Merge pull request #2013 from ImUrX/heading-extension

Add heading extension support
rust-lang · May 28, 2023 · 3a51abf · 3a51abf
2 parents 870e908 + 1db52ff
commit 3a51abf
Show file tree

Hide file tree

Showing 9 changed files with 504 additions and 125 deletions.
diff --git a/guide/src/format/markdown.md b/guide/src/format/markdown.md
@@ -220,3 +220,16 @@ To enable it, see the [`output.html.curly-quotes`] config option.
 [tables]: https://github.github.com/gfm/#tables-extension-
 [task list extension]: https://github.github.com/gfm/#task-list-items-extension-
 [`output.html.curly-quotes`]: configuration/renderers.md#html-renderer-options
+
+### Heading attributes
+
+Headings can have a custom HTML ID and classes. This let's you maintain the same ID even if you change the heading's text, it also let's you add multiple classes in the heading.
+
+Example:
+```md
+# Example heading { #first .class1 .class2 }
+```
+
+This makes the level 1 heading with the content `Example heading`, ID `first`, and classes `class1` and `class2`. Note that the attributes should be space-separated.
+
+More information can be found in the [heading attrs spec page](https://github.com/raphlinus/pulldown-cmark/blob/master/specs/heading_attrs.txt).
diff --git a/src/renderer/html_handlebars/hbs_renderer.rs b/src/renderer/html_handlebars/hbs_renderer.rs
@@ -789,8 +789,10 @@ fn make_data(
 /// Goes through the rendered HTML, making sure all header tags have
 /// an anchor respectively so people can link to sections directly.
 fn build_header_links(html: &str) -> String {
-    static BUILD_HEADER_LINKS: Lazy<Regex> =
-        Lazy::new(|| Regex::new(r"<h(\d)>(.*?)</h\d>").unwrap());
+    static BUILD_HEADER_LINKS: Lazy<Regex> = Lazy::new(|| {
+        Regex::new(r#"<h(\d)(?: id="([^"]+)")?(?: class="([^"]+)")?>(.*?)</h\d>"#).unwrap()
+    });
+    static IGNORE_CLASS: &[&str] = &["menu-title"];
 
     let mut id_counter = HashMap::new();
 
@@ -800,7 +802,22 @@ fn build_header_links(html: &str) -> String {
                 .parse()
                 .expect("Regex should ensure we only ever get numbers here");
 
-            insert_link_into_header(level, &caps[2], &mut id_counter)
+            // Ignore .menu-title because now it's getting detected by the regex.
+            if let Some(classes) = caps.get(3) {
+                for class in classes.as_str().split(" ") {
+                    if IGNORE_CLASS.contains(&class) {
+                        return caps[0].to_string();
+                    }
+                }
+            }
+
+            insert_link_into_header(
+                level,
+                &caps[4],
+                caps.get(2).map(|x| x.as_str().to_string()),
+                caps.get(3).map(|x| x.as_str().to_string()),
+                &mut id_counter,
+            )
         })
         .into_owned()
 }
@@ -810,15 +827,21 @@ fn build_header_links(html: &str) -> String {
 fn insert_link_into_header(
     level: usize,
     content: &str,
+    id: Option<String>,
+    classes: Option<String>,
     id_counter: &mut HashMap<String, usize>,
 ) -> String {
-    let id = utils::unique_id_from_content(content, id_counter);
+    let id = id.unwrap_or_else(|| utils::unique_id_from_content(content, id_counter));
+    let classes = classes
+        .map(|s| format!(" class=\"{s}\""))
+        .unwrap_or_default();
 
     format!(
-        r##"<h{level} id="{id}"><a class="header" href="#{id}">{text}</a></h{level}>"##,
+        r##"<h{level} id="{id}"{classes}><a class="header" href="#{id}">{text}</a></h{level}>"##,
         level = level,
         id = id,
-        text = content
+        text = content,
+        classes = classes
     )
 }
 
@@ -1015,6 +1038,21 @@ mod tests {
                 "<h1>Foo</h1><h3>Foo</h3>",
                 r##"<h1 id="foo"><a class="header" href="#foo">Foo</a></h1><h3 id="foo-1"><a class="header" href="#foo-1">Foo</a></h3>"##,
             ),
+            // id only
+            (
+                r##"<h1 id="foobar">Foo</h1>"##,
+                r##"<h1 id="foobar"><a class="header" href="#foobar">Foo</a></h1>"##,
+            ),
+            // class only
+            (
+                r##"<h1 class="class1 class2">Foo</h1>"##,
+                r##"<h1 id="foo" class="class1 class2"><a class="header" href="#foo">Foo</a></h1>"##,
+            ),
+            // both id and class
+            (
+                r##"<h1 id="foobar" class="class1 class2">Foo</h1>"##,
+                r##"<h1 id="foobar" class="class1 class2"><a class="header" href="#foobar">Foo</a></h1>"##,
+            ),
         ];
 
         for (src, should_be) in inputs {

diff --git a/src/renderer/html_handlebars/search.rs b/src/renderer/html_handlebars/search.rs
@@ -138,9 +138,11 @@ fn render_item(
 
