includes/WebRequest.php

   1 <?php
   2 /**
   3  * Deal with importing all those nasssty globals and things
   4  */
   5
   6 # Copyright (C) 2003 Brion Vibber <brion@pobox.com>
   7 # http://www.mediawiki.org/
   8 #
   9 # This program is free software; you can redistribute it and/or modify
  10 # it under the terms of the GNU General Public License as published by
  11 # the Free Software Foundation; either version 2 of the License, or
  12 # (at your option) any later version.
  13 #
  14 # This program is distributed in the hope that it will be useful,
  15 # but WITHOUT ANY WARRANTY; without even the implied warranty of
  16 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  17 # GNU General Public License for more details.
  18 #
  19 # You should have received a copy of the GNU General Public License along
  20 # with this program; if not, write to the Free Software Foundation, Inc.,
  21 # 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  22 # http://www.gnu.org/copyleft/gpl.html
  23
  24
  25 /**
  26  * Some entry points may use this file without first enabling the
  27  * autoloader.
  28  */
  29 if ( !function_exists( '__autoload' ) ) {
  30         require_once( dirname(__FILE__) . '/normal/UtfNormal.php' );
  31 }
  32
  33 /**
  34  * The WebRequest class encapsulates getting at data passed in the
  35  * URL or via a POSTed form, handling remove of "magic quotes" slashes,
  36  * stripping illegal input characters and normalizing Unicode sequences.
  37  *
  38  * Usually this is used via a global singleton, $wgRequest. You should
  39  * not create a second WebRequest object; make a FauxRequest object if
  40  * you want to pass arbitrary data to some function in place of the web
  41  * input.
  42  *
  43  */
  44 class WebRequest {
  45         function __construct() {
  46                 $this->checkMagicQuotes();
  47         }
  48
  49         /**
  50          * Check for title, action, and/or variant data in the URL
  51          * and interpolate it into the GET variables.
  52          * This should only be run after $wgContLang is available,
  53          * as we may need the list of language variants to determine
  54          * available variant URLs.
  55          */
  56         function interpolateTitle() {
  57                 global $wgUsePathInfo;
  58                 if ( $wgUsePathInfo ) {
  59                         // PATH_INFO is mangled due to http://bugs.php.net/bug.php?id=31892
  60                         // And also by Apache 2.x, double slashes are converted to single slashes.
  61                         // So we will use REQUEST_URI if possible.
  62                         $matches = array();
  63                         if ( !empty( $_SERVER['REQUEST_URI'] ) ) {
  64                                 // Slurp out the path portion to examine...
  65                                 $url = $_SERVER['REQUEST_URI'];
  66                                 if ( !preg_match( '!^https?://!', $url ) ) {
  67                                         $url = 'http://unused' . $url;
  68                                 }
  69                                 $a = parse_url( $url );
  70                                 if( $a ) {
  71                                         $path = $a['path'];
  72
  73                                         global $wgArticlePath;
  74                                         $matches = $this->extractTitle( $path, $wgArticlePath );
  75
  76                                         global $wgActionPaths;
  77                                         if( !$matches && $wgActionPaths) {
  78                                                 $matches = $this->extractTitle( $path, $wgActionPaths, 'action' );
  79                                         }
  80
  81                                         global $wgVariantArticlePath, $wgContLang;
  82                                         if( !$matches && $wgVariantArticlePath ) {
  83                                                 $variantPaths = array();
  84                                                 foreach( $wgContLang->getVariants() as $variant ) {
  85                                                         $variantPaths[$variant] =
  86                                                                 str_replace( '$2', $variant, $wgVariantArticlePath );
  87                                                 }
  88                                                 $matches = $this->extractTitle( $path, $variantPaths, 'variant' );
  89                                         }
  90                                 }
  91                         } elseif ( isset( $_SERVER['ORIG_PATH_INFO'] ) && $_SERVER['ORIG_PATH_INFO'] != '' ) {
  92                                 // Mangled PATH_INFO
  93                                 // http://bugs.php.net/bug.php?id=31892
  94                                 // Also reported when ini_get('cgi.fix_pathinfo')==false
  95                                 $matches['title'] = substr( $_SERVER['ORIG_PATH_INFO'], 1 );
  96
  97                         } elseif ( isset( $_SERVER['PATH_INFO'] ) && ($_SERVER['PATH_INFO'] != '') ) {
  98                                 // Regular old PATH_INFO yay
  99                                 $matches['title'] = substr( $_SERVER['PATH_INFO'], 1 );
 100                         }
 101                         foreach( $matches as $key => $val) {
 102                                 $_GET[$key] = $_REQUEST[$key] = $val;
 103                         }
 104                 }
 105         }
 106
 107         /**
 108          * Internal URL rewriting function; tries to extract page title and,
 109          * optionally, one other fixed parameter value from a URL path.
