fix!: remove adjusted_current_node_present_and_not_in_html_namespace

Conceptually the tokenizer emits tokens, which are then handled in the
tree construction stage (which this crate doesn't yet implement).
While the tokenizer can operate almost entirely based on its state
(which may be changed via Tokenizer::set_state) and its internal state,
there is the exception of the 'Markup declaration open state'[1], the third
condition of which depends on the "adjusted current node", which in turn
depends on the "stack of open elements" only known to the tree constructor.

In 82898967320f90116bbc686ab7ffc2f61ff456c4 I tried to address this
by adding the adjusted_current_node_present_and_not_in_html_namespace
method to the Emitter trait. What I missed was that adding this method
to the Emitter trait effectively crippled the composability of the API.
You should be able to do the following:

    struct TreeConstructor<R, O> {
        tokenizer: Tokenizer<R, O, SomeEmitter<O>>,
        stack_of_open_elements: Vec<NodeId>,
        // ...
    }

However this doesn't work if the implementation of SomeEmitter
depends on the stack_of_open_elements field.

This commits remedies this oversight by removing this method and
instead making the Tokenizer yield values of a new Event enum:

    enum Event<T> { Token(T), CdataOpen }

Event::CdataOpen signals that the new Tokenizer::handle_cdata_open
method has to be called, which accepts a CdataAction:

   enum CdataAction { Cdata, BogusComment }

the variants of which correspond exactly to the possible outcomes
of the third condition of the 'Markup declaration open state'.

Removing this method also has the added benefit that the DefaultEmitter
is now again spec-compliant, which lets us expose it again in the next
commit in good conscience (previously it just hard-coded the method
implementation to return false, which is why I had removed the
DefaultEmitter from the public API in the last release).

[1]: https://html.spec.whatwg.org/multipage/parsing.html#markup-declaration-open-state

5 files changed, 80 insertions, 52 deletions

diff --git a/src/emitter.rs b/src/emitter.rs
index a5ecd55..dee0aa0 100644
--- a/src/emitter.rs
+++ b/src/emitter.rs
@@ -156,13 +156,6 @@ pub trait Emitter<O> {
     ///
     /// If the current token is not a doctype, this method may panic.
     fn push_doctype_system_id(&mut self, s: &str);
-
-    /// Returns true if there is an _adjusted current node_ and it is not an element in the HTML namespace.
-    ///
-    /// See the third list item under [Markup declaration open state].
-    ///
-    /// [Markup declaration open state]: https://html.spec.whatwg.org/multipage/parsing.html#markup-declaration-open-state
-    fn adjusted_current_node_present_and_not_in_html_namespace(&mut self) -> bool;
 }
 
 /// The DefaultEmitter is not exposed in the public API because:
@@ -172,26 +165,12 @@ pub trait Emitter<O> {
 ///
 #[cfg_attr(not(feature = "integration-tests"), doc = "```ignore")]
 #[cfg_attr(feature = "integration-tests", doc = "```")]
-///   # use html5tokenizer::{DefaultEmitter, Tokenizer, Token};
+///   # use html5tokenizer::{DefaultEmitter, Event, Tokenizer, Token};
 ///   let emitter = DefaultEmitter::default();
 ///   let html = "<script><b>";
 ///   let mut tokens = Tokenizer::new(html, emitter).flatten();
-///   assert!(matches!(tokens.next().unwrap(), Token::StartTag(tag) if tag.name == "script"));
-///   assert!(matches!(tokens.next().unwrap(), Token::StartTag(tag) if tag.name == "b"));
-///   ```
-///
-/// * The DefaultEmitter implements [`Emitter::adjusted_current_node_present_and_not_in_html_namespace`]
-///   by returning false, which results in all CDATA sections being tokenized as bogus comments.
-///
-#[cfg_attr(not(feature = "integration-tests"), doc = "```ignore")]
-#[cfg_attr(feature = "integration-tests", doc = "```")]
-///   # use html5tokenizer::{DefaultEmitter, Tokenizer, Token};
-///   let emitter = DefaultEmitter::default();
-///   let html = "<svg><![CDATA[I love SVG]]>";
-///   let mut tokens = Tokenizer::new(html, emitter).flatten();
-///   assert!(matches!(tokens.next().unwrap(), Token::StartTag(tag) if tag.name == "svg"));
-///   assert!(matches!(tokens.next().unwrap(), Token::Error {..}));
-///   assert!(matches!(tokens.next().unwrap(), Token::Comment(_bogus_comment)));
+///   assert!(matches!(tokens.next().unwrap(), Event::Token(Token::StartTag(tag)) if tag.name == "script"));
+///   assert!(matches!(tokens.next().unwrap(), Event::Token(Token::StartTag(tag)) if tag.name == "b"));
 ///   ```
 ///
 /// [`Tokenizer::set_state`]: crate::Tokenizer::set_state
@@ -467,10 +446,6 @@ impl<O: Offset> Emitter<O> for DefaultEmitter<O> {
             debug_assert!(false);
         }
     }
-
-    fn adjusted_current_node_present_and_not_in_html_namespace(&mut self) -> bool {
-        false
-    }
 }
 
