includes/http/MWHttpRequest.php

   1 <?php
   2 /**
   3  * This program is free software; you can redistribute it and/or modify
   4  * it under the terms of the GNU General Public License as published by
   5  * the Free Software Foundation; either version 2 of the License, or
   6  * (at your option) any later version.
   7  *
   8  * This program is distributed in the hope that it will be useful,
   9  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  10  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  11  * GNU General Public License for more details.
  12  *
  13  * You should have received a copy of the GNU General Public License along
  14  * with this program; if not, write to the Free Software Foundation, Inc.,
  15  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  16  * http://www.gnu.org/copyleft/gpl.html
  17  *
  18  * @file
  19  */
  20
  21 use MediaWiki\Logger\LoggerFactory;
  22 use Psr\Log\LoggerInterface;
  23 use Psr\Log\LoggerAwareInterface;
  24 use Psr\Log\NullLogger;
  25
  26 /**
  27  * This wrapper class will call out to curl (if available) or fallback
  28  * to regular PHP if necessary for handling internal HTTP requests.
  29  *
  30  * Renamed from HttpRequest to MWHttpRequest to avoid conflict with
  31  * PHP's HTTP extension.
  32  */
  33 class MWHttpRequest implements LoggerAwareInterface {
  34         const SUPPORTS_FILE_POSTS = false;
  35
  36         protected $content;
  37         protected $timeout = 'default';
  38         protected $headersOnly = null;
  39         protected $postData = null;
  40         protected $proxy = null;
  41         protected $noProxy = false;
  42         protected $sslVerifyHost = true;
  43         protected $sslVerifyCert = true;
  44         protected $caInfo = null;
  45         protected $method = "GET";
  46         protected $reqHeaders = [];
  47         protected $url;
  48         protected $parsedUrl;
  49         protected $callback;
  50         protected $maxRedirects = 5;
  51         protected $followRedirects = false;
  52
  53         /**
  54          * @var CookieJar
  55          */
  56         protected $cookieJar;
  57
  58         protected $headerList = [];
  59         protected $respVersion = "0.9";
  60         protected $respStatus = "200 Ok";
  61         protected $respHeaders = [];
  62
  63         public $status;
  64
  65         /**
  66          * @var Profiler
  67          */
  68         protected $profiler;
  69
  70         /**
  71          * @var string
  72          */
  73         protected $profileName;
  74
  75         /**
  76          * @var LoggerInterface;
  77          */
  78         protected $logger;
  79
  80         /**
  81          * @param string $url Url to use. If protocol-relative, will be expanded to an http:// URL
  82          * @param array $options (optional) extra params to pass (see Http::request())
  83          * @param string $caller The method making this request, for profiling
  84          * @param Profiler $profiler An instance of the profiler for profiling, or null
  85          */
  86         protected function __construct(
  87                 $url, $options = [], $caller = __METHOD__, $profiler = null
  88         ) {
  89                 global $wgHTTPTimeout, $wgHTTPConnectTimeout;
  90
  91                 $this->url = wfExpandUrl( $url, PROTO_HTTP );
  92                 $this->parsedUrl = wfParseUrl( $this->url );
  93
  94                 if ( isset( $options['logger'] ) ) {
  95                         $this->logger = $options['logger'];
  96                 } else {
  97                         $this->logger = new NullLogger();
  98                 }
  99
 100                 if ( !$this->parsedUrl || !Http::isValidURI( $this->url ) ) {
 101                         $this->status = Status::newFatal( 'http-invalid-url', $url );
 102                 } else {
 103                         $this->status = Status::newGood( 100 ); // continue
 104                 }
 105
 106                 if ( isset( $options['timeout'] ) && $options['timeout'] != 'default' ) {
 107                         $this->timeout = $options['timeout'];
 108                 } else {
 109                         $this->timeout = $wgHTTPTimeout;
