Exclude tel scheme from being checked (#1429)

lycheeverse · May 19, 2024 · c3f7fe7 · c3f7fe7
1 parent 9e031b6
commit c3f7fe7
Show file tree

Hide file tree

Showing 5 changed files with 82 additions and 2 deletions.
diff --git a/lychee-lib/src/client.rs b/lychee-lib/src/client.rs
@@ -500,6 +500,7 @@ impl Client {
         let status = match uri.scheme() {
             _ if uri.is_file() => self.check_file(uri).await,
             _ if uri.is_mail() => self.check_mail(uri).await,
+            _ if uri.is_tel() => self.check_tel(uri).await,
             _ => self.check_website(uri, default_chain).await?,
         };
 
@@ -702,6 +703,14 @@ impl Client {
     pub async fn check_mail(&self, _uri: &Uri) -> Status {
         Status::Excluded
     }
+
+    /// Check a tel
+    ///
+    /// This implementation simply excludes all tel.
+    #[allow(clippy::unused_async)]
+    pub async fn check_tel(&self, _uri: &Uri) -> Status {
+        Status::Excluded
+    }
 }
 
 // Check if the given `Url` would cause `reqwest` to panic.
@@ -907,6 +916,14 @@ mod tests {
         }));
     }
 
+    #[tokio::test]
+    async fn test_include_tel() {
+        let client = ClientBuilder::builder().build().client().unwrap();
+        assert!(client.is_excluded(&Uri {
+            url: "tel:1234567890".try_into().unwrap()
+        }));
+    }
+
     #[tokio::test]
     async fn test_require_https() {
         let client = ClientBuilder::builder().build().client().unwrap();

diff --git a/lychee-lib/src/extract/html/html5ever.rs b/lychee-lib/src/extract/html/html5ever.rs
@@ -89,9 +89,10 @@ impl TokenSink for LinkExtractor {
                                 // This ignores links like `<img srcset="v2@1.5x.png">`
                                 let is_email = is_email_link(url);
                                 let is_mailto = url.starts_with("mailto:");
+                                let is_phone = url.starts_with("tel:");
                                 let is_href = attr.name.local.as_ref() == "href";
 
-                                !is_email || (is_mailto && is_href)
+                                !is_email || (is_mailto && is_href) || (is_phone && is_href)
                             })
                             .map(|url| RawUri {
                                 text: url.to_string(),
@@ -318,6 +319,29 @@ mod tests {
         let uris = extract_html(input, false);
         assert_eq!(uris, expected);
     }
+
+    #[test]
+    fn test_valid_tel() {
+        let input = r#"<!DOCTYPE html>
+        <html lang="en-US">
+          <head>
+            <meta charset="utf-8">
+            <title>Test</title>
+          </head>
+          <body>
+            <a href="tel:1234567890">
+          </body>
+        </html>"#;
+
+        let expected = vec![RawUri {
+            text: "tel:1234567890".to_string(),
+            element: Some("a".to_string()),
+            attribute: Some("href".to_string()),
+        }];
+        let uris = extract_html(input, false);
+        assert_eq!(uris, expected);
+    }
+
     #[test]
     fn test_exclude_email_without_mailto() {
         let input = r#"<!DOCTYPE html>

diff --git a/lychee-lib/src/extract/html/html5gum.rs b/lychee-lib/src/extract/html/html5gum.rs
@@ -172,9 +172,10 @@ impl LinkExtractor {
                         // This ignores links like `<img srcset="v2@1.5x.png">`
                         let is_email = is_email_link(url);
                         let is_mailto = url.starts_with("mailto:");
+                        let is_phone = url.starts_with("tel:");
                         let is_href = attr == "href";
 
-                        !is_email || (is_mailto && is_href)
+                        !is_email || (is_mailto && is_href) || (is_phone && is_href)
                     })
                     .map(|url| RawUri {
                         text: url.to_string(),
@@ -453,6 +454,28 @@ mod tests {
         assert_eq!(uris, expected);
     }
 
+    #[test]
+    fn test_valid_tel() {
+        let input = r#"<!DOCTYPE html>
+        <html lang="en-US">
+          <head>
+            <meta charset="utf-8">
+            <title>Test</title>
+          </head>
+          <body>
+            <a href="tel:1234567890">
+          </body>
+        </html>"#;
+
+        let expected = vec![RawUri {
+            text: "tel:1234567890".to_string(),
+            element: Some("a".to_string()),
+            attribute: Some("href".to_string()),
+        }];
+        let uris = extract_html(input, false);
+        assert_eq!(uris, expected);
+    }
+
     #[test]
     fn test_valid_email() {
         let input = r#"<!DOCTYPE html>

diff --git a/lychee-lib/src/filter/mod.rs b/lychee-lib/src/filter/mod.rs
@@ -214,6 +214,7 @@ impl Filter {
             || self.is_host_excluded(uri)
             || self.is_ip_excluded(uri)
             || self.is_mail_excluded(uri)
+            || uri.is_tel()
             || is_example_domain(uri)
             || is_unsupported_domain(uri)
         {

diff --git a/lychee-lib/src/types/uri/valid.rs b/lychee-lib/src/types/uri/valid.rs
@@ -96,6 +96,13 @@ impl Uri {
         self.scheme() == "mailto"
     }
 
+    #[inline]
+    #[must_use]
+    /// Check if the URI is a tel
+    pub fn is_tel(&self) -> bool {
+        self.scheme() == "tel"
+    }
+
     #[inline]
     #[must_use]
     /// Check if the URI is a file
@@ -325,6 +332,14 @@ mod tests {
         );
     }
 
+    #[test]
+    fn test_uri_tel() {
+        assert_eq!(
+            Uri::try_from("tel:1234567890"),
+            Ok(Uri::try_from("tel:1234567890").unwrap())
+        );
+    }
+
     #[test]
     fn test_uri_host_ip_v4() {
         assert_eq!(