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         var $data = array();
  46
  47         function __construct() {
  48                 /// @fixme This preemptive de-quoting can interfere with other web libraries
  49                 ///        and increases our memory footprint. It would be cleaner to do on
  50                 ///        demand; but currently we have no wrapper for $_SERVER etc.
  51                 $this->checkMagicQuotes();
  52
  53                 // POST overrides GET data
  54                 // We don't use $_REQUEST here to avoid interference from cookies...
  55                 $this->data = array_merge( $_GET, $_POST );
  56         }
  57
  58         /**
  59          * Check for title, action, and/or variant data in the URL
  60          * and interpolate it into the GET variables.
  61          * This should only be run after $wgContLang is available,
  62          * as we may need the list of language variants to determine
  63          * available variant URLs.
  64          */
  65         function interpolateTitle() {
  66                 global $wgUsePathInfo;
  67                 if ( $wgUsePathInfo ) {
  68                         // PATH_INFO is mangled due to http://bugs.php.net/bug.php?id=31892
  69                         // And also by Apache 2.x, double slashes are converted to single slashes.
  70                         // So we will use REQUEST_URI if possible.
  71                         $matches = array();
  72                         if ( !empty( $_SERVER['REQUEST_URI'] ) ) {
  73                                 // Slurp out the path portion to examine...
  74                                 $url = $_SERVER['REQUEST_URI'];
  75                                 if ( !preg_match( '!^https?://!', $url ) ) {
  76                                         $url = 'http://unused' . $url;
  77                                 }
  78                                 $a = parse_url( $url );
  79                                 if( $a ) {
  80                                         $path = $a['path'];
  81
  82                                         global $wgScript;
  83                                         if( $path == $wgScript ) {
  84                                                 // Script inside a rewrite path?
  85                                                 // Abort to keep from breaking...
  86                                                 return;
  87                                         }
  88                                         // Raw PATH_INFO style
  89                                         $matches = $this->extractTitle( $path, "$wgScript/$1" );
  90
  91                                         global $wgArticlePath;
  92                                         if( !$matches && $wgArticlePath ) {
  93                                                 $matches = $this->extractTitle( $path, $wgArticlePath );
  94                                         }
  95
  96                                         global $wgActionPaths;
  97                                         if( !$matches && $wgActionPaths ) {
  98                                                 $matches = $this->extractTitle( $path, $wgActionPaths, 'action' );
  99                                         }
 100
 101                                         global $wgVariantArticlePath, $wgContLang;
 102                                         if( !$matches && $wgVariantArticlePath ) {
 103                                                 $variantPaths = array();
 104                                                 foreach( $wgContLang->getVariants() as $variant ) {
 105                                                         $variantPaths[$variant] =
 106                                                                 str_replace( '$2', $variant, $wgVariantArticlePath );
 107                                                 }
 108                                                 $matches = $this->extractTitle( $path, $variantPaths, 'variant' );
 109                                         }
 110                                 }
 111                         } elseif ( isset( $_SERVER['ORIG_PATH_INFO'] ) && $_SERVER['ORIG_PATH_INFO'] != '' ) {
 112                                 // Mangled PATH_INFO
 113                                 // http://bugs.php.net/bug.php?id=31892
 114                                 // Also reported when ini_get('cgi.fix_pathinfo')==false
 115                                 $matches['title'] = substr( $_SERVER['ORIG_PATH_INFO'], 1 );
 116
 117                         } elseif ( isset( $_SERVER['PATH_INFO'] ) && ($_SERVER['PATH_INFO'] != '') ) {
 118                                 // Regular old PATH_INFO yay
 119                                 $matches['title'] = substr( $_SERVER['PATH_INFO'], 1 );
 120                         }
 121                         foreach( $matches as $key => $val) {
 122                                 $this->data[$key] = $_GET[$key] = $_REQUEST[$key] = $val;
 123                         }
 124                 }
 125         }
 126
 127         /**
 128          * Internal URL rewriting function; tries to extract page title and,
 129          * optionally, one other fixed parameter value from a URL path.
