diff options
Diffstat (limited to 'modules/aggregator/aggregator.module')
| -rw-r--r-- | modules/aggregator/aggregator.module | 504 |
1 files changed, 67 insertions, 437 deletions
diff --git a/modules/aggregator/aggregator.module b/modules/aggregator/aggregator.module index d2826a587..dd3a14f5a 100644 --- a/modules/aggregator/aggregator.module +++ b/modules/aggregator/aggregator.module @@ -133,7 +133,7 @@ function aggregator_menu() { $items['admin/content/aggregator/settings'] = array( 'title' => 'Settings', 'page callback' => 'drupal_get_form', - 'page arguments' => array('aggregator_admin_settings'), + 'page arguments' => array('aggregator_admin_form'), 'type' => MENU_LOCAL_TASK, 'weight' => 10, 'access arguments' => array('administer news feeds'), @@ -289,7 +289,7 @@ function aggregator_perm() { function aggregator_cron() { $result = db_query('SELECT * FROM {aggregator_feed} WHERE checked + refresh < :time', array(':time' => REQUEST_TIME)); foreach ($result as $feed) { - aggregator_refresh((array)$feed); + aggregator_refresh($feed); } } @@ -417,6 +417,7 @@ function aggregator_save_category($edit) { ->fields(array( 'title' => $edit['title'], 'description' => $edit['description'], + 'block' => 5, )) ->execute(); $op = 'insert'; @@ -503,456 +504,51 @@ function aggregator_save_feed($edit) { * Removes all items from a feed. * * @param $feed - * An associative array describing the feed to be cleared. + * An object describing the feed to be cleared. */ function aggregator_remove($feed) { - $iids = db_query('SELECT iid FROM {aggregator_item} WHERE fid = :fid', array(':fid' => $feed['fid']))->fetchCol(); - if ($iids) { - db_delete('aggregator_category_item') - ->condition('iid', $iids, 'IN') - ->execute(); - } - - db_delete('aggregator_item') - ->condition('fid', $feed['fid']) - ->execute(); + // Call hook_aggregator_remove() on all modules. + module_invoke_all('aggregator_remove', $feed); + // Reset feed. db_merge('aggregator_feed') - ->key(array('fid' => $feed['fid'])) + ->key(array('fid' => $feed->fid)) ->fields(array( 'checked' => 0, 'hash' => '', 'modified' => 0, - 'description' => $feed['description'], - 'image' => $feed['image'], + 'description' => $feed->description, + 'image' => $feed->image, )) ->execute(); - drupal_set_message(t('The news items from %site have been removed.', array('%site' => $feed['title']))); -} - -/** - * Callback function used by the XML parser. - */ -function aggregator_element_start($parser, $name, $attributes) { - global $item, $element, $tag, $items, $channel; - - switch ($name) { - case 'IMAGE': - case 'TEXTINPUT': - case 'CONTENT': - case 'SUMMARY': - case 'TAGLINE': - case 'SUBTITLE': - case 'LOGO': - case 'INFO': - $element = $name; - break; - case 'ID': - if ($element != 'ITEM') { - $element = $name; - } - case 'LINK': - if (!empty($attributes['REL']) && $attributes['REL'] == 'alternate') { - if ($element == 'ITEM') { - $items[$item]['LINK'] = $attributes['HREF']; - } - else { - $channel['LINK'] = $attributes['HREF']; - } - } - break; - case 'ITEM': - $element = $name; - $item += 1; - break; - case 'ENTRY': - $element = 'ITEM'; - $item += 1; - break; - } - - $tag = $name; -} - -/** - * Call-back function used by the XML parser. - */ -function aggregator_element_end($parser, $name) { - global $element; - - switch ($name) { - case 'IMAGE': - case 'TEXTINPUT': - case 'ITEM': - case 'ENTRY': - case 'CONTENT': - case 'INFO': - $element = ''; - break; - case 'ID': - if ($element == 'ID') { - $element = ''; - } - } -} - -/** - * Callback function used by the XML parser. - */ -function aggregator_element_data($parser, $data) { - global $channel, $element, $items, $item, $image, $tag; - $items += array($item => array()); - switch ($element) { - case 'ITEM': - $items[$item] += array($tag => ''); - $items[$item][$tag] .