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 $wgScript;
  74                                         if( $path == $wgScript ) {
  75                                                 // Script inside a rewrite path?
  76                                                 // Abort to keep from breaking...
  77                                                 return;
  78                                         }
  79                                         // Raw PATH_INFO style
  80                                         $matches = $this->extractTitle( $path, "$wgScript/$1" );
  81
  82                                         global $wgArticlePath;
  83                                         if( !$matches && $wgActionPaths) {
  84                                                 $matches = $this->extractTitle( $path, $wgArticlePath );
  85                                         }
  86
  87                                         global $wgActionPaths;
  88                                         if( !$matches && $wgActionPaths) {
  89                                                 $matches = $this->extractTitle( $path, $wgActionPaths, 'action' );
  90                                         }
  91
  92                                         global $wgVariantArticlePath, $wgContLang;
  93                                         if( !$matches && $wgVariantArticlePath ) {
  94                                                 $variantPaths = array();
  95                                                 foreach( $wgContLang->getVariants() as $variant ) {
  96                                                         $variantPaths[$variant] =
  97                                                                 str_replace( '$2', $variant, $wgVariantArticlePath );
  98                                                 }
  99                                                 $matches = $this->extractTitle( $path, $variantPaths, 'variant' );
 100                                         }
 101                                 }
 102                         } elseif ( isset( $_SERVER['ORIG_PATH_INFO'] ) && $_SERVER['ORIG_PATH_INFO'] != '' ) {
 103                                 // Mangled PATH_INFO
 104                                 // http://bugs.php.net/bug.php?id=31892
 105                                 // Also reported when ini_get('cgi.fix_pathinfo')==false
 106                                 $matches['title'] = substr( $_SERVER['ORIG_PATH_INFO'], 1 );
 107
 108                         } elseif ( isset( $_SERVER['PATH_INFO'] ) && ($_SERVER['PATH_INFO'] != '') ) {
 109                                 // Regular old PATH_INFO yay
 110                                 $matches['title'] = substr( $_SERVER['PATH_INFO'], 1 );
 111                         }
 112                         foreach( $matches as $key => $val) {
 113                                 $_GET[$key] = $_REQUEST[$key] = $val;
 114                         }
 115                 }
 116         }
 117
 118         /**
 119          * Internal URL rewriting function; tries to extract page title and,
 120          * optionally, one other fixed parameter value from a URL path.
 121          *
 122          * @param string $path the URL path given from the client
 123          * @param array $bases one or more URLs, optionally with $1 at the end
 124          * @param string $key if provided, the matching key in $bases will be
 125          *        passed on as the value of this URL parameter
 126          * @return array of URL variables to interpolate; empty if no match
 127          */
 128         private function extractTitle( $path, $bases, $key=false ) {
 129                 foreach( (array)$bases as $keyValue => $base ) {
 130                         // Find the part after $wgArticlePath
 131                         $base = str_replace( '$1', '', $base );
 132                         $baseLen = strlen( $base );
 133                         if( substr( $path, 0, $baseLen ) == $base ) {
 134                                 $raw = substr( $path, $baseLen );
 135                                 if( $raw !== '' ) {
 136                                         $matches = array( 'title' => rawurldecode( $raw ) );
 137                                         if( $key ) {
 138                                                 $matches[$key] = $keyValue;
 139                                         }
 140                                         return $matches;
 141                                 }
 142                         }
 143                 }
 144                 return array();
 145         }
 146
 147         private $_response;
 148
 149         /**
 150          * Recursively strips slashes from the given array;
 151          * used for undoing the evil that is magic_quotes_gpc.
 152          * @param array &$arr will be modified
 153          * @return array the original array
 154          * @private
 155          */
 156         function &fix_magic_quotes( &$arr ) {
 157                 foreach( $arr as $key => $val ) {
 158                         if( is_array( $val ) ) {
 159                                 $this->fix_magic_quotes( $arr[$key] );
 160                         } else {
 161                                 $arr[$key] = stripslashes( $val );
 162                         }
 163                 }
 164                 return $arr;
 165         }
 166
 167         /**
 168          * If magic_quotes_gpc option is on, run the global arrays
 169          * through fix_magic_quotes to strip out the stupid slashes.
