Quote bugfix (#47)

related: [query_selector cannot parse meta[property=og:title]](#46)

Co-authored-by: mehmetcan <2010841+mehmetcansahin@users.noreply.github.com>

src/queryselector/parser.rs

@@ -83,14 +83,24 @@ impl<'a> Parser<'a> {
             }
             Some(b'=') => {
                 self.stream.advance();
+                let quote = self.stream.expect_oneof_and_skip(&[b'"', b'\'']);
                 let value = self.read_identifier();
+                if let Some(quote) = quote {
+                    // Only require the given quote if the value starts with a quote
+                    self.stream.expect_and_skip(quote)?;
+                }
                 self.stream.expect_and_skip(b']')?;
                 Selector::AttributeValue(attribute, value)
             }
             Some(c @ b'~' | c @ b'^' | c @ b'$' | c @ b'*') => {
                 self.stream.advance();
                 self.stream.expect_and_skip(b'=')?;
+                let quote = self.stream.expect_oneof_and_skip(&[b'"', b'\'']);
                 let value = self.read_identifier();
+                if let Some(quote) = quote {
+                    // Only require the given quote if the value starts with a quote
+                    self.stream.expect_and_skip(quote)?;
+                }
                 self.stream.expect_and_skip(b']')?;
                 match c {
                     b'~' => Selector::AttributeValueWhitespacedContains(attribute, value),

src/tests.rs

@@ -572,6 +572,35 @@ mod query_selector {
             "PASS"
         );
     }
+
+    #[test]
+    fn query_selector_with_quote() {
+        let input = r#"<div><meta property="og:title" content="hello" /></div>"#;
+        let dom = parse(input, ParserOptions::default()).unwrap();
+        let parser = dom.parser();
+        let node_option = dom
+            .query_selector(r#"meta[property="og:title"]"#)
+            .and_then(|mut iter| iter.next());
+        let value = if let Some(node) = node_option {
+            Some(
+                node.get(parser)
+                    .unwrap()
+                    .as_tag()
+                    .unwrap()
+                    .attributes()
+                    .get("content")
+                    .flatten()
+                    .unwrap()
+                    .try_as_utf8_str()
+                    .unwrap()
+                    .to_string(),
+            )
+        } else {
+            None
+        };
+
+        assert_eq!(value, Some("hello".to_string()));
+    }
 }
 
 #[test]

src/util.rs

@@ -5,6 +5,7 @@ pub fn is_ident(c: u8) -> bool {
         || (b'a'..=b'z').contains(&c)
         || c == b'-'
         || c == b'_'
+        || c == b':'
 }
 
 #[inline(always)]