Add MIME type parameters ("type0", "type1") to the end of the plugin arguments for...

[squirrelmail.git] / functions / mime.php

<?php

/**
 * mime.php
 *
 * This contains the functions necessary to detect and decode MIME
 * messages.
 *
 * @copyright 1999-2020 The SquirrelMail Project Team
 * @license http://opensource.org/licenses/gpl-license.php GNU Public License
 * @version $Id$
 * @package squirrelmail
 */

/**
 * dependency information
   functions       dependency
   mime_structure
        class/mime/Message.class.php
            Message::parseStructure
        functions/page_header.php
            displayPageHeader
        functions/display_messages.php
            plain_error_message
   mime_fetch_body
        functions/imap_general.php
            sqimap_run_command
   mime_print_body_lines



functions/imap.php
functions/attachment_common.php
functions/display_messages.php

magicHtml => url_parser
translateText => url_parser

*/


/* -------------------------------------------------------------------------- */
/* MIME DECODING                                                              */
/* -------------------------------------------------------------------------- */

/**
 * Get the MIME structure
 *
 * This function gets the structure of a message and stores it in the "message" class.
 * It will return this object for use with all relevant header information and
 * fully parsed into the standard "message" object format.
 */
function mime_structure ($bodystructure, $flags=array()) {

    /* Isolate the body structure and remove beginning and end parenthesis. */
    $read = trim(substr ($bodystructure, strpos(strtolower($bodystructure), 'bodystructure') + 13));
    $read = trim(substr ($read, 0, -1));
    $i = 0;
    $msg = Message::parseStructure($read,$i);

    if (!is_object($msg)) {
        global $color, $mailbox;
        displayPageHeader( $color, $mailbox );
        $errormessage  = _("SquirrelMail could not decode the bodystructure of the message");
        $errormessage .= '<br />'._("The bodystructure provided by your IMAP server:").'<br /><br />';
        $errormessage .= '<pre>' . sm_encode_html_special_chars($read) . '</pre>';
        plain_error_message( $errormessage );
        echo '</body></html>';
        exit;
    }
    if (count($flags)) {
        foreach ($flags as $flag) {
//FIXME: please document why it is we have to check the first char of the flag but we then go ahead and do a full string comparison anyway.  Is this a speed enhancement?  If not, let's keep it simple and just compare the full string and forget the switch block.
            $char = strtoupper($flag{1});
            switch ($char) {
                case 'S':
                    if (strtolower($flag) == '\\seen') {
                        $msg->is_seen = true;
                    }
                    break;
                case 'A':
                    if (strtolower($flag) == '\\answered') {
                        $msg->is_answered = true;
                    }
                    break;
                case 'D':
                    if (strtolower($flag) == '\\deleted') {
                        $msg->is_deleted = true;
                    }
                    break;
                case 'F':
                    if (strtolower($flag) == '\\flagged') {
                        $msg->is_flagged = true;
                    }
                    else if (strtolower($flag) == '$forwarded') {
                        $msg->is_forwarded = true;
                    }
                    break;
                case 'M':
                    if (strtolower($flag) == '$mdnsent') {
                        $msg->is_mdnsent = true;
                    }
                    break;
                default:
                    break;
            }
        }
    }
    //    listEntities($msg);
    return $msg;
}



/* This starts the parsing of a particular structure.  It is called recursively,
 * so it can be passed different structures.  It returns an object of type
 * $message.
 * First, it checks to see if it is a multipart message.  If it is, then it
 * handles that as it sees is necessary.  If it is just a regular entity,
 * then it parses it and adds the necessary header information (by calling out
 * to mime_get_elements()
 */

function mime_fetch_body($imap_stream, $id, $ent_id=1, $fetch_size=0) {
    /* Do a bit of error correction.  If we couldn't find the entity id, just guess
     * that it is the first one.  That is usually the case anyway.
     */

    if (!$ent_id) {
        $cmd = "FETCH $id BODY[]";
    } else {
        $cmd = "FETCH $id BODY[$ent_id]";
    }

    if ($fetch_size!=0) $cmd .= "<0.$fetch_size>";

    $data = sqimap_run_command ($imap_stream, $cmd, true, $response, $message, TRUE);
    do {
        $topline = trim(array_shift($data));
    } while($topline && ($topline[0] == '*') && !preg_match('/\* [0-9]+ FETCH .*BODY.*/i', $topline)) ;
    // Matching with "BODY" above is difficult: in most cases "FETCH \(BODY" would work
    // but some servers may put other things in the same result, perhaps something such
    // as "* 23 FETCH (FLAGS (\Seen) BODY[1] {174}".  There is some small chance that
    // if the character sequence "BODY" appears in a response where it isn't actually
    // a FETCH response data item name, the current regex will break things.  The better
    // way to do this would be to parse the response correctly and not use a regex.

    $wholemessage = implode('', $data);
    if (preg_match('/\{([^\}]*)\}/', $topline, $regs)) {
        $ret = substr($wholemessage, 0, $regs[1]);
        /* There is some information in the content info header that could be important
         * in order to parse html messages. Let's get them here.
         */
//        if ($ret{0} == '<') {
//            $data = sqimap_run_command ($imap_stream, "FETCH $id BODY[$ent_id.MIME]", true, $response, $message, TRUE);
//        }
    } else if (preg_match('/"([^"]*)"/', $topline, $regs)) {
        $ret = $regs[1];
    } else if ((stristr($topline, 'nil') !== false) && (empty($wholemessage))) {
        $ret = $wholemessage;
    } else {
        global $where, $what, $mailbox, $passed_id, $startMessage;
        $par = 'mailbox=' . urlencode($mailbox) . '&amp;passed_id=' . $passed_id;
        if (isset($where) && isset($what)) {
            $par .= '&amp;where=' . urlencode($where) . '&amp;what=' . urlencode($what);
        } else {
            $par .= '&amp;startMessage=' . $startMessage . '&amp;show_more=0';
        }
        $par .= '&amp;response=' . urlencode($response) .
            '&amp;message='  . urlencode($message)  .
            '&amp;topline='  . urlencode($topline);

        echo '<tt><br />' .
            '<table width="80%"><tr>' .
            '<tr><td colspan="2">' .
            _("Body retrieval error. The reason for this is most probably that the message is malformed.") .
            '</td></tr>' .
            '<tr><td><b>' . _("Command:") . "</td><td>$cmd</td></tr>" .
            '<tr><td><b>' . _("Response:") . "</td><td>$response</td></tr>" .
            '<tr><td><b>' . _("Message:") . "</td><td>$message</td></tr>" .
            '<tr><td><b>' . _("FETCH line:") . "</td><td>$topline</td></tr>" .
            "</table><br /></tt></font><hr />";

        $data = sqimap_run_command ($imap_stream, "FETCH $passed_id BODY[]", true, $response, $message, TRUE);
        array_shift($data);
        $wholemessage = implode('', $data);

        $ret = $wholemessage;
    }
    return $ret;
}

function mime_print_body_lines ($imap_stream, $id, $ent_id=1, $encoding, $rStream='php://stdout', $force_crlf='') {

    /* Don't kill the connection if the browser is over a dialup
     * and it would take over 30 seconds to download it.
     * Don't call set_time_limit in safe mode.
     */

    if (!ini_get('safe_mode')) {
        set_time_limit(0);
    }
    /* in case of base64 encoded attachments, do not buffer them.
       Instead, echo the decoded attachment directly to screen */
    if (strtolower($encoding) == 'base64') {
        if (!$ent_id) {
            $query = "FETCH $id BODY[]";
        } else {
            $query = "FETCH $id BODY[$ent_id]";
        }
        sqimap_run_command($imap_stream,$query,true,$response,$message,TRUE,'sqimap_base64_decode',$rStream,true);
    } else {
        $body = mime_fetch_body ($imap_stream, $id, $ent_id);
        if (is_resource($rStream)) {
            fputs($rStream,decodeBody($body, $encoding, $force_crlf));
        } else {
            echo decodeBody($body, $encoding, $force_crlf);
        }
    }

    /*
       TODO, use the same method for quoted printable.
       However, I assume that quoted printable attachments aren't that large
       so the performancegain / memory usage drop will be minimal.
       If we decide to add that then we need to adapt sqimap_fread because
       we need to split te result on \n and fread doesn't stop at \n. That
       means we also should provide $results from sqimap_fread (by ref) to
       te function and set $no_return to false. The $filter function for
       quoted printable should handle unsetting of $results.
     */
    /*
       TODO 2: find out how we write to the output stream php://stdout. fwrite
       doesn't work because 'php://stdout isn't a stream.
     */

    return;
}

/* -[ END MIME DECODING ]----------------------------------------------------------- */

/* This is here for debugging purposes.  It will print out a list
 * of all the entity IDs that are in the $message object.
 */
function listEntities ($message) {
    if ($message) {
        echo "<tt>" . $message->entity_id . ' : ' . $message->type0 . '/' . $message->type1 . ' parent = '. $message->parent->entity_id. '<br />';
        for ($i = 0; isset($message->entities[$i]); $i++) {
            echo "$i : ";
            $msg = listEntities($message->entities[$i]);

            if ($msg) {
                echo "return: ";
                return $msg;
            }
        }
    }
}

function getPriorityStr($priority) {
    $priority_level = substr($priority,0,1);

    switch($priority_level) {
        /* Check for a higher then normal priority. */
        case '1':
        case '2':
            $priority_string = _("High");
            break;

        /* Check for a lower then normal priority. */
        case '4':
        case '5':
            $priority_string = _("Low");
            break;

        /* Check for a normal priority. */
        case '3':
        default:
            $priority_level = '3';
            $priority_string = _("Normal");
            break;

    }
    return $priority_string;
}

/* returns a $message object for a particular entity id */
function getEntity ($message, $ent_id) {
    return $message->getEntity($ent_id);
}

/* translateText
 * Extracted from strings.php 23/03/2002
 */

function translateText(&$body, $wrap_at, $charset) {
    global $where, $what;   /* from searching */
    global $color;          /* color theme */

    // require_once(SM_PATH . 'functions/url_parser.php');

    $body_ary = explode("\n", $body);
    for ($i=0; $i < count($body_ary); $i++) {
        $line = rtrim($body_ary[$i],"\r");

        if (strlen($line) - 2 >= $wrap_at) {
            sqWordWrap($line, $wrap_at, $charset);
        }
        $line = charset_decode($charset, $line);
        $line = str_replace("\t", '        ', $line);

        parseUrl ($line);

        $quotes = 0;
        $pos = 0;
        $j = strlen($line);

        while ($pos < $j) {
            if ($line[$pos] == ' ') {
                $pos++;
            } else if (strpos($line, '&gt;', $pos) === $pos) {
                $pos += 4;
                $quotes++;
            } else {
                break;
            }
        }

        if ($quotes % 2) {
            $line = '<span class="quote1">' . $line . '</span>';
        } elseif ($quotes) {
            $line = '<span class="quote2">' . $line . '</span>';
        }

        $body_ary[$i] = $line;
    }
    $body = '<pre>' . implode("\n", $body_ary) . '</pre>';
}

/**
 * This returns a parsed string called $body. That string can then
 * be displayed as the actual message in the HTML. It contains
 * everything needed, including HTML Tags, Attachments at the
 * bottom, etc.
 *
 * Since 1.2.0 function uses message_body hook.
 * Till 1.3.0 function included output of formatAttachments().
 *
 * @param resource $imap_stream imap connection resource
 * @param object $message squirrelmail message object
 * @param array $color squirrelmail color theme array
 * @param integer $wrap_at number of characters per line
 * @param string $ent_num (since 1.3.0) message part id
 * @param integer $id (since 1.3.0) message id
 * @param string $mailbox (since 1.3.0) imap folder name
 * @return string html formated message text
 */
function formatBody($imap_stream, $message, $color, $wrap_at, $ent_num, $id, $mailbox='INBOX') {
    /* This if statement checks for the entity to show as the
     * primary message. To add more of them, just put them in the
     * order that is their priority.
     */
    global $startMessage, $languages, $squirrelmail_language,
           $show_html_default, $sort, $has_unsafe_images, $passed_ent_id,
           $use_iframe, $iframe_height, $download_and_unsafe_link,
           $download_href, $unsafe_image_toggle_href, $unsafe_image_toggle_text,
           $oTemplate, $nbsp;

    // workaround for not updated config.php
    if (! isset($use_iframe)) $use_iframe = false;

    // If there's no "view_unsafe_images" variable in the URL, turn unsafe
    // images off by default.
    sqgetGlobalVar('view_unsafe_images', $view_unsafe_images, SQ_GET, FALSE);

    $body = '';
    $urlmailbox = urlencode($mailbox);
    $body_message = getEntity($message, $ent_num);
    if (($body_message->header->type0 == 'text') ||
            ($body_message->header->type0 == 'rfc822')) {
        $body = mime_fetch_body ($imap_stream, $id, $ent_num);
        $body = decodeBody($body, $body_message->header->encoding);

        if (isset($languages[$squirrelmail_language]['XTRA_CODE']) &&
                function_exists($languages[$squirrelmail_language]['XTRA_CODE'] . '_decode')) {
            if (mb_detect_encoding($body) != 'ASCII') {
                $body = call_user_func($languages[$squirrelmail_language]['XTRA_CODE'] . '_decode',$body);
            }
        }

