dépôts / lhc / web / wiklou.git / blob
? search:
summary | shortlog | log | commit | commitdiff | tree
history | raw | HEAD
merge latest master.
[lhc/web/wiklou.git] / includes / parser / Parser.php
1 <?php
2 /**
3 * PHP parser that converts wiki markup to HTML.
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 * @ingroup Parser
22 */
23
24 /**
25 * @defgroup Parser Parser
26 */
27
28 /**
29 * PHP Parser - Processes wiki markup (which uses a more user-friendly
30 * syntax, such as "[[link]]" for making links), and provides a one-way
31 * transformation of that wiki markup it into XHTML output / markup
32 * (which in turn the browser understands, and can display).
33 *
34 * There are seven main entry points into the Parser class:
35 *
36 * - Parser::parse()
37 * produces HTML output
38 * - Parser::preSaveTransform().
39 * produces altered wiki markup.
40 * - Parser::preprocess()
41 * removes HTML comments and expands templates
42 * - Parser::cleanSig() and Parser::cleanSigInSig()
43 * Cleans a signature before saving it to preferences
44 * - Parser::getSection()
45 * Return the content of a section from an article for section editing
46 * - Parser::replaceSection()
47 * Replaces a section by number inside an article
48 * - Parser::getPreloadText()
49 * Removes <noinclude> sections, and <includeonly> tags.
50 *
51 * Globals used:
52 * object: $wgContLang
53 *
54 * @warning $wgUser or $wgTitle or $wgRequest or $wgLang. Keep them away!
55 *
56 * @par Settings:
57 * $wgLocaltimezone
58 * $wgNamespacesWithSubpages
59 *
60 * @par Settings only within ParserOptions:
61 * $wgAllowExternalImages
62 * $wgAllowSpecialInclusion
63 * $wgInterwikiMagic
64 * $wgMaxArticleSize
65 * $wgUseDynamicDates
66 *
67 * @ingroup Parser
68 */
69 class Parser {
70 /**
71 * Update this version number when the ParserOutput format
72 * changes in an incompatible way, so the parser cache
73 * can automatically discard old data.
74 */
75 const VERSION = '1.6.4';
76
77 /**
78 * Update this version number when the output of serialiseHalfParsedText()
79 * changes in an incompatible way
80 */
81 const HALF_PARSED_VERSION = 2;
82
83 # Flags for Parser::setFunctionHook
84 # Also available as global constants from Defines.php
85 const SFH_NO_HASH = 1;
86 const SFH_OBJECT_ARGS = 2;
87
88 # Constants needed for external link processing
89 # Everything except bracket, space, or control characters
90 # \p{Zs} is unicode 'separator, space' category. It covers the space 0x20
91 # as well as U+3000 is IDEOGRAPHIC SPACE for bug 19052
92 const EXT_LINK_URL_CLASS = '[^][<>"\\x00-\\x20\\x7F\p{Zs}]';
93 const EXT_IMAGE_REGEX = '/^(http:\/\/|https:\/\/)([^][<>"\\x00-\\x20\\x7F\p{Zs}]+)
94 \\/([A-Za-z0-9_.,~%\\-+&;#*?!=()@\\x80-\\xFF]+)\\.((?i)gif|png|jpg|jpeg)$/Sxu';
95
96 # State constants for the definition list colon extraction
97 const COLON_STATE_TEXT = 0;
98 const COLON_STATE_TAG = 1;
99 const COLON_STATE_TAGSTART = 2;
100 const COLON_STATE_CLOSETAG = 3;
101 const COLON_STATE_TAGSLASH = 4;
102 const COLON_STATE_COMMENT = 5;
103 const COLON_STATE_COMMENTDASH = 6;
104 const COLON_STATE_COMMENTDASHDASH = 7;
105
106 # Flags for preprocessToDom
107 const PTD_FOR_INCLUSION = 1;
108
109 # Allowed values for $this->mOutputType
110 # Parameter to startExternalParse().
111 const OT_HTML = 1; # like parse()
112 const OT_WIKI = 2; # like preSaveTransform()
113 const OT_PREPROCESS = 3; # like preprocess()
114 const OT_MSG = 3;
115 const OT_PLAIN = 4; # like extractSections() - portions of the original are returned unchanged.
116
117 # Marker Suffix needs to be accessible staticly.
118 const MARKER_SUFFIX = "-QINU\x7f";
119
120 # Persistent:
121 var $mTagHooks = array();
122 var $mTransparentTagHooks = array();
123 var $mFunctionHooks = array();
124 var $mFunctionSynonyms = array( 0 => array(), 1 => array() );
125 var $mFunctionTagHooks = array();
126 var $mStripList = array();
127 var $mDefaultStripList = array();
128 var $mVarCache = array();
129 var $mImageParams = array();
130 var $mImageParamsMagicArray = array();
131 var $mMarkerIndex = 0;
132 var $mFirstCall = true;
133
134 # Initialised by initialiseVariables()
135
136 /**
137 * @var MagicWordArray
138 */
139 var $mVariables;
140
141 /**
142 * @var MagicWordArray
143 */
144 var $mSubstWords;
145 var $mConf, $mPreprocessor, $mExtLinkBracketedRegex, $mUrlProtocols; # Initialised in constructor
146
147 # Cleared with clearState():
148 /**
149 * @var ParserOutput
150 */
151 var $mOutput;
152 var $mAutonumber, $mDTopen;
153
154 /**
155 * @var StripState
156 */
157 var $mStripState;
158
159 var $mIncludeCount, $mArgStack, $mLastSection, $mInPre;
160 /**
161 * @var LinkHolderArray
162 */
163 var $mLinkHolders;
164
165 var $mLinkID;
166 var $mIncludeSizes, $mPPNodeCount, $mHighestExpansionDepth, $mDefaultSort;
167 var $mTplExpandCache; # empty-frame expansion cache
168 var $mTplRedirCache, $mTplDomCache, $mHeadings, $mDoubleUnderscores;
169 var $mExpensiveFunctionCount; # number of expensive parser function calls
170 var $mShowToc, $mForceTocPosition;
171
172 /**
173 * @var User
174 */
175 var $mUser; # User object; only used when doing pre-save transform
176
177 # Temporary
178 # These are variables reset at least once per parse regardless of $clearState
179
180 /**
181 * @var ParserOptions
182 */
183 var $mOptions;
184
185 /**
186 * @var Title
187 */
188 var $mTitle; # Title context, used for self-link rendering and similar things
189 var $mOutputType; # Output type, one of the OT_xxx constants
190 var $ot; # Shortcut alias, see setOutputType()
191 var $mRevisionObject; # The revision object of the specified revision ID
192 var $mRevisionId; # ID to display in {{REVISIONID}} tags
193 var $mRevisionTimestamp; # The timestamp of the specified revision ID
194 var $mRevisionUser; # User to display in {{REVISIONUSER}} tag
195 var $mRevIdForTs; # The revision ID which was used to fetch the timestamp
196
197 /**
198 * @var string
199 */
200 var $mUniqPrefix;
201
202 /**
203 * Constructor
204 *
205 * @param $conf array
206 */
207 public function __construct( $conf = array() ) {
208 $this->mConf = $conf;
209 $this->mUrlProtocols = wfUrlProtocols();
210 $this->mExtLinkBracketedRegex = '/\[(((?i)' . $this->mUrlProtocols . ')'.
211 self::EXT_LINK_URL_CLASS.'+)\p{Zs}*([^\]\\x00-\\x08\\x0a-\\x1F]*?)\]/Su';
212 if ( isset( $conf['preprocessorClass'] ) ) {
213 $this->mPreprocessorClass = $conf['preprocessorClass'];
214 } elseif ( defined( 'MW_COMPILED' ) ) {
215 # Preprocessor_Hash is much faster than Preprocessor_DOM in compiled mode
216 $this->mPreprocessorClass = 'Preprocessor_Hash';
217 } elseif ( extension_loaded( 'domxml' ) ) {
218 # PECL extension that conflicts with the core DOM extension (bug 13770)
219 wfDebug( "Warning: you have the obsolete domxml extension for PHP. Please remove it!\n" );
220 $this->mPreprocessorClass = 'Preprocessor_Hash';
221 } elseif ( extension_loaded( 'dom' ) ) {
222 $this->mPreprocessorClass = 'Preprocessor_DOM';
223 } else {
224 $this->mPreprocessorClass = 'Preprocessor_Hash';
225 }
226 wfDebug( __CLASS__ . ": using preprocessor: {$this->mPreprocessorClass}\n" );
227 }
228
229 /**
230 * Reduce memory usage to reduce the impact of circular references
231 */
232 function __destruct() {
233 if ( isset( $this->mLinkHolders ) ) {
234 unset( $this->mLinkHolders );
235 }
236 foreach ( $this as $name => $value ) {
237 unset( $this->$name );
238 }
239 }
240
241 /**
242 * Do various kinds of initialisation on the first call of the parser
243 */
244 function firstCallInit() {
245 if ( !$this->mFirstCall ) {
246 return;
247 }
248 $this->mFirstCall = false;
249
250 wfProfileIn( __METHOD__ );
251
252 CoreParserFunctions::register( $this );
253 CoreTagHooks::register( $this );
254 $this->initialiseVariables();
255
256 wfRunHooks( 'ParserFirstCallInit', array( &$this ) );
257 wfProfileOut( __METHOD__ );
258 }
259
260 /**
261 * Clear Parser state
262 *
263 * @private
264 */
265 function clearState() {
266 wfProfileIn( __METHOD__ );
267 if ( $this->mFirstCall ) {
268 $this->firstCallInit();
269 }
270 $this->mOutput = new ParserOutput;
271 $this->mOptions->registerWatcher( array( $this->mOutput, 'recordOption' ) );
272 $this->mAutonumber = 0;
273 $this->mLastSection = '';
274 $this->mDTopen = false;
275 $this->mIncludeCount = array();
276 $this->mArgStack = false;
277 $this->mInPre = false;
278 $this->mLinkHolders = new LinkHolderArray( $this );
279 $this->mLinkID = 0;
280 $this->mRevisionObject = $this->mRevisionTimestamp =
281 $this->mRevisionId = $this->mRevisionUser = null;
282 $this->mVarCache = array();
283 $this->mUser = null;
284
285 /**
286 * Prefix for temporary replacement strings for the multipass parser.
287 * \x07 should never appear in input as it's disallowed in XML.
288 * Using it at the front also gives us a little extra robustness
289 * since it shouldn't match when butted up against identifier-like
290 * string constructs.
291 *
292 * Must not consist of all title characters, or else it will change
293 * the behaviour of <nowiki> in a link.
294 */
295 $this->mUniqPrefix = "\x7fUNIQ" . self::getRandomString();
296 $this->mStripState = new StripState( $this->mUniqPrefix );
297
298
299 # Clear these on every parse, bug 4549
300 $this->mTplExpandCache = $this->mTplRedirCache = $this->mTplDomCache = array();
301
302 $this->mShowToc = true;
303 $this->mForceTocPosition = false;
304 $this->mIncludeSizes = array(
305 'post-expand' => 0,
306 'arg' => 0,
307 );
308 $this->mPPNodeCount = 0;
309 $this->mHighestExpansionDepth = 0;
310 $this->mDefaultSort = false;
311 $this->mHeadings = array();
312 $this->mDoubleUnderscores = array();
313 $this->mExpensiveFunctionCount = 0;
314
315 # Fix cloning
316 if ( isset( $this->mPreprocessor ) && $this->mPreprocessor->parser !== $this ) {
317 $this->mPreprocessor = null;
318 }
319
320 wfRunHooks( 'ParserClearState', array( &$this ) );
321 wfProfileOut( __METHOD__ );
322 }
323
324 /**
325 * Convert wikitext to HTML
326 * Do not call this function recursively.
327 *
328 * @param $text String: text we want to parse
329 * @param $title Title object
330 * @param $options ParserOptions
331 * @param $linestart boolean
332 * @param $clearState boolean
333 * @param $revid Int: number to pass in {{REVISIONID}}
334 * @return ParserOutput a ParserOutput
335 */
336 public function parse( $text, Title $title, ParserOptions $options, $linestart = true, $clearState = true, $revid = null ) {
337 /**
338 * First pass--just handle <nowiki> sections, pass the rest off
339 * to internalParse() which does all the real work.
340 */
341
342 global $wgUseTidy, $wgAlwaysUseTidy, $wgDisableLangConversion, $wgDisableTitleConversion;
343 $fname = __METHOD__.'-' . wfGetCaller();
344 wfProfileIn( __METHOD__ );
345 wfProfileIn( $fname );
346
347 $this->startParse( $title, $options, self::OT_HTML, $clearState );
348
349 # Remove the strip marker tag prefix from the input, if present.
350 if ( $clearState ) {
351 $text = str_replace( $this->mUniqPrefix, '', $text );
352 }
353
354 $oldRevisionId = $this->mRevisionId;
355 $oldRevisionObject = $this->mRevisionObject;
356 $oldRevisionTimestamp = $this->mRevisionTimestamp;
357 $oldRevisionUser = $this->mRevisionUser;
358 if ( $revid !== null ) {
359 $this->mRevisionId = $revid;
360 $this->mRevisionObject = null;
361 $this->mRevisionTimestamp = null;
362 $this->mRevisionUser = null;
363 }
364
365 wfRunHooks( 'ParserBeforeStrip', array( &$this, &$text, &$this->mStripState ) );
366 # No more strip!
367 wfRunHooks( 'ParserAfterStrip', array( &$this, &$text, &$this->mStripState ) );
368 $text = $this->internalParse( $text );
369 wfRunHooks( 'ParserAfterParse', array( &$this, &$text, &$this->mStripState ) );
370
371 $text = $this->mStripState->unstripGeneral( $text );
372
373 # Clean up special characters, only run once, next-to-last before doBlockLevels
374 $fixtags = array(
375 # french spaces, last one Guillemet-left
376 # only if there is something before the space
377 '/(.) (?=\\?|:|;|!|%|\\302\\273)/' => '\\1&#160;',
378 # french spaces, Guillemet-right
379 '/(\\302\\253) /' => '\\1&#160;',
380 '/&#160;(!\s*important)/' => ' \\1', # Beware of CSS magic word !important, bug #11874.
381 );
382 $text = preg_replace( array_keys( $fixtags ), array_values( $fixtags ), $text );
383
384 $text = $this->doBlockLevels( $text, $linestart );
385
386 $this->replaceLinkHolders( $text );
387
388 /**
389 * The input doesn't get language converted if
390 * a) It's disabled
391 * b) Content isn't converted
392 * c) It's a conversion table
393 * d) it is an interface message (which is in the user language)
394 */
395 if ( !( $wgDisableLangConversion
396 || isset( $this->mDoubleUnderscores['nocontentconvert'] )
397 || $this->mTitle->isConversionTable() ) )
398 {
399 # Run convert unconditionally in 1.18-compatible mode
400 global $wgBug34832TransitionalRollback;
401 if ( $wgBug34832TransitionalRollback || !$this->mOptions->getInterfaceMessage() ) {
402 # The position of the convert() call should not be changed. it
403 # assumes that the links are all replaced and the only thing left
404 # is the <nowiki> mark.
405 $text = $this->getConverterLanguage()->convert( $text );
406 }
407 }
408
409 /**
410 * A converted title will be provided in the output object if title and
411 * content conversion are enabled, the article text does not contain
412 * a conversion-suppressing double-underscore tag, and no
413 * {{DISPLAYTITLE:...}} is present. DISPLAYTITLE takes precedence over
414 * automatic link conversion.
415 */
416 if ( !( $wgDisableLangConversion
417 || $wgDisableTitleConversion
418 || isset( $this->mDoubleUnderscores['nocontentconvert'] )
419 || isset( $this->mDoubleUnderscores['notitleconvert'] )
420 || $this->mOutput->getDisplayTitle() !== false ) )
421 {
422 $convruletitle = $this->getConverterLanguage()->getConvRuleTitle();
423 if ( $convruletitle ) {
424 $this->mOutput->setTitleText( $convruletitle );
425 } else {
426 $titleText = $this->getConverterLanguage()->convertTitle( $title );
427 $this->mOutput->setTitleText( $titleText );
428 }
429 }
430
431 $text = $this->mStripState->unstripNoWiki( $text );
432
433 wfRunHooks( 'ParserBeforeTidy', array( &$this, &$text ) );
434
435 $text = $this->replaceTransparentTags( $text );
436 $text = $this->mStripState->unstripGeneral( $text );
437
438 $text = Sanitizer::normalizeCharReferences( $text );
439
440 if ( ( $wgUseTidy && $this->mOptions->getTidy() ) || $wgAlwaysUseTidy ) {
441 $text = MWTidy::tidy( $text );
442 } else {
443 # attempt to sanitize at least some nesting problems
444 # (bug #2702 and quite a few others)
445 $tidyregs = array(
446 # ''Something [http://www.cool.com cool''] -->
447 # <i>Something</i><a href="http://www.cool.com"..><i>cool></i></a>
448 '/(<([bi])>)(<([bi])>)?([^<]*)(<\/?a[^<]*>)([^<]*)(<\/\\4>)?(<\/\\2>)/' =>
449 '\\1\\3\\5\\8\\9\\6\\1\\3\\7\\8\\9',
450 # fix up an anchor inside another anchor, only
451 # at least for a single single nested link (bug 3695)
452 '/(<a[^>]+>)([^<]*)(<a[^>]+>[^<]*)<\/a>(.*)<\/a>/' =>
453 '\\1\\2</a>\\3</a>\\1\\4</a>',
454 # fix div inside inline elements- doBlockLevels won't wrap a line which
455 # contains a div, so fix it up here; replace
456 # div with escaped text
457 '/(<([aib]) [^>]+>)([^<]*)(<div([^>]*)>)(.*)(<\/div>)([^<]*)(<\/\\2>)/' =>
458 '\\1\\3&lt;div\\5&gt;\\6&lt;/div&gt;\\8\\9',
459 # remove empty italic or bold tag pairs, some
460 # introduced by rules above
461 '/<([bi])><\/\\1>/' => '',
462 );
463
464 $text = preg_replace(
465 array_keys( $tidyregs ),
466 array_values( $tidyregs ),
467 $text );
468 }
469
470 if ( $this->mExpensiveFunctionCount > $this->mOptions->getExpensiveParserFunctionLimit() ) {
471 $this->limitationWarn( 'expensive-parserfunction',
472 $this->mExpensiveFunctionCount,
473 $this->mOptions->getExpensiveParserFunctionLimit()
474 );
475 }
476
477 wfRunHooks( 'ParserAfterTidy', array( &$this, &$text ) );
478
479 # Information on include size limits, for the benefit of users who try to skirt them
480 if ( $this->mOptions->getEnableLimitReport() ) {
481 $max = $this->mOptions->getMaxIncludeSize();
482 $PFreport = "Expensive parser function count: {$this->mExpensiveFunctionCount}/{$this->mOptions->getExpensiveParserFunctionLimit()}\n";
483 $limitReport =
484 "NewPP limit report\n" .
485 "Preprocessor node count: {$this->mPPNodeCount}/{$this->mOptions->getMaxPPNodeCount()}\n" .
486 "Post-expand include size: {$this->mIncludeSizes['post-expand']}/$max bytes\n" .
487 "Template argument size: {$this->mIncludeSizes['arg']}/$max bytes\n".
488 "Highest expansion depth: {$this->mHighestExpansionDepth}/{$this->mOptions->getMaxPPExpandDepth()}\n".
489 $PFreport;
490 wfRunHooks( 'ParserLimitReport', array( $this, &$limitReport ) );
491 $text .= "\n<!-- \n$limitReport-->\n";
492 }
493 $this->mOutput->setText( $text );
494
495 $this->mRevisionId = $oldRevisionId;
496 $this->mRevisionObject = $oldRevisionObject;
497 $this->mRevisionTimestamp = $oldRevisionTimestamp;
498 $this->mRevisionUser = $oldRevisionUser;
499 wfProfileOut( $fname );
500 wfProfileOut( __METHOD__ );
501
502 return $this->mOutput;
503 }
504
505 /**
506 * Recursive parser entry point that can be called from an extension tag
507 * hook.
508 *
509 * If $frame is not provided, then template variables (e.g., {{{1}}}) within $text are not expanded
510 *
511 * @param $text String: text extension wants to have parsed
512 * @param $frame PPFrame: The frame to use for expanding any template variables
513 *
514 * @return string
515 */
516 function recursiveTagParse( $text, $frame=false ) {
517 wfProfileIn( __METHOD__ );
518 wfRunHooks( 'ParserBeforeStrip', array( &$this, &$text, &$this->mStripState ) );
519 wfRunHooks( 'ParserAfterStrip', array( &$this, &$text, &$this->mStripState ) );
520 $text = $this->internalParse( $text, false, $frame );
521 wfProfileOut( __METHOD__ );
522 return $text;
523 }
524
525 /**
526 * Expand templates and variables in the text, producing valid, static wikitext.
527 * Also removes comments.
528 * @return mixed|string
529 */
530 function preprocess( $text, Title $title, ParserOptions $options, $revid = null ) {
531 wfProfileIn( __METHOD__ );
532 $this->startParse( $title, $options, self::OT_PREPROCESS, true );
533 if ( $revid !== null ) {
534 $this->mRevisionId = $revid;
535 }
536 wfRunHooks( 'ParserBeforeStrip', array( &$this, &$text, &$this->mStripState ) );
537 wfRunHooks( 'ParserAfterStrip', array( &$this, &$text, &$this->mStripState ) );
538 $text = $this->replaceVariables( $text );
539 $text = $this->mStripState->unstripBoth( $text );
540 wfProfileOut( __METHOD__ );
541 return $text;
542 }
543
544 /**
545 * Recursive parser entry point that can be called from an extension tag
546 * hook.
547 *
548 * @param $text String: text to be expanded
549 * @param $frame PPFrame: The frame to use for expanding any template variables
550 * @return String
551 * @since 1.19
552 */
553 public function recursivePreprocess( $text, $frame = false ) {
554 wfProfileIn( __METHOD__ );
555 $text = $this->replaceVariables( $text, $frame );
556 $text = $this->mStripState->unstripBoth( $text );
557 wfProfileOut( __METHOD__ );
558 return $text;
559 }
560
561 /**
562 * Process the wikitext for the "?preload=" feature. (bug 5210)
563 *
564 * "<noinclude>", "<includeonly>" etc. are parsed as for template
565 * transclusion, comments, templates, arguments, tags hooks and parser
566 * functions are untouched.
567 *
568 * @param $text String
569 * @param $title Title
570 * @param $options ParserOptions
571 * @return String
572 */
573 public function getPreloadText( $text, Title $title, ParserOptions $options ) {
574 # Parser (re)initialisation
575 $this->startParse( $title, $options, self::OT_PLAIN, true );
576
577 $flags = PPFrame::NO_ARGS | PPFrame::NO_TEMPLATES;
578 $dom = $this->preprocessToDom( $text, self::PTD_FOR_INCLUSION );
579 $text = $this->getPreprocessor()->newFrame()->expand( $dom, $flags );
580 $text = $this->mStripState->unstripBoth( $text );
581 return $text;
582 }
583
584 /**
585 * Get a random string
586 *
587 * @return string
588 */
589 static public function getRandomString() {
590 return wfRandomString( 16 );
591 }
592
593 /**
594 * Set the current user.
595 * Should only be used when doing pre-save transform.
596 *
597 * @param $user Mixed: User object or null (to reset)
598 */
599 function setUser( $user ) {
600 $this->mUser = $user;
601 }
602
603 /**
604 * Accessor for mUniqPrefix.
605 *
606 * @return String
607 */
608 public function uniqPrefix() {
609 if ( !isset( $this->mUniqPrefix ) ) {
610 # @todo FIXME: This is probably *horribly wrong*
611 # LanguageConverter seems to want $wgParser's uniqPrefix, however
612 # if this is called for a parser cache hit, the parser may not
613 # have ever been initialized in the first place.
614 # Not really sure what the heck is supposed to be going on here.
615 return '';
616 # throw new MWException( "Accessing uninitialized mUniqPrefix" );
617 }
618 return $this->mUniqPrefix;
619 }
620
621 /**
622 * Set the context title
623 *
624 * @param $t Title
625 */
626 function setTitle( $t ) {
627 if ( !$t || $t instanceof FakeTitle ) {
628 $t = Title::newFromText( 'NO TITLE' );
629 }
630
631 if ( strval( $t->getFragment() ) !== '' ) {
632 # Strip the fragment to avoid various odd effects
633 $this->mTitle = clone $t;
634 $this->mTitle->setFragment( '' );
635 } else {
636 $this->mTitle = $t;
637 }
638 }
639
640 /**
641 * Accessor for the Title object
642 *
643 * @return Title object
644 */
645 function getTitle() {
646 return $this->mTitle;
647 }
648
649 /**
650 * Accessor/mutator for the Title object
651 *
652 * @param $x Title object or null to just get the current one
653 * @return Title object
654 */
655 function Title( $x = null ) {
656 return wfSetVar( $this->mTitle, $x );
657 }
658
659 /**
660 * Set the output type
661 *
662 * @param $ot Integer: new value
663 */
664 function setOutputType( $ot ) {
665 $this->mOutputType = $ot;
666 # Shortcut alias
667 $this->ot = array(
668 'html' => $ot == self::OT_HTML,
669 'wiki' => $ot == self::OT_WIKI,
670 'pre' => $ot == self::OT_PREPROCESS,
671 'plain' => $ot == self::OT_PLAIN,
672 );
673 }
674
675 /**
676 * Accessor/mutator for the output type
677 *
678 * @param $x int|null New value or null to just get the current one
679 * @return Integer
680 */
681 function OutputType( $x = null ) {
682 return wfSetVar( $this->mOutputType, $x );
683 }
684
685 /**
686 * Get the ParserOutput object
687 *
688 * @return ParserOutput object
689 */
690 function getOutput() {
691 return $this->mOutput;
692 }
693
694 /**
695 * Get the ParserOptions object
696 *
697 * @return ParserOptions object
698 */
699 function getOptions() {
700 return $this->mOptions;
701 }
702
703 /**
704 * Accessor/mutator for the ParserOptions object
705 *
706 * @param $x ParserOptions New value or null to just get the current one
707 * @return ParserOptions Current ParserOptions object
708 */
709 function Options( $x = null ) {
710 return wfSetVar( $this->mOptions, $x );
711 }
712
713 /**
714 * @return int
715 */
716 function nextLinkID() {
717 return $this->mLinkID++;
718 }
719
720 /**
721 * @param $id int
722 */
723 function setLinkID( $id ) {
724 $this->mLinkID = $id;
725 }
726
727 /**
728 * Get a language object for use in parser functions such as {{FORMATNUM:}}
729 * @return Language
730 */
731 function getFunctionLang() {
732 return $this->getTargetLanguage();
733 }
734
735 /**
736 * Get the target language for the content being parsed. This is usually the
737 * language that the content is in.
738 *
739 * @since 1.19
740 *
741 * @return Language|null
742 */
743 public function getTargetLanguage() {
744 $target = $this->mOptions->getTargetLanguage();
745
746 if ( $target !== null ) {
747 return $target;
748 } elseif( $this->mOptions->getInterfaceMessage() ) {
749 return $this->mOptions->getUserLangObj();
750 } elseif( is_null( $this->mTitle ) ) {
751 throw new MWException( __METHOD__ . ': $this->mTitle is null' );
752 }
753
754 return $this->mTitle->getPageLanguage();
755 }
756
757 /**
758 * Get the language object for language conversion
759 */
760 function getConverterLanguage() {
761 global $wgBug34832TransitionalRollback, $wgContLang;
762 if ( $wgBug34832TransitionalRollback ) {
763 return $wgContLang;
764 } else {
765 return $this->getTargetLanguage();
766 }
767 }
768
769 /**
770 * Get a User object either from $this->mUser, if set, or from the
771 * ParserOptions object otherwise
772 *
773 * @return User object
774 */
775 function getUser() {
776 if ( !is_null( $this->mUser ) ) {
777 return $this->mUser;
778 }
779 return $this->mOptions->getUser();
780 }
781
782 /**
783 * Get a preprocessor object
784 *
785 * @return Preprocessor instance
786 */
787 function getPreprocessor() {
788 if ( !isset( $this->mPreprocessor ) ) {
789 $class = $this->mPreprocessorClass;
790 $this->mPreprocessor = new $class( $this );
791 }
792 return $this->mPreprocessor;
793 }
794
795 /**
796 * Replaces all occurrences of HTML-style comments and the given tags
797 * in the text with a random marker and returns the next text. The output
798 * parameter $matches will be an associative array filled with data in
799 * the form:
800 *
801 * @code
802 * 'UNIQ-xxxxx' => array(
803 * 'element',
804 * 'tag content',
805 * array( 'param' => 'x' ),
806 * '<element param="x">tag content</element>' ) )
807 * @endcode
808 *
809 * @param $elements array list of element names. Comments are always extracted.
