/lib/weblib.php

<?php

// This file is part of Moodle - http://moodle.org/
//
// Moodle is free software: you can redistribute it and/or modify
// it under the terms of the GNU General Public License as published by
// the Free Software Foundation, either version 3 of the License, or
// (at your option) any later version.
//
// Moodle is distributed in the hope that it will be useful,
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
// GNU General Public License for more details.
//
// You should have received a copy of the GNU General Public License
// along with Moodle.  If not, see <http://www.gnu.org/licenses/>.

/**
 * Library of functions for web output
 *
 * Library of all general-purpose Moodle PHP functions and constants
 * that produce HTML output
 *
 * Other main libraries:
 * - datalib.php - functions that access the database.
 * - moodlelib.php - general-purpose Moodle functions.
 *
 * @package    core
 * @subpackage lib
 * @copyright  1999 onwards Martin Dougiamas {@link http://moodle.com}
 * @license    http://www.gnu.org/copyleft/gpl.html GNU GPL v3 or later
 */

defined('MOODLE_INTERNAL') || die();

/// Constants

/// Define text formatting types ... eventually we can add Wiki, BBcode etc

/**
 * Does all sorts of transformations and filtering
 */
define('FORMAT_MOODLE',   '0');   // Does all sorts of transformations and filtering

/**
 * Plain HTML (with some tags stripped)
 */
define('FORMAT_HTML',     '1');   // Plain HTML (with some tags stripped)

/**
 * Plain text (even tags are printed in full)
 */
define('FORMAT_PLAIN',    '2');   // Plain text (even tags are printed in full)

/**
 * Wiki-formatted text
 * Deprecated: left here just to note that '3' is not used (at the moment)
 * and to catch any latent wiki-like text (which generates an error)
 */
define('FORMAT_WIKI',     '3');   // Wiki-formatted text

/**
 * Markdown-formatted text http://daringfireball.net/projects/markdown/
 */
define('FORMAT_MARKDOWN', '4');   // Markdown-formatted text http://daringfireball.net/projects/markdown/

/**
 * A moodle_url comparison using this flag will return true if the base URLs match, params are ignored
 */
define('URL_MATCH_BASE', 0);
/**
 * A moodle_url comparison using this flag will return true if the base URLs match and the params of url1 are part of url2
 */
define('URL_MATCH_PARAMS', 1);
/**
 * A moodle_url comparison using this flag will return true if the two URLs are identical, except for the order of the params
 */
define('URL_MATCH_EXACT', 2);

/// Functions

/**
 * Add quotes to HTML characters
 *
 * Returns $var with HTML characters (like "<", ">", etc.) properly quoted.
 * This function is very similar to {@link p()}
 *
 * @todo Remove obsolete param $obsolete if not used anywhere
 *
 * @param string $var the string potentially containing HTML characters
 * @param boolean $obsolete no longer used.
 * @return string
 */
function s($var, $obsolete = false) {

    if ($var === '0' or $var === false or $var === 0) {
        return '0';
    }

    return preg_replace("/&amp;#(\d+|x[0-7a-fA-F]+);/i", "&#$1;", htmlspecialchars($var, ENT_QUOTES, 'UTF-8', true));
}

/**
 * Add quotes to HTML characters
 *
 * Prints $var with HTML characters (like "<", ">", etc.) properly quoted.
 * This function simply calls {@link s()}
 * @see s()
 *
 * @todo Remove obsolete param $obsolete if not used anywhere
 *
 * @param string $var the string potentially containing HTML characters
 * @param boolean $obsolete no longer used.
 * @return string
 */
function p($var, $obsolete = false) {
    echo s($var, $obsolete);
}

/**
 * Does proper javascript quoting.
 *
 * Do not use addslashes anymore, because it does not work when magic_quotes_sybase is enabled.
 *
 * @param mixed $var String, Array, or Object to add slashes to
 * @return mixed quoted result
 */
function addslashes_js($var) {
    if (is_string($var)) {
        $var = str_replace('\\', '\\\\', $var);
        $var = str_replace(array('\'', '"', "\n", "\r", "\0"), array('\\\'', '\\"', '\\n', '\\r', '\\0'), $var);
        $var = str_replace('</', '<\/', $var);   // XHTML compliance
    } else if (is_array($var)) {
        $var = array_map('addslashes_js', $var);
    } else if (is_object($var)) {
        $a = get_object_vars($var);
        foreach ($a as $key=>$value) {
          $a[$key] = addslashes_js($value);
        }
        $var = (object)$a;
    }
    return $var;
}

/**
 * Remove query string from url
 *
 * Takes in a URL and returns it without the querystring portion
 *
 * @param string $url the url which may have a query string attached
 * @return string The remaining URL
 */
 function strip_querystring($url) {

    if ($commapos = strpos($url, '?')) {
        return substr($url, 0, $commapos);
    } else {
        return $url;
    }
}

/**
 * Returns the URL of the HTTP_REFERER, less the querystring portion if required
 *
 * @uses $_SERVER
 * @param boolean $stripquery if true, also removes the query part of the url.
 * @return string The resulting referer or empty string
 */
function get_referer($stripquery=true) {
    if (isset($_SERVER['HTTP_REFERER'])) {
        if ($stripquery) {
            return strip_querystring($_SERVER['HTTP_REFERER']);
        } else {
            return $_SERVER['HTTP_REFERER'];
        }
    } else {
        return '';
    }
}


/**
 * Returns the name of the current script, WITH the querystring portion.
 *
 * This function is necessary because PHP_SELF and REQUEST_URI and SCRIPT_NAME
 * return different things depending on a lot of things like your OS, Web
 * server, and the way PHP is compiled (ie. as a CGI, module, ISAPI, etc.)
 * <b>NOTE:</b> This function returns false if the global variables needed are not set.
 *
 * @return mixed String, or false if the global variables needed are not set
 */
function me() {
    global $ME;
    return $ME;
}

