inc/3rdparty/humble-http-agent/CookieJar.php

   1 <?php
   2 /**
   3  * Cookie Jar
   4  *
   5  * PHP class for handling cookies, as defined by the Netscape spec:
   6  * <http://curl.haxx.se/rfc/cookie_spec.html>
   7  *
   8  * This class should be used to handle cookies (storing cookies from HTTP response messages, and
   9  * sending out cookies in HTTP request messages). This has been adapted for FiveFilters.org
  10  * from the original version used in HTTP Navigator. See http://www.keyvan.net/code/http-navigator/
  11  *
  12  * This class is mainly based on Cookies.pm <http://search.cpan.org/author/GAAS/libwww-perl-5.65/
  13  * lib/HTTP/Cookies.pm> from the libwww-perl collection <http://www.linpro.no/lwp/>.
  14  * Unlike Cookies.pm, this class only supports the Netscape cookie spec, not RFC 2965.
  15  *
  16  * @version 0.5
  17  * @date 2011-03-15
  18  * @see http://php.net/HttpRequestPool
  19  * @author Keyvan Minoukadeh
  20  * @copyright 2011 Keyvan Minoukadeh
  21  * @license http://www.gnu.org/licenses/agpl-3.0.html AGPL v3
  22  */
  23
  24 class CookieJar
  25 {
  26     /**
  27     * Cookies - array containing all cookies.
  28     *
  29     * <pre>
  30     * Cookies are stored like this:
  31     *   [domain][path][name] = array
  32     * where array is:
  33     *   0 => value, 1 => secure, 2 => expires
  34     * </pre>
  35     * @var array
  36     * @access private
  37     */
  38     public $cookies = array();
  39         public $debug = false;
  40
  41     /**
  42     * Constructor
  43     */
  44     function __construct() {
  45     }
  46
  47         protected function debug($msg, $file=null, $line=null) {
  48                 if ($this->debug) {
  49                         $mem = round(memory_get_usage()/1024, 2);
  50                         $memPeak = round(memory_get_peak_usage()/1024, 2);
  51                         echo '* ',$msg;
  52                         if (isset($file, $line)) echo " ($file line $line)";
  53                         echo ' - mem used: ',$mem," (peak: $memPeak)\n";
  54                         ob_flush();
  55                         flush();
  56                 }
  57         }
  58
  59     /**
  60     * Get matching cookies
  61     *
  62     * Only use this method if you cannot use add_cookie_header(), for example, if you want to use
  63     * this cookie jar class without using the request class.
  64     *
  65     * @param array $param associative array containing 'domain', 'path', 'secure' keys
  66     * @return string
  67     * @see add_cookie_header()
  68     */
  69     public function getMatchingCookies($url)
  70     {
  71                 if (($parts = @parse_url($url)) && isset($parts['scheme'], $parts['host'], $parts['path'])) {
  72                         $param['domain'] = $parts['host'];
  73                         $param['path'] = $parts['path'];
  74                         $param['secure'] = (strtolower($parts['scheme']) == 'https');
  75                         unset($parts);
  76                 } else {
  77                         return false;
  78                 }
  79         // RFC 2965 notes:
  80         //  If multiple cookies satisfy the criteria above, they are ordered in
  81         //  the Cookie header such that those with more specific Path attributes
  82         //  precede those with less specific.  Ordering with respect to other
  83         //  attributes (e.g., Domain) is unspecified.
