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         /** @var callable  */
  50         protected $callback;
  51         protected $maxRedirects = 5;
  52         protected $followRedirects = false;
  53         protected $connectTimeout;
  54
  55         /**
  56          * @var CookieJar
  57          */
  58         protected $cookieJar;
  59
  60         protected $headerList = [];
  61         protected $respVersion = "0.9";
  62         protected $respStatus = "200 Ok";
  63         protected $respHeaders = [];
  64
  65         /** @var StatusValue */
  66         protected $status;
  67
  68         /**
  69          * @var Profiler
  70          */
  71         protected $profiler;
  72
  73         /**
  74          * @var string
  75          */
  76         protected $profileName;
  77
  78         /**
  79          * @var LoggerInterface;
  80          */
  81         protected $logger;
  82
  83         /**
  84          * @param string $url Url to use. If protocol-relative, will be expanded to an http:// URL
  85          * @param array $options (optional) extra params to pass (see Http::request())
  86          * @param string $caller The method making this request, for profiling
  87          * @param Profiler $profiler An instance of the profiler for profiling, or null
  88          */
  89         protected function __construct(
  90                 $url, $options = [], $caller = __METHOD__, $profiler = null
  91         ) {
  92                 global $wgHTTPTimeout, $wgHTTPConnectTimeout;
  93
  94                 $this->url = wfExpandUrl( $url, PROTO_HTTP );
  95                 $this->parsedUrl = wfParseUrl( $this->url );
  96
  97                 if ( isset( $options['logger'] ) ) {
  98                         $this->logger = $options['logger'];
  99                 } else {
 100                         $this->logger = new NullLogger();
 101                 }
 102
 103                 if ( !$this->parsedUrl || !Http::isValidURI( $this->url ) ) {
 104                         $this->status = StatusValue::newFatal( 'http-invalid-url', $url );
 105                 } else {
 106                         $this->status = StatusValue::newGood( 100 ); // continue
 107                 }
 108
 109                 if ( isset( $options['timeout'] ) && $options['timeout'] != 'default' ) {
 110                         $this->timeout = $options['timeout'];
 111                 } else {
 112                         $this->timeout = $wgHTTPTimeout;
 113                 }
 114                 if ( isset( $options['connectTimeout'] ) && $options['connectTimeout'] != 'default' ) {
 115                         $this->connectTimeout = $options['connectTimeout'];
 116                 } else {
 117                         $this->connectTimeout = $wgHTTPConnectTimeout;
 118                 }
 119                 if ( isset( $options['userAgent'] ) ) {
 120                         $this->setUserAgent( $options['userAgent'] );
 121                 }
 122                 if ( isset( $options['username'] ) && isset( $options['password'] ) ) {
 123                         $this->setHeader(
 124                                 'Authorization',
 125                                 'Basic ' . base64_encode( $options['username'] . ':' . $options['password'] )
 126                         );
 127                 }
 128                 if ( isset( $options['originalRequest'] ) ) {
 129                         $this->setOriginalRequest( $options['originalRequest'] );
 130                 }
 131
 132                 $members = [ "postData", "proxy", "noProxy", "sslVerifyHost", "caInfo",
 133                                 "method", "followRedirects", "maxRedirects", "sslVerifyCert", "callback" ];
 134
 135                 foreach ( $members as $o ) {
 136                         if ( isset( $options[$o] ) ) {
 137                                 // ensure that MWHttpRequest::method is always
 138                                 // uppercased. T38137
 139                                 if ( $o == 'method' ) {
 140                                         $options[$o] = strtoupper( $options[$o] );
 141                                 }
 142                                 $this->$o = $options[$o];
 143                         }
 144                 }
 145
 146                 if ( $this->noProxy ) {
 147                         $this->proxy = ''; // noProxy takes precedence
 148                 }
 149
 150                 // Profile based on what's calling us
 151                 $this->profiler = $profiler;
 152                 $this->profileName = $caller;
 153         }
 154
 155         /**
 156          * @param LoggerInterface $logger
 157          */
 158         public function setLogger( LoggerInterface $logger ) {
 159                 $this->logger = $logger;
 160         }
 161
 162         /**
 163          * Simple function to test if we can make any sort of requests at all, using
 164          * cURL or fopen()
 165          * @return bool
 166          */
 167         public static function canMakeRequests() {
 168                 return function_exists( 'curl_init' ) || wfIniGetBool( 'allow_url_fopen' );
 169         }
 170
 171         /**
 172          * Generate a new request object
 173          * @param string $url Url to use
 174          * @param array $options (optional) extra params to pass (see Http::request())
 175          * @param string $caller The method making this request, for profiling
 176          * @throws DomainException
 177          * @return CurlHttpRequest|PhpHttpRequest
 178          * @see MWHttpRequest::__construct
 179          */
 180         public static function factory( $url, $options = null, $caller = __METHOD__ ) {
 181                 if ( !Http::$httpEngine ) {
 182                         Http::$httpEngine = function_exists( 'curl_init' ) ? 'curl' : 'php';
 183                 } elseif ( Http::$httpEngine == 'curl' && !function_exists( 'curl_init' ) ) {
 184                         throw new DomainException( __METHOD__ . ': curl (http://php.net/curl) is not installed, but' .
