Lucas_Werkmeister_WMDE created this task. Lucas_Werkmeister_WMDE added projects: PHP 8.2 support, Wikidata, wdwb-tech.
TASK DESCRIPTION `utf8_encode()` and `utf8_decode()` are deprecated in PHP 8.2, because their names are quite misleading (RFC <https://wiki.php.net/rfc/remove_utf8_decode_and_utf8_encode>): they convert strings from Latin 1 to UTF-8 and vice versa. Wikibase uses them in `SimpleCacheWithBagOStuff` (since the original revision of that class): private function serialize( $value ) { $serializedValue = serialize( $value ); $dataToStore = utf8_encode( $serializedValue ); $signature = hash_hmac( 'sha256', $dataToStore, $this->secret ); return json_encode( [ $signature, $dataToStore ] ); } private function unserialize( $string, $default, array $loggingContext ) { $result = json_decode( $string ); // ... $decodedData = utf8_decode( $data ); if ( $decodedData === serialize( false ) ) { return false; } // phpcs:disable Generic.PHP.NoSilencedErrors.Discouraged $value = @unserialize( $decodedData, [ 'allowed_classes' => [ \stdClass::class ] ] ); // ... return $value; } I’m pretty sure this use of the functions is pointless (though “correct” in that the methods perform their inverse operations). MediaWiki sets `mb_internal_encoding( 'UTF-8' );` early in `Setup.php` (since T137509 <https://phabricator.wikimedia.org/T137509>), so there is no reason to assume that PHP `serialize()` returns Latin 1 that needs to be UTF-8 encoded before being passed into `json_encode()` – we’re just putting JSONified mojibake in the inner cache. name=shell.php > $c = new HashBagOStuff() > ( new Wikibase\Lib\SimpleCacheWithBagOStuff( $c, 'prefix', 'secret' ) )->set( 'key', 'äöü' ) > $c->get( $c->makeKey( 'prefix', 'key' ) ) = "["9b94f92e4cfdcd3a282fc9398acd9a77f6d74d72fe95930c67e3edc44ece2745","s:6:\"\u00c3\u00a4\u00c3\u00b6\u00c3\u00bc\";"]" > json_decode( $c->get( $c->makeKey( 'prefix', 'key' ) ) )[1] = "s:6:"äöü";" We should remove both method calls. (But we’ll need to tweak the cache key at the same time, to ensure that old mojibake serializations don’t get deserialized without the `utf8_decode()`.) TASK DETAIL https://phabricator.wikimedia.org/T324202 EMAIL PREFERENCES https://phabricator.wikimedia.org/settings/panel/emailpreferences/ To: Lucas_Werkmeister_WMDE Cc: Umherirrender, Aklapper, Lucas_Werkmeister_WMDE, Astuthiodit_1, karapayneWMDE, Invadibot, maantietaja, ItamarWMDE, Akuckartz, Nandana, Lahi, Gq86, GoranSMilovanovic, TK-999, QZanden, LawExplorer, _jensen, rosalieper, Scott_WUaS, Wikidata-bugs, aude, Mbch331
_______________________________________________ Wikidata-bugs mailing list -- wikidata-bugs@lists.wikimedia.org To unsubscribe send an email to wikidata-bugs-le...@lists.wikimedia.org