/**
 * Guesses the full URL of the current script.
 *
 * This function is using $PAGE->url, but may fall back to $FULLME which
 * is constructed from  PHP_SELF and REQUEST_URI or SCRIPT_NAME
 *
 * @return mixed full page URL string or false if unknown
 */
function qualified_me() {
    global $FULLME, $PAGE, $CFG;

    if (isset($PAGE) and $PAGE->has_set_url()) {
        // this is the only recommended way to find out current page
        return $PAGE->url->out(false);

    } else {
        if ($FULLME === null) {
            // CLI script most probably
            return false;
        }
        if (!empty($CFG->sslproxy)) {
            // return only https links when using SSL proxy
            return preg_replace('/^http:/', 'https:', $FULLME, 1);
        } else {
            return $FULLME;
        }
    }
}

/**
 * Class for creating and manipulating urls.
 *
 * It can be used in moodle pages where config.php has been included without any further includes.
 *
 * It is useful for manipulating urls with long lists of params.
 * One situation where it will be useful is a page which links to itself to perform various actions
 * and / or to process form data. A moodle_url object :
 * can be created for a page to refer to itself with all the proper get params being passed from page call to
 * page call and methods can be used to output a url including all the params, optionally adding and overriding
 * params and can also be used to
 *     - output the url without any get params
 *     - and output the params as hidden fields to be output within a form
 *
 * @link http://docs.moodle.org/dev/lib/weblib.php_moodle_url See short write up here
 * @license http://www.gnu.org/copyleft/gpl.html GNU GPL v3 or later
 * @package moodlecore
 */
class moodle_url {
    /**
     * Scheme, ex.: http, https
     * @var string
     */
    protected $scheme = '';
    /**
     * hostname
     * @var string
     */
    protected $host = '';
    /**
     * Port number, empty means default 80 or 443 in case of http
     * @var unknown_type
     */
    protected $port = '';
    /**
     * Username for http auth
     * @var string
     */
    protected $user = '';
    /**
     * Password for http auth
     * @var string
     */
    protected $pass = '';
    /**
     * Script path
     * @var string
     */
    protected $path = '';
    /**
     * Optional slash argument value
     * @var string
     */
    protected $slashargument = '';
    /**
     * Anchor, may be also empty, null means none
     * @var string
     */
    protected $anchor = null;
    /**
     * Url parameters as associative array
     * @var array
     */
    protected $params = array(); // Associative array of query string params

    /**
     * Create new instance of moodle_url.
     *
     * @param moodle_url|string $url - moodle_url means make a copy of another
     *      moodle_url and change parameters, string means full url or shortened
     *      form (ex.: '/course/view.php'). It is strongly encouraged to not include
     *      query string because it may result in double encoded values. Use the
     *      $params instead. For admin URLs, just use /admin/script.php, this
     *      class takes care of the $CFG->admin issue.
     * @param array $params these params override current params or add new
     */
    public function __construct($url, array $params = null) {
        global $CFG;

        if ($url instanceof moodle_url) {
            $this->scheme = $url->scheme;
            $this->host = $url->host;
            $this->port = $url->port;
            $this->user = $url->user;
            $this->pass = $url->pass;
            $this->path = $url->path;
            $this->slashargument = $url->slashargument;
            $this->params = $url->params;
            $this->anchor = $url->anchor;

        } else {
            // detect if anchor used
            $apos = strpos($url, '#');
            if ($apos !== false) {
                $anchor = substr($url, $apos);
                $anchor = ltrim($anchor, '#');
                $this->set_anchor($anchor);
                $url = substr($url, 0, $apos);
            }

            // normalise shortened form of our url ex.: '/course/view.php'
            if (strpos($url, '/') === 0) {
                // we must not use httpswwwroot here, because it might be url of other page,
                // devs have to use httpswwwroot explicitly when creating new moodle_url
                $url = $CFG->wwwroot.$url;
            }

            // now fix the admin links if needed, no need to mess with httpswwwroot
            if ($CFG->admin !== 'admin') {
                if (strpos($url, "$CFG->wwwroot/admin/") === 0) {
                    $url = str_replace("$CFG->wwwroot/admin/", "$CFG->wwwroot/$CFG->admin/", $url);
                }
            }

            // parse the $url
            $parts = parse_url($url);
            if ($parts === false) {
                throw new moodle_exception('invalidurl');
            }
            if (isset($parts['query'])) {
                // note: the values may not be correctly decoded,
                //       url parameters should be always passed as array
                parse_str(str_replace('&amp;', '&', $parts['query']), $this->params);
            }
            unset($parts['query']);
            foreach ($parts as $key => $value) {
                $this->$key = $value;
            }

            // detect slashargument value from path - we do not support directory names ending with .php
            $pos = strpos($this->path, '.php/');
            if ($pos !== false) {
                $this->slashargument = substr($this->path, $pos + 4);
                $this->path = substr($this->path, 0, $pos + 4);
            }
        }

        $this->params($params);
    }

    /**
     * Add an array of params to the params for this url.
     *
     * The added params override existing ones if they have the same name.
     *
     * @param array $params Defaults to null. If null then returns all params.
     * @return array Array of Params for url.
     */
    public function params(array $params = null) {
        $params = (array)$params;

        foreach ($params as $key=>$value) {
            if (is_int($key)) {
                throw new coding_exception('Url parameters can not have numeric keys!');
            }
            if (!is_string($value)) {
                if (is_array($value)) {
                    throw new coding_exception('Url parameters values can not be arrays!');
                }
                if (is_object($value) and !method_exists($value, '__toString')) {
                    throw new coding_exception('Url parameters values can not be objects, unless __toString() is defined!');
                }
            }
            $this->params[$key] = (string)$value;
        }
        return $this->params;
    }

