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\Interwiki\InterwikiLookup;
  25 use MediaWiki\MediaWikiServices;
  26 use MediaWiki\Linker\LinkTarget;
  27
  28 /**
  29  * A codec for %MediaWiki page titles.
  30  *
  31  * @note Normalization and validation is applied while parsing, not when formatting.
  32  * It's possible to construct a TitleValue with an invalid title, and use MediaWikiTitleCodec
  33  * to generate an (invalid) title string from it. TitleValues should be constructed only
  34  * via parseTitle() or from a (semi)trusted source, such as the database.
  35  *
  36  * @see https://www.mediawiki.org/wiki/Requests_for_comment/TitleValue
  37  * @since 1.23
  38  */
  39 class MediaWikiTitleCodec implements TitleFormatter, TitleParser {
  40         /**
  41          * @var Language
  42          */
  43         protected $language;
  44
  45         /**
  46          * @var GenderCache
  47          */
  48         protected $genderCache;
  49
  50         /**
  51          * @var string[]
  52          */
  53         protected $localInterwikis;
  54
  55         /**
  56          * @var InterwikiLookup
  57          */
  58         protected $interwikiLookup;
  59
  60         /**
  61          * @param Language $language The language object to use for localizing namespace names.
  62          * @param GenderCache $genderCache The gender cache for generating gendered namespace names
  63          * @param string[]|string $localInterwikis
  64          * @param InterwikiLookup|null $interwikiLookup
  65          */
  66         public function __construct( Language $language, GenderCache $genderCache,
  67                 $localInterwikis = [], $interwikiLookup = null
  68         ) {
  69                 $this->language = $language;
  70                 $this->genderCache = $genderCache;
  71                 $this->localInterwikis = (array)$localInterwikis;
  72                 $this->interwikiLookup = $interwikiLookup ?:
  73                         MediaWikiServices::getInstance()->getInterwikiLookup();
  74         }
  75
  76         /**
  77          * @see TitleFormatter::getNamespaceName()
  78          *
  79          * @param int $namespace
  80          * @param string $text
  81          *
  82          * @throws InvalidArgumentException If the namespace is invalid
  83          * @return string
  84          */
  85         public function getNamespaceName( $namespace, $text ) {
  86                 if ( $this->language->needsGenderDistinction() &&
  87                         MWNamespace::hasGenderDistinction( $namespace )
  88                 ) {
  89                         // NOTE: we are assuming here that the title text is a user name!
  90                         $gender = $this->genderCache->getGenderOf( $text, __METHOD__ );
  91                         $name = $this->language->getGenderNsText( $namespace, $gender );
  92                 } else {
  93                         $name = $this->language->getNsText( $namespace );
  94                 }
  95
  96                 if ( $name === false ) {
  97                         throw new InvalidArgumentException( 'Unknown namespace ID: ' . $namespace );
  98                 }
  99
 100                 return $name;
 101         }
 102
 103         /**
 104          * @see TitleFormatter::formatTitle()
 105          *
 106          * @param int|bool $namespace The namespace ID (or false, if the namespace should be ignored)
 107          * @param string $text The page title. Should be valid. Only minimal normalization is applied.
 108          *        Underscores will be replaced.
 109          * @param string $fragment The fragment name (may be empty).
 110          * @param string $interwiki The interwiki name (may be empty).
 111          *
 112          * @throws InvalidArgumentException If the namespace is invalid
 113          * @return string
 114          */
 115         public function formatTitle( $namespace, $text, $fragment = '', $interwiki = '' ) {
 116                 if ( $namespace !== false ) {
 117                         // Try to get a namespace name, but fallback
 118                         // to empty string if it doesn't exist
 119                         try {
 120                                 $nsName = $this->getNamespaceName( $namespace, $text );
 121                         } catch ( InvalidArgumentException $e ) {
 122                                 $nsName = '';
 123                         }
 124
 125                         if ( $namespace !== 0 ) {
 126                                 $text = $nsName . ':' . $text;
 127                         }
 128                 }
 129
 130                 if ( $fragment !== '' ) {
 131                         $text = $text . '#' . $fragment;
 132                 }
 133
 134                 if ( $interwiki !== '' ) {
 135                         $text = $interwiki . ':' . $text;
 136                 }
 137
 138                 $text = str_replace( '_', ' ', $text );
 139
 140                 return $text;
 141         }
 142
 143         /**
 144          * Parses the given text and constructs a TitleValue. Normalization
 145          * is applied according to the rules appropriate for the form specified by $form.