        /* As of 1.5.2, $body is passed (and modified) by reference */
        do_hook('message_body', $body);

        /* If there are other types that shouldn't be formatted, add
         * them here.
         */

        if ($body_message->header->type1 == 'html') {
            if ($show_html_default <> 1) {
                $entity_conv = array('&nbsp;' => ' ',
                                     '<p>'    => "\n",
                                     '<P>'    => "\n",
                                     '<br>'   => "\n",
                                     '<BR>'   => "\n",
                                     '<br />' => "\n",
                                     '<BR />' => "\n",
                                     '&gt;'   => '>',
                                     '&lt;'   => '<');
                $body = strtr($body, $entity_conv);
                $body = strip_tags($body);
                $body = trim($body);
                translateText($body, $wrap_at,
                        $body_message->header->getParameter('charset'));
            } elseif ($use_iframe) {
                /**
                 * If we don't add html message between iframe tags,
                 * we must detect unsafe images and modify $has_unsafe_images.
                 */
                $html_body = magicHTML($body, $id, $message, $mailbox);
                // Convert character set in order to display html mails in different character set
                $html_body = charset_decode($body_message->header->getParameter('charset'),$html_body,false,true);

                // creating iframe url
                $iframeurl=sqm_baseuri().'src/view_html.php?'
                    . 'mailbox=' . $urlmailbox
                    . '&amp;passed_id=' . $id
                    . '&amp;ent_id=' . $ent_num
                    . '&amp;view_unsafe_images=' . (int) $view_unsafe_images;

                global $oTemplate;
                $oTemplate->assign('iframe_url', $iframeurl);
                $oTemplate->assign('iframe_height', $iframe_height);
                $oTemplate->assign('html_body', $html_body);

                $body = $oTemplate->fetch('read_html_iframe.tpl');
            } else {
                // old way of html rendering
                /**
                 * convert character set. charset_decode does not remove html special chars
                 * applied by magicHTML functions and does not sanitize them second time if
                 * fourth argument is true.
                 */
                $charset = $body_message->header->getParameter('charset');
                if (!empty($charset)) {
                    $body = charset_decode($charset,$body,false,true);
                }
                $body = magicHTML($body, $id, $message, $mailbox);
            }
        } else {
            translateText($body, $wrap_at,
                    $body_message->header->getParameter('charset'));
        }

        /*
         * Previously the links for downloading and unsafe images were printed
         * under the mail. By putting the links in a global variable we can
         * print it in the toolbar where it belongs. Since the original code was
         * in this place it's left here. It might be possible to move it to some
         * other place if that makes sense. The possibility to do so has not
         * been evaluated yet.
         */

        // Initialize the global variable to an empty string.
        // FIXME: To have $download_and_unsafe_link as a global variable might not be needed since the use of separate variables ($download_href, $unsafe_image_toggle_href, and $unsafe_image_toggle_text) for the templates was introduced.
        $download_and_unsafe_link = '';

        // Prepare and build a link for downloading the mail.
        $link = 'passed_id=' . $id . '&amp;ent_id='.$ent_num.
            '&amp;mailbox=' . $urlmailbox .'&amp;sort=' . $sort .
            '&amp;startMessage=' . $startMessage . '&amp;show_more=0';
        if (isset($passed_ent_id)) {
            $link .= '&amp;passed_ent_id='.$passed_ent_id;
        }
        $download_href = SM_PATH . 'src/download.php?absolute_dl=true&amp;' . $link;

        // Always add the link for downloading the mail as a file to the global
        // variable.
        $download_and_unsafe_link .= "$nbsp|$nbsp" 
            . create_hyperlink($download_href, _("Download this as a file"));

        // Find out the right text to use in the link depending on the
        // circumstances. If the unsafe images are displayed the link should
        // hide them, if they aren't displayed the link should only appear if
        // the mail really contains unsafe images.
        if ($view_unsafe_images) {
            $text = _("Hide Unsafe Images");
        } else {
            if (isset($has_unsafe_images) && $has_unsafe_images) {
                $link .= '&amp;view_unsafe_images=1';
                $text = _("View Unsafe Images");
            } else {
                $text = '';
            }
        }

        // Only create a link for unsafe images if there's need for one. If so:
        // add it to the global variable.
        if($text != '') {
            $unsafe_image_toggle_href = SM_PATH . 'src/read_body.php?'.$link;
            $unsafe_image_toggle_text = $text;
            $download_and_unsafe_link .= "$nbsp|$nbsp"
                . create_hyperlink($unsafe_image_toggle_href, $text);
        }
    }
    return $body;
}

/**
 * Generate attachments array for passing to templates.
 *
 * @since 1.5.2
 * @param object $message SquirrelMail message object
 * @param array $exclude_id message parts that are not attachments.
 * @param string $mailbox mailbox name
 * @param integer $id message id
 */
function buildAttachmentArray($message, $exclude_id, $mailbox, $id) {
    global $where, $what, $startMessage, $color, $passed_ent_id,
           $base_uri, $block_svg_download;

    $att_ar = $message->getAttachments($exclude_id);
    $urlMailbox = urlencode($mailbox);

    $attachments = array();
    foreach ($att_ar as $att) {
        $ent = $att->entity_id;
        $header = $att->header;
        $type0 = strtolower($header->type0);
        $type1 = strtolower($header->type1);
        if ($block_svg_download && strpos($type1, 'svg') === 0)
            continue;

        $name = '';
        $links = array();
        $links['download link']['text'] = _("Download");
        $links['download link']['href'] = $base_uri .
            "src/download.php?absolute_dl=true&amp;passed_id=$id&amp;mailbox=$urlMailbox&amp;ent_id=$ent";

        if ($type0 =='message' && $type1 == 'rfc822') {
            $default_page = $base_uri  . 'src/read_body.php';
            $rfc822_header = $att->rfc822_header;
            $filename = $rfc822_header->subject;
            if (trim( $filename ) == '') {
                $filename = 'untitled-[' . $ent . ']' ;
            }
            $from_o = $rfc822_header->from;
            if (is_object($from_o)) {
                $from_name = decodeHeader($from_o->getAddress(false));
            } elseif (is_array($from_o) && count($from_o) && is_object($from_o[0])) {
                // something weird happens when a digest message is opened and you return to the digest
                // now the from object is part of an array. Probably the parseHeader call overwrites the info
                // retrieved from the bodystructure in a different way. We need to fix this later.
                // possible starting point, do not fetch header we already have and inspect how
                // the rfc822_header object behaves.
                $from_name = decodeHeader($from_o[0]->getAddress(false));
            } else {
                $from_name = _("Unknown sender");
            }
            $description = _("From").': '.$from_name;
        } else {
            $default_page = $base_uri  . 'src/download.php';
            $filename = $att->getFilename();
            if ($header->description) {
                $description = decodeHeader($header->description);
            } else {
                $description = '';
            }
        }

        $display_filename = $filename;
        if (isset($passed_ent_id)) {
            $passed_ent_id_link = '&amp;passed_ent_id='.$passed_ent_id;
        } else {
            $passed_ent_id_link = '';
        }
        $defaultlink = $default_page . "?startMessage=$startMessage"
            . "&amp;passed_id=$id&amp;mailbox=$urlMailbox"
            . '&amp;ent_id='.$ent.$passed_ent_id_link;
        if ($where && $what) {
            $defaultlink .= '&amp;where='. urlencode($where).'&amp;what='.urlencode($what);
        }
        // IE does make use of mime content sniffing. Forcing a download
        // prohibit execution of XSS inside an application/octet-stream attachment
        if ($type0 == 'application' && $type1 == 'octet-stream') {
            $defaultlink .= '&amp;absolute_dl=true';
        }

        /* This executes the attachment hook with a specific MIME-type.
         * It also allows plugins to run if there's a rule for a more
         * generic type. Finally, a hook for ALL attachment types is
         * run as well.
         */
        // First remember the default link.
        $defaultlink_orig = $defaultlink;

        /* The API for this hook has changed as of 1.5.2 so that all plugin
           arguments are passed in an array instead of each their own plugin
           argument, and arguments are passed by reference, so instead of
           returning any changes, changes should simply be made to the original
           arguments themselves. */
        $temp = array(&$links, &$startMessage, &$id, &$urlMailbox, &$ent,
                    &$defaultlink, &$display_filename, &$where, &$what,
                    $type0, $temp1);
        do_hook("attachment $type0/$type1", $temp);
        /* The API for this hook has changed as of 1.5.2 so that all plugin
           arguments are passed in an array instead of each their own plugin
           argument, and arguments are passed by reference, so instead of
           returning any changes, changes should simply be made to the original
           arguments themselves. */
        $temp = array(&$links, &$startMessage, &$id, &$urlMailbox, &$ent,
                      &$defaultlink, &$display_filename, &$where, &$what);
        // Do not let a generic plugin change the default link if a more
        // specialized one already did it...
        if ($defaultlink != $defaultlink_orig) {
            $dummy = '';
            $temp[5] = &$dummy;
        }
        do_hook("attachment $type0/*", $temp);
        /* The API for this hook has changed as of 1.5.2 so that all plugin
           arguments are passed in an array instead of each their own plugin
           argument, and arguments are passed by reference, so instead of
           returning any changes, changes should simply be made to the original
           arguments themselves. */
        $temp = array(&$links, &$startMessage, &$id, &$urlMailbox, &$ent, 
                      &$defaultlink, &$display_filename, &$where, &$what);
        // Do not let a generic plugin change the default link if a more
        // specialized one already did it...
        if ($defaultlink != $defaultlink_orig) {
            $dummy = '';
            $temp[5] = &$dummy;
        }
        do_hook("attachment */*", $temp);

        $this_attachment = array();
        $this_attachment['Name'] = decodeHeader($display_filename);
        $this_attachment['Description'] = $description;
        $this_attachment['DefaultHREF'] = $defaultlink;
        $this_attachment['DownloadHREF'] = $links['download link']['href'];
        $this_attachment['ViewHREF'] = isset($links['attachment_common']) ? $links['attachment_common']['href'] : '';

        // base64 encoded file sizes are misleading, so approximate real size
        if (!empty($header->encoding) && strtolower($header->encoding) == 'base64')
            $this_attachment['Size'] = $header->size / 4 * 3;
        else
            $this_attachment['Size'] = $header->size;

        $this_attachment['ContentType'] = sm_encode_html_special_chars($type0 .'/'. $type1);
        $this_attachment['OtherLinks'] = array();
        foreach ($links as $val) {
            if ($val['text']==_("Download") || $val['text'] == _("View"))
                continue;
            if (empty($val['text']) && empty($val['extra']))
                continue;

            $temp = array();
            $temp['HREF'] = $val['href'];
            $temp['Text'] = (empty($val['text']) ? '' : $val['text']) . (empty($val['extra']) ? '' : $val['extra']);
            $this_attachment['OtherLinks'][] = $temp;
        }
        $attachments[] = $this_attachment;

        unset($links);
    }

    return $attachments;
}

/**
 * Displays attachment links and information
 *
 * Since 1.3.0 function is not included in formatBody() call.
 *
 * Since 1.0.2 uses attachment $type0/$type1 hook.
 * Since 1.2.5 uses attachment $type0/* hook.
 * Since 1.5.0 uses attachments_bottom hook.
 * Since 1.5.2 uses templates and does *not* return a value.
 *
 * @param object $message SquirrelMail message object
 * @param array $exclude_id message parts that are not attachments.
 * @param string $mailbox mailbox name
 * @param integer $id message id
 */
function formatAttachments($message, $exclude_id, $mailbox, $id) {
    global $oTemplate;

    $attach = buildAttachmentArray($message, $exclude_id, $mailbox, $id);

    $oTemplate->assign('attachments', $attach);
    $oTemplate->display('read_attachments.tpl');
}

function sqimap_base64_decode(&$string) {

    // Base64 encoded data goes in pairs of 4 bytes. To achieve on the
    // fly decoding (to reduce memory usage) you have to check if the
    // data has incomplete pairs

    // Remove the noise in order to check if the 4 bytes pairs are complete
    $string = str_replace(array("\r\n","\n", "\r", " "),array('','','',''),$string);

    $sStringRem = '';
    $iMod = strlen($string) % 4;
    if ($iMod) {
        $sStringRem = substr($string,-$iMod);
        // Check if $sStringRem contains padding characters
        if (substr($sStringRem,-1) != '=') {
            $string = substr($string,0,-$iMod);
        } else {
            $sStringRem = '';
        }
    }
    $string = base64_decode($string);
    return $sStringRem;
}

/**
 * Decodes encoded string (usually message body)
 *
 * This function decodes a string (usually the message body)
 * depending on the encoding type.  Currently quoted-printable
 * and base64 encodings are supported.
 *
 * The decode_body hook was added to this function in 1.4.2/1.5.0.
 * The $force_crlf parameter was added in 1.5.2.
 *
 * @param string $string     The encoded string
 * @param string $encoding   used encoding
 * @param string $force_crlf Whether or not to force CRLF or LF
 *                           line endings (or to leave as is).
 *                           If given as "LF", line endings will
 *                           all be converted to LF; if "CRLF",
 *                           line endings will all be converted
 *                           to CRLF.  If given as an empty value,
 *                           the global $force_crlf_default will
 *                           be consulted (it can be specified in
 *                           config/config_local.php).  Otherwise,
 *                           any other value will cause the string
 *                           to be left alone.  Note that this will
 *                           be overridden to "LF" if not using at
 *                           least PHP version 4.3.0. (OPTIONAL;
 *                           default is empty - consult global
 *                           default value)
 *
 * @return string The decoded string
 *
 * @since 1.0
 *
 */
function decodeBody($string, $encoding, $force_crlf='') {

    global $force_crlf_default;
    if (empty($force_crlf)) $force_crlf = $force_crlf_default;
    $force_crlf = strtoupper($force_crlf);

