/**
* Return an appropriately formatted LIKE query and the clause
*/
static function mungeQuery($query, $prot)
{
$field = 'el_index';
$rv = LinkFilter::makeLikeArray($query, $prot);
if ($rv === false) {
//makeLike doesn't handle wildcard in IP, so we'll have to munge here.
if (preg_match('/^(:?[0-9]{1,3}\\.)+\\*\\s*$|^(:?[0-9]{1,3}\\.){3}[0-9]{1,3}:[0-9]*\\*\\s*$/', $query)) {
$rv = array($prot . rtrim($query, " \t*"), $dbr->anyString());
$field = 'el_to';
}
}
return array($rv, $field);
}
public function execute()
{
global $IP, $wgLocalDatabases, $wgUser;
$username = wfMessage('spambot_username')->text();
$wgUser = User::newFromName($username);
if (!$wgUser) {
$this->error("Invalid username specified in 'spambot_username' message: {$username}", true);
}
// Create the user if necessary
if (!$wgUser->getId()) {
$wgUser->addToDatabase();
}
$spec = $this->getArg();
$like = LinkFilter::makeLikeArray($spec);
if (!$like) {
$this->error("Not a valid hostname specification: {$spec}", true);
}
if ($this->hasOption('all')) {
// Clean up spam on all wikis
$this->output("Finding spam on " . count($wgLocalDatabases) . " wikis\n");
$found = false;
foreach ($wgLocalDatabases as $wikiID) {
$dbr = wfGetDB(DB_SLAVE, array(), $wikiID);
$count = $dbr->selectField('externallinks', 'COUNT(*)', array('el_index' . $dbr->buildLike($like)), __METHOD__);
if ($count) {
$found = true;
$cmd = wfShellWikiCmd("{$IP}/maintenance/cleanupSpam.php", array('--wiki', $wikiID, $spec));
passthru("{$cmd} | sed 's/^/{$wikiID}: /'");
}
}
if ($found) {
$this->output("All done\n");
} else {
$this->output("None found\n");
}
} else {
// Clean up spam on this wiki
$dbr = wfGetDB(DB_SLAVE);
$res = $dbr->select('externallinks', array('DISTINCT el_from'), array('el_index' . $dbr->buildLike($like)), __METHOD__);
$count = $dbr->numRows($res);
$this->output("Found {$count} articles containing {$spec}\n");
foreach ($res as $row) {
$this->cleanupArticle($row->el_from, $spec);
}
if ($count) {
$this->output("Done\n");
}
}
}
/**
* @param $query String
* @param $protocol String
* @return null|string
*/
public function prepareUrlQuerySearchString($query = null, $protocol = null)
{
$db = $this->getDb();
if (!is_null($query) || $query != '') {
if (is_null($protocol)) {
$protocol = 'http://';
}
$likeQuery = LinkFilter::makeLikeArray($query, $protocol);
if (!$likeQuery) {
$this->dieUsage('Invalid query', 'bad_query');
}
$likeQuery = LinkFilter::keepOneWildcard($likeQuery);
return 'el_index ' . $db->buildLike($likeQuery);
} elseif (!is_null($protocol)) {
return 'el_index ' . $db->buildLike("{$protocol}", $db->anyString());
}
return null;
}
/**
* testMakeLikeArrayWithInvalidPatterns()
*
* Tests whether LinkFilter::makeLikeArray($pattern) will reject invalid search patterns
*
* @dataProvider provideInvalidPatterns
*
* @param string $pattern Invalid search pattern
*/
function testMakeLikeArrayWithInvalidPatterns($pattern)
{
$this->assertFalse(LinkFilter::makeLikeArray($pattern), "'{$pattern}' is not a valid pattern and should be rejected");
}
/**
* Return an appropriately formatted LIKE query and the clause
*
* @param string $query Search pattern to search for
* @param string $prot Protocol, e.g. 'http://'
*
* @return array
*/
static function mungeQuery($query, $prot)
{
$field = 'el_index';
$dbr = wfGetDB(DB_SLAVE);
if ($query === '*' && $prot !== '') {
// Allow queries like 'ftp://*' to find all ftp links
$rv = array($prot, $dbr->anyString());
} else {
$rv = LinkFilter::makeLikeArray($query, $prot);
}
if ($rv === false) {
// LinkFilter doesn't handle wildcard in IP, so we'll have to munge here.