810 * @param $text string Source text string.
811 * @param $matches array Out parameter, Array: extracted tags
812 * @param $uniq_prefix string
813 * @return String: stripped text
814 */
815 public static function extractTagsAndParams( $elements, $text, &$matches, $uniq_prefix = '' ) {
816 static $n = 1;
817 $stripped = '';
818 $matches = array();
819
820 $taglist = implode( '|', $elements );
821 $start = "/<($taglist)(\\s+[^>]*?|\\s*?)(\/?" . ">)|<(!--)/i";
822
823 while ( $text != '' ) {
824 $p = preg_split( $start, $text, 2, PREG_SPLIT_DELIM_CAPTURE );
825 $stripped .= $p[0];
826 if ( count( $p ) < 5 ) {
827 break;
828 }
829 if ( count( $p ) > 5 ) {
830 # comment
831 $element = $p[4];
832 $attributes = '';
833 $close = '';
834 $inside = $p[5];
835 } else {
836 # tag
837 $element = $p[1];
838 $attributes = $p[2];
839 $close = $p[3];
840 $inside = $p[4];
841 }
842
843 $marker = "$uniq_prefix-$element-" . sprintf( '%08X', $n++ ) . self::MARKER_SUFFIX;
844 $stripped .= $marker;
845
846 if ( $close === '/>' ) {
847 # Empty element tag, <tag />
848 $content = null;
849 $text = $inside;
850 $tail = null;
851 } else {
852 if ( $element === '!--' ) {
853 $end = '/(-->)/';
854 } else {
855 $end = "/(<\\/$element\\s*>)/i";
856 }
857 $q = preg_split( $end, $inside, 2, PREG_SPLIT_DELIM_CAPTURE );
858 $content = $q[0];
859 if ( count( $q ) < 3 ) {
860 # No end tag -- let it run out to the end of the text.
861 $tail = '';
862 $text = '';
863 } else {
864 $tail = $q[1];
865 $text = $q[2];
866 }
867 }
868
869 $matches[$marker] = array( $element,
870 $content,
871 Sanitizer::decodeTagAttributes( $attributes ),
872 "<$element$attributes$close$content$tail" );
873 }
874 return $stripped;
875 }
876
877 /**
878 * Get a list of strippable XML-like elements
879 *
880 * @return array
881 */
882 function getStripList() {
883 return $this->mStripList;
884 }
885
886 /**
887 * Add an item to the strip state
888 * Returns the unique tag which must be inserted into the stripped text
889 * The tag will be replaced with the original text in unstrip()
890 *
891 * @param $text string
892 *
893 * @return string
894 */
895 function insertStripItem( $text ) {
896 $rnd = "{$this->mUniqPrefix}-item-{$this->mMarkerIndex}-" . self::MARKER_SUFFIX;
897 $this->mMarkerIndex++;
898 $this->mStripState->addGeneral( $rnd, $text );
899 return $rnd;
900 }
901
902 /**
903 * parse the wiki syntax used to render tables
904 *
905 * @private
906 * @return string
907 */
908 function doTableStuff( $text ) {
909 wfProfileIn( __METHOD__ );
910
911 $lines = StringUtils::explode( "\n", $text );
912 $out = '';
913 $td_history = array(); # Is currently a td tag open?
914 $last_tag_history = array(); # Save history of last lag activated (td, th or caption)
915 $tr_history = array(); # Is currently a tr tag open?
916 $tr_attributes = array(); # history of tr attributes
917 $has_opened_tr = array(); # Did this table open a <tr> element?
918 $indent_level = 0; # indent level of the table
919
920 foreach ( $lines as $outLine ) {
921 $line = trim( $outLine );
922
923 if ( $line === '' ) { # empty line, go to next line
924 $out .= $outLine."\n";
925 continue;
926 }
927
928 $first_character = $line[0];
929 $matches = array();
930
931 if ( preg_match( '/^(:*)\{\|(.*)$/', $line , $matches ) ) {
932 # First check if we are starting a new table
933 $indent_level = strlen( $matches[1] );
934
935 $attributes = $this->mStripState->unstripBoth( $matches[2] );
936 $attributes = Sanitizer::fixTagAttributes( $attributes , 'table' );
937
938 $outLine = str_repeat( '<dl><dd>' , $indent_level ) . "<table{$attributes}>";
939 array_push( $td_history , false );
940 array_push( $last_tag_history , '' );
941 array_push( $tr_history , false );
942 array_push( $tr_attributes , '' );
943 array_push( $has_opened_tr , false );
944 } elseif ( count( $td_history ) == 0 ) {
945 # Don't do any of the following
946 $out .= $outLine."\n";
947 continue;
948 } elseif ( substr( $line , 0 , 2 ) === '|}' ) {
949 # We are ending a table
950 $line = '</table>' . substr( $line , 2 );
951 $last_tag = array_pop( $last_tag_history );
952
953 if ( !array_pop( $has_opened_tr ) ) {
954 $line = "<tr><td></td></tr>{$line}";
955 }
956
957 if ( array_pop( $tr_history ) ) {
958 $line = "</tr>{$line}";
959 }
960
961 if ( array_pop( $td_history ) ) {
962 $line = "</{$last_tag}>{$line}";
963 }
964 array_pop( $tr_attributes );
965 $outLine = $line . str_repeat( '</dd></dl>' , $indent_level );
966 } elseif ( substr( $line , 0 , 2 ) === '|-' ) {
967 # Now we have a table row
968 $line = preg_replace( '#^\|-+#', '', $line );
969
970 # Whats after the tag is now only attributes
971 $attributes = $this->mStripState->unstripBoth( $line );
972 $attributes = Sanitizer::fixTagAttributes( $attributes, 'tr' );
973 array_pop( $tr_attributes );
974 array_push( $tr_attributes, $attributes );
975
976 $line = '';
977 $last_tag = array_pop( $last_tag_history );
978 array_pop( $has_opened_tr );
979 array_push( $has_opened_tr , true );
980
981 if ( array_pop( $tr_history ) ) {
982 $line = '</tr>';
983 }
984
985 if ( array_pop( $td_history ) ) {
986 $line = "</{$last_tag}>{$line}";
987 }
988
989 $outLine = $line;
990 array_push( $tr_history , false );
991 array_push( $td_history , false );
992 array_push( $last_tag_history , '' );
993 } elseif ( $first_character === '|' || $first_character === '!' || substr( $line , 0 , 2 ) === '|+' ) {
994 # This might be cell elements, td, th or captions
995 if ( substr( $line , 0 , 2 ) === '|+' ) {
996 $first_character = '+';
997 $line = substr( $line , 1 );
998 }
999
1000 $line = substr( $line , 1 );
1001
1002 if ( $first_character === '!' ) {
1003 $line = str_replace( '!!' , '||' , $line );
1004 }
1005
1006 # Split up multiple cells on the same line.
1007 # FIXME : This can result in improper nesting of tags processed
1008 # by earlier parser steps, but should avoid splitting up eg
1009 # attribute values containing literal "||".
1010 $cells = StringUtils::explodeMarkup( '||' , $line );
1011
1012 $outLine = '';
1013
1014 # Loop through each table cell
1015 foreach ( $cells as $cell ) {
1016 $previous = '';
1017 if ( $first_character !== '+' ) {
1018 $tr_after = array_pop( $tr_attributes );
1019 if ( !array_pop( $tr_history ) ) {
1020 $previous = "<tr{$tr_after}>\n";
1021 }
1022 array_push( $tr_history , true );
1023 array_push( $tr_attributes , '' );
1024 array_pop( $has_opened_tr );
1025 array_push( $has_opened_tr , true );
1026 }
1027
1028 $last_tag = array_pop( $last_tag_history );
1029
1030 if ( array_pop( $td_history ) ) {
1031 $previous = "</{$last_tag}>\n{$previous}";
1032 }
1033
1034 if ( $first_character === '|' ) {
1035 $last_tag = 'td';
1036 } elseif ( $first_character === '!' ) {
1037 $last_tag = 'th';
1038 } elseif ( $first_character === '+' ) {
1039 $last_tag = 'caption';
1040 } else {
1041 $last_tag = '';
1042 }
1043
1044 array_push( $last_tag_history , $last_tag );
1045
1046 # A cell could contain both parameters and data
1047 $cell_data = explode( '|' , $cell , 2 );
1048
1049 # Bug 553: Note that a '|' inside an invalid link should not
1050 # be mistaken as delimiting cell parameters
1051 if ( strpos( $cell_data[0], '[[' ) !== false ) {
1052 $cell = "{$previous}<{$last_tag}>{$cell}";
1053 } elseif ( count( $cell_data ) == 1 ) {
1054 $cell = "{$previous}<{$last_tag}>{$cell_data[0]}";
1055 } else {
1056 $attributes = $this->mStripState->unstripBoth( $cell_data[0] );
1057 $attributes = Sanitizer::fixTagAttributes( $attributes , $last_tag );
1058 $cell = "{$previous}<{$last_tag}{$attributes}>{$cell_data[1]}";
1059 }
1060
1061 $outLine .= $cell;
1062 array_push( $td_history , true );
1063 }
1064 }
1065 $out .= $outLine . "\n";
1066 }
1067
1068 # Closing open td, tr && table
1069 while ( count( $td_history ) > 0 ) {
1070 if ( array_pop( $td_history ) ) {
1071 $out .= "</td>\n";
1072 }
1073 if ( array_pop( $tr_history ) ) {
1074 $out .= "</tr>\n";
1075 }
1076 if ( !array_pop( $has_opened_tr ) ) {
1077 $out .= "<tr><td></td></tr>\n" ;
1078 }
1079
1080 $out .= "</table>\n";
1081 }
1082
1083 # Remove trailing line-ending (b/c)
1084 if ( substr( $out, -1 ) === "\n" ) {
1085 $out = substr( $out, 0, -1 );
1086 }
1087
1088 # special case: don't return empty table
1089 if ( $out === "<table>\n<tr><td></td></tr>\n</table>" ) {
1090 $out = '';
1091 }
1092
1093 wfProfileOut( __METHOD__ );
1094
1095 return $out;
1096 }
1097
1098 /**
1099 * Helper function for parse() that transforms wiki markup into
1100 * HTML. Only called for $mOutputType == self::OT_HTML.
1101 *
1102 * @private
1103 *
1104 * @param $text string
1105 * @param $isMain bool
1106 * @param $frame bool
1107 *
1108 * @return string
1109 */
1110 function internalParse( $text, $isMain = true, $frame = false ) {
1111 wfProfileIn( __METHOD__ );
1112
1113 $origText = $text;
1114
1115 # Hook to suspend the parser in this state
1116 if ( !wfRunHooks( 'ParserBeforeInternalParse', array( &$this, &$text, &$this->mStripState ) ) ) {
1117 wfProfileOut( __METHOD__ );
1118 return $text ;
1119 }
1120
1121 # if $frame is provided, then use $frame for replacing any variables
1122 if ( $frame ) {
1123 # use frame depth to infer how include/noinclude tags should be handled
1124 # depth=0 means this is the top-level document; otherwise it's an included document
1125 if ( !$frame->depth ) {
1126 $flag = 0;
1127 } else {
1128 $flag = Parser::PTD_FOR_INCLUSION;
1129 }
1130 $dom = $this->preprocessToDom( $text, $flag );
1131 $text = $frame->expand( $dom );
1132 } else {
1133 # if $frame is not provided, then use old-style replaceVariables
1134 $text = $this->replaceVariables( $text );
1135 }
1136
1137 wfRunHooks( 'InternalParseBeforeSanitize', array( &$this, &$text, &$this->mStripState ) );
1138 $text = Sanitizer::removeHTMLtags( $text, array( &$this, 'attributeStripCallback' ), false, array_keys( $this->mTransparentTagHooks ) );
1139 wfRunHooks( 'InternalParseBeforeLinks', array( &$this, &$text, &$this->mStripState ) );
1140
1141 # Tables need to come after variable replacement for things to work
1142 # properly; putting them before other transformations should keep
1143 # exciting things like link expansions from showing up in surprising
1144 # places.
1145 $text = $this->doTableStuff( $text );
1146
1147 $text = preg_replace( '/(^|\n)-----*/', '\\1<hr />', $text );
1148
1149 $text = $this->doDoubleUnderscore( $text );
1150
1151 $text = $this->doHeadings( $text );
1152 if ( $this->mOptions->getUseDynamicDates() ) {
1153 $df = DateFormatter::getInstance();
1154 $text = $df->reformat( $this->mOptions->getDateFormat(), $text );
1155 }
1156 $text = $this->replaceInternalLinks( $text );
1157 $text = $this->doAllQuotes( $text );
1158 $text = $this->replaceExternalLinks( $text );
1159
1160 # replaceInternalLinks may sometimes leave behind
1161 # absolute URLs, which have to be masked to hide them from replaceExternalLinks
1162 $text = str_replace( $this->mUniqPrefix.'NOPARSE', '', $text );
1163
1164 $text = $this->doMagicLinks( $text );
1165 $text = $this->formatHeadings( $text, $origText, $isMain );
1166
1167 wfProfileOut( __METHOD__ );
1168 return $text;
1169 }
1170
1171 /**
1172 * Replace special strings like "ISBN xxx" and "RFC xxx" with
1173 * magic external links.
1174 *
1175 * DML
1176 * @private
1177 *
1178 * @param $text string
1179 *
1180 * @return string
1181 */
1182 function doMagicLinks( $text ) {
1183 wfProfileIn( __METHOD__ );
1184 $prots = wfUrlProtocolsWithoutProtRel();
1185 $urlChar = self::EXT_LINK_URL_CLASS;
1186 $text = preg_replace_callback(
1187 '!(?: # Start cases
1188 (<a[ \t\r\n>].*?</a>) | # m[1]: Skip link text
1189 (<.*?>) | # m[2]: Skip stuff inside HTML elements' . "
1190 (\\b(?i:$prots)$urlChar+) | # m[3]: Free external links" . '
1191 (?:RFC|PMID)\s+([0-9]+) | # m[4]: RFC or PMID, capture number
1192 ISBN\s+(\b # m[5]: ISBN, capture number
1193 (?: 97[89] [\ \-]? )? # optional 13-digit ISBN prefix
1194 (?: [0-9] [\ \-]? ){9} # 9 digits with opt. delimiters
1195 [0-9Xx] # check digit
1196 \b)
1197 )!xu', array( &$this, 'magicLinkCallback' ), $text );
1198 wfProfileOut( __METHOD__ );
1199 return $text;
1200 }
1201
1202 /**
1203 * @throws MWException
1204 * @param $m array
1205 * @return HTML|string
1206 */
1207 function magicLinkCallback( $m ) {
1208 if ( isset( $m[1] ) && $m[1] !== '' ) {
1209 # Skip anchor
1210 return $m[0];
1211 } elseif ( isset( $m[2] ) && $m[2] !== '' ) {
1212 # Skip HTML element
1213 return $m[0];
1214 } elseif ( isset( $m[3] ) && $m[3] !== '' ) {
1215 # Free external link
1216 return $this->makeFreeExternalLink( $m[0] );
1217 } elseif ( isset( $m[4] ) && $m[4] !== '' ) {
1218 # RFC or PMID
1219 if ( substr( $m[0], 0, 3 ) === 'RFC' ) {
1220 $keyword = 'RFC';
1221 $urlmsg = 'rfcurl';
1222 $CssClass = 'mw-magiclink-rfc';
1223 $id = $m[4];
1224 } elseif ( substr( $m[0], 0, 4 ) === 'PMID' ) {
1225 $keyword = 'PMID';
1226 $urlmsg = 'pubmedurl';
1227 $CssClass = 'mw-magiclink-pmid';
1228 $id = $m[4];
1229 } else {
1230 throw new MWException( __METHOD__.': unrecognised match type "' .
1231 substr( $m[0], 0, 20 ) . '"' );
1232 }
1233 $url = wfMessage( $urlmsg, $id )->inContentLanguage()->text();
1234 return Linker::makeExternalLink( $url, "{$keyword} {$id}", true, $CssClass );
1235 } elseif ( isset( $m[5] ) && $m[5] !== '' ) {
1236 # ISBN
1237 $isbn = $m[5];
1238 $num = strtr( $isbn, array(
1239 '-' => '',
1240 ' ' => '',
1241 'x' => 'X',
1242 ));
1243 $titleObj = SpecialPage::getTitleFor( 'Booksources', $num );
1244 return'<a href="' .
1245 htmlspecialchars( $titleObj->getLocalUrl() ) .
1246 "\" class=\"internal mw-magiclink-isbn\">ISBN $isbn</a>";
1247 } else {
1248 return $m[0];
1249 }
1250 }
1251
1252 /**
1253 * Make a free external link, given a user-supplied URL
1254 *
1255 * @param $url string
1256 *
1257 * @return string HTML
1258 * @private
1259 */
1260 function makeFreeExternalLink( $url ) {
1261 wfProfileIn( __METHOD__ );
1262
1263 $trail = '';
1264
1265 # The characters '<' and '>' (which were escaped by
1266 # removeHTMLtags()) should not be included in
1267 # URLs, per RFC 2396.
1268 $m2 = array();
1269 if ( preg_match( '/&(lt|gt);/', $url, $m2, PREG_OFFSET_CAPTURE ) ) {
1270 $trail = substr( $url, $m2[0][1] ) . $trail;
1271 $url = substr( $url, 0, $m2[0][1] );
1272 }
1273
1274 # Move trailing punctuation to $trail
1275 $sep = ',;\.:!?';
1276 # If there is no left bracket, then consider right brackets fair game too
1277 if ( strpos( $url, '(' ) === false ) {
1278 $sep .= ')';
1279 }
1280
1281 $numSepChars = strspn( strrev( $url ), $sep );
1282 if ( $numSepChars ) {
1283 $trail = substr( $url, -$numSepChars ) . $trail;
1284 $url = substr( $url, 0, -$numSepChars );
1285 }
1286
1287 $url = Sanitizer::cleanUrl( $url );
1288
1289 # Is this an external image?
1290 $text = $this->maybeMakeExternalImage( $url );
1291 if ( $text === false ) {
1292 # Not an image, make a link
1293 $text = Linker::makeExternalLink( $url,
1294 $this->getConverterLanguage()->markNoConversion($url), true, 'free',
1295 $this->getExternalLinkAttribs( $url ) );
1296 # Register it in the output object...
1297 # Replace unnecessary URL escape codes with their equivalent characters
1298 $pasteurized = self::replaceUnusualEscapes( $url );
1299 $this->mOutput->addExternalLink( $pasteurized );
1300 }
1301 wfProfileOut( __METHOD__ );
1302 return $text . $trail;
1303 }
1304
1305
1306 /**
1307 * Parse headers and return html
1308 *
1309 * @private
1310 *
1311 * @param $text string
1312 *
1313 * @return string
1314 */
1315 function doHeadings( $text ) {
1316 wfProfileIn( __METHOD__ );
1317 for ( $i = 6; $i >= 1; --$i ) {
1318 $h = str_repeat( '=', $i );
1319 $text = preg_replace( "/^$h(.+)$h\\s*$/m",
1320 "<h$i>\\1</h$i>", $text );
1321 }
1322 wfProfileOut( __METHOD__ );
1323 return $text;
1324 }
1325
1326 /**
1327 * Replace single quotes with HTML markup
1328 * @private
1329 *
1330 * @param $text string
1331 *
1332 * @return string the altered text
1333 */
1334 function doAllQuotes( $text ) {
1335 wfProfileIn( __METHOD__ );
1336 $outtext = '';
1337 $lines = StringUtils::explode( "\n", $text );
1338 foreach ( $lines as $line ) {
1339 $outtext .= $this->doQuotes( $line ) . "\n";
1340 }
1341 $outtext = substr( $outtext, 0,-1 );
1342 wfProfileOut( __METHOD__ );
1343 return $outtext;
1344 }
1345
1346 /**
1347 * Helper function for doAllQuotes()
1348 *
1349 * @param $text string
1350 *
1351 * @return string
1352 */
1353 public function doQuotes( $text ) {
1354 $arr = preg_split( "/(''+)/", $text, -1, PREG_SPLIT_DELIM_CAPTURE );
1355 if ( count( $arr ) == 1 ) {
1356 return $text;
1357 } else {
1358 # First, do some preliminary work. This may shift some apostrophes from
1359 # being mark-up to being text. It also counts the number of occurrences
1360 # of bold and italics mark-ups.
1361 $numbold = 0;
1362 $numitalics = 0;
1363 for ( $i = 0; $i < count( $arr ); $i++ ) {
1364 if ( ( $i % 2 ) == 1 ) {
1365 # If there are ever four apostrophes, assume the first is supposed to
1366 # be text, and the remaining three constitute mark-up for bold text.
1367 if ( strlen( $arr[$i] ) == 4 ) {
1368 $arr[$i-1] .= "'";
1369 $arr[$i] = "'''";
1370 } elseif ( strlen( $arr[$i] ) > 5 ) {
1371 # If there are more than 5 apostrophes in a row, assume they're all
1372 # text except for the last 5.
1373 $arr[$i-1] .= str_repeat( "'", strlen( $arr[$i] ) - 5 );
1374 $arr[$i] = "'''''";
1375 }
1376 # Count the number of occurrences of bold and italics mark-ups.
1377 # We are not counting sequences of five apostrophes.
1378 if ( strlen( $arr[$i] ) == 2 ) {
1379 $numitalics++;
1380 } elseif ( strlen( $arr[$i] ) == 3 ) {
1381 $numbold++;
1382 } elseif ( strlen( $arr[$i] ) == 5 ) {
1383 $numitalics++;
1384 $numbold++;
1385 }
1386 }
1387 }
1388
1389 # If there is an odd number of both bold and italics, it is likely
1390 # that one of the bold ones was meant to be an apostrophe followed
1391 # by italics. Which one we cannot know for certain, but it is more
1392 # likely to be one that has a single-letter word before it.
1393 if ( ( $numbold % 2 == 1 ) && ( $numitalics % 2 == 1 ) ) {
1394 $i = 0;
1395 $firstsingleletterword = -1;
1396 $firstmultiletterword = -1;
1397 $firstspace = -1;
1398 foreach ( $arr as $r ) {
1399 if ( ( $i % 2 == 1 ) and ( strlen( $r ) == 3 ) ) {
1400 $x1 = substr( $arr[$i-1], -1 );
1401 $x2 = substr( $arr[$i-1], -2, 1 );
1402 if ( $x1 === ' ' ) {
1403 if ( $firstspace == -1 ) {
1404 $firstspace = $i;
1405 }
1406 } elseif ( $x2 === ' ') {
1407 if ( $firstsingleletterword == -1 ) {
1408 $firstsingleletterword = $i;
1409 }
1410 } else {
1411 if ( $firstmultiletterword == -1 ) {
1412 $firstmultiletterword = $i;
1413 }
1414 }
1415 }
1416 $i++;
1417 }
1418
1419 # If there is a single-letter word, use it!
1420 if ( $firstsingleletterword > -1 ) {
1421 $arr[$firstsingleletterword] = "''";
1422 $arr[$firstsingleletterword-1] .= "'";
1423 } elseif ( $firstmultiletterword > -1 ) {
1424 # If not, but there's a multi-letter word, use that one.
1425 $arr[$firstmultiletterword] = "''";
1426 $arr[$firstmultiletterword-1] .= "'";
1427 } elseif ( $firstspace > -1 ) {
1428 # ... otherwise use the first one that has neither.
1429 # (notice that it is possible for all three to be -1 if, for example,
1430 # there is only one pentuple-apostrophe in the line)
1431 $arr[$firstspace] = "''";
1432 $arr[$firstspace-1] .= "'";
1433 }
1434 }
1435
1436 # Now let's actually convert our apostrophic mush to HTML!
1437 $output = '';
1438 $buffer = '';
1439 $state = '';
1440 $i = 0;
1441 foreach ( $arr as $r ) {
1442 if ( ( $i % 2 ) == 0 ) {
1443 if ( $state === 'both' ) {
1444 $buffer .= $r;
1445 } else {
1446 $output .= $r;
1447 }
1448 } else {
1449 if ( strlen( $r ) == 2 ) {
1450 if ( $state === 'i' ) {
1451 $output .= '</i>'; $state = '';
1452 } elseif ( $state === 'bi' ) {
1453 $output .= '</i>'; $state = 'b';
1454 } elseif ( $state === 'ib' ) {
1455 $output .= '</b></i><b>'; $state = 'b';
1456 } elseif ( $state === 'both' ) {
1457 $output .= '<b><i>'.$buffer.'</i>'; $state = 'b';
1458 } else { # $state can be 'b' or ''
1459 $output .= '<i>'; $state .= 'i';
1460 }
1461 } elseif ( strlen( $r ) == 3 ) {
1462 if ( $state === 'b' ) {
1463 $output .= '</b>'; $state = '';
1464 } elseif ( $state === 'bi' ) {
1465 $output .= '</i></b><i>'; $state = 'i';
1466 } elseif ( $state === 'ib' ) {
1467 $output .= '</b>'; $state = 'i';
1468 } elseif ( $state === 'both' ) {
1469 $output .= '<i><b>'.$buffer.'</b>'; $state = 'i';
1470 } else { # $state can be 'i' or ''
1471 $output .= '<b>'; $state .= 'b';
1472 }
1473 } elseif ( strlen( $r ) == 5 ) {
1474 if ( $state === 'b' ) {
1475 $output .= '</b><i>'; $state = 'i';
1476 } elseif ( $state === 'i' ) {
1477 $output .= '</i><b>'; $state = 'b';
1478 } elseif ( $state === 'bi' ) {
1479 $output .= '</i></b>'; $state = '';
1480 } elseif ( $state === 'ib' ) {
1481 $output .= '</b></i>'; $state = '';
1482 } elseif ( $state === 'both' ) {
1483 $output .= '<i><b>'.$buffer.'</b></i>'; $state = '';
1484 } else { # ($state == '')
1485 $buffer = ''; $state = 'both';
1486 }
1487 }
1488 }
1489 $i++;
1490 }
1491 # Now close all remaining tags. Notice that the order is important.
1492 if ( $state === 'b' || $state === 'ib' ) {
1493 $output .= '</b>';
1494 }
1495 if ( $state === 'i' || $state === 'bi' || $state === 'ib' ) {
1496 $output .= '</i>';
1497 }
1498 if ( $state === 'bi' ) {
1499 $output .= '</b>';
1500 }
1501 # There might be lonely ''''', so make sure we have a buffer
1502 if ( $state === 'both' && $buffer ) {
1503 $output .= '<b><i>'.$buffer.'</i></b>';
1504 }
1505 return $output;
1506 }
1507 }
1508
1509 /**
1510 * Replace external links (REL)
1511 *
1512 * Note: this is all very hackish and the order of execution matters a lot.
1513 * Make sure to run maintenance/parserTests.php if you change this code.
1514 *
1515 * @private
1516 *
1517 * @param $text string
1518 *
1519 * @return string
1520 */
1521 function replaceExternalLinks( $text ) {
1522 wfProfileIn( __METHOD__ );
1523
1524 $bits = preg_split( $this->mExtLinkBracketedRegex, $text, -1, PREG_SPLIT_DELIM_CAPTURE );
1525 if ( $bits === false ) {
1526 throw new MWException( "PCRE needs to be compiled with --enable-unicode-properties in order for MediaWiki to function" );
1527 }
1528 $s = array_shift( $bits );
1529
1530 $i = 0;
1531 while ( $i<count( $bits ) ) {
1532 $url = $bits[$i++];
1533 $protocol = $bits[$i++];
1534 $text = $bits[$i++];
1535 $trail = $bits[$i++];
1536
1537 # The characters '<' and '>' (which were escaped by
1538 # removeHTMLtags()) should not be included in
1539 # URLs, per RFC 2396.