 130          *
 131          * @param string $path the URL path given from the client
 132          * @param array $bases one or more URLs, optionally with $1 at the end
 133          * @param string $key if provided, the matching key in $bases will be
 134          *        passed on as the value of this URL parameter
 135          * @return array of URL variables to interpolate; empty if no match
 136          */
 137         private function extractTitle( $path, $bases, $key=false ) {
 138                 foreach( (array)$bases as $keyValue => $base ) {
 139                         // Find the part after $wgArticlePath
 140                         $base = str_replace( '$1', '', $base );
 141                         $baseLen = strlen( $base );
 142                         if( substr( $path, 0, $baseLen ) == $base ) {
 143                                 $raw = substr( $path, $baseLen );
 144                                 if( $raw !== '' ) {
 145                                         $matches = array( 'title' => rawurldecode( $raw ) );
 146                                         if( $key ) {
 147                                                 $matches[$key] = $keyValue;
 148                                         }
 149                                         return $matches;
 150                                 }
 151                         }
 152                 }
 153                 return array();
 154         }
 155
 156         private $_response;
 157
 158         /**
 159          * Recursively strips slashes from the given array;
 160          * used for undoing the evil that is magic_quotes_gpc.
 161          * @param array &$arr will be modified
 162          * @return array the original array
 163          * @private
 164          */
 165         function &fix_magic_quotes( &$arr ) {
 166                 foreach( $arr as $key => $val ) {
 167                         if( is_array( $val ) ) {
 168                                 $this->fix_magic_quotes( $arr[$key] );
 169                         } else {
 170                                 $arr[$key] = stripslashes( $val );
 171                         }
 172                 }
 173                 return $arr;
 174         }
 175
 176         /**
 177          * If magic_quotes_gpc option is on, run the global arrays
 178          * through fix_magic_quotes to strip out the stupid slashes.
 179          * WARNING: This should only be done once! Running a second
 180          * time could damage the values.
 181          * @private
 182          */
 183         function checkMagicQuotes() {
 184                 if ( get_magic_quotes_gpc() ) {
 185                         $this->fix_magic_quotes( $_COOKIE );
 186                         $this->fix_magic_quotes( $_ENV );
 187                         $this->fix_magic_quotes( $_GET );
 188                         $this->fix_magic_quotes( $_POST );
 189                         $this->fix_magic_quotes( $_REQUEST );
 190                         $this->fix_magic_quotes( $_SERVER );
 191                 }
 192         }
 193
 194         /**
 195          * Recursively normalizes UTF-8 strings in the given array.
 196          * @param array $data string or array
 197          * @return cleaned-up version of the given
 198          * @private
 199          */
 200         function normalizeUnicode( $data ) {
 201                 if( is_array( $data ) ) {
 202                         foreach( $data as $key => $val ) {
 203                                 $data[$key] = $this->normalizeUnicode( $val );
 204                         }
 205                 } else {
 206                         $data = UtfNormal::cleanUp( $data );
 207                 }
 208                 return $data;
 209         }
 210
 211         /**
 212          * Fetch a value from the given array or return $default if it's not set.
 213          *
 214          * @param array $arr
 215          * @param string $name
 216          * @param mixed $default
 217          * @return mixed
 218          * @private
 219          */
 220         function getGPCVal( $arr, $name, $default ) {
 221                 if( isset( $arr[$name] ) ) {
 222                         global $wgContLang;
 223                         $data = $arr[$name];
 224                         if( isset( $_GET[$name] ) && !is_array( $data ) ) {
 225                                 # Check for alternate/legacy character encoding.
 226                                 if( isset( $wgContLang ) ) {
 227                                         $data = $wgContLang->checkTitleEncoding( $data );
 228                                 }
 229                         }
 230                         $data = $this->normalizeUnicode( $data );
 231                         return $data;
 232                 } else {
 233                         return $default;
 234                 }
 235         }
 236
 237         /**
 238          * Fetch a scalar from the input or return $default if it's not set.
 239          * Returns a string. Arrays are discarded. Useful for
 240          * non-freeform text inputs (e.g. predefined internal text keys
 241          * selected by a drop-down menu). For freeform input, see getText().
 242          *
 243          * @param string $name
 244          * @param string $default optional default (or NULL)
 245          * @return string
 246          */
 247         function getVal( $name, $default = NULL ) {
 248                 $val = $this->getGPCVal( $this->data, $name, $default );
 249                 if( is_array( $val ) ) {
 250                         $val = $default;
 251                 }
 252                 if( is_null( $val ) ) {
 253                         return null;
 254                 } else {
 255                         return (string)$val;
 256                 }
 257         }
 258
 259         /**
 260          * Fetch an array from the input or return $default if it's not set.
 261          * If source was scalar, will return an array with a single element.
 262          * If no source and no default, returns NULL.
