1 files changed, 810 insertions, 779 deletions
diff --git a/inc/3rdparty/libraries/humble-http-agent/HumbleHttpAgent.php b/inc/3rdparty/libraries/humble-http-agent/HumbleHttpAgent.php
index e4f1b3b3..963f0c05 100644
--- a/inc/3rdparty/libraries/humble-http-agent/HumbleHttpAgent.php
+++ b/inc/3rdparty/libraries/humble-http-agent/HumbleHttpAgent.php
@@ -1,779 +1,810 @@
-<?php
+<?php
-/**
+/**
- * Humble HTTP Agent
+ * Humble HTTP Agent
- * 
+ * 
- * This class is designed to take advantage of parallel HTTP requests
+ * This class is designed to take advantage of parallel HTTP requests
- * offered by PHP's PECL HTTP extension or the curl_multi_* functions. 
+ * offered by PHP's PECL HTTP extension or the curl_multi_* functions. 
- * For environments which do not have these options, it reverts to standard sequential 
+ * For environments which do not have these options, it reverts to standard sequential 
- * requests (using file_get_contents())
+ * requests (using file_get_contents())
- * 
+ * 
- * @version 1.1
+ * @version 1.4
- * @date 2012-08-20
+ * @date 2013-05-10
- * @see http://php.net/HttpRequestPool
+ * @see http://php.net/HttpRequestPool
- * @author Keyvan Minoukadeh
+ * @author Keyvan Minoukadeh
- * @copyright 2011-2012 Keyvan Minoukadeh
+ * @copyright 2011-2013 Keyvan Minoukadeh
- * @license http://www.gnu.org/licenses/agpl-3.0.html AGPL v3
+ * @license http://www.gnu.org/licenses/agpl-3.0.html AGPL v3
- */
+ */
-class HumbleHttpAgent
+class HumbleHttpAgent
-{
+{
-        const METHOD_REQUEST_POOL = 1;
+        const METHOD_REQUEST_POOL = 1;
-        const METHOD_CURL_MULTI = 2;
+        const METHOD_CURL_MULTI = 2;
-        const METHOD_FILE_GET_CONTENTS = 4;
+        const METHOD_FILE_GET_CONTENTS = 4;
-        //const UA_BROWSER = 'Mozilla/5.0 (Windows NT 6.1; WOW64; rv:2.0.1) Gecko/20100101 Firefox/4.0.1';
+        //const UA_BROWSER = 'Mozilla/5.0 (Windows NT 6.1; WOW64; rv:2.0.1) Gecko/20100101 Firefox/4.0.1';
-        const UA_BROWSER = 'Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/535.2 (KHTML, like Gecko) Chrome/15.0.874.92 Safari/535.2';
+        const UA_BROWSER = 'Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/535.2 (KHTML, like Gecko) Chrome/15.0.874.92 Safari/535.2';
-        const UA_PHP = 'PHP/5.2';
+        const UA_PHP = 'PHP/5.4';
-        const REF_GOOGLE = 'http://www.google.co.uk/url?sa=t&source=web&cd=1';
+        const REF_GOOGLE = 'http://www.google.co.uk/url?sa=t&source=web&cd=1';
-        
+        
-        protected $requests = array();
+        protected $requests = array();
-        protected $redirectQueue = array();
+        protected $redirectQueue = array();
-        protected $requestOptions;
+        protected $requestOptions;
-        protected $maxParallelRequests = 5;
+        protected $maxParallelRequests = 5;
-        protected $cache = null; //TODO
+        protected $cache = null; //TODO
-        protected $httpContext;
+        protected $httpContext;
-        protected $minimiseMemoryUse = false; //TODO
+        protected $minimiseMemoryUse = false; //TODO
-        protected $method;
+        protected $method;
-        protected $cookieJar;
+        protected $cookieJar;
-        public $debug = false;
+        public $debug = false;
-        public $debugVerbose = false;
+        public $debugVerbose = false;
-        public $rewriteHashbangFragment = true; // see http://code.google.com/web/ajaxcrawling/docs/specification.html
+        public $rewriteHashbangFragment = true; // see http://code.google.com/web/ajaxcrawling/docs/specification.html
-        public $maxRedirects = 5;
+        public $maxRedirects = 5;
-        public $userAgentMap = array();
+        public $userAgentMap = array();
-        public $rewriteUrls = array();
+        public $rewriteUrls = array();
-        public $userAgentDefault;
+        public $userAgentDefault;
-        public $referer;
+        public $referer;
-        //public $userAgent = 'Mozilla/5.0';
+        //public $userAgent = 'Mozilla/5.0';
-        
+        
-        // Prevent certain file/mime types
+        // Prevent certain file/mime types
-        // HTTP responses which match these content types will
+        // HTTP responses which match these content types will
-        // be returned without body.
+        // be returned without body.
-        public $headerOnlyTypes = array();
+        public $headerOnlyTypes = array();
-        // URLs ending with one of these extensions will
+        // URLs ending with one of these extensions will
-        // prompt Humble HTTP Agent to send a HEAD request first
+        // prompt Humble HTTP Agent to send a HEAD request first
-        // to see if returned content type matches $headerOnlyTypes.
+        // to see if returned content type matches $headerOnlyTypes.
-        public $headerOnlyClues = array('pdf','mp3','zip','exe','gif','gzip','gz','jpeg','jpg','mpg','mpeg','png','ppt','mov');
+        public $headerOnlyClues = array('pdf','mp3','zip','exe','gif','gzip','gz','jpeg','jpg','mpg','mpeg','png','ppt','mov');
-        // AJAX triggers to search for.
+        // AJAX triggers to search for.
-        // for AJAX sites, e.g. Blogger with its dynamic views templates.
+        // for AJAX sites, e.g. Blogger with its dynamic views templates.
-        public $ajaxTriggers = array("<meta name='fragment' content='!'",'<meta name="fragment" content="!"',"<meta content='!' name='fragment'",'<meta content="!" name="fragment"');
+        public $ajaxTriggers = array("<meta name='fragment' content='!'",'<meta name="fragment" content="!"',"<meta content='!' name='fragment'",'<meta content="!" name="fragment"');
-        
+        
-        //TODO: set max file size
+        //TODO: set max file size
-        //TODO: normalise headers
+        //TODO: normalise headers
-        
+        
-        function __construct($requestOptions=null, $method=null) {
+        function __construct($requestOptions=null, $method=null) {
-                $this->userAgentDefault = self::UA_BROWSER;
+                $this->userAgentDefault = self::UA_BROWSER;
-                $this->referer = self::REF_GOOGLE;
+                $this->referer = self::REF_GOOGLE;
-                // set the request method
+                // set the request method
-                if (in_array($method, array(1,2,4))) {
+                if (in_array($method, array(1,2,4))) {
-                        $this->method = $method;
+                        $this->method = $method;
-                } else {
+                } else {
-                        if (class_exists('HttpRequestPool')) {
+                        if (class_exists('HttpRequestPool')) {
-                                $this->method = self::METHOD_REQUEST_POOL;
+                                $this->method = self::METHOD_REQUEST_POOL;
-                        } elseif (function_exists('curl_multi_init')) {
+                        } elseif (function_exists('curl_multi_init')) {
-                                $this->method = self::METHOD_CURL_MULTI;
+                                $this->method = self::METHOD_CURL_MULTI;
-                        } else {
+                        } else {
-                                $this->method = self::METHOD_FILE_GET_CONTENTS;
+                                $this->method = self::METHOD_FILE_GET_CONTENTS;
-                        }
+                        }
-                }
+                }
-                if ($this->method == self::METHOD_CURL_MULTI) {
+                if ($this->method == self::METHOD_CURL_MULTI) {
-                        require_once(dirname(__FILE__).'/RollingCurl.php');
+                        require_once(dirname(__FILE__).'/RollingCurl.php');
-                }
+                }
-                // create cookie jar
+                // create cookie jar
-                $this->cookieJar = new CookieJar();
+                $this->cookieJar = new CookieJar();
-                // set request options (redirect must be 0)
+                // set request options (redirect must be 0)
-                $this->requestOptions = array(
+                $this->requestOptions = array(
-                        'timeout' => 15,
+                        'timeout' => 15,
-                        'redirect' => 0 // we handle redirects manually so we can rewrite the new hashbang URLs that are creeping up over the web
+                        'connecttimeout' => 15,
-                        // TODO: test onprogress?
+                        'dns_cache_timeout' => 300,
-                );
+                        'redirect' => 0 // we handle redirects manually so we can rewrite the new hashbang URLs that are creeping up over the web
-                if (is_array($requestOptions)) {
+                        // TODO: test onprogress?