 110          *
 111          * @param string $path the URL path given from the client
 112          * @param array $bases one or more URLs, optionally with $1 at the end
 113          * @param string $key if provided, the matching key in $bases will be
 114          *        passed on as the value of this URL parameter
 115          * @return array of URL variables to interpolate; empty if no match
 116          */
 117         private function extractTitle( $path, $bases, $key=false ) {
 118                 foreach( (array)$bases as $keyValue => $base ) {
 119                         // Find the part after $wgArticlePath
 120                         $base = str_replace( '$1', '', $base );
 121                         $baseLen = strlen( $base );
 122                         if( substr( $path, 0, $baseLen ) == $base ) {
 123                                 $raw = substr( $path, $baseLen );
 124                                 if( $raw !== '' ) {
 125                                         $matches = array( 'title' => rawurldecode( $raw ) );
 126                                         if( $key ) {
 127                                                 $matches[$key] = $keyValue;
 128                                         }
 129                                         return $matches;
 130                                 }
 131                         }
 132                 }
 133                 return array();
 134         }
 135
 136         private $_response;
 137
 138         /**
 139          * Recursively strips slashes from the given array;
 140          * used for undoing the evil that is magic_quotes_gpc.
 141          * @param array &$arr will be modified
 142          * @return array the original array
 143          * @private
 144          */
 145         function &fix_magic_quotes( &$arr ) {
 146                 foreach( $arr as $key => $val ) {
 147                         if( is_array( $val ) ) {
 148                                 $this->fix_magic_quotes( $arr[$key] );
 149                         } else {
 150                                 $arr[$key] = stripslashes( $val );
 151                         }
 152                 }
 153                 return $arr;
 154         }
 155
 156         /**
 157          * If magic_quotes_gpc option is on, run the global arrays
 158          * through fix_magic_quotes to strip out the stupid slashes.
 159          * WARNING: This should only be done once! Running a second
 160          * time could damage the values.
 161          * @private
 162          */
 163         function checkMagicQuotes() {
 164                 if ( get_magic_quotes_gpc() ) {
 165                         $this->fix_magic_quotes( $_COOKIE );
 166                         $this->fix_magic_quotes( $_ENV );
 167                         $this->fix_magic_quotes( $_GET );
 168                         $this->fix_magic_quotes( $_POST );
 169                         $this->fix_magic_quotes( $_REQUEST );
 170                         $this->fix_magic_quotes( $_SERVER );
 171                 }
 172         }
 173
 174         /**
 175          * Recursively normalizes UTF-8 strings in the given array.
 176          * @param array $data string or array
 177          * @return cleaned-up version of the given
 178          * @private
 179          */
 180         function normalizeUnicode( $data ) {
 181                 if( is_array( $data ) ) {
 182                         foreach( $data as $key => $val ) {
 183                                 $data[$key] = $this->normalizeUnicode( $val );
 184                         }
 185                 } else {
 186                         $data = UtfNormal::cleanUp( $data );
 187                 }
 188                 return $data;
 189         }
 190
 191         /**
 192          * Fetch a value from the given array or return $default if it's not set.
 193          *
 194          * @param array $arr
 195          * @param string $name
 196          * @param mixed $default
 197          * @return mixed
 198          * @private
 199          */
 200         function getGPCVal( $arr, $name, $default ) {
 201                 if( isset( $arr[$name] ) ) {
 202                         global $wgContLang;
 203                         $data = $arr[$name];
 204                         if( isset( $_GET[$name] ) && !is_array( $data ) ) {
 205                                 # Check for alternate/legacy character encoding.
 206                                 if( isset( $wgContLang ) ) {
 207                                         $data = $wgContLang->checkTitleEncoding( $data );
 208                                 }
 209                         }
 210                         $data = $this->normalizeUnicode( $data );
 211                         return $data;
 212                 } else {
 213                         return $default;
 214                 }
 215         }
 216
 217         /**
 218          * Fetch a scalar from the input or return $default if it's not set.
 219          * Returns a string. Arrays are discarded. Useful for
 220          * non-freeform text inputs (e.g. predefined internal text keys
 221          * selected by a drop-down menu). For freeform input, see getText().
