Implement new approach for URI normalization (fixes #287)

Author: colinodell

src/Util/UrlEncoder.php

@@ -48,6 +48,53 @@ public static function unescapeAndEncode($uri)
     {
         $decoded = html_entity_decode($uri);
 
-        return strtr(rawurlencode(rawurldecode($decoded)), self::$dontEncode);
+        return self::encode(self::decode($decoded));
+    }
+
+    /**
+     * Decode a percent-encoded URI
+     *
+     * @param string $uri
+     *
+     * @return string
+     */
+    private static function decode($uri)
+    {
+        return preg_replace_callback('/%([0-9a-f]{2})/iu', function($matches) {
+            // Convert percent-encoded codes to uppercase
+            $upper = strtoupper($matches[0]);
+            // Keep excluded characters as-is
+            if (array_key_exists($upper, self::$dontEncode)) {
+                return $upper;
+            }
+
+            // Otherwise, return the character for this codepoint
+            return chr(hexdec($matches[1]));
+        }, $uri);
+    }
+
+    /**
+     * Encode a URI, preserving already-encoded and excluded characters
+     *
+     * @param string $uri
+     *
+     * @return string
+     */
+    private static function encode($uri)
+    {
+        return preg_replace_callback('/(%[0-9a-f]{2})|./iu', function($matches){
+            // Keep already-encoded characters as-is
+            if (count($matches) > 1) {
+                return $matches[0];
+            }
+
+            // Keep excluded characters as-is
+            if (in_array($matches[0], self::$dontEncode)) {
+                return $matches[0];
+            }
+
+            // Otherwise, encode the character
+            return rawurlencode($matches[0]);
+        }, $uri);
     }
 }

tests/unit/Util/UrlEncoderTest.php

@@ -51,6 +51,11 @@ public function unescapeAndEncodeTestProvider()
             ['<', '%3C'],
             ['>', '%3E'],
             ['?', '?'],
+            ['https://en.wikipedia.org/wiki/Markdown#CommonMark', 'https://en.wikipedia.org/wiki/Markdown#CommonMark'],
+            ['https://img.shields.io/badge/help-%23hoaproject-ff0066.svg', 'https://img.shields.io/badge/help-%23hoaproject-ff0066.svg'],
+            ['http://example.com/a%62%63%2fd%3Fe', 'http://example.com/abc%2Fd%3Fe'],
+            ['http://ko.wikipedia.org/wiki/위키백과:대문', 'http://ko.wikipedia.org/wiki/%EC%9C%84%ED%82%A4%EB%B0%B1%EA%B3%BC:%EB%8C%80%EB%AC%B8'],
+            ['http://ko.wikipedia.org/wiki/%EC%9C%84%ED%82%A4%EB%B0%B1%EA%B3%BC:%EB%8C%80%EB%AC%B8', 'http://ko.wikipedia.org/wiki/%EC%9C%84%ED%82%A4%EB%B0%B1%EA%B3%BC:%EB%8C%80%EB%AC%B8'],
         ];
     }
 }