-                        $this->requestOptions = array_merge($this->requestOptions, $requestOptions);
+                );
-                }
+                if (is_array($requestOptions)) {
-                $this->httpContext = array(
+                        $this->requestOptions = array_merge($this->requestOptions, $requestOptions);
-                        'http' => array(
+                }
-                                'ignore_errors' => true,
+                $this->httpContext = array(
-                                'timeout' => $this->requestOptions['timeout'],
+                        'http' => array(
-                                'max_redirects' => $this->requestOptions['redirect'],
+                                'ignore_errors' => true,
-                                'header' => "Accept: */*\r\n"
+                                'timeout' => $this->requestOptions['timeout'],
-                                )
+                                'max_redirects' => $this->requestOptions['redirect'],
-                        );
+                                'header' => "Accept: */*\r\n"
-        }
+                                )
-        
+                        );
-        protected function debug($msg) {
+        }
-                if ($this->debug) {
+        
-                        $mem = round(memory_get_usage()/1024, 2);
+        protected function debug($msg) {
-                        $memPeak = round(memory_get_peak_usage()/1024, 2);
+                if ($this->debug) {
-                        echo '* ',$msg;
+                        $mem = round(memory_get_usage()/1024, 2);
-                        if ($this->debugVerbose) echo ' - mem used: ',$mem," (peak: $memPeak)";
+                        $memPeak = round(memory_get_peak_usage()/1024, 2);
-                        echo "\n";
+                        echo '* ',$msg;
-                        ob_flush();
+                        if ($this->debugVerbose) echo ' - mem used: ',$mem," (peak: $memPeak)";
-                        flush();
+                        echo "\n";
-                }
+                        ob_flush();
-        }
+                        flush();
-        
+                }
-        protected function getUserAgent($url, $asArray=false) {
+        }
-                $host = @parse_url($url, PHP_URL_HOST);
+        
-                if (strtolower(substr($host, 0, 4)) == 'www.') {
+        protected function getUserAgent($url, $asArray=false) {
-                        $host = substr($host, 4);
+                $host = @parse_url($url, PHP_URL_HOST);
-                }
+                if (strtolower(substr($host, 0, 4)) == 'www.') {
-                if ($host) {
+                        $host = substr($host, 4);
-                        $try = array($host);
+                }
-                        $split = explode('.', $host);
+                if ($host) {
-                        if (count($split) > 1) {
+                        $try = array($host);
-                                array_shift($split);
+                        $split = explode('.', $host);
-                                $try[] = '.'.implode('.', $split);
+                        if (count($split) > 1) {
-                        }
+                                array_shift($split);
-                        foreach ($try as $h) {
+                                $try[] = '.'.implode('.', $split);
-                                if (isset($this->userAgentMap[$h])) {
+                        }
-                                        $ua = $this->userAgentMap[$h];
+                        foreach ($try as $h) {
-                                        break;
+                                if (isset($this->userAgentMap[$h])) {
-                                }
+                                        $ua = $this->userAgentMap[$h];
-                        }
+                                        break;
-                }
+                                }
-                if (!isset($ua)) $ua = $this->userAgentDefault;
+                        }
-                if ($asArray) {
+                }
-                        return array('User-Agent' => $ua);
+                if (!isset($ua)) $ua = $this->userAgentDefault;
-                } else {
+                if ($asArray) {
-                        return 'User-Agent: '.$ua;
+                        return array('User-Agent' => $ua);
-                }
+                } else {
-        }
+                        return 'User-Agent: '.$ua;
-        
+                }
-        public function rewriteHashbangFragment($url) {
+        }
-                // return $url if there's no '#!'
+        
-                if (strpos($url, '#!') === false) return $url;
+        public function rewriteHashbangFragment($url) {
-                // split $url and rewrite
+                // return $url if there's no '#!'
-                // TODO: is SimplePie_IRI included?
+                if (strpos($url, '#!') === false) return $url;
-                $iri = new SimplePie_IRI($url);
+                // split $url and rewrite
-                $fragment = substr($iri->fragment, 1); // strip '!'
+                // TODO: is SimplePie_IRI included?
-                $iri->fragment = null;
+                $iri = new SimplePie_IRI($url);
-                if (isset($iri->query)) {
+                $fragment = substr($iri->fragment, 1); // strip '!'
-                        parse_str($iri->query, $query);
+                $iri->fragment = null;
-                } else {
+                if (isset($iri->query)) {
-                        $query = array();
+                        parse_str($iri->query, $query);
-                }
+                } else {
-                $query['_escaped_fragment_'] = (string)$fragment;
+                        $query = array();
-                $iri->query = str_replace('%2F', '/', http_build_query($query)); // needed for some sites
+                }
-                return $iri->get_iri();
+                $query['_escaped_fragment_'] = (string)$fragment;
-        }
+                $iri->query = str_replace('%2F', '/', http_build_query($query)); // needed for some sites
-        
+                return $iri->get_iri();
-        public function getUglyURL($url, $html) {
+        }
-                if ($html == '') return false;
+        
-                $found = false;
+        public function getRedirectURLfromHTML($url, $html) {
-                foreach ($this->ajaxTriggers as $string) {
+                $redirect_url = $this->getMetaRefreshURL($url, $html);
-                        if (stripos($html, $string)) {
+                if (!$redirect_url) {
-                                $found = true;
+                        $redirect_url = $this->getUglyURL($url, $html);
-                                break;
+                }
-                        }
+                return $redirect_url;
-                }
+        }
-                if (!$found) return false;
+        
-                $iri = new SimplePie_IRI($url);
+        public function getMetaRefreshURL($url, $html) {
-                if (isset($iri->query)) {
+                if ($html == '') return false;
-                        parse_str($iri->query, $query);
+                // <meta HTTP-EQUIV="REFRESH" content="0; url=http://www.bernama.com/bernama/v6/newsindex.php?id=943513">
-                } else {
+                if (!preg_match('!<meta http-equiv=["\']?refresh["\']? content=["\']?[0-9];\s*url=["\']?([^"\'>]+)["\']*>!i', $html, $match)) {
-                        $query = array();
+                        return false;
-                }
+                }
-                $query['_escaped_fragment_'] = '';
+                $redirect_url = $match[1];
-                $iri->query = str_replace('%2F', '/', http_build_query($query)); // needed for some sites
+                if (preg_match('!^https?://!i', $redirect_url)) {
-                return $iri->get_iri();
+                        // already absolute
-        }
+                        $this->debug('Meta refresh redirect found (http-equiv="refresh"), new URL: '.$redirect_url);
-        
+                        return $redirect_url;
-        public function removeFragment($url) {
+                }
-                $pos = strpos($url, '#');
+                // absolutize redirect URL
-                if ($pos === false) {
+                $base = new SimplePie_IRI($url);
-                        return $url;
+                // remove '//' in URL path (causes URLs not to resolve properly)
-                } else {
+                if (isset($base->path)) $base->path = preg_replace('!//+!', '/', $base->path);
-                        return substr($url, 0, $pos);
+                if ($absolute = SimplePie_IRI::absolutize($base, $redirect_url)) {
-                }
+                        $this->debug('Meta refresh redirect found (http-equiv="refresh"), new URL: '.$absolute);
-        }
+                        return $absolute;
-        
+                }
-        public function rewriteUrls($url) {
+                return false;
-                foreach ($this->rewriteUrls as $find => $action) {
+        }       
-                        if (strpos($url, $find) !== false) {
+        
-                                if (is_array($action)) {
+        public function getUglyURL($url, $html) {
-                                        return strtr($url, $action);
+                if ($html == '') return false;
-                                }
+                $found = false;
-                        }
+                foreach ($this->ajaxTriggers as $string) {
-                }
+                        if (stripos($html, $string)) {
-                return $url;
+                                $found = true;
-        }
+                                break;
-        
+                        }
-        public function enableDebug($bool=true) {
+                }
-                $this->debug = (bool)$bool;
+                if (!$found) return false;
-        }
+                $iri = new SimplePie_IRI($url);
-        
+                if (isset($iri->query)) {
-        public function minimiseMemoryUse($bool = true) {
+                        parse_str($iri->query, $query);
-                $this->minimiseMemoryUse = $bool;
+                } else {
-        }
+                        $query = array();
-        
+                }
-        public function setMaxParallelRequests($max) {
+                $query['_escaped_fragment_'] = '';
-                $this->maxParallelRequests = $max;
+                $iri->query = str_replace('%2F', '/', http_build_query($query)); // needed for some sites
-        }
+                $ugly_url = $iri->get_iri();
-        
+                $this->debug('AJAX trigger (meta name="fragment" content="!") found, new URL: '.$ugly_url);
-        public function validateUrl($url) {
+                return $ugly_url;
-                $url = filter_var($url, FILTER_SANITIZE_URL);
+        }
-                $test = filter_var($url, FILTER_VALIDATE_URL, FILTER_FLAG_SCHEME_REQUIRED);
+        
-                // deal with bug http://bugs.php.net/51192 (present in PHP 5.2.13 and PHP 5.3.2)
+        public function removeFragment($url) {
-                if ($test === false) {
+                $pos = strpos($url, '#');
-                        $test = filter_var(strtr($url, '-', '_'), FILTER_VALIDATE_URL, FILTER_FLAG_SCHEME_REQUIRED);
+                if ($pos === false) {
-                }
+                        return $url;
-                if ($test !== false && $test !== null && preg_match('!^https?://!', $url)) {
+                } else {
-                        return $url;
+                        return substr($url, 0, $pos);
-                } else {
+                }
-                        return false;
+        }
-                }
+        
-        }
+        public function rewriteUrls($url) {
-        
+                foreach ($this->rewriteUrls as $find => $action) {
-        public function fetchAll(array $urls) {
+                        if (strpos($url, $find) !== false) {
-                $this->fetchAllOnce($urls, $isRedirect=false);
+                                if (is_array($action)) {
-                $redirects = 0;
+                                        return strtr($url, $action);
-                while (!empty($this->redirectQueue) && ++$redirects <= $this->maxRedirects) {
+                                }
-                        $this->debug("Following redirects #$redirects...");
+                        }
-                        $this->fetchAllOnce($this->redirectQueue, $isRedirect=true);
+                }
-                }
+                return $url;
-        }
+        }
-        
+        
-        // fetch all URLs without following redirects
+        public function enableDebug($bool=true) {
-        public function fetchAllOnce(array $urls, $isRedirect=false) {
+                $this->debug = (bool)$bool;
-                if (!$isRedirect) $urls = array_unique($urls);
+        }
-                if (empty($urls)) return;
+        
-                
+        public function minimiseMemoryUse($bool = true) {
-                //////////////////////////////////////////////////////
+                $this->minimiseMemoryUse = $bool;
-                // parallel (HttpRequestPool)
+        }
-                if ($this->method == self::METHOD_REQUEST_POOL) {
+        
-                        $this->debug('Starting parallel fetch (HttpRequestPool)');
+        public function setMaxParallelRequests($max) {
