filter/urltolink/filter.php

   1 <?php
   2
   3 // This file is part of Moodle - http://moodle.org/
   4 //
   5 // Moodle is free software: you can redistribute it and/or modify
   6 // it under the terms of the GNU General Public License as published by
   7 // the Free Software Foundation, either version 3 of the License, or
   8 // (at your option) any later version.
   9 //
  10 // Moodle is distributed in the hope that it will be useful,
  11 // but WITHOUT ANY WARRANTY; without even the implied warranty of
  12 // MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  13 // GNU General Public License for more details.
  14 //
  15 // You should have received a copy of the GNU General Public License
  16 // along with Moodle.  If not, see <http://www.gnu.org/licenses/>.
  17
  18 /**
  19  * Filter converting URLs in the text to HTML links
  20  *
  21  * @package    filter
  22  * @subpackage urltolink
  23  * @copyright  2010 David Mudrak <david@moodle.com>
  24  * @license    http://www.gnu.org/copyleft/gpl.html GNU GPL v3 or later
  25  */
  26
  27 defined('MOODLE_INTERNAL') || die();
  28
  29 class filter_urltolink extends moodle_text_filter {
  30
  31     /**
  32      * @var array global configuration for this filter
  33      *
  34      * This might be eventually moved into parent class if we found it
  35      * useful for other filters, too.
  36      */
  37     protected static $globalconfig;
  38
  39     /**
  40      * Apply the filter to the text
  41      *
  42      * @see filter_manager::apply_filter_chain()
  43      * @param string $text to be processed by the text
  44      * @param array $options filter options
  45      * @return string text after processing
  46      */
  47     public function filter($text, array $options = array()) {
  48         if (!isset($options['originalformat'])) {
  49             // if the format is not specified, we are probably called by {@see format_string()}
  50             // in that case, it would be dangerous to replace URL with the link because it could
  51             // be stripped. therefore, we do nothing
  52             return $text;
  53         }
  54         if (in_array($options['originalformat'], explode(',', $this->get_global_config('formats')))) {
  55             $this->convert_urls_into_links($text);
  56         }
  57         return $text;
  58     }
  59
  60     ////////////////////////////////////////////////////////////////////////////
  61     // internal implementation starts here
  62     ////////////////////////////////////////////////////////////////////////////
  63
  64     /**
  65      * Returns the global filter setting
  66      *
  67      * If the $name is provided, returns single value. Otherwise returns all
  68      * global settings in object. Returns null if the named setting is not
  69      * found.
  70      *
  71      * @param mixed $name optional config variable name, defaults to null for all
  72      * @return string|object|null
  73      */
  74     protected function get_global_config($name=null) {
  75         $this->load_global_config();
  76         if (is_null($name)) {
  77             return self::$globalconfig;
  78
  79         } elseif (array_key_exists($name, self::$globalconfig)) {
  80             return self::$globalconfig->{$name};
  81
  82         } else {
  83             return null;
  84         }
  85     }
  86
  87     /**
  88      * Makes sure that the global config is loaded in $this->globalconfig
  89      *
  90      * @return void
  91      */
  92     protected function load_global_config() {
  93         if (is_null(self::$globalconfig)) {
  94             self::$globalconfig = get_config('filter_urltolink');
  95         }
  96     }
  97
  98     /**
  99      * Given some text this function converts any URLs it finds into HTML links
 100      *
 101      * @param string $text Passed in by reference. The string to be searched for urls.
 102      */
 103     protected function convert_urls_into_links(&$text) {
 104         //I've added img tags to this list of tags to ignore.
 105         //See MDL-21168 for more info. A better way to ignore tags whether or not
 106         //they are escaped partially or completely would be desirable. For example:
 107         //<a href="blah">
 108         //&lt;a href="blah"&gt;
 109         //&lt;a href="blah">
 110         $filterignoretagsopen  = array('<a\s[^>]+?>');
 111         $filterignoretagsclose = array('</a>');
 112         filter_save_ignore_tags($text,$filterignoretagsopen,$filterignoretagsclose,$ignoretags);
 113
 114         // Check if we support unicode modifiers in regular expressions. Cache it.
