spans: support attribute names

3 files changed, 36 insertions, 9 deletions

diff --git a/src/emitter.rs b/src/emitter.rs
index 20bcba4..d37c8f8 100644
--- a/src/emitter.rs
+++ b/src/emitter.rs
@@ -213,7 +213,11 @@ impl<R> DefaultEmitter<R, ()> {
                         .and_modify(|_| {
                             error = Some(Error::DuplicateAttribute);
                         })
-                        .or_insert(v);
+                        .or_insert(Attribute {
+                            value: v,
+                            name_span: (),
+                            value_span: (),
+                        });
 
                     if let Some(e) = error {
                         self.emit_error(e);
@@ -444,7 +448,7 @@ pub struct StartTag<S> {
     ///
     /// Duplicate attributes are ignored after the first one as per WHATWG spec. Implement your own
     /// [`Emitter`] to tweak this behavior.
-    pub attributes: BTreeMap<String, String>,
+    pub attributes: BTreeMap<String, Attribute<S>>,
 
     /// The source code span of the tag name.
     pub name_span: S,
@@ -467,6 +471,19 @@ impl<S> StartTag<S> {
     }
 }
 
+/// A HTML attribute value (plus spans).
+#[derive(Debug, Default, Eq, PartialEq)]
+pub struct Attribute<S> {
+    /// The value of the attribute.
+    pub value: String,
+
+    /// The source code span of the attribute name.
+    pub name_span: S,
+
+    /// The source code span of the attribute value.
+    pub value_span: S,
+}
+
 /// A HTML end/close tag, such as `</p>` or `</a>`.
 #[derive(Debug, Default, Eq, PartialEq)]
 pub struct EndTag<S> {
diff --git a/src/lib.rs b/src/lib.rs
index 4b74d61..c601147 100644
--- a/src/lib.rs
+++ b/src/lib.rs
@@ -16,7 +16,7 @@ mod utils;
 #[cfg(feature = "integration-tests")]
 pub use utils::State as InternalState;
 
-pub use emitter::{DefaultEmitter, Doctype, Emitter, EndTag, StartTag, Token};
+pub use emitter::{Attribute, DefaultEmitter, Doctype, Emitter, EndTag, StartTag, Token};
 pub use error::Error;
 pub use never::Never;
 pub use reader::{BufReadReader, Readable, Reader, StringReader};
diff --git a/src/spans.rs b/src/spans.rs
index 72b30c0..300d659 100644
--- a/src/spans.rs
+++ b/src/spans.rs
@@ -5,7 +5,7 @@ use std::{
     mem,
 };
 
-use crate::{Doctype, Emitter, EndTag, Error, Reader, StartTag, Token};
+use crate::{Attribute, Doctype, Emitter, EndTag, Error, Reader, StartTag, Token};
 
 type Span = std::ops::Range<usize>;
 
@@ -58,7 +58,7 @@ pub struct SpanEmitter<R> {
     current_characters: String,
     current_token: Option<Token<Span>>,
     last_start_tag: String,
-    current_attribute: Option<(String, String)>,
+    current_attribute: Option<(String, Attribute<Span>)>,
     seen_attributes: BTreeSet<String>,
     emitted_tokens: VecDeque<Token<Span>>,
     reader: PhantomData<R>,
@@ -262,15 +262,25 @@ impl<R: GetPos> Emitter<R> for SpanEmitter<R> {
         }));
     }
 
-    fn init_attribute(&mut self, _reader: &R) {
+    fn init_attribute(&mut self, reader: &R) {
         self.flush_current_attribute();
-        self.current_attribute = Some((String::new(), String::new()));
+        self.current_attribute = Some((
+            String::new(),
+            Attribute {
+                name_span: reader.get_pos() - 1..reader.get_pos() - 1,
+                ..Default::default()
+            },
+        ));
     }
     fn push_attribute_name(&mut self, s: &str) {
-        self.current_attribute.as_mut().unwrap().0.push_str(s);
+        let current_attr = self.current_attribute.as_mut().unwrap();
+        current_attr.0.push_str(s);
+        current_attr.1.name_span.end += s.len();
     }
     fn push_attribute_value(&mut self, s: &str) {
-        self.current_attribute.as_mut().unwrap().1.push_str(s);
+        let current_attr = self.current_attribute.as_mut().unwrap();
+        current_attr.1.value.push_str(s);
+        current_attr.1.value_span.end += s.len();
     }
     fn set_doctype_public_identifier(&mut self, value: &str) {
         if let Some(Token::Doctype(Doctype {