1540 $m2 = array();
1541 if ( preg_match( '/&(lt|gt);/', $url, $m2, PREG_OFFSET_CAPTURE ) ) {
1542 $text = substr( $url, $m2[0][1] ) . ' ' . $text;
1543 $url = substr( $url, 0, $m2[0][1] );
1544 }
1545
1546 # If the link text is an image URL, replace it with an <img> tag
1547 # This happened by accident in the original parser, but some people used it extensively
1548 $img = $this->maybeMakeExternalImage( $text );
1549 if ( $img !== false ) {
1550 $text = $img;
1551 }
1552
1553 $dtrail = '';
1554
1555 # Set linktype for CSS - if URL==text, link is essentially free
1556 $linktype = ( $text === $url ) ? 'free' : 'text';
1557
1558 # No link text, e.g. [http://domain.tld/some.link]
1559 if ( $text == '' ) {
1560 # Autonumber
1561 $langObj = $this->getTargetLanguage();
1562 $text = '[' . $langObj->formatNum( ++$this->mAutonumber ) . ']';
1563 $linktype = 'autonumber';
1564 } else {
1565 # Have link text, e.g. [http://domain.tld/some.link text]s
1566 # Check for trail
1567 list( $dtrail, $trail ) = Linker::splitTrail( $trail );
1568 }
1569
1570 $text = $this->getConverterLanguage()->markNoConversion( $text );
1571
1572 $url = Sanitizer::cleanUrl( $url );
1573
1574 # Use the encoded URL
1575 # This means that users can paste URLs directly into the text
1576 # Funny characters like ö aren't valid in URLs anyway
1577 # This was changed in August 2004
1578 $s .= Linker::makeExternalLink( $url, $text, false, $linktype,
1579 $this->getExternalLinkAttribs( $url ) ) . $dtrail . $trail;
1580
1581 # Register link in the output object.
1582 # Replace unnecessary URL escape codes with the referenced character
1583 # This prevents spammers from hiding links from the filters
1584 $pasteurized = self::replaceUnusualEscapes( $url );
1585 $this->mOutput->addExternalLink( $pasteurized );
1586 }
1587
1588 wfProfileOut( __METHOD__ );
1589 return $s;
1590 }
1591
1592 /**
1593 * Get an associative array of additional HTML attributes appropriate for a
1594 * particular external link. This currently may include rel => nofollow
1595 * (depending on configuration, namespace, and the URL's domain) and/or a
1596 * target attribute (depending on configuration).
1597 *
1598 * @param $url String|bool optional URL, to extract the domain from for rel =>
1599 * nofollow if appropriate
1600 * @return Array associative array of HTML attributes
1601 */
1602 function getExternalLinkAttribs( $url = false ) {
1603 $attribs = array();
1604 global $wgNoFollowLinks, $wgNoFollowNsExceptions, $wgNoFollowDomainExceptions;
1605 $ns = $this->mTitle->getNamespace();
1606 if ( $wgNoFollowLinks && !in_array( $ns, $wgNoFollowNsExceptions ) &&
1607 !wfMatchesDomainList( $url, $wgNoFollowDomainExceptions ) )
1608 {
1609 $attribs['rel'] = 'nofollow';
1610 }
1611 if ( $this->mOptions->getExternalLinkTarget() ) {
1612 $attribs['target'] = $this->mOptions->getExternalLinkTarget();
1613 }
1614 return $attribs;
1615 }
1616
1617 /**
1618 * Replace unusual URL escape codes with their equivalent characters
1619 *
1620 * @param $url String
1621 * @return String
1622 *
1623 * @todo This can merge genuinely required bits in the path or query string,
1624 * breaking legit URLs. A proper fix would treat the various parts of
1625 * the URL differently; as a workaround, just use the output for
1626 * statistical records, not for actual linking/output.
1627 */
1628 static function replaceUnusualEscapes( $url ) {
1629 return preg_replace_callback( '/%[0-9A-Fa-f]{2}/',
1630 array( __CLASS__, 'replaceUnusualEscapesCallback' ), $url );
1631 }
1632
1633 /**
1634 * Callback function used in replaceUnusualEscapes().
1635 * Replaces unusual URL escape codes with their equivalent character
1636 *
1637 * @param $matches array
1638 *
1639 * @return string
1640 */
1641 private static function replaceUnusualEscapesCallback( $matches ) {
1642 $char = urldecode( $matches[0] );
1643 $ord = ord( $char );
1644 # Is it an unsafe or HTTP reserved character according to RFC 1738?
1645 if ( $ord > 32 && $ord < 127 && strpos( '<>"#{}|\^~[]`;/?', $char ) === false ) {
1646 # No, shouldn't be escaped
1647 return $char;
1648 } else {
1649 # Yes, leave it escaped
1650 return $matches[0];
1651 }
1652 }
1653
1654 /**
1655 * make an image if it's allowed, either through the global
1656 * option, through the exception, or through the on-wiki whitelist
1657 * @private
1658 *
1659 * $param $url string
1660 *
1661 * @return string
1662 */
1663 function maybeMakeExternalImage( $url ) {
1664 $imagesfrom = $this->mOptions->getAllowExternalImagesFrom();
1665 $imagesexception = !empty( $imagesfrom );
1666 $text = false;
1667 # $imagesfrom could be either a single string or an array of strings, parse out the latter
1668 if ( $imagesexception && is_array( $imagesfrom ) ) {
1669 $imagematch = false;
1670 foreach ( $imagesfrom as $match ) {
1671 if ( strpos( $url, $match ) === 0 ) {
1672 $imagematch = true;
1673 break;
1674 }
1675 }
1676 } elseif ( $imagesexception ) {
1677 $imagematch = ( strpos( $url, $imagesfrom ) === 0 );
1678 } else {
1679 $imagematch = false;
1680 }
1681 if ( $this->mOptions->getAllowExternalImages()
1682 || ( $imagesexception && $imagematch ) ) {
1683 if ( preg_match( self::EXT_IMAGE_REGEX, $url ) ) {
1684 # Image found
1685 $text = Linker::makeExternalImage( $url );
1686 }
1687 }
1688 if ( !$text && $this->mOptions->getEnableImageWhitelist()
1689 && preg_match( self::EXT_IMAGE_REGEX, $url ) ) {
1690 $whitelist = explode( "\n", wfMessage( 'external_image_whitelist' )->inContentLanguage()->text() );
1691 foreach ( $whitelist as $entry ) {
1692 # Sanitize the regex fragment, make it case-insensitive, ignore blank entries/comments
1693 if ( strpos( $entry, '#' ) === 0 || $entry === '' ) {
1694 continue;
1695 }
1696 if ( preg_match( '/' . str_replace( '/', '\\/', $entry ) . '/i', $url ) ) {
1697 # Image matches a whitelist entry
1698 $text = Linker::makeExternalImage( $url );
1699 break;
1700 }
1701 }
1702 }
1703 return $text;
1704 }
1705
1706 /**
1707 * Process [[ ]] wikilinks
1708 *
1709 * @param $s string
1710 *
1711 * @return String: processed text
1712 *
1713 * @private
1714 */
1715 function replaceInternalLinks( $s ) {
1716 $this->mLinkHolders->merge( $this->replaceInternalLinks2( $s ) );
1717 return $s;
1718 }
1719
1720 /**
1721 * Process [[ ]] wikilinks (RIL)
1722 * @return LinkHolderArray
1723 *
1724 * @private
1725 */
1726 function replaceInternalLinks2( &$s ) {
1727 wfProfileIn( __METHOD__ );
1728
1729 wfProfileIn( __METHOD__.'-setup' );
1730 static $tc = FALSE, $e1, $e1_img;
1731 # the % is needed to support urlencoded titles as well
1732 if ( !$tc ) {
1733 $tc = Title::legalChars() . '#%';
1734 # Match a link having the form [[namespace:link|alternate]]trail
1735 $e1 = "/^([{$tc}]+)(?:\\|(.+?))?]](.*)\$/sD";
1736 # Match cases where there is no "]]", which might still be images
1737 $e1_img = "/^([{$tc}]+)\\|(.*)\$/sD";
1738 }
1739
1740 $holders = new LinkHolderArray( $this );
1741
1742 # split the entire text string on occurrences of [[
1743 $a = StringUtils::explode( '[[', ' ' . $s );
1744 # get the first element (all text up to first [[), and remove the space we added
1745 $s = $a->current();
1746 $a->next();
1747 $line = $a->current(); # Workaround for broken ArrayIterator::next() that returns "void"
1748 $s = substr( $s, 1 );
1749
1750 $useLinkPrefixExtension = $this->getTargetLanguage()->linkPrefixExtension();
1751 $e2 = null;
1752 if ( $useLinkPrefixExtension ) {
1753 # Match the end of a line for a word that's not followed by whitespace,
1754 # e.g. in the case of 'The Arab al[[Razi]]', 'al' will be matched
1755 $e2 = wfMessage( 'linkprefix' )->inContentLanguage()->text();
1756 }
1757
1758 if ( is_null( $this->mTitle ) ) {
1759 wfProfileOut( __METHOD__.'-setup' );
1760 wfProfileOut( __METHOD__ );
1761 throw new MWException( __METHOD__.": \$this->mTitle is null\n" );
1762 }
1763 $nottalk = !$this->mTitle->isTalkPage();
1764
1765 if ( $useLinkPrefixExtension ) {
1766 $m = array();
1767 if ( preg_match( $e2, $s, $m ) ) {
1768 $first_prefix = $m[2];
1769 } else {
1770 $first_prefix = false;
1771 }
1772 } else {
1773 $prefix = '';
1774 }
1775
1776 if ( $this->getConverterLanguage()->hasVariants() ) {
1777 $selflink = $this->getConverterLanguage()->autoConvertToAllVariants(
1778 $this->mTitle->getPrefixedText() );
1779 } else {
1780 $selflink = array( $this->mTitle->getPrefixedText() );
1781 }
1782 $useSubpages = $this->areSubpagesAllowed();
1783 wfProfileOut( __METHOD__.'-setup' );
1784
1785 # Loop for each link
1786 for ( ; $line !== false && $line !== null ; $a->next(), $line = $a->current() ) {
1787 # Check for excessive memory usage
1788 if ( $holders->isBig() ) {
1789 # Too big
1790 # Do the existence check, replace the link holders and clear the array
1791 $holders->replace( $s );
1792 $holders->clear();
1793 }
1794
1795 if ( $useLinkPrefixExtension ) {
1796 wfProfileIn( __METHOD__.'-prefixhandling' );
1797 if ( preg_match( $e2, $s, $m ) ) {
1798 $prefix = $m[2];
1799 $s = $m[1];
1800 } else {
1801 $prefix='';
1802 }
1803 # first link
1804 if ( $first_prefix ) {
1805 $prefix = $first_prefix;
1806 $first_prefix = false;
1807 }
1808 wfProfileOut( __METHOD__.'-prefixhandling' );
1809 }
1810
1811 $might_be_img = false;
1812
1813 wfProfileIn( __METHOD__."-e1" );
1814 if ( preg_match( $e1, $line, $m ) ) { # page with normal text or alt
1815 $text = $m[2];
1816 # If we get a ] at the beginning of $m[3] that means we have a link that's something like:
1817 # [[Image:Foo.jpg|[http://example.com desc]]] <- having three ] in a row fucks up,
1818 # the real problem is with the $e1 regex
1819 # See bug 1300.
1820 #
1821 # Still some problems for cases where the ] is meant to be outside punctuation,
1822 # and no image is in sight. See bug 2095.
1823 #
1824 if ( $text !== '' &&
1825 substr( $m[3], 0, 1 ) === ']' &&
1826 strpos( $text, '[' ) !== false
1827 )
1828 {
1829 $text .= ']'; # so that replaceExternalLinks($text) works later
1830 $m[3] = substr( $m[3], 1 );
1831 }
1832 # fix up urlencoded title texts
1833 if ( strpos( $m[1], '%' ) !== false ) {
1834 # Should anchors '#' also be rejected?
1835 $m[1] = str_replace( array('<', '>'), array('&lt;', '&gt;'), rawurldecode( $m[1] ) );
1836 }
1837 $trail = $m[3];
1838 } elseif ( preg_match( $e1_img, $line, $m ) ) { # Invalid, but might be an image with a link in its caption
1839 $might_be_img = true;
1840 $text = $m[2];
1841 if ( strpos( $m[1], '%' ) !== false ) {
1842 $m[1] = rawurldecode( $m[1] );
1843 }
1844 $trail = "";
1845 } else { # Invalid form; output directly
1846 $s .= $prefix . '[[' . $line ;
1847 wfProfileOut( __METHOD__."-e1" );
1848 continue;
1849 }
1850 wfProfileOut( __METHOD__."-e1" );
1851 wfProfileIn( __METHOD__."-misc" );
1852
1853 # Don't allow internal links to pages containing
1854 # PROTO: where PROTO is a valid URL protocol; these
1855 # should be external links.
1856 if ( preg_match( '/^(?i:' . $this->mUrlProtocols . ')/', $m[1] ) ) {
1857 $s .= $prefix . '[[' . $line ;
1858 wfProfileOut( __METHOD__."-misc" );
1859 continue;
1860 }
1861
1862 # Make subpage if necessary
1863 if ( $useSubpages ) {
1864 $link = $this->maybeDoSubpageLink( $m[1], $text );
1865 } else {
1866 $link = $m[1];
1867 }
1868
1869 $noforce = ( substr( $m[1], 0, 1 ) !== ':' );
1870 if ( !$noforce ) {
1871 # Strip off leading ':'
1872 $link = substr( $link, 1 );
1873 }
1874
1875 wfProfileOut( __METHOD__."-misc" );
1876 wfProfileIn( __METHOD__."-title" );
1877 $nt = Title::newFromText( $this->mStripState->unstripNoWiki( $link ) );
1878 if ( $nt === null ) {
1879 $s .= $prefix . '[[' . $line;
1880 wfProfileOut( __METHOD__."-title" );
1881 continue;
1882 }
1883
1884 $ns = $nt->getNamespace();
1885 $iw = $nt->getInterWiki();
1886 wfProfileOut( __METHOD__."-title" );
1887
1888 if ( $might_be_img ) { # if this is actually an invalid link
1889 wfProfileIn( __METHOD__."-might_be_img" );
1890 if ( $ns == NS_FILE && $noforce ) { # but might be an image
1891 $found = false;
1892 while ( true ) {
1893 # look at the next 'line' to see if we can close it there
1894 $a->next();
1895 $next_line = $a->current();
1896 if ( $next_line === false || $next_line === null ) {
1897 break;
1898 }
1899 $m = explode( ']]', $next_line, 3 );
1900 if ( count( $m ) == 3 ) {
1901 # the first ]] closes the inner link, the second the image
1902 $found = true;
1903 $text .= "[[{$m[0]}]]{$m[1]}";
1904 $trail = $m[2];
1905 break;
1906 } elseif ( count( $m ) == 2 ) {
1907 # if there's exactly one ]] that's fine, we'll keep looking
1908 $text .= "[[{$m[0]}]]{$m[1]}";
1909 } else {
1910 # if $next_line is invalid too, we need look no further
1911 $text .= '[[' . $next_line;
1912 break;
1913 }
1914 }
1915 if ( !$found ) {
1916 # we couldn't find the end of this imageLink, so output it raw
1917 # but don't ignore what might be perfectly normal links in the text we've examined
1918 $holders->merge( $this->replaceInternalLinks2( $text ) );
1919 $s .= "{$prefix}[[$link|$text";
1920 # note: no $trail, because without an end, there *is* no trail
1921 wfProfileOut( __METHOD__."-might_be_img" );
1922 continue;
1923 }
1924 } else { # it's not an image, so output it raw
1925 $s .= "{$prefix}[[$link|$text";
1926 # note: no $trail, because without an end, there *is* no trail
1927 wfProfileOut( __METHOD__."-might_be_img" );
1928 continue;
1929 }
1930 wfProfileOut( __METHOD__."-might_be_img" );
1931 }
1932
1933 $wasblank = ( $text == '' );
1934 if ( $wasblank ) {
1935 $text = $link;
1936 } else {
1937 # Bug 4598 madness. Handle the quotes only if they come from the alternate part
1938 # [[Lista d''e paise d''o munno]] -> <a href="...">Lista d''e paise d''o munno</a>
1939 # [[Criticism of Harry Potter|Criticism of ''Harry Potter'']]
1940 # -> <a href="Criticism of Harry Potter">Criticism of <i>Harry Potter</i></a>
1941 $text = $this->doQuotes( $text );
1942 }
1943
1944 # Link not escaped by : , create the various objects
1945 if ( $noforce ) {
1946 # Interwikis
1947 wfProfileIn( __METHOD__."-interwiki" );
1948 if ( $iw && $this->mOptions->getInterwikiMagic() && $nottalk && Language::fetchLanguageName( $iw, null, 'mw' ) ) {
1949 // FIXME: the above check prevents links to sites with identifiers that are not language codes
1950 $this->mOutput->addLanguageLink( $nt->getFullText() );
1951 $s = rtrim( $s . $prefix );
1952 $s .= trim( $trail, "\n" ) == '' ? '': $prefix . $trail;
1953 wfProfileOut( __METHOD__."-interwiki" );
1954 continue;
1955 }
1956 wfProfileOut( __METHOD__."-interwiki" );
1957
1958 if ( $ns == NS_FILE ) {
1959 wfProfileIn( __METHOD__."-image" );
1960 if ( !wfIsBadImage( $nt->getDBkey(), $this->mTitle ) ) {
1961 if ( $wasblank ) {
1962 # if no parameters were passed, $text
1963 # becomes something like "File:Foo.png",
1964 # which we don't want to pass on to the
1965 # image generator
1966 $text = '';
1967 } else {
1968 # recursively parse links inside the image caption
1969 # actually, this will parse them in any other parameters, too,
1970 # but it might be hard to fix that, and it doesn't matter ATM
1971 $text = $this->replaceExternalLinks( $text );
1972 $holders->merge( $this->replaceInternalLinks2( $text ) );
1973 }
1974 # cloak any absolute URLs inside the image markup, so replaceExternalLinks() won't touch them
1975 $s .= $prefix . $this->armorLinks(
1976 $this->makeImage( $nt, $text, $holders ) ) . $trail;
1977 } else {
1978 $s .= $prefix . $trail;
1979 }
1980 wfProfileOut( __METHOD__."-image" );
1981 continue;
1982 }
1983
1984 if ( $ns == NS_CATEGORY ) {
1985 wfProfileIn( __METHOD__."-category" );
1986 $s = rtrim( $s . "\n" ); # bug 87
1987
1988 if ( $wasblank ) {
1989 $sortkey = $this->getDefaultSort();
1990 } else {
1991 $sortkey = $text;
1992 }
1993 $sortkey = Sanitizer::decodeCharReferences( $sortkey );
1994 $sortkey = str_replace( "\n", '', $sortkey );
1995 $sortkey = $this->getConverterLanguage()->convertCategoryKey( $sortkey );
1996 $this->mOutput->addCategory( $nt->getDBkey(), $sortkey );
1997
1998 /**
1999 * Strip the whitespace Category links produce, see bug 87
2000 * @todo We might want to use trim($tmp, "\n") here.
2001 */
2002 $s .= trim( $prefix . $trail, "\n" ) == '' ? '' : $prefix . $trail;
2003
2004 wfProfileOut( __METHOD__."-category" );
2005 continue;
2006 }
2007 }
2008
2009 # Self-link checking
2010 if ( $nt->getFragment() === '' && $ns != NS_SPECIAL ) {
2011 if ( in_array( $nt->getPrefixedText(), $selflink, true ) ) {
2012 $s .= $prefix . Linker::makeSelfLinkObj( $nt, $text, '', $trail );
2013 continue;
2014 }
2015 }
2016
2017 # NS_MEDIA is a pseudo-namespace for linking directly to a file
2018 # @todo FIXME: Should do batch file existence checks, see comment below
2019 if ( $ns == NS_MEDIA ) {
2020 wfProfileIn( __METHOD__."-media" );
2021 # Give extensions a chance to select the file revision for us
2022 $options = array();
2023 $descQuery = false;
2024 wfRunHooks( 'BeforeParserFetchFileAndTitle',
2025 array( $this, $nt, &$options, &$descQuery ) );
2026 # Fetch and register the file (file title may be different via hooks)
2027 list( $file, $nt ) = $this->fetchFileAndTitle( $nt, $options );
2028 # Cloak with NOPARSE to avoid replacement in replaceExternalLinks
2029 $s .= $prefix . $this->armorLinks(
2030 Linker::makeMediaLinkFile( $nt, $file, $text ) ) . $trail;
2031 wfProfileOut( __METHOD__."-media" );
2032 continue;
2033 }
2034
2035 wfProfileIn( __METHOD__."-always_known" );
2036 # Some titles, such as valid special pages or files in foreign repos, should
2037 # be shown as bluelinks even though they're not included in the page table
2038 #
2039 # @todo FIXME: isAlwaysKnown() can be expensive for file links; we should really do
2040 # batch file existence checks for NS_FILE and NS_MEDIA
2041 if ( $iw == '' && $nt->isAlwaysKnown() ) {
2042 $this->mOutput->addLink( $nt );
2043 $s .= $this->makeKnownLinkHolder( $nt, $text, array(), $trail, $prefix );
2044 } else {
2045 # Links will be added to the output link list after checking
2046 $s .= $holders->makeHolder( $nt, $text, array(), $trail, $prefix );
2047 }
2048 wfProfileOut( __METHOD__."-always_known" );
2049 }
2050 wfProfileOut( __METHOD__ );
2051 return $holders;
2052 }
2053
2054 /**
2055 * Render a forced-blue link inline; protect against double expansion of
2056 * URLs if we're in a mode that prepends full URL prefixes to internal links.
2057 * Since this little disaster has to split off the trail text to avoid
2058 * breaking URLs in the following text without breaking trails on the
2059 * wiki links, it's been made into a horrible function.
2060 *
2061 * @param $nt Title
2062 * @param $text String
2063 * @param $query Array or String
2064 * @param $trail String
2065 * @param $prefix String
2066 * @return String: HTML-wikitext mix oh yuck
2067 */
2068 function makeKnownLinkHolder( $nt, $text = '', $query = array(), $trail = '', $prefix = '' ) {
2069 list( $inside, $trail ) = Linker::splitTrail( $trail );
2070
2071 if ( is_string( $query ) ) {
2072 $query = wfCgiToArray( $query );
2073 }
2074 if ( $text == '' ) {
2075 $text = htmlspecialchars( $nt->getPrefixedText() );
2076 }
2077
2078 $link = Linker::linkKnown( $nt, "$prefix$text$inside", array(), $query );
2079
2080 return $this->armorLinks( $link ) . $trail;
2081 }
2082
2083 /**
2084 * Insert a NOPARSE hacky thing into any inline links in a chunk that's
2085 * going to go through further parsing steps before inline URL expansion.
2086 *
2087 * Not needed quite as much as it used to be since free links are a bit
2088 * more sensible these days. But bracketed links are still an issue.
2089 *
2090 * @param $text String: more-or-less HTML
2091 * @return String: less-or-more HTML with NOPARSE bits
2092 */
2093 function armorLinks( $text ) {
2094 return preg_replace( '/\b((?i)' . $this->mUrlProtocols . ')/',
2095 "{$this->mUniqPrefix}NOPARSE$1", $text );
2096 }
2097
2098 /**
2099 * Return true if subpage links should be expanded on this page.
2100 * @return Boolean
2101 */
2102 function areSubpagesAllowed() {
2103 # Some namespaces don't allow subpages
2104 return MWNamespace::hasSubpages( $this->mTitle->getNamespace() );
2105 }
2106
2107 /**
2108 * Handle link to subpage if necessary
2109 *
2110 * @param $target String: the source of the link
2111 * @param &$text String: the link text, modified as necessary
2112 * @return string the full name of the link
2113 * @private
2114 */
2115 function maybeDoSubpageLink( $target, &$text ) {
2116 return Linker::normalizeSubpageLink( $this->mTitle, $target, $text );
2117 }
2118
2119 /**#@+
2120 * Used by doBlockLevels()
2121 * @private
2122 *
2123 * @return string
2124 */
2125 function closeParagraph() {
2126 $result = '';
2127 if ( $this->mLastSection != '' ) {
2128 $result = '</' . $this->mLastSection . ">\n";
2129 }
2130 $this->mInPre = false;
2131 $this->mLastSection = '';
2132 return $result;
2133 }
2134
2135 /**
2136 * getCommon() returns the length of the longest common substring
2137 * of both arguments, starting at the beginning of both.
2138 * @private
2139 *
2140 * @param $st1 string
2141 * @param $st2 string
2142 *
2143 * @return int
2144 */
2145 function getCommon( $st1, $st2 ) {
2146 $fl = strlen( $st1 );
2147 $shorter = strlen( $st2 );
2148 if ( $fl < $shorter ) {
2149 $shorter = $fl;
2150 }
2151
2152 for ( $i = 0; $i < $shorter; ++$i ) {
2153 if ( $st1[$i] != $st2[$i] ) {
2154 break;
2155 }
2156 }
2157 return $i;
2158 }
2159
2160 /**
2161 * These next three functions open, continue, and close the list
2162 * element appropriate to the prefix character passed into them.
2163 * @private
2164 *
2165 * @param $char string
2166 *
2167 * @return string
2168 */
2169 function openList( $char ) {
2170 $result = $this->closeParagraph();
2171
2172 if ( '*' === $char ) {
2173 $result .= '<ul><li>';
2174 } elseif ( '#' === $char ) {
2175 $result .= '<ol><li>';
2176 } elseif ( ':' === $char ) {
2177 $result .= '<dl><dd>';
2178 } elseif ( ';' === $char ) {
2179 $result .= '<dl><dt>';
2180 $this->mDTopen = true;
2181 } else {
2182 $result = '<!-- ERR 1 -->';
2183 }
2184
2185 return $result;
2186 }
2187
2188 /**
2189 * TODO: document
2190 * @param $char String
2191 * @private
2192 *
2193 * @return string
2194 */
2195 function nextItem( $char ) {
2196 if ( '*' === $char || '#' === $char ) {
2197 return '</li><li>';
2198 } elseif ( ':' === $char || ';' === $char ) {
2199 $close = '</dd>';
2200 if ( $this->mDTopen ) {
2201 $close = '</dt>';
2202 }
2203 if ( ';' === $char ) {
2204 $this->mDTopen = true;
2205 return $close . '<dt>';
2206 } else {
2207 $this->mDTopen = false;
2208 return $close . '<dd>';
2209 }
2210 }
2211 return '<!-- ERR 2 -->';
2212 }
2213
2214 /**
2215 * TODO: document
2216 * @param $char String
2217 * @private
2218 *
2219 * @return string
2220 */
2221 function closeList( $char ) {
2222 if ( '*' === $char ) {
2223 $text = '</li></ul>';
2224 } elseif ( '#' === $char ) {
2225 $text = '</li></ol>';
2226 } elseif ( ':' === $char ) {
2227 if ( $this->mDTopen ) {
2228 $this->mDTopen = false;
2229 $text = '</dt></dl>';
2230 } else {
2231 $text = '</dd></dl>';
2232 }
2233 } else {
2234 return '<!-- ERR 3 -->';
2235 }
2236 return $text."\n";
2237 }
2238 /**#@-*/
2239
2240 /**
2241 * Make lists from lines starting with ':', '*', '#', etc. (DBL)
2242 *
2243 * @param $text String
2244 * @param $linestart Boolean: whether or not this is at the start of a line.
2245 * @private
2246 * @return string the lists rendered as HTML
2247 */
2248 function doBlockLevels( $text, $linestart ) {
2249 wfProfileIn( __METHOD__ );
2250
2251 # Parsing through the text line by line. The main thing
2252 # happening here is handling of block-level elements p, pre,
2253 # and making lists from lines starting with * # : etc.
2254 #
2255 $textLines = StringUtils::explode( "\n", $text );
2256
2257 $lastPrefix = $output = '';
2258 $this->mDTopen = $inBlockElem = false;
2259 $prefixLength = 0;
2260 $paragraphStack = false;
2261
2262 foreach ( $textLines as $oLine ) {
2263 # Fix up $linestart
2264 if ( !$linestart ) {
2265 $output .= $oLine;
2266 $linestart = true;
2267 continue;
2268 }
2269 # * = ul
2270 # # = ol
2271 # ; = dt
2272 # : = dd
2273
2274 $lastPrefixLength = strlen( $lastPrefix );
2275 $preCloseMatch = preg_match( '/<\\/pre/i', $oLine );
2276 $preOpenMatch = preg_match( '/<pre/i', $oLine );
2277 # If not in a <pre> element, scan for and figure out what prefixes are there.