 263          *
 264          * @param string $name
 265          * @param array $default optional default (or NULL)
 266          * @return array
 267          */
 268         function getArray( $name, $default = NULL ) {
 269                 $val = $this->getGPCVal( $this->data, $name, $default );
 270                 if( is_null( $val ) ) {
 271                         return null;
 272                 } else {
 273                         return (array)$val;
 274                 }
 275         }
 276
 277         /**
 278          * Fetch an array of integers, or return $default if it's not set.
 279          * If source was scalar, will return an array with a single element.
 280          * If no source and no default, returns NULL.
 281          * If an array is returned, contents are guaranteed to be integers.
 282          *
 283          * @param string $name
 284          * @param array $default option default (or NULL)
 285          * @return array of ints
 286          */
 287         function getIntArray( $name, $default = NULL ) {
 288                 $val = $this->getArray( $name, $default );
 289                 if( is_array( $val ) ) {
 290                         $val = array_map( 'intval', $val );
 291                 }
 292                 return $val;
 293         }
 294
 295         /**
 296          * Fetch an integer value from the input or return $default if not set.
 297          * Guaranteed to return an integer; non-numeric input will typically
 298          * return 0.
 299          * @param string $name
 300          * @param int $default
 301          * @return int
 302          */
 303         function getInt( $name, $default = 0 ) {
 304                 return intval( $this->getVal( $name, $default ) );
 305         }
 306
 307         /**
 308          * Fetch an integer value from the input or return null if empty.
 309          * Guaranteed to return an integer or null; non-numeric input will
 310          * typically return null.
 311          * @param string $name
 312          * @return int
 313          */
 314         function getIntOrNull( $name ) {
 315                 $val = $this->getVal( $name );
 316                 return is_numeric( $val )
 317                         ? intval( $val )
 318                         : null;
 319         }
 320
 321         /**
 322          * Fetch a boolean value from the input or return $default if not set.
 323          * Guaranteed to return true or false, with normal PHP semantics for
 324          * boolean interpretation of strings.
 325          * @param string $name
 326          * @param bool $default
 327          * @return bool
 328          */
 329         function getBool( $name, $default = false ) {
 330                 return $this->getVal( $name, $default ) ? true : false;
 331         }
 332
 333         /**
 334          * Return true if the named value is set in the input, whatever that
 335          * value is (even "0"). Return false if the named value is not set.
 336          * Example use is checking for the presence of check boxes in forms.
 337          * @param string $name
 338          * @return bool
 339          */
 340         function getCheck( $name ) {
 341                 # Checkboxes and buttons are only present when clicked
 342                 # Presence connotes truth, abscense false
 343                 $val = $this->getVal( $name, NULL );
 344                 return isset( $val );
 345         }
 346
 347         /**
 348          * Fetch a text string from the given array or return $default if it's not
 349          * set. \r is stripped from the text, and with some language modules there
 350          * is an input transliteration applied. This should generally be used for
 351          * form <textarea> and <input> fields. Used for user-supplied freeform text
 352          * input (for which input transformations may be required - e.g. Esperanto
 353          * x-coding).
 354          *
 355          * @param string $name
 356          * @param string $default optional
 357          * @return string
 358          */
 359         function getText( $name, $default = '' ) {
 360                 global $wgContLang;
 361                 $val = $this->getVal( $name, $default );
 362                 return str_replace( "\r\n", "\n",
 363                         $wgContLang->recodeInput( $val ) );
 364         }
 365
 366         /**
 367          * Extracts the given named values into an array.
 368          * If no arguments are given, returns all input values.
 369          * No transformation is performed on the values.
 370          */
 371         function getValues() {
 372                 $names = func_get_args();
 373                 if ( count( $names ) == 0 ) {
 374                         $names = array_keys( $this->data );
 375                 }
 376
 377                 $retVal = array();
 378                 foreach ( $names as $name ) {
 379                         $value = $this->getVal( $name );
 380                         if ( !is_null( $value ) ) {
 381                                 $retVal[$name] = $value;
 382                         }
 383                 }
 384                 return $retVal;
 385         }
 386
 387         /**
 388          * Returns true if the present request was reached by a POST operation,
 389          * false otherwise (GET, HEAD, or command-line).
 390          *
 391          * Note that values retrieved by the object may come from the
 392          * GET URL etc even on a POST request.
 393          *
 394          * @return bool
 395          */
 396         function wasPosted() {
 397                 return $_SERVER['REQUEST_METHOD'] == 'POST';
 398         }
 399
 400         /**
 401          * Returns true if there is a session cookie set.