  84         $domain = $param['domain'];
  85         if (strpos($domain, '.') === false) $domain .= '.local';
  86         $request_path = $param['path'];
  87         if ($request_path == '') $request_path = '/';
  88         $request_secure = $param['secure'];
  89         $now = time();
  90         $matched_cookies = array();
  91         // domain - find matching domains
  92         $this->debug('Finding matching domains for '.$domain, __FILE__, __LINE__);
  93         while (strpos($domain, '.') !== false) {
  94             if (isset($this->cookies[$domain])) {
  95                 $this->debug(' domain match found: '.$domain);
  96                 $cookies =& $this->cookies[$domain];
  97             } else {
  98                 $domain = $this->_reduce_domain($domain);
  99                 continue;
 100             }
 101             // paths - find matching paths starting from most specific
 102             $this->debug('  - Finding matching paths for '.$request_path);
 103             $paths = array_keys($cookies);
 104             usort($paths, array($this, '_cmp_length'));
 105             foreach ($paths as $path) {
 106                 // continue to next cookie if request path does not path-match cookie path
 107                 if (!$this->_path_match($request_path, $path)) continue;
 108                 // loop through cookie names
 109                 $this->debug('     path match found: '.$path);
 110                 foreach ($cookies[$path] as $name => $values) {
 111                     // if this cookie is secure but request isn't, continue to next cookie
 112                     if ($values[1] && !$request_secure) continue;
 113                     // if cookie is not a session cookie and has expired, continue to next cookie
 114                     if (is_int($values[2]) && ($values[2] < $now)) continue;
 115                     // cookie matches request
 116                     $this->debug('      cookie match: '.$name.'='.$values[0]);
 117                     $matched_cookies[] = $name.'='.$values[0];
 118                 }
 119             }
 120             $domain = $this->_reduce_domain($domain);
 121         }
 122         // return cookies
 123         return implode('; ', $matched_cookies);
 124     }
 125
 126     /**
 127     * Parse Set-Cookie values.
 128     *
 129     * Only use this method if you cannot use extract_cookies(), for example, if you want to use
 130     * this cookie jar class without using the response class.
 131     *
 132     * @param array $set_cookies array holding 1 or more "Set-Cookie" header values
 133     * @param array $param associative array containing 'host', 'path' keys
 134     * @return void
 135     * @see extract_cookies()
 136     */
 137     public function storeCookies($url, $set_cookies)
 138     {
 139         if (count($set_cookies) == 0) return;
 140                 $param = @parse_url($url);
 141                 if (!is_array($param) || !isset($param['host'])) return;
 142         $request_host = $param['host'];
 143         if (strpos($request_host, '.') === false) $request_host .= '.local';
 144         $request_path = @$param['path'];
 145         if ($request_path == '') $request_path = '/';
 146         //
 147         // loop through set-cookie headers
 148         //
 149         foreach ($set_cookies as $set_cookie) {
 150             $this->debug('Parsing: '.$set_cookie);
 151             // temporary cookie store (before adding to jar)
 152             $tmp_cookie = array();
 153             $param = explode(';', $set_cookie);
 154             // loop through params
 155             for ($x=0; $x<count($param); $x++) {
 156                 $key_val = explode('=', $param[$x], 2);
 157                 if (count($key_val) != 2) {
 158                     // if the first param isn't a name=value pair, continue to the next set-cookie
 159                     // header
 160                     if ($x == 0) continue 2;
 161                     // check for secure flag
 162                     if (strtolower(trim($key_val[0])) == 'secure') $tmp_cookie['secure'] = true;
 163                     // continue to next param
 164                     continue;
 165                 }
 166                 list($key, $val) = array_map('trim', $key_val);
 167                 // first name=value pair is the cookie name and value
 168                 // the name and value are stored under 'name' and 'value' to avoid conflicts
 169                 // with later parameters.