 /// An HTML start tag, such as `<p>` or `<a>`.
@@ -613,7 +588,7 @@ pub enum Token<O> {
 #[cfg(test)]
 mod tests {
     use super::{DefaultEmitter, Token};
-    use crate::{attr::AttrValueSyntax, Tokenizer};
+    use crate::{attr::AttrValueSyntax, Event, Tokenizer};
 
     #[test]
     fn test_attribute_value_syntax() {
@@ -622,7 +597,7 @@ mod tests {
             DefaultEmitter::default(),
         )
         .flatten();
-        let Token::StartTag(start_tag) = tokenizer.next().unwrap() else {
+        let Event::Token(Token::StartTag(start_tag)) = tokenizer.next().unwrap() else {
             panic!("expected start tag");
         };
         assert_eq!(
diff --git a/src/lib.rs b/src/lib.rs
index 9dd878c..7bc17c3 100644
--- a/src/lib.rs
+++ b/src/lib.rs
@@ -20,7 +20,7 @@ mod utils;
 pub use emitter::{Comment, Doctype, Emitter, EndTag, StartTag, Token};
 pub use error::Error;
 pub use naive_parser::NaiveParser;
-pub use tokenizer::{State, Tokenizer};
+pub use tokenizer::{CdataAction, Event, State, Tokenizer};
 
 #[cfg(feature = "integration-tests")]
 pub use utils::State as InternalState;
diff --git a/src/machine.rs b/src/machine.rs
index c5bf021..509dae5 100644
--- a/src/machine.rs
+++ b/src/machine.rs
@@ -1,6 +1,7 @@
 use crate::attr::AttrValueSyntax;
 use crate::entities::try_read_character_reference;
 use crate::offset::{Offset, Position};
+use crate::tokenizer::CdataAction;
 use crate::utils::{
     ascii_digit_pat, control_pat, ctostr, noncharacter_pat, surrogate_pat, whitespace_pat, State,
 };
@@ -9,6 +10,7 @@ use crate::{reader::Reader, Emitter, Error, Tokenizer};
 pub enum ControlToken {
     Eof,
     Continue,
+    CdataOpen,
 }
 
 // Note: This is not implemented as a method on Tokenizer because there's fields on Tokenizer that
@@ -942,21 +944,7 @@ where
                 slf.doctype_offset = slf.reader.position() - b"<!doctype".len();
                 Ok(ControlToken::Continue)
             }
-            Some('[') if slf.try_read_string("CDATA[", true)? => {
-                if slf
-                    .emitter
-                    .adjusted_current_node_present_and_not_in_html_namespace()
-                {
-                    slf.state = State::CdataSection;
-                } else {
-                    slf.emit_error(Error::CdataInHtmlContent);
-
-                    slf.emitter.init_comment(slf.reader.position());
-                    slf.emitter.push_comment("[CDATA[");
-                    slf.state = State::BogusComment;
-                }
-                Ok(ControlToken::Continue)
-            }
+            Some('[') if slf.try_read_string("CDATA[", true)? => Ok(ControlToken::CdataOpen),
             c => {
                 slf.emit_error(Error::IncorrectlyOpenedComment);
                 slf.emitter.init_comment(slf.reader.position() - 1);
@@ -1897,3 +1885,22 @@ where
         }
     }
 }
+
+#[inline]
+pub fn handle_cdata_open<O, R, E>(slf: &mut Tokenizer<R, O, E>, action: CdataAction)
+where
+    O: Offset,
+    R: Reader + Position<O>,
+    E: Emitter<O>,
+{
+    match action {
+        CdataAction::Cdata => slf.state = State::CdataSection,
+        CdataAction::BogusComment => {
+            slf.emit_error(Error::CdataInHtmlContent);
+
+            slf.emitter.init_comment(slf.reader.position());
+            slf.emitter.push_comment("[CDATA[");
+            slf.state = State::BogusComment;
+        }
+    }
+}
diff --git a/src/naive_parser.rs b/src/naive_parser.rs
index 2626209..f126dfd 100644
--- a/src/naive_parser.rs
+++ b/src/naive_parser.rs
@@ -1,7 +1,8 @@
 use crate::emitter::DefaultEmitter;
 use crate::offset::{Offset, Position};
 use crate::reader::{IntoReader, Reader};
-use crate::{Emitter, State, Tokenizer};
+use crate::tokenizer::CdataAction;
+use crate::{Emitter, Event, State, Tokenizer};
 
