| Author: Aleksander Machniak <alec@alec.pl> |
+-----------------------------------------------------------------------+
- $Id: rcube_imap.php 4389 2011-01-04 11:16:54Z alec $
+ $Id: rcube_imap.php 4643 2011-04-11 12:24:00Z alec $
*/
$this->options['port'] = $port;
- if ($this->options['debug'])
+ if ($this->options['debug']) {
$this->conn->setDebug(true, array($this, 'debug_handler'));
+ $this->options['ident'] = array(
+ 'name' => 'Roundcube Webmail',
+ 'version' => RCMAIL_VERSION,
+ 'php' => PHP_VERSION,
+ 'os' => PHP_OS,
+ 'command' => $_SERVER['REQUEST_URI'],
+ );
+ }
+
$attempt = 0;
do {
$data = rcmail::get_instance()->plugins->exec_hook('imap_connect',
return false;
}
- $struct = &$this->_structure_part($structure);
+ $struct = &$this->_structure_part($structure, 0, '', $headers);
$struct->headers = get_object_vars($headers);
// don't trust given content-type
$struct->charset = $struct->ctype_parameters['charset'];
}
+ // #1487700: workaround for lack of charset in malformed structure
+ if (empty($struct->charset) && !empty($mime_headers) && $mime_headers->charset) {
+ $struct->charset = $mime_headers->charset;
+ }
+
// read content encoding
if (!empty($part[5]) && $part[5]!='NIL') {
$struct->encoding = strtolower($part[5]);
$mime_headers = $this->conn->fetchPartHeader(
$this->mailbox, $this->_msg_id, false, $struct->mime_id);
}
- $struct->headers = $this->_parse_headers($mime_headers) + $struct->headers;
+
+ if (is_string($mime_headers))
+ $struct->headers = $this->_parse_headers($mime_headers) + $struct->headers;
+ else if (is_object($mime_headers))
+ $struct->headers = get_object_vars($mime_headers) + $struct->headers;
// get real content-type of message/rfc822
if ($struct->mimetype == 'message/rfc822') {
* @param rcube_message_part $o_part Part object created by get_structure()
* @param mixed $print True to print part, ressource to write part contents in
* @param resource $fp File pointer to save the message part
+ * @param boolean $skip_charset_conv Disables charset conversion
+ *
* @return string Message/part body if not printed
*/
- function &get_message_part($uid, $part=1, $o_part=NULL, $print=NULL, $fp=NULL)
+ function &get_message_part($uid, $part=1, $o_part=NULL, $print=NULL, $fp=NULL, $skip_charset_conv=false)
{
// get part encoding if not provided
if (!is_object($o_part)) {
return true;
}
- // convert charset (if text or message part) and part's charset is specified
- if ($body && $o_part->charset
- && preg_match('/^(text|message)$/', $o_part->ctype_primary)
+ // convert charset (if text or message part)
+ if ($body && !$skip_charset_conv &&
+ preg_match('/^(text|message)$/', $o_part->ctype_primary)
) {
+ if (!$o_part->charset || strtoupper($o_part->charset) == 'US-ASCII') {
+ $o_part->charset = $this->default_charset;
+ }
$body = rcube_charset_convert($body, $o_part->charset);
}
*/
public static function decode_mime_string($input, $fallback=null)
{
- // Initialize variable
- $out = '';
+ if (!empty($fallback)) {
+ $default_charset = $fallback;
+ }
+ else {
+ $default_charset = rcmail::get_instance()->config->get('default_charset', 'ISO-8859-1');
+ }
- // Iterate instead of recursing, this way if there are too many values we don't have stack overflows
// rfc: all line breaks or other characters not found
// in the Base64 Alphabet must be ignored by decoding software
// delete all blanks between MIME-lines, differently we can
// receive unnecessary blanks and broken utf-8 symbols
$input = preg_replace("/\?=\s+=\?/", '?==?', $input);
- // Check if there is stuff to decode
- if (strpos($input, '=?') !== false) {
- // Loop through the string to decode all occurences of =? ?= into the variable $out
- while(($pos = strpos($input, '=?')) !== false) {
- // Append everything that is before the text to be decoded
- $out .= substr($input, 0, $pos);
+ // encoded-word regexp
+ $re = '/=\?([^?]+)\?([BbQq])\?([^?\n]*)\?=/';
- // Get the location of the text to decode
- $end_cs_pos = strpos($input, "?", $pos+2);
- $end_en_pos = strpos($input, "?", $end_cs_pos+1);
- $end_pos = strpos($input, "?=", $end_en_pos+1);
+ // Find all RFC2047's encoded words
+ if (preg_match_all($re, $input, $matches, PREG_OFFSET_CAPTURE | PREG_SET_ORDER)) {
+ // Initialize variables
+ $tmp = array();
+ $out = '';
+ $start = 0;
- // Extract the encoded string
- $encstr = substr($input, $pos+2, ($end_pos-$pos-2));
- // Extract the remaining string
- $input = substr($input, $end_pos+2);
-
- // Decode the string fragement
- $out .= rcube_imap::_decode_mime_string_part($encstr);
- }
+ foreach ($matches as $idx => $m) {
+ $pos = $m[0][1];
+ $charset = $m[1][0];
+ $encoding = $m[2][0];
+ $text = $m[3][0];
+ $length = strlen($m[0][0]);
- // Deocde the rest (if any)
- if (strlen($input) != 0)
- $out .= rcube_imap::decode_mime_string($input, $fallback);
-
- // return the results
- return $out;
- }
+ // Append everything that is before the text to be decoded
+ if ($start != $pos) {
+ $substr = substr($input, $start, $pos-$start);
+ $out .= rcube_charset_convert($substr, $default_charset);
+ $start = $pos;
+ }
+ $start += $length;
+
+ // Per RFC2047, each string part "MUST represent an integral number
+ // of characters . A multi-octet character may not be split across
+ // adjacent encoded-words." However, some mailers break this, so we
+ // try to handle characters spanned across parts anyway by iterating
+ // through and aggregating sequential encoded parts with the same
+ // character set and encoding, then perform the decoding on the
+ // aggregation as a whole.
