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  * @author Daniel Kinzler
  22  */
  23 use MediaWiki\Interwiki\InterwikiLookup;
  24 use MediaWiki\MediaWikiServices;
  25 use MediaWiki\Linker\LinkTarget;
  26
  27 /**
  28  * A codec for MediaWiki page titles.
  29  *
  30  * @note Normalization and validation is applied while parsing, not when formatting.
  31  * It's possible to construct a TitleValue with an invalid title, and use MediaWikiTitleCodec
  32  * to generate an (invalid) title string from it. TitleValues should be constructed only
  33  * via parseTitle() or from a (semi)trusted source, such as the database.
  34  *
  35  * @see https://www.mediawiki.org/wiki/Requests_for_comment/TitleValue
  36  * @since 1.23
  37  */
  38 class MediaWikiTitleCodec implements TitleFormatter, TitleParser {
  39         /**
  40          * @var Language
  41          */
  42         protected $language;
  43
  44         /**
  45          * @var GenderCache
  46          */
  47         protected $genderCache;
  48
  49         /**
  50          * @var string[]
  51          */
  52         protected $localInterwikis;
  53
  54         /**
  55          * @var InterwikiLookup
  56          */
  57         protected $interwikiLookup;
  58
  59         /**
  60          * @var NamespaceInfo
  61          */
  62         protected $nsInfo;
  63
  64         /**
  65          * @param Language $language The language object to use for localizing namespace names.
  66          * @param GenderCache $genderCache The gender cache for generating gendered namespace names
  67          * @param string[]|string $localInterwikis
  68          * @param InterwikiLookup|null $interwikiLookup
  69          * @param NamespaceInfo|null $nsInfo
  70          */
  71         public function __construct( Language $language, GenderCache $genderCache,
  72                 $localInterwikis = [], InterwikiLookup $interwikiLookup = null,
  73                 NamespaceInfo $nsInfo = null
  74         ) {
  75                 if ( !$interwikiLookup ) {
  76                         wfDeprecated( __METHOD__ . ' with no InterwikiLookup argument', '1.34' );
  77                         $interwikiLookup = MediaWikiServices::getInstance()->getInterwikiLookup();
  78                 }
  79                 if ( !$nsInfo ) {
  80                         wfDeprecated( __METHOD__ . ' with no NamespaceInfo argument', '1.34' );
  81                         $nsInfo = MediaWikiServices::getInstance()->getNamespaceInfo();
  82                 }
  83                 $this->language = $language;
  84                 $this->genderCache = $genderCache;
  85                 $this->localInterwikis = (array)$localInterwikis;
  86                 $this->interwikiLookup = $interwikiLookup;
  87                 $this->nsInfo = $nsInfo;
  88         }
  89
  90         /**
  91          * @see TitleFormatter::getNamespaceName()
  92          *
  93          * @param int $namespace
  94          * @param string $text
  95          *
  96          * @throws InvalidArgumentException If the namespace is invalid
  97          * @return string Namespace name with underscores (not spaces)
  98          */
  99         public function getNamespaceName( $namespace, $text ) {
 100                 if ( $this->language->needsGenderDistinction() &&
 101                         $this->nsInfo->hasGenderDistinction( $namespace )
 102                 ) {
 103                         // NOTE: we are assuming here that the title text is a user name!
 104                         $gender = $this->genderCache->getGenderOf( $text, __METHOD__ );
 105                         $name = $this->language->getGenderNsText( $namespace, $gender );
 106                 } else {
 107                         $name = $this->language->getNsText( $namespace );
 108                 }
 109
 110                 if ( $name === false ) {
 111                         throw new InvalidArgumentException( 'Unknown namespace ID: ' . $namespace );
 112                 }
 113
 114                 return $name;
 115         }
 116
 117         /**
 118          * @see TitleFormatter::formatTitle()
 119          *
 120          * @param int|bool $namespace The namespace ID (or false, if the namespace should be ignored)
 121          * @param string $text The page title. Should be valid. Only minimal normalization is applied.
 122          *        Underscores will be replaced.
 123          * @param string $fragment The fragment name (may be empty).
 124          * @param string $interwiki The interwiki name (may be empty).
