array( array('update_fetch_data_batch', array()), ), 'finished' => 'update_fetch_data_finished', 'title' => t('Checking available update data'), 'progress_message' => t('Trying to check available update data ...'), 'error_message' => t('Error checking available update data.'), 'file' => drupal_get_path('module', 'update') . '/update.fetch.inc', ); batch_set($batch); batch_process('admin/reports/updates'); } /** * Implements callback_batch_operation(). * * Processes a step in batch for fetching available update data. * * @param $context * Reference to an array used for Batch API storage. */ function update_fetch_data_batch(&$context) { $queue = DrupalQueue::get('update_fetch_tasks'); if (empty($context['sandbox']['max'])) { $context['finished'] = 0; $context['sandbox']['max'] = $queue->numberOfItems(); $context['sandbox']['progress'] = 0; $context['message'] = t('Checking available update data ...'); $context['results']['updated'] = 0; $context['results']['failures'] = 0; $context['results']['processed'] = 0; } // Grab another item from the fetch queue. for ($i = 0; $i < 5; $i++) { if ($item = $queue->claimItem()) { if (_update_process_fetch_task($item->data)) { $context['results']['updated']++; $context['message'] = t('Checked available update data for %title.', array('%title' => $item->data['info']['name'])); } else { $context['message'] = t('Failed to check available update data for %title.', array('%title' => $item->data['info']['name'])); $context['results']['failures']++; } $context['sandbox']['progress']++; $context['results']['processed']++; $context['finished'] = $context['sandbox']['progress'] / $context['sandbox']['max']; $queue->deleteItem($item); } else { // If the queue is currently empty, we're done. It's possible that // another thread might have added new fetch tasks while we were // processing this batch. In that case, the usual 'finished' math could // get confused, since we'd end up processing more tasks that we thought // we had when we started and initialized 'max' with numberOfItems(). By // forcing 'finished' to be exactly 1 here, we ensure that batch // processing is terminated. $context['finished'] = 1; return; } } } /** * Implements callback_batch_finished(). * * Performs actions when all fetch tasks have been completed. * * @param $success * TRUE if the batch operation was successful; FALSE if there were errors. * @param $results * An associative array of results from the batch operation, including the key * 'updated' which holds the total number of projects we fetched available * update data for. */ function update_fetch_data_finished($success, $results) { if ($success) { if (!empty($results)) { if (!empty($results['updated'])) { drupal_set_message(format_plural($results['updated'], 'Checked available update data for one project.', 'Checked available update data for @count projects.')); } if (!empty($results['failures'])) { drupal_set_message(format_plural($results['failures'], 'Failed to get available update data for one project.', 'Failed to get available update data for @count projects.'), 'error'); } } } else { drupal_set_message(t('An error occurred trying to get available update data.'), 'error'); } } /** * Attempts to drain the queue of tasks for release history data to fetch. */ function _update_fetch_data() { $queue = DrupalQueue::get('update_fetch_tasks'); $end = time() + variable_get('update_max_fetch_time', UPDATE_MAX_FETCH_TIME); while (time() < $end && ($item = $queue->claimItem())) { _update_process_fetch_task($item->data); $queue->deleteItem($item); } } /** * Processes a task to fetch available update data for a single project. * * Once the release history XML data is downloaded, it is parsed and saved into * the {cache_update} table in an entry just for that project. * * @param $project * Associative array of information about the project to fetch data for. * * @return * TRUE if we fetched parsable XML, otherwise FALSE. */ function _update_process_fetch_task($project) { global $base_url; $fail = &drupal_static(__FUNCTION__, array()); // This can be in the middle of a long-running batch, so REQUEST_TIME won't // necessarily be valid. $now = time(); if (empty($fail)) { // If we have valid data about release history XML servers that we have // failed to fetch from on previous attempts, load that from the cache. if (($cache = _update_cache_get('fetch_failures')) && ($cache->expire > $now)) { $fail = $cache->data; } } $max_fetch_attempts = variable_get('update_max_fetch_attempts', UPDATE_MAX_FETCH_ATTEMPTS); $success = FALSE; $available = array(); $site_key = drupal_hmac_base64($base_url, drupal_get_private_key()); $url = _update_build_fetch_url($project, $site_key); $fetch_url_base = _update_get_fetch_url_base($project); $project_name = $project['name']; if (empty($fail[$fetch_url_base]) || $fail[$fetch_url_base] < $max_fetch_attempts) { $xml = drupal_http_request($url); if (isset($xml->error)) { watchdog('update', 'Error %errorcode (%message) occurred when trying to fetch available update data for the project %project.', array('%errorcode' => $xml->code, '%message' => $xml->error, '%project' => $project_name), WATCHDOG_ERROR); } if (!isset($xml->error) && isset($xml->data)) { $data = $xml->data; } } if (!empty($data)) { $available = update_parse_xml($data); // @todo: Purge release data we don't need (http://drupal.org/node/238950). if (!empty($available)) { // Only if we fetched and parsed something sane do we return success. $success = TRUE; } } else { $available['project_status'] = 'not-fetched'; if (empty($fail[$fetch_url_base])) { $fail[$fetch_url_base] = 1; } else { $fail[$fetch_url_base]++; } } $frequency = variable_get('update_check_frequency', 1); $cid = 'available_releases::' . $project_name; _update_cache_set($cid, $available, $now + (60 * 60 * 24 * $frequency)); // Stash the $fail data back in the DB for the next 5 minutes. _update_cache_set('fetch_failures', $fail, $now + (60 * 5)); // Whether this worked or not, we did just (try to) check for updates. variable_set('update_last_check', $now); // Now that we processed the fetch task for this project, clear out the // record in {cache_update} for this task so we're willing to fetch again. _update_cache_clear('fetch_task::' . $project_name); return $success; } /** * Clears out all the cached available update data and initiates re-fetching. */ function _update_refresh() { module_load_include('inc', 'update', 'update.compare'); // Since we're fetching new available update data, we want to clear // our cache of both the projects we care about, and the current update // status of the site. We do *not* want to clear the cache of available // releases just yet, since that data (even if it's stale) can be useful // during update_get_projects(); for example, to modules that implement // hook_system_info_alter() such as cvs_deploy. _update_cache_clear('update_project_projects'); _update_cache_clear('update_project_data'); $projects = update_get_projects(); // Now that we have the list of projects, we should also clear our cache of // available release data, since even if we fail to fetch new data, we need // to clear out the stale data at this point. _update_cache_clear('available_releases::', TRUE); foreach ($projects as $key => $project) { update_create_fetch_task($project); } } /** * Adds a task to the queue for fetching release history data for a project. * * We only create a new fetch task if there's no task already in the queue for * this particular project (based on 'fetch_task::' entries in the * {cache_update} table). * * @param $project * Associative array of information about a project as created by * update_get_projects(), including keys such as 'name' (short name), and the * 'info' array with data from a .info file for the project. * * @see update_get_projects() * @see update_get_available() * @see update_refresh() * @see update_fetch_data() * @see _update_process_fetch_task() */ function _update_create_fetch_task($project) { $fetch_tasks = &drupal_static(__FUNCTION__, array()); if (empty($fetch_tasks)) { $fetch_tasks = _update_get_cache_multiple('fetch_task'); } $cid = 'fetch_task::' . $project['name']; if (empty($fetch_tasks[$cid])) { $queue = DrupalQueue::get('update_fetch_tasks'); $queue->createItem($project); // Due to race conditions, it is possible that another process already // inserted a row into the {cache_update} table and the following query will // throw an exception. // @todo: Remove the need for the manual check by relying on a queue that // enforces unique items. try { db_insert('cache_update') ->fields(array( 'cid' => $cid, 'created' => REQUEST_TIME, )) ->execute(); } catch (Exception $e) { // The exception can be ignored safely. } $fetch_tasks[$cid] = REQUEST_TIME; } } /** * Generates the URL to fetch information about project updates. * * This figures out the right URL to use, based on the project's .info file and * the global defaults. Appends optional query arguments when the site is * configured to report usage stats. * * @param $project * The array of project information from update_get_projects(). * @param $site_key * (optional) The anonymous site key hash. Defaults to an empty string. * * @return * The URL for fetching information about updates to the specified project. * * @see update_fetch_data() * @see _update_process_fetch_task() * @see update_get_projects() */ function _update_build_fetch_url($project, $site_key = '') { $name = $project['name']; $url = _update_get_fetch_url_base($project); $url .