    // must force line endings to LF due to broken
    // quoted_printable_decode() in PHP versions
    // before 4.3.0 (see below)
    //
    if (!check_php_version(4, 3, 0) || $force_crlf == 'LF')
        $string = str_replace("\r\n", "\n", $string);
    else if ($force_crlf == 'CRLF')
        $string = str_replace("\n", "\r\n", $string);

    $encoding = strtolower($encoding);

    $encoding_handler = do_hook('decode_body', $encoding);


    // plugins get first shot at decoding the string
    //
    if (!empty($encoding_handler) && function_exists($encoding_handler)) {
        $string = $encoding_handler('decode', $string);

    } elseif ($encoding == 'quoted-printable' ||
            $encoding == 'quoted_printable') {

        // quoted_printable_decode() function is broken in older
        // php versions.  Text with \r\n decoding was fixed only
        // in php 4.3.0.  Minimal code requirement is PHP 4.0.4+
        // and the above call to:  str_replace("\r\n", "\n", $string);
        //
        $string = quoted_printable_decode($string);

    } elseif ($encoding == 'base64') {
        $string = base64_decode($string);
    }

    // All other encodings are returned raw.
    return $string;
}

/**
 * Decodes headers
 *
 * This function decodes strings that are encoded according to
 * RFC1522 (MIME Part Two: Message Header Extensions for Non-ASCII Text).
 * Patched by Christian Schmidt <christian@ostenfeld.dk>  23/03/2002
 *
 * @param string $string header string that has to be made readable
 * @param boolean $utfencode change message in order to be readable on user's charset. defaults to true
 * @param boolean $htmlsafe preserve spaces and sanitize html special characters. defaults to true
 * @param boolean $decide decide if string can be utfencoded. defaults to false
 * @return string decoded header string
 */
function decodeHeader ($string, $utfencode=true,$htmlsafe=true,$decide=false) {
    global $languages, $squirrelmail_language,$default_charset, $fix_broken_base64_encoded_messages;
    if (is_array($string)) {
        $string = implode("\n", $string);
    }

    if (isset($languages[$squirrelmail_language]['XTRA_CODE']) &&
            function_exists($languages[$squirrelmail_language]['XTRA_CODE'] . '_decodeheader')) {
        $string = call_user_func($languages[$squirrelmail_language]['XTRA_CODE'] . '_decodeheader', $string);
        // Do we need to return at this point?
        // return $string;
    }
    $i = 0;
    $iLastMatch = -2;
    $encoded = true;

// FIXME: spaces are allowed inside quoted-printable encoding, but the following line will bust up any such encoded strings
    $aString = explode(' ',$string);
    $ret = '';
    foreach ($aString as $chunk) {
        if ($encoded && $chunk === '') {
            continue;
        } elseif ($chunk === '') {
            $ret .= ' ';
            continue;
        }
        $encoded = false;
        /* if encoded words are not separated by a linear-space-white we still catch them */
        $j = $i-1;

        while ($match = preg_match('/^(.*)=\?([^?]*)\?(Q|B)\?([^?]*)\?=(.*)$/Ui',$chunk,$res)) {
            /* if the last chunk isn't an encoded string then put back the space, otherwise don't */
            if ($iLastMatch !== $j) {
                if ($htmlsafe) {
                    $ret .= '&#32;';
                } else {
                    $ret .= ' ';
                }
            }
            $iLastMatch = $i;
            $j = $i;
            if ($htmlsafe) {
                $ret .= sm_encode_html_special_chars($res[1]);
            } else {
                $ret .= $res[1];
            }
            $encoding = ucfirst($res[3]);

            /* decide about valid decoding */
            if ($decide && is_conversion_safe($res[2])) {
                $utfencode=true;
                $can_be_encoded=true;
            } else {
                $can_be_encoded=false;
            }
            switch ($encoding)
            {
                case 'B':
                    // fix broken base64-encoded strings (remove end = padding,
                    // change any = to + in middle of string, add padding back
                    // to the end)
                    if ($fix_broken_base64_encoded_messages) {
                        $encoded_string_minus_padding = strtr(rtrim($res[4], '='), '=', '+');
                        $res[4] = str_pad($encoded_string_minus_padding, strlen($res[4]), '=');
                    }
                    $replace = base64_decode($res[4]);
                    if ($utfencode) {
                        if ($can_be_encoded) {
                            /* convert string to different charset,
                             * if functions asks for it (usually in compose)
                             */
                            $ret .= charset_convert($res[2],$replace,$default_charset,$htmlsafe);
                        } else {
                            // convert string to html codes in order to display it
                            $ret .= charset_decode($res[2],$replace);
                        }
                    } else {
                        if ($htmlsafe) {
                            $replace = sm_encode_html_special_chars($replace);
                        }
                        $ret.= $replace;
                    }
                    break;
                case 'Q':
                    $replace = str_replace('_', ' ', $res[4]);
                    $replace = preg_replace_callback('/=([0-9a-f]{2})/i',
                            create_function ('$matches', 'return chr(hexdec($matches[1]));'),
                            $replace);
                    if ($utfencode) {
                        if ($can_be_encoded) {
                            /* convert string to different charset,
                             * if functions asks for it (usually in compose)
                             */
                            $replace = charset_convert($res[2], $replace,$default_charset,$htmlsafe);
                        } else {
                            // convert string to html codes in order to display it
                            $replace = charset_decode($res[2], $replace);
                        }
                    } else {
                        if ($htmlsafe) {
                            $replace = sm_encode_html_special_chars($replace);
                        }
                    }
                    $ret .= $replace;
                    break;
                default:
                    break;
            }
            $chunk = $res[5];
            $encoded = true;
        }
        if (!$encoded) {
            if ($htmlsafe) {
                $ret .= '&#32;';
            } else {
                $ret .= ' ';
            }
        }

        if (!$encoded && $htmlsafe) {
            $ret .= sm_encode_html_special_chars($chunk);
        } else {
            $ret .= $chunk;
        }
        ++$i;
    }
    /* remove the first added space */
    if ($ret) {
        if ($htmlsafe) {
            $ret = substr($ret,5);
        } else {
            $ret = substr($ret,1);
        }
    }

    return $ret;
}

/**
 * Encodes header
 *
 * Function uses XTRA_CODE _encodeheader function, if such function exists.
 *
 * Function uses Q encoding by default and encodes a string according to RFC
 * 1522 for use in headers if it contains 8-bit characters or anything that
 * looks like it should be encoded.
 *
 * Function switches to B encoding and encodeHeaderBase64() function, if
 * string is 8bit and multibyte character set supported by mbstring extension
 * is used. It can cause E_USER_NOTICE errors, if interface is used with
 * multibyte character set unsupported by mbstring extension.
 *
 * @param string $string header string, that has to be encoded
 * @return string quoted-printable encoded string
 * @todo make $mb_charsets system wide constant
 */
function encodeHeader ($string) {
    global $default_charset, $languages, $squirrelmail_language;

    if (isset($languages[$squirrelmail_language]['XTRA_CODE']) &&
            function_exists($languages[$squirrelmail_language]['XTRA_CODE'] . '_encodeheader')) {
        return  call_user_func($languages[$squirrelmail_language]['XTRA_CODE'] . '_encodeheader', $string);
    }

    // Use B encoding for multibyte charsets
    $mb_charsets = array('utf-8','big5','gb2313','euc-kr');
    if (in_array($default_charset,$mb_charsets) &&
        in_array($default_charset,sq_mb_list_encodings()) &&
        sq_is8bit($string)) {
        return encodeHeaderBase64($string,$default_charset);
    } elseif (in_array($default_charset,$mb_charsets) &&
              sq_is8bit($string) &&
              ! in_array($default_charset,sq_mb_list_encodings())) {
        // Add E_USER_NOTICE error here (can cause 'Cannot add header information' warning in compose.php)
        // trigger_error('encodeHeader: Multibyte character set unsupported by mbstring extension.',E_USER_NOTICE);
    }

    // Encode only if the string contains 8-bit characters or =?
    $j = strlen($string);
    $max_l = 75 - strlen($default_charset) - 7;
    $aRet = array();
    $ret = '';
    $iEncStart = $enc_init = false;
    $cur_l = $iOffset = 0;
    for($i = 0; $i < $j; ++$i) {
        switch($string{$i})
        {
            case '"':
            case '=':
            case '<':
            case '>':
            case ',':
            case '?':
            case '_':
                if ($iEncStart === false) {
                    $iEncStart = $i;
                }
                $cur_l+=3;
                if ($cur_l > ($max_l-2)) {
                    /* if there is an stringpart that doesn't need encoding, add it */
                    $aRet[] = substr($string,$iOffset,$iEncStart-$iOffset);
                    $aRet[] = "=?$default_charset?Q?$ret?=";
                    $iOffset = $i;
                    $cur_l = 0;
                    $ret = '';
                    $iEncStart = false;
                } else {
                    $ret .= sprintf("=%02X",ord($string{$i}));
                }
                break;
            case '(':
            case ')':
                if ($iEncStart !== false) {
                    $aRet[] = substr($string,$iOffset,$iEncStart-$iOffset);
                    $aRet[] = "=?$default_charset?Q?$ret?=";
                    $iOffset = $i;
                    $cur_l = 0;
                    $ret = '';
                    $iEncStart = false;
                }
                break;
            case ' ':
                if ($iEncStart !== false) {
                    $cur_l++;
                    if ($cur_l > $max_l) {
                        $aRet[] = substr($string,$iOffset,$iEncStart-$iOffset);
                        $aRet[] = "=?$default_charset?Q?$ret?=";
                        $iOffset = $i;
                        $cur_l = 0;
                        $ret = '';
                        $iEncStart = false;
                    } else {
                        $ret .= '_';
                    }
                }
                break;
            default:
                $k = ord($string{$i});
                if ($k > 126) {
                    if ($iEncStart === false) {
                        // do not start encoding in the middle of a string, also take the rest of the word.
                        $sLeadString = substr($string,0,$i);
                        $aLeadString = explode(' ',$sLeadString);
                        $sToBeEncoded = array_pop($aLeadString);
                        $iEncStart = $i - strlen($sToBeEncoded);
                        $ret .= $sToBeEncoded;
                        $cur_l += strlen($sToBeEncoded);
                    }
                    $cur_l += 3;
                    /* first we add the encoded string that reached it's max size */
                    if ($cur_l > ($max_l-2)) {
                        $aRet[] = substr($string,$iOffset,$iEncStart-$iOffset);
                        $aRet[] = "=?$default_charset?Q?$ret?= "; /* the next part is also encoded => separate by space */
                        $cur_l = 3;
                        $ret = '';
                        $iOffset = $i;
                        $iEncStart = $i;
                    }
                    $enc_init = true;
                    $ret .= sprintf("=%02X", $k);
                } else {
                    if ($iEncStart !== false) {
                        $cur_l++;
                        if ($cur_l > $max_l) {
                            $aRet[] = substr($string,$iOffset,$iEncStart-$iOffset);
                            $aRet[] = "=?$default_charset?Q?$ret?=";
                            $iEncStart = false;
                            $iOffset = $i;
                            $cur_l = 0;
                            $ret = '';
                        } else {
                            $ret .= $string{$i};
                        }
                    }
                }
                break;
        }
    }

    if ($enc_init) {
        if ($iEncStart !== false) {
            $aRet[] = substr($string,$iOffset,$iEncStart-$iOffset);
            $aRet[] = "=?$default_charset?Q?$ret?=";
        } else {
            $aRet[] = substr($string,$iOffset);
        }
        $string = implode('',$aRet);
    }
    return $string;
}

/**
 * Encodes string according to rfc2047 B encoding header formating rules
 *
 * It is recommended way to encode headers with character sets that store
 * symbols in more than one byte.
 *
 * Function requires mbstring support. If required mbstring functions are missing,
 * function returns false and sets E_USER_WARNING level error message.
 *
 * Minimal requirements - php 4.0.6 with mbstring extension. Please note,
 * that mbstring functions will generate E_WARNING errors, if unsupported
 * character set is used. mb_encode_mimeheader function provided by php
 * mbstring extension is not used in order to get better control of header
 * encoding.
 *
 * Used php code functions - function_exists(), trigger_error(), strlen()
 * (is used with charset names and base64 strings). Used php mbstring
 * functions - mb_strlen and mb_substr.
 *
 * Related documents: rfc 2045 (BASE64 encoding), rfc 2047 (mime header
 * encoding), rfc 2822 (header folding)
 *
 * @param string $string header string that must be encoded
 * @param string $charset character set. Must be supported by mbstring extension.
 * Use sq_mb_list_encodings() to detect supported charsets.
 * @return string string encoded according to rfc2047 B encoding formating rules
 * @since 1.5.1
 * @todo First header line can be wrapped to $iMaxLength - $HeaderFieldLength - 1
 * @todo Do we want to control max length of header?
 * @todo Do we want to control EOL (end-of-line) marker?
 * @todo Do we want to translate error message?
 */
function encodeHeaderBase64($string,$charset) {
    /**
     * Check mbstring function requirements.
     */
    if (! function_exists('mb_strlen') ||
        ! function_exists('mb_substr')) {
        // set E_USER_WARNING
        trigger_error('encodeHeaderBase64: Required mbstring functions are missing.',E_USER_WARNING);
        // return false
        return false;
    }