    /**
     * Remove all params if no arguments passed.
     * Remove selected params if arguments are passed.
     *
     * Can be called as either remove_params('param1', 'param2')
     * or remove_params(array('param1', 'param2')).
     *
     * @param mixed $params either an array of param names, or a string param name,
     * @param string $params,... any number of additional param names.
     * @return array url parameters
     */
    public function remove_params($params = null) {
        if (!is_array($params)) {
            $params = func_get_args();
        }
        foreach ($params as $param) {
            unset($this->params[$param]);
        }
        return $this->params;
    }

    /**
     * Remove all url parameters
     * @param $params
     * @return void
     */
    public function remove_all_params($params = null) {
        $this->params = array();
        $this->slashargument = '';
    }

    /**
     * Add a param to the params for this url.
     *
     * The added param overrides existing one if they have the same name.
     *
     * @param string $paramname name
     * @param string $newvalue Param value. If new value specified current value is overriden or parameter is added
     * @return mixed string parameter value, null if parameter does not exist
     */
    public function param($paramname, $newvalue = '') {
        if (func_num_args() > 1) {
            // set new value
            $this->params(array($paramname=>$newvalue));
        }
        if (isset($this->params[$paramname])) {
            return $this->params[$paramname];
        } else {
            return null;
        }
    }

    /**
     * Merges parameters and validates them
     * @param array $overrideparams
     * @return array merged parameters
     */
    protected function merge_overrideparams(array $overrideparams = null) {
        $overrideparams = (array)$overrideparams;
        $params = $this->params;
        foreach ($overrideparams as $key=>$value) {
            if (is_int($key)) {
                throw new coding_exception('Overridden parameters can not have numeric keys!');
            }
            if (is_array($value)) {
                throw new coding_exception('Overridden parameters values can not be arrays!');
            }
            if (is_object($value) and !method_exists($value, '__toString')) {
                throw new coding_exception('Overridden parameters values can not be objects, unless __toString() is defined!');
            }
            $params[$key] = (string)$value;
        }
        return $params;
    }

    /**
     * Get the params as as a query string.
     * This method should not be used outside of this method.
     *
     * @param boolean $escaped Use &amp; as params separator instead of plain &
     * @param array $overrideparams params to add to the output params, these
     *      override existing ones with the same name.
     * @return string query string that can be added to a url.
     */
    public function get_query_string($escaped = true, array $overrideparams = null) {
        $arr = array();
        if ($overrideparams !== null) {
            $params = $this->merge_overrideparams($overrideparams);
        } else {
            $params = $this->params;
        }
        foreach ($params as $key => $val) {
            if (is_array($val)) {
                foreach ($val as $index => $value) {
                    $arr[] = rawurlencode($key.'['.$index.']')."=".rawurlencode($value);
                }
            } else {
                $arr[] = rawurlencode($key)."=".rawurlencode($val);
            }
        }
        if ($escaped) {
            return implode('&amp;', $arr);
        } else {
            return implode('&', $arr);
        }
    }

    /**
     * Shortcut for printing of encoded URL.
     * @return string
     */
    public function __toString() {
        return $this->out(true);
    }

    /**
     * Output url
     *
     * If you use the returned URL in HTML code, you want the escaped ampersands. If you use
     * the returned URL in HTTP headers, you want $escaped=false.
     *
     * @param boolean $escaped Use &amp; as params separator instead of plain &
     * @param array $overrideparams params to add to the output url, these override existing ones with the same name.
     * @return string Resulting URL
     */
    public function out($escaped = true, array $overrideparams = null) {
        if (!is_bool($escaped)) {
            debugging('Escape parameter must be of type boolean, '.gettype($escaped).' given instead.');
        }

        $uri = $this->out_omit_querystring().$this->slashargument;

        $querystring = $this->get_query_string($escaped, $overrideparams);
        if ($querystring !== '') {
            $uri .= '?' . $querystring;
        }
        if (!is_null($this->anchor)) {
            $uri .= '#'.$this->anchor;
        }

        return $uri;
    }

    /**
     * Returns url without parameters, everything before '?'.
     *
     * @param bool $includeanchor if {@link self::anchor} is defined, should it be returned?
     * @return string
     */
    public function out_omit_querystring($includeanchor = false) {

        $uri = $this->scheme ? $this->scheme.':'.((strtolower($this->scheme) == 'mailto') ? '':'//'): '';
        $uri .= $this->user ? $this->user.($this->pass? ':'.$this->pass:'').'@':'';
        $uri .= $this->host ? $this->host : '';
        $uri .= $this->port ? ':'.$this->port : '';
        $uri .= $this->path ? $this->path : '';
        if ($includeanchor and !is_null($this->anchor)) {
            $uri .= '#' . $this->anchor;
        }

        return $uri;
    }

    /**
     * Compares this moodle_url with another
     * See documentation of constants for an explanation of the comparison flags.
     * @param moodle_url $url The moodle_url object to compare
     * @param int $matchtype The type of comparison (URL_MATCH_BASE, URL_MATCH_PARAMS, URL_MATCH_EXACT)
     * @return boolean
     */
    public function compare(moodle_url $url, $matchtype = URL_MATCH_EXACT) {

        $baseself = $this->out_omit_querystring();
        $baseother = $url->out_omit_querystring();

        // Append index.php if there is no specific file
        if (substr($baseself,-1)=='/') {
            $baseself .= 'index.php';
        }
        if (substr($baseother,-1)=='/') {
            $baseother .= 'index.php';
        }