 110                 }
 111                 if ( isset( $options['connectTimeout'] ) && $options['connectTimeout'] != 'default' ) {
 112                         $this->connectTimeout = $options['connectTimeout'];
 113                 } else {
 114                         $this->connectTimeout = $wgHTTPConnectTimeout;
 115                 }
 116                 if ( isset( $options['userAgent'] ) ) {
 117                         $this->setUserAgent( $options['userAgent'] );
 118                 }
 119
 120                 $members = [ "postData", "proxy", "noProxy", "sslVerifyHost", "caInfo",
 121                                 "method", "followRedirects", "maxRedirects", "sslVerifyCert", "callback" ];
 122
 123                 foreach ( $members as $o ) {
 124                         if ( isset( $options[$o] ) ) {
 125                                 // ensure that MWHttpRequest::method is always
 126                                 // uppercased. Bug 36137
 127                                 if ( $o == 'method' ) {
 128                                         $options[$o] = strtoupper( $options[$o] );
 129                                 }
 130                                 $this->$o = $options[$o];
 131                         }
 132                 }
 133
 134                 if ( $this->noProxy ) {
 135                         $this->proxy = ''; // noProxy takes precedence
 136                 }
 137
 138                 // Profile based on what's calling us
 139                 $this->profiler = $profiler;
 140                 $this->profileName = $caller;
 141         }
 142
 143         /**
 144          * @param LoggerInterface $logger
 145          */
 146         public function setLogger( LoggerInterface $logger ) {
 147                 $this->logger = $logger;
 148         }
 149
 150         /**
 151          * Simple function to test if we can make any sort of requests at all, using
 152          * cURL or fopen()
 153          * @return bool
 154          */
 155         public static function canMakeRequests() {
 156                 return function_exists( 'curl_init' ) || wfIniGetBool( 'allow_url_fopen' );
 157         }
 158
 159         /**
 160          * Generate a new request object
 161          * @param string $url Url to use
 162          * @param array $options (optional) extra params to pass (see Http::request())
 163          * @param string $caller The method making this request, for profiling
 164          * @throws MWException
 165          * @return CurlHttpRequest|PhpHttpRequest
 166          * @see MWHttpRequest::__construct
 167          */
 168         public static function factory( $url, $options = null, $caller = __METHOD__ ) {
 169                 if ( !Http::$httpEngine ) {
 170                         Http::$httpEngine = function_exists( 'curl_init' ) ? 'curl' : 'php';
 171                 } elseif ( Http::$httpEngine == 'curl' && !function_exists( 'curl_init' ) ) {
 172                         throw new MWException( __METHOD__ . ': curl (http://php.net/curl) is not installed, but' .
 173                                 ' Http::$httpEngine is set to "curl"' );
 174                 }
 175
 176                 if ( !is_array( $options ) ) {
 177                         $options = [];
 178                 }
 179
 180                 if ( !isset( $options['logger'] ) ) {
 181                         $options['logger'] = LoggerFactory::getInstance( 'http' );
 182                 }
 183
 184                 switch ( Http::$httpEngine ) {
 185                         case 'curl':
 186                                 return new CurlHttpRequest( $url, $options, $caller, Profiler::instance() );
 187                         case 'php':
 188                                 if ( !wfIniGetBool( 'allow_url_fopen' ) ) {
 189                                         throw new MWException( __METHOD__ . ': allow_url_fopen ' .
 190                                                 'needs to be enabled for pure PHP http requests to ' .
 191                                                 'work. If possible, curl should be used instead. See ' .
 192                                                 'http://php.net/curl.'