 115         // TODO: this check should be a environment requirement in Moodle 2.0, as far as unicode
 116         // chars are going to arrive to URLs officially really soon (2010?)
 117         // Original RFC regex from: http://www.bytemycode.com/snippets/snippet/796/
 118         // Various ideas from: http://alanstorm.com/url_regex_explained
 119         // Unicode check, negative assertion and other bits from Moodle.
 120         static $unicoderegexp;
 121         if (!isset($unicoderegexp)) {
 122             $unicoderegexp = @preg_match('/\pL/u', 'a'); // This will fail silently, returning false,
 123         }
 124
 125         // TODO MDL-21296 - use of unicode modifiers may cause a timeout
 126         $urlstart = '(?:http(s)?://|(?<!://)(www\.))';
 127         $domainsegment = '(?:[\pLl0-9][\pLl0-9-]*[\pLl0-9]|[\pLl0-9])';
 128         $numericip = '(?:(?:[0-9]{1,3}\.){3}[0-9]{1,3})';
 129         $port = '(?::\d*)';
 130         $pathchar = '(?:[\pL0-9\.!$&\'\(\)*+,;=_~:@-]|%[a-f0-9]{2})';
 131         $path = "(?:/$pathchar*)*";
 132         $querystring = '(?:\?(?:[\pL0-9\.!$&\'\(\)*+,;=_~:@/?-]|%[a-fA-F0-9]{2})*)';
 133         $fragment = '(?:\#(?:[\pL0-9\.!$&\'\(\)*+,;=_~:@/?-]|%[a-fA-F0-9]{2})*)';
 134
 135         // Lookbehind assertions.
 136         // Is not HTML attribute or CSS URL property. Unfortunately legit text like "url(http://...)" will not be a link.
 137         $lookbehindstart = "(?<!=[\"']|\burl\([\"' ]|\burl\()";
 138         $lookbehindend = "(?<![]),.;])";
 139
 140         $regex = "$lookbehindstart$urlstart((?:$domainsegment\.)+$domainsegment|$numericip)" .
 141                 "($port?$path$querystring?$fragment?)$lookbehindend";
 142         if ($unicoderegexp) {
 143             $regex = '#' . $regex . '#ui';
 144         } else {
 145             $regex = '#' . preg_replace(array('\pLl', '\PL'), 'a-z', $regex) . '#i';
 146         }
 147
 148         $text = preg_replace($regex, '<a href="http$1://$2$3$4" class="_blanktarget">$0</a>', $text);
 149
 150         if (!empty($ignoretags)) {
 151             $ignoretags = array_reverse($ignoretags); /// Reversed so "progressive" str_replace() will solve some nesting problems.
 152             $text = str_replace(array_keys($ignoretags),$ignoretags,$text);
 153         }
 154
 155         if ($this->get_global_config('embedimages')) {
 156             // now try to inject the images, this code was originally in the mediapluing filter
 157             // this may be useful only if somebody relies on the fact the links in FORMAT_MOODLE get converted
 158             // to URLs which in turn change to real images
 159             $search = '/<a href="([^"]+\.(jpg|png|gif))" class="_blanktarget">([^>]*)<\/a>/is';
 160             $text = preg_replace_callback($search, 'filter_urltolink_img_callback', $text);
 161         }
 162     }
 163 }
 164
 165
 166 /**
 167  * Change links to images into embedded images.
 168  *
 169  * This plugin is intended for automatic conversion of image URLs when FORMAT_MOODLE used.
 170  *
 171  * @param  $link
 172  * @return string
 173  */
 174 function filter_urltolink_img_callback($link) {
 175     if ($link[1] !== $link[3]) {
 176         // this is not a link created by this filter, because the url does not match the text
 177         return $link[0];
 178     }
 179     return '<img class="filter_urltolink_image" alt="" src="'.$link[1].'" />';
 180 }
 181