Merge pull request #5882 from nextcloud/fix/detect-char-encoding

Change how charsets are detected for emails

lib/IMAP/Charset/Converter.php

@@ -0,0 +1,87 @@
+<?php
+
+declare(strict_types=1);
+
+/**
+ * @author Anna Larch <[email protected]>
+ *
+ * Mail
+ *
+ * This code is free software: you can redistribute it and/or modify
+ * it under the terms of the GNU Affero General Public License, version 3,
+ * as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU Affero General Public License for more details.
+ *
+ * You should have received a copy of the GNU Affero General Public License, version 3,
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>
+ *
+ */
+
+namespace OCA\Mail\IMAP\Charset;
+
+use Horde_Mime_Part;
+use OCA\Mail\Exception\ServiceException;
+use function in_array;
+use function is_string;
+
+class Converter {
+
+	/**
+	 * @param Horde_Mime_Part $p
+	 * @return string
+	 * @throws ServiceException
+	 */
+	public function convert(Horde_Mime_Part $p): string {
+		/** @var null|string $data */
+		$data = $p->getContents();
+		if ($data === null) {
+			return '';
+		}
+
+		// Only convert encoding if it is explicitly specified in the header because text/calendar
+		// data is utf-8 by default.
+		$charset = $p->getCharset();
+		if ($charset !== null && strtoupper($charset) === 'UTF-8') {
+			return $data;
+		}
+
+		// The part specifies a charset
+		if ($charset !== null) {
+			if (in_array($charset, mb_list_encodings(), true)) {
+				$converted = mb_convert_encoding($data, 'UTF-8', $charset);
+			} else {
+				$converted = iconv($charset, 'UTF-8', $data);
+			}
+
+			if (is_string($converted)) {
+				return $converted;
+			}
+		}
+
+		// No charset specified, let's ask mb if this could be UTF-8
+		$detectedCharset = mb_detect_encoding($data, 'UTF-8', true);
+		if ($detectedCharset === false) {
+			// Fallback, non UTF-8
+			$detectedCharset = mb_detect_encoding($data, null, true);
+		}
+		// Still UTF8, no need to convert
+		if ($detectedCharset !== false && strtoupper($detectedCharset) === 'UTF-8') {
+			return $data;
+		}
+
+		$converted = @mb_convert_encoding($data, 'UTF-8', $charset);
+		if ($converted === false) {
+			// Might be a charset that PHP mb doesn't know how to handle, fall back to iconv
+			$converted = iconv($charset, 'UTF-8', $data);
+		}
+
+		if (!is_string($converted)) {
+			throw new ServiceException('Could not detect message charset');
+		}
+		return $converted;
+	}
+}

lib/IMAP/ImapMessageFetcher.php

@@ -39,6 +39,7 @@
 use Horde_Mime_Part;
 use OCA\Mail\AddressList;
 use OCA\Mail\Exception\ServiceException;
+use OCA\Mail\IMAP\Charset\Converter;
 use OCA\Mail\Model\IMAPMessage;
 use OCA\Mail\Service\Html;
 use OCA\Mail\Service\SmimeService;
@@ -79,7 +80,8 @@ public function __construct(int $uid,
 		Horde_Imap_Client_Base $client,
 		string $userId,
 		Html $htmlService,
-		SmimeService $smimeService) {
+		SmimeService $smimeService,
+		private Converter $converter) {
 		$this->uid = $uid;
 		$this->mailbox = $mailbox;
 		$this->client = $client;
@@ -434,6 +436,7 @@ private function handleHtmlMessage(Horde_Mime_Part $p, string $partNo, bool $isF
 	 * @throws DoesNotExistException
 	 * @throws Horde_Imap_Client_Exception
 	 * @throws Horde_Imap_Client_Exception_NoSupportExtension
+	 * @throws ServiceException
 	 */
 	private function loadBodyData(Horde_Mime_Part $p, string $partNo, bool $isFetched): string {
 		if (!$isFetched) {
@@ -461,22 +464,10 @@ private function loadBodyData(Horde_Mime_Part $p, string $partNo, bool $isFetche
 			}
 
 			$data = $fetch->getBodyPart($partNo);
-
 			$p->setContents($data);
 		}
 
-		$data = $p->getContents();
-		if ($data === null) {
-			return '';
-		}
-
-		// Only convert encoding if it is explicitly specified in the header because text/calendar
-		// data is utf-8 by default.
-		$charset = $p->getContentTypeParameter('charset');
-		if ($charset !== null && strtoupper($charset) !== 'UTF-8') {
-			$data = mb_convert_encoding($data, 'UTF-8', $charset);
-		}
-		return (string)$data;
+		return $this->converter->convert($p);
 	}
 
 	private function hasAttachments(Horde_Mime_Part $part): bool {

lib/IMAP/ImapMessageFetcherFactory.php

@@ -27,17 +27,21 @@
 namespace OCA\Mail\IMAP;
 
 use Horde_Imap_Client_Base;
+use OCA\Mail\IMAP\Charset\Converter;
 use OCA\Mail\Service\Html;
 use OCA\Mail\Service\SmimeService;
 