    // initial return array
    $aRet = array();

    /**
     * header length = 75 symbols max (same as in encodeHeader)
     * remove $charset length
     * remove =? ? ?= (5 chars)
     * remove 2 more chars (\r\n ?)
     */
    $iMaxLength = 75 - strlen($charset) - 7;

    // set first character position
    $iStartCharNum = 0;

    // loop through all characters. count characters and not bytes.
    for ($iCharNum=1; $iCharNum<=mb_strlen($string,$charset); $iCharNum++) {
        // encode string from starting character to current character.
        $encoded_string = base64_encode(mb_substr($string,$iStartCharNum,$iCharNum-$iStartCharNum,$charset));

        // Check encoded string length
        if(strlen($encoded_string)>$iMaxLength) {
            // if string exceeds max length, reduce number of encoded characters and add encoded string part to array
            $aRet[] = base64_encode(mb_substr($string,$iStartCharNum,$iCharNum-$iStartCharNum-1,$charset));

            // set new starting character
            $iStartCharNum = $iCharNum-1;

            // encode last char (in case it is last character in string)
            $encoded_string = base64_encode(mb_substr($string,$iStartCharNum,$iCharNum-$iStartCharNum,$charset));
        } // if string is shorter than max length - add next character
    }

    // add last encoded string to array
    $aRet[] = $encoded_string;

    // set initial return string
    $sRet = '';

    // loop through encoded strings
    foreach($aRet as $string) {
        // TODO: Do we want to control EOL (end-of-line) marker
        if ($sRet!='') $sRet.= " ";

        // add header tags and encoded string to return string
        $sRet.= '=?'.$charset.'?B?'.$string.'?=';
    }

    return $sRet;
}

/* This function trys to locate the entity_id of a specific mime element */
function find_ent_id($id, $message) {
    for ($i = 0, $ret = ''; $ret == '' && $i < count($message->entities); $i++) {
        if ($message->entities[$i]->header->type0 == 'multipart')  {
            $ret = find_ent_id($id, $message->entities[$i]);
        } else {
            if (strcasecmp($message->entities[$i]->header->id, $id) == 0) {
//                if (sq_check_save_extension($message->entities[$i])) {
                return $message->entities[$i]->entity_id;
//                }
            } elseif (!empty($message->entities[$i]->header->parameters['name'])) {
                /**
                 * This is part of a fix for Outlook Express 6.x generating
                 * cid URLs without creating content-id headers
                 * @@JA - 20050207
                 */
                if (strcasecmp($message->entities[$i]->header->parameters['name'], $id) == 0) {
                    return $message->entities[$i]->entity_id;
                }
            }
        }
    }
    return $ret;
}

function sq_check_save_extension($message) {
    $filename = $message->getFilename();
    $ext = substr($filename, strrpos($filename,'.')+1);
    $save_extensions = array('jpg','jpeg','gif','png','bmp');
    return in_array($ext, $save_extensions);
}


/**
 ** HTMLFILTER ROUTINES
 */

/**
 * This function checks attribute values for entity-encoded values
 * and returns them translated into 8-bit strings so we can run
 * checks on them.
 *
 * @param  $attvalue A string to run entity check against.
 * @return           Nothing, modifies a reference value.
 */
function sq_defang(&$attvalue){
    $me = 'sq_defang';
    /**
     * Skip this if there aren't ampersands or backslashes.
     */
    if (strpos($attvalue, '&') === false
        && strpos($attvalue, '\\') === false){
        return;
    }
    $m = false;
    // before deent, translate the dangerous unicode characters and ... to safe values
    // otherwise the regular expressions do not match.



    do {
        $m = false;
        $m = $m || sq_deent($attvalue, '/\&#0*(\d+);*/s');
        $m = $m || sq_deent($attvalue, '/\&#x0*((\d|[a-f])+);*/si', true);
        $m = $m || sq_deent($attvalue, '/\\\\(\d+)/s', true);
    } while ($m == true);
    $attvalue = stripslashes($attvalue);
}

/**
 * Kill any tabs, newlines, or carriage returns. Our friends the
 * makers of the browser with 95% market value decided that it'd
 * be funny to make "java[tab]script" be just as good as "javascript".
 *
 * @param  attvalue  The attribute value before extraneous spaces removed.
 * @return attvalue  Nothing, modifies a reference value.
 */
function sq_unspace(&$attvalue){
    $me = 'sq_unspace';
    if (strcspn($attvalue, "\t\r\n\0 ") != strlen($attvalue)){
        $attvalue = str_replace(Array("\t", "\r", "\n", "\0", " "),
                                Array('',   '',   '',   '',   ''), $attvalue);
    }
}

/**
 * Translate all dangerous Unicode or Shift_JIS characters which are accepted by
 * IE as regular characters.
 *
 * @param  attvalue  The attribute value before dangerous characters are translated.
 * @return attvalue  Nothing, modifies a reference value.
 * @author Marc Groot Koerkamp.
 */
function sq_fixIE_idiocy(&$attvalue) {
    // remove NUL
    $attvalue = str_replace("\0", "", $attvalue);
    // remove comments
    $attvalue = preg_replace("/(\/\*.*?\*\/)/","",$attvalue);

    // IE has the evil habit of accepting every possible value for the attribute expression.
    // The table below contains characters which are parsed by IE if they are used in the "expression"
    // attribute value.
    $aDangerousCharsReplacementTable = array(
                        array('&#x029F;', '&#0671;' ,/* L UNICODE IPA Extension */
                              '&#x0280;', '&#0640;' ,/* R UNICODE IPA Extension */
                              '&#x0274;', '&#0628;' ,/* N UNICODE IPA Extension */
                              '&#xFF25;', '&#65317;' ,/* Unicode FULLWIDTH LATIN CAPITAL LETTER E */
                              '&#xFF45;', '&#65349;' ,/* Unicode FULLWIDTH LATIN SMALL LETTER E */
                              '&#xFF38;', '&#65336;',/* Unicode FULLWIDTH LATIN CAPITAL LETTER X */
                              '&#xFF58;', '&#65368;',/* Unicode FULLWIDTH LATIN SMALL LETTER X */
                              '&#xFF30;', '&#65328;',/* Unicode FULLWIDTH LATIN CAPITAL LETTER P */
                              '&#xFF50;', '&#65360;',/* Unicode FULLWIDTH LATIN SMALL LETTER P */
                              '&#xFF32;', '&#65330;',/* Unicode FULLWIDTH LATIN CAPITAL LETTER R */
                              '&#xFF52;', '&#65362;',/* Unicode FULLWIDTH LATIN SMALL LETTER R */
                              '&#xFF33;', '&#65331;',/* Unicode FULLWIDTH LATIN CAPITAL LETTER S */
                              '&#xFF53;', '&#65363;',/* Unicode FULLWIDTH LATIN SMALL LETTER S */
                              '&#xFF29;', '&#65321;',/* Unicode FULLWIDTH LATIN CAPITAL LETTER I */
                              '&#xFF49;', '&#65353;',/* Unicode FULLWIDTH LATIN SMALL LETTER I */
                              '&#xFF2F;', '&#65327;',/* Unicode FULLWIDTH LATIN CAPITAL LETTER O */
                              '&#xFF4F;', '&#65359;',/* Unicode FULLWIDTH LATIN SMALL LETTER O */
                              '&#xFF2E;', '&#65326;',/* Unicode FULLWIDTH LATIN CAPITAL LETTER N */
                              '&#xFF4E;', '&#65358;',/* Unicode FULLWIDTH LATIN SMALL LETTER N */
                              '&#xFF2C;', '&#65324;',/* Unicode FULLWIDTH LATIN CAPITAL LETTER L */
                              '&#xFF4C;', '&#65356;',/* Unicode FULLWIDTH LATIN SMALL LETTER L */
                              '&#xFF35;', '&#65333;',/* Unicode FULLWIDTH LATIN CAPITAL LETTER U */
                              '&#xFF55;', '&#65365;',/* Unicode FULLWIDTH LATIN SMALL LETTER U */
                              '&#x207F;', '&#8319;' ,/* Unicode SUPERSCRIPT LATIN SMALL LETTER N */
                              "\xEF\xBC\xA5", /* Shift JIS FULLWIDTH LATIN CAPITAL LETTER E */   // in unicode this is some Chinese char range
                              "\xEF\xBD\x85", /* Shift JIS FULLWIDTH LATIN SMALL LETTER E */
                              "\xEF\xBC\xB8", /* Shift JIS FULLWIDTH LATIN CAPITAL LETTER X */
                              "\xEF\xBD\x98", /* Shift JIS FULLWIDTH LATIN SMALL LETTER X */
                              "\xEF\xBC\xB0", /* Shift JIS FULLWIDTH LATIN CAPITAL LETTER P */
                              "\xEF\xBD\x90", /* Shift JIS FULLWIDTH LATIN SMALL LETTER P */
                              "\xEF\xBC\xB2", /* Shift JIS FULLWIDTH LATIN CAPITAL LETTER R */
                              "\xEF\xBD\x92", /* Shift JIS FULLWIDTH LATIN SMALL LETTER R */
                              "\xEF\xBC\xB3", /* Shift JIS FULLWIDTH LATIN CAPITAL LETTER S */
                              "\xEF\xBD\x93", /* Shift JIS FULLWIDTH LATIN SMALL LETTER S */
                              "\xEF\xBC\xA9", /* Shift JIS FULLWIDTH LATIN CAPITAL LETTER I */
                              "\xEF\xBD\x89", /* Shift JIS FULLWIDTH LATIN SMALL LETTER I */
                              "\xEF\xBC\xAF", /* Shift JIS FULLWIDTH LATIN CAPITAL LETTER O */
                              "\xEF\xBD\x8F", /* Shift JIS FULLWIDTH LATIN SMALL LETTER O */
                              "\xEF\xBC\xAE", /* Shift JIS FULLWIDTH LATIN CAPITAL LETTER N */
                              "\xEF\xBD\x8E", /* Shift JIS FULLWIDTH LATIN SMALL LETTER N */
                              "\xEF\xBC\xAC", /* Shift JIS FULLWIDTH LATIN CAPITAL LETTER L */
                              "\xEF\xBD\x8C", /* Shift JIS FULLWIDTH LATIN SMALL LETTER L */
                              "\xEF\xBC\xB5", /* Shift JIS FULLWIDTH LATIN CAPITAL LETTER U */
                              "\xEF\xBD\x95", /* Shift JIS FULLWIDTH LATIN SMALL LETTER U */
                              "\xE2\x81\xBF", /* Shift JIS FULLWIDTH SUPERSCRIPT N */
                              "\xCA\x9F", /* L UNICODE IPA Extension */
                              "\xCA\x80", /* R UNICODE IPA Extension */
                              "\xC9\xB4"),  /* N UNICODE IPA Extension */
                       array('l', 'l', 'r','r','n','n',
                             'E','E','e','e','X','X','x','x','P','P','p','p','R','R','r','r','S','S','s','s','I','I',
                             'i','i','O','O','o','o','N','N','n','n','L','L','l','l','U','U','u','u','n','n',
                             'E','e','X','x','P','p','R','r','S','s','I','i','O','o','N','n','L','l','U','u','n','l','r','n'));
    $attvalue = str_replace($aDangerousCharsReplacementTable[0],$aDangerousCharsReplacementTable[1],$attvalue);

    // Escapes are useful for special characters like "{}[]()'&. In other cases they are
    // used for XSS.
    $attvalue = preg_replace("/(\\\\)([a-zA-Z]{1})/",'$2',$attvalue);
}

/**
 * This function returns the final tag out of the tag name, an array
 * of attributes, and the type of the tag. This function is called by
 * sq_sanitize internally.
 *
 * @param  $tagname  the name of the tag.
 * @param  $attary   the array of attributes and their values
 * @param  $tagtype  The type of the tag (see in comments).
 * @return           a string with the final tag representation.
 */
function sq_tagprint($tagname, $attary, $tagtype){
    $me = 'sq_tagprint';

    if ($tagtype == 2){
        $fulltag = '</' . $tagname . '>';
    } else {
        $fulltag = '<' . $tagname;
        if (is_array($attary) && sizeof($attary)){
            $atts = Array();
            while (list($attname, $attvalue) = each($attary)){
                array_push($atts, "$attname=$attvalue");
            }
            $fulltag .= ' ' . join(" ", $atts);
        }
        if ($tagtype == 3){
            $fulltag .= ' /';
        }
        $fulltag .= '>';
    }
    return $fulltag;
}

/**
 * A small helper function to use with array_walk. Modifies a by-ref
 * value and makes it lowercase.
 *
 * @param  $val a value passed by-ref.
 * @return      void since it modifies a by-ref value.
 */
function sq_casenormalize(&$val){
    $val = strtolower($val);
}