 402          * This does not necessarily mean that the user is logged in!
 403          *
 404          * If you want to check for an open session, use session_id()
 405          * instead; that will also tell you if the session was opened
 406          * during the current request (in which case the cookie will
 407          * be sent back to the client at the end of the script run).
 408          *
 409          * @return bool
 410          */
 411         function checkSessionCookie() {
 412                 return isset( $_COOKIE[session_name()] );
 413         }
 414
 415         /**
 416          * Return the path portion of the request URI.
 417          * @return string
 418          */
 419         function getRequestURL() {
 420                 if( isset( $_SERVER['REQUEST_URI'] ) ) {
 421                         $base = $_SERVER['REQUEST_URI'];
 422                 } elseif( isset( $_SERVER['SCRIPT_NAME'] ) ) {
 423                         // Probably IIS; doesn't set REQUEST_URI
 424                         $base = $_SERVER['SCRIPT_NAME'];
 425                         if( isset( $_SERVER['QUERY_STRING'] ) && $_SERVER['QUERY_STRING'] != '' ) {
 426                                 $base .= '?' . $_SERVER['QUERY_STRING'];
 427                         }
 428                 } else {
 429                         // This shouldn't happen!
 430                         throw new MWException( "Web server doesn't provide either " .
 431                                 "REQUEST_URI or SCRIPT_NAME. Report details of your " .
 432                                 "web server configuration to http://bugzilla.wikimedia.org/" );
 433                 }
 434                 // User-agents should not send a fragment with the URI, but
 435                 // if they do, and the web server passes it on to us, we
 436                 // need to strip it or we get false-positive redirect loops
 437                 // or weird output URLs
 438                 $hash = strpos( $base, '#' );
 439                 if( $hash !== false ) {
 440                         $base = substr( $base, 0, $hash );
 441                 }
 442                 if( $base{0} == '/' ) {
 443                         return $base;
 444                 } else {
 445                         // We may get paths with a host prepended; strip it.
 446                         return preg_replace( '!^[^:]+://[^/]+/!', '/', $base );
 447                 }
 448         }
 449
 450         /**
 451          * Return the request URI with the canonical service and hostname.
 452          * @return string
 453          */
 454         function getFullRequestURL() {
 455                 global $wgServer;
 456                 return $wgServer . $this->getRequestURL();
 457         }
 458
 459         /**
 460          * Take an arbitrary query and rewrite the present URL to include it
 461          * @param $query String: query string fragment; do not include initial '?'
 462          * @return string
 463          */
 464         function appendQuery( $query ) {
 465                 global $wgTitle;
 466                 $basequery = '';
 467                 foreach( $_GET as $var => $val ) {
 468                         if ( $var == 'title' )
 469                                 continue;
 470                         if ( is_array( $val ) )
 471                                 /* This will happen given a request like
 472                                  * http://en.wikipedia.org/w/index.php?title[]=Special:Userlogin&returnto[]=Main_Page
 473                                  */
 474                                 continue;
 475                         $basequery .= '&' . urlencode( $var ) . '=' . urlencode( $val );
 476                 }
 477                 $basequery .= '&' . $query;
 478
 479                 # Trim the extra &
 480                 $basequery = substr( $basequery, 1 );
 481                 return $wgTitle->getLocalURL( $basequery );
 482         }
 483
 484         /**
 485          * HTML-safe version of appendQuery().
 486          * @param $query String: query string fragment; do not include initial '?'
 487          * @return string
 488          */
 489         function escapeAppendQuery( $query ) {
 490                 return htmlspecialchars( $this->appendQuery( $query ) );
 491         }
 492
 493         /**
 494          * Check for limit and offset parameters on the input, and return sensible
 495          * defaults if not given. The limit must be positive and is capped at 5000.
 496          * Offset must be positive but is not capped.
 497          *
 498          * @param $deflimit Integer: limit to use if no input and the user hasn't set the option.
 499          * @param $optionname String: to specify an option other than rclimit to pull from.
 500          * @return array first element is limit, second is offset
 501          */
 502         function getLimitOffset( $deflimit = 50, $optionname = 'rclimit' ) {
 503                 global $wgUser;
 504
 505                 $limit = $this->getInt( 'limit', 0 );
 506                 if( $limit < 0 ) $limit = 0;
 507                 if( ( $limit == 0 ) && ( $optionname != '' ) ) {
 508                         $limit = (int)$wgUser->getOption( $optionname );
 509                 }
 510                 if( $limit <= 0 ) $limit = $deflimit;
 511                 if( $limit > 5000 ) $limit = 5000; # We have *some* limits...