 185                                 ' Http::$httpEngine is set to "curl"' );
 186                 }
 187
 188                 if ( !is_array( $options ) ) {
 189                         $options = [];
 190                 }
 191
 192                 if ( !isset( $options['logger'] ) ) {
 193                         $options['logger'] = LoggerFactory::getInstance( 'http' );
 194                 }
 195
 196                 switch ( Http::$httpEngine ) {
 197                         case 'curl':
 198                                 return new CurlHttpRequest( $url, $options, $caller, Profiler::instance() );
 199                         case 'php':
 200                                 if ( !wfIniGetBool( 'allow_url_fopen' ) ) {
 201                                         throw new DomainException( __METHOD__ . ': allow_url_fopen ' .
 202                                                 'needs to be enabled for pure PHP http requests to ' .
 203                                                 'work. If possible, curl should be used instead. See ' .
 204                                                 'http://php.net/curl.'
 205                                         );
 206                                 }
 207                                 return new PhpHttpRequest( $url, $options, $caller, Profiler::instance() );
 208                         default:
 209                                 throw new DomainException( __METHOD__ . ': The setting of Http::$httpEngine is not valid.' );
 210                 }
 211         }
 212
 213         /**
 214          * Get the body, or content, of the response to the request
 215          *
 216          * @return string
 217          */
 218         public function getContent() {
 219                 return $this->content;
 220         }
 221
 222         /**
 223          * Set the parameters of the request
 224          *
 225          * @param array $args
 226          * @todo overload the args param
 227          */
 228         public function setData( $args ) {
 229                 $this->postData = $args;
 230         }
 231
 232         /**
 233          * Take care of setting up the proxy (do nothing if "noProxy" is set)
 234          *
 235          * @return void
 236          */
 237         protected function proxySetup() {
 238                 // If there is an explicit proxy set and proxies are not disabled, then use it
 239                 if ( $this->proxy && !$this->noProxy ) {
 240                         return;
 241                 }
 242
 243                 // Otherwise, fallback to $wgHTTPProxy if this is not a machine
 244                 // local URL and proxies are not disabled
 245                 if ( self::isLocalURL( $this->url ) || $this->noProxy ) {
 246                         $this->proxy = '';
 247                 } else {
 248                         $this->proxy = Http::getProxy();
 249                 }
 250         }
 251
 252         /**
 253          * Check if the URL can be served by localhost
 254          *
 255          * @param string $url Full url to check
 256          * @return bool
 257          */
 258         private static function isLocalURL( $url ) {
 259                 global $wgCommandLineMode, $wgLocalVirtualHosts;
 260
 261                 if ( $wgCommandLineMode ) {
 262                         return false;
 263                 }
 264
 265                 // Extract host part
 266                 $matches = [];
 267                 if ( preg_match( '!^https?://([\w.-]+)[/:].*$!', $url, $matches ) ) {
 268                         $host = $matches[1];
 269                         // Split up dotwise
 270                         $domainParts = explode( '.', $host );
 271                         // Check if this domain or any superdomain is listed as a local virtual host
 272                         $domainParts = array_reverse( $domainParts );
 273
 274                         $domain = '';
 275                         $countParts = count( $domainParts );
 276                         for ( $i = 0; $i < $countParts; $i++ ) {
 277                                 $domainPart = $domainParts[$i];
 278                                 if ( $i == 0 ) {
 279                                         $domain = $domainPart;
 280                                 } else {
 281                                         $domain = $domainPart . '.' . $domain;
 282                                 }
 283
 284                                 if ( in_array( $domain, $wgLocalVirtualHosts ) ) {
 285                                         return true;
 286                                 }
 287                         }
 288                 }
 289
 290                 return false;
 291         }
 292
 293         /**
 294          * Set the user agent
 295          * @param string $UA
 296          */
 297         public function setUserAgent( $UA ) {
 298                 $this->setHeader( 'User-Agent', $UA );
 299         }
 300
 301         /**
 302          * Set an arbitrary header
 303          * @param string $name
 304          * @param string $value
 305          */
 306         public function setHeader( $name, $value ) {
 307                 // I feel like I should normalize the case here...
