Merge #8065

8065: Better handling of block doc comments r=Veykril a=Veykril

Moves doc string processing to `Attrs::docs`, as we need the indent info from all comments before being able to know how much to strip

Closes #7774

Co-authored-by: Lukas Wirth <lukastw97@gmail.com>

crates/hir_def/src/attr.rs

@@ -76,37 +76,23 @@ impl ops::Deref for Attrs {
 impl RawAttrs {
     pub(crate) const EMPTY: Self = Self { entries: None };
 
-    pub(crate) fn new(owner: &dyn AttrsOwner, hygiene: &Hygiene) -> Self {
+    pub(crate) fn new(owner: &dyn ast::AttrsOwner, hygiene: &Hygiene) -> Self {
-        let attrs: Vec<_> = collect_attrs(owner).collect();
+        let entries = collect_attrs(owner)
-        let entries = if attrs.is_empty() {
-            // Avoid heap allocation
-            None
-        } else {
-            Some(
-                attrs
-                    .into_iter()
             .enumerate()
             .flat_map(|(i, attr)| match attr {
-                        Either::Left(attr) => Attr::from_src(attr, hygiene).map(|attr| (i, attr)),
+                Either::Left(attr) => Attr::from_src(attr, hygiene, i as u32),
-                        Either::Right(comment) => comment.doc_comment().map(|doc| {
+                Either::Right(comment) => comment.doc_comment().map(|doc| Attr {
-                            (
+                    index: i as u32,
-                                i,
-                                Attr {
-                                    index: 0,
                     input: Some(AttrInput::Literal(SmolStr::new(doc))),
                     path: ModPath::from(hir_expand::name!(doc)),
-                                },
-                            )
                 }),
             })
-                    .map(|(i, attr)| Attr { index: i as u32, ..attr })
+            .collect::<Arc<_>>();
-                    .collect(),
+
-            )
+        Self { entries: if entries.is_empty() { None } else { Some(entries) } }
-        };
-        Self { entries }
     }
 
-    fn from_attrs_owner(db: &dyn DefDatabase, owner: InFile<&dyn AttrsOwner>) -> Self {
+    fn from_attrs_owner(db: &dyn DefDatabase, owner: InFile<&dyn ast::AttrsOwner>) -> Self {
         let hygiene = Hygiene::new(db.upcast(), owner.file_id);
         Self::new(owner.value, &hygiene)
     }
@@ -162,7 +148,7 @@ impl RawAttrs {
                     let attr = ast::Attr::parse(&format!("#[{}]", tree)).ok()?;
                     // FIXME hygiene
                     let hygiene = Hygiene::new_unhygienic();
-                    Attr::from_src(attr, &hygiene).map(|attr| Attr { index, ..attr })
+                    Attr::from_src(attr, &hygiene, index)
                 });
 
                 let cfg_options = &crate_graph[krate].cfg_options;
@@ -192,7 +178,7 @@ impl Attrs {
                     Some(it) => {
                         let raw_attrs = RawAttrs::from_attrs_owner(
                             db,
-                            it.as_ref().map(|it| it as &dyn AttrsOwner),
+                            it.as_ref().map(|it| it as &dyn ast::AttrsOwner),
                         );
                         match mod_data.definition_source(db) {
                             InFile { file_id, value: ModuleSource::SourceFile(file) } => raw_attrs
@@ -203,9 +189,9 @@ impl Attrs {
                     None => RawAttrs::from_attrs_owner(
                         db,
                         mod_data.definition_source(db).as_ref().map(|src| match src {
-                            ModuleSource::SourceFile(file) => file as &dyn AttrsOwner,
+                            ModuleSource::SourceFile(file) => file as &dyn ast::AttrsOwner,
-                            ModuleSource::Module(module) => module as &dyn AttrsOwner,
+                            ModuleSource::Module(module) => module as &dyn ast::AttrsOwner,
-                            ModuleSource::BlockExpr(block) => block as &dyn AttrsOwner,
+                            ModuleSource::BlockExpr(block) => block as &dyn ast::AttrsOwner,
                         }),
                     ),
                 }
@@ -263,7 +249,7 @@ impl Attrs {
         let mut res = ArenaMap::default();
 
         for (id, var) in src.value.iter() {
-            let attrs = RawAttrs::from_attrs_owner(db, src.with_value(var as &dyn AttrsOwner))
+            let attrs = RawAttrs::from_attrs_owner(db, src.with_value(var as &dyn ast::AttrsOwner))
                 .filter(db, krate);
 
             res.insert(id, attrs)
@@ -297,7 +283,7 @@ impl Attrs {
     /// Constructs a map that maps the lowered `Attr`s in this `Attrs` back to its original syntax nodes.
     ///
     /// `owner` must be the original owner of the attributes.
-    pub fn source_map(&self, owner: &dyn AttrsOwner) -> AttrSourceMap {
+    pub fn source_map(&self, owner: &dyn ast::AttrsOwner) -> AttrSourceMap {
         AttrSourceMap { attrs: collect_attrs(owner).collect() }
     }
 
