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         /** @var array */
  50         protected $reqHeaders = [];
  51         protected $url;
  52         protected $parsedUrl;
  53         /** @var callable */
  54         protected $callback;
  55         protected $maxRedirects = 5;
  56         protected $followRedirects = false;
  57         protected $connectTimeout;
  58
  59         /**
  60          * @var CookieJar
  61          */
  62         protected $cookieJar;
  63
  64         protected $headerList = [];
  65         protected $respVersion = "0.9";
  66         protected $respStatus = "200 Ok";
  67         /** @var string[][] */
  68         protected $respHeaders = [];
  69
  70         /** @var StatusValue */
  71         protected $status;
  72
  73         /**
  74          * @var Profiler
  75          */
  76         protected $profiler;
  77
  78         /**
  79          * @var string
  80          */
  81         protected $profileName;
  82
  83         /**
  84          * @var LoggerInterface
  85          */
  86         protected $logger;
  87
  88         /**
  89          * @param string $url Url to use. If protocol-relative, will be expanded to an http:// URL
  90          * @param array $options (optional) extra params to pass (see HttpRequestFactory::create())
  91          * @codingStandardsIgnoreStart
  92          * @phan-param array{timeout?:int,connectTimeout?:int,postData?:array,proxy?:string,noProxy?:bool,sslVerifyHost?:bool,sslVerifyCert?:bool,caInfo?:string,maxRedirects?:int,followRedirects?:bool,userAgent?:string,logger?:LoggerInterface,username?:string,password?:string,originalRequest?:WebRequest|array{ip:string,userAgent:string},method?:string} $options
  93          * @codingStandardsIgnoreEnd
  94          * @param string $caller The method making this request, for profiling
  95          * @param Profiler|null $profiler An instance of the profiler for profiling, or null
  96          * @throws Exception
  97          */
  98         public function __construct(
  99                 $url, array $options = [], $caller = __METHOD__, Profiler $profiler = null
 100         ) {
 101                 global $wgHTTPTimeout, $wgHTTPConnectTimeout;
 102
 103                 $this->url = wfExpandUrl( $url, PROTO_HTTP );
 104                 $this->parsedUrl = wfParseUrl( $this->url );
 105
 106                 // @phan-suppress-next-line PhanTypeInvalidDimOffset
 107                 $this->logger = $options['logger'] ?? new NullLogger();
 108
 109                 if ( !$this->parsedUrl || !Http::isValidURI( $this->url ) ) {
 110                         $this->status = StatusValue::newFatal( 'http-invalid-url', $url );
 111                 } else {
 112                         $this->status = StatusValue::newGood( 100 ); // continue
 113                 }
 114
 115                 if ( isset( $options['timeout'] ) && $options['timeout'] != 'default' ) {
 116                         $this->timeout = $options['timeout'];
 117                 } else {
 118                         $this->timeout = $wgHTTPTimeout;
 119                 }
 120                 if ( isset( $options['connectTimeout'] ) && $options['connectTimeout'] != 'default' ) {
 121                         $this->connectTimeout = $options['connectTimeout'];
 122                 } else {
 123                         $this->connectTimeout = $wgHTTPConnectTimeout;
 124                 }
 125                 if ( isset( $options['userAgent'] ) ) {
 126                         $this->setUserAgent( $options['userAgent'] );
 127                 }
 128                 if ( isset( $options['username'] ) && isset( $options['password'] ) ) {
 129                         $this->setHeader(
 130                                 'Authorization',
 131                                 'Basic ' . base64_encode( $options['username'] . ':' . $options['password'] )
 132                         );
 133                 }
 134                 if ( isset( $options['originalRequest'] ) ) {
 135                         $this->setOriginalRequest( $options['originalRequest'] );
 136                 }
 137
 138                 $this->setHeader( 'X-Request-Id', WebRequest::getRequestId() );
 139
 140                 $members = [ "postData", "proxy", "noProxy", "sslVerifyHost", "caInfo",
 141                                 "method", "followRedirects", "maxRedirects", "sslVerifyCert", "callback" ];
 142
 143                 foreach ( $members as $o ) {
 144                         if ( isset( $options[$o] ) ) {
 145                                 // ensure that MWHttpRequest::method is always
 146                                 // uppercased. T38137
