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