 146          *
 147          * @param string $text The text to parse
 148          * @param int $defaultNamespace Namespace to assume per default (usually NS_MAIN)
 149          *
 150          * @throws MalformedTitleException
 151          * @return TitleValue
 152          */
 153         public function parseTitle( $text, $defaultNamespace ) {
 154                 // NOTE: this is an ugly cludge that allows this class to share the
 155                 // code for parsing with the old Title class. The parser code should
 156                 // be refactored to avoid this.
 157                 $parts = $this->splitTitleString( $text, $defaultNamespace );
 158
 159                 // Relative fragment links are not supported by TitleValue
 160                 if ( $parts['dbkey'] === '' ) {
 161                         throw new MalformedTitleException( 'title-invalid-empty', $text );
 162                 }
 163
 164                 return new TitleValue(
 165                         $parts['namespace'],
 166                         $parts['dbkey'],
 167                         $parts['fragment'],
 168                         $parts['interwiki']
 169                 );
 170         }
 171
 172         /**
 173          * @see TitleFormatter::getText()
 174          *
 175          * @param LinkTarget $title
 176          *
 177          * @return string $title->getText()
 178          */
 179         public function getText( LinkTarget $title ) {
 180                 return $this->formatTitle( false, $title->getText(), '' );
 181         }
 182
 183         /**
 184          * @see TitleFormatter::getText()
 185          *
 186          * @param LinkTarget $title
 187          *
 188          * @return string
 189          */
 190         public function getPrefixedText( LinkTarget $title ) {
 191                 return $this->formatTitle(
 192                         $title->getNamespace(),
 193                         $title->getText(),
 194                         '',
 195                         $title->getInterwiki()
 196                 );
 197         }
 198
 199         /**
 200          * @since 1.27
 201          * @see TitleFormatter::getPrefixedDBkey()
 202          * @param LinkTarget $target
 203          * @return string
 204          */
 205         public function getPrefixedDBkey( LinkTarget $target ) {
 206                 $key = '';
 207                 if ( $target->isExternal() ) {
 208                         $key .= $target->getInterwiki() . ':';
 209                 }
 210                 // Try to get a namespace name, but fallback
 211                 // to empty string if it doesn't exist
 212                 try {
 213                         $nsName = $this->getNamespaceName(
 214                                 $target->getNamespace(),
 215                                 $target->getText()
 216                         );
 217                 } catch ( InvalidArgumentException $e ) {
 218                         $nsName = '';
 219                 }
 220
 221                 if ( $target->getNamespace() !== 0 ) {
 222                         $key .= $nsName . ':';
 223                 }
 224
 225                 $key .= $target->getText();
 226
 227                 return strtr( $key, ' ', '_' );
 228         }
 229
 230         /**
 231          * @see TitleFormatter::getText()
 232          *
 233          * @param LinkTarget $title
 234          *
 235          * @return string
 236          */
 237         public function getFullText( LinkTarget $title ) {
 238                 return $this->formatTitle(
 239                         $title->getNamespace(),
 240                         $title->getText(),
 241                         $title->getFragment(),
 242                         $title->getInterwiki()
 243                 );
 244         }
 245
 246         /**
 247          * Normalizes and splits a title string.
 248          *
 249          * This function removes illegal characters, splits off the interwiki and
 250          * namespace prefixes, sets the other forms, and canonicalizes
 251          * everything.
 252          *
 253          * @todo this method is only exposed as a temporary measure to ease refactoring.
 254          * It was copied with minimal changes from Title::secureAndSplit().