2278 if ( !$this->mInPre ) {
2279 # Multiple prefixes may abut each other for nested lists.
2280 $prefixLength = strspn( $oLine, '*#:;' );
2281 $prefix = substr( $oLine, 0, $prefixLength );
2282
2283 # eh?
2284 # ; and : are both from definition-lists, so they're equivalent
2285 # for the purposes of determining whether or not we need to open/close
2286 # elements.
2287 $prefix2 = str_replace( ';', ':', $prefix );
2288 $t = substr( $oLine, $prefixLength );
2289 $this->mInPre = (bool)$preOpenMatch;
2290 } else {
2291 # Don't interpret any other prefixes in preformatted text
2292 $prefixLength = 0;
2293 $prefix = $prefix2 = '';
2294 $t = $oLine;
2295 }
2296
2297 # List generation
2298 if ( $prefixLength && $lastPrefix === $prefix2 ) {
2299 # Same as the last item, so no need to deal with nesting or opening stuff
2300 $output .= $this->nextItem( substr( $prefix, -1 ) );
2301 $paragraphStack = false;
2302
2303 if ( substr( $prefix, -1 ) === ';') {
2304 # The one nasty exception: definition lists work like this:
2305 # ; title : definition text
2306 # So we check for : in the remainder text to split up the
2307 # title and definition, without b0rking links.
2308 $term = $t2 = '';
2309 if ( $this->findColonNoLinks( $t, $term, $t2 ) !== false ) {
2310 $t = $t2;
2311 $output .= $term . $this->nextItem( ':' );
2312 }
2313 }
2314 } elseif ( $prefixLength || $lastPrefixLength ) {
2315 # We need to open or close prefixes, or both.
2316
2317 # Either open or close a level...
2318 $commonPrefixLength = $this->getCommon( $prefix, $lastPrefix );
2319 $paragraphStack = false;
2320
2321 # Close all the prefixes which aren't shared.
2322 while ( $commonPrefixLength < $lastPrefixLength ) {
2323 $output .= $this->closeList( $lastPrefix[$lastPrefixLength-1] );
2324 --$lastPrefixLength;
2325 }
2326
2327 # Continue the current prefix if appropriate.
2328 if ( $prefixLength <= $commonPrefixLength && $commonPrefixLength > 0 ) {
2329 $output .= $this->nextItem( $prefix[$commonPrefixLength-1] );
2330 }
2331
2332 # Open prefixes where appropriate.
2333 while ( $prefixLength > $commonPrefixLength ) {
2334 $char = substr( $prefix, $commonPrefixLength, 1 );
2335 $output .= $this->openList( $char );
2336
2337 if ( ';' === $char ) {
2338 # @todo FIXME: This is dupe of code above
2339 if ( $this->findColonNoLinks( $t, $term, $t2 ) !== false ) {
2340 $t = $t2;
2341 $output .= $term . $this->nextItem( ':' );
2342 }
2343 }
2344 ++$commonPrefixLength;
2345 }
2346 $lastPrefix = $prefix2;
2347 }
2348
2349 # If we have no prefixes, go to paragraph mode.
2350 if ( 0 == $prefixLength ) {
2351 wfProfileIn( __METHOD__."-paragraph" );
2352 # No prefix (not in list)--go to paragraph mode
2353 # XXX: use a stack for nestable elements like span, table and div
2354 $openmatch = preg_match('/(?:<table|<blockquote|<h1|<h2|<h3|<h4|<h5|<h6|<pre|<tr|<p|<ul|<ol|<li|<\\/tr|<\\/td|<\\/th)/iS', $t );
2355 $closematch = preg_match(
2356 '/(?:<\\/table|<\\/blockquote|<\\/h1|<\\/h2|<\\/h3|<\\/h4|<\\/h5|<\\/h6|'.
2357 '<td|<th|<\\/?div|<hr|<\\/pre|<\\/p|'.$this->mUniqPrefix.'-pre|<\\/li|<\\/ul|<\\/ol|<\\/?center)/iS', $t );
2358 if ( $openmatch or $closematch ) {
2359 $paragraphStack = false;
2360 # TODO bug 5718: paragraph closed
2361 $output .= $this->closeParagraph();
2362 if ( $preOpenMatch and !$preCloseMatch ) {
2363 $this->mInPre = true;
2364 }
2365 $inBlockElem = !$closematch;
2366 } elseif ( !$inBlockElem && !$this->mInPre ) {
2367 if ( ' ' == substr( $t, 0, 1 ) and ( $this->mLastSection === 'pre' || trim( $t ) != '' ) ) {
2368 # pre
2369 if ( $this->mLastSection !== 'pre' ) {
2370 $paragraphStack = false;
2371 $output .= $this->closeParagraph().'<pre>';
2372 $this->mLastSection = 'pre';
2373 }
2374 $t = substr( $t, 1 );
2375 } else {
2376 # paragraph
2377 if ( trim( $t ) === '' ) {
2378 if ( $paragraphStack ) {
2379 $output .= $paragraphStack.'<br />';
2380 $paragraphStack = false;
2381 $this->mLastSection = 'p';
2382 } else {
2383 if ( $this->mLastSection !== 'p' ) {
2384 $output .= $this->closeParagraph();
2385 $this->mLastSection = '';
2386 $paragraphStack = '<p>';
2387 } else {
2388 $paragraphStack = '</p><p>';
2389 }
2390 }
2391 } else {
2392 if ( $paragraphStack ) {
2393 $output .= $paragraphStack;
2394 $paragraphStack = false;
2395 $this->mLastSection = 'p';
2396 } elseif ( $this->mLastSection !== 'p' ) {
2397 $output .= $this->closeParagraph().'<p>';
2398 $this->mLastSection = 'p';
2399 }
2400 }
2401 }
2402 }
2403 wfProfileOut( __METHOD__."-paragraph" );
2404 }
2405 # somewhere above we forget to get out of pre block (bug 785)
2406 if ( $preCloseMatch && $this->mInPre ) {
2407 $this->mInPre = false;
2408 }
2409 if ( $paragraphStack === false ) {
2410 $output .= $t."\n";
2411 }
2412 }
2413 while ( $prefixLength ) {
2414 $output .= $this->closeList( $prefix2[$prefixLength-1] );
2415 --$prefixLength;
2416 }
2417 if ( $this->mLastSection != '' ) {
2418 $output .= '</' . $this->mLastSection . '>';
2419 $this->mLastSection = '';
2420 }
2421
2422 wfProfileOut( __METHOD__ );
2423 return $output;
2424 }
2425
2426 /**
2427 * Split up a string on ':', ignoring any occurrences inside tags
2428 * to prevent illegal overlapping.
2429 *
2430 * @param $str String the string to split
2431 * @param &$before String set to everything before the ':'
2432 * @param &$after String set to everything after the ':'
2433 * @return String the position of the ':', or false if none found
2434 */
2435 function findColonNoLinks( $str, &$before, &$after ) {
2436 wfProfileIn( __METHOD__ );
2437
2438 $pos = strpos( $str, ':' );
2439 if ( $pos === false ) {
2440 # Nothing to find!
2441 wfProfileOut( __METHOD__ );
2442 return false;
2443 }
2444
2445 $lt = strpos( $str, '<' );
2446 if ( $lt === false || $lt > $pos ) {
2447 # Easy; no tag nesting to worry about
2448 $before = substr( $str, 0, $pos );
2449 $after = substr( $str, $pos+1 );
2450 wfProfileOut( __METHOD__ );
2451 return $pos;
2452 }
2453
2454 # Ugly state machine to walk through avoiding tags.
2455 $state = self::COLON_STATE_TEXT;
2456 $stack = 0;
2457 $len = strlen( $str );
2458 for( $i = 0; $i < $len; $i++ ) {
2459 $c = $str[$i];
2460
2461 switch( $state ) {
2462 # (Using the number is a performance hack for common cases)
2463 case 0: # self::COLON_STATE_TEXT:
2464 switch( $c ) {
2465 case "<":
2466 # Could be either a <start> tag or an </end> tag
2467 $state = self::COLON_STATE_TAGSTART;
2468 break;
2469 case ":":
2470 if ( $stack == 0 ) {
2471 # We found it!
2472 $before = substr( $str, 0, $i );
2473 $after = substr( $str, $i + 1 );
2474 wfProfileOut( __METHOD__ );
2475 return $i;
2476 }
2477 # Embedded in a tag; don't break it.
2478 break;
2479 default:
2480 # Skip ahead looking for something interesting
2481 $colon = strpos( $str, ':', $i );
2482 if ( $colon === false ) {
2483 # Nothing else interesting
2484 wfProfileOut( __METHOD__ );
2485 return false;
2486 }
2487 $lt = strpos( $str, '<', $i );
2488 if ( $stack === 0 ) {
2489 if ( $lt === false || $colon < $lt ) {
2490 # We found it!
2491 $before = substr( $str, 0, $colon );
2492 $after = substr( $str, $colon + 1 );
2493 wfProfileOut( __METHOD__ );
2494 return $i;
2495 }
2496 }
2497 if ( $lt === false ) {
2498 # Nothing else interesting to find; abort!
2499 # We're nested, but there's no close tags left. Abort!
2500 break 2;
2501 }
2502 # Skip ahead to next tag start
2503 $i = $lt;
2504 $state = self::COLON_STATE_TAGSTART;
2505 }
2506 break;
2507 case 1: # self::COLON_STATE_TAG:
2508 # In a <tag>
2509 switch( $c ) {
2510 case ">":
2511 $stack++;
2512 $state = self::COLON_STATE_TEXT;
2513 break;
2514 case "/":
2515 # Slash may be followed by >?
2516 $state = self::COLON_STATE_TAGSLASH;
2517 break;
2518 default:
2519 # ignore
2520 }
2521 break;
2522 case 2: # self::COLON_STATE_TAGSTART:
2523 switch( $c ) {
2524 case "/":
2525 $state = self::COLON_STATE_CLOSETAG;
2526 break;
2527 case "!":
2528 $state = self::COLON_STATE_COMMENT;
2529 break;
2530 case ">":
2531 # Illegal early close? This shouldn't happen D:
2532 $state = self::COLON_STATE_TEXT;
2533 break;
2534 default:
2535 $state = self::COLON_STATE_TAG;
2536 }
2537 break;
2538 case 3: # self::COLON_STATE_CLOSETAG:
2539 # In a </tag>
2540 if ( $c === ">" ) {
2541 $stack--;
2542 if ( $stack < 0 ) {
2543 wfDebug( __METHOD__.": Invalid input; too many close tags\n" );
2544 wfProfileOut( __METHOD__ );
2545 return false;
2546 }
2547 $state = self::COLON_STATE_TEXT;
2548 }
2549 break;
2550 case self::COLON_STATE_TAGSLASH:
2551 if ( $c === ">" ) {
2552 # Yes, a self-closed tag <blah/>
2553 $state = self::COLON_STATE_TEXT;
2554 } else {
2555 # Probably we're jumping the gun, and this is an attribute
2556 $state = self::COLON_STATE_TAG;
2557 }
2558 break;
2559 case 5: # self::COLON_STATE_COMMENT:
2560 if ( $c === "-" ) {
2561 $state = self::COLON_STATE_COMMENTDASH;
2562 }
2563 break;
2564 case self::COLON_STATE_COMMENTDASH:
2565 if ( $c === "-" ) {
2566 $state = self::COLON_STATE_COMMENTDASHDASH;
2567 } else {
2568 $state = self::COLON_STATE_COMMENT;
2569 }
2570 break;
2571 case self::COLON_STATE_COMMENTDASHDASH:
2572 if ( $c === ">" ) {
2573 $state = self::COLON_STATE_TEXT;
2574 } else {
2575 $state = self::COLON_STATE_COMMENT;
2576 }
2577 break;
2578 default:
2579 throw new MWException( "State machine error in " . __METHOD__ );
2580 }
2581 }
2582 if ( $stack > 0 ) {
2583 wfDebug( __METHOD__.": Invalid input; not enough close tags (stack $stack, state $state)\n" );
2584 wfProfileOut( __METHOD__ );
2585 return false;
2586 }
2587 wfProfileOut( __METHOD__ );
2588 return false;
2589 }
2590
2591 /**
2592 * Return value of a magic variable (like PAGENAME)
2593 *
2594 * @private
2595 *
2596 * @param $index integer
2597 * @param $frame PPFrame
2598 *
2599 * @return string
2600 */
2601 function getVariableValue( $index, $frame = false ) {
2602 global $wgContLang, $wgSitename, $wgServer;
2603 global $wgArticlePath, $wgScriptPath, $wgStylePath;
2604
2605 if ( is_null( $this->mTitle ) ) {
2606 // If no title set, bad things are going to happen
2607 // later. Title should always be set since this
2608 // should only be called in the middle of a parse
2609 // operation (but the unit-tests do funky stuff)
2610 throw new MWException( __METHOD__ . ' Should only be '
2611 . ' called while parsing (no title set)' );
2612 }
2613
2614 /**
2615 * Some of these require message or data lookups and can be
2616 * expensive to check many times.
2617 */
2618 if ( wfRunHooks( 'ParserGetVariableValueVarCache', array( &$this, &$this->mVarCache ) ) ) {
2619 if ( isset( $this->mVarCache[$index] ) ) {
2620 return $this->mVarCache[$index];
2621 }
2622 }
2623
2624 $ts = wfTimestamp( TS_UNIX, $this->mOptions->getTimestamp() );
2625 wfRunHooks( 'ParserGetVariableValueTs', array( &$this, &$ts ) );
2626
2627 # Use the time zone
2628 global $wgLocaltimezone;
2629 if ( isset( $wgLocaltimezone ) ) {
2630 $oldtz = date_default_timezone_get();
2631 date_default_timezone_set( $wgLocaltimezone );
2632 }
2633
2634 $localTimestamp = date( 'YmdHis', $ts );
2635 $localMonth = date( 'm', $ts );
2636 $localMonth1 = date( 'n', $ts );
2637 $localMonthName = date( 'n', $ts );
2638 $localDay = date( 'j', $ts );
2639 $localDay2 = date( 'd', $ts );
2640 $localDayOfWeek = date( 'w', $ts );
2641 $localWeek = date( 'W', $ts );
2642 $localYear = date( 'Y', $ts );
2643 $localHour = date( 'H', $ts );
2644 if ( isset( $wgLocaltimezone ) ) {
2645 date_default_timezone_set( $oldtz );
2646 }
2647
2648 $pageLang = $this->getFunctionLang();
2649
2650 switch ( $index ) {
2651 case 'currentmonth':
2652 $value = $pageLang->formatNum( gmdate( 'm', $ts ) );
2653 break;
2654 case 'currentmonth1':
2655 $value = $pageLang->formatNum( gmdate( 'n', $ts ) );
2656 break;
2657 case 'currentmonthname':
2658 $value = $pageLang->getMonthName( gmdate( 'n', $ts ) );
2659 break;
2660 case 'currentmonthnamegen':
2661 $value = $pageLang->getMonthNameGen( gmdate( 'n', $ts ) );
2662 break;
2663 case 'currentmonthabbrev':
2664 $value = $pageLang->getMonthAbbreviation( gmdate( 'n', $ts ) );
2665 break;
2666 case 'currentday':
2667 $value = $pageLang->formatNum( gmdate( 'j', $ts ) );
2668 break;
2669 case 'currentday2':
2670 $value = $pageLang->formatNum( gmdate( 'd', $ts ) );
2671 break;
2672 case 'localmonth':
2673 $value = $pageLang->formatNum( $localMonth );
2674 break;
2675 case 'localmonth1':
2676 $value = $pageLang->formatNum( $localMonth1 );
2677 break;
2678 case 'localmonthname':
2679 $value = $pageLang->getMonthName( $localMonthName );
2680 break;
2681 case 'localmonthnamegen':
2682 $value = $pageLang->getMonthNameGen( $localMonthName );
2683 break;
2684 case 'localmonthabbrev':
2685 $value = $pageLang->getMonthAbbreviation( $localMonthName );
2686 break;
2687 case 'localday':
2688 $value = $pageLang->formatNum( $localDay );
2689 break;
2690 case 'localday2':
2691 $value = $pageLang->formatNum( $localDay2 );
2692 break;
2693 case 'pagename':
2694 $value = wfEscapeWikiText( $this->mTitle->getText() );
2695 break;
2696 case 'pagenamee':
2697 $value = wfEscapeWikiText( $this->mTitle->getPartialURL() );
2698 break;
2699 case 'fullpagename':
2700 $value = wfEscapeWikiText( $this->mTitle->getPrefixedText() );
2701 break;
2702 case 'fullpagenamee':
2703 $value = wfEscapeWikiText( $this->mTitle->getPrefixedURL() );
2704 break;
2705 case 'subpagename':
2706 $value = wfEscapeWikiText( $this->mTitle->getSubpageText() );
2707 break;
2708 case 'subpagenamee':
2709 $value = wfEscapeWikiText( $this->mTitle->getSubpageUrlForm() );
2710 break;
2711 case 'basepagename':
2712 $value = wfEscapeWikiText( $this->mTitle->getBaseText() );
2713 break;
2714 case 'basepagenamee':
2715 $value = wfEscapeWikiText( wfUrlEncode( str_replace( ' ', '_', $this->mTitle->getBaseText() ) ) );
2716 break;
2717 case 'talkpagename':
2718 if ( $this->mTitle->canTalk() ) {
2719 $talkPage = $this->mTitle->getTalkPage();
2720 $value = wfEscapeWikiText( $talkPage->getPrefixedText() );
2721 } else {
2722 $value = '';
2723 }
2724 break;
2725 case 'talkpagenamee':
2726 if ( $this->mTitle->canTalk() ) {
2727 $talkPage = $this->mTitle->getTalkPage();
2728 $value = wfEscapeWikiText( $talkPage->getPrefixedUrl() );
2729 } else {
2730 $value = '';
2731 }
2732 break;
2733 case 'subjectpagename':
2734 $subjPage = $this->mTitle->getSubjectPage();
2735 $value = wfEscapeWikiText( $subjPage->getPrefixedText() );
2736 break;
2737 case 'subjectpagenamee':
2738 $subjPage = $this->mTitle->getSubjectPage();
2739 $value = wfEscapeWikiText( $subjPage->getPrefixedUrl() );
2740 break;
2741 case 'pageid': // requested in bug 23427
2742 $pageid = $this->getTitle()->getArticleId();
2743 if( $pageid == 0 ) {
2744 # 0 means the page doesn't exist in the database,
2745 # which means the user is previewing a new page.
2746 # The vary-revision flag must be set, because the magic word
2747 # will have a different value once the page is saved.
2748 $this->mOutput->setFlag( 'vary-revision' );
2749 wfDebug( __METHOD__ . ": {{PAGEID}} used in a new page, setting vary-revision...\n" );
2750 }
2751 $value = $pageid ? $pageid : null;
2752 break;
2753 case 'revisionid':
2754 # Let the edit saving system know we should parse the page
2755 # *after* a revision ID has been assigned.
2756 $this->mOutput->setFlag( 'vary-revision' );
2757 wfDebug( __METHOD__ . ": {{REVISIONID}} used, setting vary-revision...\n" );
2758 $value = $this->mRevisionId;
2759 break;
2760 case 'revisionday':
2761 # Let the edit saving system know we should parse the page
2762 # *after* a revision ID has been assigned. This is for null edits.
2763 $this->mOutput->setFlag( 'vary-revision' );
2764 wfDebug( __METHOD__ . ": {{REVISIONDAY}} used, setting vary-revision...\n" );
2765 $value = intval( substr( $this->getRevisionTimestamp(), 6, 2 ) );
2766 break;
2767 case 'revisionday2':
2768 # Let the edit saving system know we should parse the page
2769 # *after* a revision ID has been assigned. This is for null edits.
2770 $this->mOutput->setFlag( 'vary-revision' );
2771 wfDebug( __METHOD__ . ": {{REVISIONDAY2}} used, setting vary-revision...\n" );
2772 $value = substr( $this->getRevisionTimestamp(), 6, 2 );
2773 break;
2774 case 'revisionmonth':
2775 # Let the edit saving system know we should parse the page
2776 # *after* a revision ID has been assigned. This is for null edits.
2777 $this->mOutput->setFlag( 'vary-revision' );
2778 wfDebug( __METHOD__ . ": {{REVISIONMONTH}} used, setting vary-revision...\n" );
2779 $value = substr( $this->getRevisionTimestamp(), 4, 2 );
2780 break;
2781 case 'revisionmonth1':
2782 # Let the edit saving system know we should parse the page
2783 # *after* a revision ID has been assigned. This is for null edits.
2784 $this->mOutput->setFlag( 'vary-revision' );
2785 wfDebug( __METHOD__ . ": {{REVISIONMONTH1}} used, setting vary-revision...\n" );
2786 $value = intval( substr( $this->getRevisionTimestamp(), 4, 2 ) );
2787 break;
2788 case 'revisionyear':
2789 # Let the edit saving system know we should parse the page
2790 # *after* a revision ID has been assigned. This is for null edits.
2791 $this->mOutput->setFlag( 'vary-revision' );
2792 wfDebug( __METHOD__ . ": {{REVISIONYEAR}} used, setting vary-revision...\n" );
2793 $value = substr( $this->getRevisionTimestamp(), 0, 4 );
2794 break;
2795 case 'revisiontimestamp':
2796 # Let the edit saving system know we should parse the page
2797 # *after* a revision ID has been assigned. This is for null edits.
2798 $this->mOutput->setFlag( 'vary-revision' );
2799 wfDebug( __METHOD__ . ": {{REVISIONTIMESTAMP}} used, setting vary-revision...\n" );
2800 $value = $this->getRevisionTimestamp();
2801 break;
2802 case 'revisionuser':
2803 # Let the edit saving system know we should parse the page
2804 # *after* a revision ID has been assigned. This is for null edits.
2805 $this->mOutput->setFlag( 'vary-revision' );
2806 wfDebug( __METHOD__ . ": {{REVISIONUSER}} used, setting vary-revision...\n" );
2807 $value = $this->getRevisionUser();
2808 break;
2809 case 'namespace':
2810 $value = str_replace( '_',' ',$wgContLang->getNsText( $this->mTitle->getNamespace() ) );
2811 break;
2812 case 'namespacee':
2813 $value = wfUrlencode( $wgContLang->getNsText( $this->mTitle->getNamespace() ) );
2814 break;
2815 case 'namespacenumber':
2816 $value = $this->mTitle->getNamespace();
2817 break;
2818 case 'talkspace':
2819 $value = $this->mTitle->canTalk() ? str_replace( '_',' ',$this->mTitle->getTalkNsText() ) : '';
2820 break;
2821 case 'talkspacee':
2822 $value = $this->mTitle->canTalk() ? wfUrlencode( $this->mTitle->getTalkNsText() ) : '';
2823 break;
2824 case 'subjectspace':
2825 $value = $this->mTitle->getSubjectNsText();
2826 break;
2827 case 'subjectspacee':
2828 $value = ( wfUrlencode( $this->mTitle->getSubjectNsText() ) );
2829 break;
2830 case 'currentdayname':
2831 $value = $pageLang->getWeekdayName( gmdate( 'w', $ts ) + 1 );
2832 break;
2833 case 'currentyear':
2834 $value = $pageLang->formatNum( gmdate( 'Y', $ts ), true );
2835 break;
2836 case 'currenttime':
2837 $value = $pageLang->time( wfTimestamp( TS_MW, $ts ), false, false );
2838 break;
2839 case 'currenthour':
2840 $value = $pageLang->formatNum( gmdate( 'H', $ts ), true );
2841 break;
2842 case 'currentweek':
2843 # @bug 4594 PHP5 has it zero padded, PHP4 does not, cast to
2844 # int to remove the padding
2845 $value = $pageLang->formatNum( (int)gmdate( 'W', $ts ) );
2846 break;
2847 case 'currentdow':
2848 $value = $pageLang->formatNum( gmdate( 'w', $ts ) );
2849 break;
2850 case 'localdayname':
2851 $value = $pageLang->getWeekdayName( $localDayOfWeek + 1 );
2852 break;
2853 case 'localyear':
2854 $value = $pageLang->formatNum( $localYear, true );
2855 break;
2856 case 'localtime':
2857 $value = $pageLang->time( $localTimestamp, false, false );
2858 break;
2859 case 'localhour':
2860 $value = $pageLang->formatNum( $localHour, true );
2861 break;
2862 case 'localweek':
2863 # @bug 4594 PHP5 has it zero padded, PHP4 does not, cast to
2864 # int to remove the padding
2865 $value = $pageLang->formatNum( (int)$localWeek );
2866 break;
2867 case 'localdow':
2868 $value = $pageLang->formatNum( $localDayOfWeek );
2869 break;
2870 case 'numberofarticles':
2871 $value = $pageLang->formatNum( SiteStats::articles() );
2872 break;
2873 case 'numberoffiles':
2874 $value = $pageLang->formatNum( SiteStats::images() );
2875 break;
2876 case 'numberofusers':
2877 $value = $pageLang->formatNum( SiteStats::users() );
2878 break;
2879 case 'numberofactiveusers':
2880 $value = $pageLang->formatNum( SiteStats::activeUsers() );
2881 break;
2882 case 'numberofpages':
2883 $value = $pageLang->formatNum( SiteStats::pages() );
2884 break;
2885 case 'numberofadmins':
2886 $value = $pageLang->formatNum( SiteStats::numberingroup( 'sysop' ) );
2887 break;
2888 case 'numberofedits':
2889 $value = $pageLang->formatNum( SiteStats::edits() );
2890 break;
2891 case 'numberofviews':
2892 global $wgDisableCounters;
2893 $value = !$wgDisableCounters ? $pageLang->formatNum( SiteStats::views() ) : '';
2894 break;
2895 case 'currenttimestamp':
2896 $value = wfTimestamp( TS_MW, $ts );
2897 break;
2898 case 'localtimestamp':
2899 $value = $localTimestamp;
2900 break;
2901 case 'currentversion':
2902 $value = SpecialVersion::getVersion();
2903 break;
2904 case 'articlepath':
2905 return $wgArticlePath;
2906 case 'sitename':
2907 return $wgSitename;
2908 case 'server':
2909 return $wgServer;
2910 case 'servername':
2911 $serverParts = wfParseUrl( $wgServer );
2912 return $serverParts && isset( $serverParts['host'] ) ? $serverParts['host'] : $wgServer;
2913 case 'scriptpath':
2914 return $wgScriptPath;
2915 case 'stylepath':
2916 return $wgStylePath;
2917 case 'directionmark':
2918 return $pageLang->getDirMark();
2919 case 'contentlanguage':
2920 global $wgLanguageCode;
2921 return $wgLanguageCode;
2922 default:
2923 $ret = null;
2924 if ( wfRunHooks( 'ParserGetVariableValueSwitch', array( &$this, &$this->mVarCache, &$index, &$ret, &$frame ) ) ) {
2925 return $ret;
2926 } else {
2927 return null;
2928 }
2929 }
2930
2931 if ( $index ) {
2932 $this->mVarCache[$index] = $value;
2933 }
2934
2935 return $value;
2936 }
2937
2938 /**
2939 * initialise the magic variables (like CURRENTMONTHNAME) and substitution modifiers
2940 *
2941 * @private
2942 */
2943 function initialiseVariables() {
2944 wfProfileIn( __METHOD__ );
2945 $variableIDs = MagicWord::getVariableIDs();
2946 $substIDs = MagicWord::getSubstIDs();
2947
2948 $this->mVariables = new MagicWordArray( $variableIDs );
2949 $this->mSubstWords = new MagicWordArray( $substIDs );
2950 wfProfileOut( __METHOD__ );
2951 }
2952
2953 /**
2954 * Preprocess some wikitext and return the document tree.
2955 * This is the ghost of replace_variables().
2956 *
2957 * @param $text String: The text to parse
2958 * @param $flags Integer: bitwise combination of:
2959 * self::PTD_FOR_INCLUSION Handle "<noinclude>" and "<includeonly>" as if the text is being
2960 * included. Default is to assume a direct page view.
2961 *
2962 * The generated DOM tree must depend only on the input text and the flags.
2963 * The DOM tree must be the same in OT_HTML and OT_WIKI mode, to avoid a regression of bug 4899.
2964 *
2965 * Any flag added to the $flags parameter here, or any other parameter liable to cause a
2966 * change in the DOM tree for a given text, must be passed through the section identifier
2967 * in the section edit link and thus back to extractSections().