-                        try {
+                $this->maxParallelRequests = $max;
-                                while (count($urls) > 0) {
+        }
-                                        $this->debug('Processing set of '.min($this->maxParallelRequests, count($urls)));
+        
-                                        $subset = array_splice($urls, 0, $this->maxParallelRequests);
+        public function validateUrl($url) {
-                                        $pool = new HttpRequestPool();
+                $url = filter_var($url, FILTER_SANITIZE_URL);
-                                        foreach ($subset as $orig => $url) {
+                $test = filter_var($url, FILTER_VALIDATE_URL, FILTER_FLAG_SCHEME_REQUIRED);
-                                                if (!$isRedirect) $orig = $url;
+                // deal with bug http://bugs.php.net/51192 (present in PHP 5.2.13 and PHP 5.3.2)
-                                                unset($this->redirectQueue[$orig]);
+                if ($test === false) {
-                                                $this->debug("...$url");
+                        $test = filter_var(strtr($url, '-', '_'), FILTER_VALIDATE_URL, FILTER_FLAG_SCHEME_REQUIRED);
-                                                if (!$isRedirect && isset($this->requests[$url])) {
+                }
-                                                        $this->debug("......in memory");
+                if ($test !== false && $test !== null && preg_match('!^https?://!', $url)) {
-                                                /*
+                        return $url;
-                                                } elseif ($this->isCached($url)) {
+                } else {
-                                                        $this->debug("......is cached");
+                        return false;
-                                                        if (!$this->minimiseMemoryUse) {
+                }
-                                                                $this->requests[$url] = $this->getCached($url);
+        }
-                                                        }
+        
-                                                */
+        public function fetchAll(array $urls) {
-                                                } else {
+                $this->fetchAllOnce($urls, $isRedirect=false);
-                                                        $this->debug("......adding to pool");
+                $redirects = 0;
-                                                        $req_url = $this->rewriteUrls($url);
+                while (!empty($this->redirectQueue) && ++$redirects <= $this->maxRedirects) {
-                                                        $req_url = ($this->rewriteHashbangFragment) ? $this->rewriteHashbangFragment($req_url) : $req_url;
+                        $this->debug("Following redirects #$redirects...");
-                                                        $req_url = $this->removeFragment($req_url);
+                        $this->fetchAllOnce($this->redirectQueue, $isRedirect=true);
-                                                        if (!empty($this->headerOnlyTypes) && !isset($this->requests[$orig]['wrongGuess']) && $this->possibleUnsupportedType($req_url)) {
+                }
-                                                                $_meth = HttpRequest::METH_HEAD;
+        }
-                                                        } else {
+        
-                                                                $_meth = HttpRequest::METH_GET;
+        // fetch all URLs without following redirects
-                                                                unset($this->requests[$orig]['wrongGuess']);
+        public function fetchAllOnce(array $urls, $isRedirect=false) {
-                                                        }
+                if (!$isRedirect) $urls = array_unique($urls);
-                                                        $httpRequest = new HttpRequest($req_url, $_meth, $this->requestOptions);
+                if (empty($urls)) return;
-                                                        // send cookies, if we have any
+                
-                                                        if ($cookies = $this->cookieJar->getMatchingCookies($req_url)) {
+                //////////////////////////////////////////////////////
-                                                                $this->debug("......sending cookies: $cookies");
+                // parallel (HttpRequestPool)
-                                                                $httpRequest->addHeaders(array('Cookie' => $cookies));
+                if ($this->method == self::METHOD_REQUEST_POOL) {
-                                                        }
+                        $this->debug('Starting parallel fetch (HttpRequestPool)');
-                                                        //$httpRequest->addHeaders(array('User-Agent' => $this->userAgent));
+                        try {
-                                                        $httpRequest->addHeaders($this->getUserAgent($req_url, true));
+                                while (count($urls) > 0) {
-                                                        // add referer for picky sites
+                                        $this->debug('Processing set of '.min($this->maxParallelRequests, count($urls)));
-                                                        $httpRequest->addheaders(array('Referer' => $this->referer));
+                                        $subset = array_splice($urls, 0, $this->maxParallelRequests);
-                                                        $this->requests[$orig] = array('headers'=>null, 'body'=>null, 'httpRequest'=>$httpRequest);
+                                        $pool = new HttpRequestPool();
-                                                        $this->requests[$orig]['original_url'] = $orig;
+                                        foreach ($subset as $orig => $url) {
-                                                        $pool->attach($httpRequest);
+                                                if (!$isRedirect) $orig = $url;
-                                                }
+                                                unset($this->redirectQueue[$orig]);
-                                        }
+                                                $this->debug("...$url");
-                                        // did we get anything into the pool?
+                                                if (!$isRedirect && isset($this->requests[$url])) {
-                                        if (count($pool) > 0) {
+                                                        $this->debug("......in memory");
-                                                $this->debug('Sending request...');
+                                                /*
-                                                try {
+                                                } elseif ($this->isCached($url)) {
-                                                        $pool->send();
+                                                        $this->debug("......is cached");
-                                                } catch (HttpRequestPoolException $e) {
+                                                        if (!$this->minimiseMemoryUse) {
-                                                        // do nothing
+                                                                $this->requests[$url] = $this->getCached($url);
-                                                }
+                                                        }
-                                                $this->debug('Received responses');
+                                                */
-                                                foreach($subset as $orig => $url) {
+                                                } else {
-                                                        if (!$isRedirect) $orig = $url;
+                                                        $this->debug("......adding to pool");
-                                                        $request = $this->requests[$orig]['httpRequest'];
+                                                        $req_url = $this->rewriteUrls($url);
-                                                        //$this->requests[$orig]['headers'] = $this->headersToString($request->getResponseHeader());
+                                                        $req_url = ($this->rewriteHashbangFragment) ? $this->rewriteHashbangFragment($req_url) : $req_url;
-                                                        // getResponseHeader() doesn't return status line, so, for consistency...
+                                                        $req_url = $this->removeFragment($req_url);
-                                                        $this->requests[$orig]['headers'] = substr($request->getRawResponseMessage(), 0, $request->getResponseInfo('header_size'));
+                                                        if (!empty($this->headerOnlyTypes) && !isset($this->requests[$orig]['wrongGuess']) && $this->possibleUnsupportedType($req_url)) {
-                                                        // check content type
+                                                                $_meth = HttpRequest::METH_HEAD;
-                                                        // TODO: use getResponseHeader('content-type') or getResponseInfo()
+                                                        } else {
-                                                        if ($this->headerOnlyType($this->requests[$orig]['headers'])) {
+                                                                $_meth = HttpRequest::METH_GET;
-                                                                $this->requests[$orig]['body'] = '';
+                                                                unset($this->requests[$orig]['wrongGuess']);
-                                                                $_header_only_type = true;
+                                                        }
-                                                                $this->debug('Header only type returned');
+                                                        $httpRequest = new HttpRequest($req_url, $_meth, $this->requestOptions);
-                                                        } else {
+                                                        // send cookies, if we have any
-                                                                $this->requests[$orig]['body'] = $request->getResponseBody();
+                                                        if ($cookies = $this->cookieJar->getMatchingCookies($req_url)) {
-                                                                $_header_only_type = false;
+                                                                $this->debug("......sending cookies: $cookies");
-                                                        }
+                                                                $httpRequest->addHeaders(array('Cookie' => $cookies));
-                                                        $this->requests[$orig]['effective_url'] = $request->getResponseInfo('effective_url');
+                                                        }
-                                                        $this->requests[$orig]['status_code'] = $status_code = $request->getResponseCode();
+                                                        //$httpRequest->addHeaders(array('User-Agent' => $this->userAgent));
-                                                        // is redirect?
+                                                        $httpRequest->addHeaders($this->getUserAgent($req_url, true));
-                                                        if ((in_array($status_code, array(300, 301, 302, 303, 307)) || $status_code > 307 && $status_code < 400) && $request->getResponseHeader('location')) {
+                                                        // add referer for picky sites
-                                                                $redirectURL = $request->getResponseHeader('location');
+                                                        $httpRequest->addheaders(array('Referer' => $this->referer));
-                                                                if (!preg_match('!^https?://!i', $redirectURL)) {
+                                                        $this->requests[$orig] = array('headers'=>null, 'body'=>null, 'httpRequest'=>$httpRequest);
-                                                                        $redirectURL = SimplePie_Misc::absolutize_url($redirectURL, $url);
+                                                        $this->requests[$orig]['original_url'] = $orig;
-                                                                }
+                                                        $pool->attach($httpRequest);
-                                                                if ($this->validateURL($redirectURL)) {
+                                                }
-                                                                        $this->debug('Redirect detected. Valid URL: '.$redirectURL);
+                                        }
-                                                                        // store any cookies
+                                        // did we get anything into the pool?