 255          *
 256          * @todo This method should be split up and an appropriate interface
 257          * defined for use by the Title class.
 258          *
 259          * @param string $text
 260          * @param int $defaultNamespace
 261          *
 262          * @throws MalformedTitleException If $text is not a valid title string.
 263          * @return array A map with the fields 'interwiki', 'fragment', 'namespace',
 264          *         'user_case_dbkey', and 'dbkey'.
 265          */
 266         public function splitTitleString( $text, $defaultNamespace = NS_MAIN ) {
 267                 $dbkey = str_replace( ' ', '_', $text );
 268
 269                 # Initialisation
 270                 $parts = [
 271                         'interwiki' => '',
 272                         'local_interwiki' => false,
 273                         'fragment' => '',
 274                         'namespace' => $defaultNamespace,
 275                         'dbkey' => $dbkey,
 276                         'user_case_dbkey' => $dbkey,
 277                 ];
 278
 279                 # Strip Unicode bidi override characters.
 280                 # Sometimes they slip into cut-n-pasted page titles, where the
 281                 # override chars get included in list displays.
 282                 $dbkey = preg_replace( '/\xE2\x80[\x8E\x8F\xAA-\xAE]/S', '', $dbkey );
 283
 284                 # Clean up whitespace
 285                 # Note: use of the /u option on preg_replace here will cause
 286                 # input with invalid UTF-8 sequences to be nullified out in PHP 5.2.x,
 287                 # conveniently disabling them.
 288                 $dbkey = preg_replace(
 289                         '/[ _\xA0\x{1680}\x{180E}\x{2000}-\x{200A}\x{2028}\x{2029}\x{202F}\x{205F}\x{3000}]+/u',
 290                         '_',
 291                         $dbkey
 292                 );
 293                 $dbkey = trim( $dbkey, '_' );
 294
 295                 if ( strpos( $dbkey, UtfNormal\Constants::UTF8_REPLACEMENT ) !== false ) {
 296                         # Contained illegal UTF-8 sequences or forbidden Unicode chars.
 297                         throw new MalformedTitleException( 'title-invalid-utf8', $text );
 298                 }
 299
 300                 $parts['dbkey'] = $dbkey;
 301
 302                 # Initial colon indicates main namespace rather than specified default
 303                 # but should not create invalid {ns,title} pairs such as {0,Project:Foo}
 304                 if ( $dbkey !== '' && $dbkey[0] == ':' ) {
 305                         $parts['namespace'] = NS_MAIN;
 306                         $dbkey = substr( $dbkey, 1 ); # remove the colon but continue processing
 307                         $dbkey = trim( $dbkey, '_' ); # remove any subsequent whitespace
 308                 }
 309
 310                 if ( $dbkey == '' ) {
 311                         throw new MalformedTitleException( 'title-invalid-empty', $text );
 312                 }
 313
 314                 # Namespace or interwiki prefix
 315                 $prefixRegexp = "/^(.+?)_*:_*(.*)$/S";
 316                 do {
 317                         $m = [];
 318                         if ( preg_match( $prefixRegexp, $dbkey, $m ) ) {
 319                                 $p = $m[1];
 320                                 $ns = $this->language->getNsIndex( $p );
 321                                 if ( $ns !== false ) {
 322                                         # Ordinary namespace
 323                                         $dbkey = $m[2];
 324                                         $parts['namespace'] = $ns;
 325                                         # For Talk:X pages, check if X has a "namespace" prefix
 326                                         if ( $ns == NS_TALK && preg_match( $prefixRegexp, $dbkey, $x ) ) {
 327                                                 if ( $this->language->getNsIndex( $x[1] ) ) {
 328                                                         # Disallow Talk:File:x type titles...
 329                                                         throw new MalformedTitleException( 'title-invalid-talk-namespace', $text );
 330                                                 } elseif ( $this->interwikiLookup->isValidInterwiki( $x[1] ) ) {
 331                                                         # Disallow Talk:Interwiki:x type titles...
