includes/title/MediaWikiTitleCodec.php

   1 <?php
   2 /**
   3  * A codec for %MediaWiki page titles.
   4  *
   5  * This program is free software; you can redistribute it and/or modify
   6  * it under the terms of the GNU General Public License as published by
   7  * the Free Software Foundation; either version 2 of the License, or
   8  * (at your option) any later version.
   9  *
  10  * This program is distributed in the hope that it will be useful,
  11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  13  * GNU General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU General Public License along
  16  * with this program; if not, write to the Free Software Foundation, Inc.,
  17  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  18  * http://www.gnu.org/copyleft/gpl.html
  19  *
  20  * @file
  21  * @license GPL 2+
  22  * @author Daniel Kinzler
  23  */
  24 use MediaWiki\Linker\LinkTarget;
  25
  26 /**
  27  * A codec for %MediaWiki page titles.
  28  *
  29  * @note Normalization and validation is applied while parsing, not when formatting.
  30  * It's possible to construct a TitleValue with an invalid title, and use MediaWikiTitleCodec
  31  * to generate an (invalid) title string from it. TitleValues should be constructed only
  32  * via parseTitle() or from a (semi)trusted source, such as the database.
  33  *
  34  * @see https://www.mediawiki.org/wiki/Requests_for_comment/TitleValue
  35  * @since 1.23
  36  */
  37 class MediaWikiTitleCodec implements TitleFormatter, TitleParser {
  38         /**
  39          * @var Language
  40          */
  41         protected $language;
  42
  43         /**
  44          * @var GenderCache
  45          */
  46         protected $genderCache;
  47
  48         /**
  49          * @var string[]
  50          */
  51         protected $localInterwikis;
  52
  53         /**
  54          * @param Language $language The language object to use for localizing namespace names.
  55          * @param GenderCache $genderCache The gender cache for generating gendered namespace names
  56          * @param string[]|string $localInterwikis
  57          */
  58         public function __construct( Language $language, GenderCache $genderCache,
  59                 $localInterwikis = []
  60         ) {
  61                 $this->language = $language;
  62                 $this->genderCache = $genderCache;
  63                 $this->localInterwikis = (array)$localInterwikis;
  64         }
  65
  66         /**
  67          * @see TitleFormatter::getNamespaceName()
  68          *
  69          * @param int $namespace
  70          * @param string $text
  71          *
  72          * @throws InvalidArgumentException If the namespace is invalid
  73          * @return string
  74          */
  75         public function getNamespaceName( $namespace, $text ) {
  76                 if ( $this->language->needsGenderDistinction() &&
  77                         MWNamespace::hasGenderDistinction( $namespace )
  78                 ) {
  79
  80                         // NOTE: we are assuming here that the title text is a user name!
  81                         $gender = $this->genderCache->getGenderOf( $text, __METHOD__ );
  82                         $name = $this->language->getGenderNsText( $namespace, $gender );
  83                 } else {
  84                         $name = $this->language->getNsText( $namespace );
  85                 }
  86
  87                 if ( $name === false ) {
  88                         throw new InvalidArgumentException( 'Unknown namespace ID: ' . $namespace );
  89                 }
  90
  91                 return $name;
  92         }
  93
  94         /**
  95          * @see TitleFormatter::formatTitle()
  96          *
  97          * @param int|bool $namespace The namespace ID (or false, if the namespace should be ignored)
  98          * @param string $text The page title. Should be valid. Only minimal normalization is applied.
  99          *        Underscores will be replaced.
 100          * @param string $fragment The fragment name (may be empty).
 101          * @param string $interwiki The interwiki name (may be empty).