 222          *
 223          * @param string $name
 224          * @param string $default optional default (or NULL)
 225          * @return string
 226          */
 227         function getVal( $name, $default = NULL ) {
 228                 $val = $this->getGPCVal( $_REQUEST, $name, $default );
 229                 if( is_array( $val ) ) {
 230                         $val = $default;
 231                 }
 232                 if( is_null( $val ) ) {
 233                         return null;
 234                 } else {
 235                         return (string)$val;
 236                 }
 237         }
 238
 239         /**
 240          * Fetch an array from the input or return $default if it's not set.
 241          * If source was scalar, will return an array with a single element.
 242          * If no source and no default, returns NULL.
 243          *
 244          * @param string $name
 245          * @param array $default optional default (or NULL)
 246          * @return array
 247          */
 248         function getArray( $name, $default = NULL ) {
 249                 $val = $this->getGPCVal( $_REQUEST, $name, $default );
 250                 if( is_null( $val ) ) {
 251                         return null;
 252                 } else {
 253                         return (array)$val;
 254                 }
 255         }
 256
 257         /**
 258          * Fetch an array of integers, or return $default if it's not set.
 259          * If source was scalar, will return an array with a single element.
 260          * If no source and no default, returns NULL.
 261          * If an array is returned, contents are guaranteed to be integers.
 262          *
 263          * @param string $name
 264          * @param array $default option default (or NULL)
 265          * @return array of ints
 266          */
 267         function getIntArray( $name, $default = NULL ) {
 268                 $val = $this->getArray( $name, $default );
 269                 if( is_array( $val ) ) {
 270                         $val = array_map( 'intval', $val );
 271                 }
 272                 return $val;
 273         }
 274
 275         /**
 276          * Fetch an integer value from the input or return $default if not set.
 277          * Guaranteed to return an integer; non-numeric input will typically
 278          * return 0.
 279          * @param string $name
 280          * @param int $default
 281          * @return int
 282          */
 283         function getInt( $name, $default = 0 ) {
 284                 return intval( $this->getVal( $name, $default ) );
 285         }
 286
 287         /**
 288          * Fetch an integer value from the input or return null if empty.
 289          * Guaranteed to return an integer or null; non-numeric input will
 290          * typically return null.
 291          * @param string $name
 292          * @return int
 293          */
 294         function getIntOrNull( $name ) {
 295                 $val = $this->getVal( $name );
 296                 return is_numeric( $val )
 297                         ? intval( $val )
 298                         : null;
 299         }
 300
 301         /**
 302          * Fetch a boolean value from the input or return $default if not set.
 303          * Guaranteed to return true or false, with normal PHP semantics for
 304          * boolean interpretation of strings.
 305          * @param string $name
 306          * @param bool $default
 307          * @return bool
 308          */
 309         function getBool( $name, $default = false ) {
 310                 return $this->getVal( $name, $default ) ? true : false;
 311         }
 312
 313         /**
 314          * Return true if the named value is set in the input, whatever that
 315          * value is (even "0"). Return false if the named value is not set.
 316          * Example use is checking for the presence of check boxes in forms.
 317          * @param string $name
 318          * @return bool
 319          */
 320         function getCheck( $name ) {
 321                 # Checkboxes and buttons are only present when clicked
 322                 # Presence connotes truth, abscense false
 323                 $val = $this->getVal( $name, NULL );
 324                 return isset( $val );
 325         }
 326
 327         /**
 328          * Fetch a text string from the given array or return $default if it's not
 329          * set. \r is stripped from the text, and with some language modules there
 330          * is an input transliteration applied. This should generally be used for
 331          * form <textarea> and <input> fields. Used for user-supplied freeform text
 332          * input (for which input transformations may be required - e.g. Esperanto
 333          * x-coding).
 334          *
 335          * @param string $name
 336          * @param string $default optional
 337          * @return string
 338          */
 339         function getText( $name, $default = '' ) {
 340                 global $wgContLang;
 341                 $val = $this->getVal( $name, $default );
 342                 return str_replace( "\r\n", "\n",
 343                         $wgContLang->recodeInput( $val ) );
 344         }
 345
 346         /**
 347          * Extracts the given named values into an array.
 348          * If no arguments are given, returns all input values.
 349          * No transformation is performed on the values.
 350          */
 351         function getValues() {
 352                 $names = func_get_args();
 353                 if ( count( $names ) == 0 ) {
 354                         $names = array_keys( $_REQUEST );
 355                 }
 356
 357                 $retVal = array();
 358                 foreach ( $names as $name ) {
 359                         $value = $this->getVal( $name );
 360                         if ( !is_null( $value ) ) {
 361                                 $retVal[$name] = $value;
 362                         }
 363                 }
 364                 return $retVal;
 365         }
 366
 367         /**
 368          * Returns true if the present request was reached by a POST operation,
 369          * false otherwise (GET, HEAD, or command-line).
