includes/title/MediaWikiTitleCodec.php

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