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 MediaWiki\Logger\LoggerFactory;
  22 use Psr\Log\LoggerInterface;
  23 use Psr\Log\LoggerAwareInterface;
  24 use Psr\Log\NullLogger;
  25
  26 /**
  27  * This wrapper class will call out to curl (if available) or fallback
  28  * to regular PHP if necessary for handling internal HTTP requests.
  29  *
  30  * Renamed from HttpRequest to MWHttpRequest to avoid conflict with
  31  * PHP's HTTP extension.
  32  */
  33 class MWHttpRequest implements LoggerAwareInterface {
  34         const SUPPORTS_FILE_POSTS = false;
  35
  36         protected $content;
  37         protected $timeout = 'default';
  38         protected $headersOnly = null;
  39         protected $postData = null;
  40         protected $proxy = null;
  41         protected $noProxy = false;
  42         protected $sslVerifyHost = true;
  43         protected $sslVerifyCert = true;
  44         protected $caInfo = null;
  45         protected $method = "GET";
  46         protected $reqHeaders = [];
  47         protected $url;
  48         protected $parsedUrl;
  49         protected $callback;
  50         protected $maxRedirects = 5;
  51         protected $followRedirects = false;
  52
  53         /**
  54          * @var CookieJar
  55          */
  56         protected $cookieJar;
  57
  58         protected $headerList = [];
  59         protected $respVersion = "0.9";
  60         protected $respStatus = "200 Ok";
  61         protected $respHeaders = [];
  62
  63         public $status;
  64
  65         /**
  66          * @var Profiler
  67          */
  68         protected $profiler;
  69
  70         /**
  71          * @var string
  72          */
  73         protected $profileName;
  74
  75         /**
  76          * @var LoggerInterface;
  77          */
  78         protected $logger;
  79
  80         /**
  81          * @param string $url Url to use. If protocol-relative, will be expanded to an http:// URL
  82          * @param array $options (optional) extra params to pass (see Http::request())
  83          * @param string $caller The method making this request, for profiling
  84          * @param Profiler $profiler An instance of the profiler for profiling, or null
  85          */
  86         protected function __construct(
  87                 $url, $options = [], $caller = __METHOD__, $profiler = null
  88         ) {
  89                 global $wgHTTPTimeout, $wgHTTPConnectTimeout;
  90
  91                 $this->url = wfExpandUrl( $url, PROTO_HTTP );
  92                 $this->parsedUrl = wfParseUrl( $this->url );
  93
  94                 if ( isset( $options['logger'] ) ) {
  95                         $this->logger = $options['logger'];
  96                 } else {
  97                         $this->logger = new NullLogger();
  98                 }
  99
 100                 if ( !$this->parsedUrl || !Http::isValidURI( $this->url ) ) {
 101                         $this->status = Status::newFatal( 'http-invalid-url', $url );
 102                 } else {
 103                         $this->status = Status::newGood( 100 ); // continue
 104                 }
 105
 106                 if ( isset( $options['timeout'] ) && $options['timeout'] != 'default' ) {
 107                         $this->timeout = $options['timeout'];
 108                 } else {
 109                         $this->timeout = $wgHTTPTimeout;
 110                 }
 111                 if ( isset( $options['connectTimeout'] ) && $options['connectTimeout'] != 'default' ) {
 112                         $this->connectTimeout = $options['connectTimeout'];
 113                 } else {
 114                         $this->connectTimeout = $wgHTTPConnectTimeout;
 115                 }
 116                 if ( isset( $options['userAgent'] ) ) {
 117                         $this->setUserAgent( $options['userAgent'] );
 118                 }
 119                 if ( isset( $options['username'] ) && isset( $options['password'] ) ) {
