bump html5gum to 0.5.5

there's some perf improvements there, but one requires manual configuration. custom should_emit_errors, when inlined, eliminates some useless code that just wastes time when we don't care about errors.
lycheeverse · Jul 25, 2023 · fb95caf · fb95caf
1 parent 0d7c0a7
commit fb95caf
Show file tree

Hide file tree

Showing 3 changed files with 12 additions and 7 deletions.
diff --git a/Cargo.lock b/Cargo.lock
diff --git a/lychee-lib/Cargo.toml b/lychee-lib/Cargo.toml
@@ -25,7 +25,7 @@ futures = "0.3.27"
 glob = "0.3.1"
 headers = "0.3.8"
 html5ever = "0.26.0"
-html5gum = "0.5.4"
+html5gum = "0.5.5"
 http = "0.2.9"
 hyper = "0.14.27"
 ip_network = "0.4.1"

diff --git a/lychee-lib/src/extract/html5gum.rs b/lychee-lib/src/extract/html5gum.rs
@@ -212,6 +212,11 @@ impl Emitter for &mut LinkExtractor {
         self.flush_current_characters();
     }
     fn emit_error(&mut self, _: Error) {}
+
+    #[inline]
+    fn should_emit_errors() -> bool {
+        false
+    }
     fn pop_token(&mut self) -> Option<()> {
         None
     }
@@ -401,9 +406,9 @@ mod tests {
     #[test]
     fn test_include_nofollow() {
         let input = r#"
-        <a rel="nofollow" href="https://foo.com">do not follow me</a> 
-        <a rel="canonical,nofollow,dns-prefetch" href="https://example.com">do not follow me</a> 
-        <a href="https://example.org">i'm fine</a> 
+        <a rel="nofollow" href="https://foo.com">do not follow me</a>
+        <a rel="canonical,nofollow,dns-prefetch" href="https://example.com">do not follow me</a>
+        <a href="https://example.org">i'm fine</a>
         "#;
         let expected = vec![RawUri {
             text: "https://example.org".to_string(),
@@ -420,7 +425,7 @@ mod tests {
         <script>
         var foo = "https://example.com";
         </script>
-        <a href="https://example.org">i'm fine</a> 
+        <a href="https://example.org">i'm fine</a>
         "#;
         let expected = vec![RawUri {
             text: "https://example.org".to_string(),