[6]) (None) (Boring) (ConstraintSccIndex(1): ConstraintSccIndex(1)),
('?3: '?0) due to All(src/main.rs:15:5: 15:6 (#0)) (None) (OpaqueType) (ConstraintSccIndex(1): ConstraintSccIndex(1))
```
Here `'?3` is equal to `'?4`, but the reason why it's in the opaque is that it's related to `'?4`. With my PR this will be correctly tracked so we end up with
```
('?2: '?5) due to Single(bb0[5]) (None) (Boring) (ConstraintSccIndex(1): ConstraintSccIndex(1)),
('?5: '?3) due to Single(bb0[6]) (None) (Boring) (ConstraintSccIndex(1): ConstraintSccIndex(1)),
('?3: '?4) due to Single(bb0[6]) (None) (Assignment) (ConstraintSccIndex(1): ConstraintSccIndex(1)),
('?4: '?0) due to All(src/main.rs:15:5: 15:6 (#0)) (None) (OpaqueType) (ConstraintSccIndex(1): ConstraintSccIndex(1)),
```
This additional `Assignment` step then worsens the error message as we stop talking about the fact that the closures is returned from the function. Fixing this is hard. I've looked into this and it's making me sad :< Properly handling this requires some deeper changes to MIR borrowck diagnostics and that seems like too much for this PR. Given that this only impacts a single test, it seems acceptable to me.
r? `@ghost`
Tell LLVM about read-only captures
`&Freeze` parameters are not only `readonly` within the function, but any captures of the pointer can also only be used for reads. This can now be encoded using the `captures(address, read_provenance)` attribute.
During `let` binding parse error and encountering a block, detect if there is a likely missing `if` or `else`:
```
error: expected one of `.`, `;`, `?`, `else`, or an operator, found `{`
--> $DIR/missing-if-let-or-let-else.rs:14:25
|
LL | let Some(x) = foo() {
| ^ expected one of `.`, `;`, `?`, `else`, or an operator
|
help: you might have meant to use `if let`
|
LL | if let Some(x) = foo() {
| ++
help: alternatively, you might have meant to use `let else`
|
LL | let Some(x) = foo() else {
| ++++
```
The capture of i in assert_ne!() is now known read-only, which
enables early SROA. Block this by passing i to println, where
we currently cannot recognize this.
`&Freeze` parameters are not only `readonly` within the function,
but any captures of the pointer can also only be used for reads.
This can now be encoded using the `captures(address, read_provenance)`
attribute.
add a fallback implementation for the `prefetch_*` intrinsics
related ACP: https://github.com/rust-lang/libs-team/issues/638
The fallback is to just ignore the arguments. That is a valid implementation because this intrinsic is just a hint.
I also added the `miri::intrinsic_fallback_is_spec` annotation, so that miri now supports these operations. A prefetch intrinsic call is valid on any pointer. (specifically LLVM guarantees this https://llvm.org/docs/LangRef.html#llvm-prefetch-intrinsic)
Next, I made the `LOCALITY` argument a const generic. That argument must be const (otherwise LLVM crashes), but that was not reflected in the type.
Finally, with these changes, the intrinsic can be safe and `const` (a prefetch at const evaluation time is just a no-op).
cc `@Amanieu`
r? `@RalfJung`
Pretty print the name of an future from calling async closure
Fixes https://github.com/rust-lang/rust/issues/145606 by introducing a way to customize the path rendering of async closures' futures in the pretty printer API.
mention lint group in default level lint note
### Summary
This PR updates lint diagnostics so that default-level notes now mention the lint group they belong to, if any.
Fixes: rust-lang/rust#65464.
### Example
```rust
fn main() {
let x = 5;
}
```
Before:
```
= note: `#[warn(unused_variables)]` on by default
```
After:
```
= note: `#[warn(unused_variables)]` (part of `#[warn(unused)]`) on by default
```
### Unchanged Cases
Messages remain the same when the lint level is explicitly set, e.g.:
* Attribute on the lint `#[warn(unused_variables)]`:
```
note: the lint level is defined here
LL | #[warn(unused_variables)]
| ^^^^^^^^^^^^^^^^
```
* Attribute on the group `#[warn(unused)]:`:
```
= note: `#[warn(unused_variables)]` implied by `#[warn(unused)]`
```
* CLI option `-W unused`:
```
= note: `-W unused-variables` implied by `-W unused`
= help: to override `-W unused` add `#[allow(unused_variables)]`
```
* CLI option `-W unused-variables`:
```
= note: requested on the command line with `-W unused-variables`
```
ignore frontmatters in `TokenStream::new`
Fixesrust-lang/rust#145520 for now, we'd likely want to figure the stripping part later, so I noted it down on the list on the tracking issue.
cc `@fmease`
Do not consider a `T: !Sized` candidate to satisfy a `T: !MetaSized` obligation.