 120                         $this->setHeader(
 121                                 'Authorization',
 122                                 'Basic ' . base64_encode( $options['username'] . ':' . $options['password'] )
 123                         );
 124                 }
 125
 126                 $members = [ "postData", "proxy", "noProxy", "sslVerifyHost", "caInfo",
 127                                 "method", "followRedirects", "maxRedirects", "sslVerifyCert", "callback" ];
 128
 129                 foreach ( $members as $o ) {
 130                         if ( isset( $options[$o] ) ) {
 131                                 // ensure that MWHttpRequest::method is always
 132                                 // uppercased. Bug 36137
 133                                 if ( $o == 'method' ) {
 134                                         $options[$o] = strtoupper( $options[$o] );
 135                                 }
 136                                 $this->$o = $options[$o];
 137                         }
 138                 }
 139
 140                 if ( $this->noProxy ) {
 141                         $this->proxy = ''; // noProxy takes precedence
 142                 }
 143
 144                 // Profile based on what's calling us
 145                 $this->profiler = $profiler;
 146                 $this->profileName = $caller;
 147         }
 148
 149         /**
 150          * @param LoggerInterface $logger
 151          */
 152         public function setLogger( LoggerInterface $logger ) {
 153                 $this->logger = $logger;
 154         }
 155
 156         /**
 157          * Simple function to test if we can make any sort of requests at all, using
 158          * cURL or fopen()
 159          * @return bool
 160          */
 161         public static function canMakeRequests() {
 162                 return function_exists( 'curl_init' ) || wfIniGetBool( 'allow_url_fopen' );
 163         }
 164
 165         /**
 166          * Generate a new request object
 167          * @param string $url Url to use
 168          * @param array $options (optional) extra params to pass (see Http::request())
 169          * @param string $caller The method making this request, for profiling
 170          * @throws MWException
 171          * @return CurlHttpRequest|PhpHttpRequest
 172          * @see MWHttpRequest::__construct
 173          */
 174         public static function factory( $url, $options = null, $caller = __METHOD__ ) {
 175                 if ( !Http::$httpEngine ) {
 176                         Http::$httpEngine = function_exists( 'curl_init' ) ? 'curl' : 'php';
 177                 } elseif ( Http::$httpEngine == 'curl' && !function_exists( 'curl_init' ) ) {
 178                         throw new MWException( __METHOD__ . ': curl (http://php.net/curl) is not installed, but' .
 179                                 ' Http::$httpEngine is set to "curl"' );
 180                 }
 181
 182                 if ( !is_array( $options ) ) {
 183                         $options = [];
 184                 }
 185
 186                 if ( !isset( $options['logger'] ) ) {
 187                         $options['logger'] = LoggerFactory::getInstance( 'http' );
 188                 }
 189
 190                 switch ( Http::$httpEngine ) {
 191                         case 'curl':
 192                                 return new CurlHttpRequest( $url, $options, $caller, Profiler::instance() );
 193                         case 'php':
 194                                 if ( !wfIniGetBool( 'allow_url_fopen' ) ) {
 195                                         throw new MWException( __METHOD__ . ': allow_url_fopen ' .
 196                                                 'needs to be enabled for pure PHP http requests to ' .
 197                                                 'work. If possible, curl should be used instead. See ' .
 198                                                 'http://php.net/curl.'
