class.mailparse.php

<?php
/*********************************************************************
    class.mailparse.php

    Mail parsing helper class.
    Mail parsing will change once we move to PHP5

    Peter Rotich <peter@osticket.com>
    Copyright (c)  2006-2013 osTicket
    http://www.osticket.com

    Released under the GNU General Public License WITHOUT ANY WARRANTY.
    See LICENSE.TXT for details.

    vim: expandtab sw=4 ts=4 sts=4:
**********************************************************************/

require_once(PEAR_DIR.'Mail/mimeDecode.php');
require_once(PEAR_DIR.'Mail/RFC822.php');

class Mail_Parse {

    var $mime_message;
    var $include_bodies;
    var $decode_headers;
    var $decode_bodies;

    var $struct;

    var $charset ='UTF-8'; //Default charset.

    function Mail_parse($mimeMessage, $charset=null){

        $this->mime_message = $mimeMessage;

        if($charset)
            $this->charset = $charset;

        $this->include_bodies = true;
        $this->decode_headers = true;
        $this->decode_bodies = true;

        //Desired charset
        if($charset)
            $this->charset = $charset;
    }

    function decode() {

        $params = array('crlf'          => "\r\n",
                        'charset'       => $this->charset,
                        'input'         => $this->mime_message,
                        'include_bodies'=> $this->include_bodies,
                        'decode_headers'=> $this->decode_headers,
                        'decode_bodies' => $this->decode_bodies);
        $this->splitBodyHeader();
        $this->struct=Mail_mimeDecode::decode($params);

        return (PEAR::isError($this->struct) || !(count($this->struct->headers)>1))?FALSE:TRUE;
    }

    function splitBodyHeader() {
        $match = array();
        if (preg_match("/^(.*?)\r?\n\r?\n(.*)/s",
                $this->mime_message,
                $match)) {
            $this->header=$match[1];
        }
    }
    /**
     * Takes the header section of an email message with the form of
     * Header: Value
     * and returns a hashtable of header-name => value pairs. Also, this
     * function properly handles header values that span multiple lines
     * (such as Content-Type).
     *
     * Specify $as_array to TRUE to keep all header values. If a header is
     * specified more than once, all the values are placed in an array under
     * the header key. If left as FALSE, only the value given in the last
     * occurance of the header is retained.
     */
    /* static */ function splitHeaders($headers_text, $as_array=false) {
        $headers = preg_split("/\r?\n/", $headers_text);
        for ($i=0, $k=count($headers); $i<$k; $i++) {
            # XXX: Might tabs be used here?
            if (substr($headers[$i], 0, 1) == " ") {
                # Continuation from previous header (runon to next line)
                $j=$i-1; while (!isset($headers[$j]) && $j>0) $j--;
                $headers[$j] .= " ".ltrim($headers[$i]);
                unset($headers[$i]);
            } elseif (strlen($headers[$i]) == 0) {
                unset($headers[$i]);
            }
        }
        $array = array();
        foreach ($headers as $hdr) {
            list($name, $val) = explode(": ", $hdr, 2);
            # Create list of values if header is specified more than once
            if ($array[$name] && $as_array) {
                if (is_array($array[$name])) $array[$name][] = $val;
                else $array[$name] = array($array[$name], $val);
            } else {
                $array[$name] = $val;
            }
        }
        return $array;
    }

    /* static */
    function findHeaderEntry($headers, $name) {
        if (!is_array($headers))
            $headers = self::splitHeaders($headers);
        foreach ($headers as $key=>$val)
            if (strcasecmp($key, $name) === 0)
                return $val;
        return false;
    }

    function getStruct(){
        return $this->struct;
    }

    function getHeader() {
        if(!$this->header) $this->splitBodyHeader();

        return $this->header;
    }

    function getError(){
        return PEAR::isError($this->struct)?$this->struct->getMessage():'';
    }


    function getFromAddressList(){
        if (!($header = $this->struct->headers['from']))
            return null;

        return Mail_Parse::parseAddressList($header);
    }

    function getToAddressList(){
        // Delivered-to incase it was a BBC mail.
        if (!($header = $this->struct->headers['to']))
            if (!($header = $this->struct->headers['delivered-to']))
                return null;

        return Mail_Parse::parseAddressList($header);
    }

    function getCcAddressList(){
        if (!($header = $this->struct->headers['cc']))
            return null;

        return Mail_Parse::parseAddressList($header);
    }

    function getMessageId(){
        return $this->struct->headers['message-id'];
    }

    function getSubject(){
        return $this->struct->headers['subject'];
    }

    function getReplyTo() {
        if (!($header = $this->struct->headers['reply-to']))
            return null;

        return Mail_Parse::parseAddressList($header);
    }

    function isBounceNotice() {
        if (!($body = $this->getPart($this->struct, 'message/delivery-status')))
            return false;

