Issue #274: Some strings with default charset had UTF-8 characters an…

…d failed to convert
barbushin · May 11, 2019 · 04e01b9 · 04e01b9
1 parent aeecd5d
commit 04e01b9
Show file tree

Hide file tree

Showing 3 changed files with 13 additions and 3 deletions.
diff --git a/src/PhpImap/DataPartInfo.php b/src/PhpImap/DataPartInfo.php
@@ -36,6 +36,9 @@ function fetch() {
 		}
 
 		switch($this->encoding) {
+			case ENC7BIT:
+				$this->data = $this->data;
+				break;
 			case ENC8BIT:
 				$this->data = imap_utf8($this->data);
 				break;
@@ -49,10 +52,16 @@ function fetch() {
 			case ENCQUOTEDPRINTABLE:
 				$this->data = quoted_printable_decode($this->data);
 				break;
+			case ENCOTHER:
+				$this->data = $this->data;
+				break;
+			default:
+				$this->data = $this->data;
+				break;
 		}
 
 		if(isset($this->charset)) {
-			$this->data = $this->mail->convertStringEncoding($this->data, $this->charset, $this->mail->getServerEncoding());
+			$this->data = $this->mail->decodeMimeStr($this->data, $this->charset);
 		}
 
 		return $this->data;

diff --git a/src/PhpImap/Mailbox.php b/src/PhpImap/Mailbox.php
@@ -949,9 +949,9 @@ public function decodeMimeStr($string, $toCharset = 'utf-8') {
 		$newString = '';
 		foreach(imap_mime_header_decode($string) as $element) {
 			if(isset($element->text)) {
-				$fromCharset = !isset($element->charset) || $element->charset == 'default' ? 'iso-8859-1' : $element->charset;
+				$fromCharset = !isset($element->charset) ? 'iso-8859-1' : $element->charset;
 				// Convert to UTF-8, if string has UTF-8 characters to avoid broken strings. See https://github.com/barbushin/php-imap/issues/232
-				$toCharset = isset($element->charset) && $element->charset == 'UTF-8' ? 'UTF-8' : $toCharset;
+				$toCharset = isset($element->charset) && preg_match('/(UTF\-8)|(default)/i', $element->charset) ? 'UTF-8' : $toCharset;
 				$newString .= $this->convertStringEncoding($element->text, $fromCharset, $toCharset);
 			}
 		}

diff --git a/tests/unit/MailboxTest.php b/tests/unit/MailboxTest.php
@@ -608,6 +608,7 @@ public function testDecodeMimeStr() {
 			array("This is the Euro symbol ''.", "This is the Euro symbol ''."),
 			array('=?UTF-8?q?Some_subject_here_?= =?UTF-8?q?=F0=9F=98=98?=', 'Some subject here 😘', 'US-ASCII'),
 			array('=?UTF-8?Q?mountainguan=E6=B5=8B=E8=AF=95?=', 'mountainguan测试', 'US-ASCII'),
+			array('مقتطفات من: صن تزو. "فن الحرب". كتب أبل. Something in english', 'مقتطفات من: صن تزو. "فن الحرب". كتب أبل. Something in english', 'US-ASCII'),
 		);
 
 		foreach($test_strings as $test) {