 102          *
 103          * @throws InvalidArgumentException If the namespace is invalid
 104          * @return string
 105          */
 106         public function formatTitle( $namespace, $text, $fragment = '', $interwiki = '' ) {
 107                 if ( $namespace !== false ) {
 108                         // Try to get a namespace name, but fallback
 109                         // to empty string if it doesn't exist
 110                         try {
 111                                 $nsName = $this->getNamespaceName( $namespace, $text );
 112                         } catch ( InvalidArgumentException $e ) {
 113                                 $nsName = '';
 114                         }
 115
 116                         if ( $namespace !== 0 ) {
 117                                 $text = $nsName . ':' . $text;
 118                         }
 119                 }
 120
 121                 if ( $fragment !== '' ) {
 122                         $text = $text . '#' . $fragment;
 123                 }
 124
 125                 if ( $interwiki !== '' ) {
 126                         $text = $interwiki . ':' . $text;
 127                 }
 128
 129                 $text = str_replace( '_', ' ', $text );
 130
 131                 return $text;
 132         }
 133
 134         /**
 135          * Parses the given text and constructs a TitleValue. Normalization
 136          * is applied according to the rules appropriate for the form specified by $form.
 137          *
 138          * @param string $text The text to parse
 139          * @param int $defaultNamespace Namespace to assume per default (usually NS_MAIN)
 140          *
 141          * @throws MalformedTitleException
 142          * @return TitleValue
 143          */
 144         public function parseTitle( $text, $defaultNamespace ) {
 145                 // NOTE: this is an ugly cludge that allows this class to share the
 146                 // code for parsing with the old Title class. The parser code should
 147                 // be refactored to avoid this.
 148                 $parts = $this->splitTitleString( $text, $defaultNamespace );
 149
 150                 // Relative fragment links are not supported by TitleValue
 151                 if ( $parts['dbkey'] === '' ) {
 152                         throw new MalformedTitleException( 'title-invalid-empty', $text );
 153                 }
 154
 155                 return new TitleValue(
 156                         $parts['namespace'],
 157                         $parts['dbkey'],
 158                         $parts['fragment'],
 159                         $parts['interwiki']
 160                 );
 161         }
 162
 163         /**
 164          * @see TitleFormatter::getText()
 165          *
 166          * @param LinkTarget $title
 167          *
 168          * @return string $title->getText()
 169          */
 170         public function getText( LinkTarget $title ) {
 171                 return $this->formatTitle( false, $title->getText(), '' );
 172         }
 173
 174         /**
 175          * @see TitleFormatter::getText()
 176          *
 177          * @param LinkTarget $title
 178          *
 179          * @return string
 180          */
 181         public function getPrefixedText( LinkTarget $title ) {
 182                 return $this->formatTitle(
 183                         $title->getNamespace(),
 184                         $title->getText(),
 185                         '',
 186                         $title->getInterwiki()
 187                 );
 188         }
 189
 190         /**
 191          * @since 1.27
 192          * @see TitleFormatter::getPrefixedDBkey()
 193          * @param LinkTarget $target
 194          * @return string
 195          */
 196         public function getPrefixedDBkey( LinkTarget $target ) {
 197                 $key = '';
 198                 if ( $target->isExternal() ) {
 199                         $key .= $target->getInterwiki() . ':';
 200                 }
 201                 // Try to get a namespace name, but fallback
 202                 // to empty string if it doesn't exist
 203                 try {
 204                         $nsName = $this->getNamespaceName(
 205                                 $target->getNamespace(),
 206                                 $target->getText()
 207                         );
 208                 } catch ( InvalidArgumentException $e ) {
 209                         $nsName = '';
 210                 }
 211
 212                 if ( $target->getNamespace() !== 0 ) {
 213                         $key .= $nsName . ':';
 214                 }
 215
 216                 $key .= $target->getText();
 217
 218                 return strtr( $key, ' ', '_' );
 219         }
 220
 221         /**
 222          * @see TitleFormatter::getText()
 223          *
 224          * @param LinkTarget $title
 225          *
 226          * @return string
 227          */
 228         public function getFullText( LinkTarget $title ) {
 229                 return $this->formatTitle(
 230                         $title->getNamespace(),
 231                         $title->getText(),
 232                         $title->getFragment(),
 233                         $title->getInterwiki()
 234                 );
 235         }
 236
 237         /**
 238          * Normalizes and splits a title string.
 239          *
 240          * This function removes illegal characters, splits off the interwiki and
 241          * namespace prefixes, sets the other forms, and canonicalizes
 242          * everything.
