Commit Graph

28363 Commits

Author SHA1 Message Date
Lukas Wirth
efd76ecff1 Move AstIdMap infra to the span crate 2024-03-01 15:39:44 +01:00
Lukas Wirth
549aae7c07 Add a few more doc lines to the span crate 2024-03-01 15:20:30 +01:00
Lukas Wirth
8a5bb9d5ac Only the ROOT syntax context has None outer_expn 2024-03-01 15:14:17 +01:00
Lukas Wirth
eb7a4f293e minor: Add some basic docs for spans/hygiene handling 2024-03-01 14:39:15 +01:00
Lukas Wirth
aabaa47bfb fix: Don't highlight related assoc items of super traits 2024-03-01 14:03:12 +01:00
Lukas Wirth
ed7e9aa5d8 Simplify 2024-03-01 13:25:24 +01:00
Lukas Wirth
c8fdcea85c Fix targetDir config name 2024-03-01 13:19:55 +01:00
Lukas Wirth
83a1ad5bfe Give all threads names 2024-03-01 13:19:55 +01:00
Vladislav Mamon
1d28aecd13
fix: autocomplete constants inside format strings 2024-03-01 14:23:31 +03:00
mo8it
00a049b585 Avoid cloning deps 2024-02-29 17:01:32 +01:00
mo8it
9df73c6cb0 Convert directly into an Arc<str> 2024-02-29 16:53:48 +01:00
mo8it
4aab2d5cb4 Remove unneeded to_owned 2024-02-29 16:29:15 +01:00
mo8it
748f57c16d Only clone when required 2024-02-29 16:28:59 +01:00
bors
9efa23c4da Auto merge of #16707 - regexident:pub-sema-impl, r=Veykril
Export `SemanticsImpl` from `ra_ap_hir` crate, since it's already exposed via `Semantics.deref()`

The `SemanticsImpl` type is already de-facto exposed via `<Semantics as Deref>::Target`.

By not being part of the public crate interface it however doesn't get included in the documentation, resulting in a massive blind spot when it comes to `ra_ap_hir`'s type resolution APIs.
2024-02-29 15:12:04 +00:00
mo8it
06a883e32f Take a reference instead of cloning 2024-02-29 16:05:26 +01:00
bors
b3b9b53df8 Auto merge of #16706 - Veykril:load-cargo-ide-db, r=Veykril
internal: Remove load-cargo dependency on ide

This lightens up the dep tree for projects using r-a as a library that do not need the ide crate itself.
2024-02-29 14:59:27 +00:00
bors
ecda4642ef Auto merge of #16705 - regexident:resolve-callable-exprs, r=Veykril
Add public function for resolving callable AST exprs to their HIR equivalents

(the PR is motivated by an outside use of the `ra_ap_hir` crate that would benefit from being able to walk a `hir::Function`'s AST, resolving callable exprs within to their HIR equivalents)
2024-02-29 14:46:21 +00:00
bors
a6606d1767 Auto merge of #16698 - regexident:param-derives, r=Veykril
Derive `PartialEq`, `Eq` & `Hash` for `hir::Param`

Since `hir::SelfParam`, as well as all members of `hir::Param` already implement `PartialEq`, `Eq` & `Hash` it seems reasonable to also make `hir::Param` implement those.

(the change is motivated by an outside use of the `ra_ap_hir` crate that would benefit from being able to collect params in a `HashSet`)
2024-02-29 14:33:37 +00:00
bors
ef2acb6ff3 Auto merge of #16695 - Veykril:revert-16541, r=Veykril
internal: Revert #16541

Closes https://github.com/rust-lang/rust-analyzer/issues/16602
2024-02-29 14:20:57 +00:00
Laurențiu Nicola
a01e4f8b72 Add basic support for Native Debug 2024-02-29 16:14:58 +02:00
bors
71eb5400a3 Auto merge of #16638 - Lindronics:destructure-struct-binding, r=Veykril
feature: Add `destructure_struct_binding`

Adds an assist for destructuring a struct in a binding (#8673). I saw that #13997 has been abandoned for a while, so I thought I'd give it a go.

## Example

```rust
let foo = Foo { bar: 1, baz: 2 };
let bar2 = foo.bar;
let baz2 = foo.baz;
let foo2 = foo;

let fizz = Fizz(1, 2);
let buzz = fizz.0;
```
becomes
```rust
let Foo { bar, baz } = Foo { bar: 1, baz: 2 };
let bar2 = bar;
let baz2 = baz;
let foo2 = todo!();

let Fizz(_0, _1) = Fizz(1, 2);
let buzz = _0;
```

More examples in the tests.

