This commit adds a function that tries to determine the syntax highlighting class of NAME_REFs based on the usage.
It is used for highlighting injections (such as highlighting of doctests) as the semantic logic will most of the time result in unresolved references.
It also adds a color to unresolved references in HTML encoding.
4860: Accept relative paths in rust-project.json r=matklad a=tweksteen
If a relative path is found as part of Crate.root_module or Root.path, interpret it as relative to the location of the rust-project.json file.
Fixes: #4816
Co-authored-by: Thiébaud Weksteen <tweek@google.com>
4892: Make ra_syntax::{SyntaxNodeChildren, SyntaxElementChildren} public. r=matklad a=davidlattimore
SyntaxNode::children and SyntaxNode::children_with_tokens return these types, but there's currently no way AFAIK to name them.
Co-authored-by: David Lattimore <dml@google.com>
4893: Add Embark Studios to the sponsors list r=matklad a=matklad
Forgot to do this when updating the sponsors list on the website the
last time :(
bors r+
🤖
Co-authored-by: Aleksey Kladov <aleksey.kladov@gmail.com>
4882: _match.rs: improve comment formatting r=matklad a=jonas-schievink
This results in much nicer rustdoc output
Co-authored-by: Jonas Schievink <jonasschievink@gmail.com>
4880: "fill match arms" assist: Match on bind patterns r=flodiebold a=tobz1000
This prevents duplication of match arms where the pre-existing arm is a bind pattern.
Co-authored-by: Toby Dimmick <tobydimmick@pm.me>
4875: Introduce paths crate r=matklad a=matklad
It's a good idea to distinguish between absolute and relative paths at
the type level, to avoid accidental dependency on the cwd, which
really shouldn't matter for rust-analyzer service
bors r+
🤖
Co-authored-by: Aleksey Kladov <aleksey.kladov@gmail.com>
It's a good idea to distinguish between absolute and relative paths at
the type level, to avoid accidental dependency on the cwd, which
really shouldn't matter for rust-analyzer service
4700: Add top level keywords completion r=matklad a=mcrakhman
This fixes the following issue: https://github.com/rust-analyzer/rust-analyzer/issues/4566.
Also added simple logic which filters the keywords which can be used with unsafe on the top level.
Co-authored-by: Mikhail Rakhmanov <rakhmanov.m@gmail.com>
Co-authored-by: Aleksey Kladov <aleksey.kladov@gmail.com>
4857: Fix invalid shorthand initialization diagnostic for tuple structs r=jonas-schievink a=OptimalStrategy
Initializing tuple structs explicitly, like in the example below, produces a "Shorthand struct initialization" diagnostic that leads to a compilation error when applied:
```rust
struct S(usize);
fn main() {
let s = S { 0: 0 }; // OK, but triggers the diagnostic
// let s = S { 0 }; // Compilation error
}
```
This PR adds a check that the field name is not a literal.
Co-authored-by: OptimalStrategy <george@usan-podgornov.com>
Co-authored-by: OptimalStrategy <17456182+OptimalStrategy@users.noreply.github.com>
4858: find_path: return shorter paths for external items r=flodiebold a=jonas-schievink
If a containing module is already in scope, there's no need to use the full path to the item.
Fixes https://github.com/rust-analyzer/rust-analyzer/issues/4846
Co-authored-by: Jonas Schievink <jonasschievink@gmail.com>