 243          *
 244          * @todo this method is only exposed as a temporary measure to ease refactoring.
 245          * It was copied with minimal changes from Title::secureAndSplit().
 246          *
 247          * @todo This method should be split up and an appropriate interface
 248          * defined for use by the Title class.
 249          *
 250          * @param string $text
 251          * @param int $defaultNamespace
 252          *
 253          * @throws MalformedTitleException If $text is not a valid title string.
 254          * @return array A map with the fields 'interwiki', 'fragment', 'namespace',
 255          *         'user_case_dbkey', and 'dbkey'.
 256          */
 257         public function splitTitleString( $text, $defaultNamespace = NS_MAIN ) {
 258                 $dbkey = str_replace( ' ', '_', $text );
 259
 260                 # Initialisation
 261                 $parts = [
 262                         'interwiki' => '',
 263                         'local_interwiki' => false,
 264                         'fragment' => '',
 265                         'namespace' => $defaultNamespace,
 266                         'dbkey' => $dbkey,
 267                         'user_case_dbkey' => $dbkey,
 268                 ];
 269
 270                 # Strip Unicode bidi override characters.
 271                 # Sometimes they slip into cut-n-pasted page titles, where the
 272                 # override chars get included in list displays.
 273                 $dbkey = preg_replace( '/\xE2\x80[\x8E\x8F\xAA-\xAE]/S', '', $dbkey );
 274
 275                 # Clean up whitespace
 276                 # Note: use of the /u option on preg_replace here will cause
 277                 # input with invalid UTF-8 sequences to be nullified out in PHP 5.2.x,
 278                 # conveniently disabling them.
 279                 $dbkey = preg_replace(
 280                         '/[ _\xA0\x{1680}\x{180E}\x{2000}-\x{200A}\x{2028}\x{2029}\x{202F}\x{205F}\x{3000}]+/u',
 281                         '_',
 282                         $dbkey
 283                 );
 284                 $dbkey = trim( $dbkey, '_' );
 285
 286                 if ( strpos( $dbkey, UtfNormal\Constants::UTF8_REPLACEMENT ) !== false ) {
 287                         # Contained illegal UTF-8 sequences or forbidden Unicode chars.
 288                         throw new MalformedTitleException( 'title-invalid-utf8', $text );
 289                 }
 290
 291                 $parts['dbkey'] = $dbkey;
 292
 293                 # Initial colon indicates main namespace rather than specified default
 294                 # but should not create invalid {ns,title} pairs such as {0,Project:Foo}
 295                 if ( $dbkey !== '' && ':' == $dbkey[0] ) {
 296                         $parts['namespace'] = NS_MAIN;
 297                         $dbkey = substr( $dbkey, 1 ); # remove the colon but continue processing
 298                         $dbkey = trim( $dbkey, '_' ); # remove any subsequent whitespace
 299                 }
 300
 301                 if ( $dbkey == '' ) {
 302                         throw new MalformedTitleException( 'title-invalid-empty', $text );
 303                 }
 304
 305                 # Namespace or interwiki prefix
 306                 $prefixRegexp = "/^(.+?)_*:_*(.*)$/S";
 307                 do {
 308                         $m = [];
 309                         if ( preg_match( $prefixRegexp, $dbkey, $m ) ) {
 310                                 $p = $m[1];
 311                                 $ns = $this->language->getNsIndex( $p );
 312                                 if ( $ns !== false ) {
 313                                         # Ordinary namespace
 314                                         $dbkey = $m[2];
 315                                         $parts['namespace'] = $ns;
 316                                         # For Talk:X pages, check if X has a "namespace" prefix
 317                                         if ( $ns == NS_TALK && preg_match( $prefixRegexp, $dbkey, $x ) ) {
 318                                                 if ( $this->language->getNsIndex( $x[1] ) ) {
 319                                                         # Disallow Talk:File:x type titles...
 320                                                         throw new MalformedTitleException( 'title-invalid-talk-namespace', $text );
 321                                                 } elseif ( Interwiki::isValidInterwiki( $x[1] ) ) {
 322                                                         // TODO: get rid of global state!
