backup/moodle2/backup_xml_transformer.class.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  * Defines backup_xml_transformer class
  20  *
  21  * @package     core_backup
  22  * @subpackage  moodle2
  23  * @category    backup
  24  * @copyright   2010 onwards Eloy Lafuente (stronk7) {@link http://stronk7.com}
  25  * @license     http://www.gnu.org/copyleft/gpl.html GNU GPL v3 or later
  26  */
  27
  28 defined('MOODLE_INTERNAL') || die();
  29
  30 // Cache for storing link encoders, so that we don't need to call
  31 // register_link_encoders each time backup_xml_transformer is constructed
  32 // TODO MDL-25290 replace global with MUC code.
  33 global $LINKS_ENCODERS_CACHE;
  34
  35 $LINKS_ENCODERS_CACHE = array();
  36
  37 /**
  38  * Class implementing the @xml_contenttransformed logic to be applied in moodle2 backups
  39  *
  40  * TODO: Finish phpdocs
  41  */
  42 class backup_xml_transformer extends xml_contenttransformer {
  43
  44     private $absolute_links_encoders; // array of static methods to be called in order to
  45                                       // perform the encoding of absolute links to all the
  46                                       // contents sent to xml
  47     private $courseid;                // courseid this content belongs to
  48     private $unicoderegexp;           // to know if the site supports unicode regexp
  49
  50     public function __construct($courseid) {
  51         $this->absolute_links_encoders = array();
  52         $this->courseid = $courseid;
  53         // Check if we support unicode modifiers in regular expressions
  54         $this->unicoderegexp = @preg_match('/\pL/u', 'a'); // This will fail silently, returning false,
  55                                                            // if regexp libraries don't support unicode
  56         // Register all the available content link encoders
  57         $this->absolute_links_encoders = $this->register_link_encoders();
  58     }
  59
  60     public function process($content) {
  61
  62         // Array or object, debug and try our best recursively, shouldn't happen but...
  63         if (is_array($content)) {
  64             debugging('Backup XML transformer should not process arrays but plain content only', DEBUG_DEVELOPER);
  65             foreach($content as $key => $plaincontent) {
  66                 $content[$key] = $this->process($plaincontent);
  67             }
  68             return $content;
  69         } else if (is_object($content)) {
  70             debugging('Backup XML transformer should not process objects but plain content only', DEBUG_DEVELOPER);
  71             foreach((array)$content as $key => $plaincontent) {
  72                 $content[$key] = $this->process($plaincontent);
  73             }
  74             return (object)$content;
  75         }
  76
  77         if (is_null($content)) {  // Some cases we know we can skip complete processing
  78             return '$@NULL@$';
  79         } else if ($content === '') {
  80             return '';
  81         } else if (is_numeric($content)) {
  82             return $content;
  83         } else if (strlen($content) < 32) { // Impossible to have one link in 32cc
  84             return $content;                // (http://10.0.0.1/file.php/1/1.jpg, http://10.0.0.1/mod/url/view.php?id=)
  85         }
  86
  87         $content = $this->process_filephp_links($content); // Replace all calls to file.php by $@FILEPHP@$ in a normalised way
  88         $content = $this->encode_absolute_links($content); // Pass the content against all the found encoders
  89
  90         return $content;
  91     }
  92
  93     private function process_filephp_links($content) {
  94         global $CFG;
  95
  96         if (strpos($content, 'file.php') === false) { // No file.php, nothing to convert
  97             return $content;
  98         }
  99
 100         //First, we check for every call to file.php inside the course
 101         $search = array($CFG->wwwroot.'/file.php/' . $this->courseid,
 102                         $CFG->wwwroot.'/file.php?file=/' . $this->courseid,
 103                         $CFG->wwwroot.'/file.php?file=%2f' . $this->courseid,
 104                         $CFG->wwwroot.'/file.php?file=%2F' . $this->courseid);