 512
 513                 $offset = $this->getInt( 'offset', 0 );
 514                 if( $offset < 0 ) $offset = 0;
 515
 516                 return array( $limit, $offset );
 517         }
 518
 519         /**
 520          * Return the path to the temporary file where PHP has stored the upload.
 521          * @param $key String:
 522          * @return string or NULL if no such file.
 523          */
 524         function getFileTempname( $key ) {
 525                 if( !isset( $_FILES[$key] ) ) {
 526                         return NULL;
 527                 }
 528                 return $_FILES[$key]['tmp_name'];
 529         }
 530
 531         /**
 532          * Return the size of the upload, or 0.
 533          * @param $key String:
 534          * @return integer
 535          */
 536         function getFileSize( $key ) {
 537                 if( !isset( $_FILES[$key] ) ) {
 538                         return 0;
 539                 }
 540                 return $_FILES[$key]['size'];
 541         }
 542
 543         /**
 544          * Return the upload error or 0
 545          * @param $key String:
 546          * @return integer
 547          */
 548         function getUploadError( $key ) {
 549                 if( !isset( $_FILES[$key] ) || !isset( $_FILES[$key]['error'] ) ) {
 550                         return 0/*UPLOAD_ERR_OK*/;
 551                 }
 552                 return $_FILES[$key]['error'];
 553         }
 554
 555         /**
 556          * Return the original filename of the uploaded file, as reported by
 557          * the submitting user agent. HTML-style character entities are
 558          * interpreted and normalized to Unicode normalization form C, in part
 559          * to deal with weird input from Safari with non-ASCII filenames.
 560          *
 561          * Other than this the name is not verified for being a safe filename.
 562          *
 563          * @param $key String:
 564          * @return string or NULL if no such file.
 565          */
 566         function getFileName( $key ) {
 567                 if( !isset( $_FILES[$key] ) ) {
 568                         return NULL;
 569                 }
 570                 $name = $_FILES[$key]['name'];
 571
 572                 # Safari sends filenames in HTML-encoded Unicode form D...
 573                 # Horrid and evil! Let's try to make some kind of sense of it.
 574                 $name = Sanitizer::decodeCharReferences( $name );
 575                 $name = UtfNormal::cleanUp( $name );
 576                 wfDebug( "WebRequest::getFileName() '" . $_FILES[$key]['name'] . "' normalized to '$name'\n" );
 577                 return $name;
 578         }
 579
 580         /**
 581          * Return a handle to WebResponse style object, for setting cookies,
 582          * headers and other stuff, for Request being worked on.
 583          */
 584         function response() {
 585                 /* Lazy initialization of response object for this request */
 586                 if (!is_object($this->_response)) {
 587                         $this->_response = new WebResponse;
 588                 }
 589                 return $this->_response;
 590         }
 591
 592 }
 593
 594 /**
 595  * WebRequest clone which takes values from a provided array.
 596  *
 597  */
 598 class FauxRequest extends WebRequest {
 599         var $wasPosted = false;
 600
 601         /**
 602          * @param array $data Array of *non*-urlencoded key => value pairs, the
 603          *   fake GET/POST values
 604          * @param bool $wasPosted Whether to treat the data as POST
 605          */
 606         function FauxRequest( $data, $wasPosted = false ) {
 607                 if( is_array( $data ) ) {
 608                         $this->data = $data;
 609                 } else {
 610                         throw new MWException( "FauxRequest() got bogus data" );
 611                 }
 612                 $this->wasPosted = $wasPosted;
 613         }
 614
 615         function getText( $name, $default = '' ) {
 616                 # Override; don't recode since we're using internal data
 617                 return (string)$this->getVal( $name, $default );
 618         }
 619
 620         function getValues() {
 621                 return $this->data;
 622         }
 623
 624         function wasPosted() {
 625                 return $this->wasPosted;
 626         }
 627
 628         function checkSessionCookie() {
 629                 return false;
 630         }
 631
 632         function getRequestURL() {
 633                 throw new MWException( 'FauxRequest::getRequestURL() not implemented' );
 634         }
 635
 636         function appendQuery( $query ) {
 637                 throw new MWException( 'FauxRequest::appendQuery() not implemented' );
 638         }
 639
 640 }
 641
 642