Don't check prefix attribute

lycheeverse · Oct 19, 2024 · 18a20a2 · 18a20a2
1 parent 3a2533f
commit 18a20a2
Show file tree

Hide file tree

Showing 2 changed files with 31 additions and 0 deletions.
diff --git a/lychee-lib/src/extract/html/html5ever.rs b/lychee-lib/src/extract/html/html5ever.rs
@@ -86,6 +86,12 @@ impl TokenSink for LinkExtractor {
                     }
                 }
 
+                // Check and exclude `prefix` attribute. This attribute is used to define a prefix
+                // for the current element. It is not used to link to a resource.
+                if let Some(_prefix) = attrs.iter().find(|attr| &attr.name.local == "prefix") {
+                    return TokenSinkResult::Continue;
+                }
+
                 for attr in attrs {
                     let urls = LinkExtractor::extract_urls_from_elem_attr(
                         &attr.name.local,
@@ -416,6 +422,16 @@ mod tests {
         assert!(uris.is_empty());
     }
 
+    #[test]
+    fn test_skip_prefix() {
+        let input = r#"
+            <html lang="en-EN" prefix="og: https://ogp.me/ns#">
+        "#;
+
+        let uris = extract_html(input, false);
+        assert!(uris.is_empty());
+    }
+
     #[test]
     fn test_ignore_text_content_links() {
         let input = r#"

diff --git a/lychee-lib/src/extract/html/html5gum.rs b/lychee-lib/src/extract/html/html5gum.rs
@@ -178,6 +178,11 @@ impl LinkExtractor {
             return;
         }
 
+        if self.current_attributes.contains_key("prefix") {
+            self.current_attributes.clear();
+            return;
+        }
+
         let new_urls = self
             .extract_urls_from_elem_attr()
             .into_iter()
@@ -613,6 +618,16 @@ mod tests {
         assert!(uris.is_empty());
     }
 
+    #[test]
+    fn test_skip_prefix() {
+        let input = r#"
+            <html lang="en-EN" prefix="og: https://ogp.me/ns#">
+        "#;
+
+        let uris = extract_html(input, false);
+        assert!(uris.is_empty());
+    }
+
     #[test]
     fn test_ignore_text_content_links() {
         let input = r#"