 147                                 if ( $o == 'method' ) {
 148                                         // @phan-suppress-next-line PhanTypeInvalidDimOffset
 149                                         $options[$o] = strtoupper( $options[$o] );
 150                                 }
 151                                 $this->$o = $options[$o];
 152                         }
 153                 }
 154
 155                 if ( $this->noProxy ) {
 156                         $this->proxy = ''; // noProxy takes precedence
 157                 }
 158
 159                 // Profile based on what's calling us
 160                 $this->profiler = $profiler;
 161                 $this->profileName = $caller;
 162         }
 163
 164         /**
 165          * @param LoggerInterface $logger
 166          */
 167         public function setLogger( LoggerInterface $logger ) {
 168                 $this->logger = $logger;
 169         }
 170
 171         /**
 172          * Simple function to test if we can make any sort of requests at all, using
 173          * cURL or fopen()
 174          * @return bool
 175          */
 176         public static function canMakeRequests() {
 177                 return function_exists( 'curl_init' ) || wfIniGetBool( 'allow_url_fopen' );
 178         }
 179
 180         /**
 181          * Generate a new request object
 182          * @deprecated since 1.34, use HttpRequestFactory instead
 183          * @param string $url Url to use
 184          * @param array|null $options (optional) extra params to pass (see HttpRequestFactory::create())
 185          * @param string $caller The method making this request, for profiling
 186          * @throws DomainException
 187          * @return MWHttpRequest
 188          * @see MWHttpRequest::__construct
 189          */
 190         public static function factory( $url, array $options = null, $caller = __METHOD__ ) {
 191                 if ( $options === null ) {
 192                         $options = [];
 193                 }
 194                 return \MediaWiki\MediaWikiServices::getInstance()
 195                         ->getHttpRequestFactory()
 196                         ->create( $url, $options, $caller );
 197         }
 198
 199         /**
 200          * Get the body, or content, of the response to the request
 201          *
 202          * @return string
 203          */
 204         public function getContent() {
 205                 return $this->content;
 206         }
 207
 208         /**
 209          * Set the parameters of the request
 210          *
 211          * @param array $args
 212          * @todo overload the args param
 213          */
 214         public function setData( array $args ) {
 215                 $this->postData = $args;
 216         }
 217
 218         /**
 219          * Take care of setting up the proxy (do nothing if "noProxy" is set)
 220          *
 221          * @return void
 222          */
 223         protected function proxySetup() {
 224                 // If there is an explicit proxy set and proxies are not disabled, then use it
 225                 if ( $this->proxy && !$this->noProxy ) {
 226                         return;
 227                 }
 228
 229                 // Otherwise, fallback to $wgHTTPProxy if this is not a machine
 230                 // local URL and proxies are not disabled
 231                 if ( self::isLocalURL( $this->url ) || $this->noProxy ) {
 232                         $this->proxy = '';
 233                 } else {
 234                         global $wgHTTPProxy;
 235                         $this->proxy = (string)$wgHTTPProxy;
 236                 }
 237         }
 238
 239         /**
 240          * Check if the URL can be served by localhost
 241          *
 242          * @param string $url Full url to check
 243          * @return bool
 244          */
 245         private static function isLocalURL( $url ) {
 246                 global $wgCommandLineMode, $wgLocalVirtualHosts;
 247
 248                 if ( $wgCommandLineMode ) {
 249                         return false;
 250                 }
 251
 252                 // Extract host part
 253                 $matches = [];
 254                 if ( preg_match( '!^https?://([\w.-]+)[/:].*$!', $url, $matches ) ) {
 255                         $host = $matches[1];
 256                         // Split up dotwise
 257                         $domainParts = explode( '.', $host );
 258                         // Check if this domain or any superdomain is listed as a local virtual host
 259                         $domainParts = array_reverse( $domainParts );
 260
 261                         $domain = '';
