Ticket #17491: #17491.patch

wp-includes/formatting.php

@@ -1532 +1532 @@
 /**
  * Verifies that an email is valid.
  *
- * Does not grok i18n domains. Not RFC compliant.
+ * To comply with addr-spec of RFC 2822. (3.4.1. Addr-spec specification page 16)
+ * code from swiftmailer.org
  *
- * @since 0.71
+ * @since 3.4
  *
  * @param string $email Email address to verify.
  * @param boolean $deprecated Deprecated.
@@ -1544 +1545 @@
         if ( ! empty( $deprecated ) )
                 _deprecated_argument( __FUNCTION__, '3.0' );
 
-        // Test for the minimum length the email can be
-        if ( strlen( $email ) < 3 ) {
-                return apply_filters( 'is_email', false, $email, 'email_too_short' );
-        }
+        /*** Refer to RFC 2822 for ABNF grammar ***/
+    
+        //All basic building blocks
+        $rfc2822['NO-WS-CTL'] = '[\x01-\x08\x0B\x0C\x0E-\x19\x7F]';
+        $rfc2822['WSP'] = '[ \t]';
+        $rfc2822['CRLF'] = '(?:\r\n)';
+        $rfc2822['FWS'] = '(?:(?:' . $rfc2822['WSP'] . '*' . $rfc2822['CRLF'] . ')?' . $rfc2822['WSP'] . ')';
+        $rfc2822['text'] = '[\x00-\x08\x0B\x0C\x0E-\x7F]';
+        $rfc2822['quoted-pair'] = '(?:\\\\' . $rfc2822['text'] . ')';
+        $rfc2822['ctext'] = '(?:' . $rfc2822['NO-WS-CTL'] . '|[\x21-\x27\x2A-\x5B\x5D-\x7E])';
+        //Uses recursive PCRE (?1) -- could be a weak point??
+        $rfc2822['ccontent'] = '(?:' . $rfc2822['ctext'] . '|' . $rfc2822['quoted-pair'] . '|(?1))';
+        $rfc2822['comment'] = '(\((?:' . $rfc2822['FWS'] . '|' . $rfc2822['ccontent']. ')*' . $rfc2822['FWS'] . '?\))';
+        $rfc2822['CFWS'] = '(?:(?:' . $rfc2822['FWS'] . '?' . $rfc2822['comment'] . ')*(?:(?:' . $rfc2822['FWS'] . '?' . $rfc2822['comment'] . ')|' . $rfc2822['FWS'] . '))';
+        $rfc2822['qtext'] = '(?:' . $rfc2822['NO-WS-CTL'] . '|[\x21\x23-\x5B\x5D-\x7E])';
+        $rfc2822['qcontent'] = '(?:' . $rfc2822['qtext'] . '|' . $rfc2822['quoted-pair'] . ')';
+        $rfc2822['quoted-string'] = '(?:' . $rfc2822['CFWS'] . '?"' . '(' . $rfc2822['FWS'] . '?' . $rfc2822['qcontent'] . ')*' . $rfc2822['FWS'] . '?"' . $rfc2822['CFWS'] . '?)';
+        $rfc2822['atext'] = '[a-zA-Z0-9!#\$%&\'\*\+\-\/=\?\^_`\{\}\|~]';
+        $rfc2822['atom'] = '(?:' . $rfc2822['CFWS'] . '?' . $rfc2822['atext'] . '+' . $rfc2822['CFWS'] . '?)';
+        $rfc2822['dot-atom-text'] = '(?:' . $rfc2822['atext'] . '+' . '(\.' . $rfc2822['atext'] . '+)*)';
+        $rfc2822['dot-atom'] = '(?:' . $rfc2822['CFWS'] . '?' . $rfc2822['dot-atom-text'] . '+' . $rfc2822['CFWS'] . '?)';
+        $rfc2822['word'] = '(?:' . $rfc2822['atom'] . '|' . $rfc2822['quoted-string'] . ')';
+        $rfc2822['phrase'] = '(?:' . $rfc2822['word'] . '+?)';
+        $rfc2822['no-fold-quote'] = '(?:"(?:' . $rfc2822['qtext'] . '|' . $rfc2822['quoted-pair'] . ')*")';
+        $rfc2822['dtext'] = '(?:' . $rfc2822['NO-WS-CTL'] . '|[\x21-\x5A\x5E-\x7E])';
+        $rfc2822['no-fold-literal'] = '(?:\[(?:' . $rfc2822['dtext'] . '|' . $rfc2822['quoted-pair'] . ')*\])';
+        //Message IDs
+        $rfc2822['id-left'] = '(?:' . $rfc2822['dot-atom-text'] . '|' . $rfc2822['no-fold-quote'] . ')';
+        $rfc2822['id-right'] = '(?:' . $rfc2822['dot-atom-text'] . '|' . $rfc2822['no-fold-literal'] . ')';
+        //Addresses, mailboxes and paths
+        $rfc2822['local-part'] = '(?:' . $rfc2822['dot-atom'] . '|' . $rfc2822['quoted-string'] . ')';
+        $rfc2822['dcontent'] = '(?:' . $rfc2822['dtext'] . '|' . $rfc2822['quoted-pair'] . ')';
+        $rfc2822['domain-literal'] = '(?:' . $rfc2822['CFWS'] . '?\[(' . $rfc2822['FWS'] . '?' . $rfc2822['dcontent'] . ')*?' . $rfc2822['FWS'] . '?\]' . $rfc2822['CFWS'] . '?)';
+        $rfc2822['domain'] = '(?:' . $rfc2822['dot-atom'] . '|' . $rfc2822['domain-literal'] . ')';
+        $rfc2822['addr-spec'] = '(?:' . $rfc2822['local-part'] . '@' . $rfc2822['domain'] . ')';
 