 125          *
 126          * @throws InvalidArgumentException If the namespace is invalid
 127          * @return string
 128          */
 129         public function formatTitle( $namespace, $text, $fragment = '', $interwiki = '' ) {
 130                 $out = '';
 131                 if ( $interwiki !== '' ) {
 132                         $out = $interwiki . ':';
 133                 }
 134
 135                 if ( $namespace != 0 ) {
 136                         try {
 137                                 $nsName = $this->getNamespaceName( $namespace, $text );
 138                         } catch ( InvalidArgumentException $e ) {
 139                                 // See T165149. Awkward, but better than erroneously linking to the main namespace.
 140                                 $nsName = $this->language->getNsText( NS_SPECIAL ) . ":Badtitle/NS{$namespace}";
 141                         }
 142
 143                         $out .= $nsName . ':';
 144                 }
 145                 $out .= $text;
 146
 147                 if ( $fragment !== '' ) {
 148                         $out .= '#' . $fragment;
 149                 }
 150
 151                 $out = str_replace( '_', ' ', $out );
 152
 153                 return $out;
 154         }
 155
 156         /**
 157          * Parses the given text and constructs a TitleValue. Normalization
 158          * is applied according to the rules appropriate for the form specified by $form.
 159          *
 160          * @param string $text The text to parse
 161          * @param int $defaultNamespace Namespace to assume per default (usually NS_MAIN)
 162          *
 163          * @throws MalformedTitleException
 164          * @return TitleValue
 165          */
 166         public function parseTitle( $text, $defaultNamespace = NS_MAIN ) {
 167                 // Convert things like &eacute; &#257; or &#x3017; into normalized (T16952) text
 168                 $filteredText = Sanitizer::decodeCharReferencesAndNormalize( $text );
 169
 170                 // NOTE: this is an ugly cludge that allows this class to share the
 171                 // code for parsing with the old Title class. The parser code should
 172                 // be refactored to avoid this.
 173                 $parts = $this->splitTitleString( $filteredText, $defaultNamespace );
 174
 175                 // Fragment-only is okay, but only with no namespace
 176                 if ( $parts['dbkey'] === '' &&
 177                 ( $parts['fragment'] === '' || $parts['namespace'] !== NS_MAIN ) ) {
 178                         throw new MalformedTitleException( 'title-invalid-empty', $text );
 179                 }
 180
 181                 return new TitleValue(
 182                         $parts['namespace'],
 183                         $parts['dbkey'],
 184                         $parts['fragment'],
 185                         $parts['interwiki']
 186                 );
 187         }
 188
 189         /**
 190          * @see TitleFormatter::getText()
 191          *
 192          * @param LinkTarget $title
 193          *
 194          * @return string $title->getText()
 195          */
 196         public function getText( LinkTarget $title ) {
 197                 return $title->getText();
 198         }
 199
 200         /**
 201          * @see TitleFormatter::getText()
 202          *
 203          * @param LinkTarget $title
 204          *
 205          * @return string
 206          */
 207         public function getPrefixedText( LinkTarget $title ) {
 208                 if ( !isset( $title->prefixedText ) ) {
 209                         $title->prefixedText = $this->formatTitle(
 210                                 $title->getNamespace(),
 211                                 $title->getText(),
 212                                 '',
 213                                 $title->getInterwiki()
 214                         );
 215                 }
 216
 217                 return $title->prefixedText;
 218         }
 219
 220         /**
 221          * @since 1.27
 222          * @see TitleFormatter::getPrefixedDBkey()
 223          * @param LinkTarget $target
 224          * @return string
 225          */
 226         public function getPrefixedDBkey( LinkTarget $target ) {
 227                 return strtr( $this->formatTitle(
 228                         $target->getNamespace(),
 229                         $target->getDBkey(),
 230                         '',
 231                         $target->getInterwiki()
 232                 ), ' ', '_' );
 233         }
 234
 235         /**
 236          * @see TitleFormatter::getText()
 237          *
 238          * @param LinkTarget $title
 239          *
 240          * @return string
 241          */
 242         public function getFullText( LinkTarget $title ) {
 243                 return $this->formatTitle(
 244                         $title->getNamespace(),
 245                         $title->getText(),
 246                         $title->getFragment(),
 247                         $title->getInterwiki()
 248                 );
 249         }
 250
 251         /**
 252          * Normalizes and splits a title string.
