Fragments are now included in file links.

These fragments are ignored during the file link check.
This commit allows their use in specific cases in the future.

lychee-lib/src/utils/request.rs

@@ -70,10 +70,19 @@ pub(crate) fn create(
                     credentials,
                 )))
             } else if let InputSource::FsPath(root) = &input_content.source {
-                if is_anchor {
-                    // Silently ignore anchor links for now
-                    Ok(None)
-                } else if let Some(url) = create_uri_from_path(root, &text, base)? {
+                let path = if is_anchor {
+                    // If the link is just an anchor,
+                    // prepend the filename of the file from which the link came.
+                    root.file_name()
+                        .expect("File doesn't have a file name.")
+                        .to_str()
+                        .expect("Filename is invalid unicode")
+                        .to_string()
+                        + &text
+                } else {
+                    text
+                };
+                if let Some(url) = create_uri_from_path(root, &path, base)? {
                     let uri = Uri { url };
                     let credentials = credentials(extractor, &uri);
 
@@ -122,7 +131,7 @@ fn construct_url(base: &Option<Url>, text: &str) -> Option<Result<Url>> {
 }
 
 fn create_uri_from_path(src: &Path, dst: &str, base: &Option<Base>) -> Result<Option<Url>> {
-    let dst = url::remove_get_params_and_fragment(dst);
+    let (dst, frag) = url::remove_get_params_and_seperate_fragment(dst);
     // Avoid double-encoding already encoded destination paths by removing any
     // potential encoding (e.g. `web%20site` becomes `web site`).
     // That's because Url::from_file_path will encode the full URL in the end.
@@ -136,6 +145,10 @@ fn create_uri_from_path(src: &Path, dst: &str, base: &Option<Base>) -> Result<Op
     let resolved = path::resolve(src, &PathBuf::from(&*decoded), base)?;
     match resolved {
         Some(path) => Url::from_file_path(&path)
+            .map(|mut url| {
+                url.set_fragment(frag);
+                url
+            })
             .map(Some)
             .map_err(|_e| ErrorKind::InvalidUrlFromPath(path)),
         None => Ok(None),

lychee-lib/src/utils/url.rs

@@ -4,18 +4,18 @@ use once_cell::sync::Lazy;
 
 static LINK_FINDER: Lazy<LinkFinder> = Lazy::new(LinkFinder::new);
 
-/// Remove all GET parameters from a URL.
+/// Remove all GET parameters from a URL and seperates out the fragment.
 /// The link is not a URL but a String as it may not have a base domain.
-pub(crate) fn remove_get_params_and_fragment(url: &str) -> &str {
-    let path = match url.split_once('#') {
-        Some((path_without_fragment, _fragment)) => path_without_fragment,
-        None => url,
+pub(crate) fn remove_get_params_and_seperate_fragment(url: &str) -> (&str, Option<&str>) {
+    let (path, frag) = match url.split_once('#') {
+        Some((path, fragment)) => (path, Some(fragment)),
+        None => (url, None),
     };
     let path = match path.split_once('?') {
         Some((path_without_params, _params)) => path_without_params,
         None => path,
     };
-    path
+    (path, frag)
 }
 
 // Use `LinkFinder` to offload the raw link searching in plaintext
@@ -29,47 +29,49 @@ mod test_fs_tree {
 
     #[test]
     fn test_remove_get_params_and_fragment() {
-        assert_eq!(remove_get_params_and_fragment("/"), "/");
+        assert_eq!(remove_get_params_and_seperate_fragment("/"), ("/", None));
         assert_eq!(
-            remove_get_params_and_fragment("index.html?foo=bar"),
-            "index.html"
+            remove_get_params_and_seperate_fragment("index.html?foo=bar"),
+            ("index.html", None)
         );
         assert_eq!(
-            remove_get_params_and_fragment("/index.html?foo=bar"),
-            "/index.html"
+            remove_get_params_and_seperate_fragment("/index.html?foo=bar"),
+            ("/index.html", None)
         );
         assert_eq!(
-            remove_get_params_and_fragment("/index.html?foo=bar&baz=zorx?bla=blub"),
-            "/index.html"
+            remove_get_params_and_seperate_fragment("/index.html?foo=bar&baz=zorx?bla=blub"),
+            ("/index.html", None)
         );
         assert_eq!(
-            remove_get_params_and_fragment("https://example.com/index.html?foo=bar"),
-            "https://example.com/index.html"
+            remove_get_params_and_seperate_fragment("https://example.com/index.html?foo=bar"),
+            ("https://example.com/index.html", None)
         );
         assert_eq!(
-            remove_get_params_and_fragment("test.png?foo=bar"),
-            "test.png"
+            remove_get_params_and_seperate_fragment("test.png?foo=bar"),
+            ("test.png", None)
         );
 
         assert_eq!(
-            remove_get_params_and_fragment("https://example.com/index.html#anchor"),
-            "https://example.com/index.html"
+            remove_get_params_and_seperate_fragment("https://example.com/index.html#anchor"),
+            ("https://example.com/index.html", Some("anchor"))
         );
         assert_eq!(
-            remove_get_params_and_fragment("https://example.com/index.html?foo=bar#anchor"),
-            "https://example.com/index.html"
+            remove_get_params_and_seperate_fragment(
+                "https://example.com/index.html?foo=bar#anchor"
+            ),
+            ("https://example.com/index.html", Some("anchor"))
         );
         assert_eq!(
-            remove_get_params_and_fragment("test.png?foo=bar#anchor"),
-            "test.png"
+            remove_get_params_and_seperate_fragment("test.png?foo=bar#anchor"),
+            ("test.png", Some("anchor"))
         );
         assert_eq!(
-            remove_get_params_and_fragment("test.png#anchor?anchor!?"),
-            "test.png"
+            remove_get_params_and_seperate_fragment("test.png#anchor?anchor!?"),
+            ("test.png", Some("anchor?anchor!?"))
         );
         assert_eq!(
-            remove_get_params_and_fragment("test.png?foo=bar#anchor?anchor!"),
-            "test.png"
+            remove_get_params_and_seperate_fragment("test.png?foo=bar#anchor?anchor!"),
+            ("test.png", Some("anchor?anchor!"))
         );
     }
 }