This example should fail to compile (and does under this PR, with the old and new solvers), but currently compiles successfully ([playground](https://play.rust-lang.org/?version=nightly&mode=debug&edition=2024&gist=6e0e5d0ae0cdf0571dea97938fb4a86d)), because (IIUC) the old solver's `lazily_elaborate_sizedness_candidate`/callers and the new solver's `TraitPredicate::fast_reject_assumption`/`match_assumption` consider a `T: _ Sized` candidate to satisfy a `T: _ MetaSized` obligation, for either polarity `_`, when that should only hold for positive polarity.
```rs
#![feature(negative_bounds)]
#![feature(sized_hierarchy)]
use std::marker::MetaSized;
fn foo<T: !MetaSized>() {}
fn bar<T: !Sized + MetaSized>() {
foo::<T>();
//~^ ERROR the trait bound `T: !MetaSized` is not satisfied // error under this PR
}
```
Only observable with the internal-only `feature(negative_bounds)`, so might just be "wontfix".
This example is added as a test in this PR (as well as testing that `foo<()>` and `foo<str>` are disallowed for `fn foo<T: !MetaSized`).
cc `@davidtwco` for `feature(sized_hierarchy)`
Maybe similar to 91c53c9 from <https://github.com/rust-lang/rust/pull/143307>
Port must_use to the new target checking
This PR ports `must_use` to the new target checking logic
This also adds a tool-only suggestion to remove attributes on invalid targets, as to not immediately undo the work of https://github.com/rust-lang/rust/pull/145274
r? `@jdonszelmann`
Properly recover from parenthesized use-bounds (precise capturing lists) plus small cleanups
Fixes https://github.com/rust-lang/rust/issues/145470.
First commit fixes the issue, second one performs some desperately needed cleanups.
The fix shouldn't be a breaking change because IINM the parser always ensures that all brackets are balanced (via a buffer of brackets). Meaning even though we used to accept `(use<>` as a valid precise capturing list, it was guaranteed that we would fail in the end.
Rehome 37 `tests/ui/issues/` tests to other subdirectories under `tests/ui/`
Part of rust-lang/rust#133895
Methodology:
1. Refer to the previously written `tests/ui/SUMMARY.md`
2. Find an appropriate category for the test, using the original issue thread and the test contents.
3. Add the issue URL at the bottom (not at the top, as that would mess up stderr line numbers)
4. Rename the tests to make their purpose clearer
Inspired by the methodology that ``@Kivooeo`` was using.
r? ``@jieyouxu``
nll-relate: improve hr opaque types support
This should currently not be user-facing outside of diagnostics as even if we successfully relate the opaque types, we don't support opaque types with non-param arguments and also require all member regions to be equal to the arguments or `'static`. This means there's no way to end up with a placeholder in the hidden type.
r? types
Add `-Zindirect-branch-cs-prefix`
Cc: ``@azhogin`` ``@Darksonn``
This goes on top of https://github.com/rust-lang/rust/pull/135927, i.e. please skip the first commit here. Please feel free to inherit it there.
In fact, I am not sure if there is any use case for the flag without `-Zretpoline*`. GCC and Clang allow it, though.
There is a `FIXME` for two `ignore`s in the test that I took from another test I did in the past -- they may be needed or not here since I didn't run the full CI. Either way, it is not critical.
Tracking issue: https://github.com/rust-lang/rust/issues/116852.
MCP: https://github.com/rust-lang/compiler-team/issues/868.
Extend `QueryStability` to handle `IntoIterator` implementations
This PR extends the `rustc::potential_query_instability` lint to check values passed as `IntoIterator` implementations.
