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 HttpRequestFactory::create())
  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 since 1.34, use HttpRequestFactory instead
 176          * @param string $url Url to use
 177          * @param array|null $options (optional) extra params to pass (see HttpRequestFactory::create())
 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                         global $wgHTTPProxy;
 228                         $this->proxy = (string)$wgHTTPProxy;
 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                 return $this->doSetCallback( $callback );
 333         }
 334
 335         /**
 336          * Worker function for setting callbacks.  Calls can originate both internally and externally
 337          * via setCallback).  Defaults to the internal read callback if $callback is null.
 338          *
 339          * @param callable|null $callback
 340          * @throws InvalidArgumentException
 341          */
 342         protected function doSetCallback( $callback ) {
 343                 if ( is_null( $callback ) ) {
 344                         $callback = [ $this, 'read' ];
 345                 } elseif ( !is_callable( $callback ) ) {
 346                         $this->status->fatal( 'http-internal-error' );
 347                         throw new InvalidArgumentException( __METHOD__ . ': invalid callback' );
 348                 }
 349                 $this->callback = $callback;
 350         }
 351
 352         /**
 353          * A generic callback to read the body of the response from a remote
 354          * server.
 355          *
 356          * @param resource $fh
 357          * @param string $content
 358          * @return int
 359          * @internal
 360          */
 361         public function read( $fh, $content ) {
 362                 $this->content .= $content;
 363                 return strlen( $content );
 364         }
 365
 366         /**
 367          * Take care of whatever is necessary to perform the URI request.
 368          *
 369          * @return StatusValue
 370          * @note currently returns Status for B/C
 371          */
 372         public function execute() {
 373                 throw new LogicException( 'children must override this' );
 374         }
 375
 376         protected function prepare() {
 377                 $this->content = "";
 378
 379                 if ( strtoupper( $this->method ) == "HEAD" ) {
 380                         $this->headersOnly = true;
 381                 }
 382
 383                 $this->proxySetup(); // set up any proxy as needed
 384
 385                 if ( !$this->callback ) {
 386                         $this->doSetCallback( null );
 387                 }
 388
 389                 if ( !isset( $this->reqHeaders['User-Agent'] ) ) {
 390                         $this->setUserAgent( Http::userAgent() );
 391                 }
 392         }
 393
 394         /**
 395          * Parses the headers, including the HTTP status code and any
 396          * Set-Cookie headers.  This function expects the headers to be
 397          * found in an array in the member variable headerList.
 398          */
 399         protected function parseHeader() {
 400                 $lastname = "";
 401
 402                 // Failure without (valid) headers gets a response status of zero
 403                 if ( !$this->status->isOK() ) {
 404                         $this->respStatus = '0 Error';
 405                 }
 406
 407                 foreach ( $this->headerList as $header ) {
 408                         if ( preg_match( "#^HTTP/([0-9.]+) (.*)#", $header, $match ) ) {
 409                                 $this->respVersion = $match[1];
 410                                 $this->respStatus = $match[2];
 411                         } elseif ( preg_match( "#^[ \t]#", $header ) ) {
 412                                 $last = count( $this->respHeaders[$lastname] ) - 1;
 413                                 $this->respHeaders[$lastname][$last] .= "\r\n$header";
 414                         } elseif ( preg_match( "#^([^:]*):[\t ]*(.*)#", $header, $match ) ) {
 415                                 $this->respHeaders[strtolower( $match[1] )][] = $match[2];
 416                                 $lastname = strtolower( $match[1] );
 417                         }
 418                 }
 419
 420                 $this->parseCookies();
 421         }
 422
 423         /**
 424          * Sets HTTPRequest status member to a fatal value with the error
 425          * message if the returned integer value of the status code was
 426          * not successful (1-299) or a redirect (300-399).
 427          * See RFC2616, section 10, http://www.w3.org/Protocols/rfc2616/rfc2616-sec10.html
 428          * for a list of status codes.
 429          */
 430         protected function setStatus() {
 431                 if ( !$this->respHeaders ) {
 432                         $this->parseHeader();
 433                 }
 434
 435                 if ( ( (int)$this->respStatus > 0 && (int)$this->respStatus < 400 ) ) {
 436                         $this->status->setResult( true, (int)$this->respStatus );
 437                 } else {
 438                         list( $code, $message ) = explode( " ", $this->respStatus, 2 );
 439                         $this->status->setResult( false, (int)$this->respStatus );
 440                         $this->status->fatal( "http-bad-status", $code, $message );
 441                 }
 442         }
 443
 444         /**
 445          * Get the integer value of the HTTP status code (e.g. 200 for "200 Ok")
 446          * (see RFC2616, section 10, http://www.w3.org/Protocols/rfc2616/rfc2616-sec10.html
 447          * for a list of status codes.)
 448          *
 449          * @return int
 450          */
 451         public function getStatus() {
 452                 if ( !$this->respHeaders ) {
 453                         $this->parseHeader();
 454                 }
 455
 456                 return (int)$this->respStatus;
 457         }
 458
 459         /**
 460          * Returns true if the last status code was a redirect.
 461          *
 462          * @return bool
 463          */
 464         public function isRedirect() {
 465                 if ( !$this->respHeaders ) {
 466                         $this->parseHeader();
 467                 }
 468
 469                 $status = (int)$this->respStatus;
 470
 471                 if ( $status >= 300 && $status <= 303 ) {
 472                         return true;
 473                 }
 474
 475                 return false;
 476         }
 477
 478         /**
 479          * Returns an associative array of response headers after the
 480          * request has been executed.  Because some headers
 481          * (e.g. Set-Cookie) can appear more than once the, each value of
 482          * the associative array is an array of the values given.
 483          * Header names are always in lowercase.
 484          *
 485          * @return array
 486          */
 487         public function getResponseHeaders() {
 488                 if ( !$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 case-insensitive
 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( CookieJar $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, array $attr = [] ) {
 548                 if ( !$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
 667         /**
 668          * Check that the given URI is a valid one.
 669          *
 670          * This hardcodes a small set of protocols only, because we want to
 671          * deterministically reject protocols not supported by all HTTP-transport
 672          * methods.
 673          *
 674          * "file://" specifically must not be allowed, for security reasons
 675          * (see <https://www.mediawiki.org/wiki/Special:Code/MediaWiki/r67684>).
 676          *
 677          * @todo FIXME this is wildly inaccurate and fails to actually check most stuff
 678          *
 679          * @since 1.34
 680          * @param string $uri URI to check for validity
 681          * @return bool
 682          */
 683         public static function isValidURI( $uri ) {
 684                 return (bool)preg_match(
 685                         '/^https?:\/\/[^\/\s]\S*$/D',
 686                         $uri
 687                 );
 688         }
 689 }