 170          * WARNING: This should only be done once! Running a second
 171          * time could damage the values.
 172          * @private
 173          */
 174         function checkMagicQuotes() {
 175                 if ( get_magic_quotes_gpc() ) {
 176                         $this->fix_magic_quotes( $_COOKIE );
 177                         $this->fix_magic_quotes( $_ENV );
 178                         $this->fix_magic_quotes( $_GET );
 179                         $this->fix_magic_quotes( $_POST );
 180                         $this->fix_magic_quotes( $_REQUEST );
 181                         $this->fix_magic_quotes( $_SERVER );
 182                 }
 183         }
 184
 185         /**
 186          * Recursively normalizes UTF-8 strings in the given array.
 187          * @param array $data string or array
 188          * @return cleaned-up version of the given
 189          * @private
 190          */
 191         function normalizeUnicode( $data ) {
 192                 if( is_array( $data ) ) {
 193                         foreach( $data as $key => $val ) {
 194                                 $data[$key] = $this->normalizeUnicode( $val );
 195                         }
 196                 } else {
 197                         $data = UtfNormal::cleanUp( $data );
 198                 }
 199                 return $data;
 200         }
 201
 202         /**
 203          * Fetch a value from the given array or return $default if it's not set.
 204          *
 205          * @param array $arr
 206          * @param string $name
 207          * @param mixed $default
 208          * @return mixed
 209          * @private
 210          */
 211         function getGPCVal( $arr, $name, $default ) {
 212                 if( isset( $arr[$name] ) ) {
 213                         global $wgContLang;
 214                         $data = $arr[$name];
 215                         if( isset( $_GET[$name] ) && !is_array( $data ) ) {
 216                                 # Check for alternate/legacy character encoding.
 217                                 if( isset( $wgContLang ) ) {
 218                                         $data = $wgContLang->checkTitleEncoding( $data );
 219                                 }
 220                         }
 221                         $data = $this->normalizeUnicode( $data );
 222                         return $data;
 223                 } else {
 224                         return $default;
 225                 }
 226         }
 227
 228         /**
 229          * Fetch a scalar from the input or return $default if it's not set.
 230          * Returns a string. Arrays are discarded. Useful for
 231          * non-freeform text inputs (e.g. predefined internal text keys
 232          * selected by a drop-down menu). For freeform input, see getText().
 233          *
 234          * @param string $name
 235          * @param string $default optional default (or NULL)
 236          * @return string
 237          */
 238         function getVal( $name, $default = NULL ) {
 239                 $val = $this->getGPCVal( $_REQUEST, $name, $default );
 240                 if( is_array( $val ) ) {
 241                         $val = $default;
 242                 }
 243                 if( is_null( $val ) ) {
 244                         return null;
 245                 } else {
 246                         return (string)$val;
 247                 }
 248         }
 249
 250         /**
 251          * Fetch an array from the input or return $default if it's not set.
 252          * If source was scalar, will return an array with a single element.
 253          * If no source and no default, returns NULL.
 254          *
 255          * @param string $name
 256          * @param array $default optional default (or NULL)
 257          * @return array
 258          */
 259         function getArray( $name, $default = NULL ) {
 260                 $val = $this->getGPCVal( $_REQUEST, $name, $default );
 261                 if( is_null( $val ) ) {
 262                         return null;
 263                 } else {
 264                         return (array)$val;
 265                 }
 266         }
 267
 268         /**
 269          * Fetch an array of integers, or return $default if it's not set.
 270          * If source was scalar, will return an array with a single element.
 271          * If no source and no default, returns NULL.
 272          * If an array is returned, contents are guaranteed to be integers.
 273          *
 274          * @param string $name
 275          * @param array $default option default (or NULL)
 276          * @return array of ints
 277          */
 278         function getIntArray( $name, $default = NULL ) {
 279                 $val = $this->getArray( $name, $default );
 280                 if( is_array( $val ) ) {
 281                         $val = array_map( 'intval', $val );
 282                 }
 283                 return $val;
 284         }
 285
 286         /**
 287          * Fetch an integer value from the input or return $default if not set.