 262                         $countParts = count( $domainParts );
 263                         for ( $i = 0; $i < $countParts; $i++ ) {
 264                                 $domainPart = $domainParts[$i];
 265                                 if ( $i == 0 ) {
 266                                         $domain = $domainPart;
 267                                 } else {
 268                                         $domain = $domainPart . '.' . $domain;
 269                                 }
 270
 271                                 if ( in_array( $domain, $wgLocalVirtualHosts ) ) {
 272                                         return true;
 273                                 }
 274                         }
 275                 }
 276
 277                 return false;
 278         }
 279
 280         /**
 281          * Set the user agent
 282          * @param string $UA
 283          */
 284         public function setUserAgent( $UA ) {
 285                 $this->setHeader( 'User-Agent', $UA );
 286         }
 287
 288         /**
 289          * Set an arbitrary header
 290          * @param string $name
 291          * @param string $value
 292          */
 293         public function setHeader( $name, $value ) {
 294                 // I feel like I should normalize the case here...
 295                 $this->reqHeaders[$name] = $value;
 296         }
 297
 298         /**
 299          * Get an array of the headers
 300          * @return array
 301          */
 302         protected function getHeaderList() {
 303                 $list = [];
 304
 305                 if ( $this->cookieJar ) {
 306                         $this->reqHeaders['Cookie'] =
 307                                 $this->cookieJar->serializeToHttpRequest(
 308                                         $this->parsedUrl['path'],
 309                                         $this->parsedUrl['host']
 310                                 );
 311                 }
 312
 313                 foreach ( $this->reqHeaders as $name => $value ) {
 314                         $list[] = "$name: $value";
 315                 }
 316
 317                 return $list;
 318         }
 319
 320         /**
 321          * Set a read callback to accept data read from the HTTP request.
 322          * By default, data is appended to an internal buffer which can be
 323          * retrieved through $req->getContent().
 324          *
 325          * To handle data as it comes in -- especially for large files that
 326          * would not fit in memory -- you can instead set your own callback,
 327          * in the form function($resource, $buffer) where the first parameter
 328          * is the low-level resource being read (implementation specific),
 329          * and the second parameter is the data buffer.
 330          *
 331          * You MUST return the number of bytes handled in the buffer; if fewer
 332          * bytes are reported handled than were passed to you, the HTTP fetch
 333          * will be aborted.
 334          *
 335          * @param callable|null $callback
 336          * @throws InvalidArgumentException
 337          */
 338         public function setCallback( $callback ) {
 339                 return $this->doSetCallback( $callback );
 340         }
 341
 342         /**
 343          * Worker function for setting callbacks.  Calls can originate both internally and externally
 344          * via setCallback).  Defaults to the internal read callback if $callback is null.
 345          *
 346          * @param callable|null $callback
 347          * @throws InvalidArgumentException
 348          */
 349         protected function doSetCallback( $callback ) {
 350                 if ( is_null( $callback ) ) {
 351                         $callback = [ $this, 'read' ];
 352                 } elseif ( !is_callable( $callback ) ) {
 353                         $this->status->fatal( 'http-internal-error' );
 354                         throw new InvalidArgumentException( __METHOD__ . ': invalid callback' );
 355                 }
 356                 $this->callback = $callback;
 357         }
 358
 359         /**
 360          * A generic callback to read the body of the response from a remote
 361          * server.
 362          *
 363          * @param resource $fh
 364          * @param string $content
 365          * @return int
 366          * @internal
 367          */
 368         public function read( $fh, $content ) {
 369                 $this->content .= $content;
 370                 return strlen( $content );
 371         }
 372
 373         /**
 374          * Take care of whatever is necessary to perform the URI request.