 308                 $this->reqHeaders[$name] = $value;
 309         }
 310
 311         /**
 312          * Get an array of the headers
 313          * @return array
 314          */
 315         protected function getHeaderList() {
 316                 $list = [];
 317
 318                 if ( $this->cookieJar ) {
 319                         $this->reqHeaders['Cookie'] =
 320                                 $this->cookieJar->serializeToHttpRequest(
 321                                         $this->parsedUrl['path'],
 322                                         $this->parsedUrl['host']
 323                                 );
 324                 }
 325
 326                 foreach ( $this->reqHeaders as $name => $value ) {
 327                         $list[] = "$name: $value";
 328                 }
 329
 330                 return $list;
 331         }
 332
 333         /**
 334          * Set a read callback to accept data read from the HTTP request.
 335          * By default, data is appended to an internal buffer which can be
 336          * retrieved through $req->getContent().
 337          *
 338          * To handle data as it comes in -- especially for large files that
 339          * would not fit in memory -- you can instead set your own callback,
 340          * in the form function($resource, $buffer) where the first parameter
 341          * is the low-level resource being read (implementation specific),
 342          * and the second parameter is the data buffer.
 343          *
 344          * You MUST return the number of bytes handled in the buffer; if fewer
 345          * bytes are reported handled than were passed to you, the HTTP fetch
 346          * will be aborted.
 347          *
 348          * @param callable|null $callback
 349          * @throws InvalidArgumentException
 350          */
 351         public function setCallback( $callback ) {
 352                 if ( is_null( $callback ) ) {
 353                         $callback = [ $this, 'read' ];
 354                 } elseif ( !is_callable( $callback ) ) {
 355                         throw new InvalidArgumentException( __METHOD__ . ': invalid callback' );
 356                 }
 357                 $this->callback = $callback;
 358         }
 359
 360         /**
 361          * A generic callback to read the body of the response from a remote
 362          * server.
 363          *
 364          * @param resource $fh
 365          * @param string $content
 366          * @return int
 367          * @internal
 368          */
 369         public function read( $fh, $content ) {
 370                 $this->content .= $content;
 371                 return strlen( $content );
 372         }
 373
 374         /**
 375          * Take care of whatever is necessary to perform the URI request.
 376          *
 377          * @return StatusValue
 378          * @note currently returns Status for B/C
 379          */
 380         public function execute() {
 381                 throw new LogicException( 'children must override this' );
 382         }
 383
 384         protected function prepare() {
 385                 $this->content = "";
 386
 387                 if ( strtoupper( $this->method ) == "HEAD" ) {
 388                         $this->headersOnly = true;
 389                 }
 390
 391                 $this->proxySetup(); // set up any proxy as needed
 392
 393                 if ( !$this->callback ) {
 394                         $this->setCallback( null );
 395                 }
 396
 397                 if ( !isset( $this->reqHeaders['User-Agent'] ) ) {
 398                         $this->setUserAgent( Http::userAgent() );
 399                 }
 400         }
 401
 402         /**
 403          * Parses the headers, including the HTTP status code and any
 404          * Set-Cookie headers.  This function expects the headers to be
 405          * found in an array in the member variable headerList.