+
+ $tmp[] = $text;
+ if ($next_match = $matches[$idx+1]) {
+ if ($next_match[0][1] == $start
+ && $next_match[1][0] == $charset
+ && $next_match[2][0] == $encoding
+ ) {
+ continue;
+ }
+ }
- // no encoding information, use fallback
- return rcube_charset_convert($input,
- !empty($fallback) ? $fallback : rcmail::get_instance()->config->get('default_charset', 'ISO-8859-1'));
- }
+ $count = count($tmp);
+ $text = '';
+ // Decode and join encoded-word's chunks
+ if ($encoding == 'B' || $encoding == 'b') {
+ // base64 must be decoded a segment at a time
+ for ($i=0; $i<$count; $i++)
+ $text .= base64_decode($tmp[$i]);
+ }
+ else { //if ($encoding == 'Q' || $encoding == 'q') {
+ // quoted printable can be combined and processed at once
+ for ($i=0; $i<$count; $i++)
+ $text .= $tmp[$i];
- /**
- * Decode a part of a mime-encoded string
- *
- * @param string $str String to decode
- * @return string Decoded string
- * @access private
- */
- private function _decode_mime_string_part($str)
- {
- $a = explode('?', $str);
- $count = count($a);
+ $text = str_replace('_', ' ', $text);
+ $text = quoted_printable_decode($text);
+ }
- // should be in format "charset?encoding?base64_string"
- if ($count >= 3) {
- for ($i=2; $i<$count; $i++)
- $rest .= $a[$i];
+ $out .= rcube_charset_convert($text, $charset);
+ $tmp = array();
+ }
- if (($a[1]=='B') || ($a[1]=='b'))
- $rest = base64_decode($rest);
- else if (($a[1]=='Q') || ($a[1]=='q')) {
- $rest = str_replace('_', ' ', $rest);
- $rest = quoted_printable_decode($rest);
+ // add the last part of the input string
+ if ($start != strlen($input)) {
+ $out .= rcube_charset_convert(substr($input, $start), $default_charset);
}
- return rcube_charset_convert($rest, $a[0]);
+ // return the results
+ return $out;
}
- // we dont' know what to do with this
- return $str;
+ // no encoding information, use fallback
+ return rcube_charset_convert($input, $default_charset);
}
private function _parse_address_list($str, $decode=true)
{
// remove any newlines and carriage returns before
- $a = rcube_explode_quoted_string('[,;]', preg_replace( "/[\r\n]/", " ", $str));
+ $str = preg_replace('/\r?\n(\s|\t)?/', ' ', $str);
+
+ // extract list items, remove comments
+ $str = self::explode_header_string(',;', $str, true);
$result = array();
- foreach ($a as $key => $val) {
+ foreach ($str as $key => $val) {
$name = '';
$address = '';
$val = trim($val);
}
+ /**
+ * Explodes header (e.g. address-list) string into array of strings
+ * using specified separator characters with proper handling
+ * of quoted-strings and comments (RFC2822)
+ *
+ * @param string $separator String containing separator characters
+ * @param string $str Header string
+ * @param bool $remove_comments Enable to remove comments
+ *
+ * @return array Header items
+ */
+ static function explode_header_string($separator, $str, $remove_comments=false)
+ {
+ $length = strlen($str);
+ $result = array();
+ $quoted = false;
+ $comment = 0;
+ $out = '';
+
+ for ($i=0; $i<$length; $i++) {
+ // we're inside a quoted string
+ if ($quoted) {
+ if ($str[$i] == '"') {
+ $quoted = false;
+ }
+ else if ($str[$i] == '\\') {
+ if ($comment <= 0) {
+ $out .= '\\';
+ }
+ $i++;
+ }
+ }
+ // we're inside a comment string
+ else if ($comment > 0) {
+ if ($str[$i] == ')') {
+ $comment--;
+ }
+ else if ($str[$i] == '(') {
+ $comment++;
+ }
+ else if ($str[$i] == '\\') {
+ $i++;
+ }
+ continue;
+ }
+ // separator, add to result array
+ else if (strpos($separator, $str[$i]) !== false) {
+ if ($out) {
+ $result[] = $out;
+ }
+ $out = '';
+ continue;
+ }
+ // start of quoted string
+ else if ($str[$i] == '"') {
+ $quoted = true;
+ }
+ // start of comment
+ else if ($remove_comments && $str[$i] == '(') {
+ $comment++;
+ }
+
+ if ($comment <= 0) {
+ $out .= $str[$i];
+ }
+ }
+
+ if ($out && $comment <= 0) {
+ $result[] = $out;
+ }
+
+ return $result;
+ }
+
+
/**
* This is our own debug handler for the IMAP connection
* @access public