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 Psr\Log\LoggerInterface;
  22 use Psr\Log\LoggerAwareInterface;
  23 use Psr\Log\NullLogger;
  24
  25 /**
  26  * This wrapper class will call out to curl (if available) or fallback
  27  * to regular PHP if necessary for handling internal HTTP requests.
  28  *
  29  * Renamed from HttpRequest to MWHttpRequest to avoid conflict with
  30  * PHP's HTTP extension.
  31  */
  32 abstract class MWHttpRequest implements LoggerAwareInterface {
  33         const SUPPORTS_FILE_POSTS = false;
  34
  35         /**
  36          * @var int|string
  37          */
  38         protected $timeout = 'default';
  39
  40         protected $content;
  41         protected $headersOnly = null;
  42         protected $postData = null;
  43         protected $proxy = null;
  44         protected $noProxy = false;
  45         protected $sslVerifyHost = true;
  46         protected $sslVerifyCert = true;
  47         protected $caInfo = null;
  48         protected $method = "GET";
  49         protected $reqHeaders = [];
  50         protected $url;
  51         protected $parsedUrl;
  52         /** @var callable */
  53         protected $callback;
  54         protected $maxRedirects = 5;
  55         protected $followRedirects = false;
  56         protected $connectTimeout;
  57
  58         /**
  59          * @var CookieJar
  60          */
  61         protected $cookieJar;
  62
  63         protected $headerList = [];
  64         protected $respVersion = "0.9";
  65         protected $respStatus = "200 Ok";
  66         protected $respHeaders = [];
  67
  68         /** @var StatusValue */
  69         protected $status;
  70
  71         /**
  72          * @var Profiler
  73          */
  74         protected $profiler;
  75
  76         /**
  77          * @var string
  78          */
  79         protected $profileName;
  80
  81         /**
  82          * @var LoggerInterface
  83          */
  84         protected $logger;
  85
  86         /**
  87          * @param string $url Url to use. If protocol-relative, will be expanded to an http:// URL
  88          * @param array $options (optional) extra params to pass (see Http::request())
  89          * @param string $caller The method making this request, for profiling
  90          * @param Profiler|null $profiler An instance of the profiler for profiling, or null
  91          * @throws Exception
  92          */
  93         public function __construct(
  94                 $url, array $options = [], $caller = __METHOD__, $profiler = null
  95         ) {
  96                 global $wgHTTPTimeout, $wgHTTPConnectTimeout;
  97
  98                 $this->url = wfExpandUrl( $url, PROTO_HTTP );
  99                 $this->parsedUrl = wfParseUrl( $this->url );
 100
 101                 $this->logger = $options['logger'] ?? new NullLogger();
 102
 103                 if ( !$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          * Deprecated: @see HttpRequestFactory::create
 174          * @param string $url Url to use
 175          * @param array|null $options (optional) extra params to pass (see Http::request())
 176          * @param string $caller The method making this request, for profiling
 177          * @throws DomainException
 178          * @return MWHttpRequest
 179          * @see MWHttpRequest::__construct
 180          */
 181         public static function factory( $url, array $options = null, $caller = __METHOD__ ) {
 182                 if ( $options === null ) {
 183                         $options = [];
 184                 }
 185                 return \MediaWiki\MediaWikiServices::getInstance()
 186                         ->getHttpRequestFactory()
 187                         ->create( $url, $options, $caller );
 188         }
 189
 190         /**
 191          * Get the body, or content, of the response to the request
 192          *
 193          * @return string
 194          */
 195         public function getContent() {
 196                 return $this->content;
 197         }
 198
 199         /**
 200          * Set the parameters of the request
 201          *
 202          * @param array $args
 203          * @todo overload the args param
 204          */
 205         public function setData( $args ) {
 206                 $this->postData = $args;
 207         }
 208
 209         /**
 210          * Take care of setting up the proxy (do nothing if "noProxy" is set)
 211          *
 212          * @return void
 213          */
 214         protected function proxySetup() {
 215                 // If there is an explicit proxy set and proxies are not disabled, then use it
 216                 if ( $this->proxy && !$this->noProxy ) {
 217                         return;
 218                 }
 219
 220                 // Otherwise, fallback to $wgHTTPProxy if this is not a machine
 221                 // local URL and proxies are not disabled
 222                 if ( self::isLocalURL( $this->url ) || $this->noProxy ) {
 223                         $this->proxy = '';
 224                 } else {
 225                         $this->proxy = Http::getProxy();
 226                 }
 227         }
 228
 229         /**
 230          * Check if the URL can be served by localhost
 231          *
 232          * @param string $url Full url to check
 233          * @return bool
 234          */
 235         private static function isLocalURL( $url ) {
 236                 global $wgCommandLineMode, $wgLocalVirtualHosts;
 237
 238                 if ( $wgCommandLineMode ) {
 239                         return false;
 240                 }
 241
 242                 // Extract host part
 243                 $matches = [];