 288          * Guaranteed to return an integer; non-numeric input will typically
 289          * return 0.
 290          * @param string $name
 291          * @param int $default
 292          * @return int
 293          */
 294         function getInt( $name, $default = 0 ) {
 295                 return intval( $this->getVal( $name, $default ) );
 296         }
 297
 298         /**
 299          * Fetch an integer value from the input or return null if empty.
 300          * Guaranteed to return an integer or null; non-numeric input will
 301          * typically return null.
 302          * @param string $name
 303          * @return int
 304          */
 305         function getIntOrNull( $name ) {
 306                 $val = $this->getVal( $name );
 307                 return is_numeric( $val )
 308                         ? intval( $val )
 309                         : null;
 310         }
 311
 312         /**
 313          * Fetch a boolean value from the input or return $default if not set.
 314          * Guaranteed to return true or false, with normal PHP semantics for
 315          * boolean interpretation of strings.
 316          * @param string $name
 317          * @param bool $default
 318          * @return bool
 319          */
 320         function getBool( $name, $default = false ) {
 321                 return $this->getVal( $name, $default ) ? true : false;
 322         }
 323
 324         /**
 325          * Return true if the named value is set in the input, whatever that
 326          * value is (even "0"). Return false if the named value is not set.
 327          * Example use is checking for the presence of check boxes in forms.
 328          * @param string $name
 329          * @return bool
 330          */
 331         function getCheck( $name ) {
 332                 # Checkboxes and buttons are only present when clicked
 333                 # Presence connotes truth, abscense false
 334                 $val = $this->getVal( $name, NULL );
 335                 return isset( $val );
 336         }
 337
 338         /**
 339          * Fetch a text string from the given array or return $default if it's not
 340          * set. \r is stripped from the text, and with some language modules there
 341          * is an input transliteration applied. This should generally be used for
 342          * form <textarea> and <input> fields. Used for user-supplied freeform text
 343          * input (for which input transformations may be required - e.g. Esperanto
 344          * x-coding).
 345          *
 346          * @param string $name
 347          * @param string $default optional
 348          * @return string
 349          */
 350         function getText( $name, $default = '' ) {
 351                 global $wgContLang;
 352                 $val = $this->getVal( $name, $default );
 353                 return str_replace( "\r\n", "\n",
 354                         $wgContLang->recodeInput( $val ) );
 355         }
 356
 357         /**
 358          * Extracts the given named values into an array.
 359          * If no arguments are given, returns all input values.
 360          * No transformation is performed on the values.
 361          */
 362         function getValues() {
 363                 $names = func_get_args();
 364                 if ( count( $names ) == 0 ) {
 365                         $names = array_keys( $_REQUEST );
 366                 }
 367
 368                 $retVal = array();
 369                 foreach ( $names as $name ) {
 370                         $value = $this->getVal( $name );
 371                         if ( !is_null( $value ) ) {
 372                                 $retVal[$name] = $value;
 373                         }
 374                 }
 375                 return $retVal;
 376         }
 377
 378         /**
 379          * Returns true if the present request was reached by a POST operation,
 380          * false otherwise (GET, HEAD, or command-line).
 381          *
 382          * Note that values retrieved by the object may come from the
 383          * GET URL etc even on a POST request.
 384          *
 385          * @return bool
 386          */
 387         function wasPosted() {
 388                 return $_SERVER['REQUEST_METHOD'] == 'POST';
 389         }
 390
 391         /**
 392          * Returns true if there is a session cookie set.
 393          * This does not necessarily mean that the user is logged in!
 394          *
 395          * If you want to check for an open session, use session_id()
 396          * instead; that will also tell you if the session was opened
 397          * during the current request (in which case the cookie will
 398          * be sent back to the client at the end of the script run).
 399          *
 400          * @return bool
 401          */
 402         function checkSessionCookie() {
 403                 return isset( $_COOKIE[session_name()] );
 404         }
 405
 406         /**
 407          * Return the path portion of the request URI.
