break!: make Doctype name field optional

4 files changed, 24 insertions, 4 deletions

diff --git a/CHANGELOG.md b/CHANGELOG.md
index 7b0b5e7..0deb60e 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -33,6 +33,8 @@
   * `StartTag`/`EndTag`: Added `name_span` fields
     (and removed the same-named methods).
 
+  * `Doctype`: The `name` field is now optional.
+
   * `AttributeOwned`: The `name_offset` and `value_offset` fields have
     been replaced with `name_span` and `value_span` respectively.
 
diff --git a/integration_tests/tests/test_html5lib.rs b/integration_tests/tests/test_html5lib.rs
index f351f85..f13e580 100644
--- a/integration_tests/tests/test_html5lib.rs
+++ b/integration_tests/tests/test_html5lib.rs
@@ -146,7 +146,7 @@ fn run_test_inner<R: Reader>(
             Token::String(data) => actual.tokens.push(TestToken::Character(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()),
+                name: doctype.name,
                 public_id: doctype.public_id,
                 system_id: doctype.system_id,
                 force_quirks: doctype.force_quirks,
diff --git a/src/emitter.rs b/src/emitter.rs
index db3da78..69baec2 100644
--- a/src/emitter.rs
+++ b/src/emitter.rs
@@ -153,6 +153,11 @@ pub trait Emitter<O> {
     /// If there is no current attribute, this method may panic.
     fn push_attribute_value(&mut self, s: &str);
 
+    /// Assuming the _current token_ is a doctype, set its name to the empty string.
+    ///
+    /// If the current token is not a doctype, this method may panic.
+    fn init_doctype_name(&mut self, offset: 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.
@@ -411,15 +416,26 @@ impl<O: Offset> Emitter<O> for DefaultEmitter<O> {
         }
     }
 
+    fn init_doctype_name(&mut self, offset: O) {
+        let Some(Token::Doctype(doctype)) = &mut self.current_token else {
+            debug_assert!(false);
+            return;
+        };
+        doctype.name = Some("".into());
+    }
+
     fn push_doctype_name(&mut self, s: &str) {
         match self.current_token {
-            Some(Token::Doctype(ref mut doctype)) => doctype.name.push_str(s),
+            Some(Token::Doctype(Doctype {
+                name: Some(ref mut name),
+                ..
+            })) => name.push_str(s),
             _ => debug_assert!(false),
         }
     }
     fn init_doctype(&mut self, offset: O) {
         self.current_token = Some(Token::Doctype(Doctype {
-            name: String::new(),
+            name: None,
             force_quirks: false,
             public_id: None,
             system_id: None,
@@ -601,7 +617,7 @@ pub struct Doctype<O> {
 
     /// The doctype's name. Uppercase ASCII characters (A-Z) have been
     /// converted to lowercase. For HTML documents this should be "html".
-    pub name: String,
+    pub name: Option<String>,
 
     /// The doctype's public identifier.
     pub public_id: Option<String>,
diff --git a/src/machine.rs b/src/machine.rs
index 26e1652..6c4558c 100644
--- a/src/machine.rs
+++ b/src/machine.rs
@@ -1208,6 +1208,7 @@ where
             Some('\0') => {
                 slf.emit_error(Error::UnexpectedNullCharacter);
                 slf.init_doctype();
+                slf.emitter.init_doctype_name(slf.position_before_match);
                 slf.emitter.push_doctype_name("\u{fffd}");
                 slf.state = State::DoctypeName;
                 Ok(ControlToken::Continue)
@@ -1229,6 +1230,7 @@ where
             }
             Some(x) => {
                 slf.init_doctype();
+                slf.emitter.init_doctype_name(slf.position_before_match);
                 slf.emitter
                     .push_doctype_name(ctostr!(x.to_ascii_lowercase()));
                 slf.state = State::DoctypeName;