## What is included?

- [x] Destructure record, tuple, and unit struct bindings
- [x] Edit field usages
- [x] Non-exhaustive structs in foreign crates and private fields get hidden behind `..`
- [x] Nested bindings
- [x] Carry over `mut` and `ref mut` in nested bindings to fields, i.e. `let Foo { ref mut bar } = ...` becomes `let Foo { bar: Bar { baz: ref mut baz } } = ...`
- [x] Attempt to resolve collisions with other names in the scope
- [x] If the binding is to a reference, field usages are dereferenced if required
- [x] Use shorthand notation if possible

## Known limitations

- `let foo = Foo { bar: 1 }; foo;` currently results in `let Foo { bar } = Foo { bar: 1 }; todo!();` instead of reassembling the struct. This requires user intervention.
- Unused fields are not currently omitted. I thought that this is more ergonomic, as there already is a quick fix action for adding `: _` to unused field patterns.
2024-02-29 14:07:50 +00:00
bors
0ec6015b6e Auto merge of #16709 - ShoyuVanilla:fix-goto-index-mut, r=Veykril
fix: Goto definition for `index_mut`

Mostly same with #16696.
0ac05c0527/crates/hir-ty/src/infer/mutability.rs (L103-L133)
Thankfully, we are doing similar method resolutions so we can use them like the mentioned PR.
As there are only three `LangItem`s having `Mut` in there names; `FnMut`, `DerefMut` and `IndexMut`, I think that this is the last one 😄
2024-02-29 13:55:23 +00:00
Niklas Lindorfer
dc7b502689
Fix usage in other assist 2024-02-29 13:17:45 +00:00
Niklas Lindorfer
7c30c70658
Attempt resolving name collisions 2024-02-29 13:17:45 +00:00
Niklas Lindorfer
b203a07d92
Handle bindings to refs 2024-02-29 13:17:45 +00:00
Niklas Lindorfer
ed230048dc
Add destructure_struct_binding action
Separate into create and apply edit

Rename usages

Hacky name map

Add more tests

Handle non-exhaustive

Add some more TODOs

Private fields

Use todo

Nesting

Improve rest token generation

Cleanup

Doc -> regular comment

Support mut
2024-02-29 13:17:44 +00:00
Lukas Wirth
4ee0dbdc04 internal: Remove load-cargo dependency on ide 2024-02-28 16:59:04 +01:00
Lukas Wirth
03b02e6bd0 internal: Move ide-assists codegen tests into an xtask codegen command 2024-02-28 16:54:44 +01:00
Shoyu Vanilla
bf23deecb3 Fix goto index_mut 2024-02-29 00:52:58 +09:00
Shoyu Vanilla
83f6dd14f8 Add a new failing test for Goto index_mut definition 2024-02-29 00:52:30 +09:00
Vincent Esche
ed3497883c Add public function for resolving ast::Expr to hir::Callable 2024-02-28 16:44:07 +01:00
Vincent Esche
74531e4d80 Export SemanticsImpl from ra_ap_hir crate, since it's already exposed via Semantics.deref() 2024-02-28 16:16:54 +01:00
Lukas Wirth
ab533d887d fix: Ignore generic arguments in intra doc link path resolution 2024-02-28 14:47:47 +01:00
Vincent Esche
fac8a14de8 Add a more methods for resolving definitions from AST to their corresponding HIR types 2024-02-28 10:27:28 +01:00
Vincent Esche
6112ddfabb Add prefix file_ to Semantics's to_module_defs()/to_module_def() methods 2024-02-28 10:27:28 +01:00
Vincent Esche
c061a9e84e Derive PartialEq, Eq & Hash for hir::Param 2024-02-27 20:53:29 +01:00
bors
0ac05c0527 Auto merge of #16697 - regexident:relpath-to-relpathbuf, r=Veykril
Add `to_path_buf()` method for `RelPath`

There seems to be no ergonomic way to obtain a `RelPathBuf` from a corresponding `&RelPath` at the moment, making the latter sort of a dead end.