 370          *
 371          * Note that values retrieved by the object may come from the
 372          * GET URL etc even on a POST request.
 373          *
 374          * @return bool
 375          */
 376         function wasPosted() {
 377                 return $_SERVER['REQUEST_METHOD'] == 'POST';
 378         }
 379
 380         /**
 381          * Returns true if there is a session cookie set.
 382          * This does not necessarily mean that the user is logged in!
 383          *
 384          * If you want to check for an open session, use session_id()
 385          * instead; that will also tell you if the session was opened
 386          * during the current request (in which case the cookie will
 387          * be sent back to the client at the end of the script run).
 388          *
 389          * @return bool
 390          */
 391         function checkSessionCookie() {
 392                 return isset( $_COOKIE[session_name()] );
 393         }
 394
 395         /**
 396          * Return the path portion of the request URI.
 397          * @return string
 398          */
 399         function getRequestURL() {
 400                 if( isset( $_SERVER['REQUEST_URI'] ) ) {
 401                         $base = $_SERVER['REQUEST_URI'];
 402                 } elseif( isset( $_SERVER['SCRIPT_NAME'] ) ) {
 403                         // Probably IIS; doesn't set REQUEST_URI
 404                         $base = $_SERVER['SCRIPT_NAME'];
 405                         if( isset( $_SERVER['QUERY_STRING'] ) && $_SERVER['QUERY_STRING'] != '' ) {
 406                                 $base .= '?' . $_SERVER['QUERY_STRING'];
 407                         }
 408                 } else {
 409                         // This shouldn't happen!
 410                         throw new MWException( "Web server doesn't provide either " .
 411                                 "REQUEST_URI or SCRIPT_NAME. Report details of your " .
 412                                 "web server configuration to http://bugzilla.wikimedia.org/" );
 413                 }
 414                 // User-agents should not send a fragment with the URI, but
 415                 // if they do, and the web server passes it on to us, we
 416                 // need to strip it or we get false-positive redirect loops
 417                 // or weird output URLs
 418                 $hash = strpos( $base, '#' );
 419                 if( $hash !== false ) {
 420                         $base = substr( $base, 0, $hash );
 421                 }
 422                 if( $base{0} == '/' ) {
 423                         return $base;
 424                 } else {
 425                         // We may get paths with a host prepended; strip it.
 426                         return preg_replace( '!^[^:]+://[^/]+/!', '/', $base );
 427                 }
 428         }
 429
 430         /**
 431          * Return the request URI with the canonical service and hostname.
 432          * @return string
 433          */
 434         function getFullRequestURL() {
 435                 global $wgServer;
 436                 return $wgServer . $this->getRequestURL();
 437         }
 438
 439         /**
 440          * Take an arbitrary query and rewrite the present URL to include it
 441          * @param $query String: query string fragment; do not include initial '?'
 442          * @return string
 443          */
 444         function appendQuery( $query ) {
 445                 global $wgTitle;
 446                 $basequery = '';
 447                 foreach( $_GET as $var => $val ) {
 448                         if ( $var == 'title' )
 449                                 continue;
 450                         if ( is_array( $val ) )
 451                                 /* This will happen given a request like
 452                                  * http://en.wikipedia.org/w/index.php?title[]=Special:Userlogin&returnto[]=Main_Page
 453                                  */
 454                                 continue;
 455                         $basequery .= '&' . urlencode( $var ) . '=' . urlencode( $val );
 456                 }
 457                 $basequery .= '&' . $query;
 458
 459                 # Trim the extra &
 460                 $basequery = substr( $basequery, 1 );
 461                 return $wgTitle->getLocalURL( $basequery );
 462         }
 463
 464         /**
 465          * HTML-safe version of appendQuery().
 466          * @param $query String: query string fragment; do not include initial '?'
 467          * @return string
 468          */
 469         function escapeAppendQuery( $query ) {
 470                 return htmlspecialchars( $this->appendQuery( $query ) );
 471         }
 472
 473         /**
 474          * Check for limit and offset parameters on the input, and return sensible
 475          * defaults if not given. The limit must be positive and is capped at 5000.
 476          * Offset must be positive but is not capped.
 477          *
 478          * @param $deflimit Integer: limit to use if no input and the user hasn't set the option.
 479          * @param $optionname String: to specify an option other than rclimit to pull from.