/**
 * This function skips any whitespace from the current position within
 * a string and to the next non-whitespace value.
 *
 * @param  $body   the string
 * @param  $offset the offset within the string where we should start
 *                 looking for the next non-whitespace character.
 * @return         the location within the $body where the next
 *                 non-whitespace char is located.
 */
function sq_skipspace($body, $offset){
    $me = 'sq_skipspace';
    preg_match('/^(\s*)/s', substr($body, $offset), $matches);
    if (!empty($matches[1])){
        $offset += strlen($matches[1]);
    }
    return $offset;
}

/**
 * This function looks for the next character within a string.  It's
 * really just a glorified "strpos", except it catches if failures
 * nicely.
 *
 * @param  $body   The string to look for needle in.
 * @param  $offset Start looking from this position.
 * @param  $needle The character/string to look for.
 * @return         location of the next occurance of the needle, or
 *                 strlen($body) if needle wasn't found.
 */
function sq_findnxstr($body, $offset, $needle){
    $me  = 'sq_findnxstr';
    $pos = strpos($body, $needle, $offset);
    if ($pos === FALSE){
        $pos = strlen($body);
    }
    return $pos;
}

/**
 * This function takes a PCRE-style regexp and tries to match it
 * within the string.
 *
 * @param  $body   The string to look for needle in.
 * @param  $offset Start looking from here.
 * @param  $reg    A PCRE-style regex to match.
 * @return         Returns a false if no matches found, or an array
 *                 with the following members:
 *                 - integer with the location of the match within $body
 *                 - string with whatever content between offset and the match
 *                 - string with whatever it is we matched
 */
function sq_findnxreg($body, $offset, $reg){
    $me = 'sq_findnxreg';
    $matches = Array();
    $retarr = Array();
    preg_match("%^(.*?)($reg)%si", substr($body, $offset), $matches);
    if (!isset($matches{0}) || !$matches{0}){
        $retarr = false;
    } else {
        $retarr{0} = $offset + strlen($matches{1});
        $retarr{1} = $matches{1};
        $retarr{2} = $matches{2};
    }
    return $retarr;
}

/**
 * This function looks for the next tag.
 *
 * @param  $body   String where to look for the next tag.
 * @param  $offset Start looking from here.
 * @return         false if no more tags exist in the body, or
 *                 an array with the following members:
 *                 - string with the name of the tag
 *                 - array with attributes and their values
 *                 - integer with tag type (1, 2, or 3)
 *                 - integer where the tag starts (starting "<")
 *                 - integer where the tag ends (ending ">")
 *                 first three members will be false, if the tag is invalid.
 */
function sq_getnxtag($body, $offset){
    $me = 'sq_getnxtag';
    if ($offset > strlen($body)){
        return false;
    }
    $lt = sq_findnxstr($body, $offset, "<");
    if ($lt == strlen($body)){
        return false;
    }
    /**
     * We are here:
     * blah blah <tag attribute="value">
     * \---------^
     */
    $pos = sq_skipspace($body, $lt+1);
    if ($pos >= strlen($body)){
        return Array(false, false, false, $lt, strlen($body));
    }
    /**
     * There are 3 kinds of tags:
     * 1. Opening tag, e.g.:
     *    <a href="blah">
     * 2. Closing tag, e.g.:
     *    </a>
     * 3. XHTML-style content-less tag, e.g.:
     *    <img src="blah" />
     */
    $tagtype = false;
    switch (substr($body, $pos, 1)){
        case '/':
            $tagtype = 2;
            $pos++;
            break;
        case '!':
            /**
             * A comment or an SGML declaration.
             */
            if (substr($body, $pos+1, 2) == "--"){
                $gt = strpos($body, "-->", $pos);
                if ($gt === false){
                    $gt = strlen($body);
                } else {
                    $gt += 2;
                }
                return Array(false, false, false, $lt, $gt);
            } else {
                $gt = sq_findnxstr($body, $pos, ">");
                return Array(false, false, false, $lt, $gt);
            }
            break;
        default:
            /**
             * Assume tagtype 1 for now. If it's type 3, we'll switch values
             * later.
             */
            $tagtype = 1;
            break;
    }

    $tag_start = $pos;
    $tagname = '';
    /**
     * Look for next [\W-_], which will indicate the end of the tag name.
     */
    $regary = sq_findnxreg($body, $pos, "[^\w\-_]");
    if ($regary == false){
        return Array(false, false, false, $lt, strlen($body));
    }
    list($pos, $tagname, $match) = $regary;
    $tagname = strtolower($tagname);

    /**
     * $match can be either of these:
     * '>'  indicating the end of the tag entirely.
     * '\s' indicating the end of the tag name.
     * '/'  indicating that this is type-3 xhtml tag.
     *
     * Whatever else we find there indicates an invalid tag.
     */
    switch ($match){
        case '/':
            /**
             * This is an xhtml-style tag with a closing / at the
             * end, like so: <img src="blah" />. Check if it's followed
             * by the closing bracket. If not, then this tag is invalid
             */
            if (substr($body, $pos, 2) == "/>"){
                $pos++;
                $tagtype = 3;
            } else {
                $gt = sq_findnxstr($body, $pos, ">");
                $retary = Array(false, false, false, $lt, $gt);
                return $retary;
            }
        case '>':
            return Array($tagname, false, $tagtype, $lt, $pos);
            break;
        default:
            /**
             * Check if it's whitespace
             */
            if (!preg_match('/\s/', $match)){
                /**
                 * This is an invalid tag! Look for the next closing ">".
                 */
                $gt = sq_findnxstr($body, $lt, ">");
                return Array(false, false, false, $lt, $gt);
            }
            break;
    }

    /**
     * At this point we're here:
     * <tagname  attribute='blah'>
     * \-------^
     *
     * At this point we loop in order to find all attributes.
     */
    $attname = '';
    $atttype = false;
    $attary = Array();

    while ($pos <= strlen($body)){
        $pos = sq_skipspace($body, $pos);
        if ($pos == strlen($body)){
            /**
             * Non-closed tag.
             */
            return Array(false, false, false, $lt, $pos);
        }
        /**
         * See if we arrived at a ">" or "/>", which means that we reached
         * the end of the tag.
         */
        $matches = Array();
        if (preg_match("%^(\s*)(>|/>)%s", substr($body, $pos), $matches)) {
            /**
             * Yep. So we did.
             */
            $pos += strlen($matches{1});
            if ($matches{2} == "/>"){
                $tagtype = 3;
                $pos++;
            }
            return Array($tagname, $attary, $tagtype, $lt, $pos);
        }

        /**
         * There are several types of attributes, with optional
         * [:space:] between members.
         * Type 1:
         *   attrname[:space:]=[:space:]'CDATA'
         * Type 2:
         *   attrname[:space:]=[:space:]"CDATA"
         * Type 3:
         *   attr[:space:]=[:space:]CDATA
         * Type 4:
         *   attrname
         *
         * We leave types 1 and 2 the same, type 3 we check for
         * '"' and convert to "&quot" if needed, then wrap in
         * double quotes. Type 4 we convert into:
         * attrname="yes".
         */
        $regary = sq_findnxreg($body, $pos, "[^:\w\-_]");
        if ($regary == false){
            /**
             * Looks like body ended before the end of tag.
             */
            return Array(false, false, false, $lt, strlen($body));
        }
        list($pos, $attname, $match) = $regary;
        $attname = strtolower($attname);
        /**
         * We arrived at the end of attribute name. Several things possible
         * here:
         * '>'  means the end of the tag and this is attribute type 4
         * '/'  if followed by '>' means the same thing as above
         * '\s' means a lot of things -- look what it's followed by.
         *      anything else means the attribute is invalid.
         */
        switch($match){
            case '/':
                /**
                 * This is an xhtml-style tag with a closing / at the
                 * end, like so: <img src="blah" />. Check if it's followed
                 * by the closing bracket. If not, then this tag is invalid
                 */
                if (substr($body, $pos, 2) == "/>"){
                    $pos++;
                    $tagtype = 3;
                } else {
                    $gt = sq_findnxstr($body, $pos, ">");
                    $retary = Array(false, false, false, $lt, $gt);
                    return $retary;
                }
            case '>':
                $attary{$attname} = '"yes"';
                return Array($tagname, $attary, $tagtype, $lt, $pos);
                break;
            default:
                /**
                 * Skip whitespace and see what we arrive at.
                 */
                $pos = sq_skipspace($body, $pos);
                $char = substr($body, $pos, 1);
                /**
                 * Two things are valid here:
                 * '=' means this is attribute type 1 2 or 3.
                 * \w means this was attribute type 4.
                 * anything else we ignore and re-loop. End of tag and
                 * invalid stuff will be caught by our checks at the beginning
                 * of the loop.
                 */
                if ($char == "="){
                    $pos++;
                    $pos = sq_skipspace($body, $pos);
                    /**
                     * Here are 3 possibilities:
                     * "'"  attribute type 1
                     * '"'  attribute type 2
                     * everything else is the content of tag type 3
                     */
                    $quot = substr($body, $pos, 1);
                    if ($quot == "'"){
                        $regary = sq_findnxreg($body, $pos+1, "\'");
                        if ($regary == false){
                            return Array(false, false, false, $lt, strlen($body));
                        }
                        list($pos, $attval, $match) = $regary;
                        $pos++;
                        $attary{$attname} = "'" . $attval . "'";
                    } else if ($quot == '"'){
                        $regary = sq_findnxreg($body, $pos+1, '\"');
                        if ($regary == false){
                            return Array(false, false, false, $lt, strlen($body));
                        }
                        list($pos, $attval, $match) = $regary;
                        $pos++;
                        $attary{$attname} = '"' . $attval . '"';
                    } else {
                        /**
                         * These are hateful. Look for \s, or >.
                         */
                        $regary = sq_findnxreg($body, $pos, "[\s>]");
                        if ($regary == false){
                            return Array(false, false, false, $lt, strlen($body));
                        }
                        list($pos, $attval, $match) = $regary;
                        /**
                         * If it's ">" it will be caught at the top.
                         */
                        $attval = preg_replace("/\"/s", "&quot;", $attval);
                        $attary{$attname} = '"' . $attval . '"';
                    }
                } else if (preg_match("|[\w/>]|", $char)) {
                    /**
                     * That was attribute type 4.
                     */
                    $attary{$attname} = '"yes"';
                } else {
                    /**
                     * An illegal character. Find next '>' and return.
                     */
                    $gt = sq_findnxstr($body, $pos, ">");
                    return Array(false, false, false, $lt, $gt);
                }
                break;
        }
    }
    /**
     * The fact that we got here indicates that the tag end was never
     * found. Return invalid tag indication so it gets stripped.
     */
    return Array(false, false, false, $lt, strlen($body));
}

/**
 * Translates entities into literal values so they can be checked.
 *
 * @param $attvalue the by-ref value to check.
 * @param $regex    the regular expression to check against.
 * @param $hex      whether the entites are hexadecimal.
 * @return          True or False depending on whether there were matches.
 */
function sq_deent(&$attvalue, $regex, $hex=false){
    $me = 'sq_deent';
    $ret_match = false;
    // remove comments
    //$attvalue = preg_replace("/(\/\*.*\*\/)/","",$attvalue);
    preg_match_all($regex, $attvalue, $matches);
    if (is_array($matches) && sizeof($matches[0]) > 0){
        $repl = Array();
        for ($i = 0; $i < sizeof($matches[0]); $i++){
            $numval = $matches[1][$i];
            if ($hex){
                $numval = hexdec($numval);
            }
            $repl{$matches[0][$i]} = chr($numval);
        }
        $attvalue = strtr($attvalue, $repl);
        return true;
    } else {
        return false;
    }
}

/**
 * This function runs various checks against the attributes.
 *
 * @param  $tagname         String with the name of the tag.
 * @param  $attary          Array with all tag attributes.
 * @param  $rm_attnames     See description for sq_sanitize
 * @param  $bad_attvals     See description for sq_sanitize
 * @param  $add_attr_to_tag See description for sq_sanitize
 * @param  $message         message object
 * @param  $id              message id
 * @return                  Array with modified attributes.
 */
function sq_fixatts($tagname,
                    $attary,
                    $rm_attnames,
                    $bad_attvals,
                    $add_attr_to_tag,
                    $message,
                    $id,
                    $mailbox
                    ){
    $me = 'sq_fixatts';
    while (list($attname, $attvalue) = each($attary)){
        /**
         * See if this attribute should be removed.
         */
        foreach ($rm_attnames as $matchtag=>$matchattrs){
            if (preg_match($matchtag, $tagname)){
                foreach ($matchattrs as $matchattr){
                    if (preg_match($matchattr, $attname)){
                        unset($attary{$attname});
                        continue;
                    }
                }
            }
        }
        /**
         * Workaround for IE quirks
         */
        sq_fixIE_idiocy($attvalue);

        /**
         * Remove any backslashes, entities, and extraneous whitespace.
         */

        $oldattvalue = $attvalue;
        sq_defang($attvalue);
        if ($attname == 'style' && $attvalue !== $oldattvalue) {
            // entities are used in the attribute value. In 99% of the cases it's there as XSS
            // i.e.<div style="{ left:exp&#x0280;essio&#x0274;( alert('XSS') ) }">
            $attvalue = "idiocy";
            $attary{$attname} = $attvalue;
        }
        sq_unspace($attvalue);