@@ -325,15 +311,34 @@ impl Attrs {
             AttrInput::Literal(s) => Some(s),
             AttrInput::TokenTree(_) => None,
         });
-        // FIXME: Replace `Itertools::intersperse` with `Iterator::intersperse[_with]` until the
+        let indent = docs
-        // libstd api gets stabilized (https://github.com/rust-lang/rust/issues/79524).
+            .clone()
-        let docs = Itertools::intersperse(docs, &SmolStr::new_inline("\n"))
+            .flat_map(|s| s.lines())
-            .map(|it| it.as_str())
+            .filter(|line| !line.chars().all(|c| c.is_whitespace()))
-            .collect::<String>();
+            .map(|line| line.chars().take_while(|c| c.is_whitespace()).count())
-        if docs.is_empty() {
+            .min()
+            .unwrap_or(0);
+        let mut buf = String::new();
+        for doc in docs {
+            // str::lines doesn't yield anything for the empty string
+            if !doc.is_empty() {
+                buf.extend(Itertools::intersperse(
+                    doc.lines().map(|line| {
+                        line.char_indices()
+                            .nth(indent)
+                            .map_or(line, |(offset, _)| &line[offset..])
+                            .trim_end()
+                    }),
+                    "\n",
+                ));
+            }
+            buf.push('\n');
+        }
+        buf.pop();
+        if buf.is_empty() {
             None
         } else {
-            Some(Documentation(docs))
+            Some(Documentation(buf))
         }
     }
 }
@@ -407,7 +412,7 @@ pub enum AttrInput {
 }
 
 impl Attr {
-    fn from_src(ast: ast::Attr, hygiene: &Hygiene) -> Option<Attr> {
+    fn from_src(ast: ast::Attr, hygiene: &Hygiene, index: u32) -> Option<Attr> {
         let path = ModPath::from_src(ast.path()?, hygiene)?;
         let input = if let Some(lit) = ast.literal() {
             let value = match lit.kind() {
@@ -420,7 +425,7 @@ impl Attr {
         } else {
             None
         };
-        Some(Attr { index: 0, path, input })
+        Some(Attr { index, path, input })
     }
 
     /// Maps this lowered `Attr` back to its original syntax node.
@@ -429,7 +434,7 @@ impl Attr {
     ///
     /// Note that the returned syntax node might be a `#[cfg_attr]`, or a doc comment, instead of
     /// the attribute represented by `Attr`.
-    pub fn to_src(&self, owner: &dyn AttrsOwner) -> Either<ast::Attr, ast::Comment> {
+    pub fn to_src(&self, owner: &dyn ast::AttrsOwner) -> Either<ast::Attr, ast::Comment> {
         collect_attrs(owner).nth(self.index as usize).unwrap_or_else(|| {
             panic!("cannot find `Attr` at index {} in {}", self.index, owner.syntax())
         })
@@ -508,7 +513,7 @@ impl<'a> AttrQuery<'a> {
         self.attrs().next().is_some()
     }
 
-    pub fn attrs(self) -> impl Iterator<Item = &'a Attr> {
+    pub fn attrs(self) -> impl Iterator<Item = &'a Attr> + Clone {
         let key = self.key;
         self.attrs
             .iter()
@@ -521,7 +526,7 @@ where
     N: ast::AttrsOwner,
 {
     let src = InFile::new(src.file_id, src.to_node(db.upcast()));
-    RawAttrs::from_attrs_owner(db, src.as_ref().map(|it| it as &dyn AttrsOwner))
+    RawAttrs::from_attrs_owner(db, src.as_ref().map(|it| it as &dyn ast::AttrsOwner))
 }
 
 fn attrs_from_item_tree<N: ItemTreeNode>(id: ItemTreeId<N>, db: &dyn DefDatabase) -> RawAttrs {
@@ -530,7 +535,9 @@ fn attrs_from_item_tree<N: ItemTreeNode>(id: ItemTreeId<N>, db: &dyn DefDatabase
     tree.raw_attrs(mod_item.into()).clone()
 }
 
