feat!: add offset to comments

4 files changed, 58 insertions, 6 deletions

diff --git a/integration_tests/tests/test_html5lib.rs b/integration_tests/tests/test_html5lib.rs
index 209e199..0f96063 100644
--- a/integration_tests/tests/test_html5lib.rs
+++ b/integration_tests/tests/test_html5lib.rs
@@ -135,7 +135,7 @@ fn run_test_inner<R: Reader>(
             }),
             Token::EndTag(tag) => actual.tokens.push(TestToken::EndTag { name: tag.name }),
             Token::String(data) => actual.tokens.push(TestToken::Character(data)),
-            Token::Comment(data) => actual.tokens.push(TestToken::Comment(data)),
+            Token::Comment(comment) => actual.tokens.push(TestToken::Comment(comment.data)),
             Token::Doctype(doctype) => actual.tokens.push(TestToken::Doctype {
                 name: Some(doctype.name).filter(|name| !name.is_empty()),
                 public_id: doctype.public_identifier,
diff --git a/src/emitter.rs b/src/emitter.rs
index b3fdb99..caf7b55 100644
--- a/src/emitter.rs
+++ b/src/emitter.rs
@@ -270,8 +270,11 @@ impl<O: Offset, R: Position<O>> Emitter<R> for DefaultEmitter<R, O> {
         self.seen_attributes.clear();
     }
 
-    fn init_comment(&mut self, _reader: &R) {
-        self.current_token = Some(Token::Comment(String::new()));
+    fn init_comment(&mut self, reader: &R) {
+        self.current_token = Some(Token::Comment(Comment {
+            data: String::new(),
+            data_offset: reader.position(),
+        }));
     }
     fn emit_current_tag(&mut self) {
         self.flush_current_attribute();
@@ -348,7 +351,7 @@ impl<O: Offset, R: Position<O>> Emitter<R> for DefaultEmitter<R, O> {
 
     fn push_comment(&mut self, s: &str) {
         match self.current_token {
-            Some(Token::Comment(ref mut data)) => data.push_str(s),
+            Some(Token::Comment(Comment { ref mut data, .. })) => data.push_str(s),
             _ => debug_assert!(false),
         }
     }
@@ -483,6 +486,22 @@ pub struct EndTag<O> {
     pub name_span: Range<O>,
 }
 
+/// An HTML comment.
+#[derive(PartialEq, Eq, Debug)]
+pub struct Comment<O> {
+    /// The text within the comment.
+    pub data: String,
+    /// The source offset of the comment data.
+    pub data_offset: O,
+}
+
+impl<O: Offset> Comment<O> {
+    /// Calculates the span for the comment data and returns it.
+    pub fn data_span(&self) -> Range<O> {
+        self.data_offset..self.data_offset + self.data.len()
+    }
+}
+
 /// A doctype. Some examples:
 ///
 /// * `<!DOCTYPE {name}>`
@@ -515,7 +534,7 @@ pub enum Token<O> {
     /// A literal string.
     String(String),
     /// A HTML comment.
-    Comment(String),
+    Comment(Comment<O>),
     /// A HTML doctype declaration.
     Doctype(Doctype),
     /// A HTML parsing error.
diff --git a/src/lib.rs b/src/lib.rs
index fd0349c..1105141 100644
--- a/src/lib.rs
+++ b/src/lib.rs
@@ -15,6 +15,6 @@ mod utils;
 #[cfg(feature = "integration-tests")]
 pub use utils::State as InternalState;
 
-pub use emitter::{Attribute, DefaultEmitter, Doctype, Emitter, EndTag, StartTag, Token};
+pub use emitter::{Attribute, Comment, DefaultEmitter, Doctype, Emitter, EndTag, StartTag, Token};
 pub use error::Error;
 pub use tokenizer::{State, Tokenizer};
diff --git a/tests/test_spans.rs b/tests/test_spans.rs
index b41b1b9..21882a3 100644
--- a/tests/test_spans.rs
+++ b/tests/test_spans.rs
@@ -105,6 +105,39 @@ fn attribute_value_span() {
     "###);
 }
 
+#[test]
+fn comment_proper_data_span() {
+    let html = "<!-- Why are you looking at the source code? -->";
+    let Token::Comment(comment) = tokenizer(html).next().unwrap() else {
+        panic!("expected comment");
+    };
+    // FIXME: this span is wrong (starts one byte too soon)
+    assert_eq!(comment.data, html[1..][comment.data_span()]);
+    let labels = vec![(comment.data_span(), "")];
+    assert_snapshot!(annotate(html, labels), @r###"
+    <!-- Why are you looking at the source code? -->
+       ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+    "###);
+}
+
+#[test]
+fn comment_bogus_data_span() {
+    let html = "<! Why are you looking at the source code? -->";
+    let Token::Comment(comment) = tokenizer(html)
+        .filter(|t| !matches!(t, Token::Error { .. }))
+        .next()
+        .unwrap()
+    else {
+        panic!("expected comment");
+    };
+    assert_eq!(comment.data, html[comment.data_span()]);
+    let labels = vec![(comment.data_span(), "")];
+    assert_snapshot!(annotate(html, labels), @r###"
+    <! Why are you looking at the source code? -->
+      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+    "###);
+}
+
 fn annotate_errors(html: &'static str) -> String {
     let mut labels = Vec::new();
     for token in tokenizer(html) {