 class ImapMessageFetcherFactory {
 	private Html $htmlService;
 	private SmimeService $smimeService;
+	private Converter $charsetConverter;
 
 	public function __construct(Html         $htmlService,
-		SmimeService $smimeService) {
+		SmimeService $smimeService,
+		Converter $charsetConverter) {
 		$this->htmlService = $htmlService;
 		$this->smimeService = $smimeService;
+		$this->charsetConverter = $charsetConverter;
 	}
 
 	public function build(int $uid,
@@ -51,6 +55,7 @@ public function build(int $uid,
 			$userId,
 			$this->htmlService,
 			$this->smimeService,
+			$this->charsetConverter,
 		);
 	}
 }

tests/Unit/IMAP/Charset/ConverterTest.php

@@ -0,0 +1,101 @@
+<?php
+
+declare(strict_types=1);
+
+/**
+ * @copyright 2023 Anna Larch <[email protected]>
+ * @author Anna Larch <[email protected]>
+ *
+ * Mail
+ *
+ * This code is free software: you can redistribute it and/or modify
+ * it under the terms of the GNU Affero General Public License, version 3,
+ * as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU Affero General Public License for more details.
+ *
+ * You should have received a copy of the GNU Affero General Public License, version 3,
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>
+ *
+ */
+
+namespace OCA\Mail\Tests\Unit\IMAP\Charset;
+
+use ChristophWurst\Nextcloud\Testing\TestCase;
+use Horde_Mime_Part;
+use OCA\Mail\IMAP\Charset\Converter;
+use function fopen;
+use function fwrite;
+use function mb_convert_encoding;
+
+class ConverterTest extends TestCase {
+
+	public Converter $converter;
+	protected function setUp(): void {
+		parent::setUp();
+
+		$this->converter = new Converter();
+	}
+
+	/**
+	 * @dataProvider dataProviderMimeParts
+	 */
+	public function testConvert($mimePart, $expected): void {
+		$actual = $this->converter->convert($mimePart);
+		$this->assertEquals($expected, $actual);
+		$isUtf8 = mb_check_encoding($actual, 'UTF-8');
+		$this->assertTrue($isUtf8);
+	}
+
+	public function dataProviderMimeParts(): array {
+		// UTF8
+		$utfMimePart = new Horde_Mime_Part();
+		$utfMimePart->setType('text/plain');
+		$utfMimePart->setCharset('UTF-8');
+		$utfMimePart->setContents('😊');
+		// UTF8 stream
+		$utfMimeStreamPart = new Horde_Mime_Part();
+		$utfMimeStreamPart->setType('text/plain');
+		$utfMimeStreamPart->setCharset('UTF-8');
+		$fh = fopen("php://temp", 'r+');
+		fwrite($fh, '💦');
+		$utfMimeStreamPart->setContents($fh, [ 'usestream' => true, ]);
+		// Hebrew
+		$iso88591MimePart = new Horde_Mime_Part();
+		$iso88591MimePart->setType('text/plain');
+		$iso88591MimePart->setCharset('ISO-8859-1');
+		$iso88591MimePart->setContents(mb_convert_encoding('Ümlaut', 'ISO-8859-1', 'UTF-8'));
+		$iso88591MimePart_noCharset = new Horde_Mime_Part();
+		$iso88591MimePart_noCharset->setContents('בה בדף לחבר ממונרכיה, בקר בגרסה ואמנות דת');
+		// Japanese
+		$iso2022jpMimePart = new Horde_Mime_Part();
+		$iso2022jpMimePart->setType('text/plain');
+		$iso2022jpMimePart->setCharset('ISO-2022-JP');
+		$iso2022jpMimePart->setContents(mb_convert_encoding('外せ園査リツハワ題', 'ISO-2022-JP', 'UTF-8'));
+		$iso2022jpMimePart_noCharset = new Horde_Mime_Part();
+		$iso2022jpMimePart_noCharset->setContents('外せ園査リツハワ題');
+		// Korean - not in mb nor iconv
+		// $iso106461MimePart = new Horde_Mime_Part();
+		// $iso106461MimePart->setType('text/plain');
+		// $iso106461MimePart->setCharset('ISO 10646-1');
+		//$iso106461MimePart->setContents(iconv('UTF-8', 'ISO 10646-1', '언론·출판은 타인의 명'));
+		// Arabic - not in mb
+		$windowsMimePart = new Horde_Mime_Part();
+		$windowsMimePart->setType('text/plain');
+		$windowsMimePart->setCharset('Windows-1256');
+		$windowsMimePart->setContents(iconv('UTF-8', 'Windows-1256', 'قام زهاء أوراقهم ما,'));
+
+		return[
+			[$utfMimePart, '😊'],
+			[$utfMimeStreamPart, '💦'],
+			[$iso88591MimePart, 'Ümlaut'],
+			[$iso2022jpMimePart, '外せ園査リツハワ題'],
+			[$iso88591MimePart_noCharset, 'בה בדף לחבר ממונרכיה, בקר בגרסה ואמנות דת'],
+			// [$iso106461MimePart, '언론·출판은 타인의 명'],
+			[$windowsMimePart, 'قام زهاء أوراقهم ما,']
+		];
+	}
+}