2968 *
2969 * The output of this function is currently only cached in process memory, but a persistent
2970 * cache may be implemented at a later date which takes further advantage of these strict
2971 * dependency requirements.
2972 *
2973 * @private
2974 *
2975 * @return PPNode
2976 */
2977 function preprocessToDom( $text, $flags = 0 ) {
2978 $dom = $this->getPreprocessor()->preprocessToObj( $text, $flags );
2979 return $dom;
2980 }
2981
2982 /**
2983 * Return a three-element array: leading whitespace, string contents, trailing whitespace
2984 *
2985 * @param $s string
2986 *
2987 * @return array
2988 */
2989 public static function splitWhitespace( $s ) {
2990 $ltrimmed = ltrim( $s );
2991 $w1 = substr( $s, 0, strlen( $s ) - strlen( $ltrimmed ) );
2992 $trimmed = rtrim( $ltrimmed );
2993 $diff = strlen( $ltrimmed ) - strlen( $trimmed );
2994 if ( $diff > 0 ) {
2995 $w2 = substr( $ltrimmed, -$diff );
2996 } else {
2997 $w2 = '';
2998 }
2999 return array( $w1, $trimmed, $w2 );
3000 }
3001
3002 /**
3003 * Replace magic variables, templates, and template arguments
3004 * with the appropriate text. Templates are substituted recursively,
3005 * taking care to avoid infinite loops.
3006 *
3007 * Note that the substitution depends on value of $mOutputType:
3008 * self::OT_WIKI: only {{subst:}} templates
3009 * self::OT_PREPROCESS: templates but not extension tags
3010 * self::OT_HTML: all templates and extension tags
3011 *
3012 * @param $text String the text to transform
3013 * @param $frame PPFrame Object describing the arguments passed to the template.
3014 * Arguments may also be provided as an associative array, as was the usual case before MW1.12.
3015 * Providing arguments this way may be useful for extensions wishing to perform variable replacement explicitly.
3016 * @param $argsOnly Boolean only do argument (triple-brace) expansion, not double-brace expansion
3017 * @private
3018 *
3019 * @return string
3020 */
3021 function replaceVariables( $text, $frame = false, $argsOnly = false ) {
3022 # Is there any text? Also, Prevent too big inclusions!
3023 if ( strlen( $text ) < 1 || strlen( $text ) > $this->mOptions->getMaxIncludeSize() ) {
3024 return $text;
3025 }
3026 wfProfileIn( __METHOD__ );
3027
3028 if ( $frame === false ) {
3029 $frame = $this->getPreprocessor()->newFrame();
3030 } elseif ( !( $frame instanceof PPFrame ) ) {
3031 wfDebug( __METHOD__." called using plain parameters instead of a PPFrame instance. Creating custom frame.\n" );
3032 $frame = $this->getPreprocessor()->newCustomFrame( $frame );
3033 }
3034
3035 $dom = $this->preprocessToDom( $text );
3036 $flags = $argsOnly ? PPFrame::NO_TEMPLATES : 0;
3037 $text = $frame->expand( $dom, $flags );
3038
3039 wfProfileOut( __METHOD__ );
3040 return $text;
3041 }
3042
3043 /**
3044 * Clean up argument array - refactored in 1.9 so parserfunctions can use it, too.
3045 *
3046 * @param $args array
3047 *
3048 * @return array
3049 */
3050 static function createAssocArgs( $args ) {
3051 $assocArgs = array();
3052 $index = 1;
3053 foreach ( $args as $arg ) {
3054 $eqpos = strpos( $arg, '=' );
3055 if ( $eqpos === false ) {
3056 $assocArgs[$index++] = $arg;
3057 } else {
3058 $name = trim( substr( $arg, 0, $eqpos ) );
3059 $value = trim( substr( $arg, $eqpos+1 ) );
3060 if ( $value === false ) {
3061 $value = '';
3062 }
3063 if ( $name !== false ) {
3064 $assocArgs[$name] = $value;
3065 }
3066 }
3067 }
3068
3069 return $assocArgs;
3070 }
3071
3072 /**
3073 * Warn the user when a parser limitation is reached
3074 * Will warn at most once the user per limitation type
3075 *
3076 * @param $limitationType String: should be one of:
3077 * 'expensive-parserfunction' (corresponding messages:
3078 * 'expensive-parserfunction-warning',
3079 * 'expensive-parserfunction-category')
3080 * 'post-expand-template-argument' (corresponding messages:
3081 * 'post-expand-template-argument-warning',
3082 * 'post-expand-template-argument-category')
3083 * 'post-expand-template-inclusion' (corresponding messages:
3084 * 'post-expand-template-inclusion-warning',
3085 * 'post-expand-template-inclusion-category')
3086 * @param $current int|null Current value
3087 * @param $max int|null Maximum allowed, when an explicit limit has been
3088 * exceeded, provide the values (optional)
3089 */
3090 function limitationWarn( $limitationType, $current = '', $max = '' ) {
3091 # does no harm if $current and $max are present but are unnecessary for the message
3092 $warning = wfMessage( "$limitationType-warning" )->numParams( $current, $max )
3093 ->inContentLanguage()->escaped();
3094 $this->mOutput->addWarning( $warning );
3095 $this->addTrackingCategory( "$limitationType-category" );
3096 }
3097
3098 /**
3099 * Return the text of a template, after recursively
3100 * replacing any variables or templates within the template.
3101 *
3102 * @param $piece Array: the parts of the template
3103 * $piece['title']: the title, i.e. the part before the |
3104 * $piece['parts']: the parameter array
3105 * $piece['lineStart']: whether the brace was at the start of a line
3106 * @param $frame PPFrame The current frame, contains template arguments
3107 * @return String: the text of the template
3108 * @private
3109 */
3110 function braceSubstitution( $piece, $frame ) {
3111 global $wgContLang;
3112 wfProfileIn( __METHOD__ );
3113 wfProfileIn( __METHOD__.'-setup' );
3114
3115 # Flags
3116 $found = false; # $text has been filled
3117 $nowiki = false; # wiki markup in $text should be escaped
3118 $isHTML = false; # $text is HTML, armour it against wikitext transformation
3119 $forceRawInterwiki = false; # Force interwiki transclusion to be done in raw mode not rendered
3120 $isChildObj = false; # $text is a DOM node needing expansion in a child frame
3121 $isLocalObj = false; # $text is a DOM node needing expansion in the current frame
3122
3123 # Title object, where $text came from
3124 $title = false;
3125
3126 # $part1 is the bit before the first |, and must contain only title characters.
3127 # Various prefixes will be stripped from it later.
3128 $titleWithSpaces = $frame->expand( $piece['title'] );
3129 $part1 = trim( $titleWithSpaces );
3130 $titleText = false;
3131
3132 # Original title text preserved for various purposes
3133 $originalTitle = $part1;
3134
3135 # $args is a list of argument nodes, starting from index 0, not including $part1
3136 # @todo FIXME: If piece['parts'] is null then the call to getLength() below won't work b/c this $args isn't an object
3137 $args = ( null == $piece['parts'] ) ? array() : $piece['parts'];
3138 wfProfileOut( __METHOD__.'-setup' );
3139
3140 $titleProfileIn = null; // profile templates
3141
3142 # SUBST
3143 wfProfileIn( __METHOD__.'-modifiers' );
3144 if ( !$found ) {
3145
3146 $substMatch = $this->mSubstWords->matchStartAndRemove( $part1 );
3147
3148 # Possibilities for substMatch: "subst", "safesubst" or FALSE
3149 # Decide whether to expand template or keep wikitext as-is.
3150 if ( $this->ot['wiki'] ) {
3151 if ( $substMatch === false ) {
3152 $literal = true; # literal when in PST with no prefix
3153 } else {
3154 $literal = false; # expand when in PST with subst: or safesubst:
3155 }
3156 } else {
3157 if ( $substMatch == 'subst' ) {
3158 $literal = true; # literal when not in PST with plain subst:
3159 } else {
3160 $literal = false; # expand when not in PST with safesubst: or no prefix
3161 }
3162 }
3163 if ( $literal ) {
3164 $text = $frame->virtualBracketedImplode( '{{', '|', '}}', $titleWithSpaces, $args );
3165 $isLocalObj = true;
3166 $found = true;
3167 }
3168 }
3169
3170 # Variables
3171 if ( !$found && $args->getLength() == 0 ) {
3172 $id = $this->mVariables->matchStartToEnd( $part1 );
3173 if ( $id !== false ) {
3174 $text = $this->getVariableValue( $id, $frame );
3175 if ( MagicWord::getCacheTTL( $id ) > -1 ) {
3176 $this->mOutput->updateCacheExpiry( MagicWord::getCacheTTL( $id ) );
3177 }
3178 $found = true;
3179 }
3180 }
3181
3182 # MSG, MSGNW and RAW
3183 if ( !$found ) {
3184 # Check for MSGNW:
3185 $mwMsgnw = MagicWord::get( 'msgnw' );
3186 if ( $mwMsgnw->matchStartAndRemove( $part1 ) ) {
3187 $nowiki = true;
3188 } else {
3189 # Remove obsolete MSG:
3190 $mwMsg = MagicWord::get( 'msg' );
3191 $mwMsg->matchStartAndRemove( $part1 );
3192 }
3193
3194 # Check for RAW:
3195 $mwRaw = MagicWord::get( 'raw' );
3196 if ( $mwRaw->matchStartAndRemove( $part1 ) ) {
3197 $forceRawInterwiki = true;
3198 }
3199 }
3200 wfProfileOut( __METHOD__.'-modifiers' );
3201
3202 # Parser functions
3203 if ( !$found ) {
3204 wfProfileIn( __METHOD__ . '-pfunc' );
3205
3206 $colonPos = strpos( $part1, ':' );
3207 if ( $colonPos !== false ) {
3208 # Case sensitive functions
3209 $function = substr( $part1, 0, $colonPos );
3210 if ( isset( $this->mFunctionSynonyms[1][$function] ) ) {
3211 $function = $this->mFunctionSynonyms[1][$function];
3212 } else {
3213 # Case insensitive functions
3214 $function = $wgContLang->lc( $function );
3215 if ( isset( $this->mFunctionSynonyms[0][$function] ) ) {
3216 $function = $this->mFunctionSynonyms[0][$function];
3217 } else {
3218 $function = false;
3219 }
3220 }
3221 if ( $function ) {
3222 wfProfileIn( __METHOD__ . '-pfunc-' . $function );
3223 list( $callback, $flags ) = $this->mFunctionHooks[$function];
3224 $initialArgs = array( &$this );
3225 $funcArgs = array( trim( substr( $part1, $colonPos + 1 ) ) );
3226 if ( $flags & SFH_OBJECT_ARGS ) {
3227 # Add a frame parameter, and pass the arguments as an array
3228 $allArgs = $initialArgs;
3229 $allArgs[] = $frame;
3230 for ( $i = 0; $i < $args->getLength(); $i++ ) {
3231 $funcArgs[] = $args->item( $i );
3232 }
3233 $allArgs[] = $funcArgs;
3234 } else {
3235 # Convert arguments to plain text
3236 for ( $i = 0; $i < $args->getLength(); $i++ ) {
3237 $funcArgs[] = trim( $frame->expand( $args->item( $i ) ) );
3238 }
3239 $allArgs = array_merge( $initialArgs, $funcArgs );
3240 }
3241
3242 # Workaround for PHP bug 35229 and similar
3243 if ( !is_callable( $callback ) ) {
3244 wfProfileOut( __METHOD__ . '-pfunc-' . $function );
3245 wfProfileOut( __METHOD__ . '-pfunc' );
3246 wfProfileOut( __METHOD__ );
3247 throw new MWException( "Tag hook for $function is not callable\n" );
3248 }
3249 $result = call_user_func_array( $callback, $allArgs );
3250 $found = true;
3251 $noparse = true;
3252 $preprocessFlags = 0;
3253
3254 if ( is_array( $result ) ) {
3255 if ( isset( $result[0] ) ) {
3256 $text = $result[0];
3257 unset( $result[0] );
3258 }
3259
3260 # Extract flags into the local scope
3261 # This allows callers to set flags such as nowiki, found, etc.
3262 extract( $result );
3263 } else {
3264 $text = $result;
3265 }
3266 if ( !$noparse ) {
3267 $text = $this->preprocessToDom( $text, $preprocessFlags );
3268 $isChildObj = true;
3269 }
3270 wfProfileOut( __METHOD__ . '-pfunc-' . $function );
3271 }
3272 }
3273 wfProfileOut( __METHOD__ . '-pfunc' );
3274 }
3275
3276 # Finish mangling title and then check for loops.
3277 # Set $title to a Title object and $titleText to the PDBK
3278 if ( !$found ) {
3279 $ns = NS_TEMPLATE;
3280 # Split the title into page and subpage
3281 $subpage = '';
3282 $part1 = $this->maybeDoSubpageLink( $part1, $subpage );
3283 if ( $subpage !== '' ) {
3284 $ns = $this->mTitle->getNamespace();
3285 }
3286 $title = Title::newFromText( $part1, $ns );
3287 if ( $title ) {
3288 $titleText = $title->getPrefixedText();
3289 # Check for language variants if the template is not found
3290 if ( $this->getConverterLanguage()->hasVariants() && $title->getArticleID() == 0 ) {
3291 $this->getConverterLanguage()->findVariantLink( $part1, $title, true );
3292 }
3293 # Do recursion depth check
3294 $limit = $this->mOptions->getMaxTemplateDepth();
3295 if ( $frame->depth >= $limit ) {
3296 $found = true;
3297 $text = '<span class="error">'
3298 . wfMessage( 'parser-template-recursion-depth-warning' )
3299 ->numParams( $limit )->inContentLanguage()->text()
3300 . '</span>';
3301 }
3302 }
3303 }
3304
3305 # Load from database
3306 if ( !$found && $title ) {
3307 if ( !Profiler::instance()->isPersistent() ) {
3308 # Too many unique items can kill profiling DBs/collectors
3309 $titleProfileIn = __METHOD__ . "-title-" . $title->getDBKey();
3310 wfProfileIn( $titleProfileIn ); // template in
3311 }
3312 wfProfileIn( __METHOD__ . '-loadtpl' );
3313 if ( !$title->isExternal() ) {
3314 if ( $title->isSpecialPage()
3315 && $this->mOptions->getAllowSpecialInclusion()
3316 && $this->ot['html'] )
3317 {
3318 // Pass the template arguments as URL parameters.
3319 // "uselang" will have no effect since the Language object
3320 // is forced to the one defined in ParserOptions.
3321 $pageArgs = array();
3322 for ( $i = 0; $i < $args->getLength(); $i++ ) {
3323 $bits = $args->item( $i )->splitArg();
3324 if ( strval( $bits['index'] ) === '' ) {
3325 $name = trim( $frame->expand( $bits['name'], PPFrame::STRIP_COMMENTS ) );
3326 $value = trim( $frame->expand( $bits['value'] ) );
3327 $pageArgs[$name] = $value;
3328 }
3329 }
3330
3331 // Create a new context to execute the special page
3332 $context = new RequestContext;
3333 $context->setTitle( $title );
3334 $context->setRequest( new FauxRequest( $pageArgs ) );
3335 $context->setUser( $this->getUser() );
3336 $context->setLanguage( $this->mOptions->getUserLangObj() );
3337 $ret = SpecialPageFactory::capturePath( $title, $context );
3338 if ( $ret ) {
3339 $text = $context->getOutput()->getHTML();
3340 $this->mOutput->addOutputPageMetadata( $context->getOutput() );
3341 $found = true;
3342 $isHTML = true;
3343 $this->disableCache();
3344 }
3345 } elseif ( MWNamespace::isNonincludable( $title->getNamespace() ) ) {
3346 $found = false; # access denied
3347 wfDebug( __METHOD__.": template inclusion denied for " . $title->getPrefixedDBkey() );
3348 } else {
3349 list( $text, $title ) = $this->getTemplateDom( $title );
3350 if ( $text !== false ) {
3351 $found = true;
3352 $isChildObj = true;
3353 }
3354 }
3355
3356 # If the title is valid but undisplayable, make a link to it
3357 if ( !$found && ( $this->ot['html'] || $this->ot['pre'] ) ) {
3358 $text = "[[:$titleText]]";
3359 $found = true;
3360 }
3361 } elseif ( $title->isTrans() ) {
3362 # Interwiki transclusion
3363 if ( $this->ot['html'] && !$forceRawInterwiki ) {
3364 $text = $this->interwikiTransclude( $title, 'render' );
3365 $isHTML = true;
3366 } else {
3367 $text = $this->interwikiTransclude( $title, 'raw' );
3368 # Preprocess it like a template
3369 $text = $this->preprocessToDom( $text, self::PTD_FOR_INCLUSION );
3370 $isChildObj = true;
3371 }
3372 $found = true;
3373 }
3374
3375 # Do infinite loop check
3376 # This has to be done after redirect resolution to avoid infinite loops via redirects
3377 if ( !$frame->loopCheck( $title ) ) {
3378 $found = true;
3379 $text = '<span class="error">'
3380 . wfMessage( 'parser-template-loop-warning', $titleText )->inContentLanguage()->text()
3381 . '</span>';
3382 wfDebug( __METHOD__.": template loop broken at '$titleText'\n" );
3383 }
3384 wfProfileOut( __METHOD__ . '-loadtpl' );
3385 }
3386
3387 # If we haven't found text to substitute by now, we're done
3388 # Recover the source wikitext and return it
3389 if ( !$found ) {
3390 $text = $frame->virtualBracketedImplode( '{{', '|', '}}', $titleWithSpaces, $args );
3391 if ( $titleProfileIn ) {
3392 wfProfileOut( $titleProfileIn ); // template out
3393 }
3394 wfProfileOut( __METHOD__ );
3395 return array( 'object' => $text );
3396 }
3397
3398 # Expand DOM-style return values in a child frame
3399 if ( $isChildObj ) {
3400 # Clean up argument array
3401 $newFrame = $frame->newChild( $args, $title );
3402
3403 if ( $nowiki ) {
3404 $text = $newFrame->expand( $text, PPFrame::RECOVER_ORIG );
3405 } elseif ( $titleText !== false && $newFrame->isEmpty() ) {
3406 # Expansion is eligible for the empty-frame cache
3407 if ( isset( $this->mTplExpandCache[$titleText] ) ) {
3408 $text = $this->mTplExpandCache[$titleText];
3409 } else {
3410 $text = $newFrame->expand( $text );
3411 $this->mTplExpandCache[$titleText] = $text;
3412 }
3413 } else {
3414 # Uncached expansion
3415 $text = $newFrame->expand( $text );
3416 }
3417 }
3418 if ( $isLocalObj && $nowiki ) {
3419 $text = $frame->expand( $text, PPFrame::RECOVER_ORIG );
3420 $isLocalObj = false;
3421 }
3422
3423 if ( $titleProfileIn ) {
3424 wfProfileOut( $titleProfileIn ); // template out
3425 }
3426
3427 # Replace raw HTML by a placeholder
3428 if ( $isHTML ) {
3429 $text = $this->insertStripItem( $text );
3430 } elseif ( $nowiki && ( $this->ot['html'] || $this->ot['pre'] ) ) {
3431 # Escape nowiki-style return values
3432 $text = wfEscapeWikiText( $text );
3433 } elseif ( is_string( $text )
3434 && !$piece['lineStart']
3435 && preg_match( '/^(?:{\\||:|;|#|\*)/', $text ) )
3436 {
3437 # Bug 529: if the template begins with a table or block-level
3438 # element, it should be treated as beginning a new line.
3439 # This behaviour is somewhat controversial.
3440 $text = "\n" . $text;
3441 }
3442
3443 if ( is_string( $text ) && !$this->incrementIncludeSize( 'post-expand', strlen( $text ) ) ) {
3444 # Error, oversize inclusion
3445 if ( $titleText !== false ) {
3446 # Make a working, properly escaped link if possible (bug 23588)
3447 $text = "[[:$titleText]]";
3448 } else {
3449 # This will probably not be a working link, but at least it may
3450 # provide some hint of where the problem is
3451 preg_replace( '/^:/', '', $originalTitle );
3452 $text = "[[:$originalTitle]]";
3453 }
3454 $text .= $this->insertStripItem( '<!-- WARNING: template omitted, post-expand include size too large -->' );
3455 $this->limitationWarn( 'post-expand-template-inclusion' );
3456 }
3457
3458 if ( $isLocalObj ) {
3459 $ret = array( 'object' => $text );
3460 } else {
3461 $ret = array( 'text' => $text );
3462 }
3463
3464 wfProfileOut( __METHOD__ );
3465 return $ret;
3466 }
3467
3468 /**
3469 * Get the semi-parsed DOM representation of a template with a given title,
3470 * and its redirect destination title. Cached.
3471 *
3472 * @param $title Title
3473 *
3474 * @return array
3475 */
3476 function getTemplateDom( $title ) {
3477 $cacheTitle = $title;
3478 $titleText = $title->getPrefixedDBkey();
3479
3480 if ( isset( $this->mTplRedirCache[$titleText] ) ) {
3481 list( $ns, $dbk ) = $this->mTplRedirCache[$titleText];
3482 $title = Title::makeTitle( $ns, $dbk );
3483 $titleText = $title->getPrefixedDBkey();
3484 }
3485 if ( isset( $this->mTplDomCache[$titleText] ) ) {
3486 return array( $this->mTplDomCache[$titleText], $title );
3487 }
3488
3489 # Cache miss, go to the database
3490 list( $text, $title ) = $this->fetchTemplateAndTitle( $title );
3491
3492 if ( $text === false ) {
3493 $this->mTplDomCache[$titleText] = false;
3494 return array( false, $title );
3495 }
3496
3497 $dom = $this->preprocessToDom( $text, self::PTD_FOR_INCLUSION );
3498 $this->mTplDomCache[ $titleText ] = $dom;
3499
3500 if ( !$title->equals( $cacheTitle ) ) {
3501 $this->mTplRedirCache[$cacheTitle->getPrefixedDBkey()] =
3502 array( $title->getNamespace(), $cdb = $title->getDBkey() );
3503 }
3504
3505 return array( $dom, $title );
3506 }
3507
3508 /**
3509 * Fetch the unparsed text of a template and register a reference to it.
3510 * @param Title $title
3511 * @return Array ( string or false, Title )
3512 */
3513 function fetchTemplateAndTitle( $title ) {
3514 $templateCb = $this->mOptions->getTemplateCallback(); # Defaults to Parser::statelessFetchTemplate()
3515 $stuff = call_user_func( $templateCb, $title, $this );
3516 $text = $stuff['text'];
3517 $finalTitle = isset( $stuff['finalTitle'] ) ? $stuff['finalTitle'] : $title;
3518 if ( isset( $stuff['deps'] ) ) {
3519 foreach ( $stuff['deps'] as $dep ) {
3520 $this->mOutput->addTemplate( $dep['title'], $dep['page_id'], $dep['rev_id'] );
3521 }
3522 }
3523 return array( $text, $finalTitle );
3524 }
3525
3526 /**
3527 * Fetch the unparsed text of a template and register a reference to it.
3528 * @param Title $title
3529 * @return mixed string or false
3530 */
3531 function fetchTemplate( $title ) {
3532 $rv = $this->fetchTemplateAndTitle( $title );
3533 return $rv[0];
3534 }
3535
3536 /**
3537 * Static function to get a template
3538 * Can be overridden via ParserOptions::setTemplateCallback().
3539 *
3540 * @param $title Title
3541 * @param $parser Parser
3542 *
3543 * @return array
3544 */
3545 static function statelessFetchTemplate( $title, $parser = false ) {
3546 $text = $skip = false;
3547 $finalTitle = $title;
3548 $deps = array();
3549
3550 # Loop to fetch the article, with up to 1 redirect
3551 for ( $i = 0; $i < 2 && is_object( $title ); $i++ ) {
3552 # Give extensions a chance to select the revision instead
3553 $id = false; # Assume current
3554 wfRunHooks( 'BeforeParserFetchTemplateAndtitle',
3555 array( $parser, $title, &$skip, &$id ) );
3556
3557 if ( $skip ) {
3558 $text = false;
3559 $deps[] = array(
3560 'title' => $title,
3561 'page_id' => $title->getArticleID(),
3562 'rev_id' => null
3563 );
3564 break;
3565 }
3566 # Get the revision
3567 $rev = $id
3568 ? Revision::newFromId( $id )
3569 : Revision::newFromTitle( $title, false, Revision::READ_NORMAL );
3570 $rev_id = $rev ? $rev->getId() : 0;
3571 # If there is no current revision, there is no page
3572 if ( $id === false && !$rev ) {
3573 $linkCache = LinkCache::singleton();
3574 $linkCache->addBadLinkObj( $title );
3575 }
3576
3577 $deps[] = array(
3578 'title' => $title,
3579 'page_id' => $title->getArticleID(),
3580 'rev_id' => $rev_id );
3581 if ( $rev && !$title->equals( $rev->getTitle() ) ) {
3582 # We fetched a rev from a different title; register it too...
3583 $deps[] = array(
3584 'title' => $rev->getTitle(),
3585 'page_id' => $rev->getPage(),
3586 'rev_id' => $rev_id );
3587 }
3588
3589 if ( $rev ) {
3590 $content = $rev->getContent();
3591 $text = $content->getWikitextForTransclusion();
3592
3593 if ( $text === false || $text === null ) {
3594 $text = false;
3595 break;
3596 }
3597 } elseif ( $title->getNamespace() == NS_MEDIAWIKI ) {
3598 global $wgContLang;
3599 $message = wfMessage( $wgContLang->lcfirst( $title->getText() ) )->inContentLanguage();
3600 if ( !$message->exists() ) {
3601 $text = false;
3602 break;
3603 }
3604 $content = $message->content();
3605 $text = $message->plain();
3606 } else {
3607 break;
3608 }
3609 if ( !$content ) {
3610 break;
3611 }
3612 # Redirect?
3613 $finalTitle = $title;
3614 $title = $content->getRedirectTarget();
3615 }
3616 return array(
3617 'text' => $text,
3618 'finalTitle' => $finalTitle,
3619 'deps' => $deps );
3620 }
3621
3622 /**
3623 * Fetch a file and its title and register a reference to it.
3624 * If 'broken' is a key in $options then the file will appear as a broken thumbnail.
3625 * @param Title $title
3626 * @param Array $options Array of options to RepoGroup::findFile
3627 * @return File|bool
3628 */
3629 function fetchFile( $title, $options = array() ) {
3630 $res = $this->fetchFileAndTitle( $title, $options );
3631 return $res[0];
3632 }
3633
3634 /**
3635 * Fetch a file and its title and register a reference to it.
3636 * If 'broken' is a key in $options then the file will appear as a broken thumbnail.
3637 * @param Title $title
3638 * @param Array $options Array of options to RepoGroup::findFile
3639 * @return Array ( File or false, Title of file )
3640 */
3641 function fetchFileAndTitle( $title, $options = array() ) {
3642 if ( isset( $options['broken'] ) ) {
3643 $file = false; // broken thumbnail forced by hook
3644 } elseif ( isset( $options['sha1'] ) ) { // get by (sha1,timestamp)
3645 $file = RepoGroup::singleton()->findFileFromKey( $options['sha1'], $options );
3646 } else { // get by (name,timestamp)
3647 $file = wfFindFile( $title, $options );
3648 }
3649 $time = $file ? $file->getTimestamp() : false;
3650 $sha1 = $file ? $file->getSha1() : false;
3651 # Register the file as a dependency...
3652 $this->mOutput->addImage( $title->getDBkey(), $time, $sha1 );
3653 if ( $file && !$title->equals( $file->getTitle() ) ) {
3654 # Update fetched file title
3655 $title = $file->getTitle();
3656 if ( is_null( $file->getRedirectedTitle() ) ) {
3657 # This file was not a redirect, but the title does not match.
3658 # Register under the new name because otherwise the link will
3659 # get lost.
3660 $this->mOutput->addImage( $title->getDBkey(), $time, $sha1 );
3661 }
3662 }
3663 return array( $file, $title );
3664 }
3665
3666 /**
3667 * Transclude an interwiki link.
