Chinese language negotiation

fisharebest · Mar 9, 2019 · e384ba5 · e384ba5
1 parent bb79927
commit e384ba5
Show file tree

Hide file tree

Showing 2 changed files with 113 additions and 8 deletions.
diff --git a/src/Locale.php b/src/Locale.php
@@ -14,6 +14,19 @@
  */
 class Locale
 {
+    /**
+     * Some browsers let the user choose "Chinese, Traditional", but add headers for "zh-HK"...
+     *
+     * @var array
+     */
+    private static $http_accept_chinese = array(
+        'zh-cn' => 'zh-hans-cn',
+        'zh-sg' => 'zh-hans-sg',
+        'zh-hk' => 'zh-hant-hk',
+        'zh-mo' => 'zh-hant-mo',
+        'zh-tw' => 'zh-hant-tw',
+    );
+
     /**
      * Callback for PHP sort functions - allows lists of locales to be sorted.
      * Diacritics are removed and text is capitalized to allow fast/simple sorting.
@@ -44,9 +57,9 @@ public static function create($code)
 
         if (class_exists($class)) {
             return new $class();
-        } else {
-            throw new DomainException($code);
         }
+
+        throw new DomainException($code);
     }
 
     /**
@@ -78,12 +91,9 @@ public static function httpAcceptLanguage(array $server, array $available, Local
                 return $x[0];
             }, $preferences);
 
-            // If "de-DE" requested, but not "de", then add it at a lower priority
-            foreach ($preferences as $code => $priority) {
-                if (preg_match('/^([a-z]+)[^a-z]/', $code, $match) && !isset($preferences[$match[1]])) {
-                    $preferences[$match[1]] = $priority * 0.5;
-                }
-            }
+            // "Common sense" logic for badly configured clients.
+            $preferences = self::httpAcceptChinese($preferences);
+            $preferences = self::httpAcceptDowngrade($preferences);
 
             foreach (array_keys($preferences) as $code) {
                 try {
@@ -99,4 +109,51 @@ public static function httpAcceptLanguage(array $server, array $available, Local
 
         return $default;
     }
+
+    /**
+     * If a client requests "de-DE" (but not "de"), then add "de" as a lower-priority fallback.
+     *
+     * @param $preferences
+     *
+     * @return int[]
+     */
+    private static function httpAcceptDowngrade($preferences)
+    {
+        foreach ($preferences as $code => $priority) {
+            // Three parts: "zh-hans-cn" => "zh-hans" and "zh"
+            if (preg_match('/^(([a-z]+)-[a-z]+)-[a-z]+$/', $code, $match)) {
+                if (!isset($preferences[$match[2]])) {
+                    $preferences[$match[2]] = $priority * 0.5;
+                }
+                if (!isset($preferences[$match[1]])) {
+                    $preferences[$match[1]] = $priority * 0.5;
+                }
+            }
+            // Two parts: "de-de" => "de"
+            if (preg_match('/^([a-z]+)-[a-z]+$/', $code, $match) && !isset($preferences[$match[1]])) {
+                $preferences[$match[1]] = $priority * 0.5;
+            }
+        }
+
+        return $preferences;
+    }
+
+    /**
+     * Some browsers allow the user to select "Chinese (simplified)", but then use zh-CN instead of zh-Hans.
+     * This goes against the advice of w3.org.
+     *
+     * @param int[] $preferences
+     *
+     * @return int[]
+     */
+    private static function httpAcceptChinese($preferences)
+    {
+        foreach (self::$http_accept_chinese as $old => $new) {
+            if (array_key_exists($old, $preferences) && !array_key_exists($new, $preferences)) {
+                $preferences[$new] = $preferences[$old] * 0.5;
+            }
+        }
+
+        return $preferences;
+    }
 }
diff --git a/test/LocaleTest.php b/test/LocaleTest.php
@@ -138,6 +138,54 @@ public function testHttpAcceptLanguageDowngrade()
         $this->assertEquals(Locale::create('de'), $locale);
     }
 
+    /**
+     * Test language negotiation
+     */
+    public function testHttpAcceptLanguageDoubleDowngrade()
+    {
+        $available = array(
+            Locale::create('zh'),
+        );
+        $server    = array('HTTP_ACCEPT_LANGUAGE' => 'zh-Hans-CN');
+        $default   = Locale::create('en-US');
+        $locale    = Locale::httpAcceptLanguage($server, $available, $default);
+
+        $this->assertEquals(Locale::create('zh'), $locale);
+    }
+
+    /**
+     * Test language negotiation
+     */
+    public function testHttpAcceptChinese()
+    {
+        $available = array(
+            Locale::create('zh-Hans'),
+            Locale::create('zh-Hant'),
+        );
+
+        $default   = Locale::create('en-US');
+
+        $server    = array('HTTP_ACCEPT_LANGUAGE' => 'zh-CN');
+        $locale    = Locale::httpAcceptLanguage($server, $available, $default);
+        $this->assertEquals(Locale::create('zh-Hans'), $locale);
+
+        $server    = array('HTTP_ACCEPT_LANGUAGE' => 'zh-SG');
+        $locale    = Locale::httpAcceptLanguage($server, $available, $default);
+        $this->assertEquals(Locale::create('zh-Hans'), $locale);
+
+        $server    = array('HTTP_ACCEPT_LANGUAGE' => 'zh-HK');
+        $locale    = Locale::httpAcceptLanguage($server, $available, $default);
+        $this->assertEquals(Locale::create('zh-Hant'), $locale);
+
+        $server    = array('HTTP_ACCEPT_LANGUAGE' => 'zh-MO');
+        $locale    = Locale::httpAcceptLanguage($server, $available, $default);
+        $this->assertEquals(Locale::create('zh-Hant'), $locale);
+
+        $server    = array('HTTP_ACCEPT_LANGUAGE' => 'zh-TW');
+        $locale    = Locale::httpAcceptLanguage($server, $available, $default);
+        $this->assertEquals(Locale::create('zh-Hant'), $locale);
+    }
+
     /**
      * Test language negotiation
      */