 105         $replace = array('$@FILEPHP@$', '$@FILEPHP@$', '$@FILEPHP@$', '$@FILEPHP@$');
 106         $content = str_replace($search, $replace, $content);
 107
 108         // Now we look for any '$@FILEPHP@$' URLs, replacing:
 109         //     - slashes and %2F by $@SLASH@$
 110         //     - &forcedownload=1 &amp;forcedownload=1 and ?forcedownload=1 by $@FORCEDOWNLOAD@$
 111         // This way, backup contents will be neutral and independent of slasharguments configuration. MDL-18799
 112         // Based in $this->unicoderegexp, decide the regular expression to use
 113         if ($this->unicoderegexp) { //We can use unicode modifiers
 114             $search = '/(\$@FILEPHP@\$)((?:(?:\/|%2f|%2F))(?:(?:\([-;:@#&=\pL0-9\$~_.+!*\',]*?\))|[-;:@#&=\pL0-9\$~_.+!*\',]|%[a-fA-F0-9]{2}|\/)*)?(\?(?:(?:(?:\([-;:@#&=\pL0-9\$~_.+!*\',]*?\))|[-;:@#&=?\pL0-9\$~_.+!*\',]|%[a-fA-F0-9]{2}|\/)*))?(?<![,.;])/u';
 115         } else { //We cannot ue unicode modifiers
 116             $search = '/(\$@FILEPHP@\$)((?:(?:\/|%2f|%2F))(?:(?:\([-;:@#&=a-zA-Z0-9\$~_.+!*\',]*?\))|[-;:@#&=a-zA-Z0-9\$~_.+!*\',]|%[a-fA-F0-9]{2}|\/)*)?(\?(?:(?:(?:\([-;:@#&=a-zA-Z0-9\$~_.+!*\',]*?\))|[-;:@#&=?a-zA-Z0-9\$~_.+!*\',]|%[a-fA-F0-9]{2}|\/)*))?(?<![,.;])/';
 117         }
 118         $content = preg_replace_callback($search, array('backup_xml_transformer', 'process_filephp_uses'), $content);
 119
 120         return $content;
 121     }
 122
 123     private function encode_absolute_links($content) {
 124         foreach ($this->absolute_links_encoders as $classname => $methodname) {
 125             $content = call_user_func(array($classname, $methodname), $content);
 126         }
 127         return $content;
 128     }
 129
 130     static private function process_filephp_uses($matches) {
 131
 132         // Replace slashes (plain and encoded) and forcedownload=1 parameter
 133         $search = array('/', '%2f', '%2F', '?forcedownload=1', '&forcedownload=1', '&amp;forcedownload=1');
 134         $replace = array('$@SLASH@$', '$@SLASH@$', '$@SLASH@$', '$@FORCEDOWNLOAD@$', '$@FORCEDOWNLOAD@$', '$@FORCEDOWNLOAD@$');
 135
 136         $result = $matches[1] . (isset($matches[2]) ? str_replace($search, $replace, $matches[2]) : '') . (isset($matches[3]) ? str_replace($search, $replace, $matches[3]) : '');
 137
 138         return $result;
 139     }
 140
 141     /**
 142      * Register all available content link encoders
 143      *
 144      * @return array encoder
 145      * @todo MDL-25290 replace LINKS_ENCODERS_CACHE global with MUC code
 146      */
 147     private function register_link_encoders() {
 148         global $LINKS_ENCODERS_CACHE;
 149         // If encoder is linked, then return cached encoder.
 150         if (!empty($LINKS_ENCODERS_CACHE)) {
 151             return $LINKS_ENCODERS_CACHE;
 152         }
 153
 154         $encoders = array();
 155
 156         // Add the course encoder
 157         $encoders['backup_course_task'] = 'encode_content_links';
 158
 159         // Add the module ones. Each module supporting moodle2 backups MUST have it
 160         $mods = core_component::get_plugin_list('mod');
 161         foreach ($mods as $mod => $moddir) {
 162             if (plugin_supports('mod', $mod, FEATURE_BACKUP_MOODLE2) && class_exists('backup_' . $mod . '_activity_task')) {
 163                 $encoders['backup_' . $mod . '_activity_task'] = 'encode_content_links';
 164             }
 165         }
 166
 167         // Add the block encoders
 168         $blocks = core_component::get_plugin_list('block');
 169         foreach ($blocks as $block => $blockdir) {
 170             if (class_exists('backup_' . $block . '_block_task')) {
 171                 $encoders['backup_' . $block . '_block_task'] = 'encode_content_links';
 172             }
 173         }
 174
 175         // Add the course format encodes
 176         // TODO: Same than blocks, need to know how courseformats are going to handle backup
 177         //       (1.9 was based in backuplib function, see code)
 178
 179         // Add local encodes
 180         // TODO: Any interest? 1.9 never had that.
 181
 182         $LINKS_ENCODERS_CACHE = $encoders;
 183         return $encoders;
 184     }
 185 }