        $info = self::splitHeaders($body);
        if (!isset($info['Action']))
            return false;

        return strcasecmp($info['Action'], 'failed') === 0;
    }

    function getDeliveryStatusMessage() {
        $ctype = @strtolower($this->struct->ctype_primary.'/'.$this->struct->ctype_secondary);
        if ($ctype == 'multipart/report'
            && isset($this->struct->ctype_parameters['report-type'])
            && $this->struct->ctype_parameters['report-type'] == 'delivery-status'
        ) {
            return sprintf('<pre>%s</pre>',
                Format::htmlchars(
                    $this->getPart($this->struct, 'text/plain', 1)
                ));
        }
        return false;
    }

    function getOriginalMessage() {
        foreach ($this->struct->parts as $p) {
            $ctype = $p->ctype_primary.'/'.$p->ctype_secondary;
            if (strtolower($ctype) === 'message/rfc822')
                return $p->parts[0];
        }
        return null;
    }

    function getBody(){
        global $cfg;

        if ($cfg->isHtmlThreadEnabled()) {
            if ($body=$this->getPart($this->struct,'text/html')) {
                // Cleanup the html -- Balance html tags & neutralize unsafe tags.
                $body = (trim($body, " <>br/\t\n\r"))
                    ? Format::safe_html($body)
                    : '--';
            }
            elseif ($body=$this->getPart($this->struct,'text/plain')) {
                $body = trim($body)
                    ? sprintf('<pre>%s</pre>',
                        Format::htmlchars($body))
                    : '--';
            }
        }
        else {
            if (!($body=$this->getPart($this->struct,'text/plain'))) {
                if ($body=$this->getPart($this->struct,'text/html')) {
                    $body = Format::html2text(Format::safe_html($body), 100, false);
                }
            }
            $body = trim($body) ? $body : '--';
        }
        return $body;
    }

    function getPart($struct, $ctypepart, $recurse=-1) {

        if($struct && !$struct->parts) {
            $ctype = @strtolower($struct->ctype_primary.'/'.$struct->ctype_secondary);
            if($ctype && strcasecmp($ctype,$ctypepart)==0) {
                $content = $struct->body;
                //Encode to desired encoding - ONLY if charset is known??
                if (isset($struct->ctype_parameters['charset']))
                    $content = Format::encode($content,
                        $struct->ctype_parameters['charset'], $this->charset);

                return $content;
            }
        }

        $data='';
        if($struct && $struct->parts && $recurse) {
            foreach($struct->parts as $i=>$part) {
                if($part && !$part->disposition
                        && ($text=$this->getPart($part,$ctypepart,$recurse - 1)))
                    $data.=$text;
            }
        }
        return $data;
    }


    function mime_encode($text, $charset=null, $encoding='utf-8') {
        return Format::encode($text, $charset, $encoding);
    }

    function getAttachments($part=null){

        /* Consider this part as an attachment if
         *   * It has a Content-Disposition header
         *     * AND it is specified as either 'attachment' or 'inline'
         *   * The Content-Type header specifies
         *     * type is image/* or application/*
         *     * has a name parameter
         */
        if($part && (
                ($part->disposition
                    && (!strcasecmp($part->disposition,'attachment')
                        || !strcasecmp($part->disposition,'inline'))
                )
                || (!strcasecmp($part->ctype_primary,'image')
                    || !strcasecmp($part->ctype_primary,'application')))) {

            if (isset($part->d_parameters['filename']))
                $filename = $part->d_parameters['filename'];
            elseif (isset($part->d_parameters['filename*']))
                // Support RFC 6266, section 4.3 and RFC, and RFC 5987
                $filename = Format::decodeRfc5987(
                    $part->d_parameters['filename*']);

            // Support attachments that do not specify a content-disposition
            // but do specify a "name" parameter in the content-type header.
            elseif (isset($part->ctype_parameters['name']))
                $filename=$part->ctype_parameters['name'];
            elseif (isset($part->ctype_parameters['name*']))
                $filename = Format::decodeRfc5987(
                    $part->ctype_parameters['name*']);
            else
                // Not an attachment?
                return false;

            $file=array(
                    'name'  => $filename,
                    'type'  => strtolower($part->ctype_primary.'/'.$part->ctype_secondary),
                    );

            if ($part->ctype_parameters['charset'])
                $file['data'] = $this->mime_encode($part->body,
                    $part->ctype_parameters['charset']);
            else
                $file['data'] = $part->body;

            if(!$this->decode_bodies && $part->headers['content-transfer-encoding'])
                $file['encoding'] = $part->headers['content-transfer-encoding'];