 408          * @return string
 409          */
 410         function getRequestURL() {
 411                 if( isset( $_SERVER['REQUEST_URI'] ) ) {
 412                         $base = $_SERVER['REQUEST_URI'];
 413                 } elseif( isset( $_SERVER['SCRIPT_NAME'] ) ) {
 414                         // Probably IIS; doesn't set REQUEST_URI
 415                         $base = $_SERVER['SCRIPT_NAME'];
 416                         if( isset( $_SERVER['QUERY_STRING'] ) && $_SERVER['QUERY_STRING'] != '' ) {
 417                                 $base .= '?' . $_SERVER['QUERY_STRING'];
 418                         }
 419                 } else {
 420                         // This shouldn't happen!
 421                         throw new MWException( "Web server doesn't provide either " .
 422                                 "REQUEST_URI or SCRIPT_NAME. Report details of your " .
 423                                 "web server configuration to http://bugzilla.wikimedia.org/" );
 424                 }
 425                 // User-agents should not send a fragment with the URI, but
 426                 // if they do, and the web server passes it on to us, we
 427                 // need to strip it or we get false-positive redirect loops
 428                 // or weird output URLs
 429                 $hash = strpos( $base, '#' );
 430                 if( $hash !== false ) {
 431                         $base = substr( $base, 0, $hash );
 432                 }
 433                 if( $base{0} == '/' ) {
 434                         return $base;
 435                 } else {
 436                         // We may get paths with a host prepended; strip it.
 437                         return preg_replace( '!^[^:]+://[^/]+/!', '/', $base );
 438                 }
 439         }
 440
 441         /**
 442          * Return the request URI with the canonical service and hostname.
 443          * @return string
 444          */
 445         function getFullRequestURL() {
 446                 global $wgServer;
 447                 return $wgServer . $this->getRequestURL();
 448         }
 449
 450         /**
 451          * Take an arbitrary query and rewrite the present URL to include it
 452          * @param $query String: query string fragment; do not include initial '?'
 453          * @return string
 454          */
 455         function appendQuery( $query ) {
 456                 global $wgTitle;
 457                 $basequery = '';
 458                 foreach( $_GET as $var => $val ) {
 459                         if ( $var == 'title' )
 460                                 continue;
 461                         if ( is_array( $val ) )
 462                                 /* This will happen given a request like
 463                                  * http://en.wikipedia.org/w/index.php?title[]=Special:Userlogin&returnto[]=Main_Page
 464                                  */
 465                                 continue;
 466                         $basequery .= '&' . urlencode( $var ) . '=' . urlencode( $val );
 467                 }
 468                 $basequery .= '&' . $query;
 469
 470                 # Trim the extra &
 471                 $basequery = substr( $basequery, 1 );
 472                 return $wgTitle->getLocalURL( $basequery );
 473         }
 474
 475         /**
 476          * HTML-safe version of appendQuery().
 477          * @param $query String: query string fragment; do not include initial '?'
 478          * @return string
 479          */
 480         function escapeAppendQuery( $query ) {
 481                 return htmlspecialchars( $this->appendQuery( $query ) );
 482         }
 483
 484         /**
 485          * Check for limit and offset parameters on the input, and return sensible
 486          * defaults if not given. The limit must be positive and is capped at 5000.
 487          * Offset must be positive but is not capped.
 488          *
 489          * @param $deflimit Integer: limit to use if no input and the user hasn't set the option.
 490          * @param $optionname String: to specify an option other than rclimit to pull from.
 491          * @return array first element is limit, second is offset
 492          */
 493         function getLimitOffset( $deflimit = 50, $optionname = 'rclimit' ) {
 494                 global $wgUser;
 495
 496                 $limit = $this->getInt( 'limit', 0 );
 497                 if( $limit < 0 ) $limit = 0;
 498                 if( ( $limit == 0 ) && ( $optionname != '' ) ) {
 499                         $limit = (int)$wgUser->getOption( $optionname );
 500                 }
 501                 if( $limit <= 0 ) $limit = $deflimit;
 502                 if( $limit > 5000 ) $limit = 5000; # We have *some* limits...
 503
 504                 $offset = $this->getInt( 'offset', 0 );
 505                 if( $offset < 0 ) $offset = 0;
 506
 507                 return array( $limit, $offset );
 508         }
 509
 510         /**
 511          * Return the path to the temporary file where PHP has stored the upload.