= $data; - break; - case 'IMAGE': - case 'LOGO': - $image += array($tag => ''); - $image[$tag] .= $data; - break; - case 'LINK': - if ($data) { - $items[$item] += array($tag => ''); - $items[$item][$tag] .= $data; - } - break; - case 'CONTENT': - $items[$item] += array('CONTENT' => ''); - $items[$item]['CONTENT'] .= $data; - break; - case 'SUMMARY': - $items[$item] += array('SUMMARY' => ''); - $items[$item]['SUMMARY'] .= $data; - break; - case 'TAGLINE': - case 'SUBTITLE': - $channel += array('DESCRIPTION' => ''); - $channel['DESCRIPTION'] .= $data; - break; - case 'INFO': - case 'ID': - case 'TEXTINPUT': - // The sub-element is not supported. However, we must recognize - // it or its contents will end up in the item array. - break; - default: - $channel += array($tag => ''); - $channel[$tag] .= $data; - } } /** * Checks a news feed for new items. * * @param $feed - * An associative array describing the feed to be refreshed. + * An object describing the feed to be refreshed. */ function aggregator_refresh($feed) { - global $channel, $image; - - // Generate conditional GET headers. - $headers = array(); - if ($feed['etag']) { - $headers['If-None-Match'] = $feed['etag']; - } - if ($feed['modified']) { - $headers['If-Modified-Since'] = gmdate('D, d M Y H:i:s', $feed['modified']) . ' GMT'; - } - - // Request feed. - $result = drupal_http_request($feed['url'], $headers); - - // Process HTTP response code. - switch ($result->code) { - case 304: - db_update('aggregator_feed') - ->fields(array('checked' => REQUEST_TIME)) - ->condition('fid', $feed['fid']) - ->execute(); - drupal_set_message(t('There is no new syndicated content from %site.', array('%site' => $feed['title']))); - break; - case 301: - $feed['url'] = $result->redirect_url; - // Do not break here. - case 200: - case 302: - case 307: - // We store the md5 hash of feed data in the database. When refreshing a - // feed we compare stored hash and new hash calculated from downloaded - // data. If both are equal we say that feed is not updated. - $md5 = md5($result->data); - if ($feed['hash'] == $md5) { - db_update('aggregator_feed') - ->condition('fid', $feed['fid']) - ->fields(array('checked' => REQUEST_TIME)) - ->execute(); - drupal_set_message(t('There is no new syndicated content from %site.', array('%site' => $feed['title']))); - break; - } - - // Filter the input data. - if (aggregator_parse_feed($result->data, $feed)) { - $modified = empty($result->headers['Last-Modified']) ? 0 : strtotime($result->headers['Last-Modified']); - - // Prepare the channel data. - foreach ($channel as $key => $value) { - $channel[$key] = trim($value); - } - - // Prepare the image data (if any). - foreach ($image as $key => $value) { - $image[$key] = trim($value); - } - - if (!empty($image['LINK']) && !empty($image['URL']) && !empty($image['TITLE'])) { - $image = l(theme('image', $image['URL'], $image['TITLE']), $image['LINK'], array('html' => TRUE)); - } - else { - $image = ''; - } - - $etag = empty($result->headers['ETag']) ? '' : $result->headers['ETag']; - // Update the feed data. - db_merge('aggregator_feed') - ->key(array('fid' => $feed['fid'])) - ->fields(array( - 'url' => $feed['url'], - 'checked' => REQUEST_TIME, - 'link' => $channel['LINK'], - 'description' => $channel['DESCRIPTION'], - 'image' => $image, - 'hash' => $md5, - 'etag' => $etag, - 'modified' => $modified, - )) - ->execute(); - - // Clear the cache. - cache_clear_all(); - - if (isset($result->redirect_url)) { - watchdog('aggregator', 'Updated URL for feed %title to %url.', array('%title' => $feed['title'], '%url' => $feed['url'])); - } - - watchdog('aggregator', 'There is new syndicated content from %site.', array('%site' => $feed['title'])); - drupal_set_message(t('There is new syndicated content from %site.', array('%site' => $feed['title']))); - } - break; - default: - watchdog('aggregator', 'The feed from %site seems to be broken, due to "%error".', array('%site' => $feed['title'], '%error' => $result->code . ' ' . $result->error), WATCHDOG_WARNING); - drupal_set_message(t('The feed from %site seems to be broken, because of error "%error".', array('%site' => $feed['title'], '%error' => $result->code . ' ' . $result->error))); - module_invoke('system', 'check_http_request'); - } -} - -/** - * Parse the W3C date/time format, a subset of ISO 8601. - * - * PHP date parsing functions do not handle this format. - * See http://www.w3.org/TR/NOTE-datetime for more information. - * Originally from MagpieRSS (http://magpierss.sourceforge.net/). - * - * @param $date_str - * A string with a potentially W3C DTF date. - * @return - * A timestamp if parsed successfully or FALSE if not. - */ -function aggregator_parse_w3cdtf($date_str) { - if (preg_match('/(\d{4})-(\d{2})-(\d{2})T(\d{2}):(\d{2})(:(\d{2}))?(?:([-+])(\d{2}):?(\d{2})|(Z))?/', $date_str, $match)) { - list($year, $month, $day, $hours, $minutes, $seconds) = array($match[1], $match[2], $match[3], $match[4], $match[5], $match[6]); - // Calculate the epoch for current date assuming GMT. - $epoch = gmmktime($hours, $minutes, $seconds, $month, $day, $year); - if ($match[10] != 'Z') { // Z is zulu time, aka GMT - list($tz_mod, $tz_hour, $tz_min) = array($match[8], $match[9], $match[10]); - // Zero out the variables. - if (!$tz_hour) { - $tz_hour = 0; - } - if (!$tz_min) { - $tz_min = 0; - } - $offset_secs = (($tz_hour * 60) + $tz_min) * 60; - // Is timezone ahead of GMT? If yes, subtract offset. - if ($tz_mod == '+') { - $offset_secs *= -1; - } - $epoch += $offset_secs; - } - return $epoch; - } - else { - return FALSE; - } -} - -/** - * Parse a feed and store its items. - * - * @param $data - * The feed data. - * @param $feed - * An associative array describing the feed to be parsed. - * @return - * FALSE on error, TRUE otherwise. - */ -function aggregator_parse_feed(&$data, $feed) { - global $items, $image, $channel; - - // Unset the global variables before we use them. - unset($GLOBALS['element'], $GLOBALS['item'], $GLOBALS['tag']); - $items = array(); - $image = array(); - $channel = array(); - - // Parse the data. - $xml_parser = drupal_xml_parser_create($data); - xml_set_element_handler($xml_parser, 'aggregator_element_start', 'aggregator_element_end'); - xml_set_character_data_handler($xml_parser, 'aggregator_element_data'); - - if (!xml_parse($xml_parser, $data, 1)) { - watchdog('aggregator', 'The feed from %site seems to be broken, due to an error "%error" on line %line.', array('%site' => $feed['title'], '%error' => xml_error_string(xml_get_error_code($xml_parser)), '%line' => xml_get_current_line_number($xml_parser)), WATCHDOG_WARNING); - drupal_set_message(t('The feed from %site seems to be broken, because of error "%error" on line %line.', array('%site' => $feed['title'], '%error' => xml_error_string(xml_get_error_code($xml_parser)), '%line' => xml_get_current_line_number($xml_parser))), 'error'); - return FALSE; - } - xml_parser_free($xml_parser); - - // We reverse the array such that we store the first item last, and the last - // item first. In the database, the newest item should be at the top. - $items = array_reverse($items); - - // Initialize variables. - $title = $link = $author = $description = $guid = NULL; - foreach ($items as $item) { - unset($title, $link, $author, $description, $guid); - - // Prepare the item: - foreach ($item as $key => $value) { - $item[$key] = trim($value); - } - - // Resolve the item's title. If no title is found, we use up to 40 - // characters of the description ending at a word boundary, but not - // splitting potential entities. - if (!empty($item['TITLE'])) { - $title = $item['TITLE']; - } - elseif (!empty($item['DESCRIPTION'])) { - $title = preg_replace('/^(.*)[^\w;&].*?$/', "\\1", truncate_utf8($item['DESCRIPTION'], 40)); - } - else { - $title = ''; - } - - // Resolve the items link. - if (!empty($item['LINK'])) { - $link = $item['LINK']; - } - else { - $link = $feed['link']; - } - $guid = isset($item['GUID']) ? $item['GUID'] : ''; - - // Atom feeds have a CONTENT and/or SUMMARY tag instead of a DESCRIPTION tag. - if (!empty($item['CONTENT:ENCODED'])) { - $item['DESCRIPTION'] = $item['CONTENT:ENCODED']; - } - elseif (!empty($item['SUMMARY'])) { - $item['DESCRIPTION'] = $item['SUMMARY']; - } - elseif (!empty($item['CONTENT'])) { - $item['DESCRIPTION'] = $item['CONTENT']; - } - - // Try to resolve and parse the item's publication date. - $date = ''; - foreach (array('PUBDATE', 'DC:DATE', 'DCTERMS:ISSUED', 'DCTERMS:CREATED', 'DCTERMS:MODIFIED', 'ISSUED', 'CREATED', 'MODIFIED', 'PUBLISHED', 'UPDATED') as $key) { - if (!empty($item[$key])) { - $date = $item[$key]; - break; + // Fetch the feed. + $fetcher = variable_get('aggregator_fetcher', 'aggregator'); + module_invoke($fetcher, 'aggregator_fetch', $feed); + + if ($feed->source_string !== FALSE) { + // Parse the feed. + $parser = variable_get('aggregator_parser', 'aggregator'); + module_invoke($parser, 'aggregator_parse', $feed); + + // If there are items on the feed, let all enabled processors do their work on it. + if (@count($feed->items)) { + $processors = variable_get('aggregator_processors', array('aggregator')); + foreach ($processors as $processor) { + module_invoke($processor, 'aggregator_process', $feed); } } - - $timestamp = strtotime($date); - - if ($timestamp === FALSE) { - $timestamp = aggregator_parse_w3cdtf($date); // Aggregator_parse_w3cdtf() returns FALSE on failure. - } - - // Save this item. Try to avoid duplicate entries as much as possible. If - // we find a duplicate entry, we resolve it and pass along its ID is such - // that we can update it if needed. - if (!empty($guid)) { - $entry = db_query("SELECT iid, timestamp FROM {aggregator_item} WHERE fid = :fid AND guid = :guid", array(':fid' => $feed['fid'], ':guid' => $guid))->fetchObject(); - } - elseif ($link && $link != $feed['link'] && $link != $feed['url']) { - $entry = db_query("SELECT iid, timestamp FROM {aggregator_item} WHERE fid = :fid AND link = :link", array(':fid' => $feed['fid'], ':link' => $link))->fetchObject(); - } - else { - $entry = db_query("SELECT iid, timestamp FROM {aggregator_item} WHERE fid = :fid AND title = :title", array(':fid' => $feed['fid'], ':title' => $title))->fetchObject(); - } - - if (!