        // Compare the two base URLs
        if ($baseself != $baseother) {
            return false;
        }

        if ($matchtype == URL_MATCH_BASE) {
            return true;
        }

        $urlparams = $url->params();
        foreach ($this->params() as $param => $value) {
            if ($param == 'sesskey') {
                continue;
            }
            if (!array_key_exists($param, $urlparams) || $urlparams[$param] != $value) {
                return false;
            }
        }

        if ($matchtype == URL_MATCH_PARAMS) {
            return true;
        }

        foreach ($urlparams as $param => $value) {
            if ($param == 'sesskey') {
                continue;
            }
            if (!array_key_exists($param, $this->params()) || $this->param($param) != $value) {
                return false;
            }
        }

        return true;
    }

    /**
     * Sets the anchor for the URI (the bit after the hash)
     * @param string $anchor null means remove previous
     */
    public function set_anchor($anchor) {
        if (is_null($anchor)) {
            // remove
            $this->anchor = null;
        } else if ($anchor === '') {
            // special case, used as empty link
            $this->anchor = '';
        } else if (preg_match('|[a-zA-Z\_\:][a-zA-Z0-9\_\-\.\:]*|', $anchor)) {
            // Match the anchor against the NMTOKEN spec
            $this->anchor = $anchor;
        } else {
            // bad luck, no valid anchor found
            $this->anchor = null;
        }
    }

    /**
     * Sets the url slashargument value
     * @param string $path usually file path
     * @param string $parameter name of page parameter if slasharguments not supported
     * @param bool $supported usually null, then it depends on $CFG->slasharguments, use true or false for other servers
     * @return void
     */
    public function set_slashargument($path, $parameter = 'file', $supported = NULL) {
        global $CFG;
        if (is_null($supported)) {
            $supported = $CFG->slasharguments;
        }

        if ($supported) {
            $parts = explode('/', $path);
            $parts = array_map('rawurlencode', $parts);
            $path  = implode('/', $parts);
            $this->slashargument = $path;
            unset($this->params[$parameter]);

        } else {
            $this->slashargument = '';
            $this->params[$parameter] = $path;
        }
    }

    // == static factory methods ==

    /**
     * General moodle file url.
     * @param string $urlbase the script serving the file
     * @param string $path
     * @param bool $forcedownload
     * @return moodle_url
     */
    public static function make_file_url($urlbase, $path, $forcedownload = false) {
        global $CFG;

        $params = array();
        if ($forcedownload) {
            $params['forcedownload'] = 1;
        }

        $url = new moodle_url($urlbase, $params);
        $url->set_slashargument($path);

        return $url;
    }

    /**
     * Factory method for creation of url pointing to plugin file.
     * Please note this method can be used only from the plugins to
     * create urls of own files, it must not be used outside of plugins!
     * @param int $contextid
     * @param string $component
     * @param string $area
     * @param int $itemid
     * @param string $pathname
     * @param string $filename
     * @param bool $forcedownload
     * @return moodle_url
     */
    public static function make_pluginfile_url($contextid, $component, $area, $itemid, $pathname, $filename, $forcedownload = false) {
        global $CFG;
        $urlbase = "$CFG->httpswwwroot/pluginfile.php";
        if ($itemid === NULL) {
            return self::make_file_url($urlbase, "/$contextid/$component/$area".$pathname.$filename, $forcedownload);
        } else {
            return self::make_file_url($urlbase, "/$contextid/$component/$area/$itemid".$pathname.$filename, $forcedownload);
        }
    }

    /**
     * Factory method for creation of url pointing to draft
     * file of current user.
     * @param int $draftid draft item id
     * @param string $pathname
     * @param string $filename
     * @param bool $forcedownload
     * @return moodle_url
     */
    public static function make_draftfile_url($draftid, $pathname, $filename, $forcedownload = false) {
        global $CFG, $USER;
        $urlbase = "$CFG->httpswwwroot/draftfile.php";
        $context = context_user::instance($USER->id);

        return self::make_file_url($urlbase, "/$context->id/user/draft/$draftid".$pathname.$filename, $forcedownload);
    }

    /**
     * Factory method for creating of links to legacy
     * course files.
     * @param int $courseid
     * @param string $filepath
     * @param bool $forcedownload
     * @return moodle_url
     */
    public static function make_legacyfile_url($courseid, $filepath, $forcedownload = false) {
        global $CFG;

        $urlbase = "$CFG->wwwroot/file.php";
        return self::make_file_url($urlbase, '/'.$courseid.'/'.$filepath, $forcedownload);
    }

    /**
     * Returns URL a relative path from $CFG->wwwroot
     *
     * Can be used for passing around urls with the wwwroot stripped
     *
     * @param boolean $escaped Use &amp; as params separator instead of plain &
     * @param array $overrideparams params to add to the output url, these override existing ones with the same name.
     * @return string Resulting URL
     * @throws coding_exception if called on a non-local url
     */
    public function out_as_local_url($escaped = true, array $overrideparams = null) {
        global $CFG;

        $url = $this->out($escaped, $overrideparams);
        $httpswwwroot = str_replace("http://", "https://", $CFG->wwwroot);

        // $url should be equal to wwwroot or httpswwwroot. If not then throw exception.
        if (($url === $CFG->wwwroot) || (strpos($url, $CFG->wwwroot.'/') === 0)) {
            $localurl = substr($url, strlen($CFG->wwwroot));
            return !empty($localurl) ? $localurl : '';
        } else if (($url === $httpswwwroot) || (strpos($url, $httpswwwroot.'/') === 0)) {
            $localurl = substr($url, strlen($httpswwwroot));
            return !empty($localurl) ? $localurl : '';
        } else {
            throw new coding_exception('out_as_local_url called on a non-local URL');
        }
    }

