3 * A codec for %MediaWiki page titles.
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.
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.
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
22 * @author Daniel Kinzler
26 * A codec for %MediaWiki page titles.
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.
33 * @see https://www.mediawiki.org/wiki/Requests_for_comment/TitleValue
35 class MediaWikiTitleCodec
implements TitleFormatter
, TitleParser
{
45 protected $genderCache;
50 protected $localInterwikis;
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
57 public function __construct( Language
$language, GenderCache
$genderCache, $localInterwikis = array() ) {
58 $this->language
= $language;
59 $this->genderCache
= $genderCache;
60 $this->localInterwikis
= (array)$localInterwikis;
64 * @see TitleFormatter::getNamespaceName()
66 * @param int $namespace
69 * @throws InvalidArgumentException if the namespace is invalid
72 public function getNamespaceName( $namespace, $text ) {
73 if ( $this->language
->needsGenderDistinction() &&
74 MWNamespace
::hasGenderDistinction( $namespace ) ) {
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 );
80 $name = $this->language
->getNsText( $namespace );
83 if ( $name === false ) {
84 throw new InvalidArgumentException( 'Unknown namespace ID: ' . $namespace );
91 * @see TitleFormatter::formatTitle()
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).
98 * @throws InvalidArgumentException if the namespace is invalid
101 public function formatTitle( $namespace, $text, $fragment = '' ) {
102 if ( $namespace !== false ) {
103 $namespace = $this->getNamespaceName( $namespace, $text );
105 if ( $namespace !== '' ) {
106 $text = $namespace . ':' . $text;
110 if ( $fragment !== '' ) {
111 $text = $text . '#' . $fragment;
114 $text = str_replace( '_', ' ', $text );
120 * Parses the given text and constructs a TitleValue. Normalization
121 * is applied according to the rules appropriate for the form specified by $form.
123 * @param string $text The text to parse
124 * @param int $defaultNamespace Namespace to assume per default (usually NS_MAIN)
126 * @throws MalformedTitleException
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 );
135 // Interwiki links are not supported by TitleValue
136 if ( $parts['interwiki'] !== '' ) {
137 throw new MalformedTitleException( 'Title must not contain an interwiki prefix: ' . $text );
140 // Relative fragment links are not supported by TitleValue
141 if ( $parts['dbkey'] === '' ) {
142 throw new MalformedTitleException( 'Title must not be empty: ' . $text );
145 return new TitleValue( $parts['namespace'], $parts['dbkey'], $parts['fragment'] );
149 * @see TitleFormatter::getText()
151 * @param TitleValue $title
153 * @return string $title->getText()
155 public function getText( TitleValue
$title ) {
156 return $this->formatTitle( false, $title->getText(), '' );
160 * @see TitleFormatter::getText()
162 * @param TitleValue $title
166 public function getPrefixedText( TitleValue
$title ) {
167 return $this->formatTitle( $title->getNamespace(), $title->getText(), '' );
171 * @see TitleFormatter::getText()
173 * @param TitleValue $title
177 public function getFullText( TitleValue
$title ) {
178 return $this->formatTitle( $title->getNamespace(), $title->getText(), $title->getFragment() );
182 * Normalizes and splits a title string.
184 * This function removes illegal characters, splits off the interwiki and
185 * namespace prefixes, sets the other forms, and canonicalizes
188 * @todo: this method is only exposed as a temporary measure to ease refactoring.
189 * It was copied with minimal changes from Title::secureAndSplit().
191 * @todo: This method should be split up and an appropriate interface
192 * defined for use by the Title class.
194 * @param string $text
195 * @param int $defaultNamespace
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'.
201 public function splitTitleString( $text, $defaultNamespace = NS_MAIN
) {
202 $dbkey = str_replace( ' ', '_', $text );
208 'namespace' => $defaultNamespace,
210 'user_case_dbkey' => $dbkey,
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 );
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, '_' );
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 );
230 $parts['dbkey'] = $dbkey;
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
240 if ( $dbkey == '' ) {
241 throw new MalformedTitleException( 'Empty title: ' . $text );
244 # Namespace or interwiki prefix
245 $prefixRegexp = "/^(.+?)_*:_*(.*)$/S";
248 if ( preg_match( $prefixRegexp, $dbkey, $m ) ) {
250 if ( ( $ns = $this->language
->getNsIndex( $p ) ) !== false ) {
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 );
265 } elseif ( Interwiki
::isValidInterwiki( $p ) ) {
268 $parts['interwiki'] = $this->language
->lc( $p );
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 # Empty self-links should point to the Main Page, to ensure
275 # compatibility with cross-wiki transclusions and the like.