-                                                                        $cookies = $request->getResponseHeader('set-cookie');
+                                        if (count($pool) > 0) {
-                                                                        if ($cookies && !is_array($cookies)) $cookies = array($cookies);
+                                                $this->debug('Sending request...');
-                                                                        if ($cookies) $this->cookieJar->storeCookies($url, $cookies);
+                                                try {
-                                                                        $this->redirectQueue[$orig] = $redirectURL;
+                                                        $pool->send();
-                                                                } else {
+                                                } catch (HttpRequestPoolException $e) {
-                                                                        $this->debug('Redirect detected. Invalid URL: '.$redirectURL);
+                                                        // do nothing
-                                                                }
+                                                }
-                                                        } elseif (!$_header_only_type && $request->getMethod() === HttpRequest::METH_HEAD) {
+                                                $this->debug('Received responses');
-                                                                // the response content-type did not match our 'header only' types, 
+                                                foreach($subset as $orig => $url) {
-                                                                // but we'd issues a HEAD request because we assumed it would. So
+                                                        if (!$isRedirect) $orig = $url;
-                                                                // let's queue a proper GET request for this item...
+                                                        $request = $this->requests[$orig]['httpRequest'];
-                                                                $this->debug('Wrong guess at content-type, queing GET request');
+                                                        //$this->requests[$orig]['headers'] = $this->headersToString($request->getResponseHeader());
-                                                                $this->requests[$orig]['wrongGuess'] = true;
+                                                        // getResponseHeader() doesn't return status line, so, for consistency...
-                                                                $this->redirectQueue[$orig] = $this->requests[$orig]['effective_url'];
+                                                        $this->requests[$orig]['headers'] = substr($request->getRawResponseMessage(), 0, $request->getResponseInfo('header_size'));
-                                                        } elseif (strpos($this->requests[$orig]['effective_url'], '_escaped_fragment_') === false) {
+                                                        // check content type
-                                                                // check for <meta name='fragment' content='!'/>
+                                                        // TODO: use getResponseHeader('content-type') or getResponseInfo()
-                                                                // for AJAX sites, e.g. Blogger with its dynamic views templates.
+                                                        if ($this->headerOnlyType($this->requests[$orig]['headers'])) {
-                                                                // Based on Google's spec: https://developers.google.com/webmasters/ajax-crawling/docs/specification
+                                                                $this->requests[$orig]['body'] = '';
-                                                                if (isset($this->requests[$orig]['body'])) {
+                                                                $_header_only_type = true;
-                                                                        $redirectURL = $this->getUglyURL($this->requests[$orig]['effective_url'], substr($this->requests[$orig]['body'], 0, 4000));
+                                                                $this->debug('Header only type returned');
-                                                                        if ($redirectURL) {
+                                                        } else {
-                                                                                $this->debug('AJAX trigger (meta name="fragment" content="!") found. Queueing '.$redirectURL);
+                                                                $this->requests[$orig]['body'] = $request->getResponseBody();
-                                                                                $this->redirectQueue[$orig] = $redirectURL;
+                                                                $_header_only_type = false;
-                                                                        }
+                                                        }
-                                                                }
+                                                        $this->requests[$orig]['effective_url'] = $request->getResponseInfo('effective_url');
-                                                        }
+                                                        $this->requests[$orig]['status_code'] = $status_code = $request->getResponseCode();
-                                                        //die($url.' -multi- '.$request->getResponseInfo('effective_url'));
+                                                        // is redirect?
-                                                        $pool->detach($request);
+                                                        if ((in_array($status_code, array(300, 301, 302, 303, 307)) || $status_code > 307 && $status_code < 400) && $request->getResponseHeader('location')) {
-                                                        unset($this->requests[$orig]['httpRequest'], $request);
+                                                                $redirectURL = $request->getResponseHeader('location');
-                                                        /*
+                                                                if (!preg_match('!^https?://!i', $redirectURL)) {
-                                                        if ($this->minimiseMemoryUse) {
+                                                                        $redirectURL = SimplePie_Misc::absolutize_url($redirectURL, $url);
-                                                                if ($this->cache($url)) {
+                                                                }
-                                                                        unset($this->requests[$url]);
+                                                                if ($this->validateURL($redirectURL)) {
-                                                                }
+                                                                        $this->debug('Redirect detected. Valid URL: '.$redirectURL);
-                                                        }
+                                                                        // store any cookies
-                                                        */
+                                                                        $cookies = $request->getResponseHeader('set-cookie');
-                                                }
+                                                                        if ($cookies && !is_array($cookies)) $cookies = array($cookies);
-                                        }
+                                                                        if ($cookies) $this->cookieJar->storeCookies($url, $cookies);
-                                }
+                                                                        $this->redirectQueue[$orig] = $redirectURL;
-                        } catch (HttpException $e) {
+                                                                } else {
-                                $this->debug($e);
+                                                                        $this->debug('Redirect detected. Invalid URL: '.$redirectURL);
-                                return false;
+                                                                }
-                        }
+                                                        } elseif (!$_header_only_type && $request->getMethod() === HttpRequest::METH_HEAD) {
-                }
+                                                                // the response content-type did not match our 'header only' types, 
-                
+                                                                // but we'd issues a HEAD request because we assumed it would. So
-                //////////////////////////////////////////////////////////
+                                                                // let's queue a proper GET request for this item...
-                // parallel (curl_multi_*)
+                                                                $this->debug('Wrong guess at content-type, queing GET request');
-                elseif ($this->method == self::METHOD_CURL_MULTI) {
+                                                                $this->requests[$orig]['wrongGuess'] = true;
-                        $this->debug('Starting parallel fetch (curl_multi_*)');
+                                                                $this->redirectQueue[$orig] = $this->requests[$orig]['effective_url'];
-                        while (count($urls) > 0) {
+                                                        } elseif (strpos($this->requests[$orig]['effective_url'], '_escaped_fragment_') === false) {
-                                $this->debug('Processing set of '.min($this->maxParallelRequests, count($urls)));
+                                                                // check for <meta name='fragment' content='!'/>
-                                $subset = array_splice($urls, 0, $this->maxParallelRequests);
+                                                                // for AJAX sites, e.g. Blogger with its dynamic views templates.
-                                $pool = new RollingCurl(array($this, 'handleCurlResponse'));
+                                                                // Based on Google's spec: https://developers.google.com/webmasters/ajax-crawling/docs/specification
-                                $pool->window_size = count($subset);            
+                                                                if (isset($this->requests[$orig]['body'])) {
-                                
+                                                                        $redirectURL = $this->getRedirectURLfromHTML($this->requests[$orig]['effective_url'], substr($this->requests[$orig]['body'], 0, 4000));
-                                foreach ($subset as $orig => $url) {
+                                                                        if ($redirectURL) {
-                                        if (!$isRedirect) $orig = $url;
+                                                                                $this->redirectQueue[$orig] = $redirectURL;
-                                        unset($this->redirectQueue[$orig]);
+                                                                        }
-                                        $this->debug("...$url");
+                                                                }
-                                        if (!$isRedirect && isset($this->requests[$url])) {
+                                                        }
-                                                $this->debug("......in memory");
+                                                        //die($url.' -multi- '.$request->getResponseInfo('effective_url'));
-                                        /*
+                                                        $pool->detach($request);
-                                        } elseif ($this->isCached($url)) {
+                                                        unset($this->requests[$orig]['httpRequest'], $request);
-                                                $this->debug("......is cached");
+                                                        /*
-                                                if (!$this->minimiseMemoryUse) {
+                                                        if ($this->minimiseMemoryUse) {
-                                                        $this->requests[$url] = $this->getCached($url);
+                                                                if ($this->cache($url)) {
-                                                }
+                                                                        unset($this->requests[$url]);
-                                        */
+                                                                }
-                                        } else {
+                                                        }
-                                                $this->debug("......adding to pool");
+                                                        */
-                                                $req_url = $this->rewriteUrls($url);
+                                                }
-                                                $req_url = ($this->rewriteHashbangFragment) ? $this->rewriteHashbangFragment($req_url) : $req_url;
+                                        }
-                                                $req_url = $this->removeFragment($req_url);
+                                }
-                                                if (!empty($this->headerOnlyTypes) && !isset($this->requests[$orig]['wrongGuess']) && $this->possibleUnsupportedType($req_url)) {
+                        } catch (HttpException $e) {
-                                                        $_meth = 'HEAD';
+                                $this->debug($e);
-                                                } else {
+                                return false;
-                                                        $_meth = 'GET';
+                        }
-                                                        unset($this->requests[$orig]['wrongGuess']);
+                }
-                                                }                                               
+                
-                                                $headers = array();
+                //////////////////////////////////////////////////////////
-                                                //$headers[] = 'User-Agent: '.$this->userAgent;
+                // parallel (curl_multi_*)
-                                                $headers[] = $this->getUserAgent($req_url);
+                elseif ($this->method == self::METHOD_CURL_MULTI) {
-                                                // add referer for picky sites
+                        $this->debug('Starting parallel fetch (curl_multi_*)');
-                                                $headers[] = 'Referer: '.$this->referer;
+                        while (count($urls) > 0) {
-                                                // send cookies, if we have any
+                                $this->debug('Processing set of '.min($this->maxParallelRequests, count($urls)));
-                                                if ($cookies = $this->cookieJar->getMatchingCookies($req_url)) {
+                                $subset = array_splice($urls, 0, $this->maxParallelRequests);
-                                                        $this->debug("......sending cookies: $cookies");
+                                $pool = new RollingCurl(array($this, 'handleCurlResponse'));
-                                                        $headers[] = 'Cookie: '.$cookies;
+                                $pool->window_size = count($subset);            
-                                                }
+                                
-                                                $httpRequest = new RollingCurlRequest($req_url, $_meth, null, $headers, array(
+                                foreach ($subset as $orig => $url) {
-                                                        CURLOPT_CONNECTTIMEOUT => $this->requestOptions['timeout'],
+                                        if (!$isRedirect) $orig = $url;
-                                                        CURLOPT_TIMEOUT => $this->requestOptions['timeout']
+                                        unset($this->redirectQueue[$orig]);
-                                                        ));
+                                        $this->debug("...$url");
-                                                $httpRequest->set_original_url($orig);
+                                        if (!$isRedirect && isset($this->requests[$url])) {
-                                                $this->requests[$orig] = array('headers'=>null, 'body'=>null, 'httpRequest'=>$httpRequest);
+                                                $this->debug("......in memory");
-                                                $this->requests[$orig]['original_url'] = $orig; // TODO: is this needed anymore?