 253          *
 254          * This function removes illegal characters, splits off the interwiki and
 255          * namespace prefixes, sets the other forms, and canonicalizes
 256          * everything.
 257          *
 258          * @todo this method is only exposed as a temporary measure to ease refactoring.
 259          * It was copied with minimal changes from Title::secureAndSplit().
 260          *
 261          * @todo This method should be split up and an appropriate interface
 262          * defined for use by the Title class.
 263          *
 264          * @param string $text
 265          * @param int $defaultNamespace
 266          *
 267          * @throws MalformedTitleException If $text is not a valid title string.
 268          * @return array A map with the fields 'interwiki', 'fragment', 'namespace',
 269          *         'user_case_dbkey', and 'dbkey'.
 270          */
 271         public function splitTitleString( $text, $defaultNamespace = NS_MAIN ) {
 272                 $dbkey = str_replace( ' ', '_', $text );
 273
 274                 # Initialisation
 275                 $parts = [
 276                         'interwiki' => '',
 277                         'local_interwiki' => false,
 278                         'fragment' => '',
 279                         'namespace' => $defaultNamespace,
 280                         'dbkey' => $dbkey,
 281                         'user_case_dbkey' => $dbkey,
 282                 ];
 283
 284                 # Strip Unicode bidi override characters.
 285                 # Sometimes they slip into cut-n-pasted page titles, where the
 286                 # override chars get included in list displays.
 287                 $dbkey = preg_replace( '/[\x{200E}\x{200F}\x{202A}-\x{202E}]+/u', '', $dbkey );
 288
 289                 # Clean up whitespace
 290                 # Note: use of the /u option on preg_replace here will cause
 291                 # input with invalid UTF-8 sequences to be nullified out in PHP 5.2.x,
 292                 # conveniently disabling them.
 293                 $dbkey = preg_replace(
 294                         '/[ _\xA0\x{1680}\x{180E}\x{2000}-\x{200A}\x{2028}\x{2029}\x{202F}\x{205F}\x{3000}]+/u',
 295                         '_',
 296                         $dbkey
 297                 );
 298                 $dbkey = trim( $dbkey, '_' );
 299
 300                 if ( strpos( $dbkey, UtfNormal\Constants::UTF8_REPLACEMENT ) !== false ) {
 301                         # Contained illegal UTF-8 sequences or forbidden Unicode chars.
 302                         throw new MalformedTitleException( 'title-invalid-utf8', $text );
 303                 }
 304
 305                 $parts['dbkey'] = $dbkey;
 306
 307                 # Initial colon indicates main namespace rather than specified default
 308                 # but should not create invalid {ns,title} pairs such as {0,Project:Foo}
 309                 if ( $dbkey !== '' && $dbkey[0] == ':' ) {
 310                         $parts['namespace'] = NS_MAIN;
 311                         $dbkey = substr( $dbkey, 1 ); # remove the colon but continue processing
 312                         $dbkey = trim( $dbkey, '_' ); # remove any subsequent whitespace
 313                 }
 314
 315                 if ( $dbkey == '' ) {
 316                         throw new MalformedTitleException( 'title-invalid-empty', $text );
 317                 }
 318
 319                 # Namespace or interwiki prefix
 320                 $prefixRegexp = "/^(.+?)_*:_*(.*)$/S";
 321                 do {
 322                         $m = [];
 323                         if ( preg_match( $prefixRegexp, $dbkey, $m ) ) {
 324                                 $p = $m[1];
 325                                 $ns = $this->language->getNsIndex( $p );
 326                                 if ( $ns !== false ) {
 327                                         # Ordinary namespace
 328                                         $dbkey = $m[2];
 329                                         $parts['namespace'] = $ns;
 330                                         # For Talk:X pages, check if X has a "namespace" prefix
 331                                         if ( $ns == NS_TALK && preg_match( $prefixRegexp, $dbkey, $x ) ) {
 332                                                 if ( $this->language->getNsIndex( $x[1] ) ) {
 333                                                         # Disallow Talk:File:x type titles...