        /**
         * Now let's run checks on the attvalues.
         * I don't expect anyone to comprehend this. If you do,
         * get in touch with me so I can drive to where you live and
         * shake your hand personally. :)
         */
        foreach ($bad_attvals as $matchtag=>$matchattrs){
            if (preg_match($matchtag, $tagname)){
                foreach ($matchattrs as $matchattr=>$valary){
                    if (preg_match($matchattr, $attname)){
                        /**
                         * There are two arrays in valary.
                         * First is matches.
                         * Second one is replacements
                         */
                        list($valmatch, $valrepl) = $valary;
                        $newvalue =
                            preg_replace($valmatch, $valrepl, $attvalue);
                        if ($newvalue != $attvalue){
                            $attary{$attname} = $newvalue;
                            $attvalue = $newvalue;
                        }
                    }
                }
            }
        }
        if ($attname == 'style') {
            if (preg_match('/[\0-\37\200-\377]+/',$attvalue)) {
                // 8bit and control characters in style attribute values can be used for XSS, remove them
                $attary{$attname} = '"disallowed character"';
            }
            preg_match_all("/url\s*\((.+)\)/si",$attvalue,$aMatch);
            if (count($aMatch)) {
                foreach($aMatch[1] as $sMatch) {
                    // url value
                    $urlvalue = $sMatch;
                    sq_fix_url($attname, $urlvalue, $message, $id, $mailbox,"'");
                    $attary{$attname} = str_replace($sMatch,$urlvalue,$attvalue);
                }
            }
        }
        /**
         * Use white list based filtering on attributes which can contain url's
         */
        else if ($attname == 'href' || $attname == 'xlink:href' || $attname == 'src'
              || $attname == 'poster' || $attname == 'formaction'
              || $attname == 'background' || $attname == 'action') {
            sq_fix_url($attname, $attvalue, $message, $id, $mailbox);
            $attary{$attname} = $attvalue;
        }
    }
    /**
     * See if we need to append any attributes to this tag.
     */
    foreach ($add_attr_to_tag as $matchtag=>$addattary){
        if (preg_match($matchtag, $tagname)){
            $attary = array_merge($attary, $addattary);
        }
    }
    return $attary;
}

/**
 * This function filters url's
 *
 * @param  $attvalue        String with attribute value to filter
 * @param  $message         message object
 * @param  $id               message id
 * @param  $mailbox         mailbox
 * @param  $sQuote          quoting characters around url's
 */
function sq_fix_url($attname, &$attvalue, $message, $id, $mailbox,$sQuote = '"') {
    $attvalue = trim($attvalue);
    if ($attvalue && ($attvalue[0] =='"'|| $attvalue[0] == "'")) {
        // remove the double quotes
        $sQuote = $attvalue[0];
        $attvalue = trim(substr($attvalue,1,-1));
    }

    // If there's no "view_unsafe_images" variable in the URL, turn unsafe
    // images off by default.
    sqgetGlobalVar('view_unsafe_images', $view_unsafe_images, SQ_GET, FALSE);

    global $use_transparent_security_image;
    if ($use_transparent_security_image) $secremoveimg = '../images/spacer.png';
    else $secremoveimg = '../images/' . _("sec_remove_eng.png");

    /**
     * Replace empty src tags with the blank image.  src is only used
     * for frames, images, and image inputs.  Doing a replace should
     * not affect them working as should be, however it will stop
     * IE from being kicked off when src for img tags are not set
     */
    if ($attvalue == '') {
        $attvalue = '"' . SM_PATH . 'images/blank.png"';
    } else {
        // first, disallow 8 bit characters and control characters
        if (preg_match('/[\0-\37\200-\377]+/',$attvalue)) {
            switch ($attname) {
                case 'href':
                    $attvalue = $sQuote . 'http://invalid-stuff-detected.example.com' . $sQuote;
                    break;
                default:
                    $attvalue = $sQuote . SM_PATH . 'images/blank.png'. $sQuote;
                    break;
            }
        } else {
            $aUrl = parse_url($attvalue);
            if (isset($aUrl['scheme'])) {
                switch(strtolower($aUrl['scheme'])) {
                    case 'mailto':
                    case 'http':
                    case 'https':
                    case 'ftp':
                        if ($attname != 'href') {
                            if ($view_unsafe_images == false) {
                                $attvalue = $sQuote . $secremoveimg . $sQuote;
                            } else {
                                if (isset($aUrl['path'])) {

                                    // No one has been able to show that image URIs
                                    // can be exploited, so for now, no restrictions
                                    // are made at all.  If this proves to be a problem,
                                    // the commented-out code below can be of help.
                                    // (One consideration is that I see nothing in this
                                    // function that specifically says that we will
                                    // only ever arrive here when inspecting an image
                                    // tag, although that does seem to be the end
                                    // result - e.g., <script src="..."> where malicious
                                    // image URIs are in fact a problem are already
                                    // filtered out elsewhere.
                                    /* ---------------------------------
                                    // validate image extension.
                                    $ext = strtolower(substr($aUrl['path'],strrpos($aUrl['path'],'.')));
                                    if (!in_array($ext,array('.jpeg','.jpg','xjpeg','.gif','.bmp','.jpe','.png','.xbm'))) {
                                        // If URI is to something other than
                                        // a regular image file, get the contents
                                        // and try to see if it is an image.
                                        // Don't use Fileinfo (finfo_file()) because
                                        // we'd need to make the admin configure the
                                        // location of the magic.mime file (FIXME: add finfo_file() support later?)
                                        //
                                        $mime_type = '';
                                        if (function_exists('mime_content_type')
                                         && ($FILE = @fopen($attvalue, 'rb', FALSE))) {

                                            // fetch file
                                            //
                                            $file_contents = '';
                                            while (!feof($FILE)) {
                                                $file_contents .= fread($FILE, 8192);
                                            }
                                            fclose($FILE);

                                            // store file locally
                                            //
                                            global $attachment_dir, $username;
                                            $hashed_attachment_dir = getHashedDir($username, $attachment_dir);
                                            $localfilename = GenerateRandomString(32, '', 7);
                                            $full_localfilename = "$hashed_attachment_dir/$localfilename";
                                            while (file_exists($full_localfilename)) {
                                                $localfilename = GenerateRandomString(32, '', 7);
                                                $full_localfilename = "$hashed_attachment_dir/$localfilename";
                                            }
                                            $FILE = fopen("$hashed_attachment_dir/$localfilename", 'wb');
                                            fwrite($FILE, $file_contents);
                                            fclose($FILE);

                                            // get mime type and remove file
                                            //
                                            $mime_type = mime_content_type("$hashed_attachment_dir/$localfilename");
                                            unlink("$hashed_attachment_dir/$localfilename");
                                        }
                                        // debug: echo "$attvalue FILE TYPE IS $mime_type<HR>";
                                        if (substr(strtolower($mime_type), 0, 5) != 'image') {
                                            $attvalue = $sQuote . SM_PATH . 'images/blank.png'. $sQuote;
                                        }
                                    }
                                    --------------------------------- */
                                } else {
                                    $attvalue = $sQuote . SM_PATH . 'images/blank.png'. $sQuote;
                                }
                            }
                        } else {
                            $attvalue = $sQuote . $attvalue . $sQuote;
                        }
                        break;
                    case 'outbind':
                        /**
                         * "Hack" fix for Outlook using propriatary outbind:// protocol in img tags.
                         * One day MS might actually make it match something useful, for now, falling
                         * back to using cid2http, so we can grab the blank.png.
                         */
                        $attvalue = $sQuote . sq_cid2http($message, $id, $attvalue, $mailbox) . $sQuote;
                        break;
                    case 'cid':
                        /**
                            * Turn cid: urls into http-friendly ones.
                            */
                        $attvalue = $sQuote . sq_cid2http($message, $id, $attvalue, $mailbox) . $sQuote;
                        break;
                    default:
                        $attvalue = $sQuote . SM_PATH . 'images/blank.png' . $sQuote;
                        break;
                }
            } else {
                if (!isset($aUrl['path']) || $aUrl['path'] != $secremoveimg) {
                    // parse_url did not lead to satisfying result
                    $attvalue = $sQuote . SM_PATH . 'images/blank.png' . $sQuote;
                }
            }
        }
    }
}

/**
 * This function edits the style definition to make them friendly and
 * usable in SquirrelMail.
 *
 * @param  $message  the message object
 * @param  $id       the message id
 * @param  $content  a string with whatever is between <style> and </style>
 * @param  $mailbox  the message mailbox
 * @return           a string with edited content.
 */
function sq_fixstyle($body, $pos, $message, $id, $mailbox){
    $me = 'sq_fixstyle';
    // workaround for </style> in between comments
    $iCurrentPos = $pos;
    $content = '';
    $sToken = '';
    $bSucces = false;
    $bEndTag = false;
    for ($i=$pos,$iCount=strlen($body);$i<$iCount;++$i) {
        $char = $body{$i};
        switch ($char) {
            case '<':
                $sToken = $char;
                break;
            case '/':
                 if ($sToken == '<') {
                    $sToken .= $char;
                    $bEndTag = true;
                 } else {
                    $content .= $char;
                 }
                 break;
            case '>':
                 if ($bEndTag) {
                    $sToken .= $char;
                    if (preg_match('/\<\/\s*style\s*\>/i',$sToken,$aMatch)) {
                        $newpos = $i + 1;
                        $bSucces = true;
                        break 2;
                    } else {
                        $content .= $sToken;
                    }
                    $bEndTag = false;
                 } else {
                    $content .= $char;
                 }
                 break;
            case '!':
                if ($sToken == '<') {
                    // possible comment
                    if (isset($body{$i+2}) && substr($body,$i,3) == '!--') {
                        $i = strpos($body,'-->',$i+3);
                        if ($i === false) { // no end comment
                            $i = strlen($body);
                        }
                        $sToken = '';
                    }
                } else {
                    $content .= $char;
                }
                break;
            default:
                if ($bEndTag) {
                    $sToken .= $char;
                } else {
                    $content .= $char;
                }
                break;
        }
    }
    if ($bSucces == FALSE){
        return array(FALSE, strlen($body));
    }



    /**
     * First look for general BODY style declaration, which would be
     * like so:
     * body {background: blah-blah}
     * and change it to .bodyclass so we can just assign it to a <div>
     */
    // $content = preg_replace("|body(\s*\{.*?\})|si", ".bodyclass\\1", $content);
    // Nah, this is even better - try to preface all CSS selectors with
    // our <div> class ID "bodyclass" then correct generic "body" selectors
    // TODO: this works pretty good but breaks stuff like this:
    //       @media print { body { font-size: 10pt; } }
    //       but there isn't an easy way to make this regex skip @media
    //       definitions... though lots of the ones in the wild will be
    //       correctly handled because they tend to end with a parenthesis, like:
    //       @media screen and (max-width:480px) { ...
    $content = preg_replace('/([a-z0-9._-][a-z0-9 >+~|:._-]*\s*(?:,|{.*?}))/si', '.bodyclass $1', $content);
    $content = str_replace('.bodyclass body', '.bodyclass', $content);

    global $use_transparent_security_image;
    if ($use_transparent_security_image) $secremoveimg = '../images/spacer.png';
    else $secremoveimg = '../images/' . _("sec_remove_eng.png");

    /**
    * Fix url('blah') declarations.
    */
    //   $content = preg_replace("|url\s*\(\s*([\'\"])\s*\S+script\s*:.*?([\'\"])\s*\)|si",
    //                           "url(\\1$secremoveimg\\2)", $content);

    // first check for 8bit sequences and disallowed control characters
    if (preg_match('/[\16-\37\200-\377]+/',$content)) {
        $content = '<!-- style block removed by html filter due to presence of 8bit characters -->';
        return array($content, $newpos);
    }

    // IE Sucks hard. We have a special function for it.
    sq_fixIE_idiocy($content);

    // remove @import line
    $content = preg_replace("/^\s*(@import.*)$/mi","\n<!-- @import rules forbidden -->\n",$content);

    // translate ur\l and variations (IE parses that)
    // TODO check if the sq_fixIE_idiocy function already handles this.
    $content = preg_replace("/(\\\\)?u(\\\\)?r(\\\\)?l(\\\\)?/i", 'url', $content);
    preg_match_all("/url\s*\((.+)\)/si",$content,$aMatch);
    if (count($aMatch)) {
        $aValue = $aReplace = array();
        foreach($aMatch[1] as $sMatch) {
            // url value
            $urlvalue = $sMatch;
            sq_fix_url('style',$urlvalue, $message, $id, $mailbox,"'");
            $aValue[] = $sMatch;
            $aReplace[] = $urlvalue;
        }
        $content = str_replace($aValue,$aReplace,$content);
    }

    /**
     * Remove any backslashes, entities, and extraneous whitespace.
     */
    $contentTemp = $content;
    sq_defang($contentTemp);
    sq_unspace($contentTemp);

