improve validation and escaping of e-mail addresses: use same method for validation in check_email and address_error; encode and validate IDNs; make (more) robust functions to split address lists and emails

2 files changed, 140 insertions, 33 deletions

diff --git a/it_mail.class b/it_mail.class
index 4119994..94801a5 100644
--- a/it_mail.class
+++ b/it_mail.class
@@ -301,6 +301,70 @@ function send($p = array())
 
 
 /**
+ * INTERNAL: Split address lists in headers like From, To, Cc, Bcc into
+ * name/email pairs
+ * @param $string String containing address list
+ * @return Generator returning name/email tuples
+ */
+static function addrlist_split($string)
+{
+	$quoted = false;
+	foreach (str_split($string) as $char)
+	{
+		if ($char == '"')
+			$quoted = !$quoted;
+
+		if ($char == ',' && !$quoted)
+			$n++;
+		else
+			$mailboxes[$n] .= $char;
+	}
+
+	foreach ((array)$mailboxes as $mailbox)
+	{
+		$mode = 'name';
+		$name = $email = '';
+		$quoted = false;
+		foreach (str_split($mailbox) as $char)
+		{
+			if ($char == '"')
+				$quoted = !$quoted;
+			switch ($mode)
+			{
+			case 'name':
+				if ($char == '<' && !$quoted)
+					$mode = 'email';
+				else
+					$name .= $char;
+				break;
+			case 'email':
+				if ($char == '>' && !$quoted)
+				{
+					$mode = 'finish';
+					break 2;
+				}
+				else
+					$email .= $char;
+				break;
+			}
+		}
+		if ($mode == 'email')
+		{
+			$name = $name . '<' . $email;
+			$email = '';
+			$mode = 'name';
+		}
+		if ($mode == 'name' && !$email)
+		{
+			$email = it::replace(['^"(.*)"$' => '$1'], trim($name));
+			$name = '';
+		}
+		yield [$name, trim($email)];
+	}
+}
+
+
+/**
  * INTERNAL: Escape header line with ?=?<charset>?Q if necessary, e.g. in Subject line
  * @param $string String to be escaped
  * @return String escape suitable for sending in header line
@@ -328,23 +392,13 @@ function addrlist_escape($string)
 {
 	# Exclude e-mail addresses from being encoded as
 	# e.g. GMail or Exchange have problems with that
-	foreach (str_split($string) as $char)
+	foreach (self::addrlist_split($string) as list($name, $email))
 	{
-		if ($char == '"')
-			$quoted = !$quoted;
-
-		if ($char == ',' && !$quoted)
-			$n++;
-		else
-			$mailboxes[$n] .= $char;
-	}
-
-	foreach ((array)$mailboxes as $mailbox)
-	{
-		if (preg_match('/^(.*)(\s+?<[^>]+@[^>]+>\s*)$/', $mailbox, $matches))
-			$result[] = $this->header_escape(trim($matches[1])) . $matches[2];
-		else
-			$result[] = trim($mailbox);
+		$email = self::email_escape($email);
+		if ($name && $email)
+			$result[] = $this->header_escape(trim($name)) . " <$email>";
+		else if (!$name && $email)
+			$result[] = $email;
 	}
 
 	return implode(', ', $result);
@@ -364,6 +418,48 @@ static function fullname_escape($string)
 
 
 /**
+ * INTERNAL: Split email-address into local part and doomain
+ * @return Array with to elements: local part and domain
+ */
+static function email_split($email)
+{
+	$mode = 'local';
+	$local = $domain = '';
+	$quoted = false;
+	foreach (str_split($email) as $char)
+	{
+		if ($char == '"')
+			$quoted = !$quoted;
+		switch ($mode)
+		{
+		case 'local':
+			if ($char == '@' && !$quoted)
+				$mode = 'domain';
+			else
+				$local .= $char;
+			break;
+		case 'domain':
+			$domain .= $char;
+		}
+	}
+
+	return [$local, $domain];
+}
+
+
+/**
+ * INTERNAL: Convert domain part of email address into ascii idn form
+ * @return String with converted email address
+ */
+static function email_escape($email)
+{
+	list($local, $domain) = self::email_split($email);
+
+	return $local . ($domain ? '@' . idn_to_ascii($domain) : '');
+}
+
+
+/**
  * INTERNAL: Send SMTP command and return true if result is 2XX
  * @param $fp Connection to SMTP server opened using fsockopen
  * @param $command Command to send, e.g. HELO gna.ch
@@ -408,20 +504,12 @@ static function send_smtp_cmd($fp, $cmd, &$answer, $timeoutok = false, $failcode
  */
 static function address_error($addresslist)
 {
-	foreach (str_split($addresslist) as $char)
+	foreach (self::addrlist_split($addresslist) as list($dummy, $email))
 	{
-		if ($char == '"')
-			$quoted = !$quoted;
-		else if ($char == ',' && !$quoted)
-			$n++;
-		else if (!$quoted)
-			$addresses[$n] .= $char;
+		$email = self::email_escape($email);
+		if (filter_var($email, FILTER_VALIDATE_EMAIL) === false && !it::match('^\s*[a-z][-a-z0-9]*\s*$', $email))
+			return "invalid format on $email";
 	}
-
-	foreach ($addresses as $address)
-		if (($email = it::match('<([^>]*)>', $address) ?? trim($address)))
-			if (filter_var($email, FILTER_VALIDATE_EMAIL) === false && !it::match('^\s*[a-z][-a-z0-9]*\s*$', $email))
-				return "invalid format on $email";
 }
 