 193                                         );
 194                                 }
 195                                 return new PhpHttpRequest( $url, $options, $caller, Profiler::instance() );
 196                         default:
 197                                 throw new MWException( __METHOD__ . ': The setting of Http::$httpEngine is not valid.' );
 198                 }
 199         }
 200
 201         /**
 202          * Get the body, or content, of the response to the request
 203          *
 204          * @return string
 205          */
 206         public function getContent() {
 207                 return $this->content;
 208         }
 209
 210         /**
 211          * Set the parameters of the request
 212          *
 213          * @param array $args
 214          * @todo overload the args param
 215          */
 216         public function setData( $args ) {
 217                 $this->postData = $args;
 218         }
 219
 220         /**
 221          * Take care of setting up the proxy (do nothing if "noProxy" is set)
 222          *
 223          * @return void
 224          */
 225         public function proxySetup() {
 226                 // If there is an explicit proxy set and proxies are not disabled, then use it
 227                 if ( $this->proxy && !$this->noProxy ) {
 228                         return;
 229                 }
 230
 231                 // Otherwise, fallback to $wgHTTPProxy if this is not a machine
 232                 // local URL and proxies are not disabled
 233                 if ( self::isLocalURL( $this->url ) || $this->noProxy ) {
 234                         $this->proxy = '';
 235                 } else {
 236                         $this->proxy = Http::getProxy();
 237                 }
 238         }
 239
 240         /**
 241          * Check if the URL can be served by localhost
 242          *
 243          * @param string $url Full url to check
 244          * @return bool
 245          */
 246         private static function isLocalURL( $url ) {
 247                 global $wgCommandLineMode, $wgLocalVirtualHosts;
 248
 249                 if ( $wgCommandLineMode ) {
 250                         return false;
 251                 }
 252
 253                 // Extract host part
 254                 $matches = [];
 255                 if ( preg_match( '!^https?://([\w.-]+)[/:].*$!', $url, $matches ) ) {
 256                         $host = $matches[1];
 257                         // Split up dotwise
 258                         $domainParts = explode( '.', $host );
 259                         // Check if this domain or any superdomain is listed as a local virtual host
 260                         $domainParts = array_reverse( $domainParts );
 261
 262                         $domain = '';
 263                         $countParts = count( $domainParts );
 264                         for ( $i = 0; $i < $countParts; $i++ ) {
 265                                 $domainPart = $domainParts[$i];
 266                                 if ( $i == 0 ) {
 267                                         $domain = $domainPart;
 268                                 } else {
 269                                         $domain = $domainPart . '.' . $domain;
 270                                 }
 271
 272                                 if ( in_array( $domain, $wgLocalVirtualHosts ) ) {
 273                                         return true;
 274                                 }
 275                         }
 276                 }
 277
 278                 return false;
 279         }
 280
 281         /**
 282          * Set the user agent
 283          * @param string $UA
 284          */
 285         public function setUserAgent( $UA ) {
 286                 $this->setHeader( 'User-Agent', $UA );
 287         }
 288
 289         /**
 290          * Set an arbitrary header
 291          * @param string $name
 292          * @param string $value
 293          */
 294         public function setHeader( $name, $value ) {
 295                 // I feel like I should normalize the case here...
 296                 $this->reqHeaders[$name] = $value;
 297         }
 298
 299         /**
 300          * Get an array of the headers
 301          * @return array
 302          */
 303         public function getHeaderList() {
 304                 $list = [];
 305
 306                 if ( $this->cookieJar ) {
 307                         $this->reqHeaders['Cookie'] =
 308                                 $this->cookieJar->serializeToHttpRequest(
 309                                         $this->parsedUrl['path'],
 310                                         $this->parsedUrl['host']
 311                                 );
 312                 }
 313
 314                 foreach ( $this->reqHeaders as $name => $value ) {
 315                         $list[] = "$name: $value";
 316                 }
 317
 318                 return $list;
 319         }
 320
 321         /**
 322          * Set a read callback to accept data read from the HTTP request.
 323          * By default, data is appended to an internal buffer which can be
 324          * retrieved through $req->getContent().
 325          *
 326          * To handle data as it comes in -- especially for large files that
 327          * would not fit in memory -- you can instead set your own callback,
 328          * in the form function($resource, $buffer) where the first parameter
 329          * is the low-level resource being read (implementation specific),
 330          * and the second parameter is the data buffer.
 331          *
 332          * You MUST return the number of bytes handled in the buffer; if fewer
 333          * bytes are reported handled than were passed to you, the HTTP fetch
 334          * will be aborted.
 335          *
 336          * @param callable $callback
 337          * @throws MWException
 338          */
 339         public function setCallback( $callback ) {
 340                 if ( !is_callable( $callback ) ) {
 341                         throw new MWException( 'Invalid MwHttpRequest callback' );
 342                 }
 343                 $this->callback = $callback;
 344         }
 345
 346         /**
 347          * A generic callback to read the body of the response from a remote
 348          * server.
 349          *
 350          * @param resource $fh
 351          * @param string $content
 352          * @return int
 353          */
 354         public function read( $fh, $content ) {
 355                 $this->content .= $content;
 356                 return strlen( $content );
 357         }
 358
 359         /**
 360          * Take care of whatever is necessary to perform the URI request.