 375          *
 376          * @return StatusValue
 377          * @note currently returns Status for B/C
 378          */
 379         public function execute() {
 380                 throw new LogicException( 'children must override this' );
 381         }
 382
 383         protected function prepare() {
 384                 $this->content = "";
 385
 386                 if ( strtoupper( $this->method ) == "HEAD" ) {
 387                         $this->headersOnly = true;
 388                 }
 389
 390                 $this->proxySetup(); // set up any proxy as needed
 391
 392                 if ( !$this->callback ) {
 393                         $this->doSetCallback( null );
 394                 }
 395
 396                 if ( !isset( $this->reqHeaders['User-Agent'] ) ) {
 397                         $this->setUserAgent( Http::userAgent() );
 398                 }
 399         }
 400
 401         /**
 402          * Parses the headers, including the HTTP status code and any
 403          * Set-Cookie headers.  This function expects the headers to be
 404          * found in an array in the member variable headerList.
 405          */
 406         protected function parseHeader() {
 407                 $lastname = "";
 408
 409                 // Failure without (valid) headers gets a response status of zero
 410                 if ( !$this->status->isOK() ) {
 411                         $this->respStatus = '0 Error';
 412                 }
 413
 414                 foreach ( $this->headerList as $header ) {
 415                         if ( preg_match( "#^HTTP/([0-9.]+) (.*)#", $header, $match ) ) {
 416                                 $this->respVersion = $match[1];
 417                                 $this->respStatus = $match[2];
 418                         } elseif ( preg_match( "#^[ \t]#", $header ) ) {
 419                                 $last = count( $this->respHeaders[$lastname] ) - 1;
 420                                 $this->respHeaders[$lastname][$last] .= "\r\n$header";
 421                         } elseif ( preg_match( "#^([^:]*):[\t ]*(.*)#", $header, $match ) ) {
 422                                 $this->respHeaders[strtolower( $match[1] )][] = $match[2];
 423                                 $lastname = strtolower( $match[1] );
 424                         }
 425                 }
 426
 427                 $this->parseCookies();
 428         }
 429
 430         /**
 431          * Sets HTTPRequest status member to a fatal value with the error
 432          * message if the returned integer value of the status code was
 433          * not successful (1-299) or a redirect (300-399).
 434          * See RFC2616, section 10, http://www.w3.org/Protocols/rfc2616/rfc2616-sec10.html
 435          * for a list of status codes.
 436          */
 437         protected function setStatus() {
 438                 if ( !$this->respHeaders ) {
 439                         $this->parseHeader();
 440                 }
 441
 442                 if ( ( (int)$this->respStatus > 0 && (int)$this->respStatus < 400 ) ) {
 443                         $this->status->setResult( true, (int)$this->respStatus );
 444                 } else {
 445                         list( $code, $message ) = explode( " ", $this->respStatus, 2 );
 446                         $this->status->setResult( false, (int)$this->respStatus );
 447                         $this->status->fatal( "http-bad-status", $code, $message );
 448                 }
 449         }
 450
 451         /**
 452          * Get the integer value of the HTTP status code (e.g. 200 for "200 Ok")
 453          * (see RFC2616, section 10, http://www.w3.org/Protocols/rfc2616/rfc2616-sec10.html
 454          * for a list of status codes.)
 455          *
 456          * @return int
 457          */
 458         public function getStatus() {
 459                 if ( !$this->respHeaders ) {
 460                         $this->parseHeader();
 461                 }
 462
 463                 return (int)$this->respStatus;
 464         }
 465
 466         /**
 467          * Returns true if the last status code was a redirect.
