Percent-encode unreserved characters in URL path (#29)

webignition · Oct 9, 2018 · d26f181 · d26f181
1 parent 7c0bf4d
commit d26f181
Show file tree

Hide file tree

Showing 10 changed files with 44 additions and 150 deletions.
diff --git a/src/NormalisedUrl/Path/Normaliser.php b/src/NormalisedUrl/Path/Normaliser.php
@@ -2,6 +2,8 @@
 
 namespace webignition\NormalisedUrl\Path;
 
+use webignition\Url\PercentEncoder;
+
 class Normaliser
 {
     /**
@@ -14,6 +16,7 @@ class Normaliser
      */
     public function __construct($path)
     {
+        $path = PercentEncoder::encodeUnreservedCharacters($path);
         $this->path = (string)$path;
         $this->normalise();
     }

diff --git a/src/NormalisedUrl/Path/Path.php b/src/NormalisedUrl/Path/Path.php
@@ -16,7 +16,7 @@ public function __construct($path)
     {
         parent::__construct($path);
 
-        $normaliser = new Normaliser($path);
+        $normaliser = new Normaliser(parent::get());
 
         $this->set($normaliser->get());
     }

diff --git a/src/Url/Encoder.php b/src/Url/Encoder.php
diff --git a/src/Url/Path/Encoder.php b/src/Url/Path/Encoder.php
diff --git a/src/Url/Path/Path.php b/src/Url/Path/Path.php
@@ -2,6 +2,8 @@
 
 namespace webignition\Url\Path;
 
+use webignition\Url\PercentEncoder;
+
 /**
  * Represents the path part of a URL
  */
@@ -19,6 +21,8 @@ class Path
      */
     public function __construct($path)
     {
+        $path = PercentEncoder::encodeUnreservedCharacters($path);
+
         $this->set($path);
     }
 

diff --git a/src/Url/PercentEncoder.php b/src/Url/PercentEncoder.php
@@ -0,0 +1,20 @@
+<?php
+
+namespace webignition\Url;
+
+class PercentEncoder
+{
+    private static $charUnreserved = 'a-zA-Z0-9_\-\.~';
+    private static $charSubDelims = '!\$&\'\(\)\*\+,;=';
+
+    public static function encodeUnreservedCharacters($path)
+    {
+        return preg_replace_callback(
+            '/(?:[^' . self::$charUnreserved . self::$charSubDelims . '%:@\/]++|%(?![A-Fa-f0-9]{2}))/',
+            function (array $match) {
+                return rawurlencode($match[0]);
+            },
+            $path
+        );
+    }
+}
diff --git a/tests/DataProvider/PathNormalisationDataProviderTrait.php b/tests/DataProvider/PathNormalisationDataProviderTrait.php
@@ -82,6 +82,10 @@ public function pathNormalisationDataProvider()
                 'path' => '/one//two',
                 'expectedNormalisedPath' => '/one//two',
             ],
+            'percent-encode unicode characters' => [
+                'path' => '/Nattō',
+                'expectedNormalisedPath' => '/Natt%C5%8D',
+            ],
         ];
     }
 }
diff --git a/tests/Url/EncoderTest.php b/tests/Url/EncoderTest.php
diff --git a/tests/Url/Path/EncoderTest.php b/tests/Url/Path/EncoderTest.php
diff --git a/tests/Url/Path/PathTest.php b/tests/Url/Path/PathTest.php
@@ -9,13 +9,14 @@ class PathTest extends \PHPUnit_Framework_TestCase
     /**
      * @dataProvider createDataProvider
      *
-     * @param $pathString
+     * @param string|null $path
+     * @param string $expectedPath
      */
-    public function testCreate($pathString)
+    public function testCreate($path, $expectedPath)
     {
-        $path = new Path($pathString);
+        $path = new Path($path);
 
-        $this->assertEquals($pathString, (string)$path);
+        $this->assertEquals($expectedPath, (string)$path);
     }
 
     /**
@@ -26,12 +27,19 @@ public function createDataProvider()
         return [
             'null' => [
                 'pathString' => null,
+                'expectedPath' => '',
             ],
             'empty' => [
                 'pathString' => '',
+                'expectedPath' => '',
             ],
             'non-empty' => [
                 'pathString' => '/foo',
+                'expectedPath' => '/foo',
+            ],
+            'percent-encode unicode characters' => [
+                'path' => '/Nattō',
+                'expectedPath' => '/Natt%C5%8D',
             ],
         ];
     }