3668 *
3669 * @param $title Title
3670 * @param $action
3671 *
3672 * @return string
3673 */
3674 function interwikiTransclude( $title, $action ) {
3675 global $wgEnableScaryTranscluding;
3676
3677 if ( !$wgEnableScaryTranscluding ) {
3678 return wfMessage('scarytranscludedisabled')->inContentLanguage()->text();
3679 }
3680
3681 $url = $title->getFullUrl( "action=$action" );
3682
3683 if ( strlen( $url ) > 255 ) {
3684 return wfMessage( 'scarytranscludetoolong' )->inContentLanguage()->text();
3685 }
3686 return $this->fetchScaryTemplateMaybeFromCache( $url );
3687 }
3688
3689 /**
3690 * @param $url string
3691 * @return Mixed|String
3692 */
3693 function fetchScaryTemplateMaybeFromCache( $url ) {
3694 global $wgTranscludeCacheExpiry;
3695 $dbr = wfGetDB( DB_SLAVE );
3696 $tsCond = $dbr->timestamp( time() - $wgTranscludeCacheExpiry );
3697 $obj = $dbr->selectRow( 'transcache', array('tc_time', 'tc_contents' ),
3698 array( 'tc_url' => $url, "tc_time >= " . $dbr->addQuotes( $tsCond ) ) );
3699 if ( $obj ) {
3700 return $obj->tc_contents;
3701 }
3702
3703 $text = Http::get( $url );
3704 if ( !$text ) {
3705 return wfMessage( 'scarytranscludefailed', $url )->inContentLanguage()->text();
3706 }
3707
3708 $dbw = wfGetDB( DB_MASTER );
3709 $dbw->replace( 'transcache', array('tc_url'), array(
3710 'tc_url' => $url,
3711 'tc_time' => $dbw->timestamp( time() ),
3712 'tc_contents' => $text)
3713 );
3714 return $text;
3715 }
3716
3717 /**
3718 * Triple brace replacement -- used for template arguments
3719 * @private
3720 *
3721 * @param $piece array
3722 * @param $frame PPFrame
3723 *
3724 * @return array
3725 */
3726 function argSubstitution( $piece, $frame ) {
3727 wfProfileIn( __METHOD__ );
3728
3729 $error = false;
3730 $parts = $piece['parts'];
3731 $nameWithSpaces = $frame->expand( $piece['title'] );
3732 $argName = trim( $nameWithSpaces );
3733 $object = false;
3734 $text = $frame->getArgument( $argName );
3735 if ( $text === false && $parts->getLength() > 0
3736 && (
3737 $this->ot['html']
3738 || $this->ot['pre']
3739 || ( $this->ot['wiki'] && $frame->isTemplate() )
3740 )
3741 ) {
3742 # No match in frame, use the supplied default
3743 $object = $parts->item( 0 )->getChildren();
3744 }
3745 if ( !$this->incrementIncludeSize( 'arg', strlen( $text ) ) ) {
3746 $error = '<!-- WARNING: argument omitted, expansion size too large -->';
3747 $this->limitationWarn( 'post-expand-template-argument' );
3748 }
3749
3750 if ( $text === false && $object === false ) {
3751 # No match anywhere
3752 $object = $frame->virtualBracketedImplode( '{{{', '|', '}}}', $nameWithSpaces, $parts );
3753 }
3754 if ( $error !== false ) {
3755 $text .= $error;
3756 }
3757 if ( $object !== false ) {
3758 $ret = array( 'object' => $object );
3759 } else {
3760 $ret = array( 'text' => $text );
3761 }
3762
3763 wfProfileOut( __METHOD__ );
3764 return $ret;
3765 }
3766
3767 /**
3768 * Return the text to be used for a given extension tag.
3769 * This is the ghost of strip().
3770 *
3771 * @param $params array Associative array of parameters:
3772 * name PPNode for the tag name
3773 * attr PPNode for unparsed text where tag attributes are thought to be
3774 * attributes Optional associative array of parsed attributes
3775 * inner Contents of extension element
3776 * noClose Original text did not have a close tag
3777 * @param $frame PPFrame
3778 *
3779 * @return string
3780 */
3781 function extensionSubstitution( $params, $frame ) {
3782 $name = $frame->expand( $params['name'] );
3783 $attrText = !isset( $params['attr'] ) ? null : $frame->expand( $params['attr'] );
3784 $content = !isset( $params['inner'] ) ? null : $frame->expand( $params['inner'] );
3785 $marker = "{$this->mUniqPrefix}-$name-" . sprintf( '%08X', $this->mMarkerIndex++ ) . self::MARKER_SUFFIX;
3786
3787 $isFunctionTag = isset( $this->mFunctionTagHooks[strtolower($name)] ) &&
3788 ( $this->ot['html'] || $this->ot['pre'] );
3789 if ( $isFunctionTag ) {
3790 $markerType = 'none';
3791 } else {
3792 $markerType = 'general';
3793 }
3794 if ( $this->ot['html'] || $isFunctionTag ) {
3795 $name = strtolower( $name );
3796 $attributes = Sanitizer::decodeTagAttributes( $attrText );
3797 if ( isset( $params['attributes'] ) ) {
3798 $attributes = $attributes + $params['attributes'];
3799 }
3800
3801 if ( isset( $this->mTagHooks[$name] ) ) {
3802 # Workaround for PHP bug 35229 and similar
3803 if ( !is_callable( $this->mTagHooks[$name] ) ) {
3804 throw new MWException( "Tag hook for $name is not callable\n" );
3805 }
3806 $output = call_user_func_array( $this->mTagHooks[$name],
3807 array( $content, $attributes, $this, $frame ) );
3808 } elseif ( isset( $this->mFunctionTagHooks[$name] ) ) {
3809 list( $callback, $flags ) = $this->mFunctionTagHooks[$name];
3810 if ( !is_callable( $callback ) ) {
3811 throw new MWException( "Tag hook for $name is not callable\n" );
3812 }
3813
3814 $output = call_user_func_array( $callback, array( &$this, $frame, $content, $attributes ) );
3815 } else {
3816 $output = '<span class="error">Invalid tag extension name: ' .
3817 htmlspecialchars( $name ) . '</span>';
3818 }
3819
3820 if ( is_array( $output ) ) {
3821 # Extract flags to local scope (to override $markerType)
3822 $flags = $output;
3823 $output = $flags[0];
3824 unset( $flags[0] );
3825 extract( $flags );
3826 }
3827 } else {
3828 if ( is_null( $attrText ) ) {
3829 $attrText = '';
3830 }
3831 if ( isset( $params['attributes'] ) ) {
3832 foreach ( $params['attributes'] as $attrName => $attrValue ) {
3833 $attrText .= ' ' . htmlspecialchars( $attrName ) . '="' .
3834 htmlspecialchars( $attrValue ) . '"';
3835 }
3836 }
3837 if ( $content === null ) {
3838 $output = "<$name$attrText/>";
3839 } else {
3840 $close = is_null( $params['close'] ) ? '' : $frame->expand( $params['close'] );
3841 $output = "<$name$attrText>$content$close";
3842 }
3843 }
3844
3845 if ( $markerType === 'none' ) {
3846 return $output;
3847 } elseif ( $markerType === 'nowiki' ) {
3848 $this->mStripState->addNoWiki( $marker, $output );
3849 } elseif ( $markerType === 'general' ) {
3850 $this->mStripState->addGeneral( $marker, $output );
3851 } else {
3852 throw new MWException( __METHOD__.': invalid marker type' );
3853 }
3854 return $marker;
3855 }
3856
3857 /**
3858 * Increment an include size counter
3859 *
3860 * @param $type String: the type of expansion
3861 * @param $size Integer: the size of the text
3862 * @return Boolean: false if this inclusion would take it over the maximum, true otherwise
3863 */
3864 function incrementIncludeSize( $type, $size ) {
3865 if ( $this->mIncludeSizes[$type] + $size > $this->mOptions->getMaxIncludeSize() ) {
3866 return false;
3867 } else {
3868 $this->mIncludeSizes[$type] += $size;
3869 return true;
3870 }
3871 }
3872
3873 /**
3874 * Increment the expensive function count
3875 *
3876 * @return Boolean: false if the limit has been exceeded
3877 */
3878 function incrementExpensiveFunctionCount() {
3879 $this->mExpensiveFunctionCount++;
3880 return $this->mExpensiveFunctionCount <= $this->mOptions->getExpensiveParserFunctionLimit();
3881 }
3882
3883 /**
3884 * Strip double-underscore items like __NOGALLERY__ and __NOTOC__
3885 * Fills $this->mDoubleUnderscores, returns the modified text
3886 *
3887 * @param $text string
3888 *
3889 * @return string
3890 */
3891 function doDoubleUnderscore( $text ) {
3892 wfProfileIn( __METHOD__ );
3893
3894 # The position of __TOC__ needs to be recorded
3895 $mw = MagicWord::get( 'toc' );
3896 if ( $mw->match( $text ) ) {
3897 $this->mShowToc = true;
3898 $this->mForceTocPosition = true;
3899
3900 # Set a placeholder. At the end we'll fill it in with the TOC.
3901 $text = $mw->replace( '<!--MWTOC-->', $text, 1 );
3902
3903 # Only keep the first one.
3904 $text = $mw->replace( '', $text );
3905 }
3906
3907 # Now match and remove the rest of them
3908 $mwa = MagicWord::getDoubleUnderscoreArray();
3909 $this->mDoubleUnderscores = $mwa->matchAndRemove( $text );
3910
3911 if ( isset( $this->mDoubleUnderscores['nogallery'] ) ) {
3912 $this->mOutput->mNoGallery = true;
3913 }
3914 if ( isset( $this->mDoubleUnderscores['notoc'] ) && !$this->mForceTocPosition ) {
3915 $this->mShowToc = false;
3916 }
3917 if ( isset( $this->mDoubleUnderscores['hiddencat'] ) && $this->mTitle->getNamespace() == NS_CATEGORY ) {
3918 $this->addTrackingCategory( 'hidden-category-category' );
3919 }
3920 # (bug 8068) Allow control over whether robots index a page.
3921 #
3922 # @todo FIXME: Bug 14899: __INDEX__ always overrides __NOINDEX__ here! This
3923 # is not desirable, the last one on the page should win.
3924 if ( isset( $this->mDoubleUnderscores['noindex'] ) && $this->mTitle->canUseNoindex() ) {
3925 $this->mOutput->setIndexPolicy( 'noindex' );
3926 $this->addTrackingCategory( 'noindex-category' );
3927 }
3928 if ( isset( $this->mDoubleUnderscores['index'] ) && $this->mTitle->canUseNoindex() ) {
3929 $this->mOutput->setIndexPolicy( 'index' );
3930 $this->addTrackingCategory( 'index-category' );
3931 }
3932
3933 # Cache all double underscores in the database
3934 foreach ( $this->mDoubleUnderscores as $key => $val ) {
3935 $this->mOutput->setProperty( $key, '' );
3936 }
3937
3938 wfProfileOut( __METHOD__ );
3939 return $text;
3940 }
3941
3942 /**
3943 * Add a tracking category, getting the title from a system message,
3944 * or print a debug message if the title is invalid.
3945 *
3946 * @param $msg String: message key
3947 * @return Boolean: whether the addition was successful
3948 */
3949 public function addTrackingCategory( $msg ) {
3950 if ( $this->mTitle->getNamespace() === NS_SPECIAL ) {
3951 wfDebug( __METHOD__.": Not adding tracking category $msg to special page!\n" );
3952 return false;
3953 }
3954 // Important to parse with correct title (bug 31469)
3955 $cat = wfMessage( $msg )
3956 ->title( $this->getTitle() )
3957 ->inContentLanguage()
3958 ->text();
3959
3960 # Allow tracking categories to be disabled by setting them to "-"
3961 if ( $cat === '-' ) {
3962 return false;
3963 }
3964
3965 $containerCategory = Title::makeTitleSafe( NS_CATEGORY, $cat );
3966 if ( $containerCategory ) {
3967 $this->mOutput->addCategory( $containerCategory->getDBkey(), $this->getDefaultSort() );
3968 return true;
3969 } else {
3970 wfDebug( __METHOD__.": [[MediaWiki:$msg]] is not a valid title!\n" );
3971 return false;
3972 }
3973 }
3974
3975 /**
3976 * This function accomplishes several tasks:
3977 * 1) Auto-number headings if that option is enabled
3978 * 2) Add an [edit] link to sections for users who have enabled the option and can edit the page
3979 * 3) Add a Table of contents on the top for users who have enabled the option
3980 * 4) Auto-anchor headings
3981 *
3982 * It loops through all headlines, collects the necessary data, then splits up the
3983 * string and re-inserts the newly formatted headlines.
3984 *
3985 * @param $text String
3986 * @param $origText String: original, untouched wikitext
3987 * @param $isMain Boolean
3988 * @return mixed|string
3989 * @private
3990 */
3991 function formatHeadings( $text, $origText, $isMain=true ) {
3992 global $wgMaxTocLevel, $wgHtml5, $wgExperimentalHtmlIds;
3993
3994 # Inhibit editsection links if requested in the page
3995 if ( isset( $this->mDoubleUnderscores['noeditsection'] ) ) {
3996 $maybeShowEditLink = $showEditLink = false;
3997 } else {
3998 $maybeShowEditLink = true; /* Actual presence will depend on ParserOptions option */
3999 $showEditLink = $this->mOptions->getEditSection();
4000 }
4001 if ( $showEditLink ) {
4002 $this->mOutput->setEditSectionTokens( true );
4003 }
4004
4005 # Get all headlines for numbering them and adding funky stuff like [edit]
4006 # links - this is for later, but we need the number of headlines right now
4007 $matches = array();
4008 $numMatches = preg_match_all( '/<H(?P<level>[1-6])(?P<attrib>.*?'.'>)(?P<header>.*?)<\/H[1-6] *>/i', $text, $matches );
4009
4010 # if there are fewer than 4 headlines in the article, do not show TOC
4011 # unless it's been explicitly enabled.
4012 $enoughToc = $this->mShowToc &&
4013 ( ( $numMatches >= 4 ) || $this->mForceTocPosition );
4014
4015 # Allow user to stipulate that a page should have a "new section"
4016 # link added via __NEWSECTIONLINK__
4017 if ( isset( $this->mDoubleUnderscores['newsectionlink'] ) ) {
4018 $this->mOutput->setNewSection( true );
4019 }
4020
4021 # Allow user to remove the "new section"
4022 # link via __NONEWSECTIONLINK__
4023 if ( isset( $this->mDoubleUnderscores['nonewsectionlink'] ) ) {
4024 $this->mOutput->hideNewSection( true );
4025 }
4026
4027 # if the string __FORCETOC__ (not case-sensitive) occurs in the HTML,
4028 # override above conditions and always show TOC above first header
4029 if ( isset( $this->mDoubleUnderscores['forcetoc'] ) ) {
4030 $this->mShowToc = true;
4031 $enoughToc = true;
4032 }
4033
4034 # headline counter
4035 $headlineCount = 0;
4036 $numVisible = 0;
4037
4038 # Ugh .. the TOC should have neat indentation levels which can be
4039 # passed to the skin functions. These are determined here
4040 $toc = '';
4041 $full = '';
4042 $head = array();
4043 $sublevelCount = array();
4044 $levelCount = array();
4045 $level = 0;
4046 $prevlevel = 0;
4047 $toclevel = 0;
4048 $prevtoclevel = 0;
4049 $markerRegex = "{$this->mUniqPrefix}-h-(\d+)-" . self::MARKER_SUFFIX;
4050 $baseTitleText = $this->mTitle->getPrefixedDBkey();
4051 $oldType = $this->mOutputType;
4052 $this->setOutputType( self::OT_WIKI );
4053 $frame = $this->getPreprocessor()->newFrame();
4054 $root = $this->preprocessToDom( $origText );
4055 $node = $root->getFirstChild();
4056 $byteOffset = 0;
4057 $tocraw = array();
4058 $refers = array();
4059
4060 foreach ( $matches[3] as $headline ) {
4061 $isTemplate = false;
4062 $titleText = false;
4063 $sectionIndex = false;
4064 $numbering = '';
4065 $markerMatches = array();
4066 if ( preg_match("/^$markerRegex/", $headline, $markerMatches ) ) {
4067 $serial = $markerMatches[1];
4068 list( $titleText, $sectionIndex ) = $this->mHeadings[$serial];
4069 $isTemplate = ( $titleText != $baseTitleText );
4070 $headline = preg_replace( "/^$markerRegex/", "", $headline );
4071 }
4072
4073 if ( $toclevel ) {
4074 $prevlevel = $level;
4075 }
4076 $level = $matches[1][$headlineCount];
4077
4078 if ( $level > $prevlevel ) {
4079 # Increase TOC level
4080 $toclevel++;
4081 $sublevelCount[$toclevel] = 0;
4082 if ( $toclevel<$wgMaxTocLevel ) {
4083 $prevtoclevel = $toclevel;
4084 $toc .= Linker::tocIndent();
4085 $numVisible++;
4086 }
4087 } elseif ( $level < $prevlevel && $toclevel > 1 ) {
4088 # Decrease TOC level, find level to jump to
4089
4090 for ( $i = $toclevel; $i > 0; $i-- ) {
4091 if ( $levelCount[$i] == $level ) {
4092 # Found last matching level
4093 $toclevel = $i;
4094 break;
4095 } elseif ( $levelCount[$i] < $level ) {
4096 # Found first matching level below current level
4097 $toclevel = $i + 1;
4098 break;
4099 }
4100 }
4101 if ( $i == 0 ) {
4102 $toclevel = 1;
4103 }
4104 if ( $toclevel<$wgMaxTocLevel ) {
4105 if ( $prevtoclevel < $wgMaxTocLevel ) {
4106 # Unindent only if the previous toc level was shown :p
4107 $toc .= Linker::tocUnindent( $prevtoclevel - $toclevel );
4108 $prevtoclevel = $toclevel;
4109 } else {
4110 $toc .= Linker::tocLineEnd();
4111 }
4112 }
4113 } else {
4114 # No change in level, end TOC line
4115 if ( $toclevel<$wgMaxTocLevel ) {
4116 $toc .= Linker::tocLineEnd();
4117 }
4118 }
4119
4120 $levelCount[$toclevel] = $level;
4121
4122 # count number of headlines for each level
4123 @$sublevelCount[$toclevel]++;
4124 $dot = 0;
4125 for( $i = 1; $i <= $toclevel; $i++ ) {
4126 if ( !empty( $sublevelCount[$i] ) ) {
4127 if ( $dot ) {
4128 $numbering .= '.';
4129 }
4130 $numbering .= $this->getTargetLanguage()->formatNum( $sublevelCount[$i] );
4131 $dot = 1;
4132 }
4133 }
4134
4135 # The safe header is a version of the header text safe to use for links
4136
4137 # Remove link placeholders by the link text.
4138 # <!--LINK number-->
4139 # turns into
4140 # link text with suffix
4141 # Do this before unstrip since link text can contain strip markers
4142 $safeHeadline = $this->replaceLinkHoldersText( $headline );
4143
4144 # Avoid insertion of weird stuff like <math> by expanding the relevant sections
4145 $safeHeadline = $this->mStripState->unstripBoth( $safeHeadline );
4146
4147 # Strip out HTML (first regex removes any tag not allowed)
4148 # Allowed tags are <sup> and <sub> (bug 8393), <i> (bug 26375) and <b> (r105284)
4149 # We strip any parameter from accepted tags (second regex)
4150 $tocline = preg_replace(
4151 array( '#<(?!/?(sup|sub|i|b)(?: [^>]*)?>).*?'.'>#', '#<(/?(sup|sub|i|b))(?: .*?)?'.'>#' ),
4152 array( '', '<$1>' ),
4153 $safeHeadline
4154 );
4155 $tocline = trim( $tocline );
4156
4157 # For the anchor, strip out HTML-y stuff period
4158 $safeHeadline = preg_replace( '/<.*?'.'>/', '', $safeHeadline );
4159 $safeHeadline = Sanitizer::normalizeSectionNameWhitespace( $safeHeadline );
4160
4161 # Save headline for section edit hint before it's escaped
4162 $headlineHint = $safeHeadline;
4163
4164 if ( $wgHtml5 && $wgExperimentalHtmlIds ) {
4165 # For reverse compatibility, provide an id that's
4166 # HTML4-compatible, like we used to.
4167 #
4168 # It may be worth noting, academically, that it's possible for
4169 # the legacy anchor to conflict with a non-legacy headline
4170 # anchor on the page. In this case likely the "correct" thing
4171 # would be to either drop the legacy anchors or make sure
4172 # they're numbered first. However, this would require people
4173 # to type in section names like "abc_.D7.93.D7.90.D7.A4"
4174 # manually, so let's not bother worrying about it.
4175 $legacyHeadline = Sanitizer::escapeId( $safeHeadline,
4176 array( 'noninitial', 'legacy' ) );
4177 $safeHeadline = Sanitizer::escapeId( $safeHeadline );
4178
4179 if ( $legacyHeadline == $safeHeadline ) {
4180 # No reason to have both (in fact, we can't)
4181 $legacyHeadline = false;
4182 }
4183 } else {
4184 $legacyHeadline = false;
4185 $safeHeadline = Sanitizer::escapeId( $safeHeadline,
4186 'noninitial' );
4187 }
4188
4189 # HTML names must be case-insensitively unique (bug 10721).
4190 # This does not apply to Unicode characters per
4191 # http://dev.w3.org/html5/spec/infrastructure.html#case-sensitivity-and-string-comparison
4192 # @todo FIXME: We may be changing them depending on the current locale.
4193 $arrayKey = strtolower( $safeHeadline );
4194 if ( $legacyHeadline === false ) {
4195 $legacyArrayKey = false;
4196 } else {
4197 $legacyArrayKey = strtolower( $legacyHeadline );
4198 }
4199
4200 # count how many in assoc. array so we can track dupes in anchors
4201 if ( isset( $refers[$arrayKey] ) ) {
4202 $refers[$arrayKey]++;
4203 } else {
4204 $refers[$arrayKey] = 1;
4205 }
4206 if ( isset( $refers[$legacyArrayKey] ) ) {
4207 $refers[$legacyArrayKey]++;
4208 } else {
4209 $refers[$legacyArrayKey] = 1;
4210 }
4211
4212 # Don't number the heading if it is the only one (looks silly)
4213 if ( count( $matches[3] ) > 1 && $this->mOptions->getNumberHeadings() ) {
4214 # the two are different if the line contains a link
4215 $headline = Html::element( 'span', array( 'class' => 'mw-headline-number' ), $numbering ) . ' ' . $headline;
4216 }
4217
4218 # Create the anchor for linking from the TOC to the section
4219 $anchor = $safeHeadline;
4220 $legacyAnchor = $legacyHeadline;
4221 if ( $refers[$arrayKey] > 1 ) {
4222 $anchor .= '_' . $refers[$arrayKey];
4223 }
4224 if ( $legacyHeadline !== false && $refers[$legacyArrayKey] > 1 ) {
4225 $legacyAnchor .= '_' . $refers[$legacyArrayKey];
4226 }
4227 if ( $enoughToc && ( !isset( $wgMaxTocLevel ) || $toclevel < $wgMaxTocLevel ) ) {
4228 $toc .= Linker::tocLine( $anchor, $tocline,
4229 $numbering, $toclevel, ( $isTemplate ? false : $sectionIndex ) );
4230 }
4231
4232 # Add the section to the section tree
4233 # Find the DOM node for this header
4234 while ( $node && !$isTemplate ) {
4235 if ( $node->getName() === 'h' ) {
4236 $bits = $node->splitHeading();
4237 if ( $bits['i'] == $sectionIndex ) {
4238 break;
4239 }
4240 }
4241 $byteOffset += mb_strlen( $this->mStripState->unstripBoth(
4242 $frame->expand( $node, PPFrame::RECOVER_ORIG ) ) );
4243 $node = $node->getNextSibling();
4244 }
4245 $tocraw[] = array(
4246 'toclevel' => $toclevel,
4247 'level' => $level,
4248 'line' => $tocline,
4249 'number' => $numbering,
4250 'index' => ( $isTemplate ? 'T-' : '' ) . $sectionIndex,
4251 'fromtitle' => $titleText,
4252 'byteoffset' => ( $isTemplate ? null : $byteOffset ),
4253 'anchor' => $anchor,
4254 );
4255
4256 # give headline the correct <h#> tag
4257 if ( $maybeShowEditLink && $sectionIndex !== false ) {
4258 // Output edit section links as markers with styles that can be customized by skins
4259 if ( $isTemplate ) {
4260 # Put a T flag in the section identifier, to indicate to extractSections()
4261 # that sections inside <includeonly> should be counted.
4262 $editlinkArgs = array( $titleText, "T-$sectionIndex"/*, null */ );
4263 } else {
4264 $editlinkArgs = array( $this->mTitle->getPrefixedText(), $sectionIndex, $headlineHint );
4265 }
4266 // We use a bit of pesudo-xml for editsection markers. The language converter is run later on
4267 // Using a UNIQ style marker leads to the converter screwing up the tokens when it converts stuff
4268 // And trying to insert strip tags fails too. At this point all real inputted tags have already been escaped
4269 // so we don't have to worry about a user trying to input one of these markers directly.
4270 // We use a page and section attribute to stop the language converter from converting these important bits
4271 // of data, but put the headline hint inside a content block because the language converter is supposed to
4272 // be able to convert that piece of data.
4273 $editlink = '<mw:editsection page="' . htmlspecialchars($editlinkArgs[0]);
4274 $editlink .= '" section="' . htmlspecialchars($editlinkArgs[1]) .'"';
4275 if ( isset($editlinkArgs[2]) ) {
4276 $editlink .= '>' . $editlinkArgs[2] . '</mw:editsection>';
4277 } else {
4278 $editlink .= '/>';
4279 }
4280 } else {
4281 $editlink = '';
4282 }
4283 $head[$headlineCount] = Linker::makeHeadline( $level,
4284 $matches['attrib'][$headlineCount], $anchor, $headline,
4285 $editlink, $legacyAnchor );
4286
4287 $headlineCount++;
4288 }
4289
4290 $this->setOutputType( $oldType );
4291
4292 # Never ever show TOC if no headers
4293 if ( $numVisible < 1 ) {
4294 $enoughToc = false;
4295 }
4296
4297 if ( $enoughToc ) {
4298 if ( $prevtoclevel > 0 && $prevtoclevel < $wgMaxTocLevel ) {
4299 $toc .= Linker::tocUnindent( $prevtoclevel - 1 );
4300 }
4301 $toc = Linker::tocList( $toc, $this->mOptions->getUserLangObj() );
4302 $this->mOutput->setTOCHTML( $toc );
4303 }
4304
4305 if ( $isMain ) {
4306 $this->mOutput->setSections( $tocraw );
4307 }
4308
4309 # split up and insert constructed headlines
4310 $blocks = preg_split( '/<H[1-6].*?' . '>.*?<\/H[1-6]>/i', $text );
4311 $i = 0;
4312
4313 // build an array of document sections
4314 $sections = array();
4315 foreach ( $blocks as $block ) {
4316 // $head is zero-based, sections aren't.
4317 if ( empty( $head[$i - 1] ) ) {
4318 $sections[$i] = $block;
4319 } else {
4320 $sections[$i] = $head[$i - 1] . $block;
4321 }
4322
4323 /**
4324 * Send a hook, one per section.
4325 * The idea here is to be able to make section-level DIVs, but to do so in a
4326 * lower-impact, more correct way than r50769
4327 *
4328 * $this : caller
4329 * $section : the section number
4330 * &$sectionContent : ref to the content of the section
4331 * $showEditLinks : boolean describing whether this section has an edit link
4332 */
4333 wfRunHooks( 'ParserSectionCreate', array( $this, $i, &$sections[$i], $showEditLink ) );
4334
4335 $i++;
4336 }
4337
4338 if ( $enoughToc && $isMain && !$this->mForceTocPosition ) {
4339 // append the TOC at the beginning
4340 // Top anchor now in skin
4341 $sections[0] = $sections[0] . $toc . "\n";
4342 }
4343
4344 $full .= join( '', $sections );
4345
4346 if ( $this->mForceTocPosition ) {
4347 return str_replace( '<!--MWTOC-->', $toc, $full );
4348 } else {
4349 return $full;
4350 }
4351 }
4352
4353 /**
4354 * Transform wiki markup when saving a page by doing "\r\n" -> "\n"
4355 * conversion, substitting signatures, {{subst:}} templates, etc.