 468          *
 469          * @return bool
 470          */
 471         public function isRedirect() {
 472                 if ( !$this->respHeaders ) {
 473                         $this->parseHeader();
 474                 }
 475
 476                 $status = (int)$this->respStatus;
 477
 478                 if ( $status >= 300 && $status <= 303 ) {
 479                         return true;
 480                 }
 481
 482                 return false;
 483         }
 484
 485         /**
 486          * Returns an associative array of response headers after the
 487          * request has been executed.  Because some headers
 488          * (e.g. Set-Cookie) can appear more than once the, each value of
 489          * the associative array is an array of the values given.
 490          * Header names are always in lowercase.
 491          *
 492          * @return array
 493          */
 494         public function getResponseHeaders() {
 495                 if ( !$this->respHeaders ) {
 496                         $this->parseHeader();
 497                 }
 498
 499                 return $this->respHeaders;
 500         }
 501
 502         /**
 503          * Returns the value of the given response header.
 504          *
 505          * @param string $header case-insensitive
 506          * @return string|null
 507          */
 508         public function getResponseHeader( $header ) {
 509                 if ( !$this->respHeaders ) {
 510                         $this->parseHeader();
 511                 }
 512
 513                 if ( isset( $this->respHeaders[strtolower( $header )] ) ) {
 514                         $v = $this->respHeaders[strtolower( $header )];
 515                         return $v[count( $v ) - 1];
 516                 }
 517
 518                 return null;
 519         }
 520
 521         /**
 522          * Tells the MWHttpRequest object to use this pre-loaded CookieJar.
 523          *
 524          * To read response cookies from the jar, getCookieJar must be called first.
 525          *
 526          * @param CookieJar $jar
 527          */
 528         public function setCookieJar( CookieJar $jar ) {
 529                 $this->cookieJar = $jar;
 530         }
 531
 532         /**
 533          * Returns the cookie jar in use.
 534          *
 535          * @return CookieJar
 536          */
 537         public function getCookieJar() {
 538                 if ( !$this->respHeaders ) {
 539                         $this->parseHeader();
 540                 }
 541
 542                 return $this->cookieJar;
 543         }
 544
 545         /**
 546          * Sets a cookie. Used before a request to set up any individual
 547          * cookies. Used internally after a request to parse the
 548          * Set-Cookie headers.
 549          * @see Cookie::set
 550          * @param string $name
 551          * @param string $value
 552          * @param array $attr
 553          */
 554         public function setCookie( $name, $value, array $attr = [] ) {
 555                 if ( !$this->cookieJar ) {
 556                         $this->cookieJar = new CookieJar;
 557                 }
 558
 559                 if ( $this->parsedUrl && !isset( $attr['domain'] ) ) {
 560                         $attr['domain'] = $this->parsedUrl['host'];
 561                 }
 562
 563                 $this->cookieJar->setCookie( $name, $value, $attr );
 564         }
 565
 566         /**
 567          * Parse the cookies in the response headers and store them in the cookie jar.
 568          */
 569         protected function parseCookies() {
 570                 if ( !$this->cookieJar ) {
 571                         $this->cookieJar = new CookieJar;
 572                 }
 573
 574                 if ( isset( $this->respHeaders['set-cookie'] ) ) {
 575                         $url = parse_url( $this->getFinalUrl() );
 576                         foreach ( $this->respHeaders['set-cookie'] as $cookie ) {
 577                                 $this->cookieJar->parseCookieResponseHeader( $cookie, $url['host'] );
 578                         }
 579                 }
 580         }
 581
 582         /**
 583          * Returns the final URL after all redirections.
 584          *
 585          * Relative values of the "Location" header are incorrect as
 586          * stated in RFC, however they do happen and modern browsers
 587          * support them.  This function loops backwards through all
 588          * locations in order to build the proper absolute URI - Marooned
 589          * at wikia-inc.com
 590          *
 591          * Note that the multiple Location: headers are an artifact of
 592          * CURL -- they shouldn't actually get returned this way. Rewrite
 593          * this when T31232 is taken care of (high-level redirect
 594          * handling rewrite).