 199                                         );
 200                                 }
 201                                 return new PhpHttpRequest( $url, $options, $caller, Profiler::instance() );
 202                         default:
 203                                 throw new MWException( __METHOD__ . ': The setting of Http::$httpEngine is not valid.' );
 204                 }
 205         }
 206
 207         /**
 208          * Get the body, or content, of the response to the request
 209          *
 210          * @return string
 211          */
 212         public function getContent() {
 213                 return $this->content;
 214         }
 215
 216         /**
 217          * Set the parameters of the request
 218          *
 219          * @param array $args
 220          * @todo overload the args param
 221          */
 222         public function setData( $args ) {
 223                 $this->postData = $args;
 224         }
 225
 226         /**
 227          * Take care of setting up the proxy (do nothing if "noProxy" is set)
 228          *
 229          * @return void
 230          */
 231         public function proxySetup() {
 232                 // If there is an explicit proxy set and proxies are not disabled, then use it
 233                 if ( $this->proxy && !$this->noProxy ) {
 234                         return;
 235                 }
 236
 237                 // Otherwise, fallback to $wgHTTPProxy if this is not a machine
 238                 // local URL and proxies are not disabled
 239                 if ( self::isLocalURL( $this->url ) || $this->noProxy ) {
 240                         $this->proxy = '';
 241                 } else {
 242                         $this->proxy = Http::getProxy();
 243                 }
 244         }
 245
 246         /**
 247          * Check if the URL can be served by localhost
 248          *
 249          * @param string $url Full url to check
 250          * @return bool
 251          */
 252         private static function isLocalURL( $url ) {
 253                 global $wgCommandLineMode, $wgLocalVirtualHosts;
 254
 255                 if ( $wgCommandLineMode ) {
 256                         return false;
 257                 }
 258
 259                 // Extract host part
 260                 $matches = [];
 261                 if ( preg_match( '!^https?://([\w.-]+)[/:].*$!', $url, $matches ) ) {
 262                         $host = $matches[1];
 263                         // Split up dotwise
 264                         $domainParts = explode( '.', $host );
 265                         // Check if this domain or any superdomain is listed as a local virtual host
 266                         $domainParts = array_reverse( $domainParts );
 267
 268                         $domain = '';
 269                         $countParts = count( $domainParts );
 270                         for ( $i = 0; $i < $countParts; $i++ ) {
 271                                 $domainPart = $domainParts[$i];
 272                                 if ( $i == 0 ) {
 273                                         $domain = $domainPart;
 274                                 } else {
 275                                         $domain = $domainPart . '.' . $domain;
 276                                 }
 277
 278                                 if ( in_array( $domain, $wgLocalVirtualHosts ) ) {
 279                                         return true;
 280                                 }
 281                         }
 282                 }
 283
 284                 return false;
 285         }
 286
 287         /**
 288          * Set the user agent
 289          * @param string $UA
 290          */
 291         public function setUserAgent( $UA ) {
 292                 $this->setHeader( 'User-Agent', $UA );
 293         }
 294
 295         /**
 296          * Set an arbitrary header
 297          * @param string $name
 298          * @param string $value
 299          */
 300         public function setHeader( $name, $value ) {
 301                 // I feel like I should normalize the case here...
 302                 $this->reqHeaders[$name] = $value;
 303         }
 304
 305         /**
 306          * Get an array of the headers
 307          * @return array
 308          */
 309         public function getHeaderList() {
 310                 $list = [];
 311
 312                 if ( $this->cookieJar ) {
 313                         $this->reqHeaders['Cookie'] =
 314                                 $this->cookieJar->serializeToHttpRequest(
 315                                         $this->parsedUrl['path'],
 316                                         $this->parsedUrl['host']
 317                                 );
 318                 }
 319
 320                 foreach ( $this->reqHeaders as $name => $value ) {
 321                         $list[] = "$name: $value";
 322                 }
 323
 324                 return $list;
 325         }
 326
 327         /**
 328          * Set a read callback to accept data read from the HTTP request.
 329          * By default, data is appended to an internal buffer which can be
 330          * retrieved through $req->getContent().
 331          *
 332          * To handle data as it comes in -- especially for large files that
 333          * would not fit in memory -- you can instead set your own callback,
 334          * in the form function($resource, $buffer) where the first parameter
 335          * is the low-level resource being read (implementation specific),
 336          * and the second parameter is the data buffer.
 337          *
 338          * You MUST return the number of bytes handled in the buffer; if fewer
 339          * bytes are reported handled than were passed to you, the HTTP fetch
 340          * will be aborted.
 341          *
 342          * @param callable $callback
 343          * @throws MWException
 344          */
 345         public function setCallback( $callback ) {
 346                 if ( !is_callable( $callback ) ) {
 347                         throw new MWException( 'Invalid MwHttpRequest 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          */
 360         public function read( $fh, $content ) {
 361                 $this->content .= $content;
 362                 return strlen( $content );
 363         }
 364
 365         /**
 366          * Take care of whatever is necessary to perform the URI request.