4356 *
4357 * @param $text String: the text to transform
4358 * @param $title Title: the Title object for the current article
4359 * @param $user User: the User object describing the current user
4360 * @param $options ParserOptions: parsing options
4361 * @param $clearState Boolean: whether to clear the parser state first
4362 * @return String: the altered wiki markup
4363 */
4364 public function preSaveTransform( $text, Title $title, User $user, ParserOptions $options, $clearState = true ) {
4365 $this->startParse( $title, $options, self::OT_WIKI, $clearState );
4366 $this->setUser( $user );
4367
4368 $pairs = array(
4369 "\r\n" => "\n",
4370 );
4371 $text = str_replace( array_keys( $pairs ), array_values( $pairs ), $text );
4372 if( $options->getPreSaveTransform() ) {
4373 $text = $this->pstPass2( $text, $user );
4374 }
4375 $text = $this->mStripState->unstripBoth( $text );
4376
4377 $this->setUser( null ); #Reset
4378
4379 return $text;
4380 }
4381
4382 /**
4383 * Pre-save transform helper function
4384 * @private
4385 *
4386 * @param $text string
4387 * @param $user User
4388 *
4389 * @return string
4390 */
4391 function pstPass2( $text, $user ) {
4392 global $wgContLang, $wgLocaltimezone;
4393
4394 # Note: This is the timestamp saved as hardcoded wikitext to
4395 # the database, we use $wgContLang here in order to give
4396 # everyone the same signature and use the default one rather
4397 # than the one selected in each user's preferences.
4398 # (see also bug 12815)
4399 $ts = $this->mOptions->getTimestamp();
4400 if ( isset( $wgLocaltimezone ) ) {
4401 $tz = $wgLocaltimezone;
4402 } else {
4403 $tz = date_default_timezone_get();
4404 }
4405
4406 $unixts = wfTimestamp( TS_UNIX, $ts );
4407 $oldtz = date_default_timezone_get();
4408 date_default_timezone_set( $tz );
4409 $ts = date( 'YmdHis', $unixts );
4410 $tzMsg = date( 'T', $unixts ); # might vary on DST changeover!
4411
4412 # Allow translation of timezones through wiki. date() can return
4413 # whatever crap the system uses, localised or not, so we cannot
4414 # ship premade translations.
4415 $key = 'timezone-' . strtolower( trim( $tzMsg ) );
4416 $msg = wfMessage( $key )->inContentLanguage();
4417 if ( $msg->exists() ) {
4418 $tzMsg = $msg->text();
4419 }
4420
4421 date_default_timezone_set( $oldtz );
4422
4423 $d = $wgContLang->timeanddate( $ts, false, false ) . " ($tzMsg)";
4424
4425 # Variable replacement
4426 # Because mOutputType is OT_WIKI, this will only process {{subst:xxx}} type tags
4427 $text = $this->replaceVariables( $text );
4428
4429 # This works almost by chance, as the replaceVariables are done before the getUserSig(),
4430 # which may corrupt this parser instance via its wfMessage()->text() call-
4431
4432 # Signatures
4433 $sigText = $this->getUserSig( $user );
4434 $text = strtr( $text, array(
4435 '~~~~~' => $d,
4436 '~~~~' => "$sigText $d",
4437 '~~~' => $sigText
4438 ) );
4439
4440 # Context links: [[|name]] and [[name (context)|]]
4441 $tc = '[' . Title::legalChars() . ']';
4442 $nc = '[ _0-9A-Za-z\x80-\xff-]'; # Namespaces can use non-ascii!
4443
4444 $p1 = "/\[\[(:?$nc+:|:|)($tc+?)( ?\\($tc+\\))\\|]]/"; # [[ns:page (context)|]]
4445 $p4 = "/\[\[(:?$nc+:|:|)($tc+?)( ?（$tc+）)\\|]]/"; # [[ns:page（context）|]]
4446 $p3 = "/\[\[(:?$nc+:|:|)($tc+?)( ?\\($tc+\\)|)((?:, |，)$tc+|)\\|]]/"; # [[ns:page (context), context|]]
4447 $p2 = "/\[\[\\|($tc+)]]/"; # [[|page]]
4448
4449 # try $p1 first, to turn "[[A, B (C)|]]" into "[[A, B (C)|A, B]]"
4450 $text = preg_replace( $p1, '[[\\1\\2\\3|\\2]]', $text );
4451 $text = preg_replace( $p4, '[[\\1\\2\\3|\\2]]', $text );
4452 $text = preg_replace( $p3, '[[\\1\\2\\3\\4|\\2]]', $text );
4453
4454 $t = $this->mTitle->getText();
4455 $m = array();
4456 if ( preg_match( "/^($nc+:|)$tc+?( \\($tc+\\))$/", $t, $m ) ) {
4457 $text = preg_replace( $p2, "[[$m[1]\\1$m[2]|\\1]]", $text );
4458 } elseif ( preg_match( "/^($nc+:|)$tc+?(, $tc+|)$/", $t, $m ) && "$m[1]$m[2]" != '' ) {
4459 $text = preg_replace( $p2, "[[$m[1]\\1$m[2]|\\1]]", $text );
4460 } else {
4461 # if there's no context, don't bother duplicating the title
4462 $text = preg_replace( $p2, '[[\\1]]', $text );
4463 }
4464
4465 # Trim trailing whitespace
4466 $text = rtrim( $text );
4467
4468 return $text;
4469 }
4470
4471 /**
4472 * Fetch the user's signature text, if any, and normalize to
4473 * validated, ready-to-insert wikitext.
4474 * If you have pre-fetched the nickname or the fancySig option, you can
4475 * specify them here to save a database query.
4476 * Do not reuse this parser instance after calling getUserSig(),
4477 * as it may have changed if it's the $wgParser.
4478 *
4479 * @param $user User
4480 * @param $nickname String|bool nickname to use or false to use user's default nickname
4481 * @param $fancySig Boolean|null whether the nicknname is the complete signature
4482 * or null to use default value
4483 * @return string
4484 */
4485 function getUserSig( &$user, $nickname = false, $fancySig = null ) {
4486 global $wgMaxSigChars;
4487
4488 $username = $user->getName();
4489
4490 # If not given, retrieve from the user object.
4491 if ( $nickname === false )
4492 $nickname = $user->getOption( 'nickname' );
4493
4494 if ( is_null( $fancySig ) ) {
4495 $fancySig = $user->getBoolOption( 'fancysig' );
4496 }
4497
4498 $nickname = $nickname == null ? $username : $nickname;
4499
4500 if ( mb_strlen( $nickname ) > $wgMaxSigChars ) {
4501 $nickname = $username;
4502 wfDebug( __METHOD__ . ": $username has overlong signature.\n" );
4503 } elseif ( $fancySig !== false ) {
4504 # Sig. might contain markup; validate this
4505 if ( $this->validateSig( $nickname ) !== false ) {
4506 # Validated; clean up (if needed) and return it
4507 return $this->cleanSig( $nickname, true );
4508 } else {
4509 # Failed to validate; fall back to the default
4510 $nickname = $username;
4511 wfDebug( __METHOD__.": $username has bad XML tags in signature.\n" );
4512 }
4513 }
4514
4515 # Make sure nickname doesnt get a sig in a sig
4516 $nickname = self::cleanSigInSig( $nickname );
4517
4518 # If we're still here, make it a link to the user page
4519 $userText = wfEscapeWikiText( $username );
4520 $nickText = wfEscapeWikiText( $nickname );
4521 $msgName = $user->isAnon() ? 'signature-anon' : 'signature';
4522
4523 return wfMessage( $msgName, $userText, $nickText )->inContentLanguage()->title( $this->getTitle() )->text();
4524 }
4525
4526 /**
4527 * Check that the user's signature contains no bad XML
4528 *
4529 * @param $text String
4530 * @return mixed An expanded string, or false if invalid.
4531 */
4532 function validateSig( $text ) {
4533 return( Xml::isWellFormedXmlFragment( $text ) ? $text : false );
4534 }
4535
4536 /**
4537 * Clean up signature text
4538 *
4539 * 1) Strip ~~~, ~~~~ and ~~~~~ out of signatures @see cleanSigInSig
4540 * 2) Substitute all transclusions
4541 *
4542 * @param $text String
4543 * @param $parsing bool Whether we're cleaning (preferences save) or parsing
4544 * @return String: signature text
4545 */
4546 public function cleanSig( $text, $parsing = false ) {
4547 if ( !$parsing ) {
4548 global $wgTitle;
4549 $this->startParse( $wgTitle, new ParserOptions, self::OT_PREPROCESS, true );
4550 }
4551
4552 # Option to disable this feature
4553 if ( !$this->mOptions->getCleanSignatures() ) {
4554 return $text;
4555 }
4556
4557 # @todo FIXME: Regex doesn't respect extension tags or nowiki
4558 # => Move this logic to braceSubstitution()
4559 $substWord = MagicWord::get( 'subst' );
4560 $substRegex = '/\{\{(?!(?:' . $substWord->getBaseRegex() . '))/x' . $substWord->getRegexCase();
4561 $substText = '{{' . $substWord->getSynonym( 0 );
4562
4563 $text = preg_replace( $substRegex, $substText, $text );
4564 $text = self::cleanSigInSig( $text );
4565 $dom = $this->preprocessToDom( $text );
4566 $frame = $this->getPreprocessor()->newFrame();
4567 $text = $frame->expand( $dom );
4568
4569 if ( !$parsing ) {
4570 $text = $this->mStripState->unstripBoth( $text );
4571 }
4572
4573 return $text;
4574 }
4575
4576 /**
4577 * Strip ~~~, ~~~~ and ~~~~~ out of signatures
4578 *
4579 * @param $text String
4580 * @return String: signature text with /~{3,5}/ removed
4581 */
4582 public static function cleanSigInSig( $text ) {
4583 $text = preg_replace( '/~{3,5}/', '', $text );
4584 return $text;
4585 }
4586
4587 /**
4588 * Set up some variables which are usually set up in parse()
4589 * so that an external function can call some class members with confidence
4590 *
4591 * @param $title Title|null
4592 * @param $options ParserOptions
4593 * @param $outputType
4594 * @param $clearState bool
4595 */
4596 public function startExternalParse( Title $title = null, ParserOptions $options, $outputType, $clearState = true ) {
4597 $this->startParse( $title, $options, $outputType, $clearState );
4598 }
4599
4600 /**
4601 * @param $title Title|null
4602 * @param $options ParserOptions
4603 * @param $outputType
4604 * @param $clearState bool
4605 */
4606 private function startParse( Title $title = null, ParserOptions $options, $outputType, $clearState = true ) {
4607 $this->setTitle( $title );
4608 $this->mOptions = $options;
4609 $this->setOutputType( $outputType );
4610 if ( $clearState ) {
4611 $this->clearState();
4612 }
4613 }
4614
4615 /**
4616 * Wrapper for preprocess()
4617 *
4618 * @param $text String: the text to preprocess
4619 * @param $options ParserOptions: options
4620 * @param $title Title object or null to use $wgTitle
4621 * @return String
4622 */
4623 public function transformMsg( $text, $options, $title = null ) {
4624 static $executing = false;
4625
4626 # Guard against infinite recursion
4627 if ( $executing ) {
4628 return $text;
4629 }
4630 $executing = true;
4631
4632 wfProfileIn( __METHOD__ );
4633 if ( !$title ) {
4634 global $wgTitle;
4635 $title = $wgTitle;
4636 }
4637 if ( !$title ) {
4638 # It's not uncommon having a null $wgTitle in scripts. See r80898
4639 # Create a ghost title in such case
4640 $title = Title::newFromText( 'Dwimmerlaik' );
4641 }
4642 $text = $this->preprocess( $text, $title, $options );
4643
4644 $executing = false;
4645 wfProfileOut( __METHOD__ );
4646 return $text;
4647 }
4648
4649 /**
4650 * Create an HTML-style tag, e.g. "<yourtag>special text</yourtag>"
4651 * The callback should have the following form:
4652 * function myParserHook( $text, $params, $parser, $frame ) { ... }
4653 *
4654 * Transform and return $text. Use $parser for any required context, e.g. use
4655 * $parser->getTitle() and $parser->getOptions() not $wgTitle or $wgOut->mParserOptions
4656 *
4657 * Hooks may return extended information by returning an array, of which the
4658 * first numbered element (index 0) must be the return string, and all other
4659 * entries are extracted into local variables within an internal function
4660 * in the Parser class.
4661 *
4662 * This interface (introduced r61913) appears to be undocumented, but
4663 * 'markerName' is used by some core tag hooks to override which strip
4664 * array their results are placed in. **Use great caution if attempting
4665 * this interface, as it is not documented and injudicious use could smash
4666 * private variables.**
4667 *
4668 * @param $tag Mixed: the tag to use, e.g. 'hook' for "<hook>"
4669 * @param $callback Mixed: the callback function (and object) to use for the tag
4670 * @return Mixed|null The old value of the mTagHooks array associated with the hook
4671 */
4672 public function setHook( $tag, $callback ) {
4673 $tag = strtolower( $tag );
4674 if ( preg_match( '/[<>\r\n]/', $tag, $m ) ) {
4675 throw new MWException( "Invalid character {$m[0]} in setHook('$tag', ...) call" );
4676 }
4677 $oldVal = isset( $this->mTagHooks[$tag] ) ? $this->mTagHooks[$tag] : null;
4678 $this->mTagHooks[$tag] = $callback;
4679 if ( !in_array( $tag, $this->mStripList ) ) {
4680 $this->mStripList[] = $tag;
4681 }
4682
4683 return $oldVal;
4684 }
4685
4686 /**
4687 * As setHook(), but letting the contents be parsed.
4688 *
4689 * Transparent tag hooks are like regular XML-style tag hooks, except they
4690 * operate late in the transformation sequence, on HTML instead of wikitext.
4691 *
4692 * This is probably obsoleted by things dealing with parser frames?
4693 * The only extension currently using it is geoserver.
4694 *
4695 * @since 1.10
4696 * @todo better document or deprecate this
4697 *
4698 * @param $tag Mixed: the tag to use, e.g. 'hook' for "<hook>"
4699 * @param $callback Mixed: the callback function (and object) to use for the tag
4700 * @return Mixed|null The old value of the mTagHooks array associated with the hook
4701 */
4702 function setTransparentTagHook( $tag, $callback ) {
4703 $tag = strtolower( $tag );
4704 if ( preg_match( '/[<>\r\n]/', $tag, $m ) ) {
4705 throw new MWException( "Invalid character {$m[0]} in setTransparentHook('$tag', ...) call" );
4706 }
4707 $oldVal = isset( $this->mTransparentTagHooks[$tag] ) ? $this->mTransparentTagHooks[$tag] : null;
4708 $this->mTransparentTagHooks[$tag] = $callback;
4709
4710 return $oldVal;
4711 }
4712
4713 /**
4714 * Remove all tag hooks
4715 */
4716 function clearTagHooks() {
4717 $this->mTagHooks = array();
4718 $this->mFunctionTagHooks = array();
4719 $this->mStripList = $this->mDefaultStripList;
4720 }
4721
4722 /**
4723 * Create a function, e.g. {{sum:1|2|3}}
4724 * The callback function should have the form:
4725 * function myParserFunction( &$parser, $arg1, $arg2, $arg3 ) { ... }
4726 *
4727 * Or with SFH_OBJECT_ARGS:
4728 * function myParserFunction( $parser, $frame, $args ) { ... }
4729 *
4730 * The callback may either return the text result of the function, or an array with the text
4731 * in element 0, and a number of flags in the other elements. The names of the flags are
4732 * specified in the keys. Valid flags are:
4733 * found The text returned is valid, stop processing the template. This
4734 * is on by default.
4735 * nowiki Wiki markup in the return value should be escaped
4736 * isHTML The returned text is HTML, armour it against wikitext transformation
4737 *
4738 * @param $id String: The magic word ID
4739 * @param $callback Mixed: the callback function (and object) to use
4740 * @param $flags Integer: a combination of the following flags:
4741 * SFH_NO_HASH No leading hash, i.e. {{plural:...}} instead of {{#if:...}}
4742 *
4743 * SFH_OBJECT_ARGS Pass the template arguments as PPNode objects instead of text. This
4744 * allows for conditional expansion of the parse tree, allowing you to eliminate dead
4745 * branches and thus speed up parsing. It is also possible to analyse the parse tree of
4746 * the arguments, and to control the way they are expanded.
4747 *
4748 * The $frame parameter is a PPFrame. This can be used to produce expanded text from the
4749 * arguments, for instance:
4750 * $text = isset( $args[0] ) ? $frame->expand( $args[0] ) : '';
4751 *
4752 * For technical reasons, $args[0] is pre-expanded and will be a string. This may change in
4753 * future versions. Please call $frame->expand() on it anyway so that your code keeps
4754 * working if/when this is changed.
4755 *
4756 * If you want whitespace to be trimmed from $args, you need to do it yourself, post-
4757 * expansion.
4758 *
4759 * Please read the documentation in includes/parser/Preprocessor.php for more information
4760 * about the methods available in PPFrame and PPNode.
4761 *
4762 * @return string|callback The old callback function for this name, if any
4763 */
4764 public function setFunctionHook( $id, $callback, $flags = 0 ) {
4765 global $wgContLang;
4766
4767 $oldVal = isset( $this->mFunctionHooks[$id] ) ? $this->mFunctionHooks[$id][0] : null;
4768 $this->mFunctionHooks[$id] = array( $callback, $flags );
4769
4770 # Add to function cache
4771 $mw = MagicWord::get( $id );
4772 if ( !$mw )
4773 throw new MWException( __METHOD__.'() expecting a magic word identifier.' );
4774
4775 $synonyms = $mw->getSynonyms();
4776 $sensitive = intval( $mw->isCaseSensitive() );
4777
4778 foreach ( $synonyms as $syn ) {
4779 # Case
4780 if ( !$sensitive ) {
4781 $syn = $wgContLang->lc( $syn );
4782 }
4783 # Add leading hash
4784 if ( !( $flags & SFH_NO_HASH ) ) {
4785 $syn = '#' . $syn;
4786 }
4787 # Remove trailing colon
4788 if ( substr( $syn, -1, 1 ) === ':' ) {
4789 $syn = substr( $syn, 0, -1 );
4790 }
4791 $this->mFunctionSynonyms[$sensitive][$syn] = $id;
4792 }
4793 return $oldVal;
4794 }
4795
4796 /**
4797 * Get all registered function hook identifiers
4798 *
4799 * @return Array
4800 */
4801 function getFunctionHooks() {
4802 return array_keys( $this->mFunctionHooks );
4803 }
4804
4805 /**
4806 * Create a tag function, e.g. "<test>some stuff</test>".
4807 * Unlike tag hooks, tag functions are parsed at preprocessor level.
4808 * Unlike parser functions, their content is not preprocessed.
4809 * @return null
4810 */
4811 function setFunctionTagHook( $tag, $callback, $flags ) {
4812 $tag = strtolower( $tag );
4813 if ( preg_match( '/[<>\r\n]/', $tag, $m ) ) throw new MWException( "Invalid character {$m[0]} in setFunctionTagHook('$tag', ...) call" );
4814 $old = isset( $this->mFunctionTagHooks[$tag] ) ?
4815 $this->mFunctionTagHooks[$tag] : null;
4816 $this->mFunctionTagHooks[$tag] = array( $callback, $flags );
4817
4818 if ( !in_array( $tag, $this->mStripList ) ) {
4819 $this->mStripList[] = $tag;
4820 }
4821
4822 return $old;
4823 }
4824
4825 /**
4826 * @todo FIXME: Update documentation. makeLinkObj() is deprecated.
4827 * Replace "<!--LINK-->" link placeholders with actual links, in the buffer
4828 * Placeholders created in Skin::makeLinkObj()
4829 *
4830 * @param $text string
4831 * @param $options int
4832 *
4833 * @return array of link CSS classes, indexed by PDBK.
4834 */
4835 function replaceLinkHolders( &$text, $options = 0 ) {
4836 return $this->mLinkHolders->replace( $text );
4837 }
4838
4839 /**
4840 * Replace "<!--LINK-->" link placeholders with plain text of links
4841 * (not HTML-formatted).
4842 *
4843 * @param $text String
4844 * @return String
4845 */
4846 function replaceLinkHoldersText( $text ) {
4847 return $this->mLinkHolders->replaceText( $text );
4848 }
4849
4850 /**
4851 * Renders an image gallery from a text with one line per image.
4852 * text labels may be given by using |-style alternative text. E.g.
4853 * Image:one.jpg|The number "1"
4854 * Image:tree.jpg|A tree
4855 * given as text will return the HTML of a gallery with two images,
4856 * labeled 'The number "1"' and
4857 * 'A tree'.
4858 *
4859 * @param string $text
4860 * @param array $params
4861 * @return string HTML
4862 */
4863 function renderImageGallery( $text, $params ) {
4864 $ig = new ImageGallery();
4865 $ig->setContextTitle( $this->mTitle );
4866 $ig->setShowBytes( false );
4867 $ig->setShowFilename( false );
4868 $ig->setParser( $this );
4869 $ig->setHideBadImages();
4870 $ig->setAttributes( Sanitizer::validateTagAttributes( $params, 'table' ) );
4871
4872 if ( isset( $params['showfilename'] ) ) {
4873 $ig->setShowFilename( true );
4874 } else {
4875 $ig->setShowFilename( false );
4876 }
4877 if ( isset( $params['caption'] ) ) {
4878 $caption = $params['caption'];
4879 $caption = htmlspecialchars( $caption );
4880 $caption = $this->replaceInternalLinks( $caption );
4881 $ig->setCaptionHtml( $caption );
4882 }
4883 if ( isset( $params['perrow'] ) ) {
4884 $ig->setPerRow( $params['perrow'] );
4885 }
4886 if ( isset( $params['widths'] ) ) {
4887 $ig->setWidths( $params['widths'] );
4888 }
4889 if ( isset( $params['heights'] ) ) {
4890 $ig->setHeights( $params['heights'] );
4891 }
4892
4893 wfRunHooks( 'BeforeParserrenderImageGallery', array( &$this, &$ig ) );
4894
4895 $lines = StringUtils::explode( "\n", $text );
4896 foreach ( $lines as $line ) {
4897 # match lines like these:
4898 # Image:someimage.jpg|This is some image
4899 $matches = array();
4900 preg_match( "/^([^|]+)(\\|(.*))?$/", $line, $matches );
4901 # Skip empty lines
4902 if ( count( $matches ) == 0 ) {
4903 continue;
4904 }
4905
4906 if ( strpos( $matches[0], '%' ) !== false ) {
4907 $matches[1] = rawurldecode( $matches[1] );
4908 }
4909 $title = Title::newFromText( $matches[1], NS_FILE );
4910 if ( is_null( $title ) ) {
4911 # Bogus title. Ignore these so we don't bomb out later.
4912 continue;
4913 }
4914
4915 $label = '';
4916 $alt = '';
4917 $link = '';
4918 if ( isset( $matches[3] ) ) {
4919 // look for an |alt= definition while trying not to break existing
4920 // captions with multiple pipes (|) in it, until a more sensible grammar
4921 // is defined for images in galleries
4922
4923 $matches[3] = $this->recursiveTagParse( trim( $matches[3] ) );
4924 $parameterMatches = StringUtils::explode('|', $matches[3]);
4925 $magicWordAlt = MagicWord::get( 'img_alt' );
4926 $magicWordLink = MagicWord::get( 'img_link' );
4927
4928 foreach ( $parameterMatches as $parameterMatch ) {
4929 if ( $match = $magicWordAlt->matchVariableStartToEnd( $parameterMatch ) ) {
4930 $alt = $this->stripAltText( $match, false );
4931 }
4932 elseif( $match = $magicWordLink->matchVariableStartToEnd( $parameterMatch ) ){
4933 $link = strip_tags($this->replaceLinkHoldersText($match));
4934 $chars = self::EXT_LINK_URL_CLASS;
4935 $prots = $this->mUrlProtocols;
4936 //check to see if link matches an absolute url, if not then it must be a wiki link.
4937 if(!preg_match( "/^($prots)$chars+$/u", $link)){
4938 $localLinkTitle = Title::newFromText($link);
4939 $link = $localLinkTitle->getLocalURL();
4940 }
4941 }
4942 else {
4943 // concatenate all other pipes
4944 $label .= '|' . $parameterMatch;
4945 }
4946 }
4947 // remove the first pipe
4948 $label = substr( $label, 1 );
4949 }
4950
4951 $ig->add( $title, $label, $alt ,$link);
4952 }
4953 return $ig->toHTML();
4954 }
4955
4956 /**
4957 * @param $handler
4958 * @return array
4959 */
4960 function getImageParams( $handler ) {
4961 if ( $handler ) {
4962 $handlerClass = get_class( $handler );
4963 } else {
4964 $handlerClass = '';
4965 }
4966 if ( !isset( $this->mImageParams[$handlerClass] ) ) {
4967 # Initialise static lists
4968 static $internalParamNames = array(
4969 'horizAlign' => array( 'left', 'right', 'center', 'none' ),
4970 'vertAlign' => array( 'baseline', 'sub', 'super', 'top', 'text-top', 'middle',
4971 'bottom', 'text-bottom' ),
4972 'frame' => array( 'thumbnail', 'manualthumb', 'framed', 'frameless',
4973 'upright', 'border', 'link', 'alt', 'class' ),
4974 );
4975 static $internalParamMap;
4976 if ( !$internalParamMap ) {
4977 $internalParamMap = array();
4978 foreach ( $internalParamNames as $type => $names ) {
4979 foreach ( $names as $name ) {
4980 $magicName = str_replace( '-', '_', "img_$name" );
4981 $internalParamMap[$magicName] = array( $type, $name );
4982 }
4983 }
4984 }
4985
4986 # Add handler params
4987 $paramMap = $internalParamMap;
4988 if ( $handler ) {
4989 $handlerParamMap = $handler->getParamMap();
4990 foreach ( $handlerParamMap as $magic => $paramName ) {
4991 $paramMap[$magic] = array( 'handler', $paramName );
4992 }
4993 }
4994 $this->mImageParams[$handlerClass] = $paramMap;
4995 $this->mImageParamsMagicArray[$handlerClass] = new MagicWordArray( array_keys( $paramMap ) );
4996 }
4997 return array( $this->mImageParams[$handlerClass], $this->mImageParamsMagicArray[$handlerClass] );
4998 }
4999
5000 /**
5001 * Parse image options text and use it to make an image
5002 *
5003 * @param $title Title
5004 * @param $options String
5005 * @param $holders LinkHolderArray|bool
5006 * @return string HTML
5007 */
5008 function makeImage( $title, $options, $holders = false ) {
5009 # Check if the options text is of the form "options|alt text"
5010 # Options are:
5011 # * thumbnail make a thumbnail with enlarge-icon and caption, alignment depends on lang
5012 # * left no resizing, just left align. label is used for alt= only
5013 # * right same, but right aligned
5014 # * none same, but not aligned
5015 # * ___px scale to ___ pixels width, no aligning. e.g. use in taxobox
5016 # * center center the image
5017 # * frame Keep original image size, no magnify-button.