 361          *
 362          * @return Status
 363          */
 364         public function execute() {
 365
 366                 $this->content = "";
 367
 368                 if ( strtoupper( $this->method ) == "HEAD" ) {
 369                         $this->headersOnly = true;
 370                 }
 371
 372                 $this->proxySetup(); // set up any proxy as needed
 373
 374                 if ( !$this->callback ) {
 375                         $this->setCallback( [ $this, 'read' ] );
 376                 }
 377
 378                 if ( !isset( $this->reqHeaders['User-Agent'] ) ) {
 379                         $this->setUserAgent( Http::userAgent() );
 380                 }
 381
 382         }
 383
 384         /**
 385          * Parses the headers, including the HTTP status code and any
 386          * Set-Cookie headers.  This function expects the headers to be
 387          * found in an array in the member variable headerList.
 388          */
 389         protected function parseHeader() {
 390
 391                 $lastname = "";
 392
 393                 foreach ( $this->headerList as $header ) {
 394                         if ( preg_match( "#^HTTP/([0-9.]+) (.*)#", $header, $match ) ) {
 395                                 $this->respVersion = $match[1];
 396                                 $this->respStatus = $match[2];
 397                         } elseif ( preg_match( "#^[ \t]#", $header ) ) {
 398                                 $last = count( $this->respHeaders[$lastname] ) - 1;
 399                                 $this->respHeaders[$lastname][$last] .= "\r\n$header";
 400                         } elseif ( preg_match( "#^([^:]*):[\t ]*(.*)#", $header, $match ) ) {
 401                                 $this->respHeaders[strtolower( $match[1] )][] = $match[2];
 402                                 $lastname = strtolower( $match[1] );
 403                         }
 404                 }
 405
 406                 $this->parseCookies();
 407
 408         }
 409
 410         /**
 411          * Sets HTTPRequest status member to a fatal value with the error
 412          * message if the returned integer value of the status code was
 413          * not successful (< 300) or a redirect (>=300 and < 400).  (see
 414          * RFC2616, section 10,
 415          * http://www.w3.org/Protocols/rfc2616/rfc2616-sec10.html for a
 416          * list of status codes.)
 417          */
 418         protected function setStatus() {
 419                 if ( !$this->respHeaders ) {
 420                         $this->parseHeader();
 421                 }
 422
 423                 if ( (int)$this->respStatus > 399 ) {
 424                         list( $code, $message ) = explode( " ", $this->respStatus, 2 );
 425                         $this->status->fatal( "http-bad-status", $code, $message );
 426                 }
 427         }
 428
 429         /**
 430          * Get the integer value of the HTTP status code (e.g. 200 for "200 Ok")
 431          * (see RFC2616, section 10, http://www.w3.org/Protocols/rfc2616/rfc2616-sec10.html
 432          * for a list of status codes.)
 433          *
 434          * @return int
 435          */
 436         public function getStatus() {
 437                 if ( !$this->respHeaders ) {
 438                         $this->parseHeader();
 439                 }
 440
 441                 return (int)$this->respStatus;
 442         }
 443
 444         /**
 445          * Returns true if the last status code was a redirect.
 446          *
 447          * @return bool
 448          */
 449         public function isRedirect() {
 450                 if ( !$this->respHeaders ) {
 451                         $this->parseHeader();
 452                 }
 453
 454                 $status = (int)$this->respStatus;
 455
 456                 if ( $status >= 300 && $status <= 303 ) {
 457                         return true;
 458                 }
 459
 460                 return false;
 461         }
 462
 463         /**
 464          * Returns an associative array of response headers after the
 465          * request has been executed.  Because some headers
 466          * (e.g. Set-Cookie) can appear more than once the, each value of
 467          * the associative array is an array of the values given.
 468          *
 469          * @return array
 470          */
 471         public function getResponseHeaders() {
 472                 if ( !$this->respHeaders ) {
 473                         $this->parseHeader();
 474                 }
 475
 476                 return $this->respHeaders;
 477         }
 478
 479         /**
 480          * Returns the value of the given response header.