    /**
     * Returns the 'path' portion of a URL. For example, if the URL is
     * http://www.example.org:447/my/file/is/here.txt?really=1 then this will
     * return '/my/file/is/here.txt'.
     *
     * By default the path includes slash-arguments (for example,
     * '/myfile.php/extra/arguments') so it is what you would expect from a
     * URL path. If you don't want this behaviour, you can opt to exclude the
     * slash arguments. (Be careful: if the $CFG variable slasharguments is
     * disabled, these URLs will have a different format and you may need to
     * look at the 'file' parameter too.)
     *
     * @param bool $includeslashargument If true, includes slash arguments
     * @return string Path of URL
     */
    public function get_path($includeslashargument = true) {
        return $this->path . ($includeslashargument ? $this->slashargument : '');
    }

    /**
     * Returns a given parameter value from the URL.
     *
     * @param string $name Name of parameter
     * @return string Value of parameter or null if not set
     */
    public function get_param($name) {
        if (array_key_exists($name, $this->params)) {
            return $this->params[$name];
        } else {
            return null;
        }
    }
}

/**
 * Determine if there is data waiting to be processed from a form
 *
 * Used on most forms in Moodle to check for data
 * Returns the data as an object, if it's found.
 * This object can be used in foreach loops without
 * casting because it's cast to (array) automatically
 *
 * Checks that submitted POST data exists and returns it as object.
 *
 * @uses $_POST
 * @return mixed false or object
 */
function data_submitted() {

    if (empty($_POST)) {
        return false;
    } else {
        return (object)fix_utf8($_POST);
    }
}

/**
 * Given some normal text this function will break up any
 * long words to a given size by inserting the given character
 *
 * It's multibyte savvy and doesn't change anything inside html tags.
 *
 * @param string $string the string to be modified
 * @param int $maxsize maximum length of the string to be returned
 * @param string $cutchar the string used to represent word breaks
 * @return string
 */
function break_up_long_words($string, $maxsize=20, $cutchar=' ') {

/// First of all, save all the tags inside the text to skip them
    $tags = array();
    filter_save_tags($string,$tags);

/// Process the string adding the cut when necessary
    $output = '';
    $length = textlib::strlen($string);
    $wordlength = 0;

    for ($i=0; $i<$length; $i++) {
        $char = textlib::substr($string, $i, 1);
        if ($char == ' ' or $char == "\t" or $char == "\n" or $char == "\r" or $char == "<" or $char == ">") {
            $wordlength = 0;
        } else {
            $wordlength++;
            if ($wordlength > $maxsize) {
                $output .= $cutchar;
                $wordlength = 0;
            }
        }
        $output .= $char;
    }

/// Finally load the tags back again
    if (!empty($tags)) {
        $output = str_replace(array_keys($tags), $tags, $output);
    }

    return $output;
}

/**
 * Try and close the current window using JavaScript, either immediately, or after a delay.
 *
 * Echo's out the resulting XHTML & javascript
 *
 * @global object
 * @global object
 * @param integer $delay a delay in seconds before closing the window. Default 0.
 * @param boolean $reloadopener if true, we will see if this window was a pop-up, and try
 *      to reload the parent window before this one closes.
 */
function close_window($delay = 0, $reloadopener = false) {
    global $PAGE, $OUTPUT;

    if (!$PAGE->headerprinted) {
        $PAGE->set_title(get_string('closewindow'));
        echo $OUTPUT->header();
    } else {
        $OUTPUT->container_end_all(false);
    }

    if ($reloadopener) {
        // Trigger the reload immediately, even if the reload is after a delay.
        $PAGE->requires->js_function_call('window.opener.location.reload', array(true));
    }
    $OUTPUT->notification(get_string('windowclosing'), 'notifysuccess');

    $PAGE->requires->js_function_call('close_window', array(new stdClass()), false, $delay);

    echo $OUTPUT->footer();
    exit;
}

/**
 * Returns a string containing a link to the user documentation for the current
 * page. Also contains an icon by default. Shown to teachers and admin only.
 *
 * @global object
 * @global object
 * @param string $text The text to be displayed for the link
 * @param string $iconpath The path to the icon to be displayed
 * @return string The link to user documentation for this current page
 */
function page_doc_link($text='') {
    global $CFG, $PAGE, $OUTPUT;

    if (empty($CFG->docroot) || during_initial_install()) {
        return '';
    }
    if (!has_capability('moodle/site:doclinks', $PAGE->context)) {
        return '';
    }

    $path = $PAGE->docspath;
    if (!$path) {
        return '';
    }
    return $OUTPUT->doc_link($path, $text);
}


/**
 * Validates an email to make sure it makes sense.
 *
 * @param string $address The email address to validate.
 * @return boolean
 */
function validate_email($address) {

    return (preg_match('#^[-!\#$%&\'*+\\/0-9=?A-Z^_`a-z{|}~]+'.
                 '(\.[-!\#$%&\'*+\\/0-9=?A-Z^_`a-z{|}~]+)*'.
                  '@'.
                  '[-!\#$%&\'*+\\/0-9=?A-Z^_`a-z{|}~]+\.'.
                  '[-!\#$%&\'*+\\./0-9=?A-Z^_`a-z{|}~]+$#',
                  $address));
}

/**
 * Extracts file argument either from file parameter or PATH_INFO
 * Note: $scriptname parameter is not needed anymore
 *
 * @global string
 * @uses $_SERVER
 * @uses PARAM_PATH
 * @return string file path (only safe characters)
 */
function get_file_argument() {
    global $SCRIPT;