 332                                                         throw new MalformedTitleException( 'title-invalid-talk-namespace', $text );
 333                                                 }
 334                                         }
 335                                 } elseif ( $this->interwikiLookup->isValidInterwiki( $p ) ) {
 336                                         # Interwiki link
 337                                         $dbkey = $m[2];
 338                                         $parts['interwiki'] = $this->language->lc( $p );
 339
 340                                         # Redundant interwiki prefix to the local wiki
 341                                         foreach ( $this->localInterwikis as $localIW ) {
 342                                                 if ( 0 == strcasecmp( $parts['interwiki'], $localIW ) ) {
 343                                                         if ( $dbkey == '' ) {
 344                                                                 # Empty self-links should point to the Main Page, to ensure
 345                                                                 # compatibility with cross-wiki transclusions and the like.
 346                                                                 $mainPage = Title::newMainPage();
 347                                                                 return [
 348                                                                         'interwiki' => $mainPage->getInterwiki(),
 349                                                                         'local_interwiki' => true,
 350                                                                         'fragment' => $mainPage->getFragment(),
 351                                                                         'namespace' => $mainPage->getNamespace(),
 352                                                                         'dbkey' => $mainPage->getDBkey(),
 353                                                                         'user_case_dbkey' => $mainPage->getUserCaseDBKey()
 354                                                                 ];
 355                                                         }
 356                                                         $parts['interwiki'] = '';
 357                                                         # local interwikis should behave like initial-colon links
 358                                                         $parts['local_interwiki'] = true;
 359
 360                                                         # Do another namespace split...
 361                                                         continue 2;
 362                                                 }
 363                                         }
 364
 365                                         # If there's an initial colon after the interwiki, that also
 366                                         # resets the default namespace
 367                                         if ( $dbkey !== '' && $dbkey[0] == ':' ) {
 368                                                 $parts['namespace'] = NS_MAIN;
 369                                                 $dbkey = substr( $dbkey, 1 );
 370                                                 $dbkey = trim( $dbkey, '_' );
 371                                         }
 372                                 }
 373                                 # If there's no recognized interwiki or namespace,
 374                                 # then let the colon expression be part of the title.
 375                         }
 376                         break;
 377                 } while ( true );
 378
 379                 $fragment = strstr( $dbkey, '#' );
 380                 if ( false !== $fragment ) {
 381                         $parts['fragment'] = str_replace( '_', ' ', substr( $fragment, 1 ) );
 382                         $dbkey = substr( $dbkey, 0, strlen( $dbkey ) - strlen( $fragment ) );
 383                         # remove whitespace again: prevents "Foo_bar_#"
 384                         # becoming "Foo_bar_"
 385                         $dbkey = preg_replace( '/_*$/', '', $dbkey );
 386                 }
 387
 388                 # Reject illegal characters.
 389                 $rxTc = self::getTitleInvalidRegex();
 390                 $matches = [];
 391                 if ( preg_match( $rxTc, $dbkey, $matches ) ) {
 392                         throw new MalformedTitleException( 'title-invalid-characters', $text, [ $matches[0] ] );
 393                 }
 394
 395                 # Pages with "/./" or "/../" appearing in the URLs will often be un-
 396                 # reachable due to the way web browsers deal with 'relative' URLs.
 397                 # Also, they conflict with subpage syntax.  Forbid them explicitly.
 398                 if (
 399                         strpos( $dbkey, '.' ) !== false &&
 400                         (
 401                                 $dbkey === '.' || $dbkey === '..' ||
 402                                 strpos( $dbkey, './' ) === 0 ||
 403                                 strpos( $dbkey, '../' ) === 0 ||
 404                                 strpos( $dbkey, '/./' ) !== false ||
 405                                 strpos( $dbkey, '/../' ) !== false ||
 406                                 substr( $dbkey, -2 ) == '/.' ||
 407                                 substr( $dbkey, -3 ) == '/..'
 408                         )
 409                 ) {
 410                         throw new MalformedTitleException( 'title-invalid-relative', $text );
 411                 }
 412
 413                 # Magic tilde sequences? Nu-uh!