276 $mainPage = Title
::newMainPage();
278 'interwiki' => $mainPage->getInterwiki(),
279 'fragment' => $mainPage->getFragment(),
280 'namespace' => $mainPage->getNamespace(),
281 'dbkey' => $mainPage->getDBkey(),
282 'user_case_dbkey' => $mainPage->getUserCaseDBKey()
285 $parts['interwiki'] = '';
287 # Do another namespace split...
292 # If there's an initial colon after the interwiki, that also
293 # resets the default namespace
294 if ( $dbkey !== '' && $dbkey[0] == ':' ) {
295 $parts['namespace'] = NS_MAIN
;
296 $dbkey = substr( $dbkey, 1 );
299 # If there's no recognized interwiki or namespace,
300 # then let the colon expression be part of the title.
305 $fragment = strstr( $dbkey, '#' );
306 if ( false !== $fragment ) {
307 $parts['fragment'] = str_replace( '_', ' ', substr( $fragment, 1 ) );
308 $dbkey = substr( $dbkey, 0, strlen( $dbkey ) - strlen( $fragment ) );
309 # remove whitespace again: prevents "Foo_bar_#"
310 # becoming "Foo_bar_"
311 $dbkey = preg_replace( '/_*$/', '', $dbkey );
314 # Reject illegal characters.
315 $rxTc = Title
::getTitleInvalidRegex();
316 if ( preg_match( $rxTc, $dbkey ) ) {
317 throw new MalformedTitleException( 'Illegal characters found in title: ' . $text );
320 # Pages with "/./" or "/../" appearing in the URLs will often be un-
321 # reachable due to the way web browsers deal with 'relative' URLs.
322 # Also, they conflict with subpage syntax. Forbid them explicitly.
324 strpos( $dbkey, '.' ) !== false &&
326 $dbkey === '.' ||
$dbkey === '..' ||
327 strpos( $dbkey, './' ) === 0 ||
328 strpos( $dbkey, '../' ) === 0 ||
329 strpos( $dbkey, '/./' ) !== false ||
330 strpos( $dbkey, '/../' ) !== false ||
331 substr( $dbkey, -2 ) == '/.' ||
332 substr( $dbkey, -3 ) == '/..'
335 throw new MalformedTitleException( 'Bad title: ' . $text );
338 # Magic tilde sequences? Nu-uh!
339 if ( strpos( $dbkey, '~~~' ) !== false ) {
340 throw new MalformedTitleException( 'Bad title: ' . $text );
343 # Limit the size of titles to 255 bytes. This is typically the size of the
344 # underlying database field. We make an exception for special pages, which
345 # don't need to be stored in the database, and may edge over 255 bytes due
346 # to subpage syntax for long titles, e.g. [[Special:Block/Long name]]
348 ( $parts['namespace'] != NS_SPECIAL
&& strlen( $dbkey ) > 255 )
349 ||
strlen( $dbkey ) > 512
351 throw new MalformedTitleException( 'Title too long: ' . substr( $dbkey, 0, 255 ) . '...' );
354 # Normally, all wiki links are forced to have an initial capital letter so [[foo]]
355 # and [[Foo]] point to the same place. Don't force it for interwikis, since the
356 # other site might be case-sensitive.
357 $parts['user_case_dbkey'] = $dbkey;
358 if ( $parts['interwiki'] === '' ) {
359 $dbkey = Title
::capitalize( $dbkey, $parts['namespace'] );
362 # Can't make a link to a namespace alone... "empty" local links can only be
363 # self-links with a fragment identifier.
364 if ( $dbkey == '' && $parts['interwiki'] === '' ) {
365 if ( $parts['namespace'] != NS_MAIN
) {
366 throw new MalformedTitleException( 'Empty title: ' . $text );
370 // Allow IPv6 usernames to start with '::' by canonicalizing IPv6 titles.
371 // IP names are not allowed for accounts, and can only be referring to
372 // edits from the IP. Given '::' abbreviations and caps/lowercaps,
373 // there are numerous ways to present the same IP. Having sp:contribs scan
374 // them all is silly and having some show the edits and others not is
375 // inconsistent. Same for talk/userpages. Keep them normalized instead.
376 if ( $parts['namespace'] == NS_USER ||
$parts['namespace'] == NS_USER_TALK
) {
377 $dbkey = IP
::sanitizeIP( $dbkey );
380 // Any remaining initial :s are illegal.
381 if ( $dbkey !== '' && ':' == $dbkey[0] ) {
382 throw new MalformedTitleException( 'Title must not start with a colon: ' . $text );
386 $parts['dbkey'] = $dbkey;