 170                 if ($x == 0) {
 171                     $tmp_cookie = array('name'=>$key, 'value'=>$val);
 172                     continue;
 173                 }
 174                 $key = strtolower($key);
 175                 if (in_array($key, array('expires', 'path', 'domain', 'secure'))) {
 176                     $tmp_cookie[$key] = $val;
 177                 }
 178             }
 179             //
 180             // set cookie
 181             //
 182             // check domain
 183             if (isset($tmp_cookie['domain']) && ($tmp_cookie['domain'] != $request_host) &&
 184                     ($tmp_cookie['domain'] != ".$request_host")) {
 185                 $domain = $tmp_cookie['domain'];
 186                 if ((strpos($domain, '.') === false) && ($domain != 'local')) {
 187                     $this->debug(' - domain "'.$domain.'" has no dot and is not a local domain');
 188                     continue;
 189                 }
 190                 if (preg_match('/\.[0-9]+$/', $domain)) {
 191                     $this->debug(' - domain "'.$domain.'" appears to be an ip address');
 192                     continue;
 193                 }
 194                 if (substr($domain, 0, 1) != '.') $domain = ".$domain";
 195                 if (!$this->_domain_match($request_host, $domain)) {
 196                     $this->debug(' - request host "'.$request_host.'" does not domain-match "'.$domain.'"');
 197                     continue;
 198                 }
 199             } else {
 200                 // if domain is not specified in the set-cookie header, domain will default to
 201                 // the request host
 202                 $domain = $request_host;
 203             }
 204             // check path
 205             if (isset($tmp_cookie['path']) && ($tmp_cookie['path'] != '')) {
 206                 $path = urldecode($tmp_cookie['path']);
 207                 if (!$this->_path_match($request_path, $path)) {
 208                     $this->debug(' - request path "'.$request_path.'" does not path-match "'.$path.'"');
 209                     continue;
 210                 }
 211             } else {
 212                 $path = $request_path;
 213                 $path = substr($path, 0, strrpos($path, '/'));
 214                 if ($path == '') $path = '/';
 215             }
 216             // check if secure
 217             $secure = (isset($tmp_cookie['secure'])) ? true : false;
 218             // check expiry
 219             if (isset($tmp_cookie['expires'])) {
 220                 if (($expires = strtotime($tmp_cookie['expires'])) < 0) {
 221                     $expires = null;
 222                 }
 223             } else {
 224                 $expires = null;
 225             }
 226             // set cookie
 227             $this->set_cookie($domain, $path, $tmp_cookie['name'], $tmp_cookie['value'], $secure, $expires);
 228         }
 229     }
 230
 231         // return array of set-cookie values extracted from HTTP response headers (string $h)
 232         public function extractCookies($h) {
 233         $x = 0;
 234         $lines = 0;
 235         $headers = array();
 236         $last_match = false;
 237                 $h = explode("\n", $h);
 238         foreach ($h as $line) {
 239                         $line = rtrim($line);
 240             $lines++;
 241
 242             $trimmed_line = trim($line);
 243             if (isset($line_last)) {
 244                 // check if we have \r\n\r\n (indicating the end of headers)
 245                 // some servers will not use CRLF (\r\n), so we make CR (\r) optional.
 246                 // if (preg_match('/\015?\012\015?\012/', $line_last.$line)) {
 247                 //     break;
 248                 // }
 249                 // As an alternative, we can check if the current trimmed line is empty
 250                 if ($trimmed_line == '') {
 251                     break;
 252                 }
 253
 254                 // check for continuation line...
 255                 // RFC 2616 Section 2.2 "Basic Rules":
 256                 // HTTP/1.1 header field values can be folded onto multiple lines if the
 257                 // continuation line begins with a space or horizontal tab. All linear
 258                 // white space, including folding, has the same semantics as SP. A
 259                 // recipient MAY replace any linear white space with a single SP before
 260                 // interpreting the field value or forwarding the message downstream.
 261                 if ($last_match && preg_match('/^\s+(.*)/', $line, $match)) {
 262                     // append to previous header value
 263                     $headers[$x-1] .= ' '.rtrim($match[1]);
 264                     continue;
 265                 }
 266             }
 267             $line_last = $line;
 268
 269             // split header name and value
 270             if (preg_match('/^Set-Cookie\s*:\s*(.*)/i', $line, $match)) {
 271                 $headers[$x++] = rtrim($match[1]);
 272                 $last_match = true;
 273             } else {
 274                 $last_match = false;
 275             }
 276         }
 277         return $headers;
 278         }
 279
 280     /**
 281     * Set Cookie
 282     * @param string $domain
 283     * @param string $path
 284     * @param string $name cookie name
 285     * @param string $value cookie value
 286     * @param bool $secure
 287     * @param int $expires expiry time (null if session cookie, <= 0 will delete cookie)
 288     * @return void
 289     */
 290     function set_cookie($domain, $path, $name, $value, $secure=false, $expires=null)
 291     {
 292         if ($domain == '') return;
 293         if ($path == '') return;
 294         if ($name == '') return;
 295         // check if cookie needs to go
 296         if (isset($expires) && ($expires <= 0)) {
 297             if (isset($this->cookies[$domain][$path][$name])) unset($this->cookies[$domain][$path][$name]);
 298             return;
 299         }
 300         if ($value == '') return;
 301         $this->cookies[$domain][$path][$name] = array($value, $secure, $expires);
 302         return;
 303     }
 304
 305     /**
 306     * Clear cookies - [domain [,path [,name]]] - call method with no arguments to clear all cookies.
