Rollup merge of #146195 - nixxo:urlencoding-fix, r=ehuss

Zalathar · web-flow · commit 3bd603b2394a · 2025-09-09T14:35:03.000+10:00
fix partial urlencoded link support Hello Rust community. This is my first contribution, hope is useful. While translating in Italian the rust book https://github.com/nixxo/rust-lang-book-it I noticed that the linkchecker tool was failing reporting broken links on some pages even if the link worked properly in the browser. Upon inspection I noticed that mdbook basically urlencoded the links, but not urlencoded the heading IDs resulting in a non-identical anchor/IDs pairing that linkchecker reports as non-valid. looking at the source code for the linkchecker tool I noticed that urlencoding was done by the `small_url_encode` function in a partial way, as the name suggests. Replacing this function with a full urlencoding fixes the issue and the links are properly reported as valid. - added full urlencoding to properly check urlencoded anchor links against non-urlencoded heading IDs - added tests urlecoding provided by https://crates.io/crates/urlencoding
diff --git a/Cargo.lock b/Cargo.lock
@@ -2167,6 +2167,7 @@ version = "0.1.0"
 dependencies = [
  "html5ever",
  "regex",
+ "urlencoding",
 ]
 
 [[package]]
@@ -5824,6 +5825,12 @@ dependencies = [
  "percent-encoding",
 ]
 
+[[package]]
+name = "urlencoding"
+version = "2.1.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "daf8dba3b7eb870caf1ddeed7bc9d2a049f3cfdfae7cb521b087cc33ae4c49da"
+
 [[package]]
 name = "utf-8"
 version = "0.7.6"
diff --git a/src/tools/linkchecker/Cargo.toml b/src/tools/linkchecker/Cargo.toml
@@ -10,3 +10,4 @@ path = "main.rs"
 [dependencies]
 regex = "1"
 html5ever = "0.29.0"
+urlencoding = "2.1.3"
diff --git a/src/tools/linkchecker/main.rs b/src/tools/linkchecker/main.rs
@@ -232,18 +232,7 @@ enum FileEntry {
 type Cache = HashMap<String, FileEntry>;
 
 fn small_url_encode(s: &str) -> String {
-    s.replace('<', "%3C")
-        .replace('>', "%3E")
-        .replace(' ', "%20")
-        .replace('?', "%3F")
-        .replace('\'', "%27")
-        .replace('&', "%26")
-        .replace(',', "%2C")
-        .replace(':', "%3A")
-        .replace(';', "%3B")
-        .replace('[', "%5B")
-        .replace(']', "%5D")
-        .replace('\"', "%22")
+    urlencoding::encode(s).to_string()
 }
 
 impl Checker {
diff --git a/src/tools/linkchecker/tests/valid/inner/bar.html b/src/tools/linkchecker/tests/valid/inner/bar.html
@@ -3,5 +3,8 @@
 
   <h2 id="barfrag">Bar</h2>
 
+  <!-- testing urlecoded anchor link against a non-urlencoded heading IDs -->
+  <h2 id="barfrag-è">Bar</h2>
+
 </body>
 </html>
diff --git a/src/tools/linkchecker/tests/valid/inner/foo.html b/src/tools/linkchecker/tests/valid/inner/foo.html
@@ -8,7 +8,15 @@
   <a href="https://example.com/doesnotexist">external links not validated</a>
   <a href="redir.html#redirfrag">Redirect</a>
 
+  <!-- testing urlecoded anchor link against a non-urlencoded heading IDs -->
+  <a href="#localfrag-%C3%A8"></a>
+  <a href="bar.html#barfrag-%C3%A8"></a>
+  <a href="redir.html#redirfrag-%C3%A8"></a>
+
   <h2 id="localfrag">Local</h2>
 
+  <!-- testing urlecoded anchor link against a non-urlencoded heading IDs -->
+  <h2 id="localfrag-è">Local</h2>
+
 </body>
 </html>
diff --git a/src/tools/linkchecker/tests/valid/inner/redir-target.html b/src/tools/linkchecker/tests/valid/inner/redir-target.html
@@ -1,5 +1,8 @@
 <html>
 <body>
   <h2 id="redirfrag">Redir</h2>
+
+  <!-- testing urlecoded anchor link against a non-urlencoded heading IDs -->
+  <h2 id="redirfrag-è">Redir</h2>
 </body>
 </html>