 323                                                         # Disallow Talk:Interwiki:x type titles...
 324                                                         throw new MalformedTitleException( 'title-invalid-talk-namespace', $text );
 325                                                 }
 326                                         }
 327                                 } elseif ( Interwiki::isValidInterwiki( $p ) ) {
 328                                         # Interwiki link
 329                                         $dbkey = $m[2];
 330                                         $parts['interwiki'] = $this->language->lc( $p );
 331
 332                                         # Redundant interwiki prefix to the local wiki
 333                                         foreach ( $this->localInterwikis as $localIW ) {
 334                                                 if ( 0 == strcasecmp( $parts['interwiki'], $localIW ) ) {
 335                                                         if ( $dbkey == '' ) {
 336                                                                 # Empty self-links should point to the Main Page, to ensure
 337                                                                 # compatibility with cross-wiki transclusions and the like.
 338                                                                 $mainPage = Title::newMainPage();
 339                                                                 return [
 340                                                                         'interwiki' => $mainPage->getInterwiki(),
 341                                                                         'local_interwiki' => true,
 342                                                                         'fragment' => $mainPage->getFragment(),
 343                                                                         'namespace' => $mainPage->getNamespace(),
 344                                                                         'dbkey' => $mainPage->getDBkey(),
 345                                                                         'user_case_dbkey' => $mainPage->getUserCaseDBKey()
 346                                                                 ];
 347                                                         }
 348                                                         $parts['interwiki'] = '';
 349                                                         # local interwikis should behave like initial-colon links
 350                                                         $parts['local_interwiki'] = true;
 351
 352                                                         # Do another namespace split...
 353                                                         continue 2;
 354                                                 }
 355                                         }
 356
 357                                         # If there's an initial colon after the interwiki, that also
 358                                         # resets the default namespace
 359                                         if ( $dbkey !== '' && $dbkey[0] == ':' ) {
 360                                                 $parts['namespace'] = NS_MAIN;
 361                                                 $dbkey = substr( $dbkey, 1 );
 362                                         }
 363                                 }
 364                                 # If there's no recognized interwiki or namespace,
 365                                 # then let the colon expression be part of the title.
 366                         }
 367                         break;
 368                 } while ( true );
 369
 370                 $fragment = strstr( $dbkey, '#' );
 371                 if ( false !== $fragment ) {
 372                         $parts['fragment'] = str_replace( '_', ' ', substr( $fragment, 1 ) );
 373                         $dbkey = substr( $dbkey, 0, strlen( $dbkey ) - strlen( $fragment ) );
 374                         # remove whitespace again: prevents "Foo_bar_#"
 375                         # becoming "Foo_bar_"
 376                         $dbkey = preg_replace( '/_*$/', '', $dbkey );
 377                 }
 378
 379                 # Reject illegal characters.
 380                 $rxTc = self::getTitleInvalidRegex();
 381                 $matches = [];
 382                 if ( preg_match( $rxTc, $dbkey, $matches ) ) {
 383                         throw new MalformedTitleException( 'title-invalid-characters', $text, [ $matches[0] ] );
 384                 }
 385
 386                 # Pages with "/./" or "/../" appearing in the URLs will often be un-
 387                 # reachable due to the way web browsers deal with 'relative' URLs.
 388                 # Also, they conflict with subpage syntax.  Forbid them explicitly.
 389                 if (
 390                         strpos( $dbkey, '.' ) !== false &&
 391                         (
 392                                 $dbkey === '.' || $dbkey === '..' ||
 393                                 strpos( $dbkey, './' ) === 0 ||
 394                                 strpos( $dbkey, '../' ) === 0 ||
 395                                 strpos( $dbkey, '/./' ) !== false ||
 396                                 strpos( $dbkey, '/../' ) !== false ||
 397                                 substr( $dbkey, -2 ) == '/.' ||
 398                                 substr( $dbkey, -3 ) == '/..'