 406          */
 407         protected function parseHeader() {
 408                 $lastname = "";
 409
 410                 foreach ( $this->headerList as $header ) {
 411                         if ( preg_match( "#^HTTP/([0-9.]+) (.*)#", $header, $match ) ) {
 412                                 $this->respVersion = $match[1];
 413                                 $this->respStatus = $match[2];
 414                         } elseif ( preg_match( "#^[ \t]#", $header ) ) {
 415                                 $last = count( $this->respHeaders[$lastname] ) - 1;
 416                                 $this->respHeaders[$lastname][$last] .= "\r\n$header";
 417                         } elseif ( preg_match( "#^([^:]*):[\t ]*(.*)#", $header, $match ) ) {
 418                                 $this->respHeaders[strtolower( $match[1] )][] = $match[2];
 419                                 $lastname = strtolower( $match[1] );
 420                         }
 421                 }
 422
 423                 $this->parseCookies();
 424         }
 425
 426         /**
 427          * Sets HTTPRequest status member to a fatal value with the error
 428          * message if the returned integer value of the status code was
 429          * not successful (< 300) or a redirect (>=300 and < 400).  (see
 430          * RFC2616, section 10,
 431          * http://www.w3.org/Protocols/rfc2616/rfc2616-sec10.html for a
 432          * list of status codes.)
 433          */
 434         protected function setStatus() {
 435                 if ( !$this->respHeaders ) {
 436                         $this->parseHeader();
 437                 }
 438
 439                 if ( (int)$this->respStatus > 399 ) {
 440                         list( $code, $message ) = explode( " ", $this->respStatus, 2 );
 441                         $this->status->fatal( "http-bad-status", $code, $message );
 442                 }
 443         }
 444
 445         /**
 446          * Get the integer value of the HTTP status code (e.g. 200 for "200 Ok")
 447          * (see RFC2616, section 10, http://www.w3.org/Protocols/rfc2616/rfc2616-sec10.html
 448          * for a list of status codes.)
 449          *
 450          * @return int
 451          */
 452         public function getStatus() {
 453                 if ( !$this->respHeaders ) {
 454                         $this->parseHeader();
 455                 }
 456
 457                 return (int)$this->respStatus;
 458         }
 459
 460         /**
 461          * Returns true if the last status code was a redirect.
 462          *
 463          * @return bool
 464          */
 465         public function isRedirect() {
 466                 if ( !$this->respHeaders ) {
 467                         $this->parseHeader();
 468                 }
 469
 470                 $status = (int)$this->respStatus;
 471
 472                 if ( $status >= 300 && $status <= 303 ) {
 473                         return true;
 474                 }
 475
 476                 return false;
 477         }
 478
 479         /**
 480          * Returns an associative array of response headers after the
 481          * request has been executed.  Because some headers
 482          * (e.g. Set-Cookie) can appear more than once the, each value of
 483          * the associative array is an array of the values given.
 484          *
 485          * @return array
 486          */
 487         public function getResponseHeaders() {
 488                 if ( !$this->respHeaders ) {
 489                         $this->parseHeader();
 490                 }
 491
 492                 return $this->respHeaders;
 493         }
 494
 495         /**
 496          * Returns the value of the given response header.
 497          *
 498          * @param string $header
 499          * @return string|null
 500          */
 501         public function getResponseHeader( $header ) {
 502                 if ( !$this->respHeaders ) {
 503                         $this->parseHeader();
 504                 }
 505
 506                 if ( isset( $this->respHeaders[strtolower( $header )] ) ) {
 507                         $v = $this->respHeaders[strtolower( $header )];
 508                         return $v[count( $v ) - 1];
 509                 }
 510
 511                 return null;
 512         }
 513
 514         /**
 515          * Tells the MWHttpRequest object to use this pre-loaded CookieJar.
 516          *
 517          * To read response cookies from the jar, getCookieJar must be called first.
 518          *
 519          * @param CookieJar $jar
 520          */
 521         public function setCookieJar( $jar ) {
 522                 $this->cookieJar = $jar;
 523         }
 524
 525         /**
 526          * Returns the cookie jar in use.
 527          *
 528          * @return CookieJar
 529          */
 530         public function getCookieJar() {
 531                 if ( !$this->respHeaders ) {
 532                         $this->parseHeader();
 533                 }
 534
 535                 return $this->cookieJar;
 536         }
 537
 538         /**
 539          * Sets a cookie. Used before a request to set up any individual
 540          * cookies. Used internally after a request to parse the
 541          * Set-Cookie headers.
 542          * @see Cookie::set
 543          * @param string $name
 544          * @param string $value
 545          * @param array $attr
 546          */
 547         public function setCookie( $name, $value, $attr = [] ) {
 548                 if ( !$this->cookieJar ) {
 549                         $this->cookieJar = new CookieJar;
 550                 }
 551
 552                 if ( $this->parsedUrl && !isset( $attr['domain'] ) ) {
 553                         $attr['domain'] = $this->parsedUrl['host'];
 554                 }
 555
 556                 $this->cookieJar->setCookie( $name, $value, $attr );
 557         }
 558
 559         /**
 560          * Parse the cookies in the response headers and store them in the cookie jar.