                 in_heading = true;
             }
-            Event::End(Tag::Heading(i, ..)) if i as u32 <= max_section_depth => {
+            Event::End(Tag::Heading(i, id, _classes)) if i as u32 <= max_section_depth => {
                 in_heading = false;
-                section_id = Some(utils::unique_id_from_content(&heading, &mut id_counter));
+                section_id = id
+                    .map(|id| id.to_string())
+                    .or_else(|| Some(utils::unique_id_from_content(&heading, &mut id_counter)));
                 breadcrumbs.push(heading.clone());
             }
             Event::Start(Tag::FootnoteDefinition(name)) => {

diff --git a/src/utils/mod.rs b/src/utils/mod.rs
@@ -183,6 +183,7 @@ pub fn new_cmark_parser(text: &str, curly_quotes: bool) -> Parser<'_, '_> {
     opts.insert(Options::ENABLE_FOOTNOTES);
     opts.insert(Options::ENABLE_STRIKETHROUGH);
     opts.insert(Options::ENABLE_TASKLISTS);
+    opts.insert(Options::ENABLE_HEADING_ATTRIBUTES);
     if curly_quotes {
         opts.insert(Options::ENABLE_SMART_PUNCTUATION);
     }

diff --git a/test_book/src/individual/heading.md b/test_book/src/individual/heading.md
@@ -13,3 +13,9 @@
 ##### Really Small Heading
 
 ###### Is it even a heading anymore - heading
+
+## Custom id {#example-id}
+
+## Custom class {.class1 .class2}
+
+## Both id and class {#example-id2 .class1 .class2}
diff --git a/tests/dummy_book/src/SUMMARY.md b/tests/dummy_book/src/SUMMARY.md
@@ -14,6 +14,7 @@
     - [Unicode](first/unicode.md)
     - [No Headers](first/no-headers.md)
     - [Duplicate Headers](first/duplicate-headers.md)
+    - [Heading Attributes](first/heading-attributes.md)
 - [Second Chapter](second.md)
     - [Nested Chapter](second/nested.md)
 

diff --git a/tests/dummy_book/src/first/heading-attributes.md b/tests/dummy_book/src/first/heading-attributes.md
@@ -0,0 +1,5 @@
+# Heading Attributes {#attrs}
+
+## Heading with classes {.class1 .class2}
+
+## Heading with id and classes {#both .class1 .class2}
diff --git a/tests/rendered_output.rs b/tests/rendered_output.rs
@@ -35,6 +35,7 @@ const TOC_SECOND_LEVEL: &[&str] = &[
     "1.5. Unicode",
     "1.6. No Headers",
     "1.7. Duplicate Headers",
+    "1.8. Heading Attributes",
     "2.1. Nested Chapter",
 ];
 
@@ -754,6 +755,7 @@ mod search {
         let no_headers = get_doc_ref("first/no-headers.html");
         let duplicate_headers_1 = get_doc_ref("first/duplicate-headers.html#header-text-1");
         let conclusion = get_doc_ref("conclusion.html#conclusion");
+        let heading_attrs = get_doc_ref("first/heading-attributes.html#both");
 
         let bodyidx = &index["index"]["index"]["body"]["root"];
         let textidx = &bodyidx["t"]["e"]["x"]["t"];
@@ -766,7 +768,7 @@ mod search {
         assert_eq!(docs[&some_section]["body"], "");
         assert_eq!(
             docs[&summary]["body"],
-            "Dummy Book Introduction First Chapter Nested Chapter Includes Recursive Markdown Unicode No Headers Duplicate Headers Second Chapter Nested Chapter Conclusion"
+            "Dummy Book Introduction First Chapter Nested Chapter Includes Recursive Markdown Unicode No Headers Duplicate Headers Heading Attributes Second Chapter Nested Chapter Conclusion"
         );
         assert_eq!(
             docs[&summary]["breadcrumbs"],
@@ -785,6 +787,10 @@ mod search {
             docs[&no_headers]["body"],
             "Capybara capybara capybara. Capybara capybara capybara. ThisLongWordIsIncludedSoWeCanCheckThatSufficientlyLongWordsAreOmittedFromTheSearchIndex."
         );
+        assert_eq!(
+            docs[&heading_attrs]["breadcrumbs"],
+            "First Chapter » Heading Attributes » Heading with id and classes"
+        );
     }
 
     // Setting this to `true` may cause issues with `cargo watch`,
@@ -946,3 +952,19 @@ fn custom_fonts() {
         &["fonts.css", "myfont.woff"]
     );
 }
+
+#[test]
+fn custom_header_attributes() {
+    let temp = DummyBook::new().build().unwrap();
+    let md = MDBook::load(temp.path()).unwrap();
+    md.build().unwrap();
+
+    let contents = temp.path().join("book/first/heading-attributes.html");
+
+    let summary_strings = &[
+        r##"<h1 id="attrs"><a class="header" href="#attrs">Heading Attributes</a></h1>"##,
+        r##"<h2 id="heading-with-classes" class="class1 class2"><a class="header" href="#heading-with-classes">Heading with classes</a></h2>"##,
+        r##"<h2 id="both" class="class1 class2"><a class="header" href="#both">Heading with id and classes</a></h2>"##,
+    ];
+    assert_contains_strings(&contents, summary_strings);
+}