-fn collect_attrs(owner: &dyn AttrsOwner) -> impl Iterator<Item = Either<ast::Attr, ast::Comment>> {
+fn collect_attrs(
+    owner: &dyn ast::AttrsOwner,
+) -> impl Iterator<Item = Either<ast::Attr, ast::Comment>> {
     let (inner_attrs, inner_docs) = inner_attributes(owner.syntax())
         .map_or((None, None), |(attrs, docs)| ((Some(attrs), Some(docs))));
 

crates/ide/src/goto_definition.rs

@@ -95,12 +95,10 @@ fn extract_positioned_link_from_comment(
     let comment_range = comment.syntax().text_range();
     let doc_comment = comment.doc_comment()?;
     let def_links = extract_definitions_from_markdown(doc_comment);
+    let start = comment_range.start() + TextSize::from(comment.prefix().len() as u32);
     let (def_link, ns, _) = def_links.iter().min_by_key(|(_, _, def_link_range)| {
-        let matched_position = comment_range.start() + TextSize::from(def_link_range.start as u32);
+        let matched_position = start + TextSize::from(def_link_range.start as u32);
-        match position.offset.checked_sub(matched_position) {
+        position.offset.checked_sub(matched_position).unwrap_or_else(|| comment_range.end())
-            Some(distance) => distance,
-            None => comment_range.end(),
-        }
     })?;
     Some((def_link.to_string(), *ns))
 }

crates/ide/src/hover.rs

@@ -1533,12 +1533,21 @@ fn my() {}
     fn test_hover_struct_doc_comment() {
         check(
             r#"
-/// bar docs
+/// This is an example
+/// multiline doc
+///
+/// # Example
+///
+/// ```
+/// let five = 5;
+///
+/// assert_eq!(6, my_crate::add_one(5));
+/// ```
 struct Bar;
 
 fn foo() { let bar = Ba$0r; }
 "#,
-            expect![[r#"
+            expect![[r##"
                 *Bar*
 
                 ```rust
@@ -1551,8 +1560,17 @@ fn foo() { let bar = Ba$0r; }
 
                 ---
 
-                bar docs
+                This is an example
-            "#]],
+                multiline doc
+
+                # Example
+
+                ```
+                let five = 5;
+
+                assert_eq!(6, my_crate::add_one(5));
+                ```
+            "##]],
         );
     }
 
@@ -3423,6 +3441,40 @@ mod Foo$0 {
         );
     }
 
+    #[test]
+    fn hover_doc_block_style_indentend() {
+        check(
+            r#"
+/**
+    foo
+    ```rust
+    let x = 3;
+    ```
+*/
+fn foo$0() {}
+"#,
+            expect![[r#"
+                *foo*
+
+                ```rust
+                test
+                ```
+
+                ```rust
+                fn foo()
+                ```
+
+                ---
+
+                foo
+
+                ```rust
+                let x = 3;
+                ```
+            "#]],
+        );
+    }
+
     #[test]
     fn hover_comments_dont_highlight_parent() {
         check_hover_no_result(

crates/ide/src/runnables.rs

@@ -576,6 +576,20 @@ fn should_have_runnable_1() {}
 /// ```
 fn should_have_runnable_2() {}
 
+/**
+```rust
+let z = 55;
+```
+*/
+fn should_have_no_runnable_3() {}
+
+/**
+    ```rust
+    let z = 55;
+    ```
+*/
+fn should_have_no_runnable_4() {}
+
 /// ```no_run
 /// let z = 55;
 /// ```
@@ -616,7 +630,7 @@ fn should_have_no_runnable_6() {}
 struct StructWithRunnable(String);
 
 "#,
-            &[&BIN, &DOCTEST, &DOCTEST, &DOCTEST, &DOCTEST],
+            &[&BIN, &DOCTEST, &DOCTEST, &DOCTEST, &DOCTEST, &DOCTEST, &DOCTEST],
             expect![[r#"
                 [
                     Runnable {
@@ -682,7 +696,37 @@ struct StructWithRunnable(String);
                             file_id: FileId(
                                 0,
                             ),
-                            full_range: 756..821,
+                            full_range: 256..320,
+                            name: "should_have_no_runnable_3",
+                        },
+                        kind: DocTest {
+                            test_id: Path(
+                                "should_have_no_runnable_3",
+                            ),
+                        },
+                        cfg: None,
+                    },
+                    Runnable {
+                        nav: NavigationTarget {
+                            file_id: FileId(
+                                0,
+                            ),
+                            full_range: 322..398,
+                            name: "should_have_no_runnable_4",
+                        },
+                        kind: DocTest {
+                            test_id: Path(
+                                "should_have_no_runnable_4",
+                            ),
+                        },
+                        cfg: None,
+                    },
+                    Runnable {
+                        nav: NavigationTarget {
+                            file_id: FileId(
+                                0,
+                            ),
+                            full_range: 900..965,
                             name: "StructWithRunnable",
                         },
                         kind: DocTest {

