Language::isValidBuiltInCode() should not accept uppercase input

The results of this function are used to decide whether a code is
valid for loading an i18n file without any further normalization.

Partially reverts 93348f3 which made the regular expression case-
insensitive. Per IRC discussion, language codes should always be
lowercase and it's up to callers to deal with that.

Change-Id: I8975c3374a37935080d9f7eca6a602e32f67a87b

diff --git a/languages/Language.php b/languages/Language.php
index 0a19d61..887490b 100644 (file)
--- a/languages/Language.php
+++ b/languages/Language.php
@@ -339,7 +339,7 @@ class Language {
                        throw new MWException( __METHOD__ . " must be passed a string, $type given$addmsg" );
                }
 
-               return (bool)preg_match( '/^[a-z0-9-]{2,}$/i', $code );
+               return (bool)preg_match( '/^[a-z0-9-]{2,}$/', $code );
        }
 
        /**

diff --git a/tests/phpunit/languages/LanguageTest.php b/tests/phpunit/languages/LanguageTest.php
index ef670df..ec51441 100644 (file)
--- a/tests/phpunit/languages/LanguageTest.php
+++ b/tests/phpunit/languages/LanguageTest.php
@@ -457,32 +457,22 @@ class LanguageTest extends LanguageClassesTestCase {
         * @dataProvider provideLanguageCodes
         * @covers Language::isValidBuiltInCode
         */
-       public function testBuiltInCodeValidation( $code, $message = '' ) {
-               $this->assertTrue(
+       public function testBuiltInCodeValidation( $code, $expected, $message = '' ) {
+               $this->assertEquals( $expected,
                        (bool)Language::isValidBuiltInCode( $code ),
                        "validating code $code $message"
                );
        }
 
-       /**
-        * @covers Language::isValidBuiltInCode
-        */
-       public function testBuiltInCodeValidationRejectUnderscore() {
-               $this->assertFalse(
-                       (bool)Language::isValidBuiltInCode( 'be_tarask' ),
-                       "reject underscore in language code"
-               );
-       }
-
        public static function provideLanguageCodes() {
                return array(
-                       array( 'fr', 'Two letters, minor case' ),
-                       array( 'EN', 'Two letters, upper case' ),
-                       array( 'tyv', 'Three letters' ),
-                       array( 'tokipona', 'long language code' ),
-                       array( 'be-tarask', 'With dash' ),
-                       array( 'Zh-classical', 'Begin with upper case, dash' ),
-                       array( 'Be-x-old', 'With extension (two dashes)' ),
+                       array( 'fr', true, 'Two letters, minor case' ),
+                       array( 'EN', false, 'Two letters, upper case' ),
+                       array( 'tyv', true, 'Three letters' ),
+                       array( 'tokipona', true, 'long language code' ),
+                       array( 'be-tarask', true, 'With dash' ),
+                       array( 'be-x-old', true, 'With extension (two dashes)' ),
+                       array( 'be_tarask', false, 'Reject underscores' ),
                );
        }