    /**
     * Fix stupid css declarations which lead to vulnerabilities
     * in IE.
     *
     * Also remove "position" attribute, as it can easily be set
     * to "fixed" or "absolute" with "left" and "top" attributes
     * of zero, taking over the whole content frame.  It can also
     * be set to relative and move itself anywhere it wants to,
     * displaying content in areas it shouldn't be allowed to touch.
     */
    $match   = Array('/\/\*.*\*\//', // removes /* blah blah */
                    '/expression/i',
                    '/behaviou*r/i',
                    '/binding/i',
                    '/include-source/i',
                    '/javascript/i',
                    '/script/i',
                    '/position/i');
    $replace = Array('','idiocy', 'idiocy', 'idiocy', 'idiocy', 'idiocy', 'idiocy', '');
    $contentNew = preg_replace($match, $replace, $contentTemp);
    if ($contentNew !== $contentTemp) {
        // insecure css declarations are used. From now on we don't care
        // anymore if the css is destroyed by sq_deent, sq_unspace or sq_unbackslash
        $content = $contentNew;
    }
    return array($content, $newpos);
}


/**
 * This function converts cid: url's into the ones that can be viewed in
 * the browser.
 *
 * @param  $message  the message object
 * @param  $id       the message id
 * @param  $cidurl   the cid: url.
 * @param  $mailbox  the message mailbox
 * @return           a string with a http-friendly url
 */
function sq_cid2http($message, $id, $cidurl, $mailbox){
    /**
     * Get rid of quotes.
     */
    $quotchar = substr($cidurl, 0, 1);
    if ($quotchar == '"' || $quotchar == "'"){
        $cidurl = str_replace($quotchar, "", $cidurl);
    } else {
        $quotchar = '';
    }
    $cidurl = substr(trim($cidurl), 4);

    $match_str = '/\{.*?\}\//';
    $str_rep = '';
    $cidurl = preg_replace($match_str, $str_rep, $cidurl);

    $linkurl = find_ent_id($cidurl, $message);
    /* in case of non-safe cid links $httpurl should be replaced by a sort of
       unsafe link image */
    $httpurl = '';

    /**
     * This is part of a fix for Outlook Express 6.x generating
     * cid URLs without creating content-id headers. These images are
     * not part of the multipart/related html mail. The html contains
     * <img src="cid:{some_id}/image_filename.ext"> references to
     * attached images with as goal to render them inline although
     * the attachment disposition property is not inline.
     */

    if (empty($linkurl)) {
        if (preg_match('/{.*}\//', $cidurl)) {
            $cidurl = preg_replace('/{.*}\//','', $cidurl);
            if (!empty($cidurl)) {
                $linkurl = find_ent_id($cidurl, $message);
            }
        }
    }

    if (!empty($linkurl)) {
        $httpurl = $quotchar . sqm_baseuri() . 'src/download.php?absolute_dl=true&amp;' .
            "passed_id=$id&amp;mailbox=" . urlencode($mailbox) .
            '&amp;ent_id=' . $linkurl . $quotchar;
    } else {
        /**
         * If we couldn't generate a proper img url, drop in a blank image
         * instead of sending back empty, otherwise it causes unusual behaviour
         */
        $httpurl = $quotchar . SM_PATH . 'images/blank.png' . $quotchar;
    }

    return $httpurl;
}

/**
 * This function changes the <body> tag into a <div> tag since we
 * can't really have a body-within-body.
 *
 * @param  $attary   an array of attributes and values of <body>
 * @param  $mailbox  mailbox we're currently reading (for cid2http)
 * @param  $message  current message (for cid2http)
 * @param  $id       current message id (for cid2http)
 * @return           a modified array of attributes to be set for <div>
 */
function sq_body2div($attary, $mailbox, $message, $id){
    $me = 'sq_body2div';
    $divattary = Array('class' => "'bodyclass'");
    $text = '#000000';
    $has_bgc_stl = $has_txt_stl = false;
    $styledef = '';
    if (is_array($attary) && sizeof($attary) > 0){
        foreach ($attary as $attname=>$attvalue){
            $quotchar = substr($attvalue, 0, 1);
            $attvalue = str_replace($quotchar, "", $attvalue);
            switch ($attname){
                case 'background':
                    $attvalue = sq_cid2http($message, $id, $attvalue, $mailbox);
                    $styledef .= "background-image: url('$attvalue'); ";
                    break;
                case 'bgcolor':
                    $has_bgc_stl = true;
                    $styledef .= "background-color: $attvalue; ";
                    break;
                case 'text':
                    $has_txt_stl = true;
                    $styledef .= "color: $attvalue; ";
                    break;
            }
        }
        // Outlook defines a white bgcolor and no text color. This can lead to
        // white text on a white bg with certain themes.
        if ($has_bgc_stl && !$has_txt_stl) {
            $styledef .= "color: $text; ";
        }
        if (strlen($styledef) > 0){
            $divattary{"style"} = "\"$styledef\"";
        }
    }
    return $divattary;
}

/**
 * This is the main function and the one you should actually be calling.
 * There are several variables you should be aware of an which need
 * special description.
 *
 * Since the description is quite lengthy, see it here:
 * http://linux.duke.edu/projects/mini/htmlfilter/
 *
 * @param $body                 the string with HTML you wish to filter
 * @param $tag_list             see description above
 * @param $rm_tags_with_content see description above
 * @param $self_closing_tags    see description above
 * @param $force_tag_closing    see description above
 * @param $rm_attnames          see description above
 * @param $bad_attvals          see description above
 * @param $add_attr_to_tag      see description above
 * @param $message              message object
 * @param $id                   message id
 * @param $recursively_called   boolean flag for recursive calls into this function (optional; default FALSE)
 * @return                      sanitized html safe to show on your pages.
 */
function sq_sanitize($body,
                     $tag_list,
                     $rm_tags_with_content,
                     $self_closing_tags,
                     $force_tag_closing,
                     $rm_attnames,
                     $bad_attvals,
                     $add_attr_to_tag,
                     $message,
                     $id,
                     $mailbox,
                     $recursively_called=FALSE
                     ){
    $me = 'sq_sanitize';

    /**
     * See if tag_list is of tags to remove or tags to allow.
     * false  means remove these tags
     * true   means allow these tags
     */
    $orig_tag_list = $tag_list;
    $rm_tags = array_shift($tag_list);

    /**
     * Normalize rm_tags and rm_tags_with_content.
     */
    @array_walk($tag_list, 'sq_casenormalize');
    @array_walk($rm_tags_with_content, 'sq_casenormalize');
    @array_walk($self_closing_tags, 'sq_casenormalize');

    $curpos = 0;
    $open_tags = Array();
    $trusted = "\n<!-- begin sanitized html -->\n";
    $skip_content = false;
    /**
     * Take care of netscape's stupid javascript entities like
     * &{alert('boo')};
     */
    $body = preg_replace("/&(\{.*?\};)/si", "&amp;\\1", $body);

    while (($curtag = sq_getnxtag($body, $curpos)) != FALSE){
        list($tagname, $attary, $tagtype, $lt, $gt) = $curtag;

        /**
         * RCDATA and RAWTEXT tags are handled differently:
         * next instance of closing tag is used, whether or not
         * the HTML is well formed before that
         */
        global $rcdata_rawtext_tags;
        if (!$recursively_called
         && in_array($tagname, $rcdata_rawtext_tags)
         && $tagtype === 1){
            $closing_tag = false;
            $closing_tag_offset = $curpos;
            // seek out the closing tag for the current RCDATA/RAWTEXT tag
            while (1) {
                // first we need to move forward to next available closing tag
                // (intentionally leave off the closing > and let sq_getnxtag() validate a proper tag syntax)
                $next_tag = sq_findnxreg($body, $closing_tag_offset, "</\s*$tagname");
                if ($next_tag === false) {
                    $closing_tag = false;
                    break;
                }
                // but then we have to make sure it's a well-formed tag
                $closing_tag = sq_getnxtag($body, $next_tag[0]);
                if ($closing_tag === false)
                    break;
                else if ($closing_tag[0] !== false
                 // these should be redundant
                 && $closing_tag[0] === $tagname && $closing_tag[2] === 2) {
                    $trusted .= sq_sanitize(substr($body, $curpos, $closing_tag[4] - $curpos + 1),
                                            $orig_tag_list, $rm_tags_with_content, $self_closing_tags,
                                            $force_tag_closing, $rm_attnames, $bad_attvals, $add_attr_to_tag,
                                            $message, $id, $mailbox, true);
                    $curpos = $closing_tag[4] + 1;
                    continue 2;
                }
                $closing_tag_offset = $next_tag[0] + 1;
            }
            if ($closing_tag === false)
            { /* no-op... there was no closing tag for this RCDATA/RAWTEXT tag - we could probably set $curpos to the end of $body, but this HTML is malformed anyway and should just fall apart on its own */ }
        }

        $free_content = substr($body, $curpos, $lt-$curpos);
        /**
         * Take care of <style>
         */
        if ($tagname == "style" && $tagtype == 1){
            list($free_content, $curpos) =
                sq_fixstyle($body, $gt+1, $message, $id, $mailbox);
            if ($free_content != FALSE){
                if ( !empty($attary) ) {
                    $attary = sq_fixatts($tagname,
                                         $attary,
                                         $rm_attnames,
                                         $bad_attvals,
                                         $add_attr_to_tag,
                                         $message,
                                         $id,
                                         $mailbox
                                         );
                }
                $trusted .= sq_tagprint($tagname, $attary, $tagtype);
                $trusted .= $free_content;
                $trusted .= sq_tagprint($tagname, false, 2);
            }
            continue;
        }
        if ($skip_content == false){
            $trusted .= $free_content;
        }
        if ($tagname != FALSE){
            if ($tagtype == 2){
                if ($skip_content == $tagname){
                    /**
                     * Got to the end of tag we needed to remove.
                     */
                    $tagname = false;
                    $skip_content = false;
                } else {
                    if ($skip_content == false){
                        if ($tagname == "body"){
                            $tagname = "div";
                        }
                        if (isset($open_tags{$tagname}) &&
                                $open_tags{$tagname} > 0){
                            $open_tags{$tagname}--;
                        } else {
                            $tagname = false;
                        }
                    }
                }
            } else {
                /**
                 * $rm_tags_with_content
                 */
                if ($skip_content == false){
                    /**
                     * See if this is a self-closing type and change
                     * tagtype appropriately.
                     */
                    if ($tagtype == 1
                            && in_array($tagname, $self_closing_tags)){
                        $tagtype = 3;
                    }
                    /**
                     * See if we should skip this tag and any content
                     * inside it.
                     */
                    if ($tagtype == 1 &&
                            in_array($tagname, $rm_tags_with_content)){
                        $skip_content = $tagname;
                    } else {
                        if (($rm_tags == false
                                    && in_array($tagname, $tag_list)) ||
                                ($rm_tags == true &&
                                 !in_array($tagname, $tag_list))){
                            $tagname = false;
                        } else {
                            /**
                             * Convert body into div.
                             */
                            if ($tagname == "body"){
                                $tagname = "div";
                                $attary = sq_body2div($attary, $mailbox,
                                        $message, $id);
                            }
                            if ($tagtype == 1){
                                if (isset($open_tags{$tagname})){
                                    $open_tags{$tagname}++;
                                } else {
                                    $open_tags{$tagname}=1;
                                }
                            }
                            /**
                             * This is where we run other checks.
                             */
                            if (is_array($attary) && sizeof($attary) > 0){
                                $attary = sq_fixatts($tagname,
                                                     $attary,
                                                     $rm_attnames,
                                                     $bad_attvals,
                                                     $add_attr_to_tag,
                                                     $message,
                                                     $id,
                                                     $mailbox
                                                     );
                            }
                        }
                    }
                }
            }
            if ($tagname != false && $skip_content == false){
                $trusted .= sq_tagprint($tagname, $attary, $tagtype);
            }
        }
        $curpos = $gt+1;
    }
    $trusted .= substr($body, $curpos, strlen($body)-$curpos);
    if ($force_tag_closing == true){
        foreach ($open_tags as $tagname=>$opentimes){
            while ($opentimes > 0){
                $trusted .= '</' . $tagname . '>';
                $opentimes--;
            }
        }
        $trusted .= "\n";
    }
    $trusted .= "<!-- end sanitized html -->\n";
    return $trusted;
}

/**
 * This is a wrapper function to call html sanitizing routines.
 *
 * @param  $body  the body of the message
 * @param  $id    the id of the message

 * @param  $message
 * @param  $mailbox
 * @param  boolean $take_mailto_links When TRUE, converts mailto: links
 *                                    into internal SM compose links
 *                                    (optional; default = TRUE)
 * @return        a string with html safe to display in the browser.
 */
function magicHTML($body, $id, $message, $mailbox = 'INBOX', $take_mailto_links =true) {

    // require_once(SM_PATH . 'functions/url_parser.php');  // for $MailTo_PReg_Match

    global $attachment_common_show_images, $view_unsafe_images,
           $has_unsafe_images, $allow_svg_display, $rcdata_rawtext_tags,
           $remove_rcdata_rawtext_tags_and_content;

    $rcdata_rawtext_tags = array(
        "noscript",
        "noframes",
        "noembed",
        "textarea",
        // also "title", "xmp", "script", "iframe", "plaintext" which we already remove below
    );