Full disclosure: I want the lint to warn about this line (please see #138871 for why): aa8f0fd716/src/librustdoc/json/mod.rs (L261)
However, the lint warns about several other lines as well.
Final note: the functions `get_callee_generic_args_and_args` and `get_input_traits_and_projections` were copied directly from [Clippy's source code](4fd8c04da0/src/tools/clippy/clippy_lints/src/methods/unnecessary_to_owned.rs (L445-L496)).
Detect missing `derive` on unresolved attribute even when not imported
When encountering unresolved attributes, ensure the proc-macros for every crate in scope are added to the `macro_map` so that typos and missing `derive`s are properly detected.
```
error: cannot find attribute `sede` in this scope
--> $DIR/missing-derive-3.rs:20:7
|
LL | #[sede(untagged)]
| ^^^^
|
help: the derive macros `Deserialize` and `Serialize` accept the similarly named `serde` attribute
|
LL | #[serde(untagged)]
| +
error: cannot find attribute `serde` in this scope
--> $DIR/missing-derive-3.rs:14:7
|
LL | #[serde(untagged)]
| ^^^^^
|
note: `serde` is imported here, but it is a crate, not an attribute
--> $DIR/missing-derive-3.rs:4:1
|
LL | extern crate serde;
| ^^^^^^^^^^^^^^^^^^^
help: `serde` is an attribute that can be used by the derive macros `Deserialize` and `Serialize`, you might be missing a `derive` attribute
|
LL + #[derive(Deserialize, Serialize)]
LL | enum B {
|
```
Follow up to rust-lang/rust#134841. Fixrust-lang/rust#47608.
Using an error type instead of `()` avoids the duplicated errors
on `struct SUnsizedField` in `deriving-from-wrong-target.rs`. It also
improves the expanded output from this:
```
struct S2(u32, u32);
impl ::core::convert::From<()> for S2 {
#[inline]
fn from(value: ()) -> S2 { (/*ERROR*/) }
}
```
to this:
```
struct S2(u32, u32);
impl ::core::convert::From<(/*ERROR*/)> for S2 {
#[inline]
fn from(value: (/*ERROR*/)) -> S2 { (/*ERROR*/) }
}
```
The new code also only matchs on `item.kind` once.
Remove the `From` derive macro from prelude
The new `#[derive(From)]` functionality (implemented in https://github.com/rust-lang/rust/pull/144922) caused name resolution ambiguity issues (https://github.com/rust-lang/rust/issues/145524). The reproducer looks e.g. like this:
```rust
mod foo {
pub use derive_more::From;
}
use foo::*;
#[derive(From)] // ERROR: `From` is ambiguous
struct S(u32);
```
It's pretty unfortunate that it works like this, but I guess that there's not much to be done here, and we'll have to wait for the next edition to put the `From` macro into the prelude. That will probably require https://github.com/rust-lang/rust/pull/139493 to land.
I created a new module in core (and re-exported it in std) called `from`, where I re-exported the `From` macro. I *think* that since this is a new module, it should not have the same backwards incompatibility issue.
Happy to hear suggestions about the naming - maybe it would make sense as `core::macros::from::From`? But we already had a precedent in the `core::assert_matches` module, so I just followed suit.
Fixes: https://github.com/rust-lang/rust/issues/145524
r? ``@petrochenkov``
cleanup: use run_in_tmpdir in run-make/rustdoc-scrape-examples-paths
We had an issue on our LLVM-head Rust builder where it got stuck with this test failing because it was reusing the tmpdir between runs and something broke the incremental compile. Everything seems to work fine with run_in_tmpdir in this test. tests/run-make/uefi-qemu also uses the same tmpdir across runs, but I don't have the right environment to test that so I didn't try fixing it. That is the only use of std::env::temp_dir left in run-make tests after this fix.
take attr style into account in diagnostics
when the original attribute was specified as an inner attribute, the suggestion will now match that attribute style
rework GAT borrowck limitation error
The old one depends on the `ConstraintCategory` of the constraint which meant we did not emit this note if we had to prove the higher ranked trait bound due to e.g. normalization.
