feat!: add span and offsets to Doctype

1 files changed, 40 insertions, 11 deletions

diff --git a/src/emitter.rs b/src/emitter.rs
index f665f47..5b64acd 100644
--- a/src/emitter.rs
+++ b/src/emitter.rs
@@ -69,12 +69,12 @@ pub trait Emitter<O> {
     /// Emit the _current token_, assuming it is a comment.
     ///
     /// If the current token is not a comment, this method may panic.
-    fn emit_current_comment(&mut self);
+    fn emit_current_comment(&mut self, offset: O);
 
     /// Emit the _current token_, assuming it is a doctype.
     ///
     /// If the current token is not a doctype, this method may panic.
-    fn emit_current_doctype(&mut self);
+    fn emit_current_doctype(&mut self, offset: O);
 
     /// Assuming the _current token_ is a start tag, set the self-closing flag.
     ///
@@ -140,12 +140,12 @@ pub trait Emitter<O> {
     /// Assuming the _current token_ is a doctype, set its "public identifier" to the empty string.
     ///
     /// If the current token is not a doctype, this method may panic.
-    fn init_doctype_public_id(&mut self);
+    fn init_doctype_public_id(&mut self, offset: O);
 
     /// Assuming the _current token_ is a doctype, set its "system identifier" to the empty string.
     ///
     /// If the current token is not a doctype, this method may panic.
-    fn init_doctype_system_id(&mut self);
+    fn init_doctype_system_id(&mut self, offset: O);
 
     /// Assuming the _current token_ is a doctype, append a string to its "public identifier" to the given string.
     ///
@@ -308,17 +308,18 @@ impl<O: Offset> Emitter<O> for DefaultEmitter<O> {
         }
         self.emit_token(token);
     }
-    fn emit_current_comment(&mut self) {
+    fn emit_current_comment(&mut self, _offset: O) {
         let comment = self.current_token.take().unwrap();
         debug_assert!(matches!(comment, Token::Comment(_)));
         self.emit_token(comment);
     }
 
-    fn emit_current_doctype(&mut self) {
+    fn emit_current_doctype(&mut self, offset: O) {
         let Some(Token::Doctype(mut doctype)) = self.current_token.take() else {
             debug_assert!(false);
             return;
         };
+        doctype.span.end = offset;
         self.emit_token(Token::Doctype(doctype));
     }
 
@@ -370,12 +371,15 @@ impl<O: Offset> Emitter<O> for DefaultEmitter<O> {
             _ => debug_assert!(false),
         }
     }
-    fn init_doctype(&mut self, _offset: O) {
+    fn init_doctype(&mut self, offset: O) {
         self.current_token = Some(Token::Doctype(Doctype {
             name: String::new(),
             force_quirks: false,
             public_id: None,
             system_id: None,
+            span: offset..O::default(),
+            public_id_offset: O::default(),
+            system_id_offset: O::default(),
         }));
     }
 
@@ -405,19 +409,21 @@ impl<O: Offset> Emitter<O> for DefaultEmitter<O> {
         let current_attr = self.current_attribute.as_mut().unwrap();
         current_attr.1.value.push_str(s);
     }
-    fn init_doctype_public_id(&mut self) {
+    fn init_doctype_public_id(&mut self, offset: O) {
         let Some(Token::Doctype(doctype)) = &mut self.current_token else {
             debug_assert!(false);
             return;
         };
         doctype.public_id = Some("".to_owned());
+        doctype.public_id_offset = offset;
     }
-    fn init_doctype_system_id(&mut self) {
+    fn init_doctype_system_id(&mut self, offset: O) {
         let Some(Token::Doctype(doctype)) = &mut self.current_token else {
             debug_assert!(false);
             return;
         };
         doctype.system_id = Some("".to_owned());
+        doctype.system_id_offset = offset;
     }
     fn push_doctype_public_id(&mut self, s: &str) {
         if let Some(Token::Doctype(Doctype {
@@ -512,7 +518,7 @@ impl<O: Offset> Comment<O> {
 /// * `<!DOCTYPE {name} SYSTEM '{system_id}'>`
 /// * `<!DOCTYPE {name} PUBLIC '{public_id}' '{system_id}'>`
 #[derive(Debug, Eq, PartialEq)]
-pub struct Doctype {
+pub struct Doctype<O> {
     /// The ["force quirks"](https://html.spec.whatwg.org/#force-quirks-flag) flag.
     pub force_quirks: bool,
 
@@ -524,6 +530,29 @@ pub struct Doctype {
 
     /// The doctype's system identifier.
     pub system_id: Option<String>,
+
+    /// The source code span of the doctype.
+    pub span: Range<O>,
+
+    /// The source offset of the pulic identifier.
+    public_id_offset: O,
+
+    /// The source offset of the system identifier.
+    system_id_offset: O,
+}
+
+impl<O: Offset> Doctype<O> {
+    /// Calculates the span of the public identifier and returns it.
+    pub fn public_id_span(&self) -> Option<Range<O>> {
+        let public_id = self.public_id.as_ref()?;
+        Some(self.public_id_offset..self.public_id_offset + public_id.len())
+    }
+
+    /// Calculates the span of the system identifier and returns it.
+    pub fn system_id_span(&self) -> Option<Range<O>> {
+        let system_id = self.system_id.as_ref()?;
+        Some(self.system_id_offset..self.system_id_offset + system_id.len())
+    }
 }
 
 /// The token type used by default. You can define your own token type by implementing the
@@ -539,7 +568,7 @@ pub enum Token<O> {
     /// A HTML comment.
     Comment(Comment<O>),
     /// A HTML doctype declaration.
-    Doctype(Doctype),
+    Doctype(Doctype<O>),
     /// A HTML parsing error.
     ///
     /// Can be skipped over, the tokenizer is supposed to recover from the error and continues with