+                                        /*
-                                                $pool->add($httpRequest);
+                                        } elseif ($this->isCached($url)) {
-                                        }
+                                                $this->debug("......is cached");
-                                }
+                                                if (!$this->minimiseMemoryUse) {
-                                // did we get anything into the pool?
+                                                        $this->requests[$url] = $this->getCached($url);
-                                if (count($pool) > 0) {
+                                                }
-                                        $this->debug('Sending request...');
+                                        */
-                                        $pool->execute(); // this will call handleCurlResponse() and populate $this->requests[$orig]
+                                        } else {
-                                        $this->debug('Received responses');
+                                                $this->debug("......adding to pool");
-                                        foreach($subset as $orig => $url) {
+                                                $req_url = $this->rewriteUrls($url);
-                                                if (!$isRedirect) $orig = $url;
+                                                $req_url = ($this->rewriteHashbangFragment) ? $this->rewriteHashbangFragment($req_url) : $req_url;
-                                                // $this->requests[$orig]['headers']
+                                                $req_url = $this->removeFragment($req_url);
-                                                // $this->requests[$orig]['body']
+                                                if (!empty($this->headerOnlyTypes) && !isset($this->requests[$orig]['wrongGuess']) && $this->possibleUnsupportedType($req_url)) {
-                                                // $this->requests[$orig]['effective_url']
+                                                        $_meth = 'HEAD';
-                                                // check content type
+                                                } else {
-                                                if ($this->headerOnlyType($this->requests[$orig]['headers'])) {
+                                                        $_meth = 'GET';
-                                                        $this->requests[$orig]['body'] = '';
+                                                        unset($this->requests[$orig]['wrongGuess']);
-                                                        $_header_only_type = true;
+                                                }                                               
-                                                        $this->debug('Header only type returned');
+                                                $headers = array();
-                                                } else {
+                                                //$headers[] = 'User-Agent: '.$this->userAgent;
-                                                        $_header_only_type = false;
+                                                $headers[] = $this->getUserAgent($req_url);
-                                                }
+                                                // add referer for picky sites
-                                                $status_code = $this->requests[$orig]['status_code'];
+                                                $headers[] = 'Referer: '.$this->referer;
-                                                if ((in_array($status_code, array(300, 301, 302, 303, 307)) || $status_code > 307 && $status_code < 400) && isset($this->requests[$orig]['location'])) {
+                                                // send cookies, if we have any
-                                                        $redirectURL = $this->requests[$orig]['location'];
+                                                if ($cookies = $this->cookieJar->getMatchingCookies($req_url)) {
-                                                        if (!preg_match('!^https?://!i', $redirectURL)) {
+                                                        $this->debug("......sending cookies: $cookies");
-                                                                $redirectURL = SimplePie_Misc::absolutize_url($redirectURL, $url);
+                                                        $headers[] = 'Cookie: '.$cookies;
-                                                        }
+                                                }
-                                                        if ($this->validateURL($redirectURL)) {
+                                                $httpRequest = new RollingCurlRequest($req_url, $_meth, null, $headers, array(
-                                                                $this->debug('Redirect detected. Valid URL: '.$redirectURL);
+                                                        CURLOPT_CONNECTTIMEOUT => $this->requestOptions['timeout'],
-                                                                // store any cookies
+                                                        CURLOPT_TIMEOUT => $this->requestOptions['timeout']
-                                                                $cookies = $this->cookieJar->extractCookies($this->requests[$orig]['headers']);
+                                                        ));
-                                                                if (!empty($cookies)) $this->cookieJar->storeCookies($url, $cookies);                                                   
+                                                $httpRequest->set_original_url($orig);
-                                                                $this->redirectQueue[$orig] = $redirectURL;
+                                                $this->requests[$orig] = array('headers'=>null, 'body'=>null, 'httpRequest'=>$httpRequest);
-                                                        } else {
+                                                $this->requests[$orig]['original_url'] = $orig; // TODO: is this needed anymore?
-                                                                $this->debug('Redirect detected. Invalid URL: '.$redirectURL);
+                                                $pool->add($httpRequest);
-                                                        }
+                                        }
-                                                } elseif (!$_header_only_type && $this->requests[$orig]['method'] == 'HEAD') {
+                                }
-                                                        // the response content-type did not match our 'header only' types, 
+                                // did we get anything into the pool?
-                                                        // but we'd issues a HEAD request because we assumed it would. So
+                                if (count($pool) > 0) {
-                                                        // let's queue a proper GET request for this item...
+                                        $this->debug('Sending request...');
-                                                        $this->debug('Wrong guess at content-type, queing GET request');
+                                        $pool->execute(); // this will call handleCurlResponse() and populate $this->requests[$orig]
-                                                        $this->requests[$orig]['wrongGuess'] = true;
+                                        $this->debug('Received responses');
-                                                        $this->redirectQueue[$orig] = $this->requests[$orig]['effective_url'];
+                                        foreach($subset as $orig => $url) {
-                                                } elseif (strpos($this->requests[$orig]['effective_url'], '_escaped_fragment_') === false) {
+                                                if (!$isRedirect) $orig = $url;
-                                                        // check for <meta name='fragment' content='!'/>
+                                                // $this->requests[$orig]['headers']
-                                                        // for AJAX sites, e.g. Blogger with its dynamic views templates.
+                                                // $this->requests[$orig]['body']
-                                                        // Based on Google's spec: https://developers.google.com/webmasters/ajax-crawling/docs/specification
+                                                // $this->requests[$orig]['effective_url']
-                                                        if (isset($this->requests[$orig]['body'])) {
+                                                // check content type
-                                                                $redirectURL = $this->getUglyURL($this->requests[$orig]['effective_url'], substr($this->requests[$orig]['body'], 0, 4000));
+                                                if ($this->headerOnlyType($this->requests[$orig]['headers'])) {
-                                                                if ($redirectURL) {
+                                                        $this->requests[$orig]['body'] = '';
-                                                                        $this->debug('AJAX trigger (meta name="fragment" content="!") found. Queueing '.$redirectURL);
+                                                        $_header_only_type = true;
-                                                                        $this->redirectQueue[$orig] = $redirectURL;
+                                                        $this->debug('Header only type returned');
-                                                                }
+                                                } else {
-                                                        }
+                                                        $_header_only_type = false;
-                                                }
+                                                }
-                                                // die($url.' -multi- '.$request->getResponseInfo('effective_url'));
+                                                $status_code = $this->requests[$orig]['status_code'];
-                                                unset($this->requests[$orig]['httpRequest'], $this->requests[$orig]['method']);
+                                                if ((in_array($status_code, array(300, 301, 302, 303, 307)) || $status_code > 307 && $status_code < 400) && isset($this->requests[$orig]['location'])) {
-                                        }
+                                                        $redirectURL = $this->requests[$orig]['location'];
-                                }
+                                                        if (!preg_match('!^https?://!i', $redirectURL)) {
-                        }
+                                                                $redirectURL = SimplePie_Misc::absolutize_url($redirectURL, $url);
-                }
+                                                        }
+                                                        if ($this->validateURL($redirectURL)) {
-                //////////////////////////////////////////////////////
+                                                                $this->debug('Redirect detected. Valid URL: '.$redirectURL);
-                // sequential (file_get_contents)
+                                                                // store any cookies
-                else {
+                                                                $cookies = $this->cookieJar->extractCookies($this->requests[$orig]['headers']);
-                        $this->debug('Starting sequential fetch (file_get_contents)');
+                                                                if (!empty($cookies)) $this->cookieJar->storeCookies($url, $cookies);                                                   
-                        $this->debug('Processing set of '.count($urls));
+                                                                $this->redirectQueue[$orig] = $redirectURL;
-                        foreach ($urls as $orig => $url) {
+                                                        } else {
-                                if (!$isRedirect) $orig = $url;
+                                                                $this->debug('Redirect detected. Invalid URL: '.$redirectURL);
-                                unset($this->redirectQueue[$orig]);
+                                                        }
-                                $this->debug("...$url");
+                                                } elseif (!$_header_only_type && $this->requests[$orig]['method'] == 'HEAD') {
-                                if (!$isRedirect && isset($this->requests[$url])) {
+                                                        // the response content-type did not match our 'header only' types, 
-                                        $this->debug("......in memory");
+                                                        // but we'd issues a HEAD request because we assumed it would. So
-                                /*
+                                                        // let's queue a proper GET request for this item...