 334                                                         throw new MalformedTitleException( 'title-invalid-talk-namespace', $text );
 335                                                 } elseif ( $this->interwikiLookup->isValidInterwiki( $x[1] ) ) {
 336                                                         # Disallow Talk:Interwiki:x type titles...
 337                                                         throw new MalformedTitleException( 'title-invalid-talk-namespace', $text );
 338                                                 }
 339                                         }
 340                                 } elseif ( $this->interwikiLookup->isValidInterwiki( $p ) ) {
 341                                         # Interwiki link
 342                                         $dbkey = $m[2];
 343                                         $parts['interwiki'] = $this->language->lc( $p );
 344
 345                                         # Redundant interwiki prefix to the local wiki
 346                                         foreach ( $this->localInterwikis as $localIW ) {
 347                                                 if ( strcasecmp( $parts['interwiki'], $localIW ) == 0 ) {
 348                                                         if ( $dbkey == '' ) {
 349                                                                 # Empty self-links should point to the Main Page, to ensure
 350                                                                 # compatibility with cross-wiki transclusions and the like.
 351                                                                 $mainPage = Title::newMainPage();
 352                                                                 return [
 353                                                                         'interwiki' => $mainPage->getInterwiki(),
 354                                                                         'local_interwiki' => true,
 355                                                                         'fragment' => $mainPage->getFragment(),
 356                                                                         'namespace' => $mainPage->getNamespace(),
 357                                                                         'dbkey' => $mainPage->getDBkey(),
 358                                                                         'user_case_dbkey' => $mainPage->getUserCaseDBKey()
 359                                                                 ];
 360                                                         }
 361                                                         $parts['interwiki'] = '';
 362                                                         # local interwikis should behave like initial-colon links
 363                                                         $parts['local_interwiki'] = true;
 364
 365                                                         # Do another namespace split...
 366                                                         continue 2;
 367                                                 }
 368                                         }
 369
 370                                         # If there's an initial colon after the interwiki, that also
 371                                         # resets the default namespace
 372                                         if ( $dbkey !== '' && $dbkey[0] == ':' ) {
 373                                                 $parts['namespace'] = NS_MAIN;
 374                                                 $dbkey = substr( $dbkey, 1 );
 375                                                 $dbkey = trim( $dbkey, '_' );
 376                                         }
 377                                 }
 378                                 # If there's no recognized interwiki or namespace,
 379                                 # then let the colon expression be part of the title.
 380                         }
 381                         break;
 382                 } while ( true );
 383
 384                 $fragment = strstr( $dbkey, '#' );
 385                 if ( $fragment !== false ) {
 386                         $parts['fragment'] = str_replace( '_', ' ', substr( $fragment, 1 ) );
 387                         $dbkey = substr( $dbkey, 0, strlen( $dbkey ) - strlen( $fragment ) );
 388                         # remove whitespace again: prevents "Foo_bar_#"
 389                         # becoming "Foo_bar_"
 390                         $dbkey = preg_replace( '/_*$/', '', $dbkey );
 391                 }
 392
 393                 # Reject illegal characters.
 394                 $rxTc = self::getTitleInvalidRegex();
 395                 $matches = [];
 396                 if ( preg_match( $rxTc, $dbkey, $matches ) ) {
 397                         throw new MalformedTitleException( 'title-invalid-characters', $text, [ $matches[0] ] );
 398                 }
 399
 400                 # Pages with "/./" or "/../" appearing in the URLs will often be un-
 401                 # reachable due to the way web browsers deal with 'relative' URLs.
 402                 # Also, they conflict with subpage syntax.  Forbid them explicitly.
 403                 if (
 404                         strpos( $dbkey, '.' ) !== false &&
 405                         (
 406                                 $dbkey === '.' || $dbkey === '..' ||
 407                                 strpos( $dbkey, './' ) === 0 ||
 408                                 strpos( $dbkey, '../' ) === 0 ||
 409                                 strpos( $dbkey, '/./' ) !== false ||
 410                                 strpos( $dbkey, '/../' ) !== false ||
 411                                 substr( $dbkey, -2 ) == '/.' ||
 412                                 substr( $dbkey, -3 ) == '/..'