crates/ide_db/src/call_info.rs

@@ -53,15 +53,15 @@ pub fn call_info(db: &RootDatabase, position: FilePosition) -> Option<CallInfo>
 
     match callable.kind() {
         hir::CallableKind::Function(func) => {
-            res.doc = func.docs(db).map(|it| it.as_str().to_string());
+            res.doc = func.docs(db).map(|it| it.into());
             format_to!(res.signature, "fn {}", func.name(db));
         }
         hir::CallableKind::TupleStruct(strukt) => {
-            res.doc = strukt.docs(db).map(|it| it.as_str().to_string());
+            res.doc = strukt.docs(db).map(|it| it.into());
             format_to!(res.signature, "struct {}", strukt.name(db));
         }
         hir::CallableKind::TupleEnumVariant(variant) => {
-            res.doc = variant.docs(db).map(|it| it.as_str().to_string());
+            res.doc = variant.docs(db).map(|it| it.into());
             format_to!(
                 res.signature,
                 "enum {}::{}",

crates/syntax/src/ast.rs

@@ -118,7 +118,7 @@ fn test_doc_comment_none() {
     .ok()
     .unwrap();
     let module = file.syntax().descendants().find_map(Module::cast).unwrap();
-    assert!(module.doc_comment_text().is_none());
+    assert!(module.doc_comments().doc_comment_text().is_none());
 }
 
 #[test]
@@ -133,7 +133,7 @@ fn test_outer_doc_comment_of_items() {
     .ok()
     .unwrap();
     let module = file.syntax().descendants().find_map(Module::cast).unwrap();
-    assert_eq!("doc", module.doc_comment_text().unwrap());
+    assert_eq!(" doc", module.doc_comments().doc_comment_text().unwrap());
 }
 
 #[test]
@@ -148,7 +148,7 @@ fn test_inner_doc_comment_of_items() {
     .ok()
     .unwrap();
     let module = file.syntax().descendants().find_map(Module::cast).unwrap();
-    assert!(module.doc_comment_text().is_none());
+    assert!(module.doc_comments().doc_comment_text().is_none());
 }
 
 #[test]
@@ -162,7 +162,7 @@ fn test_doc_comment_of_statics() {
     .ok()
     .unwrap();
     let st = file.syntax().descendants().find_map(Static::cast).unwrap();
-    assert_eq!("Number of levels", st.doc_comment_text().unwrap());
+    assert_eq!(" Number of levels", st.doc_comments().doc_comment_text().unwrap());
 }
 
 #[test]
@@ -181,7 +181,10 @@ fn test_doc_comment_preserves_indents() {
     .ok()
     .unwrap();
     let module = file.syntax().descendants().find_map(Module::cast).unwrap();
-    assert_eq!("doc1\n```\nfn foo() {\n    // ...\n}\n```", module.doc_comment_text().unwrap());
+    assert_eq!(
+        " doc1\n ```\n fn foo() {\n     // ...\n }\n ```",
+        module.doc_comments().doc_comment_text().unwrap()
+    );
 }
 
 #[test]
@@ -198,7 +201,7 @@ fn test_doc_comment_preserves_newlines() {
     .ok()
     .unwrap();
     let module = file.syntax().descendants().find_map(Module::cast).unwrap();
-    assert_eq!("this\nis\nmod\nfoo", module.doc_comment_text().unwrap());
+    assert_eq!(" this\n is\n mod\n foo", module.doc_comments().doc_comment_text().unwrap());
 }
 
 #[test]
@@ -212,7 +215,7 @@ fn test_doc_comment_single_line_block_strips_suffix() {
     .ok()
     .unwrap();
     let module = file.syntax().descendants().find_map(Module::cast).unwrap();
-    assert_eq!("this is mod foo", module.doc_comment_text().unwrap());
+    assert_eq!(" this is mod foo", module.doc_comments().doc_comment_text().unwrap());
 }
 