-                                } elseif ($this->isCached($url)) {
+                                                        $this->debug('Wrong guess at content-type, queing GET request');
-                                        $this->debug("......is cached");
+                                                        $this->requests[$orig]['wrongGuess'] = true;
-                                        if (!$this->minimiseMemoryUse) {
+                                                        $this->redirectQueue[$orig] = $this->requests[$orig]['effective_url'];
-                                                $this->requests[$url] = $this->getCached($url);
+                                                } elseif (strpos($this->requests[$orig]['effective_url'], '_escaped_fragment_') === false) {
-                                        }
+                                                        // check for <meta name='fragment' content='!'/>
-                                */
+                                                        // for AJAX sites, e.g. Blogger with its dynamic views templates.
-                                } else {
+                                                        // Based on Google's spec: https://developers.google.com/webmasters/ajax-crawling/docs/specification
-                                        $this->debug("Sending request for $url");
+                                                        if (isset($this->requests[$orig]['body'])) {
-                                        $this->requests[$orig]['original_url'] = $orig;
+                                                                $redirectURL = $this->getRedirectURLfromHTML($this->requests[$orig]['effective_url'], substr($this->requests[$orig]['body'], 0, 4000));
-                                        $req_url = $this->rewriteUrls($url);
+                                                                if ($redirectURL) {
-                                        $req_url = ($this->rewriteHashbangFragment) ? $this->rewriteHashbangFragment($req_url) : $req_url;
+                                                                        $this->redirectQueue[$orig] = $redirectURL;
-                                        $req_url = $this->removeFragment($req_url);
+                                                                }
-                                        // send cookies, if we have any
+                                                        }
-                                        $httpContext = $this->httpContext;
+                                                }
-                                        $httpContext['http']['header'] .= $this->getUserAgent($req_url)."\r\n";
+                                                // die($url.' -multi- '.$request->getResponseInfo('effective_url'));
-                                        // add referer for picky sites
+                                                unset($this->requests[$orig]['httpRequest'], $this->requests[$orig]['method']);
-                                        $httpContext['http']['header'] .= 'Referer: '.$this->referer."\r\n";
+                                        }
-                                        if ($cookies = $this->cookieJar->getMatchingCookies($req_url)) {
+                                }
-                                                $this->debug("......sending cookies: $cookies");
+                        }
-                                                $httpContext['http']['header'] .= 'Cookie: '.$cookies."\r\n";
+                }
-                                        }
-                                        if (false !== ($html = @file_get_contents($req_url, false, stream_context_create($httpContext)))) {
+                //////////////////////////////////////////////////////
-                                                $this->debug('Received response');
+                // sequential (file_get_contents)
-                                                // get status code
+                else {
-                                                if (!isset($http_response_header[0]) || !preg_match('!^HTTP/\d+\.\d+\s+(\d+)!', trim($http_response_header[0]), $match)) {
+                        $this->debug('Starting sequential fetch (file_get_contents)');
-                                                        $this->debug('Error: no status code found');
+                        $this->debug('Processing set of '.count($urls));
-                                                        // TODO: handle error - no status code
+                        foreach ($urls as $orig => $url) {
-                                                } else {
+                                if (!$isRedirect) $orig = $url;
-                                                        $this->requests[$orig]['headers'] = $this->headersToString($http_response_header, false);
+                                unset($this->redirectQueue[$orig]);
-                                                        // check content type
+                                $this->debug("...$url");
-                                                        if ($this->headerOnlyType($this->requests[$orig]['headers'])) {
+                                if (!$isRedirect && isset($this->requests[$url])) {
-                                                                $this->requests[$orig]['body'] = '';
+                                        $this->debug("......in memory");
-                                                        } else {
+                                /*
-                                                                $this->requests[$orig]['body'] = $html;
+                                } elseif ($this->isCached($url)) {
-                                                        }
+                                        $this->debug("......is cached");
-                                                        $this->requests[$orig]['effective_url'] = $req_url;
+                                        if (!$this->minimiseMemoryUse) {
-                                                        $this->requests[$orig]['status_code'] = $status_code = (int)$match[1];
+                                                $this->requests[$url] = $this->getCached($url);
-                                                        unset($match);
+                                        }
-                                                        // handle redirect
+                                */
-                                                        if (preg_match('/^Location:(.*?)$/mi', $this->requests[$orig]['headers'], $match)) {
+                                } else {
-                                                                $this->requests[$orig]['location'] =  trim($match[1]);
+                                        $this->debug("Sending request for $url");
-                                                        }
+                                        $this->requests[$orig]['original_url'] = $orig;
-                                                        if ((in_array($status_code, array(300, 301, 302, 303, 307)) || $status_code > 307 && $status_code < 400) && isset($this->requests[$orig]['location'])) {
+                                        $req_url = $this->rewriteUrls($url);
-                                                                $redirectURL = $this->requests[$orig]['location'];
+                                        $req_url = ($this->rewriteHashbangFragment) ? $this->rewriteHashbangFragment($req_url) : $req_url;
-                                                                if (!preg_match('!^https?://!i', $redirectURL)) {
+                                        $req_url = $this->removeFragment($req_url);
-                                                                        $redirectURL = SimplePie_Misc::absolutize_url($redirectURL, $url);
+                                        // send cookies, if we have any
-                                                                }
+                                        $httpContext = $this->httpContext;
-                                                                if ($this->validateURL($redirectURL)) {
+                                        $httpContext['http']['header'] .= $this->getUserAgent($req_url)."\r\n";
-                                                                        $this->debug('Redirect detected. Valid URL: '.$redirectURL);
+                                        // add referer for picky sites
-                                                                        // store any cookies
+                                        $httpContext['http']['header'] .= 'Referer: '.$this->referer."\r\n";
-                                                                        $cookies = $this->cookieJar->extractCookies($this->requests[$orig]['headers']);
+                                        if ($cookies = $this->cookieJar->getMatchingCookies($req_url)) {
-                                                                        if (!empty($cookies)) $this->cookieJar->storeCookies($url, $cookies);
+                                                $this->debug("......sending cookies: $cookies");
-                                                                        $this->redirectQueue[$orig] = $redirectURL;
+                                                $httpContext['http']['header'] .= 'Cookie: '.$cookies."\r\n";
-                                                                } else {
+                                        }
-                                                                        $this->debug('Redirect detected. Invalid URL: '.$redirectURL);
+                                        if (false !== ($html = @file_get_contents($req_url, false, stream_context_create($httpContext)))) {
-                                                                }
+                                                $this->debug('Received response');
-                                                        } elseif (strpos($this->requests[$orig]['effective_url'], '_escaped_fragment_') === false) {
+                                                // get status code
-                                                                // check for <meta name='fragment' content='!'/>
+                                                if (!isset($http_response_header[0]) || !preg_match('!^HTTP/\d+\.\d+\s+(\d+)!', trim($http_response_header[0]), $match)) {
-                                                                // for AJAX sites, e.g. Blogger with its dynamic views templates.