 512          * @param $key String:
 513          * @return string or NULL if no such file.
 514          */
 515         function getFileTempname( $key ) {
 516                 if( !isset( $_FILES[$key] ) ) {
 517                         return NULL;
 518                 }
 519                 return $_FILES[$key]['tmp_name'];
 520         }
 521
 522         /**
 523          * Return the size of the upload, or 0.
 524          * @param $key String:
 525          * @return integer
 526          */
 527         function getFileSize( $key ) {
 528                 if( !isset( $_FILES[$key] ) ) {
 529                         return 0;
 530                 }
 531                 return $_FILES[$key]['size'];
 532         }
 533
 534         /**
 535          * Return the upload error or 0
 536          * @param $key String:
 537          * @return integer
 538          */
 539         function getUploadError( $key ) {
 540                 if( !isset( $_FILES[$key] ) || !isset( $_FILES[$key]['error'] ) ) {
 541                         return 0/*UPLOAD_ERR_OK*/;
 542                 }
 543                 return $_FILES[$key]['error'];
 544         }
 545
 546         /**
 547          * Return the original filename of the uploaded file, as reported by
 548          * the submitting user agent. HTML-style character entities are
 549          * interpreted and normalized to Unicode normalization form C, in part
 550          * to deal with weird input from Safari with non-ASCII filenames.
 551          *
 552          * Other than this the name is not verified for being a safe filename.
 553          *
 554          * @param $key String:
 555          * @return string or NULL if no such file.
 556          */
 557         function getFileName( $key ) {
 558                 if( !isset( $_FILES[$key] ) ) {
 559                         return NULL;
 560                 }
 561                 $name = $_FILES[$key]['name'];
 562
 563                 # Safari sends filenames in HTML-encoded Unicode form D...
 564                 # Horrid and evil! Let's try to make some kind of sense of it.
 565                 $name = Sanitizer::decodeCharReferences( $name );
 566                 $name = UtfNormal::cleanUp( $name );
 567                 wfDebug( "WebRequest::getFileName() '" . $_FILES[$key]['name'] . "' normalized to '$name'\n" );
 568                 return $name;
 569         }
 570
 571         /**
 572          * Return a handle to WebResponse style object, for setting cookies,
 573          * headers and other stuff, for Request being worked on.
 574          */
 575         function response() {
 576                 /* Lazy initialization of response object for this request */
 577                 if (!is_object($this->_response)) {
 578                         $this->_response = new WebResponse;
 579                 }
 580                 return $this->_response;
 581         }
 582
 583 }
 584
 585 /**
 586  * WebRequest clone which takes values from a provided array.
 587  *
 588  */
 589 class FauxRequest extends WebRequest {
 590         var $data = null;
 591         var $wasPosted = false;
 592
 593         /**
 594          * @param array $data Array of *non*-urlencoded key => value pairs, the
 595          *   fake GET/POST values
 596          * @param bool $wasPosted Whether to treat the data as POST
 597          */
 598         function FauxRequest( $data, $wasPosted = false ) {
 599                 if( is_array( $data ) ) {
 600                         $this->data = $data;
 601                 } else {
 602                         throw new MWException( "FauxRequest() got bogus data" );
 603                 }
 604                 $this->wasPosted = $wasPosted;
 605         }
 606
 607         function getVal( $name, $default = NULL ) {
 608                 return $this->getGPCVal( $this->data, $name, $default );
 609         }
 610
 611         function getText( $name, $default = '' ) {
 612                 # Override; don't recode since we're using internal data
 613                 return $this->getVal( $name, $default );
 614         }
 615
 616         function getValues() {
 617                 return $this->data;
 618         }
 619
 620         function wasPosted() {
 621                 return $this->wasPosted;
 622         }
 623
 624         function checkSessionCookie() {
 625                 return false;
 626         }
 627
 628         function getRequestURL() {
 629                 throw new MWException( 'FauxRequest::getRequestURL() not implemented' );
 630         }
 631
 632         function appendQuery( $query ) {
 633                 throw new MWException( 'FauxRequest::appendQuery() not implemented' );
 634         }
 635
 636 }
 637
 638