 367          *
 368          * @return Status
 369          */
 370         public function execute() {
 371                 $this->content = "";
 372
 373                 if ( strtoupper( $this->method ) == "HEAD" ) {
 374                         $this->headersOnly = true;
 375                 }
 376
 377                 $this->proxySetup(); // set up any proxy as needed
 378
 379                 if ( !$this->callback ) {
 380                         $this->setCallback( [ $this, 'read' ] );
 381                 }
 382
 383                 if ( !isset( $this->reqHeaders['User-Agent'] ) ) {
 384                         $this->setUserAgent( Http::userAgent() );
 385                 }
 386         }
 387
 388         /**
 389          * Parses the headers, including the HTTP status code and any
 390          * Set-Cookie headers.  This function expects the headers to be
 391          * found in an array in the member variable headerList.
 392          */
 393         protected function parseHeader() {
 394                 $lastname = "";
 395
 396                 foreach ( $this->headerList as $header ) {
 397                         if ( preg_match( "#^HTTP/([0-9.]+) (.*)#", $header, $match ) ) {
 398                                 $this->respVersion = $match[1];
 399                                 $this->respStatus = $match[2];
 400                         } elseif ( preg_match( "#^[ \t]#", $header ) ) {
 401                                 $last = count( $this->respHeaders[$lastname] ) - 1;
 402                                 $this->respHeaders[$lastname][$last] .= "\r\n$header";
 403                         } elseif ( preg_match( "#^([^:]*):[\t ]*(.*)#", $header, $match ) ) {
 404                                 $this->respHeaders[strtolower( $match[1] )][] = $match[2];
 405                                 $lastname = strtolower( $match[1] );
 406                         }
 407                 }
 408
 409                 $this->parseCookies();
 410         }
 411
 412         /**
 413          * Sets HTTPRequest status member to a fatal value with the error
 414          * message if the returned integer value of the status code was
 415          * not successful (< 300) or a redirect (>=300 and < 400).  (see
 416          * RFC2616, section 10,
 417          * http://www.w3.org/Protocols/rfc2616/rfc2616-sec10.html for a
 418          * list of status codes.)
 419          */
 420         protected function setStatus() {
 421                 if ( !$this->respHeaders ) {
 422                         $this->parseHeader();
 423                 }
 424
 425                 if ( (int)$this->respStatus > 399 ) {
 426                         list( $code, $message ) = explode( " ", $this->respStatus, 2 );
 427                         $this->status->fatal( "http-bad-status", $code, $message );
 428                 }
 429         }
 430
 431         /**
 432          * Get the integer value of the HTTP status code (e.g. 200 for "200 Ok")
 433          * (see RFC2616, section 10, http://www.w3.org/Protocols/rfc2616/rfc2616-sec10.html
 434          * for a list of status codes.)
 435          *
 436          * @return int
 437          */
 438         public function getStatus() {
 439                 if ( !$this->respHeaders ) {
 440                         $this->parseHeader();
 441                 }
 442
 443                 return (int)$this->respStatus;
 444         }
 445
 446         /**
 447          * Returns true if the last status code was a redirect.
 448          *
 449          * @return bool
 450          */
 451         public function isRedirect() {
 452                 if ( !$this->respHeaders ) {
 453                         $this->parseHeader();
 454                 }
 455
 456                 $status = (int)$this->respStatus;
 457
 458                 if ( $status >= 300 && $status <= 303 ) {
 459                         return true;
 460                 }
 461
 462                 return false;
 463         }
 464
 465         /**
 466          * Returns an associative array of response headers after the
 467          * request has been executed.  Because some headers
 468          * (e.g. Set-Cookie) can appear more than once the, each value of
 469          * the associative array is an array of the values given.
 470          *
 471          * @return array
 472          */
 473         public function getResponseHeaders() {
 474                 if ( !$this->respHeaders ) {
 475                         $this->parseHeader();
 476                 }
 477
 478                 return $this->respHeaders;
 479         }
 480
 481         /**
 482          * Returns the value of the given response header.