+                                                        $this->debug('Error: no status code found');
-                                                                // Based on Google's spec: https://developers.google.com/webmasters/ajax-crawling/docs/specification
+                                                        // TODO: handle error - no status code
-                                                                if (isset($this->requests[$orig]['body'])) {
+                                                } else {
-                                                                        $redirectURL = $this->getUglyURL($this->requests[$orig]['effective_url'], substr($this->requests[$orig]['body'], 0, 4000));
+                                                        $this->requests[$orig]['headers'] = $this->headersToString($http_response_header, false);
-                                                                        if ($redirectURL) {
+                                                        // check content type
-                                                                                $this->debug('AJAX trigger (meta name="fragment" content="!") found. Queueing '.$redirectURL);
+                                                        if ($this->headerOnlyType($this->requests[$orig]['headers'])) {
-                                                                                $this->redirectQueue[$orig] = $redirectURL;
+                                                                $this->requests[$orig]['body'] = '';
-                                                                        }
+                                                        } else {
-                                                                }
+                                                                $this->requests[$orig]['body'] = $html;
-                                                        }
+                                                        }
-                                                }
+                                                        $this->requests[$orig]['effective_url'] = $req_url;
-                                        } else {
+                                                        $this->requests[$orig]['status_code'] = $status_code = (int)$match[1];
-                                                $this->debug('Error retrieving URL');
+                                                        unset($match);
-                                                //print_r($req_url);
+                                                        // handle redirect
-                                                //print_r($http_response_header);
+                                                        if (preg_match('/^Location:(.*?)$/mi', $this->requests[$orig]['headers'], $match)) {
-                                                //print_r($html);
+                                                                $this->requests[$orig]['location'] =  trim($match[1]);
-                                                
+                                                        }
-                                                // TODO: handle error - failed to retrieve URL
+                                                        if ((in_array($status_code, array(300, 301, 302, 303, 307)) || $status_code > 307 && $status_code < 400) && isset($this->requests[$orig]['location'])) {
-                                        }
+                                                                $redirectURL = $this->requests[$orig]['location'];
-                                }
+                                                                if (!preg_match('!^https?://!i', $redirectURL)) {
-                        }
+                                                                        $redirectURL = SimplePie_Misc::absolutize_url($redirectURL, $url);
-                }
+                                                                }
-        }
+                                                                if ($this->validateURL($redirectURL)) {
-        
+                                                                        $this->debug('Redirect detected. Valid URL: '.$redirectURL);
-        public function handleCurlResponse($response, $info, $request) {
+                                                                        // store any cookies
-                $orig = $request->url_original;
+                                                                        $cookies = $this->cookieJar->extractCookies($this->requests[$orig]['headers']);
-                $this->requests[$orig]['headers'] = substr($response, 0, $info['header_size']);
+                                                                        if (!empty($cookies)) $this->cookieJar->storeCookies($url, $cookies);
-                $this->requests[$orig]['body'] = substr($response, $info['header_size']);
+                                                                        $this->redirectQueue[$orig] = $redirectURL;
-                $this->requests[$orig]['method'] = $request->method;
+                                                                } else {
-                $this->requests[$orig]['effective_url'] = $info['url'];
+                                                                        $this->debug('Redirect detected. Invalid URL: '.$redirectURL);
-                $this->requests[$orig]['status_code'] = (int)$info['http_code'];
+                                                                }
-                if (preg_match('/^Location:(.*?)$/mi', $this->requests[$orig]['headers'], $match)) {
+                                                        } elseif (strpos($this->requests[$orig]['effective_url'], '_escaped_fragment_') === false) {
-                        $this->requests[$orig]['location'] =  trim($match[1]);
+                                                                // check for <meta name='fragment' content='!'/>
-                }
+                                                                // for AJAX sites, e.g. Blogger with its dynamic views templates.
-        }
+                                                                // Based on Google's spec: https://developers.google.com/webmasters/ajax-crawling/docs/specification
-        
+                                                                if (isset($this->requests[$orig]['body'])) {
-        protected function headersToString(array $headers, $associative=true) {
+                                                                        $redirectURL = $this->getRedirectURLfromHTML($this->requests[$orig]['effective_url'], substr($this->requests[$orig]['body'], 0, 4000));
-                if (!$associative) {
+                                                                        if ($redirectURL) {
-                        return implode("\n", $headers);
+                                                                                $this->redirectQueue[$orig] = $redirectURL;
-                } else {
+                                                                        }
-                        $str = '';
+                                                                }
-                        foreach ($headers as $key => $val) {
+                                                        }
-                                if (is_array($val)) {
+                                                }
-                                        foreach ($val as $v) $str .= "$key: $v\n";
+                                        } else {
-                                } else {
+                                                $this->debug('Error retrieving URL');
-                                        $str .= "$key: $val\n";
+                                                //print_r($req_url);
-                                }
+                                                //print_r($http_response_header);
-                        }
+                                                //print_r($html);
-                        return rtrim($str);
+                                                
-                }
+                                                // TODO: handle error - failed to retrieve URL
-        }
+                                        }
-        
+                                }
-        public function get($url, $remove=false, $gzdecode=true) {
+                        }
-                $url = "$url";
+                }
-                if (isset($this->requests[$url]) && isset($this->requests[$url]['body'])) {
+        }
-                        $this->debug("URL already fetched - in memory ($url, effective: {$this->requests[$url]['effective_url']})");
+        
-                        $response = $this->requests[$url];
+        public function handleCurlResponse($response, $info, $request) {
-                /*
+                $orig = $request->url_original;
-                } elseif ($this->isCached($url)) {
+                $this->requests[$orig]['headers'] = substr($response, 0, $info['header_size']);
-                        $this->debug("URL already fetched - in disk cache ($url)");
+                $this->requests[$orig]['body'] = substr($response, $info['header_size']);
-                        $response = $this->getCached($url);
+                $this->requests[$orig]['method'] = $request->method;
-                        $this->requests[$url] = $response;
+                $this->requests[$orig]['effective_url'] = $info['url'];
-                */
+                $this->requests[$orig]['status_code'] = (int)$info['http_code'];
-                } else {
+                if (preg_match('/^Location:(.*?)$/mi', $this->requests[$orig]['headers'], $match)) {
-                        $this->debug("Fetching URL ($url)");
+                        $this->requests[$orig]['location'] =  trim($match[1]);
-                        $this->fetchAll(array($url));
+                }
-                        if (isset($this->requests[$url]) && isset($this->requests[$url]['body'])) {
+        }
-                                $response = $this->requests[$url];
+        
-                        } else {
+        protected function headersToString(array $headers, $associative=true) {
-                                $this->debug("Request failed");
+                if (!$associative) {
-                                $response = false;
+                        return implode("\n", $headers);
-                        }
+                } else {
-                }
+                        $str = '';
-                /*
+                        foreach ($headers as $key => $val) {
-                if ($this->minimiseMemoryUse && $response) {
+                                if (is_array($val)) {
-                        $this->cache($url);
+                                        foreach ($val as $v) $str .= "$key: $v\n";
-                        unset($this->requests[$url]);
+                                } else {
-                }
+                                        $str .= "$key: $val\n";
-                */
+                                }
-                if ($remove && $response) unset($this->requests[$url]);
+                        }
-                if ($gzdecode && stripos($response['headers'], 'Content-Encoding: gzip')) {
+                        return rtrim($str);
-                        if ($html = gzdecode($response['body'])) {
+                }
-                                $response['body'] = $html;
+        }
-                        }
+        
-                }
+        public function get($url, $remove=false, $gzdecode=true) {
-                return $response;
+                $url = "$url";
-        }
+                if (isset($this->requests[$url]) && isset($this->requests[$url]['body'])) {
-        
+                        $this->debug("URL already fetched - in memory ($url, effective: {$this->requests[$url]['effective_url']})");
-        public function parallelSupport() {
+                        $response = $this->requests[$url];
-                return class_exists('HttpRequestPool') || function_exists('curl_multi_init');
+                /*
-        }
+                } elseif ($this->isCached($url)) {
-        
+                        $this->debug("URL already fetched - in disk cache ($url)");
-        private function headerOnlyType($headers) {
+                        $response = $this->getCached($url);
-                if (preg_match('!^Content-Type:\s*(([a-z-]+)/([^;\r\n ]+))!im', $headers, $match)) {
+                        $this->requests[$url] = $response;
-                        // look for full mime type (e.g. image/jpeg) or just type (e.g. image)
+                */
-                        $match[1] = strtolower(trim($match[1]));
+                } else {
-                        $match[2] = strtolower(trim($match[2]));
+                        $this->debug("Fetching URL ($url)");
-                        foreach (array($match[1], $match[2]) as $mime) {
+                        $this->fetchAll(array($url));
-                                if (in_array($mime, $this->headerOnlyTypes)) return true;
+                        if (isset($this->requests[$url]) && isset($this->requests[$url]['body'])) {
-                        }
+                                $response = $this->requests[$url];
-                }
+                        } else {
-                return false;