= '/' . $name . '/' . DRUPAL_CORE_COMPATIBILITY; // Only append usage information if we have a site key and the project is // enabled. We do not want to record usage statistics for disabled projects. if (!empty($site_key) && (strpos($project['project_type'], 'disabled') === FALSE)) { // Append the site key. $url .= (strpos($url, '?') !== FALSE) ? '&' : '?'; $url .= 'site_key='; $url .= rawurlencode($site_key); // Append the version. if (!empty($project['info']['version'])) { $url .= '&version='; $url .= rawurlencode($project['info']['version']); } // Append the list of modules or themes enabled. $list = array_keys($project['includes']); $url .= '&list='; $url .= rawurlencode(implode(',', $list)); } return $url; } /** * Returns the base of the URL to fetch available update data for a project. * * @param $project * The array of project information from update_get_projects(). * * @return * The base of the URL used for fetching available update data. This does * not include the path elements to specify a particular project, version, * site_key, etc. * * @see _update_build_fetch_url() */ function _update_get_fetch_url_base($project) { return isset($project['info']['project status url']) ? $project['info']['project status url'] : variable_get('update_fetch_url', UPDATE_DEFAULT_URL); } /** * Performs any notifications that should be done once cron fetches new data. * * This method checks the status of the site using the new data and, depending * on the configuration of the site, notifies administrators via e-mail if there * are new releases or missing security updates. * * @see update_requirements() */ function _update_cron_notify() { module_load_install('update'); $status = update_requirements('runtime'); $params = array(); $notify_all = (variable_get('update_notification_threshold', 'all') == 'all'); foreach (array('core', 'contrib') as $report_type) { $type = 'update_' . $report_type; if (isset($status[$type]['severity']) && ($status[$type]['severity'] == REQUIREMENT_ERROR || ($notify_all && $status[$type]['reason'] == UPDATE_NOT_CURRENT))) { $params[$report_type] = $status[$type]['reason']; } } if (!empty($params)) { $notify_list = variable_get('update_notify_emails', ''); if (!empty($notify_list)) { $default_language = language_default(); foreach ($notify_list as $target) { if ($target_user = user_load_by_mail($target)) { $target_language = user_preferred_language($target_user); } else { $target_language = $default_language; } $message = drupal_mail('update', 'status_notify', $target, $target_language, $params); // Track when the last mail was successfully sent to avoid sending // too many e-mails. if ($message['result']) { variable_set('update_last_email_notification', REQUEST_TIME); } } } } } /** * Parses the XML of the Drupal release history info files. * * @param $raw_xml * A raw XML string of available release data for a given project. * * @return * Array of parsed data about releases for a given project, or NULL if there * was an error parsing the string. */ function update_parse_xml($raw_xml) { try { $xml = new SimpleXMLElement($raw_xml); } catch (Exception $e) { // SimpleXMLElement::__construct produces an E_WARNING error message for // each error found in the XML data and throws an exception if errors // were detected. Catch any exception and return failure (NULL). return; } // If there is no valid project data, the XML is invalid, so return failure. if (!isset($xml->short_name)) { return; } $short_name = (string) $xml->short_name; $data = array(); foreach ($xml as $k => $v) { $data[$k] = (string) $v; } $data['releases'] = array(); if (isset($xml->releases)) { foreach ($xml->releases->children() as $release) { $version = (string) $release->version; $data['releases'][$version] = array(); foreach ($release->children() as $k => $v) { $data['releases'][$version][$k] = (string) $v; } $data['releases'][$version]['terms'] = array(); if ($release->terms) { foreach ($release->terms->children() as $term) { if (!isset($data['releases'][$version]['terms'][(string) $term->name])) { $data['releases'][$version]['terms'][(string) $term->name] = array(); } $data['releases'][$version]['terms'][(string) $term->name][] = (string) $term->value; } } } } return $data; }