    $relativepath = optional_param('file', FALSE, PARAM_PATH);

    if ($relativepath !== false and $relativepath !== '') {
        return $relativepath;
    }
    $relativepath = false;

    // then try extract file from the slasharguments
    if (stripos($_SERVER['SERVER_SOFTWARE'], 'iis') !== false) {
        // NOTE: ISS tends to convert all file paths to single byte DOS encoding,
        //       we can not use other methods because they break unicode chars,
        //       the only way is to use URL rewriting
        if (isset($_SERVER['PATH_INFO']) and $_SERVER['PATH_INFO'] !== '') {
            // check that PATH_INFO works == must not contain the script name
            if (strpos($_SERVER['PATH_INFO'], $SCRIPT) === false) {
                $relativepath = clean_param(urldecode($_SERVER['PATH_INFO']), PARAM_PATH);
            }
        }
    } else {
        // all other apache-like servers depend on PATH_INFO
        if (isset($_SERVER['PATH_INFO'])) {
            if (isset($_SERVER['SCRIPT_NAME']) and strpos($_SERVER['PATH_INFO'], $_SERVER['SCRIPT_NAME']) === 0) {
                $relativepath = substr($_SERVER['PATH_INFO'], strlen($_SERVER['SCRIPT_NAME']));
            } else {
                $relativepath = $_SERVER['PATH_INFO'];
            }
            $relativepath = clean_param($relativepath, PARAM_PATH);
        }
    }


    return $relativepath;
}

/**
 * Just returns an array of text formats suitable for a popup menu
 *
 * @uses FORMAT_MOODLE
 * @uses FORMAT_HTML
 * @uses FORMAT_PLAIN
 * @uses FORMAT_MARKDOWN
 * @return array
 */
function format_text_menu() {
    return array (FORMAT_MOODLE => get_string('formattext'),
                  FORMAT_HTML   => get_string('formathtml'),
                  FORMAT_PLAIN  => get_string('formatplain'),
                  FORMAT_MARKDOWN  => get_string('formatmarkdown'));
}

/**
 * Given text in a variety of format codings, this function returns
 * the text as safe HTML.
 *
 * This function should mainly be used for long strings like posts,
 * answers, glossary items etc. For short strings @see format_string().
 *
 * <pre>
 * Options:
 *      trusted     :   If true the string won't be cleaned. Default false required noclean=true.
 *      noclean     :   If true the string won't be cleaned. Default false required trusted=true.
 *      nocache     :   If true the strign will not be cached and will be formatted every call. Default false.
 *      filter      :   If true the string will be run through applicable filters as well. Default true.
 *      para        :   If true then the returned string will be wrapped in div tags. Default true.
 *      newlines    :   If true then lines newline breaks will be converted to HTML newline breaks. Default true.
 *      context     :   The context that will be used for filtering.
 *      overflowdiv :   If set to true the formatted text will be encased in a div
 *                      with the class no-overflow before being returned. Default false.
 *      allowid     :   If true then id attributes will not be removed, even when
 *                      using htmlpurifier. Default false.
 * </pre>
 *
 * @todo Finish documenting this function
 *
 * @staticvar array $croncache
 * @param string $text The text to be formatted. This is raw text originally from user input.
 * @param int $format Identifier of the text format to be used
 *            [FORMAT_MOODLE, FORMAT_HTML, FORMAT_PLAIN, FORMAT_MARKDOWN]
 * @param object/array $options text formatting options
 * @param int $courseid_do_not_use deprecated course id, use context option instead
 * @return string
 */
function format_text($text, $format = FORMAT_MOODLE, $options = NULL, $courseid_do_not_use = NULL) {
    global $CFG, $COURSE, $DB, $PAGE;
    static $croncache = array();

    if ($text === '' || is_null($text)) {
        return ''; // no need to do any filters and cleaning
    }

    $options = (array)$options; // detach object, we can not modify it

    if (!isset($options['trusted'])) {
        $options['trusted'] = false;
    }
    if (!isset($options['noclean'])) {
        if ($options['trusted'] and trusttext_active()) {
            // no cleaning if text trusted and noclean not specified
            $options['noclean'] = true;
        } else {
            $options['noclean'] = false;
        }
    }
    if (!isset($options['nocache'])) {
        $options['nocache'] = false;
    }
    if (!isset($options['filter'])) {
        $options['filter'] = true;
    }
    if (!isset($options['para'])) {
        $options['para'] = true;
    }
    if (!isset($options['newlines'])) {
        $options['newlines'] = true;
    }
    if (!isset($options['overflowdiv'])) {
        $options['overflowdiv'] = false;
    }

    // Calculate best context
    if (empty($CFG->version) or $CFG->version < 2010072800 or during_initial_install()) {
        // do not filter anything during installation or before upgrade completes
        $context = null;

    } else if (isset($options['context'])) { // first by explicit passed context option
        if (is_object($options['context'])) {
            $context = $options['context'];
        } else {
            $context = context::instance_by_id($options['context']);
        }
    } else if ($courseid_do_not_use) {
        // legacy courseid
        $context = context_course::instance($courseid_do_not_use);
    } else {
        // fallback to $PAGE->context this may be problematic in CLI and other non-standard pages :-(
        $context = $PAGE->context;
    }

    if (!$context) {
        // either install/upgrade or something has gone really wrong because context does not exist (yet?)
        $options['nocache'] = true;
        $options['filter']  = false;
    }

    if ($options['filter']) {
        $filtermanager = filter_manager::instance();
        $filtermanager->setup_page_for_filters($PAGE, $context); // Setup global stuff filters may have.
    } else {
        $filtermanager = new null_filter_manager();
    }

    if (!empty($CFG->cachetext) and empty($options['nocache'])) {
        $hashstr = $text.'-'.$filtermanager->text_filtering_hash($context).'-'.$context->id.'-'.current_language().'-'.
                (int)$format.(int)$options['trusted'].(int)$options['noclean'].
                (int)$options['para'].(int)$options['newlines'];