 399                         )
 400                 ) {
 401                         throw new MalformedTitleException( 'title-invalid-relative', $text );
 402                 }
 403
 404                 # Magic tilde sequences? Nu-uh!
 405                 if ( strpos( $dbkey, '~~~' ) !== false ) {
 406                         throw new MalformedTitleException( 'title-invalid-magic-tilde', $text );
 407                 }
 408
 409                 # Limit the size of titles to 255 bytes. This is typically the size of the
 410                 # underlying database field. We make an exception for special pages, which
 411                 # don't need to be stored in the database, and may edge over 255 bytes due
 412                 # to subpage syntax for long titles, e.g. [[Special:Block/Long name]]
 413                 $maxLength = ( $parts['namespace'] != NS_SPECIAL ) ? 255 : 512;
 414                 if ( strlen( $dbkey ) > $maxLength ) {
 415                         throw new MalformedTitleException( 'title-invalid-too-long', $text,
 416                                 [ Message::numParam( $maxLength ) ] );
 417                 }
 418
 419                 # Normally, all wiki links are forced to have an initial capital letter so [[foo]]
 420                 # and [[Foo]] point to the same place.  Don't force it for interwikis, since the
 421                 # other site might be case-sensitive.
 422                 $parts['user_case_dbkey'] = $dbkey;
 423                 if ( $parts['interwiki'] === '' ) {
 424                         $dbkey = Title::capitalize( $dbkey, $parts['namespace'] );
 425                 }
 426
 427                 # Can't make a link to a namespace alone... "empty" local links can only be
 428                 # self-links with a fragment identifier.
 429                 if ( $dbkey == '' && $parts['interwiki'] === '' ) {
 430                         if ( $parts['namespace'] != NS_MAIN ) {
 431                                 throw new MalformedTitleException( 'title-invalid-empty', $text );
 432                         }
 433                 }
 434
 435                 // Allow IPv6 usernames to start with '::' by canonicalizing IPv6 titles.
 436                 // IP names are not allowed for accounts, and can only be referring to
 437                 // edits from the IP. Given '::' abbreviations and caps/lowercaps,
 438                 // there are numerous ways to present the same IP. Having sp:contribs scan
 439                 // them all is silly and having some show the edits and others not is
 440                 // inconsistent. Same for talk/userpages. Keep them normalized instead.
 441                 if ( $parts['namespace'] == NS_USER || $parts['namespace'] == NS_USER_TALK ) {
 442                         $dbkey = IP::sanitizeIP( $dbkey );
 443                 }
 444
 445                 // Any remaining initial :s are illegal.
 446                 if ( $dbkey !== '' && ':' == $dbkey[0] ) {
 447                         throw new MalformedTitleException( 'title-invalid-leading-colon', $text );
 448                 }
 449
 450                 # Fill fields
 451                 $parts['dbkey'] = $dbkey;
 452
 453                 return $parts;
 454         }
 455
 456         /**
 457          * Returns a simple regex that will match on characters and sequences invalid in titles.
 458          * Note that this doesn't pick up many things that could be wrong with titles, but that
 459          * replacing this regex with something valid will make many titles valid.
 460          * Previously Title::getTitleInvalidRegex()
 461          *
 462          * @return string Regex string
 463          * @since 1.25
 464          */
 465         public static function getTitleInvalidRegex() {
 466                 static $rxTc = false;
 467                 if ( !$rxTc ) {
 468                         # Matching titles will be held as illegal.
 469                         $rxTc = '/' .
 470                                 # Any character not allowed is forbidden...
 471                                 '[^' . Title::legalChars() . ']' .
 472                                 # URL percent encoding sequences interfere with the ability
 473                                 # to round-trip titles -- you can't link to them consistently.
 474                                 '|%[0-9A-Fa-f]{2}' .
 475                                 # XML/HTML character references produce similar issues.
 476                                 '|&[A-Za-z0-9\x80-\xff]+;' .
 477                                 '|&#[0-9]+;' .
 478                                 '|&#x[0-9A-Fa-f]+;' .
 479                                 '/S';
 480                 }
 481
 482                 return $rxTc;
 483         }
 484 }