 561          */
 562         protected function parseCookies() {
 563                 if ( !$this->cookieJar ) {
 564                         $this->cookieJar = new CookieJar;
 565                 }
 566
 567                 if ( isset( $this->respHeaders['set-cookie'] ) ) {
 568                         $url = parse_url( $this->getFinalUrl() );
 569                         foreach ( $this->respHeaders['set-cookie'] as $cookie ) {
 570                                 $this->cookieJar->parseCookieResponseHeader( $cookie, $url['host'] );
 571                         }
 572                 }
 573         }
 574
 575         /**
 576          * Returns the final URL after all redirections.
 577          *
 578          * Relative values of the "Location" header are incorrect as
 579          * stated in RFC, however they do happen and modern browsers
 580          * support them.  This function loops backwards through all
 581          * locations in order to build the proper absolute URI - Marooned
 582          * at wikia-inc.com
 583          *
 584          * Note that the multiple Location: headers are an artifact of
 585          * CURL -- they shouldn't actually get returned this way. Rewrite
 586          * this when T31232 is taken care of (high-level redirect
 587          * handling rewrite).
 588          *
 589          * @return string
 590          */
 591         public function getFinalUrl() {
 592                 $headers = $this->getResponseHeaders();
 593
 594                 // return full url (fix for incorrect but handled relative location)
 595                 if ( isset( $headers['location'] ) ) {
 596                         $locations = $headers['location'];
 597                         $domain = '';
 598                         $foundRelativeURI = false;
 599                         $countLocations = count( $locations );
 600
 601                         for ( $i = $countLocations - 1; $i >= 0; $i-- ) {
 602                                 $url = parse_url( $locations[$i] );
 603
 604                                 if ( isset( $url['host'] ) ) {
 605                                         $domain = $url['scheme'] . '://' . $url['host'];
 606                                         break; // found correct URI (with host)
 607                                 } else {
 608                                         $foundRelativeURI = true;
 609                                 }
 610                         }
 611
 612                         if ( !$foundRelativeURI ) {
 613                                 return $locations[$countLocations - 1];
 614                         }
 615                         if ( $domain ) {
 616                                 return $domain . $locations[$countLocations - 1];
 617                         }
 618                         $url = parse_url( $this->url );
 619                         if ( isset( $url['host'] ) ) {
 620                                 return $url['scheme'] . '://' . $url['host'] .
 621                                         $locations[$countLocations - 1];
 622                         }
 623                 }
 624
 625                 return $this->url;
 626         }
 627
 628         /**
 629          * Returns true if the backend can follow redirects. Overridden by the
 630          * child classes.
 631          * @return bool
 632          */
 633         public function canFollowRedirects() {
 634                 return true;
 635         }
 636
 637         /**
 638          * Set information about the original request. This can be useful for
 639          * endpoints/API modules which act as a proxy for some service, and
 640          * throttling etc. needs to happen in that service.
 641          * Calling this will result in the X-Forwarded-For and X-Original-User-Agent
 642          * headers being set.
 643          * @param WebRequest|array $originalRequest When in array form, it's
 644          *   expected to have the keys 'ip' and 'userAgent'.
 645          * @note IP/user agent is personally identifiable information, and should
 646          *   only be set when the privacy policy of the request target is
 647          *   compatible with that of the MediaWiki installation.
 648          */
 649         public function setOriginalRequest( $originalRequest ) {
 650                 if ( $originalRequest instanceof WebRequest ) {
 651                         $originalRequest = [
 652                                 'ip' => $originalRequest->getIP(),
 653                                 'userAgent' => $originalRequest->getHeader( 'User-Agent' ),
 654                         ];
 655                 } elseif (
 656                         !is_array( $originalRequest )
 657                         || array_diff( [ 'ip', 'userAgent' ], array_keys( $originalRequest ) )
 658                 ) {
 659                         throw new InvalidArgumentException( __METHOD__ . ': $originalRequest must be a '
 660                                 . "WebRequest or an array with 'ip' and 'userAgent' keys" );
 661                 }
 662
 663                 $this->reqHeaders['X-Forwarded-For'] = $originalRequest['ip'];
 664                 $this->reqHeaders['X-Original-User-Agent'] = $originalRequest['userAgent'];
 665         }
 666 }