 483          *
 484          * @param string $header
 485          * @return string|null
 486          */
 487         public function getResponseHeader( $header ) {
 488                 if ( !$this->respHeaders ) {
 489                         $this->parseHeader();
 490                 }
 491
 492                 if ( isset( $this->respHeaders[strtolower( $header )] ) ) {
 493                         $v = $this->respHeaders[strtolower( $header )];
 494                         return $v[count( $v ) - 1];
 495                 }
 496
 497                 return null;
 498         }
 499
 500         /**
 501          * Tells the MWHttpRequest object to use this pre-loaded CookieJar.
 502          *
 503          * @param CookieJar $jar
 504          */
 505         public function setCookieJar( $jar ) {
 506                 $this->cookieJar = $jar;
 507         }
 508
 509         /**
 510          * Returns the cookie jar in use.
 511          *
 512          * @return CookieJar
 513          */
 514         public function getCookieJar() {
 515                 if ( !$this->respHeaders ) {
 516                         $this->parseHeader();
 517                 }
 518
 519                 return $this->cookieJar;
 520         }
 521
 522         /**
 523          * Sets a cookie. Used before a request to set up any individual
 524          * cookies. Used internally after a request to parse the
 525          * Set-Cookie headers.
 526          * @see Cookie::set
 527          * @param string $name
 528          * @param mixed $value
 529          * @param array $attr
 530          */
 531         public function setCookie( $name, $value = null, $attr = null ) {
 532                 if ( !$this->cookieJar ) {
 533                         $this->cookieJar = new CookieJar;
 534                 }
 535
 536                 $this->cookieJar->setCookie( $name, $value, $attr );
 537         }
 538
 539         /**
 540          * Parse the cookies in the response headers and store them in the cookie jar.
 541          */
 542         protected function parseCookies() {
 543                 if ( !$this->cookieJar ) {
 544                         $this->cookieJar = new CookieJar;
 545                 }
 546
 547                 if ( isset( $this->respHeaders['set-cookie'] ) ) {
 548                         $url = parse_url( $this->getFinalUrl() );
 549                         foreach ( $this->respHeaders['set-cookie'] as $cookie ) {
 550                                 $this->cookieJar->parseCookieResponseHeader( $cookie, $url['host'] );
 551                         }
 552                 }
 553         }
 554
 555         /**
 556          * Returns the final URL after all redirections.
 557          *
 558          * Relative values of the "Location" header are incorrect as
 559          * stated in RFC, however they do happen and modern browsers
 560          * support them.  This function loops backwards through all
 561          * locations in order to build the proper absolute URI - Marooned
 562          * at wikia-inc.com
 563          *
 564          * Note that the multiple Location: headers are an artifact of
 565          * CURL -- they shouldn't actually get returned this way. Rewrite
 566          * this when bug 29232 is taken care of (high-level redirect
 567          * handling rewrite).
 568          *
 569          * @return string
 570          */
 571         public function getFinalUrl() {
 572                 $headers = $this->getResponseHeaders();
 573
 574                 // return full url (fix for incorrect but handled relative location)
 575                 if ( isset( $headers['location'] ) ) {
 576                         $locations = $headers['location'];
 577                         $domain = '';
 578                         $foundRelativeURI = false;
 579                         $countLocations = count( $locations );
 580
 581                         for ( $i = $countLocations - 1; $i >= 0; $i-- ) {
 582                                 $url = parse_url( $locations[$i] );
 583
 584                                 if ( isset( $url['host'] ) ) {
 585                                         $domain = $url['scheme'] . '://' . $url['host'];
 586                                         break; // found correct URI (with host)
 587                                 } else {
 588                                         $foundRelativeURI = true;
 589                                 }
 590                         }
 591
 592                         if ( $foundRelativeURI ) {
 593                                 if ( $domain ) {
 594                                         return $domain . $locations[$countLocations - 1];
 595                                 } else {
 596                                         $url = parse_url( $this->url );
 597                                         if ( isset( $url['host'] ) ) {
 598                                                 return $url['scheme'] . '://' . $url['host'] .
 599                                                         $locations[$countLocations - 1];
 600                                         }
 601                                 }
 602                         } else {
 603                                 return $locations[$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 }