            // Include Content-Id (for inline-images), stripping the <>
            $file['cid'] = (isset($part->headers['content-id']))
                ? rtrim(ltrim($part->headers['content-id'], '<'), '>') : false;

            return array($file);
        }

        if($part==null)
            $part=$this->getStruct();

        $files=array();
        if($part->parts){
            foreach($part->parts as $k=>$p){
                if($p && ($result=$this->getAttachments($p))) {
                    $files=array_merge($files,$result);
                }
            }
        }

        return $files;
    }

    function getPriority(){
        return Mail_Parse::parsePriority($this->getHeader());
    }

    function parsePriority($header=null){

        $priority=0;
        if($header && ($begin=strpos($header,'X-Priority:'))!==false){
            $begin+=strlen('X-Priority:');
            $xpriority=preg_replace("/[^0-9]/", "",substr($header, $begin, strpos($header,"\n",$begin) - $begin));
            if(!is_numeric($xpriority))
                $priority=0;
            elseif($xpriority>4)
                $priority=1;
            elseif($xpriority>=3)
                $priority=2;
            elseif($xpriority>0)
                $priority=3;
        }
        return $priority;
    }

    function parseAddressList($address){
        if (!$address)
            return false;
        return Mail_RFC822::parseAddressList($address, null, null,false);
    }

    function parse($rawemail) {
        $parser= new Mail_Parse($rawemail);
        return ($parser && $parser->decode())?$parser:null;
    }
}

class EmailDataParser {
    var $stream;
    var $error;

    function EmailDataParser($stream=null) {
        $this->stream = $stream;
    }

    function parse($stream) {
        global $cfg;

        $contents ='';
        if(is_resource($stream)) {
            while(!feof($stream))
                $contents .= fread($stream, 8192);

        } else {
            $contents = $stream;
        }

        $parser= new Mail_Parse($contents);
        if(!$parser->decode()) //Decode...returns false on decoding errors
            return $this->err('Email parse failed ['.$parser->getError().']');

        $data =array();
        //FROM address: who sent the email.
        if(($fromlist = $parser->getFromAddressList()) && !PEAR::isError($fromlist)) {
            $from=$fromlist[0]; //Default.
            foreach($fromlist as $fromobj) {
                if(!Validator::is_email($fromobj->mailbox.'@'.$fromobj->host)) continue;
                $from = $fromobj;
                break;
            }

            $data['email'] = $from->mailbox.'@'.$from->host;

            $data['name'] = trim($from->personal,'"');
            if($from->comment && $from->comment[0])
                $data['name'].= ' ('.$from->comment[0].')';

            //Use email address as name  when FROM address doesn't  have a name.
            if(!$data['name'] && $data['email'])
                $data['name'] = $data['email'];
        }

        //TO Address:Try to figure out the email address... associated with the incoming email.
        $emailId = 0;
        if(($tolist = $parser->getToAddressList())) {
            foreach ($tolist as $toaddr) {
                if(($emailId=Email::getIdByEmail($toaddr->mailbox.'@'.$toaddr->host)))
                    break;
            }
        }
        //maybe we got CC'ed??
        if(!$emailId && ($cclist=$parser->getCcAddressList())) {
            foreach ($cclist as $ccaddr) {
                if(($emailId=Email::getIdByEmail($ccaddr->mailbox.'@'.$ccaddr->host)))
                    break;
            }
        }

        if ($parser->isBounceNotice()) {
            // Fetch the original References and assign to 'references'
            if ($msg = $parser->getOriginalMessage())
                $data['references'] = $msg->headers['references'];
            // Fetch deliver status report
            $data['message'] = $parser->getDeliveryStatusMessage();
            $data['thread-type'] = 'N';
        }
        else {
            // Typical email
            $data['message'] = Format::stripEmptyLines($parser->getBody());
            $data['in-reply-to'] = $parser->struct->headers['in-reply-to'];
            $data['references'] = $parser->struct->headers['references'];
        }

        $data['subject'] = $parser->getSubject();
        $data['header'] = $parser->getHeader();
        $data['mid'] = $parser->getMessageId();
        $data['priorityId'] = $parser->getPriority();
        $data['emailId'] = $emailId;

        if (($replyto = $parser->getReplyTo()) && !PEAR::isError($replyto)) {
            $replyto = $replyto[0];
            $data['reply-to'] = $replyto->mailbox.'@'.$replyto->host;
            if ($replyto->personal)
                $data['reply-to-name'] = trim($replyto->personal, " \t\n\r\0\x0B\x22");
        }

        if($cfg && $cfg->allowEmailAttachments())
            $data['attachments'] = $parser->getAttachments();

        return $data;
    }

    function err($error) {
        $this->error = $error;

        return false;
    }

    function getError() {
        return $this->lastError();
    }

    function lastError() {
        return $this->error;
    }
}
?>