 244                 if ( preg_match( '!^https?://([\w.-]+)[/:].*$!', $url, $matches ) ) {
 245                         $host = $matches[1];
 246                         // Split up dotwise
 247                         $domainParts = explode( '.', $host );
 248                         // Check if this domain or any superdomain is listed as a local virtual host
 249                         $domainParts = array_reverse( $domainParts );
 250
 251                         $domain = '';
 252                         $countParts = count( $domainParts );
 253                         for ( $i = 0; $i < $countParts; $i++ ) {
 254                                 $domainPart = $domainParts[$i];
 255                                 if ( $i == 0 ) {
 256                                         $domain = $domainPart;
 257                                 } else {
 258                                         $domain = $domainPart . '.' . $domain;
 259                                 }
 260
 261                                 if ( in_array( $domain, $wgLocalVirtualHosts ) ) {
 262                                         return true;
 263                                 }
 264                         }
 265                 }
 266
 267                 return false;
 268         }
 269
 270         /**
 271          * Set the user agent
 272          * @param string $UA
 273          */
 274         public function setUserAgent( $UA ) {
 275                 $this->setHeader( 'User-Agent', $UA );
 276         }
 277
 278         /**
 279          * Set an arbitrary header
 280          * @param string $name
 281          * @param string $value
 282          */
 283         public function setHeader( $name, $value ) {
 284                 // I feel like I should normalize the case here...
 285                 $this->reqHeaders[$name] = $value;
 286         }
 287
 288         /**
 289          * Get an array of the headers
 290          * @return array
 291          */
 292         protected function getHeaderList() {
 293                 $list = [];
 294
 295                 if ( $this->cookieJar ) {
 296                         $this->reqHeaders['Cookie'] =
 297                                 $this->cookieJar->serializeToHttpRequest(
 298                                         $this->parsedUrl['path'],
 299                                         $this->parsedUrl['host']
 300                                 );
 301                 }
 302
 303                 foreach ( $this->reqHeaders as $name => $value ) {
 304                         $list[] = "$name: $value";
 305                 }
 306
 307                 return $list;
 308         }
 309
 310         /**
 311          * Set a read callback to accept data read from the HTTP request.
 312          * By default, data is appended to an internal buffer which can be
 313          * retrieved through $req->getContent().
 314          *
 315          * To handle data as it comes in -- especially for large files that
 316          * would not fit in memory -- you can instead set your own callback,
 317          * in the form function($resource, $buffer) where the first parameter
 318          * is the low-level resource being read (implementation specific),
 319          * and the second parameter is the data buffer.
 320          *
 321          * You MUST return the number of bytes handled in the buffer; if fewer
 322          * bytes are reported handled than were passed to you, the HTTP fetch
 323          * will be aborted.
 324          *
 325          * @param callable|null $callback
 326          * @throws InvalidArgumentException
 327          */
 328         public function setCallback( $callback ) {
 329                 if ( is_null( $callback ) ) {
 330                         $callback = [ $this, 'read' ];
 331                 } elseif ( !is_callable( $callback ) ) {
 332                         $this->status->fatal( 'http-internal-error' );
 333                         throw new InvalidArgumentException( __METHOD__ . ': invalid callback' );
 334                 }
 335                 $this->callback = $callback;
 336         }
 337
 338         /**
 339          * A generic callback to read the body of the response from a remote
 340          * server.
 341          *
 342          * @param resource $fh
 343          * @param string $content
 344          * @return int
 345          * @internal
 346          */
 347         public function read( $fh, $content ) {
 348                 $this->content .= $content;
 349                 return strlen( $content );
 350         }
 351
 352         /**
 353          * Take care of whatever is necessary to perform the URI request.
 354          *
 355          * @return StatusValue
 356          * @note currently returns Status for B/C
 357          */
 358         public function execute() {
 359                 throw new LogicException( 'children must override this' );
 360         }
 361
 362         protected function prepare() {
 363                 $this->content = "";
 364
 365                 if ( strtoupper( $this->method ) == "HEAD" ) {
 366                         $this->headersOnly = true;
 367                 }
 368
 369                 $this->proxySetup(); // set up any proxy as needed
 370
 371                 if ( !$this->callback ) {
 372                         $this->setCallback( null );
 373                 }
 374
 375                 if ( !isset( $this->reqHeaders['User-Agent'] ) ) {
 376                         $this->setUserAgent( Http::userAgent() );
 377                 }
 378         }
 379
 380         /**
 381          * Parses the headers, including the HTTP status code and any
 382          * Set-Cookie headers.  This function expects the headers to be
 383          * found in an array in the member variable headerList.