This made it annoying brittle and caused MIR borrowck errors to be order dependent, fixes the issue in https://github.com/rust-lang/rust/pull/140737#discussion_r2259592651.
r? types cc ```@amandasystems```
overhaul `&mut` suggestions in borrowck errors
* This refactors the logic so that it does not use fuzzy string matching for suggestions; it instead uses information directly from MIR.
* If something comes from a custom `Index` impl for which the `IndexMut` trait does not apply, do not suggest adding `mut` after `&`.
* Suggest `get_mut` with `unwrap` if error is fired on `BTreeMap` or `HashMap`.
Supersedes rust-lang/rust#144018 cc ```@xizheyin```
Closesrust-lang/rust#143732
Don't warn on never to any `as` casts as unreachable
I'm doing this while being sleep deprived on a night train, let's hope this is coherent.
Fixesrust-lang/rust#67227
Fix RISC-V Test Failures in ./x test for Multiple Codegen Cases
This PR resolves several test failures encountered when running `./x test` on the RISC-V architecture. These failures were caused by platform-specific behavior, ABI differences, or codegen inconsistencies unique to RISC-V.
The following test cases have been fixed to ensure compatibility with RISC-V:
* `codegen-llvm/enum/enum-match.rs`
* `codegen-llvm/enum/enum-transparent-extract.rs`
* `codegen-llvm/repeat-operand-zero-len.rs`
* `codegen-llvm/enum/enum-aggregate.rs`
* `codegen-llvm/uninhabited-transparent-return-abi.rs`
In addition, this PR adjusts `tests/codegen-llvm/transmute-scalar.rs` to explicitly specify the target architecture:
```rust
//@ compile-flags: --target=x86_64-unknown-linux-gnu
//@ needs-llvm-components: x86
```
As suggested by ```@nikic,``` this test is not target-specific and already uses `minicore`, implying it is meant to run against a stable triple regardless of the host architecture. Explicitly setting the target ensures consistent codegen behavior, particularly when testing on non-x86 platforms such as riscv64.
All changes have been tested locally on a RISC-V target and now pass as expected.
### Notes:
* These fixes are scoped specifically to enable full test suite compliance for RISC-V.
* No changes impact other architectures.
* The change to `transmute-scalar.rs` aligns with the intent of [[#143915](https://github.com/rust-lang/rust/pull/143915)](https://github.com/rust-lang/rust/pull/143915) and prevents architecture-dependent discrepancies during CI or local testing.
rustdoc-search: search backend with partitioned suffix tree
Before:
- https://notriddle.com/windows-docs-rs/doc-old/windows/
- https://doc.rust-lang.org/1.89.0/std/index.html
- https://doc.rust-lang.org/1.89.0/nightly-rustc/rustc_hir/index.html
After:
- https://notriddle.com/windows-docs-rs/doc/windows/
- https://notriddle.com/rustdoc-html-demo-12/stringdex/doc/std/index.html
- https://notriddle.com/rustdoc-html-demo-12/stringdex/compiler-doc/rustc_hir/index.html
## Summary
Rewrites the rustdoc search engine to use an indexed data structure, factored out as a crate called [stringdex](https://crates.io/crates/stringdex), that allows it to perform modified-levenshtein distance calculations, substring matches, and prefix matches in a reasonably efficient, and, more importantly, *incremental* algorithm.
## Motivation
Fixes https://github.com/rust-lang/rust/issues/131156
While the windows-rs crate is definitely the worst offender, I've noticed performance problems with the compiler crates as well. It makes no sense for rustdoc-search to have poor performance: it's basically a spell checker, and those have been usable since the 90's.
Stringdex is particularly designed to quickly return exact matches, to always report those matches first, and to never, ever [place new matches on top of old ones](https://web.dev/articles/cls). It also tries to yield to the event loop occasionally as it runs. This way, you can click the exactly-matched result before the rest of the search finishes running.
## Explanation
A longer description of how name search works can be found in stringdex's [HACKING.md](https://gitlab.com/notriddle/stringdex/-/blob/main/HACKING.md) file.
Type search is done by performing a name search on each element, then performing bitmap operations to narrow down a list of potential matches, then performing type unification on each match.
## Drawbacks
It's rather complex, and takes up more disk space than the current flat list of strings.