+                                $this->debug("Request failed");
-        }
+                                $response = false;
-        
+                        }
-        private function possibleUnsupportedType($url) {
+                }
-                $path = @parse_url($url, PHP_URL_PATH);
+                /*
-                if ($path && strpos($path, '.') !== false) {
+                if ($this->minimiseMemoryUse && $response) {
-                        $ext = strtolower(trim(pathinfo($path, PATHINFO_EXTENSION)));
+                        $this->cache($url);
-                        return in_array($ext, $this->headerOnlyClues);
+                        unset($this->requests[$url]);
-                }
+                }
-                return false;
+                */
-        }
+                if ($remove && $response) unset($this->requests[$url]);
-}
+                if ($gzdecode && stripos($response['headers'], 'Content-Encoding: gzip')) {
+                        if ($html = gzdecode($response['body'])) {
-// gzdecode from http://www.php.net/manual/en/function.gzdecode.php#82930
+                                $response['body'] = $html;
-if (!function_exists('gzdecode')) {
+                        }
-        function gzdecode($data,&$filename='',&$error='',$maxlength=null) 
+                }
-        {
+                return $response;
-                $len = strlen($data);
+        }
-                if ($len < 18 || strcmp(substr($data,0,2),"\x1f\x8b")) {
+        
-                        $error = "Not in GZIP format.";
+        public function parallelSupport() {
-                        return null;  // Not GZIP format (See RFC 1952)
+                return class_exists('HttpRequestPool') || function_exists('curl_multi_init');
-                }
+        }
-                $method = ord(substr($data,2,1));  // Compression method
+        
-                $flags  = ord(substr($data,3,1));  // Flags
+        private function headerOnlyType($headers) {
-                if ($flags & 31 != $flags) {
+                if (preg_match('!^Content-Type:\s*(([a-z-]+)/([^;\r\n ]+))!im', $headers, $match)) {
-                        $error = "Reserved bits not allowed.";
+                        // look for full mime type (e.g. image/jpeg) or just type (e.g. image)
-                        return null;
+                        $match[1] = strtolower(trim($match[1]));
-                }
+                        $match[2] = strtolower(trim($match[2]));
-                // NOTE: $mtime may be negative (PHP integer limitations)
+                        foreach (array($match[1], $match[2]) as $mime) {
-                $mtime = unpack("V", substr($data,4,4));
+                                if (in_array($mime, $this->headerOnlyTypes)) return true;
-                $mtime = $mtime[1];
+                        }
-                $xfl   = substr($data,8,1);
+                }
-                $os    = substr($data,8,1);
+                return false;
-                $headerlen = 10;
+        }
-                $extralen  = 0;
+        
-                $extra     = "";
+        private function possibleUnsupportedType($url) {
-                if ($flags & 4) {
+                $path = @parse_url($url, PHP_URL_PATH);
-                        // 2-byte length prefixed EXTRA data in header
+                if ($path && strpos($path, '.') !== false) {
-                        if ($len - $headerlen - 2 < 8) {
+                        $ext = strtolower(trim(pathinfo($path, PATHINFO_EXTENSION)));
-                                return false;  // invalid
+                        return in_array($ext, $this->headerOnlyClues);
-                        }
+                }
-                        $extralen = unpack("v",substr($data,8,2));
+                return false;
-                        $extralen = $extralen[1];
+        }
-                        if ($len - $headerlen - 2 - $extralen < 8) {
+}
-                                return false;  // invalid
-                        }
+// gzdecode from http://www.php.net/manual/en/function.gzdecode.php#82930
-                        $extra = substr($data,10,$extralen);
+if (!function_exists('gzdecode')) {
-                        $headerlen += 2 + $extralen;
+        function gzdecode($data,&$filename='',&$error='',$maxlength=null) 
-                }
+        {
-                $filenamelen = 0;
+                $len = strlen($data);
-                $filename = "";
+                if ($len < 18 || strcmp(substr($data,0,2),"\x1f\x8b")) {
-                if ($flags & 8) {
+                        $error = "Not in GZIP format.";
-                        // C-style string
+                        return null;  // Not GZIP format (See RFC 1952)
-                        if ($len - $headerlen - 1 < 8) {
+                }
-                                return false; // invalid
+                $method = ord(substr($data,2,1));  // Compression method
-                        }
+                $flags  = ord(substr($data,3,1));  // Flags
-                        $filenamelen = strpos(substr($data,$headerlen),chr(0));
+                if ($flags & 31 != $flags) {
-                        if ($filenamelen === false || $len - $headerlen - $filenamelen - 1 < 8) {
+                        $error = "Reserved bits not allowed.";
-                                return false; // invalid
+                        return null;
-                        }
+                }
-                        $filename = substr($data,$headerlen,$filenamelen);
+                // NOTE: $mtime may be negative (PHP integer limitations)
-                        $headerlen += $filenamelen + 1;
+                $mtime = unpack("V", substr($data,4,4));
-                }
+                $mtime = $mtime[1];
-                $commentlen = 0;
+                $xfl   = substr($data,8,1);
-                $comment = "";
+                $os    = substr($data,8,1);
-                if ($flags & 16) {
+                $headerlen = 10;
-                        // C-style string COMMENT data in header
+                $extralen  = 0;
-                        if ($len - $headerlen - 1 < 8) {
+                $extra     = "";
-                                return false;    // invalid
+                if ($flags & 4) {
-                        }
+                        // 2-byte length prefixed EXTRA data in header
-                        $commentlen = strpos(substr($data,$headerlen),chr(0));
+                        if ($len - $headerlen - 2 < 8) {
-                        if ($commentlen === false || $len - $headerlen - $commentlen - 1 < 8) {
+                                return false;  // invalid
-                                return false;    // Invalid header format
+                        }
-                        }
+                        $extralen = unpack("v",substr($data,8,2));
-                        $comment = substr($data,$headerlen,$commentlen);
+                        $extralen = $extralen[1];
-                        $headerlen += $commentlen + 1;
+                        if ($len - $headerlen - 2 - $extralen < 8) {
-                }
+                                return false;  // invalid
-                $headercrc = "";
+                        }
-                if ($flags & 2) {
+                        $extra = substr($data,10,$extralen);
-                        // 2-bytes (lowest order) of CRC32 on header present
+                        $headerlen += 2 + $extralen;
-                        if ($len - $headerlen - 2 < 8) {
+                }
-                                return false;    // invalid
+                $filenamelen = 0;
-                        }
+                $filename = "";
-                        $calccrc = crc32(substr($data,0,$headerlen)) & 0xffff;
+                if ($flags & 8) {
-                        $headercrc = unpack("v", substr($data,$headerlen,2));
+                        // C-style string
-                        $headercrc = $headercrc[1];
+                        if ($len - $headerlen - 1 < 8) {
-                        if ($headercrc != $calccrc) {
+                                return false; // invalid
-                                $error = "Header checksum failed.";
+                        }
-                                return false;    // Bad header CRC
+                        $filenamelen = strpos(substr($data,$headerlen),chr(0));
-                        }
+                        if ($filenamelen === false || $len - $headerlen - $filenamelen - 1 < 8) {
-                        $headerlen += 2;
+                                return false; // invalid
-                }
+                        }
-                // GZIP FOOTER
+                        $filename = substr($data,$headerlen,$filenamelen);
-                $datacrc = unpack("V",substr($data,-8,4));
+                        $headerlen += $filenamelen + 1;
-                $datacrc = sprintf('%u',$datacrc[1] & 0xFFFFFFFF);
+                }
-                $isize = unpack("V",substr($data,-4));
+                $commentlen = 0;
-                $isize = $isize[1];
+                $comment = "";
-                // decompression:
+                if ($flags & 16) {
-                $bodylen = $len-$headerlen-8;
+                        // C-style string COMMENT data in header
-                if ($bodylen < 1) {
+                        if ($len - $headerlen - 1 < 8) {
-                        // IMPLEMENTATION BUG!
+                                return false;    // invalid
-                        return null;
+                        }
-                }
+                        $commentlen = strpos(substr($data,$headerlen),chr(0));
-                $body = substr($data,$headerlen,$bodylen);
+                        if ($commentlen === false || $len - $headerlen - $commentlen - 1 < 8) {
-                $data = "";
+                                return false;    // Invalid header format
-                if ($bodylen > 0) {
+                        }
-                        switch ($method) {
+                        $comment = substr($data,$headerlen,$commentlen);
-                        case 8:
+                        $headerlen += $commentlen + 1;
-                                // Currently the only supported compression method:
+                }
-                                $data = gzinflate($body,$maxlength);
+                $headercrc = "";
-                                break;
+                if ($flags & 2) {
-                        default:
+                        // 2-bytes (lowest order) of CRC32 on header present
-                                $error = "Unknown compression method.";
+                        if ($len - $headerlen - 2 < 8) {
-                                return false;
+                                return false;    // invalid
-                        }
+                        }
-                }  // zero-byte body content is allowed
+                        $calccrc = crc32(substr($data,0,$headerlen)) & 0xffff;
-                // Verifiy CRC32
+                        $headercrc = unpack("v", substr($data,$headerlen,2));
-                $crc   = sprintf("%u",crc32($data));
+                        $headercrc = $headercrc[1];
-                $crcOK = $crc == $datacrc;
+                        if ($headercrc != $calccrc) {
-                $lenOK = $isize == strlen($data);
+                                $error = "Header checksum failed.";
-                if (!$lenOK || !$crcOK) {
+                                return false;    // Bad header CRC
-                        $error = ( $lenOK ? '' : 'Length check FAILED. ') . ( $crcOK ? '' : 'Checksum FAILED.');
+                        }
-                        return false;
+                        $headerlen += 2;
-                }
+                }
-                return $data;
+                // GZIP FOOTER
-        }
+                $datacrc = unpack("V",substr($data,-8,4));
-}
+                $datacrc = sprintf('%u',$datacrc[1] & 0xFFFFFFFF);
-?>
-\ No newline at end of file
+                $isize = unpack("V",substr($data,-4));
+                $isize = $isize[1];
+                // decompression:
+                $bodylen = $len-$headerlen-8;
+                if ($bodylen < 1) {
+                        // IMPLEMENTATION BUG!
+                        return null;
+                }
+                $body = substr($data,$headerlen,$bodylen);
+                $data = "";
+                if ($bodylen > 0) {
+                        switch ($method) {
+                        case 8:
+                                // Currently the only supported compression method:
+                                $data = gzinflate($body,$maxlength);
+                                break;
+                        default:
+                                $error = "Unknown compression method.";
+                                return false;
+                        }
+                }  // zero-byte body content is allowed
+                // Verifiy CRC32
+                $crc   = sprintf("%u",crc32($data));
+                $crcOK = $crc == $datacrc;
+                $lenOK = $isize == strlen($data);
+                if (!$lenOK || !$crcOK) {
+                        $error = ( $lenOK ? '' : 'Length check FAILED. ') . ( $crcOK ? '' : 'Checksum FAILED.');
+                        return false;
+                }
+                return $data;
+        }
+}
+\ No newline at end of file