$pattern = '/^(:?[0-9]{1,3}\\.)+\\*\\s*$|^(:?[0-9]{1,3}\\.){3}[0-9]{1,3}:[0-9]*\\*\\s*$/';
if (preg_match($pattern, $query)) {
$rv = array($prot . rtrim($query, " \t*"), $dbr->anyString());
$field = 'el_to';
}
}
return array($rv, $field);
}
private function run($resultPageSet = null)
{
$params = $this->extractRequestParams();
$protocol = $params['protocol'];
$query = $params['query'];
// Find the right prefix
global $wgUrlProtocols;
if ($protocol && !in_array($protocol, $wgUrlProtocols)) {
foreach ($wgUrlProtocols as $p) {
if (substr($p, 0, strlen($protocol)) === $protocol) {
$protocol = $p;
break;
}
}
} else {
$protocol = null;
}
$db = $this->getDB();
$this->addTables(array('page', 'externallinks'));
// must be in this order for 'USE INDEX'
$this->addOption('USE INDEX', 'el_index');
$this->addWhere('page_id=el_from');
$this->addWhereFld('page_namespace', $params['namespace']);
if (!is_null($query) || $query != '') {
if (is_null($protocol)) {
$protocol = 'http://';
}
$likeQuery = LinkFilter::makeLikeArray($query, $protocol);
if (!$likeQuery) {
$this->dieUsage('Invalid query', 'bad_query');
}
$likeQuery = LinkFilter::keepOneWildcard($likeQuery);
$this->addWhere('el_index ' . $db->buildLike($likeQuery));
} elseif (!is_null($protocol)) {
$this->addWhere('el_index ' . $db->buildLike("{$protocol}", $db->anyString()));
}
$prop = array_flip($params['prop']);
$fld_ids = isset($prop['ids']);
$fld_title = isset($prop['title']);
$fld_url = isset($prop['url']);
if (is_null($resultPageSet)) {
$this->addFields(array('page_id', 'page_namespace', 'page_title'));
$this->addFieldsIf('el_to', $fld_url);
} else {
$this->addFields($resultPageSet->getPageTableFields());
}
$limit = $params['limit'];
$offset = $params['offset'];
$this->addOption('LIMIT', $limit + 1);
if (isset($offset)) {
$this->addOption('OFFSET', $offset);
}
$res = $this->select(__METHOD__);
$result = $this->getResult();
$count = 0;
foreach ($res as $row) {
if (++$count > $limit) {
// We've reached the one extra which shows that there are additional pages to be had. Stop here...
$this->setContinueEnumParameter('offset', $offset + $limit);
break;
}
if (is_null($resultPageSet)) {
$vals = array();
if ($fld_ids) {
$vals['pageid'] = intval($row->page_id);
}
if ($fld_title) {
$title = Title::makeTitle($row->page_namespace, $row->page_title);
ApiQueryBase::addTitleInfo($vals, $title);
}
if ($fld_url) {
$vals['url'] = $row->el_to;
}
$fit = $result->addValue(array('query', $this->getModuleName()), null, $vals);
if (!$fit) {
$this->setContinueEnumParameter('offset', $offset + $count - 1);
break;
}
} else {
$resultPageSet->processDbRow($row);
}
}
if (is_null($resultPageSet)) {
$result->setIndexedTagName_internal(array('query', $this->getModuleName()), $this->getModulePrefix());
}
}