 595          *
 596          * @return string
 597          */
 598         public function getFinalUrl() {
 599                 $headers = $this->getResponseHeaders();
 600
 601                 // return full url (fix for incorrect but handled relative location)
 602                 if ( isset( $headers['location'] ) ) {
 603                         $locations = $headers['location'];
 604                         $domain = '';
 605                         $foundRelativeURI = false;
 606                         $countLocations = count( $locations );
 607
 608                         for ( $i = $countLocations - 1; $i >= 0; $i-- ) {
 609                                 $url = parse_url( $locations[$i] );
 610
 611                                 if ( isset( $url['host'] ) ) {
 612                                         $domain = $url['scheme'] . '://' . $url['host'];
 613                                         break; // found correct URI (with host)
 614                                 } else {
 615                                         $foundRelativeURI = true;
 616                                 }
 617                         }
 618
 619                         if ( !$foundRelativeURI ) {
 620                                 return $locations[$countLocations - 1];
 621                         }
 622                         if ( $domain ) {
 623                                 return $domain . $locations[$countLocations - 1];
 624                         }
 625                         $url = parse_url( $this->url );
 626                         if ( isset( $url['host'] ) ) {
 627                                 return $url['scheme'] . '://' . $url['host'] .
 628                                         $locations[$countLocations - 1];
 629                         }
 630                 }
 631
 632                 return $this->url;
 633         }
 634
 635         /**
 636          * Returns true if the backend can follow redirects. Overridden by the
 637          * child classes.
 638          * @return bool
 639          */
 640         public function canFollowRedirects() {
 641                 return true;
 642         }
 643
 644         /**
 645          * Set information about the original request. This can be useful for
 646          * endpoints/API modules which act as a proxy for some service, and
 647          * throttling etc. needs to happen in that service.
 648          * Calling this will result in the X-Forwarded-For and X-Original-User-Agent
 649          * headers being set.
 650          * @param WebRequest|array $originalRequest When in array form, it's
 651          *   expected to have the keys 'ip' and 'userAgent'.
 652          * @note IP/user agent is personally identifiable information, and should
 653          *   only be set when the privacy policy of the request target is
 654          *   compatible with that of the MediaWiki installation.
 655          */
 656         public function setOriginalRequest( $originalRequest ) {
 657                 if ( $originalRequest instanceof WebRequest ) {
 658                         $originalRequest = [
 659                                 'ip' => $originalRequest->getIP(),
 660                                 'userAgent' => $originalRequest->getHeader( 'User-Agent' ),
 661                         ];
 662                 } elseif (
 663                         !is_array( $originalRequest )
 664                         || array_diff( [ 'ip', 'userAgent' ], array_keys( $originalRequest ) )
 665                 ) {
 666                         throw new InvalidArgumentException( __METHOD__ . ': $originalRequest must be a '
 667                                 . "WebRequest or an array with 'ip' and 'userAgent' keys" );
 668                 }
 669
 670                 $this->reqHeaders['X-Forwarded-For'] = $originalRequest['ip'];
 671                 $this->reqHeaders['X-Original-User-Agent'] = $originalRequest['userAgent'];
 672         }
 673
 674         /**
 675          * Check that the given URI is a valid one.
 676          *
 677          * This hardcodes a small set of protocols only, because we want to
 678          * deterministically reject protocols not supported by all HTTP-transport
 679          * methods.
 680          *
 681          * "file://" specifically must not be allowed, for security reasons
 682          * (see <https://www.mediawiki.org/wiki/Special:Code/MediaWiki/r67684>).
 683          *
 684          * @todo FIXME this is wildly inaccurate and fails to actually check most stuff
 685          *
 686          * @since 1.34
 687          * @param string $uri URI to check for validity
 688          * @return bool
 689          */
 690         public static function isValidURI( $uri ) {
 691                 return (bool)preg_match(
 692                         '/^https?:\/\/[^\/\s]\S*$/D',
 693                         $uri
 694                 );
 695         }
 696 }