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