 384          */
 385         protected function parseHeader() {
 386                 $lastname = "";
 387
 388                 // Failure without (valid) headers gets a response status of zero
 389                 if ( !$this->status->isOK() ) {
 390                         $this->respStatus = '0 Error';
 391                 }
 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                 $this->parseCookies();
 407         }
 408
 409         /**
 410          * Sets HTTPRequest status member to a fatal value with the error
 411          * message if the returned integer value of the status code was
 412          * not successful (1-299) or a redirect (300-399).
 413          * See RFC2616, section 10, http://www.w3.org/Protocols/rfc2616/rfc2616-sec10.html
 414          * for a list of status codes.
 415          */
 416         protected function setStatus() {
 417                 if ( !$this->respHeaders ) {
 418                         $this->parseHeader();
 419                 }
 420
 421                 if ( ( (int)$this->respStatus > 0 && (int)$this->respStatus < 400 ) ) {
 422                         $this->status->setResult( true, (int)$this->respStatus );
 423                 } else {
 424                         list( $code, $message ) = explode( " ", $this->respStatus, 2 );
 425                         $this->status->setResult( false, (int)$this->respStatus );
 426                         $this->status->fatal( "http-bad-status", $code, $message );
 427                 }
 428         }
 429
 430         /**
 431          * Get the integer value of the HTTP status code (e.g. 200 for "200 Ok")
 432          * (see RFC2616, section 10, http://www.w3.org/Protocols/rfc2616/rfc2616-sec10.html
 433          * for a list of status codes.)
 434          *
 435          * @return int
 436          */
 437         public function getStatus() {
 438                 if ( !$this->respHeaders ) {
 439                         $this->parseHeader();
 440                 }
 441
 442                 return (int)$this->respStatus;
 443         }
 444
 445         /**
 446          * Returns true if the last status code was a redirect.
 447          *
 448          * @return bool
 449          */
 450         public function isRedirect() {
 451                 if ( !$this->respHeaders ) {
 452                         $this->parseHeader();
 453                 }
 454
 455                 $status = (int)$this->respStatus;
 456
 457                 if ( $status >= 300 && $status <= 303 ) {
 458                         return true;
 459                 }
 460
 461                 return false;
 462         }
 463
 464         /**
 465          * Returns an associative array of response headers after the
 466          * request has been executed.  Because some headers
 467          * (e.g. Set-Cookie) can appear more than once the, each value of
 468          * the associative array is an array of the values given.
 469          *
 470          * @return array
 471          */
 472         public function getResponseHeaders() {
 473                 if ( !$this->respHeaders ) {
 474                         $this->parseHeader();
 475                 }
 476
 477                 return $this->respHeaders;
 478         }
 479
 480         /**
 481          * Returns the value of the given response header.
 482          *
 483          * @param string $header
 484          * @return string|null
 485          */
 486         public function getResponseHeader( $header ) {
 487                 if ( !$this->respHeaders ) {
 488                         $this->parseHeader();
 489                 }
 490
 491                 if ( isset( $this->respHeaders[strtolower( $header )] ) ) {
 492                         $v = $this->respHeaders[strtolower( $header )];
 493                         return $v[count( $v ) - 1];
 494                 }
 495
 496                 return null;
 497         }
 498
 499         /**
 500          * Tells the MWHttpRequest object to use this pre-loaded CookieJar.
 501          *
 502          * To read response cookies from the jar, getCookieJar must be called first.
 503          *
 504          * @param CookieJar $jar
 505          */
 506         public function setCookieJar( $jar ) {
 507                 $this->cookieJar = $jar;
 508         }
 509
 510         /**
 511          * Returns the cookie jar in use.
 512          *
 513          * @return CookieJar
 514          */
 515         public function getCookieJar() {
 516                 if ( !$this->respHeaders ) {
 517                         $this->parseHeader();
 518                 }
 519
 520                 return $this->cookieJar;
 521         }
 522
 523         /**
 524          * Sets a cookie. Used before a request to set up any individual
 525          * cookies. Used internally after a request to parse the
 526          * Set-Cookie headers.