 481          *
 482          * @param string $header
 483          * @return string|null
 484          */
 485         public function getResponseHeader( $header ) {
 486                 if ( !$this->respHeaders ) {
 487                         $this->parseHeader();
 488                 }
 489
 490                 if ( isset( $this->respHeaders[strtolower( $header )] ) ) {
 491                         $v = $this->respHeaders[strtolower( $header )];
 492                         return $v[count( $v ) - 1];
 493                 }
 494
 495                 return null;
 496         }
 497
 498         /**
 499          * Tells the MWHttpRequest object to use this pre-loaded CookieJar.
 500          *
 501          * @param CookieJar $jar
 502          */
 503         public function setCookieJar( $jar ) {
 504                 $this->cookieJar = $jar;
 505         }
 506
 507         /**
 508          * Returns the cookie jar in use.
 509          *
 510          * @return CookieJar
 511          */
 512         public function getCookieJar() {
 513                 if ( !$this->respHeaders ) {
 514                         $this->parseHeader();
 515                 }
 516
 517                 return $this->cookieJar;
 518         }
 519
 520         /**
 521          * Sets a cookie. Used before a request to set up any individual
 522          * cookies. Used internally after a request to parse the
 523          * Set-Cookie headers.
 524          * @see Cookie::set
 525          * @param string $name
 526          * @param mixed $value
 527          * @param array $attr
 528          */
 529         public function setCookie( $name, $value = null, $attr = null ) {
 530                 if ( !$this->cookieJar ) {
 531                         $this->cookieJar = new CookieJar;
 532                 }
 533
 534                 $this->cookieJar->setCookie( $name, $value, $attr );
 535         }
 536
 537         /**
 538          * Parse the cookies in the response headers and store them in the cookie jar.
 539          */
 540         protected function parseCookies() {
 541
 542                 if ( !$this->cookieJar ) {
 543                         $this->cookieJar = new CookieJar;
 544                 }
 545
 546                 if ( isset( $this->respHeaders['set-cookie'] ) ) {
 547                         $url = parse_url( $this->getFinalUrl() );
 548                         foreach ( $this->respHeaders['set-cookie'] as $cookie ) {
 549                                 $this->cookieJar->parseCookieResponseHeader( $cookie, $url['host'] );
 550                         }
 551                 }
 552
 553         }
 554
 555         /**
 556          * Returns the final URL after all redirections.
 557          *
 558          * Relative values of the "Location" header are incorrect as
 559          * stated in RFC, however they do happen and modern browsers
 560          * support them.  This function loops backwards through all
 561          * locations in order to build the proper absolute URI - Marooned
 562          * at wikia-inc.com
 563          *
 564          * Note that the multiple Location: headers are an artifact of
 565          * CURL -- they shouldn't actually get returned this way. Rewrite
 566          * this when bug 29232 is taken care of (high-level redirect
 567          * handling rewrite).
 568          *
 569          * @return string
 570          */
 571         public function getFinalUrl() {
 572                 $headers = $this->getResponseHeaders();
 573
 574                 // return full url (fix for incorrect but handled relative location)
 575                 if ( isset( $headers['location'] ) ) {
 576                         $locations = $headers['location'];
 577                         $domain = '';
 578                         $foundRelativeURI = false;
 579                         $countLocations = count( $locations );
 580
 581                         for ( $i = $countLocations - 1; $i >= 0; $i-- ) {
 582                                 $url = parse_url( $locations[$i] );
 583
 584                                 if ( isset( $url['host'] ) ) {
 585                                         $domain = $url['scheme'] . '://' . $url['host'];
 586                                         break; // found correct URI (with host)
 587                                 } else {
 588                                         $foundRelativeURI = true;
 589                                 }
 590                         }
 591
 592                         if ( $foundRelativeURI ) {
 593                                 if ( $domain ) {
 594                                         return $domain . $locations[$countLocations - 1];
 595                                 } else {
 596                                         $url = parse_url( $this->url );
 597                                         if ( isset( $url['host'] ) ) {
 598                                                 return $url['scheme'] . '://' . $url['host'] .
 599                                                         $locations[$countLocations - 1];
 600                                         }
 601                                 }
 602                         } else {
 603                                 return $locations[$countLocations - 1];
 604                         }
 605                 }
 606
 607                 return $this->url;
 608         }
 609
 610         /**
 611          * Returns true if the backend can follow redirects. Overridden by the
 612          * child classes.
 613          * @return bool
 614          */
 615         public function canFollowRedirects() {
 616                 return true;
 617         }
 618 }