 307     * @param string $domain
 308     * @param string $path
 309     * @param string $name
 310     * @return void
 311     */
 312     function clear($domain=null, $path=null, $name=null)
 313     {
 314         if (!isset($domain)) {
 315             $this->cookies = array();
 316         } elseif (!isset($path)) {
 317             if (isset($this->cookies[$domain])) unset($this->cookies[$domain]);
 318         } elseif (!isset($name)) {
 319             if (isset($this->cookies[$domain][$path])) unset($this->cookies[$domain][$path]);
 320         } elseif (isset($name)) {
 321             if (isset($this->cookies[$domain][$path][$name])) unset($this->cookies[$domain][$path][$name]);
 322         }
 323     }
 324
 325     /**
 326     * Compare string length - used for sorting
 327     * @access private
 328     * @return int
 329     */
 330     function _cmp_length($a, $b)
 331     {
 332         $la = strlen($a); $lb = strlen($b);
 333         if ($la == $lb) return 0;
 334         return ($la > $lb) ? -1 : 1;
 335     }
 336
 337     /**
 338     * Reduce domain
 339     * @param string $domain
 340     * @return string
 341     * @access private
 342     */
 343     function _reduce_domain($domain)
 344     {
 345         if ($domain == '') return '';
 346         if (substr($domain, 0, 1) == '.') return substr($domain, 1);
 347         return substr($domain, strpos($domain, '.'));
 348     }
 349
 350     /**
 351     * Path match - check if path1 path-matches path2
 352     *
 353     * From RFC 2965:
 354     *   <i>For two strings that represent paths, P1 and P2, P1 path-matches P2
 355     *   if P2 is a prefix of P1 (including the case where P1 and P2 string-
 356     *   compare equal).  Thus, the string /tec/waldo path-matches /tec.</i>
 357     * @param string $path1
 358     * @param string $path2
 359     * @return bool
 360     * @access private
 361     */
 362     function _path_match($path1, $path2)
 363     {
 364         return (substr($path1, 0, strlen($path2)) == $path2);
 365     }
 366
 367     /**
 368     * Domain match - check if domain1 domain-matches domain2
 369     *
 370     * A few extracts from RFC 2965:
 371     *  -  A Set-Cookie2 from request-host y.x.foo.com for Domain=.foo.com
 372     *     would be rejected, because H is y.x and contains a dot.
 373     *
 374     *  -  A Set-Cookie2 from request-host x.foo.com for Domain=.foo.com
 375     *     would be accepted.
 376     *
 377     *  -  A Set-Cookie2 with Domain=.com or Domain=.com., will always be
 378     *     rejected, because there is no embedded dot.
 379     *
 380     *  -  A Set-Cookie2 from request-host example for Domain=.local will
 381     *     be accepted, because the effective host name for the request-
 382     *     host is example.local, and example.local domain-matches .local.
 383     *
 384     * I'm ignoring the first point for now (must check to see how other browsers handle
 385     * this rule for Set-Cookie headers)
 386     *
 387     * @param string $domain1
 388     * @param string $domain2
 389     * @return bool
 390     * @access private
 391     */
 392     function _domain_match($domain1, $domain2)
 393     {
 394         $domain1 = strtolower($domain1);
 395         $domain2 = strtolower($domain2);
 396         while (strpos($domain1, '.') !== false) {
 397             if ($domain1 == $domain2) return true;
 398             $domain1 = $this->_reduce_domain($domain1);
 399             continue;
 400         }
 401         return false;
 402     }
 403 }
 404 ?>