 414                 if ( strpos( $dbkey, '~~~' ) !== false ) {
 415                         throw new MalformedTitleException( 'title-invalid-magic-tilde', $text );
 416                 }
 417
 418                 # Limit the size of titles to 255 bytes. This is typically the size of the
 419                 # underlying database field. We make an exception for special pages, which
 420                 # don't need to be stored in the database, and may edge over 255 bytes due
 421                 # to subpage syntax for long titles, e.g. [[Special:Block/Long name]]
 422                 $maxLength = ( $parts['namespace'] != NS_SPECIAL ) ? 255 : 512;
 423                 if ( strlen( $dbkey ) > $maxLength ) {
 424                         throw new MalformedTitleException( 'title-invalid-too-long', $text,
 425                                 [ Message::numParam( $maxLength ) ] );
 426                 }
 427
 428                 # Normally, all wiki links are forced to have an initial capital letter so [[foo]]
 429                 # and [[Foo]] point to the same place.  Don't force it for interwikis, since the
 430                 # other site might be case-sensitive.
 431                 $parts['user_case_dbkey'] = $dbkey;
 432                 if ( $parts['interwiki'] === '' ) {
 433                         $dbkey = Title::capitalize( $dbkey, $parts['namespace'] );
 434                 }
 435
 436                 # Can't make a link to a namespace alone... "empty" local links can only be
 437                 # self-links with a fragment identifier.
 438                 if ( $dbkey == '' && $parts['interwiki'] === '' ) {
 439                         if ( $parts['namespace'] != NS_MAIN ) {
 440                                 throw new MalformedTitleException( 'title-invalid-empty', $text );
 441                         }
 442                 }
 443
 444                 // Allow IPv6 usernames to start with '::' by canonicalizing IPv6 titles.
 445                 // IP names are not allowed for accounts, and can only be referring to
 446                 // edits from the IP. Given '::' abbreviations and caps/lowercaps,
 447                 // there are numerous ways to present the same IP. Having sp:contribs scan
 448                 // them all is silly and having some show the edits and others not is
 449                 // inconsistent. Same for talk/userpages. Keep them normalized instead.
 450                 if ( $parts['namespace'] == NS_USER || $parts['namespace'] == NS_USER_TALK ) {
 451                         $dbkey = IP::sanitizeIP( $dbkey );
 452                 }
 453
 454                 // Any remaining initial :s are illegal.
 455                 if ( $dbkey !== '' && ':' == $dbkey[0] ) {
 456                         throw new MalformedTitleException( 'title-invalid-leading-colon', $text );
 457                 }
 458
 459                 # Fill fields
 460                 $parts['dbkey'] = $dbkey;
 461
 462                 return $parts;
 463         }
 464
 465         /**
 466          * Returns a simple regex that will match on characters and sequences invalid in titles.
 467          * Note that this doesn't pick up many things that could be wrong with titles, but that
 468          * replacing this regex with something valid will make many titles valid.
 469          * Previously Title::getTitleInvalidRegex()
 470          *
 471          * @return string Regex string
 472          * @since 1.25
 473          */
 474         public static function getTitleInvalidRegex() {
 475                 static $rxTc = false;
 476                 if ( !$rxTc ) {
 477                         # Matching titles will be held as illegal.
 478                         $rxTc = '/' .
 479                                 # Any character not allowed is forbidden...
 480                                 '[^' . Title::legalChars() . ']' .
 481                                 # URL percent encoding sequences interfere with the ability
 482                                 # to round-trip titles -- you can't link to them consistently.
 483                                 '|%[0-9A-Fa-f]{2}' .
 484                                 # XML/HTML character references produce similar issues.
 485                                 '|&[A-Za-z0-9\x80-\xff]+;' .
 486                                 '|&#[0-9]+;' .
 487                                 '|&#x[0-9A-Fa-f]+;' .
 488                                 '/S';
 489                 }
 490
 491                 return $rxTc;
 492         }
 493 }