 
@@ -438,11 +526,11 @@ static function check_email($email, $checkmailbox = false)
 {
 	$result = IT_MAIL_CHECKEMAIL_INVALID;
 
+	$email = self::email_escape($email);
+	list($dummy, $domain) = self::email_split($email);
 	/* Check if username starts with www. or not well-formed => reject */
-	if (!preg_match('/^www\./', $email) && preg_match('/^[a-z0-9&_+.-]+@([a-z0-9.-]+\.[a-z]+)$/i', $email, $regs))
+	if (!preg_match('/^www\./', $email) && filter_var($email, FILTER_VALIDATE_EMAIL) !== false && $domain)
 	{
-		$domain = $regs[1];
-
 		if (!getmxrr($domain, $mxhosts, $weights))
 		{
 			/* If we find no MX we check if domain exists */
diff --git a/test/it_mail.t b/test/it_mail.t
index 7e892c2..6aefd84 100755
--- a/test/it_mail.t
+++ b/test/it_mail.t
@@ -56,6 +56,12 @@ is(
 	"Don't remove quoting characters from realname when it contains a quotable character"
 );
 
+is(
+	$mail->addrlist_escape('"Alfred E. Neuman" <t€st@sör.ch>', true),
+	'"Alfred E. Neuman" <t€st@xn--sr-fka.ch>',
+	"Convert domain in email to idn version"
+);
+
 $mail = new it_mail([
 	'From' => 'Someone Ïmportant <ïmportant@search.ch>',
 	'To' => 'éxample@example.com, example@example.com, Sömeone Ëlse <sömeone@example.com>, "Alfred E. Neuman" <neuman@example.com>, "Schmitt, Sören" <schmitt@example.com>',
@@ -95,10 +101,23 @@ is(
 );
 
 is(
-	it_mail::address_error('neuman@example.com, <neuman@example.com>, "Neuman, Alfred E." <neuman@example.com>, aneuman, "<neuman@example>" <neuman@example.com>', " neumann@example.com"),
+	it_mail::address_error('neuman@example.com, <neuman@example.com>, "Neuman, Alfred E." <neuman@example.com>, aneuman, "<neuman@example>" <neuman@example.com>, " neumann@example.com", "www@search.ch".foo.bar@local.ch, test@sör.ch'),
 	null,
 	"Accept valid e-mail addresses"
 );
 
-foreach (['neuman@example', '"neuman@example.com" <neuman@example>', '<neuma@example.com'] as $email)
+foreach (['neuman@example', '"neuman@example.com" <neuman@example>', '<neuma@example.com', 'shaqroot.@yahoo.com'] as $email)
+{
 	isnt(it_mail::address_error($email), null, "Reject invalid e-mail address $email");
+	foreach (it_mail::addrlist_split($email) as list($dummy, $email))
+		is(it_mail::check_email($email), IT_MAIL_CHECKEMAIL_INVALID, "check_email rejects invalid email address $email too");
+}
+
+is(it_mail::check_email('test@sör.ch'), IT_MAIL_CHECKEMAIL_OK, "check_email accepts valid email address with idn domain");
+
+is(it_mail::email_split('test@sör.ch'), ['test', 'sör.ch'], 'email_split test@sör.ch');
+is(it_mail::email_split('"test@local.ch".foo.bar@search.ch'), ['"test@local.ch".foo.bar', 'search.ch'], 'email_split "test@local.ch".foo.bar@search.ch');
+is(it_mail::email_split('aneuman'), ['aneuman', ''], 'email_split aneuman');
+
+is(it_mail::email_escape('test@sör.ch'), 'test@xn--sr-fka.ch', 'idn encode domain sör.ch');
+is(it_mail::email_escape('أحمد@مثال.آزمایشی'), 'أحمد@xn--mgbh0fb.xn--hgbk6aj7f53bba', 'idn encode domain ﻢﺛﺎﻟ.ﺁﺰﻣﺎﯿﺸﯾ');