        $time = time() - $CFG->cachetext;
        $md5key = md5($hashstr);
        if (CLI_SCRIPT) {
            if (isset($croncache[$md5key])) {
                return $croncache[$md5key];
            }
        }

        if ($oldcacheitem = $DB->get_record('cache_text', array('md5key'=>$md5key), '*', IGNORE_MULTIPLE)) {
            if ($oldcacheitem->timemodified >= $time) {
                if (CLI_SCRIPT) {
                    if (count($croncache) > 150) {
                        reset($croncache);
                        $key = key($croncache);
                        unset($croncache[$key]);
                    }
                    $croncache[$md5key] = $oldcacheitem->formattedtext;
                }
                return $oldcacheitem->formattedtext;
            }
        }
    }

    switch ($format) {
        case FORMAT_HTML:
            if (!$options['noclean']) {
                $text = clean_text($text, FORMAT_HTML, $options);
            }
            $text = $filtermanager->filter_text($text, $context, array('originalformat' => FORMAT_HTML, 'noclean' => $options['noclean']));
            break;

        case FORMAT_PLAIN:
            $text = s($text); // cleans dangerous JS
            $text = rebuildnolinktag($text);
            $text = str_replace('  ', '&nbsp; ', $text);
            $text = nl2br($text);
            break;

        case FORMAT_WIKI:
            // this format is deprecated
            $text = '<p>NOTICE: Wiki-like formatting has been removed from Moodle.  You should not be seeing
                     this message as all texts should have been converted to Markdown format instead.
                     Please post a bug report to http://moodle.org/bugs with information about where you
                     saw this message.</p>'.s($text);
            break;

        case FORMAT_MARKDOWN:
            $text = markdown_to_html($text);
            if (!$options['noclean']) {
                $text = clean_text($text, FORMAT_HTML, $options);
            }
            $text = $filtermanager->filter_text($text, $context, array('originalformat' => FORMAT_MARKDOWN, 'noclean' => $options['noclean']));
            break;

        default:  // FORMAT_MOODLE or anything else
            $text = text_to_html($text, null, $options['para'], $options['newlines']);
            if (!$options['noclean']) {
                $text = clean_text($text, FORMAT_HTML, $options);
            }
            $text = $filtermanager->filter_text($text, $context, array('originalformat' => $format, 'noclean' => $options['noclean']));
            break;
    }
    if ($options['filter']) {
        // at this point there should not be any draftfile links any more,
        // this happens when developers forget to post process the text.
        // The only potential problem is that somebody might try to format
        // the text before storing into database which would be itself big bug.
        $text = str_replace("\"$CFG->httpswwwroot/draftfile.php", "\"$CFG->httpswwwroot/brokenfile.php#", $text);
    }

    // Warn people that we have removed this old mechanism, just in case they
    // were stupid enough to rely on it.
    if (isset($CFG->currenttextiscacheable)) {
        debugging('Once upon a time, Moodle had a truly evil use of global variables ' .
                'called $CFG->currenttextiscacheable. The good news is that this no ' .
                'longer exists. The bad news is that you seem to be using a filter that '.
                'relies on it. Please seek out and destroy that filter code.', DEBUG_DEVELOPER);
    }

    if (!empty($options['overflowdiv'])) {
        $text = html_writer::tag('div', $text, array('class'=>'no-overflow'));
    }

    if (empty($options['nocache']) and !empty($CFG->cachetext)) {
        if (CLI_SCRIPT) {
            // special static cron cache - no need to store it in db if its not already there
            if (count($croncache) > 150) {
                reset($croncache);
                $key = key($croncache);
                unset($croncache[$key]);
            }
            $croncache[$md5key] = $text;
            return $text;
        }

        $newcacheitem = new stdClass();
        $newcacheitem->md5key = $md5key;
        $newcacheitem->formattedtext = $text;
        $newcacheitem->timemodified = time();
        if ($oldcacheitem) {                               // See bug 4677 for discussion
            $newcacheitem->id = $oldcacheitem->id;
            try {
                $DB->update_record('cache_text', $newcacheitem);   // Update existing record in the cache table
            } catch (dml_exception $e) {
               // It's unlikely that the cron cache cleaner could have
               // deleted this entry in the meantime, as it allows
               // some extra time to cover these cases.
            }
        } else {
            try {
                $DB->insert_record('cache_text', $newcacheitem);   // Insert a new record in the cache table
            } catch (dml_exception $e) {
               // Again, it's possible that another user has caused this
               // record to be created already in the time that it took
               // to traverse this function.  That's OK too, as the
               // call above handles duplicate entries, and eventually
               // the cron cleaner will delete them.
            }
        }
    }

    return $text;
}

/**
 * Resets all data related to filters, called during upgrade or when filter settings change.
 *
 * @param bool $phpunitreset true means called from our PHPUnit integration test reset
 * @return void
 */
function reset_text_filters_cache($phpunitreset = false) {
    global $CFG, $DB;

    if (!$phpunitreset) {
        $DB->delete_records('cache_text');
    }

    $purifdir = $CFG->cachedir.'/htmlpurifier';
    remove_dir($purifdir, true);
}

/**
 * Given a simple string, this function returns the string
 * processed by enabled string filters if $CFG->filterall is enabled
 *
 * This function should be used to print short strings (non html) that
 * need filter processing e.g. activity titles, post subjects,
 * glossary concepts.
 *
 * @staticvar bool $strcache
 * @param string $string The string to be filtered. Should be plain text, expect
 * possibly for multilang tags.
 * @param boolean $striplinks To strip any link in the result text.
                              Moodle 1.8 default changed from false to true! MDL-8713
 * @param array $options options array/object or courseid
 * @return string
 */
function format_string($string, $striplinks = true, $options = NULL) {
    global $CFG, $COURSE, $PAGE;

