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