add method to shorten a string to given byte size to Utils

Signed-off-by: Arthur Schiwon <blizzz@arthur-schiwon.de>

2 files changed, 31 insertions, 0 deletions

diff --git a/lib/public/Util.php b/lib/public/Util.php
index 5165846707a..103b65fe874 100644
--- a/lib/public/Util.php
+++ b/lib/public/Util.php
@@ -513,4 +513,28 @@ class Util {
 		}
 		return self::$needUpgradeCache;
 	}
+
+	/**
+	 * Sometimes a string has to be shortened to fit within a certain maximum
+	 * data length in bytes. substr() you may break multibyte characters,
+	 * because it operates on single byte level. mb_substr() operates on
+	 * characters, so does not ensure that the shortend string satisfies the
+	 * max length in bytes.
+	 *
+	 * For example, json_encode is messing with multibyte characters a lot,
+	 * replacing them with something along "\u1234".
+	 *
+	 * This function shortens the string with by $accurancy (-5) from
+	 * $dataLength characters, until it fits within $dataLength bytes.
+	 *
+	 * @since 23.0.0
+	 */
+	public static function shortenMultibyteString(string $subject, int $dataLength, int $accuracy = 5): string {
+		$temp = mb_substr($subject, 0, $dataLength);
+		// json encodes encapsulates the string in double quotes, they need to be substracted
+		while ((strlen(json_encode($temp)) - 2) > $dataLength) {
+			$temp = mb_substr($temp, 0, -$accuracy);
+		}
+		return $temp;
+	}
 }
diff --git a/tests/lib/UtilTest.php b/tests/lib/UtilTest.php
index bb328c5998d..ca7a4ad1442 100644
--- a/tests/lib/UtilTest.php
+++ b/tests/lib/UtilTest.php
@@ -310,4 +310,11 @@ class UtilTest extends \Test\TestCase {
 			'myApp/vendor/myFancyCSSFile2',
 		], \OC_Util::$styles);
 	}
+
+	public function testShortenMultibyteString() {
+		$this->assertEquals('Short nuff', \OCP\Util::shortenMultibyteString('Short nuff', 255));
+		$this->assertEquals('ABC', \OCP\Util::shortenMultibyteString('ABCDEF', 3));
+		// each of the characters is 12 bytes
+		$this->assertEquals('🙈', \OCP\Util::shortenMultibyteString('🙈🙊🙉', 16, 2));
+	}
 }