## Rationale and alternatives
Instead of a suffix tree, I could've used a different [approximate matching data structure](https://en.wikipedia.org/wiki/Approximate_string_matching). I didn't do that because I wanted to keep the current behavior (for example, a straightforward trigram index won't match [oepn](https://doc.rust-lang.org/nightly/std/?search=oepn) like the current system does).
## Prior art
[Sherlodoc](https://github.com/art-w/sherlodoc) is based on a similar concept, but they:
- use edit distance over a suffix tree for type-based search, instead of the binary matching that's implemented here
- use substring matching for name-based search, [but not fuzzy name matching](https://github.com/art-w/sherlodoc/issues/21)
- actually implement body text search, which is a potential-future feature, but not implemented in this PR
## Future possibilities
### Low-level optimization in stringdex
There are half a dozen low-level optimizations that I still need to explore. I haven't done them yet, because I've been working on bug fixes and rebasing on rustdoc's side, and a more solid and diverse test suite for stringdex itself.
- Stringdex decides whether to bundle two nodes into the same file based on size. To figure out a node's size, I have to run compression on it. This is probably slower than it needs to be.
- Stack compression is limited to the same 256-slot sliding windows as backref compression, and it doesn't have to be. (stack and backref compression are used to optimize the representation of the edge pointer from a parent node to its child; backref uses one byte, while stack is entirely implicit)
- The JS-side decoder is pretty naive. It performs unnecessary hash table lookups when decoding compressed nodes, and retains a list of hashes that it doesn't need. It needs to calculate the hashes in order to construct the merkle tree correctly, but it doesn't need to keep them.
- Data compression happens at the end, while emitting the node. This means it's not being counted when deciding on how to bundle, which is pretty dumb.
### Improved recall in type-driven search
Right now, type-driven search performs very strict matching. It's very precise, but misses a lot of things people would want.
What I'm not sure about is whether to focus more on edit-distance-based approaches, or to focus on type-theoretical approaches. Both gives avenues to improve, but edit distance is going to be faster while type checking is going to be more precise.
For example, a type theoretical improvement would fix [`Iterator<T>, (T -> U) -> Iterator<U>`](https://doc.rust-lang.org/nightly/std/?search=Iterator%3CT%3E%2C%20(T%20-%3E%20U)%20-%3E%20Iterator%3CU%3E) to give [`Iterator::map`](https://doc.rust-lang.org/nightly/std/iter/trait.Iterator.html#method.map), because it would recognize that the Map struct implements the Iterator trait. I don't know of any clean way to get this result to work without implementing significant type checking logic in search.js, and an edit-distance-based "dirty" approach would likely give a bunch of other results on top of this one.
## Full-text search
Once you've got this fuzzy dictionary matching to work, the logical next step is to implement some kind of information retrieval-based approach to phrase matching.
Like applying edit distance to types, phrase search gets you significantly better recall, but with a few major drawbacks:
- You have to pick between index bloat and the use of stopwords. Stopwords are bad because they might actually be important (try searching "if let" in mdBook if you're feeling brave), but without them, you spend a lot of space on text that doesn't matter.
- Example code also tends to have a lot of irrelevant stuff in it. Like stop words, we'd have to pick potentially-confusing or bloat.
Neither of these problems are deal-breakers, but they're worth keeping in mind.
Remove the `#[no_sanitize]` attribute in favor of `#[sanitize(xyz = "on|off")]`
This came up during the sanitizer stabilization (rust-lang/rust#123617). Instead of a `#[no_sanitize(xyz)]` attribute, we would like to have a `#[sanitize(xyz = "on|off")]` attribute, which is more powerful and allows to be extended in the future (instead
of just focusing on turning sanitizers off). The implementation is done according to what was [discussed on Zulip](https://rust-lang.zulipchat.com/#narrow/channel/343119-project-exploit-mitigations/topic/Stabilize.20the.20.60no_sanitize.60.20attribute/with/495377292)).
The new attribute also works on modules, traits and impl items and thus enables usage as the following:
```rust
#[sanitize(address = "off")]
mod foo {
fn unsanitized(..) {}
#[sanitize(address = "on")]
fn sanitized(..) {}
}
trait MyTrait {
#[sanitize(address = "off")]
fn unsanitized_default(..) {}
}
#[sanitize(thread = "off")]
impl MyTrait for () {
...
}
```
r? ```@rcvalle```