$timestamp) { - $timestamp = isset($entry->timestamp) ? $entry->timestamp : REQUEST_TIME; - } - $item += array('AUTHOR' => '', 'DESCRIPTION' => ''); - aggregator_save_item(array('iid' => (isset($entry->iid) ? $entry->iid : ''), 'fid' => $feed['fid'], 'timestamp' => $timestamp, 'title' => $title, 'link' => $link, 'author' => $item['AUTHOR'], 'description' => $item['DESCRIPTION'], 'guid' => $guid)); - } - - // Remove all items that are older than flush item timer. - $age = REQUEST_TIME - variable_get('aggregator_clear', 9676800); - $iids = db_query('SELECT iid FROM {aggregator_item} WHERE fid = :fid AND timestamp < :timestamp', array(':fid' => $feed['fid'], ':timestamp' => $age))->fetchCol(); - if ($iids) { - db_delete('aggregator_category_item') - ->condition('iid', $iids, 'IN') - ->execute(); - db_delete('aggregator_item') - ->condition('iid', $iids, 'IN') - ->execute(); - } - - return TRUE; -} - -/** - * Add/edit/delete an aggregator item. - * - * @param $edit - * An associative array describing the item to be added/edited/deleted. - */ -function aggregator_save_item($edit) { - if ($edit['title'] && empty($edit['iid'])) { - $edit['iid'] = db_insert('aggregator_item') - ->fields(array( - 'title' => $edit['title'], - 'link' => $edit['link'], - 'author' => $edit['author'], - 'description' => $edit['description'], - 'guid' => $edit['guid'], - 'timestamp' => $edit['timestamp'], - 'fid' => $edit['fid'], - )) - ->execute(); - } - if ($edit['iid'] && !$edit['title']) { - db_delete('aggregator_item') - ->condition('iid', $edit['iid']) - ->execute(); - db_delete('aggregator_category_item') - ->condition('iid', $edit['iid']) - ->execute(); } - elseif ($edit['title'] && $edit['link']) { - // file the items in the categories indicated by the feed - $result = db_query('SELECT cid FROM {aggregator_category_feed} WHERE fid = :fid', array(':fid' => $edit['fid'])); - foreach ($result as $category) { - db_merge('aggregator_category_item') - ->fields(array( - 'cid' => $category->cid, - 'iid' => $edit['iid'], - )) - ->execute(); - } + // Expire old feed items. + if (drupal_function_exists('aggregator_expire')) { + aggregator_expire($feed); } } @@ -962,12 +558,12 @@ function aggregator_save_item($edit) { * @param $fid * The feed id. * @return - * An associative array describing the feed. + * An object describing the feed. */ function aggregator_feed_load($fid) { static $feeds; if (!isset($feeds[$fid])) { - $feeds[$fid] = db_query('SELECT * FROM {aggregator_feed} WHERE fid = :fid', array(':fid' => $fid))->fetchAssoc(); + $feeds[$fid] = db_query('SELECT * FROM {aggregator_feed} WHERE fid = :fid', array(':fid' => $fid))->fetchObject(); } return $feeds[$fid]; @@ -1004,9 +600,8 @@ function aggregator_category_load($cid) { function theme_aggregator_block_item($item, $feed = 0) { // Display the external link to the item. - $output .= '<a href="' . check_url($item->link) . '">' . check_plain($item->title) . "</a>\n"; - - return $output; + return '<a href="' . check_url($item->link) . '">' . check_plain($item->title) . "</a>\n"; + } /** @@ -1022,6 +617,41 @@ function aggregator_filter_xss($value) { } /** + * Check and sanitize aggregator configuration. + * + * Goes through all fetchers, parsers and processors and checks whether they are available. + * If one is missing resets to standard configuration. + * + * @return + * TRUE if this function reset the configuration FALSE if not. + */ +function aggregator_sanitize_configuration() { + $reset = FALSE; + $fetcher = variable_get('aggregator_fetcher', 'aggregator'); + if (!module_exists($fetcher)) { + $reset = TRUE; + } + $parser = variable_get('aggregator_parser', 'aggregator'); + if (!module_exists($parser)) { + $reset = TRUE; + } + $processors = variable_get('aggregator_processors', array('aggregator')); + foreach ($processors as $processor) { + if (!module_exists($processor)) { + $reset = TRUE; + break; + } + } + if ($reset) { + variable_del('aggregator_fetcher'); + variable_del('aggregator_parser'); + variable_del('aggregator_processors'); + return TRUE; + } + return FALSE; +} + +/** * Helper function for drupal_map_assoc. * * @param $count |