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