5018 # * framed Same as "frame"
5019 # * frameless like 'thumb' but without a frame. Keeps user preferences for width
5020 # * upright reduce width for upright images, rounded to full __0 px
5021 # * border draw a 1px border around the image
5022 # * alt Text for HTML alt attribute (defaults to empty)
5023 # * class Set a class for img node
5024 # * link Set the target of the image link. Can be external, interwiki, or local
5025 # vertical-align values (no % or length right now):
5026 # * baseline
5027 # * sub
5028 # * super
5029 # * top
5030 # * text-top
5031 # * middle
5032 # * bottom
5033 # * text-bottom
5034
5035 $parts = StringUtils::explode( "|", $options );
5036
5037 # Give extensions a chance to select the file revision for us
5038 $options = array();
5039 $descQuery = false;
5040 wfRunHooks( 'BeforeParserFetchFileAndTitle',
5041 array( $this, $title, &$options, &$descQuery ) );
5042 # Fetch and register the file (file title may be different via hooks)
5043 list( $file, $title ) = $this->fetchFileAndTitle( $title, $options );
5044
5045 # Get parameter map
5046 $handler = $file ? $file->getHandler() : false;
5047
5048 list( $paramMap, $mwArray ) = $this->getImageParams( $handler );
5049
5050 if ( !$file ) {
5051 $this->addTrackingCategory( 'broken-file-category' );
5052 }
5053
5054 # Process the input parameters
5055 $caption = '';
5056 $params = array( 'frame' => array(), 'handler' => array(),
5057 'horizAlign' => array(), 'vertAlign' => array() );
5058 foreach ( $parts as $part ) {
5059 $part = trim( $part );
5060 list( $magicName, $value ) = $mwArray->matchVariableStartToEnd( $part );
5061 $validated = false;
5062 if ( isset( $paramMap[$magicName] ) ) {
5063 list( $type, $paramName ) = $paramMap[$magicName];
5064
5065 # Special case; width and height come in one variable together
5066 if ( $type === 'handler' && $paramName === 'width' ) {
5067 $parsedWidthParam = $this->parseWidthParam( $value );
5068 if( isset( $parsedWidthParam['width'] ) ) {
5069 $width = $parsedWidthParam['width'];
5070 if ( $handler->validateParam( 'width', $width ) ) {
5071 $params[$type]['width'] = $width;
5072 $validated = true;
5073 }
5074 }
5075 if( isset( $parsedWidthParam['height'] ) ) {
5076 $height = $parsedWidthParam['height'];
5077 if ( $handler->validateParam( 'height', $height ) ) {
5078 $params[$type]['height'] = $height;
5079 $validated = true;
5080 }
5081 }
5082 # else no validation -- bug 13436
5083 } else {
5084 if ( $type === 'handler' ) {
5085 # Validate handler parameter
5086 $validated = $handler->validateParam( $paramName, $value );
5087 } else {
5088 # Validate internal parameters
5089 switch( $paramName ) {
5090 case 'manualthumb':
5091 case 'alt':
5092 case 'class':
5093 # @todo FIXME: Possibly check validity here for
5094 # manualthumb? downstream behavior seems odd with
5095 # missing manual thumbs.
5096 $validated = true;
5097 $value = $this->stripAltText( $value, $holders );
5098 break;
5099 case 'link':
5100 $chars = self::EXT_LINK_URL_CLASS;
5101 $prots = $this->mUrlProtocols;
5102 if ( $value === '' ) {
5103 $paramName = 'no-link';
5104 $value = true;
5105 $validated = true;
5106 } elseif ( preg_match( "/^(?i)$prots/", $value ) ) {
5107 if ( preg_match( "/^((?i)$prots)$chars+$/u", $value, $m ) ) {
5108 $paramName = 'link-url';
5109 $this->mOutput->addExternalLink( $value );
5110 if ( $this->mOptions->getExternalLinkTarget() ) {
5111 $params[$type]['link-target'] = $this->mOptions->getExternalLinkTarget();
5112 }
5113 $validated = true;
5114 }
5115 } else {
5116 $linkTitle = Title::newFromText( $value );
5117 if ( $linkTitle ) {
5118 $paramName = 'link-title';
5119 $value = $linkTitle;
5120 $this->mOutput->addLink( $linkTitle );
5121 $validated = true;
5122 }
5123 }
5124 break;
5125 default:
5126 # Most other things appear to be empty or numeric...
5127 $validated = ( $value === false || is_numeric( trim( $value ) ) );
5128 }
5129 }
5130
5131 if ( $validated ) {
5132 $params[$type][$paramName] = $value;
5133 }
5134 }
5135 }
5136 if ( !$validated ) {
5137 $caption = $part;
5138 }
5139 }
5140
5141 # Process alignment parameters
5142 if ( $params['horizAlign'] ) {
5143 $params['frame']['align'] = key( $params['horizAlign'] );
5144 }
5145 if ( $params['vertAlign'] ) {
5146 $params['frame']['valign'] = key( $params['vertAlign'] );
5147 }
5148
5149 $params['frame']['caption'] = $caption;
5150
5151 # Will the image be presented in a frame, with the caption below?
5152 $imageIsFramed = isset( $params['frame']['frame'] ) ||
5153 isset( $params['frame']['framed'] ) ||
5154 isset( $params['frame']['thumbnail'] ) ||
5155 isset( $params['frame']['manualthumb'] );
5156
5157 # In the old days, [[Image:Foo|text...]] would set alt text. Later it
5158 # came to also set the caption, ordinary text after the image -- which
5159 # makes no sense, because that just repeats the text multiple times in
5160 # screen readers. It *also* came to set the title attribute.
5161 #
5162 # Now that we have an alt attribute, we should not set the alt text to
5163 # equal the caption: that's worse than useless, it just repeats the
5164 # text. This is the framed/thumbnail case. If there's no caption, we
5165 # use the unnamed parameter for alt text as well, just for the time be-
5166 # ing, if the unnamed param is set and the alt param is not.
5167 #
5168 # For the future, we need to figure out if we want to tweak this more,
5169 # e.g., introducing a title= parameter for the title; ignoring the un-
5170 # named parameter entirely for images without a caption; adding an ex-
5171 # plicit caption= parameter and preserving the old magic unnamed para-
5172 # meter for BC; ...
5173 if ( $imageIsFramed ) { # Framed image
5174 if ( $caption === '' && !isset( $params['frame']['alt'] ) ) {
5175 # No caption or alt text, add the filename as the alt text so
5176 # that screen readers at least get some description of the image
5177 $params['frame']['alt'] = $title->getText();
5178 }
5179 # Do not set $params['frame']['title'] because tooltips don't make sense
5180 # for framed images
5181 } else { # Inline image
5182 if ( !isset( $params['frame']['alt'] ) ) {
5183 # No alt text, use the "caption" for the alt text
5184 if ( $caption !== '') {
5185 $params['frame']['alt'] = $this->stripAltText( $caption, $holders );
5186 } else {
5187 # No caption, fall back to using the filename for the
5188 # alt text
5189 $params['frame']['alt'] = $title->getText();
5190 }
5191 }
5192 # Use the "caption" for the tooltip text
5193 $params['frame']['title'] = $this->stripAltText( $caption, $holders );
5194 }
5195
5196 wfRunHooks( 'ParserMakeImageParams', array( $title, $file, &$params, $this ) );
5197
5198 # Linker does the rest
5199 $time = isset( $options['time'] ) ? $options['time'] : false;
5200 $ret = Linker::makeImageLink( $this, $title, $file, $params['frame'], $params['handler'],
5201 $time, $descQuery, $this->mOptions->getThumbSize() );
5202
5203 # Give the handler a chance to modify the parser object
5204 if ( $handler ) {
5205 $handler->parserTransformHook( $this, $file );
5206 }
5207
5208 return $ret;
5209 }
5210
5211 /**
5212 * @param $caption
5213 * @param $holders LinkHolderArray
5214 * @return mixed|String
5215 */
5216 protected function stripAltText( $caption, $holders ) {
5217 # Strip bad stuff out of the title (tooltip). We can't just use
5218 # replaceLinkHoldersText() here, because if this function is called
5219 # from replaceInternalLinks2(), mLinkHolders won't be up-to-date.
5220 if ( $holders ) {
5221 $tooltip = $holders->replaceText( $caption );
5222 } else {
5223 $tooltip = $this->replaceLinkHoldersText( $caption );
5224 }
5225
5226 # make sure there are no placeholders in thumbnail attributes
5227 # that are later expanded to html- so expand them now and
5228 # remove the tags
5229 $tooltip = $this->mStripState->unstripBoth( $tooltip );
5230 $tooltip = Sanitizer::stripAllTags( $tooltip );
5231
5232 return $tooltip;
5233 }
5234
5235 /**
5236 * Set a flag in the output object indicating that the content is dynamic and
5237 * shouldn't be cached.
5238 */
5239 function disableCache() {
5240 wfDebug( "Parser output marked as uncacheable.\n" );
5241 if ( !$this->mOutput ) {
5242 throw new MWException( __METHOD__ .
5243 " can only be called when actually parsing something" );
5244 }
5245 $this->mOutput->setCacheTime( -1 ); // old style, for compatibility
5246 $this->mOutput->updateCacheExpiry( 0 ); // new style, for consistency
5247 }
5248
5249 /**
5250 * Callback from the Sanitizer for expanding items found in HTML attribute
5251 * values, so they can be safely tested and escaped.
5252 *
5253 * @param $text String
5254 * @param $frame PPFrame
5255 * @return String
5256 */
5257 function attributeStripCallback( &$text, $frame = false ) {
5258 $text = $this->replaceVariables( $text, $frame );
5259 $text = $this->mStripState->unstripBoth( $text );
5260 return $text;
5261 }
5262
5263 /**
5264 * Accessor
5265 *
5266 * @return array
5267 */
5268 function getTags() {
5269 return array_merge( array_keys( $this->mTransparentTagHooks ), array_keys( $this->mTagHooks ), array_keys( $this->mFunctionTagHooks ) );
5270 }
5271
5272 /**
5273 * Replace transparent tags in $text with the values given by the callbacks.
5274 *
5275 * Transparent tag hooks are like regular XML-style tag hooks, except they
5276 * operate late in the transformation sequence, on HTML instead of wikitext.
5277 *
5278 * @param $text string
5279 *
5280 * @return string
5281 */
5282 function replaceTransparentTags( $text ) {
5283 $matches = array();
5284 $elements = array_keys( $this->mTransparentTagHooks );
5285 $text = self::extractTagsAndParams( $elements, $text, $matches, $this->mUniqPrefix );
5286 $replacements = array();
5287
5288 foreach ( $matches as $marker => $data ) {
5289 list( $element, $content, $params, $tag ) = $data;
5290 $tagName = strtolower( $element );
5291 if ( isset( $this->mTransparentTagHooks[$tagName] ) ) {
5292 $output = call_user_func_array( $this->mTransparentTagHooks[$tagName], array( $content, $params, $this ) );
5293 } else {
5294 $output = $tag;
5295 }
5296 $replacements[$marker] = $output;
5297 }
5298 return strtr( $text, $replacements );
5299 }
5300
5301 /**
5302 * Break wikitext input into sections, and either pull or replace
5303 * some particular section's text.
5304 *
5305 * External callers should use the getSection and replaceSection methods.
5306 *
5307 * @param $text String: Page wikitext
5308 * @param $section String: a section identifier string of the form:
5309 * "<flag1> - <flag2> - ... - <section number>"
5310 *
5311 * Currently the only recognised flag is "T", which means the target section number
5312 * was derived during a template inclusion parse, in other words this is a template
5313 * section edit link. If no flags are given, it was an ordinary section edit link.
5314 * This flag is required to avoid a section numbering mismatch when a section is
5315 * enclosed by "<includeonly>" (bug 6563).
5316 *
5317 * The section number 0 pulls the text before the first heading; other numbers will
5318 * pull the given section along with its lower-level subsections. If the section is
5319 * not found, $mode=get will return $newtext, and $mode=replace will return $text.
5320 *
5321 * Section 0 is always considered to exist, even if it only contains the empty
5322 * string. If $text is the empty string and section 0 is replaced, $newText is
5323 * returned.
5324 *
5325 * @param $mode String: one of "get" or "replace"
5326 * @param $newText String: replacement text for section data.
5327 * @return String: for "get", the extracted section text.
5328 * for "replace", the whole page with the section replaced.
5329 */
5330 private function extractSections( $text, $section, $mode, $newText='' ) {
5331 global $wgTitle; # not generally used but removes an ugly failure mode
5332 $this->startParse( $wgTitle, new ParserOptions, self::OT_PLAIN, true );
5333 $outText = '';
5334 $frame = $this->getPreprocessor()->newFrame();
5335
5336 # Process section extraction flags
5337 $flags = 0;
5338 $sectionParts = explode( '-', $section );
5339 $sectionIndex = array_pop( $sectionParts );
5340 foreach ( $sectionParts as $part ) {
5341 if ( $part === 'T' ) {
5342 $flags |= self::PTD_FOR_INCLUSION;
5343 }
5344 }
5345
5346 # Check for empty input
5347 if ( strval( $text ) === '' ) {
5348 # Only sections 0 and T-0 exist in an empty document
5349 if ( $sectionIndex == 0 ) {
5350 if ( $mode === 'get' ) {
5351 return '';
5352 } else {
5353 return $newText;
5354 }
5355 } else {
5356 if ( $mode === 'get' ) {
5357 return $newText;
5358 } else {
5359 return $text;
5360 }
5361 }
5362 }
5363
5364 # Preprocess the text
5365 $root = $this->preprocessToDom( $text, $flags );
5366
5367 # <h> nodes indicate section breaks
5368 # They can only occur at the top level, so we can find them by iterating the root's children
5369 $node = $root->getFirstChild();
5370
5371 # Find the target section
5372 if ( $sectionIndex == 0 ) {
5373 # Section zero doesn't nest, level=big
5374 $targetLevel = 1000;
5375 } else {
5376 while ( $node ) {
5377 if ( $node->getName() === 'h' ) {
5378 $bits = $node->splitHeading();
5379 if ( $bits['i'] == $sectionIndex ) {
5380 $targetLevel = $bits['level'];
5381 break;
5382 }
5383 }
5384 if ( $mode === 'replace' ) {
5385 $outText .= $frame->expand( $node, PPFrame::RECOVER_ORIG );
5386 }
5387 $node = $node->getNextSibling();
5388 }
5389 }
5390
5391 if ( !$node ) {
5392 # Not found
5393 if ( $mode === 'get' ) {
5394 return $newText;
5395 } else {
5396 return $text;
5397 }
5398 }
5399
5400 # Find the end of the section, including nested sections
5401 do {
5402 if ( $node->getName() === 'h' ) {
5403 $bits = $node->splitHeading();
5404 $curLevel = $bits['level'];
5405 if ( $bits['i'] != $sectionIndex && $curLevel <= $targetLevel ) {
5406 break;
5407 }
5408 }
5409 if ( $mode === 'get' ) {
5410 $outText .= $frame->expand( $node, PPFrame::RECOVER_ORIG );
5411 }
5412 $node = $node->getNextSibling();
5413 } while ( $node );
5414
5415 # Write out the remainder (in replace mode only)
5416 if ( $mode === 'replace' ) {
5417 # Output the replacement text
5418 # Add two newlines on -- trailing whitespace in $newText is conventionally
5419 # stripped by the editor, so we need both newlines to restore the paragraph gap
5420 # Only add trailing whitespace if there is newText
5421 if ( $newText != "" ) {
5422 $outText .= $newText . "\n\n";
5423 }
5424
5425 while ( $node ) {
5426 $outText .= $frame->expand( $node, PPFrame::RECOVER_ORIG );
5427 $node = $node->getNextSibling();
5428 }
5429 }
5430
5431 if ( is_string( $outText ) ) {
5432 # Re-insert stripped tags
5433 $outText = rtrim( $this->mStripState->unstripBoth( $outText ) );
5434 }
5435
5436 return $outText;
5437 }
5438
5439 /**
5440 * This function returns the text of a section, specified by a number ($section).
5441 * A section is text under a heading like == Heading == or \<h1\>Heading\</h1\>, or
5442 * the first section before any such heading (section 0).
5443 *
5444 * If a section contains subsections, these are also returned.
5445 *
5446 * @param $text String: text to look in
5447 * @param $section String: section identifier
5448 * @param $deftext String: default to return if section is not found
5449 * @return string text of the requested section
5450 */
5451 public function getSection( $text, $section, $deftext='' ) {
5452 return $this->extractSections( $text, $section, "get", $deftext );
5453 }
5454
5455 /**
5456 * This function returns $oldtext after the content of the section
5457 * specified by $section has been replaced with $text. If the target
5458 * section does not exist, $oldtext is returned unchanged.
5459 *
5460 * @param $oldtext String: former text of the article
5461 * @param $section int section identifier
5462 * @param $text String: replacing text
5463 * @return String: modified text
5464 */
5465 public function replaceSection( $oldtext, $section, $text ) {
5466 return $this->extractSections( $oldtext, $section, "replace", $text );
5467 }
5468
5469 /**
5470 * Get the ID of the revision we are parsing
5471 *
5472 * @return Mixed: integer or null
5473 */
5474 function getRevisionId() {
5475 return $this->mRevisionId;
5476 }
5477
5478 /**
5479 * Get the revision object for $this->mRevisionId
5480 *
5481 * @return Revision|null either a Revision object or null
5482 */
5483 protected function getRevisionObject() {
5484 if ( !is_null( $this->mRevisionObject ) ) {
5485 return $this->mRevisionObject;
5486 }
5487 if ( is_null( $this->mRevisionId ) ) {
5488 return null;
5489 }
5490
5491 $this->mRevisionObject = Revision::newFromId( $this->mRevisionId );
5492 return $this->mRevisionObject;
5493 }
5494
5495 /**
5496 * Get the timestamp associated with the current revision, adjusted for
5497 * the default server-local timestamp
5498 */
5499 function getRevisionTimestamp() {
5500 if ( is_null( $this->mRevisionTimestamp ) ) {
5501 wfProfileIn( __METHOD__ );
5502
5503 global $wgContLang;
5504
5505 $revObject = $this->getRevisionObject();
5506 $timestamp = $revObject ? $revObject->getTimestamp() : wfTimestampNow();
5507
5508 # The cryptic '' timezone parameter tells to use the site-default
5509 # timezone offset instead of the user settings.
5510 #
5511 # Since this value will be saved into the parser cache, served
5512 # to other users, and potentially even used inside links and such,
5513 # it needs to be consistent for all visitors.
5514 $this->mRevisionTimestamp = $wgContLang->userAdjust( $timestamp, '' );
5515
5516 wfProfileOut( __METHOD__ );
5517 }
5518 return $this->mRevisionTimestamp;
5519 }
5520
5521 /**
5522 * Get the name of the user that edited the last revision
5523 *
5524 * @return String: user name
5525 */
5526 function getRevisionUser() {
5527 if( is_null( $this->mRevisionUser ) ) {
5528 $revObject = $this->getRevisionObject();
5529
5530 # if this template is subst: the revision id will be blank,
5531 # so just use the current user's name
5532 if( $revObject ) {
5533 $this->mRevisionUser = $revObject->getUserText();
5534 } elseif( $this->ot['wiki'] || $this->mOptions->getIsPreview() ) {
5535 $this->mRevisionUser = $this->getUser()->getName();
5536 }
5537 }
5538 return $this->mRevisionUser;
5539 }
5540
5541 /**
5542 * Mutator for $mDefaultSort
5543 *
5544 * @param $sort string New value
5545 */
5546 public function setDefaultSort( $sort ) {
5547 $this->mDefaultSort = $sort;
5548 $this->mOutput->setProperty( 'defaultsort', $sort );
5549 }
5550
5551 /**
5552 * Accessor for $mDefaultSort
5553 * Will use the empty string if none is set.
5554 *
5555 * This value is treated as a prefix, so the
5556 * empty string is equivalent to sorting by
5557 * page name.
5558 *
5559 * @return string
5560 */
5561 public function getDefaultSort() {
5562 if ( $this->mDefaultSort !== false ) {
5563 return $this->mDefaultSort;
5564 } else {
5565 return '';
5566 }
5567 }
5568
5569 /**
5570 * Accessor for $mDefaultSort
5571 * Unlike getDefaultSort(), will return false if none is set
5572 *
5573 * @return string or false
5574 */
5575 public function getCustomDefaultSort() {
5576 return $this->mDefaultSort;
5577 }
5578
5579 /**
5580 * Try to guess the section anchor name based on a wikitext fragment
5581 * presumably extracted from a heading, for example "Header" from
5582 * "== Header ==".
5583 *
5584 * @param $text string
5585 *
5586 * @return string
5587 */
5588 public function guessSectionNameFromWikiText( $text ) {
5589 # Strip out wikitext links(they break the anchor)
5590 $text = $this->stripSectionName( $text );
5591 $text = Sanitizer::normalizeSectionNameWhitespace( $text );
5592 return '#' . Sanitizer::escapeId( $text, 'noninitial' );
5593 }
5594
5595 /**
5596 * Same as guessSectionNameFromWikiText(), but produces legacy anchors
5597 * instead. For use in redirects, since IE6 interprets Redirect: headers
5598 * as something other than UTF-8 (apparently?), resulting in breakage.
5599 *
5600 * @param $text String: The section name
5601 * @return string An anchor
5602 */
5603 public function guessLegacySectionNameFromWikiText( $text ) {
5604 # Strip out wikitext links(they break the anchor)
5605 $text = $this->stripSectionName( $text );
5606 $text = Sanitizer::normalizeSectionNameWhitespace( $text );
5607 return '#' . Sanitizer::escapeId( $text, array( 'noninitial', 'legacy' ) );
5608 }
5609
5610 /**
5611 * Strips a text string of wikitext for use in a section anchor
5612 *
5613 * Accepts a text string and then removes all wikitext from the
5614 * string and leaves only the resultant text (i.e. the result of
5615 * [[User:WikiSysop|Sysop]] would be "Sysop" and the result of
5616 * [[User:WikiSysop]] would be "User:WikiSysop") - this is intended
5617 * to create valid section anchors by mimicing the output of the
5618 * parser when headings are parsed.
5619 *
5620 * @param $text String: text string to be stripped of wikitext
5621 * for use in a Section anchor
5622 * @return string Filtered text string
5623 */
5624 public function stripSectionName( $text ) {
5625 # Strip internal link markup
5626 $text = preg_replace( '/\[\[:?([^[|]+)\|([^[]+)\]\]/', '$2', $text );
5627 $text = preg_replace( '/\[\[:?([^[]+)\|?\]\]/', '$1', $text );
5628
5629 # Strip external link markup
5630 # @todo FIXME: Not tolerant to blank link text
5631 # I.E. [http://www.mediawiki.org] will render as [1] or something depending
5632 # on how many empty links there are on the page - need to figure that out.
5633 $text = preg_replace( '/\[(?i:' . $this->mUrlProtocols . ')([^ ]+?) ([^[]+)\]/', '$2', $text );
5634
5635 # Parse wikitext quotes (italics & bold)
5636 $text = $this->doQuotes( $text );
5637
5638 # Strip HTML tags
5639 $text = StringUtils::delimiterReplace( '<', '>', '', $text );
5640 return $text;
5641 }
5642
5643 /**
5644 * strip/replaceVariables/unstrip for preprocessor regression testing
5645 *
5646 * @param $text string
5647 * @param $title Title
5648 * @param $options ParserOptions
5649 * @param $outputType int
5650 *
5651 * @return string
5652 */
5653 function testSrvus( $text, Title $title, ParserOptions $options, $outputType = self::OT_HTML ) {
5654 $this->startParse( $title, $options, $outputType, true );
5655
5656 $text = $this->replaceVariables( $text );
5657 $text = $this->mStripState->unstripBoth( $text );
5658 $text = Sanitizer::removeHTMLtags( $text );
5659 return $text;
5660 }
5661
5662 /**
5663 * @param $text string
5664 * @param $title Title
5665 * @param $options ParserOptions
5666 * @return string
5667 */
5668 function testPst( $text, Title $title, ParserOptions $options ) {
5669 return $this->preSaveTransform( $text, $title, $options->getUser(), $options );
5670 }
5671
5672 /**
5673 * @param $text
5674 * @param $title Title
5675 * @param $options ParserOptions
5676 * @return string
5677 */
5678 function testPreprocess( $text, Title $title, ParserOptions $options ) {
5679 return $this->testSrvus( $text, $title, $options, self::OT_PREPROCESS );
5680 }
5681
5682 /**
5683 * Call a callback function on all regions of the given text that are not
5684 * inside strip markers, and replace those regions with the return value
5685 * of the callback. For example, with input:
5686 *
5687 * aaa<MARKER>bbb
5688 *
5689 * This will call the callback function twice, with 'aaa' and 'bbb'. Those
5690 * two strings will be replaced with the value returned by the callback in
5691 * each case.
5692 *
5693 * @param $s string
5694 * @param $callback
5695 *
5696 * @return string
5697 */
5698 function markerSkipCallback( $s, $callback ) {
5699 $i = 0;
5700 $out = '';
5701 while ( $i < strlen( $s ) ) {
5702 $markerStart = strpos( $s, $this->mUniqPrefix, $i );
5703 if ( $markerStart === false ) {
5704 $out .= call_user_func( $callback, substr( $s, $i ) );
5705 break;
5706 } else {
5707 $out .= call_user_func( $callback, substr( $s, $i, $markerStart - $i ) );
5708 $markerEnd = strpos( $s, self::MARKER_SUFFIX, $markerStart );
5709 if ( $markerEnd === false ) {
5710 $out .= substr( $s, $markerStart );
5711 break;
5712 } else {
5713 $markerEnd += strlen( self::MARKER_SUFFIX );
5714 $out .= substr( $s, $markerStart, $markerEnd - $markerStart );
5715 $i = $markerEnd;
5716 }
5717 }
5718 }
5719 return $out;
5720 }
5721
5722 /**
5723 * Remove any strip markers found in the given text.
5724 *
5725 * @param $text Input string
5726 * @return string
5727 */
5728 function killMarkers( $text ) {
5729 return $this->mStripState->killMarkers( $text );
5730 }
5731
5732 /**
5733 * Save the parser state required to convert the given half-parsed text to
5734 * HTML. "Half-parsed" in this context means the output of
5735 * recursiveTagParse() or internalParse(). This output has strip markers
5736 * from replaceVariables (extensionSubstitution() etc.), and link
5737 * placeholders from replaceLinkHolders().
5738 *
5739 * Returns an array which can be serialized and stored persistently. This
5740 * array can later be loaded into another parser instance with
5741 * unserializeHalfParsedText(). The text can then be safely incorporated into
5742 * the return value of a parser hook.
5743 *
5744 * @param $text string
5745 *
5746 * @return array
5747 */
5748 function serializeHalfParsedText( $text ) {
5749 wfProfileIn( __METHOD__ );
5750 $data = array(
5751 'text' => $text,
5752 'version' => self::HALF_PARSED_VERSION,
5753 'stripState' => $this->mStripState->getSubState( $text ),
5754 'linkHolders' => $this->mLinkHolders->getSubArray( $text )
5755 );
5756 wfProfileOut( __METHOD__ );
5757 return $data;
5758 }
5759
5760 /**
5761 * Load the parser state given in the $data array, which is assumed to
5762 * have been generated by serializeHalfParsedText(). The text contents is
5763 * extracted from the array, and its markers are transformed into markers
5764 * appropriate for the current Parser instance. This transformed text is
5765 * returned, and can be safely included in the return value of a parser
5766 * hook.
5767 *
5768 * If the $data array has been stored persistently, the caller should first
5769 * check whether it is still valid, by calling isValidHalfParsedText().
5770 *
5771 * @param $data array Serialized data
5772 * @return String
5773 */
5774 function unserializeHalfParsedText( $data ) {
5775 if ( !isset( $data['version'] ) || $data['version'] != self::HALF_PARSED_VERSION ) {
5776 throw new MWException( __METHOD__.': invalid version' );
5777 }
5778
5779 # First, extract the strip state.
5780 $texts = array( $data['text'] );
5781 $texts = $this->mStripState->merge( $data['stripState'], $texts );
5782
5783 # Now renumber links
5784 $texts = $this->mLinkHolders->mergeForeign( $data['linkHolders'], $texts );
5785
5786 # Should be good to go.
5787 return $texts[0];
5788 }
5789
5790 /**
5791 * Returns true if the given array, presumed to be generated by
5792 * serializeHalfParsedText(), is compatible with the current version of the
5793 * parser.
5794 *
5795 * @param $data Array
5796 *
5797 * @return bool
5798 */
5799 function isValidHalfParsedText( $data ) {
5800 return isset( $data['version'] ) && $data['version'] == self::HALF_PARSED_VERSION;
5801 }
5802
5803 /**
5804 * Parsed a width param of imagelink like 300px or 200x300px
5805 *
5806 * @param $value String
5807 *
5808 * @return array
5809 * @since 1.20
5810 */
5811 public function parseWidthParam( $value ) {
5812 $parsedWidthParam = array();
5813 if( $value === '' ) {
5814 return $parsedWidthParam;
5815 }
5816 $m = array();
5817 # (bug 13500) In both cases (width/height and width only),
5818 # permit trailing "px" for backward compatibility.
5819 if ( preg_match( '/^([0-9]*)x([0-9]*)\s*(?:px)?\s*$/', $value, $m ) ) {
5820 $width = intval( $m[1] );
5821 $height = intval( $m[2] );
5822 $parsedWidthParam['width'] = $width;
5823 $parsedWidthParam['height'] = $height;
5824 } elseif ( preg_match( '/^[0-9]*\s*(?:px)?\s*$/', $value ) ) {
5825 $width = intval( $value );
5826 $parsedWidthParam['width'] = $width;
5827 }
5828 return $parsedWidthParam;
5829 }
5830 }
Wiklou, le Wiki du Wiklou