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