-        // Test for an @ character after the first position
-        if ( strpos( $email, '@', 1 ) === false ) {
-                return apply_filters( 'is_email', false, $email, 'email_no_at' );
-        }
+        if (preg_match( '/^' . $rfc2822['addr-spec'] . '$/D', $email)) return apply_filters( 'is_email', $email, $email, null );
 
-        // Split out the local and domain parts
-        list( $local, $domain ) = explode( '@', $email, 2 );
-
-        // LOCAL PART
-        // Test for invalid characters
-        if ( !preg_match( '/^[a-zA-Z0-9!#$%&\'*+\/=?^_`{|}~\.-]+$/', $local ) ) {
-                return apply_filters( 'is_email', false, $email, 'local_invalid_chars' );
-        }
-
-        // DOMAIN PART
-        // Test for sequences of periods
-        if ( preg_match( '/\.{2,}/', $domain ) ) {
-                return apply_filters( 'is_email', false, $email, 'domain_period_sequence' );
-        }
-
-        // Test for leading and trailing periods and whitespace
-        if ( trim( $domain, " \t\n\r\0\x0B." ) !== $domain ) {
-                return apply_filters( 'is_email', false, $email, 'domain_period_limits' );
-        }
-
-        // Split the domain into subs
-        $subs = explode( '.', $domain );
-
-        // Assume the domain will have at least two subs
-        if ( 2 > count( $subs ) ) {
-                return apply_filters( 'is_email', false, $email, 'domain_no_periods' );
-        }
-
-        // Loop through each sub
-        foreach ( $subs as $sub ) {
-                // Test for leading and trailing hyphens and whitespace
-                if ( trim( $sub, " \t\n\r\0\x0B-" ) !== $sub ) {
-                        return apply_filters( 'is_email', false, $email, 'sub_hyphen_limits' );
-                }
-
-                // Test for invalid characters
-                if ( !preg_match('/^[a-z0-9-]+$/i', $sub ) ) {
-                        return apply_filters( 'is_email', false, $email, 'sub_invalid_chars' );
-                }
-        }
-
-        // Congratulations your email made it!
-        return apply_filters( 'is_email', $email, $email, null );
+        return apply_filters( 'is_email', false, $email, 'not_RFC_2822_compliant' );
 }
 
 /**