 413                         )
 414                 ) {
 415                         throw new MalformedTitleException( 'title-invalid-relative', $text );
 416                 }
 417
 418                 # Magic tilde sequences? Nu-uh!
 419                 if ( strpos( $dbkey, '~~~' ) !== false ) {
 420                         throw new MalformedTitleException( 'title-invalid-magic-tilde', $text );
 421                 }
 422
 423                 # Limit the size of titles to 255 bytes. This is typically the size of the
 424                 # underlying database field. We make an exception for special pages, which
 425                 # don't need to be stored in the database, and may edge over 255 bytes due
 426                 # to subpage syntax for long titles, e.g. [[Special:Block/Long name]]
 427                 $maxLength = ( $parts['namespace'] != NS_SPECIAL ) ? 255 : 512;
 428                 if ( strlen( $dbkey ) > $maxLength ) {
 429                         throw new MalformedTitleException( 'title-invalid-too-long', $text,
 430                                 [ Message::numParam( $maxLength ) ] );
 431                 }
 432
 433                 # Normally, all wiki links are forced to have an initial capital letter so [[foo]]
 434                 # and [[Foo]] point to the same place.  Don't force it for interwikis, since the
 435                 # other site might be case-sensitive.
 436                 $parts['user_case_dbkey'] = $dbkey;
 437                 if ( $parts['interwiki'] === '' ) {
 438                         $dbkey = Title::capitalize( $dbkey, $parts['namespace'] );
 439                 }
 440
 441                 # Can't make a link to a namespace alone... "empty" local links can only be
 442                 # self-links with a fragment identifier.
 443                 if ( $dbkey == '' && $parts['interwiki'] === '' && $parts['namespace'] != NS_MAIN ) {
 444                         throw new MalformedTitleException( 'title-invalid-empty', $text );
 445                 }
 446
 447                 // Allow IPv6 usernames to start with '::' by canonicalizing IPv6 titles.
 448                 // IP names are not allowed for accounts, and can only be referring to
 449                 // edits from the IP. Given '::' abbreviations and caps/lowercaps,
 450                 // there are numerous ways to present the same IP. Having sp:contribs scan
 451                 // them all is silly and having some show the edits and others not is
 452                 // inconsistent. Same for talk/userpages. Keep them normalized instead.
 453                 if ( $parts['namespace'] == NS_USER || $parts['namespace'] == NS_USER_TALK ) {
 454                         $dbkey = IP::sanitizeIP( $dbkey );
 455                 }
 456
 457                 // Any remaining initial :s are illegal.
 458                 if ( $dbkey !== '' && $dbkey[0] == ':' ) {
 459                         throw new MalformedTitleException( 'title-invalid-leading-colon', $text );
 460                 }
 461
 462                 # Fill fields
 463                 $parts['dbkey'] = $dbkey;
 464
 465                 return $parts;
 466         }
 467
 468         /**
 469          * Returns a simple regex that will match on characters and sequences invalid in titles.
 470          * Note that this doesn't pick up many things that could be wrong with titles, but that
 471          * replacing this regex with something valid will make many titles valid.
 472          * Previously Title::getTitleInvalidRegex()
 473          *
 474          * @return string Regex string
 475          * @since 1.25
 476          */
 477         public static function getTitleInvalidRegex() {
 478                 static $rxTc = false;
 479                 if ( !$rxTc ) {
 480                         # Matching titles will be held as illegal.
 481                         $rxTc = '/' .
 482                                 # Any character not allowed is forbidden...
 483                                 '[^' . Title::legalChars() . ']' .
 484                                 # URL percent encoding sequences interfere with the ability
 485                                 # to round-trip titles -- you can't link to them consistently.
 486                                 '|%[0-9A-Fa-f]{2}' .
 487                                 # XML/HTML character references produce similar issues.
 488                                 '|&[A-Za-z0-9\x80-\xff]+;' .
 489                                 '|&#[0-9]+;' .
 490                                 '|&#x[0-9A-Fa-f]+;' .
 491                                 '/S';
 492                 }
 493
 494                 return $rxTc;
 495         }
 496 }