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