function processPage($pageContent)
{
try {
$links = array();
$html = str_get_html($pageContent);
//reparam html stricat
if (!$html->find('body', 0, true)) {
$html = $this->fixHtml($html);
}
$body = $html->find('body', 0, true);
$this->extractText($body);
foreach ($body->find("a") as $link) {
$links[] = $link->href;
}
//cata memorie consuma
//si eliberare referinte pierdute
$html->clear();
MemoryManagement::showUsage('before cleaning', true, 'KB');
MemoryManagement::clean(true);
MemoryManagement::showUsage('after cleaning', true, 'KB');
return $links;
} catch (Exception $ex) {
Applog::exceptionLog($ex);
}
}
function start()
{
Applog::log("INSIDE " . __FILE__ . ' - ' . __CLASS__ . '::' . __FUNCTION__ . '() - ' . 'line ' . __LINE__, 4);
while (($file = $this->getNextFile()) != null) {
$this->processFile($file);
MemoryManagement::clean();
}
Applog::log("Finished");
}