 /// A naive HTML parser (**not** spec-compliant since it doesn't do tree construction).
 ///
@@ -53,9 +54,21 @@ impl<R: Reader + Position<O>, O: Offset, E: Emitter<O>> Iterator for NaiveParser
     type Item = Result<E::Token, R::Error>;
 
     fn next(&mut self) -> Option<Self::Item> {
-        // A proper parser would follow the steps described under section '13.2.6 Tree construction'
-        // of the spec. Since this parser is naive, we directly return the token instead.
-        self.tokenizer.next()
+        loop {
+            let event = self.tokenizer.next()?;
+            match event {
+                Err(e) => return Some(Err(e)),
+                Ok(Event::Token(t)) => {
+                    // A proper parser would follow the steps described under section '13.2.6 Tree construction'
+                    // of the spec. Since this parser is naive, we directly return the token instead.
+                    return Some(Ok(t));
+                }
+                Ok(Event::CdataOpen) => {
+                    // Naively parse any CDATA sections as bogus comments.
+                    self.tokenizer.handle_cdata_open(CdataAction::BogusComment)
+                }
+            }
+        }
     }
 }
 
diff --git a/src/tokenizer.rs b/src/tokenizer.rs
index 96d1c34..5b11db0 100644
--- a/src/tokenizer.rs
+++ b/src/tokenizer.rs
@@ -79,6 +79,38 @@ impl<R: Reader + Position<O>, O: Offset, E: Emitter<O>> Tokenizer<R, O, E> {
             naively_switch_state: false,
         }
     }
+
+    /// To be called when the tokenizer iterator implementation yields [`Event::CdataOpen`].
+    ///
+    /// For spec-compliant parsing *action* must be [`CdataAction::Cdata`],
+    /// if there is an _adjusted current node_ and it is not an element in
+    /// the HTML namespace, or [`CdataAction::BogusComment`] otherwise
+    /// (as per the third condition under [Markup declaration open state]).
+    ///
+    /// [Markup declaration open state]: https://html.spec.whatwg.org/multipage/parsing.html#markup-declaration-open-state
+    pub fn handle_cdata_open(&mut self, action: CdataAction) {
+        machine::handle_cdata_open(self, action);
+    }
+}
+
+/// Used by [`Tokenizer::handle_cdata_open`] to determine how to process `<![CDATA[`
+///
+/// (Since as per the spec this depends on the _adjusted current node_).
+pub enum CdataAction {
+    /// Process it as CDATA.
+    Cdata,
+    /// Process it as a bogus comment.
+    BogusComment,
+}
+
+/// An event yielded by the [`Iterator`] implementation for the [`Tokenizer`].
+#[derive(Debug)]
+pub enum Event<T> {
+    /// A token emitted by the [`Emitter`].
+    Token(T),
+    /// The state machine encountered `<![CDATA[`. You must call [`Tokenizer::handle_cdata_open`],
+    /// before advancing the tokenizer iterator again.
+    CdataOpen,
 }
 
 /// The states you can set the tokenizer to.
@@ -307,12 +339,12 @@ where
     R: Reader + Position<O>,
     E: Emitter<O>,
 {
-    type Item = Result<E::Token, R::Error>;
+    type Item = Result<Event<E::Token>, R::Error>;
 
     fn next(&mut self) -> Option<Self::Item> {
         loop {
             if let Some(token) = self.emitter.pop_token() {
-                return Some(Ok(token));
+                return Some(Ok(Event::Token(token)));
             }
 
             if self.eof {
@@ -326,6 +358,7 @@ where
                     self.eof = true;
                     self.emitter.emit_eof();
                 }
+                Ok(ControlToken::CdataOpen) => return Some(Ok(Event::CdataOpen)),
             }
         }
     }