    /**
     * Don't display attached images in HTML mode.
     *
     * SB: why?
     */
    $attachment_common_show_images = false;
    $tag_list = Array(
            false, // remove these tags
            "meta",
            "html",
            "head",
            "base",
            "link",
            "frame",
            "iframe",
            "plaintext",
            "marquee",
            );

    $rm_tags_with_content = Array(
            "script",
            "object",
            "applet",
            "embed",
            "title",
            "frameset",
            "xmp",
            "xml",
            );
    if (!$allow_svg_display)
        $rm_tags_with_content[] = 'svg';
    /**
     * SquirrelMail will parse RCDATA and RAWTEXT tags and handle them as the special
     * case that they are, but if you prefer to remove them and their contents entirely
     * (in most cases, should be a safe thing with minimal impact), you can add the
     * following to config/config_local.php
     *    $remove_rcdata_rawtext_tags_and_content = TRUE; 
     */
    if ($remove_rcdata_rawtext_tags_and_content)
        $rm_tags_with_content = array_merge($rm_tags_with_content, $rcdata_rawtext_tags);

    $self_closing_tags =  Array(
            "img",
            "br",
            "hr",
            "input",
            "outbind",
            );

    $force_tag_closing = true;

    $rm_attnames = Array(
            "/.*/" =>
            Array(
                "/target/i",
                "/^on.*/i",
                "/^dynsrc/i",
                "/^data.*/i",
                "/^lowsrc.*/i",
                )
            );

    global $use_transparent_security_image;
    if ($use_transparent_security_image) $secremoveimg = '../images/spacer.png';
    else $secremoveimg = '../images/' . _("sec_remove_eng.png");

    $bad_attvals = Array(
            "/.*/" =>
            Array(
                "/^src|background/i" =>
                Array(
                    Array(
                        "/^([\'\"])\s*\S+script\s*:.*([\'\"])/si",
                        "/^([\'\"])\s*mocha\s*:*.*([\'\"])/si",
                        "/^([\'\"])\s*about\s*:.*([\'\"])/si"
                        ),
                    Array(
                        "\\1$secremoveimg\\2",
                        "\\1$secremoveimg\\2",
                        "\\1$secremoveimg\\2",
                        )
                    ),
                "/^href|action/i" =>
                Array(
                    Array(
                        "/^([\'\"])\s*\S+script\s*:.*([\'\"])/si",
                        "/^([\'\"])\s*mocha\s*:*.*([\'\"])/si",
                        "/^([\'\"])\s*about\s*:.*([\'\"])/si"
                        ),
                    Array(
                        "\\1#\\1",
                        "\\1#\\1",
                        "\\1#\\1"
                        )
                    ),
        "/^style/i" =>
            Array(
                Array(
                    "/\/\*.*\*\//",
                    "/expression/i",
                    "/binding/i",
                    "/behaviou*r/i",
                    "/include-source/i",

                    // position:relative can also be exploited
                    // to put content outside of email body area
                    // and position:fixed is similarly exploitable
                    // as position:absolute, so we'll remove it
                    // altogether....
                    //
                    // Does this screw up legitimate HTML messages?
                    // If so, the only fix I see is to allow position
                    // attributes (any values?  I think we still have
                    // to block static and fixed) only if $use_iframe
                    // is enabled (1.5.0+)
                    //
                    // was:   "/position\s*:\s*absolute/i",
                    //
                    "/position\s*:/i",

                    "/(\\\\)?u(\\\\)?r(\\\\)?l(\\\\)?/i",
                    "/url\s*\(\s*([\'\"])\s*\S+script\s*:.*([\'\"])\s*\)/si",
                    "/url\s*\(\s*([\'\"])\s*mocha\s*:.*([\'\"])\s*\)/si",
                    "/url\s*\(\s*([\'\"])\s*about\s*:.*([\'\"])\s*\)/si",
                    "/(.*)\s*:\s*url\s*\(\s*([\'\"]*)\s*\S+script\s*:.*([\'\"]*)\s*\)/si",
                    ),
                Array(
                    "",
                    "idiocy",
                    "idiocy",
                    "idiocy",
                    "idiocy",
                    "idiocy",
                    "url",
                    "url(\\1#\\1)",
                    "url(\\1#\\1)",
                    "url(\\1#\\1)",
                    "\\1:url(\\2#\\3)"
                    )
                )
            )
        );

    // If there's no "view_unsafe_images" variable in the URL, turn unsafe
    // images off by default.
    sqgetGlobalVar('view_unsafe_images', $view_unsafe_images, SQ_GET, FALSE);

    if (!$view_unsafe_images){
        /**
         * Remove any references to http/https if view_unsafe_images set
         * to false.
         */
        array_push($bad_attvals{'/.*/'}{'/^src|background/i'}[0],
                '/^([\'\"])\s*https*:.*([\'\"])/si');
        array_push($bad_attvals{'/.*/'}{'/^src|background/i'}[1],
                "\\1$secremoveimg\\1");
        array_push($bad_attvals{'/.*/'}{'/^style/i'}[0],
                '/url\([\'\"]?https?:[^\)]*[\'\"]?\)/si');
        array_push($bad_attvals{'/.*/'}{'/^style/i'}[1],
                "url(\\1$secremoveimg\\1)");
    }

    $add_attr_to_tag = Array(
            "/^a$/i" =>
            Array('target'=>'"_blank"',
                'title'=>'"'._("This external link will open in a new window").'"'
                )
            );
    $trusted = sq_sanitize($body,
                           $tag_list,
                           $rm_tags_with_content,
                           $self_closing_tags,
                           $force_tag_closing,
                           $rm_attnames,
                           $bad_attvals,
                           $add_attr_to_tag,
                           $message,
                           $id,
                           $mailbox
                           );
    if (strpos($trusted,$secremoveimg)){
        $has_unsafe_images = true;
    }

    // we want to parse mailto's in HTML output, change to SM compose links
    // this is a modified version of code from url_parser.php... but Marc is
    // right: we need a better filtering implementation; adding this randomly
    // here is not a great solution
    //
    if ($take_mailto_links) {
        // parseUrl($trusted);   // this even parses URLs inside of tags... too aggressive
        global $MailTo_PReg_Match;
        // some mailers (Microsoft, surprise surprise) produce mailto strings without being
        // inside an anchor (link) tag, so we have to make sure the regex looks for the
        // quote before mailto, and we'll also try to convert the non-links back into links
        $MailTo_PReg_Match = '/([\'"])?mailto:' . substr($MailTo_PReg_Match, 1) ;
        if ((preg_match_all($MailTo_PReg_Match, $trusted, $regs)) && ($regs[0][0] != '')) {
            foreach ($regs[0] as $i => $mailto_before) {
                $mailto_params = $regs[11][$i];

                // get rid of any leading quote we may have captured but don't care about
                //
                $mailto_before = ltrim($mailto_before, '"\'');

                // get rid of any tailing quote since we have to add send_to to the end
                //
                $mailto_before = rtrim($mailto_before, '"\'');
                $mailto_params = rtrim($mailto_params, '"\'');

                if ($regs[2][$i]) {    //if there is an email addr before '?', we need to merge it with the params
                    $to = 'to=' . $regs[2][$i];
                    if (strpos($mailto_params, 'to=') > -1)    //already a 'to='
                        $mailto_params = str_replace('to=', $to . '%2C%20', $mailto_params);
                    else {
                        if ($mailto_params)    //already some params, append to them
                            $mailto_params .= '&amp;' . $to;
                        else
                            $mailto_params .= '?' . $to;
                    }
                }

                $url_str = preg_replace(array('/to=/i', '/(?<!b)cc=/i', '/bcc=/i'), array('send_to=', 'send_to_cc=', 'send_to_bcc='), $mailto_params);

                // we'll already have target=_blank, no need to allow comp_in_new
                // here (which would be a lot more work anyway)
                //
                global $compose_new_win;
                $temp_comp_in_new = $compose_new_win;
                $compose_new_win = 0;
                $comp_uri = makeComposeLink('src/compose.php' . $url_str, $mailto_before);
                $compose_new_win = $temp_comp_in_new;

                // remove <a href=" and anything after the next quote (we only
                // need the uri, not the link HTML) in compose uri
                //
                // but only do this if the original mailto was in a real anchor tag
                //
                if (!empty($regs[1][$i])) {
                    $comp_uri = substr($comp_uri, 9);
                    $comp_uri = substr($comp_uri, 0, strpos($comp_uri, '"', 1));
                }
                $trusted = str_replace($mailto_before, $comp_uri, $trusted);
            }
        }
    }

    return $trusted;
}

/**
 * function SendDownloadHeaders - send file to the browser
 *
 * Original Source: SM core src/download.php
 * moved here to make it available to other code, and separate
 * front end from back end functionality.
 *
 * @param string $type0 first half of mime type
 * @param string $type1 second half of mime type
 * @param string $filename filename to tell the browser for downloaded file
 * @param boolean $force whether to force the download dialog to pop
 * @param optional integer $filesize send the Content-Header and length to the browser
 * @return void
 */
function SendDownloadHeaders($type0, $type1, $filename, $force, $filesize=0) {
    global $languages, $squirrelmail_language;
    $isIE = $isIE6plus = false;

    sqgetGlobalVar('HTTP_USER_AGENT', $HTTP_USER_AGENT, SQ_SERVER);

    if (strstr($HTTP_USER_AGENT, 'compatible; MSIE ') !== false &&
            strstr($HTTP_USER_AGENT, 'Opera') === false) {
        $isIE = true;
    }

    if (preg_match('/compatible; MSIE ([0-9]+)/', $HTTP_USER_AGENT, $match) &&
        ((int)$match[1]) >= 6 && strstr($HTTP_USER_AGENT, 'Opera') === false) {
        $isIE6plus = true;
    }

    if (isset($languages[$squirrelmail_language]['XTRA_CODE']) &&
            function_exists($languages[$squirrelmail_language]['XTRA_CODE'] . '_downloadfilename')) {
        $filename =
            call_user_func($languages[$squirrelmail_language]['XTRA_CODE'] . '_downloadfilename', $filename, $HTTP_USER_AGENT);
    } else {
        $filename = preg_replace('/[\\\\\/:*?"<>|;]/', '_', str_replace('&nbsp;', ' ', $filename));
    }

    // A Pox on Microsoft and it's Internet Explorer!
    //
    // IE has lots of bugs with file downloads.
    // It also has problems with SSL.  Both of these cause problems
    // for us in this function.
    //
    // See this article on Cache Control headers and SSL
    // http://support.microsoft.com/default.aspx?scid=kb;en-us;323308
    //
    // The best thing you can do for IE is to upgrade to the latest
    // version
    //set all the Cache Control Headers for IE
    if ($isIE) {
        $filename=rawurlencode($filename);
        header ("Pragma: public");
        header ("Cache-Control: no-store, max-age=0, no-cache, must-revalidate"); // HTTP/1.1
        // does nothing - see: https://blogs.msdn.microsoft.com/ieinternals/2009/07/20/internet-explorers-cache-control-extensions/
        // header ("Cache-Control: post-check=0, pre-check=0", false);
        header ("Cache-Control: private");

        //set the inline header for IE, we'll add the attachment header later if we need it
        header ("Content-Disposition: inline; filename=$filename");
    }

    if (!$force) {
        // Try to show in browser window
        header ("Content-Disposition: inline; filename=\"$filename\"");
        header ("Content-Type: $type0/$type1; name=\"$filename\"");
    } else {
        // Try to pop up the "save as" box

        // IE makes this hard.  It pops up 2 save boxes, or none.
        // http://support.microsoft.com/support/kb/articles/Q238/5/88.ASP
        // http://support.microsoft.com/default.aspx?scid=kb;EN-US;260519
        // But, according to Microsoft, it is "RFC compliant but doesn't
        // take into account some deviations that allowed within the
        // specification."  Doesn't that mean RFC non-compliant?
        // http://support.microsoft.com/support/kb/articles/Q258/4/52.ASP

        // all browsers need the application/octet-stream header for this
        header ("Content-Type: application/octet-stream; name=\"$filename\"");

        // http://support.microsoft.com/support/kb/articles/Q182/3/15.asp
        // Do not have quotes around filename, but that applied to
        // "attachment"... does it apply to inline too?
        header ("Content-Disposition: attachment; filename=\"$filename\"");

        if ($isIE && !$isIE6plus) {
            // This combination seems to work mostly.  IE 5.5 SP 1 has
            // known issues (see the Microsoft Knowledge Base)

            // This works for most types, but doesn't work with Word files
            header ("Content-Type: application/download; name=\"$filename\"");
            header ("Content-Type: application/force-download; name=\"$filename\"");
            // These are spares, just in case.  :-)
            //header("Content-Type: $type0/$type1; name=\"$filename\"");
            //header("Content-Type: application/x-msdownload; name=\"$filename\"");
            //header("Content-Type: application/octet-stream; name=\"$filename\"");
        } else if ($isIE) {
             // This is to prevent IE for MIME sniffing and auto open a file in IE
             header ("Content-Type: application/force-download; name=\"$filename\"");
        } else {
            // another application/octet-stream forces download for Netscape
            header ("Content-Type: application/octet-stream; name=\"$filename\"");
        }
    }

    //send the content-length header if the calling function provides it
    if ($filesize > 0) {
        header("Content-Length: $filesize");
    }

}  // end fn SendDownloadHeaders