 527          * @see Cookie::set
 528          * @param string $name
 529          * @param string $value
 530          * @param array $attr
 531          */
 532         public function setCookie( $name, $value, $attr = [] ) {
 533                 if ( !$this->cookieJar ) {
 534                         $this->cookieJar = new CookieJar;
 535                 }
 536
 537                 if ( $this->parsedUrl && !isset( $attr['domain'] ) ) {
 538                         $attr['domain'] = $this->parsedUrl['host'];
 539                 }
 540
 541                 $this->cookieJar->setCookie( $name, $value, $attr );
 542         }
 543
 544         /**
 545          * Parse the cookies in the response headers and store them in the cookie jar.
 546          */
 547         protected function parseCookies() {
 548                 if ( !$this->cookieJar ) {
 549                         $this->cookieJar = new CookieJar;
 550                 }
 551
 552                 if ( isset( $this->respHeaders['set-cookie'] ) ) {
 553                         $url = parse_url( $this->getFinalUrl() );
 554                         foreach ( $this->respHeaders['set-cookie'] as $cookie ) {
 555                                 $this->cookieJar->parseCookieResponseHeader( $cookie, $url['host'] );
 556                         }
 557                 }
 558         }
 559
 560         /**
 561          * Returns the final URL after all redirections.
 562          *
 563          * Relative values of the "Location" header are incorrect as
 564          * stated in RFC, however they do happen and modern browsers
 565          * support them.  This function loops backwards through all
 566          * locations in order to build the proper absolute URI - Marooned
 567          * at wikia-inc.com
 568          *
 569          * Note that the multiple Location: headers are an artifact of
 570          * CURL -- they shouldn't actually get returned this way. Rewrite
 571          * this when T31232 is taken care of (high-level redirect
 572          * handling rewrite).
 573          *
 574          * @return string
 575          */
 576         public function getFinalUrl() {
 577                 $headers = $this->getResponseHeaders();
 578
 579                 // return full url (fix for incorrect but handled relative location)
 580                 if ( isset( $headers['location'] ) ) {
 581                         $locations = $headers['location'];
 582                         $domain = '';
 583                         $foundRelativeURI = false;
 584                         $countLocations = count( $locations );
 585
 586                         for ( $i = $countLocations - 1; $i >= 0; $i-- ) {
 587                                 $url = parse_url( $locations[$i] );
 588
 589                                 if ( isset( $url['host'] ) ) {
 590                                         $domain = $url['scheme'] . '://' . $url['host'];
 591                                         break; // found correct URI (with host)
 592                                 } else {
 593                                         $foundRelativeURI = true;
 594                                 }
 595                         }
 596
 597                         if ( !$foundRelativeURI ) {
 598                                 return $locations[$countLocations - 1];
 599                         }
 600                         if ( $domain ) {
 601                                 return $domain . $locations[$countLocations - 1];
 602                         }
 603                         $url = parse_url( $this->url );
 604                         if ( isset( $url['host'] ) ) {
 605                                 return $url['scheme'] . '://' . $url['host'] .
 606                                         $locations[$countLocations - 1];
 607                         }
 608                 }
 609
 610                 return $this->url;
 611         }
 612
 613         /**
 614          * Returns true if the backend can follow redirects. Overridden by the
 615          * child classes.
 616          * @return bool
 617          */
 618         public function canFollowRedirects() {
 619                 return true;
 620         }
 621
 622         /**
 623          * Set information about the original request. This can be useful for
 624          * endpoints/API modules which act as a proxy for some service, and
 625          * throttling etc. needs to happen in that service.
 626          * Calling this will result in the X-Forwarded-For and X-Original-User-Agent
 627          * headers being set.
 628          * @param WebRequest|array $originalRequest When in array form, it's
 629          *   expected to have the keys 'ip' and 'userAgent'.
 630          * @note IP/user agent is personally identifiable information, and should
 631          *   only be set when the privacy policy of the request target is
 632          *   compatible with that of the MediaWiki installation.
 633          */
 634         public function setOriginalRequest( $originalRequest ) {
 635                 if ( $originalRequest instanceof WebRequest ) {
 636                         $originalRequest = [
 637                                 'ip' => $originalRequest->getIP(),
 638                                 'userAgent' => $originalRequest->getHeader( 'User-Agent' ),
 639                         ];
 640                 } elseif (
 641                         !is_array( $originalRequest )
 642                         || array_diff( [ 'ip', 'userAgent' ], array_keys( $originalRequest ) )
 643                 ) {
 644                         throw new InvalidArgumentException( __METHOD__ . ': $originalRequest must be a '
 645                                 . "WebRequest or an array with 'ip' and 'userAgent' keys" );
 646                 }
 647
 648                 $this->reqHeaders['X-Forwarded-For'] = $originalRequest['ip'];
 649                 $this->reqHeaders['X-Original-User-Agent'] = $originalRequest['userAgent'];
 650         }
 651 }