 #[test]
@@ -226,7 +229,7 @@ fn test_doc_comment_single_line_block_strips_suffix_whitespace() {
     .ok()
     .unwrap();
     let module = file.syntax().descendants().find_map(Module::cast).unwrap();
-    assert_eq!("this is mod foo ", module.doc_comment_text().unwrap());
+    assert_eq!(" this is mod foo ", module.doc_comments().doc_comment_text().unwrap());
 }
 
 #[test]
@@ -245,8 +248,8 @@ fn test_doc_comment_multi_line_block_strips_suffix() {
     .unwrap();
     let module = file.syntax().descendants().find_map(Module::cast).unwrap();
     assert_eq!(
-        "        this\n        is\n        mod foo\n        ",
+        "\n        this\n        is\n        mod foo\n        ",
-        module.doc_comment_text().unwrap()
+        module.doc_comments().doc_comment_text().unwrap()
     );
 }
 
@@ -260,7 +263,7 @@ fn test_comments_preserve_trailing_whitespace() {
     let def = file.syntax().descendants().find_map(Struct::cast).unwrap();
     assert_eq!(
         " Representation of a Realm.   \n In the specification these are called Realm Records.",
-        def.doc_comment_text().unwrap()
+        def.doc_comments().doc_comment_text().unwrap()
     );
 }
 
@@ -276,7 +279,7 @@ fn test_four_slash_line_comment() {
     .ok()
     .unwrap();
     let module = file.syntax().descendants().find_map(Module::cast).unwrap();
-    assert_eq!("doc comment", module.doc_comment_text().unwrap());
+    assert_eq!(" doc comment", module.doc_comments().doc_comment_text().unwrap());
 }
 
 #[test]

crates/syntax/src/ast/token_ext.rs

@@ -33,23 +33,20 @@ impl ast::Comment {
         prefix
     }
 
-    /// Returns the textual content of a doc comment block as a single string.
+    /// Returns the textual content of a doc comment node as a single string with prefix and suffix
-    /// That is, strips leading `///` (+ optional 1 character of whitespace),
+    /// removed.
-    /// trailing `*/`, trailing whitespace and then joins the lines.
     pub fn doc_comment(&self) -> Option<&str> {
         let kind = self.kind();
         match kind {
             CommentKind { shape, doc: Some(_) } => {
                 let prefix = kind.prefix();
                 let text = &self.text()[prefix.len()..];
-                let ws = text.chars().next().filter(|c| c.is_whitespace());
+                let text = if shape == CommentShape::Block {
-                let text = ws.map_or(text, |ws| &text[ws.len_utf8()..]);
+                    text.strip_suffix("*/").unwrap_or(text)
-                match shape {
+                } else {
-                    CommentShape::Block if text.ends_with("*/") => {
+                    text
-                        Some(&text[..text.len() - "*/".len()])
+                };
-                    }
+                Some(text)
-                    _ => Some(text),
-                }
             }
             _ => None,
         }

crates/syntax/src/ast/traits.rs

@@ -1,8 +1,6 @@
 //! Various traits that are implemented by ast nodes.
 //!
 //! The implementations are usually trivial, and live in generated.rs
-use itertools::Itertools;
-
 use crate::{
     ast::{self, support, AstChildren, AstNode, AstToken},
     syntax_node::SyntaxElementChildren,
@@ -76,10 +74,6 @@ pub trait DocCommentsOwner: AttrsOwner {
     fn doc_comments(&self) -> CommentIter {
         CommentIter { iter: self.syntax().children_with_tokens() }
     }
-
-    fn doc_comment_text(&self) -> Option<String> {
-        self.doc_comments().doc_comment_text()
-    }
 }
 
 impl CommentIter {
@@ -87,12 +81,12 @@ impl CommentIter {
         CommentIter { iter: syntax_node.children_with_tokens() }
     }
 
-    /// Returns the textual content of a doc comment block as a single string.
+    #[cfg(test)]
-    /// That is, strips leading `///` (+ optional 1 character of whitespace),
-    /// trailing `*/`, trailing whitespace and then joins the lines.
     pub fn doc_comment_text(self) -> Option<String> {
-        let docs =
+        let docs = itertools::Itertools::join(
-            self.filter_map(|comment| comment.doc_comment().map(ToOwned::to_owned)).join("\n");
+            &mut self.filter_map(|comment| comment.doc_comment().map(ToOwned::to_owned)),
+            "\n",
+        );
         if docs.is_empty() {
             None
         } else {