    //We'll use a in-memory cache here to speed up repeated strings
    static $strcache = false;

    if (empty($CFG->version) or $CFG->version < 2010072800 or during_initial_install()) {
        // do not filter anything during installation or before upgrade completes
        return $string = strip_tags($string);
    }

    if ($strcache === false or count($strcache) > 2000) { // this number might need some tuning to limit memory usage in cron
        $strcache = array();
    }

    if (is_numeric($options)) {
        // legacy courseid usage
        $options  = array('context'=>context_course::instance($options));
    } else {
        $options = (array)$options; // detach object, we can not modify it
    }

    if (empty($options['context'])) {
        // fallback to $PAGE->context this may be problematic in CLI and other non-standard pages :-(
        $options['context'] = $PAGE->context;
    } else if (is_numeric($options['context'])) {
        $options['context'] = context::instance_by_id($options['context']);
    }

    if (!$options['context']) {
        // we did not find any context? weird
        return $string = strip_tags($string);
    }

    //Calculate md5
    $md5 = md5($string.'<+>'.$striplinks.'<+>'.$options['context']->id.'<+>'.current_language());

    //Fetch from cache if possible
    if (isset($strcache[$md5])) {
        return $strcache[$md5];
    }

    // First replace all ampersands not followed by html entity code
    // Regular expression moved to its own method for easier unit testing
    $string = replace_ampersands_not_followed_by_entity($string);

    if (!empty($CFG->filterall)) {
        $filtermanager = filter_manager::instance();
        $filtermanager->setup_page_for_filters($PAGE, $options['context']); // Setup global stuff filters may have.
        $string = $filtermanager->filter_string($string, $options['context']);
    }

    // If the site requires it, strip ALL tags from this string
    if (!empty($CFG->formatstringstriptags)) {
        $string = str_replace(array('<', '>'), array('&lt;', '&gt;'), strip_tags($string));

    } else {
        // Otherwise strip just links if that is required (default)
        if ($striplinks) {  //strip links in string
            $string = strip_links($string);
        }
        $string = clean_text($string);
    }

    //Store to cache
    $strcache[$md5] = $string;

    return $string;
}

/**
 * Given a string, performs a negative lookahead looking for any ampersand character
 * that is not followed by a proper HTML entity. If any is found, it is replaced
 * by &amp;. The string is then returned.
 *
 * @param string $string
 * @return string
 */
function replace_ampersands_not_followed_by_entity($string) {
    return preg_replace("/\&(?![a-zA-Z0-9#]{1,8};)/", "&amp;", $string);
}

/**
 * Given a string, replaces all <a>.*</a> by .* and returns the string.
 *
 * @param string $string
 * @return string
 */
function strip_links($string) {
    return preg_replace('/(<a\s[^>]+?>)(.+?)(<\/a>)/is','$2',$string);
}

/**
 * This expression turns links into something nice in a text format. (Russell Jungwirth)
 *
 * @param string $string
 * @return string
 */
function wikify_links($string) {
    return preg_replace('~(<a [^<]*href=["|\']?([^ "\']*)["|\']?[^>]*>([^<]*)</a>)~i','$3 [ $2 ]', $string);
}

/**
 * Given text in a variety of format codings, this function returns
 * the text as plain text suitable for plain email.
 *
 * @uses FORMAT_MOODLE
 * @uses FORMAT_HTML
 * @uses FORMAT_PLAIN
 * @uses FORMAT_WIKI
 * @uses FORMAT_MARKDOWN
 * @param string $text The text to be formatted. This is raw text originally from user input.
 * @param int $format Identifier of the text format to be used
 *            [FORMAT_MOODLE, FORMAT_HTML, FORMAT_PLAIN, FORMAT_WIKI, FORMAT_MARKDOWN]
 * @return string
 */
function format_text_email($text, $format) {

    switch ($format) {

        case FORMAT_PLAIN:
            return $text;
            break;

        case FORMAT_WIKI:
            // there should not be any of these any more!
            $text = wikify_links($text);
            return textlib::entities_to_utf8(strip_tags($text), true);
            break;

        case FORMAT_HTML:
            return html_to_text($text);
            break;

        case FORMAT_MOODLE:
        case FORMAT_MARKDOWN:
        default:
            $text = wikify_links($text);
            return textlib::entities_to_utf8(strip_tags($text), true);
            break;
    }
}

/**
 * Formats activity intro text
 *
 * @global object
 * @uses CONTEXT_MODULE
 * @param string $module name of module
 * @param object $activity instance of activity
 * @param int $cmid course module id
 * @param bool $filter filter resulting html text
 * @return text
 */
function format_module_intro($module, $activity, $cmid, $filter=true) {
    global $CFG;
    require_once("$CFG->libdir/filelib.php");
    $context = context_module::instance($cmid);
    $options = array('noclean'=>true, 'para'=>false, 'filter'=>$filter, 'context'=>$context, 'overflowdiv'=>true);
    $intro = file_rewrite_pluginfile_urls($activity->intro, 'pluginfile.php', $context->id, 'mod_'.$module, 'intro', null);
    return trim(format_text($intro, $activity->introformat, $options, null));
}

/**
 * Legacy function, used for cleaning of old forum and glossary text only.
 *
 * @global object
 * @param string $text text that may contain legacy TRUSTTEXT marker
 * @return text without legacy TRUSTTEXT marker
 */
function trusttext_strip($text) {
    while (true) { //removing nested T