 480          * @return array first element is limit, second is offset
 481          */
 482         function getLimitOffset( $deflimit = 50, $optionname = 'rclimit' ) {
 483                 global $wgUser;
 484
 485                 $limit = $this->getInt( 'limit', 0 );
 486                 if( $limit < 0 ) $limit = 0;
 487                 if( ( $limit == 0 ) && ( $optionname != '' ) ) {
 488                         $limit = (int)$wgUser->getOption( $optionname );
 489                 }
 490                 if( $limit <= 0 ) $limit = $deflimit;
 491                 if( $limit > 5000 ) $limit = 5000; # We have *some* limits...
 492
 493                 $offset = $this->getInt( 'offset', 0 );
 494                 if( $offset < 0 ) $offset = 0;
 495
 496                 return array( $limit, $offset );
 497         }
 498
 499         /**
 500          * Return the path to the temporary file where PHP has stored the upload.
 501          * @param $key String:
 502          * @return string or NULL if no such file.
 503          */
 504         function getFileTempname( $key ) {
 505                 if( !isset( $_FILES[$key] ) ) {
 506                         return NULL;
 507                 }
 508                 return $_FILES[$key]['tmp_name'];
 509         }
 510
 511         /**
 512          * Return the size of the upload, or 0.
 513          * @param $key String:
 514          * @return integer
 515          */
 516         function getFileSize( $key ) {
 517                 if( !isset( $_FILES[$key] ) ) {
 518                         return 0;
 519                 }
 520                 return $_FILES[$key]['size'];
 521         }
 522
 523         /**
 524          * Return the upload error or 0
 525          * @param $key String:
 526          * @return integer
 527          */
 528         function getUploadError( $key ) {
 529                 if( !isset( $_FILES[$key] ) || !isset( $_FILES[$key]['error'] ) ) {
 530                         return 0/*UPLOAD_ERR_OK*/;
 531                 }
 532                 return $_FILES[$key]['error'];
 533         }
 534
 535         /**
 536          * Return the original filename of the uploaded file, as reported by
 537          * the submitting user agent. HTML-style character entities are
 538          * interpreted and normalized to Unicode normalization form C, in part
 539          * to deal with weird input from Safari with non-ASCII filenames.
 540          *
 541          * Other than this the name is not verified for being a safe filename.
 542          *
 543          * @param $key String:
 544          * @return string or NULL if no such file.
 545          */
 546         function getFileName( $key ) {
 547                 if( !isset( $_FILES[$key] ) ) {
 548                         return NULL;
 549                 }
 550                 $name = $_FILES[$key]['name'];
 551
 552                 # Safari sends filenames in HTML-encoded Unicode form D...
 553                 # Horrid and evil! Let's try to make some kind of sense of it.
 554                 $name = Sanitizer::decodeCharReferences( $name );
 555                 $name = UtfNormal::cleanUp( $name );
 556                 wfDebug( "WebRequest::getFileName() '" . $_FILES[$key]['name'] . "' normalized to '$name'\n" );
 557                 return $name;
 558         }
 559
 560         /**
 561          * Return a handle to WebResponse style object, for setting cookies,
 562          * headers and other stuff, for Request being worked on.
 563          */
 564         function response() {
 565                 /* Lazy initialization of response object for this request */
 566                 if (!is_object($this->_response)) {
 567                         $this->_response = new WebResponse;
 568                 }
 569                 return $this->_response;
 570         }
 571
 572 }
 573
 574 /**
 575  * WebRequest clone which takes values from a provided array.
 576  *
 577  */
 578 class FauxRequest extends WebRequest {
 579         var $data = null;
 580         var $wasPosted = false;
 581
 582         function FauxRequest( $data, $wasPosted = false ) {
 583                 if( is_array( $data ) ) {
 584                         $this->data = $data;
 585                 } else {
 586                         throw new MWException( "FauxRequest() got bogus data" );
 587                 }
 588                 $this->wasPosted = $wasPosted;
 589         }
 590
 591         function getVal( $name, $default = NULL ) {
 592                 return $this->getGPCVal( $this->data, $name, $default );
 593         }
 594
 595         function getText( $name, $default = '' ) {
 596                 # Override; don't recode since we're using internal data
 597                 return $this->getVal( $name, $default );
 598         }
 599
 600         function getValues() {
 601                 return $this->data;
 602         }
 603
 604         function wasPosted() {
 605                 return $this->wasPosted;
 606         }
 607
 608         function checkSessionCookie() {
 609                 return false;
 610         }
 611
 612         function getRequestURL() {
 613                 throw new MWException( 'FauxRequest::getRequestURL() not implemented' );
 614         }
 615
 616         function appendQuery( $query ) {
 617                 throw new MWException( 'FauxRequest::appendQuery() not implemented' );
 618         }
 619
 620 }
 621
 622