The `AbsPath` type provides the following:

```rust
impl AbsPath {
    // ...

    /// Equivalent of [`Path::to_path_buf`] for `AbsPath`.
    pub fn to_path_buf(&self) -> AbsPathBuf {
        AbsPathBuf::try_from(self.0.to_path_buf()).unwrap()
    }

    // ...
}
```

So I took the liberty of adding a corresponding equivalent for `RelPath:

```rust
impl RelPath {
    // ...

    /// Equivalent of [`Path::to_path_buf`] for `RelPath`.
    pub fn to_path_buf(&self) -> RelPathBuf {
        RelPathBuf::try_from(self.0.to_path_buf()).unwrap()
    }

    // ...
}
```

(the change is motivated by an outside use of the `ra_ap_paths` crate that would benefit from being able to use `RelPath` and `AbsPath` over `Path`)
2024-02-27 18:19:43 +00:00
Vincent Esche
0005794770 Add to_path_buf() method for RelPath 2024-02-27 18:33:47 +01:00
bors
a41cec9f7c Auto merge of #16696 - ShoyuVanilla:fix-goto-deref-mut, r=Veykril
fix: Goto definition for `deref_mut`

Fixes #16520

a3236be9d7/crates/hir/src/source_analyzer.rs (L375-L393)

As we can see from the above, current implementation routes all dereferencing prefix operations to `Deref::deref` implementation, not regarding mutabilities.

a3236be9d7/crates/hir-ty/src/infer/mutability.rs (L134-L151)

Since we are resolving them already in mutability inferences, we can use those results for proper `deref` / `deref_mut` routing.
2024-02-27 16:01:45 +00:00
Shoyu Vanilla
612443111b Remove unused imports 2024-02-28 00:10:08 +09:00
Shoyu Vanilla
0f4e313028 Fix goto deref_mut 2024-02-27 23:55:19 +09:00
Shoyu Vanilla
d1bdebf2b9 Add a new failing test for goto deref_mut 2024-02-27 23:54:05 +09:00
Lukas Wirth
30ad0ae209 internal: Revert #16541 2024-02-27 15:48:30 +01:00
bors
a3236be9d7 Auto merge of #16630 - ShoyuVanilla:fix-closure-kind-inference, r=Veykril
fix: Wrong closure kind deduction for closures with predicates

Completes #16472, fixes #16421

The changed closure kind deduction is mostly simlar to `rustc_hir_typeck/src/closure.rs`.
Porting closure sig deduction from it seems possible too and I'm considering doing it with another PR
2024-02-27 14:41:23 +00:00
bors
c031246546 Auto merge of #16693 - Veykril:system-rustc-sysroot, r=Veykril
fix: rust-project.json projects not preferring sysroot rustc
2024-02-27 14:28:36 +00:00
roife
76b86b24bd fix: clippy and format 2024-02-27 20:53:09 +08:00
roife
ec5236f3a8 test: use 4 spaces for indetation in macro expansion 2024-02-27 20:53:07 +08:00
roife
48966268fc fix: use 4 spaces for indentation in macro expansion 2024-02-27 20:49:09 +08:00
roife
ce4ae41605 internal: simplify the process of inserting spaces after mut 2024-02-27 20:49:09 +08:00
bors
6b250a22c4 Auto merge of #16687 - kilpkonn:master, r=Veykril
feat: Add "make tuple" tactic to term search

Follow up to https://github.com/rust-lang/rust-analyzer/pull/16092

Now term search also supports tuples.
```rust
let a: i32 = 1;
let b: f64 = 0.0;
let c: (i32, (f64, i32)) = todo!(); // Finds (a, (b, a))
```
In addition to new tactic that handles tuples I changed how the generics are handled.
Previously it tried all possible options from types we had in scope but now it only tries useful ones that help us directly towards the goal or at least towards calling some other function.
This changes O(2^n) to O(n^2) where n is amount of rounds which in practice allows using types that take generics for multiple rounds (previously limited to 1). Average case that also used to be exponential is now roughly linear.
This means that deeply nested generics also work.
````rust
// Finds all valid combos, including `Some(Some(Some(...)))`
let a: Option<Option<Option<bool>>> = todo!();
````

_Note that although the complexity is smaller allowing more types with generics the search overall slows down considerably. I hope it's fine tho as the autocomplete is disabled by default and for code actions it's not super slow. Might have to tweak the depth hyper parameter tho_

This resulted in a huge increase of results found (benchmarks on `ripgrep` crate):
Before
````
Tail Expr syntactic hits: 149/1692 (8%)
Tail Exprs found: 749/1692 (44%)
Term search avg time: 18ms
```
After
```
Tail Expr syntactic hits: 291/1692 (17%)
Tail Exprs found: 1253/1692 (74%)
Term search avg time: 139ms
````

Most changes are local to term search except some tuple related stuff on `hir::Type`.
2024-02-27 09:41:14 +00:00