dépôts / lhc / web / wiklou.git / blob
? search:
summary | shortlog | log | commit | commitdiff | tree
history | raw | HEAD
revert r76464
[lhc/web/wiklou.git] / includes / parser / Parser.php
1 <?php
2 /**
3 * @defgroup Parser Parser
4 *
5 * @file
6 * @ingroup Parser
7 * File for Parser and related classes
8 */
9
10
11 /**
12 * PHP Parser - Processes wiki markup (which uses a more user-friendly
13 * syntax, such as "[[link]]" for making links), and provides a one-way
14 * transformation of that wiki markup it into XHTML output / markup
15 * (which in turn the browser understands, and can display).
16 *
17 * <pre>
18 * There are five main entry points into the Parser class:
19 * parse()
20 * produces HTML output
21 * preSaveTransform().
22 * produces altered wiki markup.
23 * preprocess()
24 * removes HTML comments and expands templates
25 * cleanSig()
26 * Cleans a signature before saving it to preferences
27 * extractSections()
28 * Extracts sections from an article for section editing
29 * getPreloadText()
30 * Removes <noinclude> sections, and <includeonly> tags.
31 *
32 * Globals used:
33 * objects: $wgLang, $wgContLang
34 *
35 * NOT $wgArticle, $wgUser or $wgTitle. Keep them away!
36 *
37 * settings:
38 * $wgUseTex*, $wgUseDynamicDates*, $wgInterwikiMagic*,
39 * $wgNamespacesWithSubpages, $wgAllowExternalImages*,
40 * $wgLocaltimezone, $wgAllowSpecialInclusion*,
41 * $wgMaxArticleSize*
42 *
43 * * only within ParserOptions
44 * </pre>
45 *
46 * @ingroup Parser
47 */
48 class Parser {
49 /**
50 * Update this version number when the ParserOutput format
51 * changes in an incompatible way, so the parser cache
52 * can automatically discard old data.
53 */
54 const VERSION = '1.6.4';
55
56 # Flags for Parser::setFunctionHook
57 # Also available as global constants from Defines.php
58 const SFH_NO_HASH = 1;
59 const SFH_OBJECT_ARGS = 2;
60
61 # Constants needed for external link processing
62 # Everything except bracket, space, or control characters
63 const EXT_LINK_URL_CLASS = '[^][<>"\\x00-\\x20\\x7F]';
64 const EXT_IMAGE_REGEX = '/^(http:\/\/|https:\/\/)([^][<>"\\x00-\\x20\\x7F]+)
65 \\/([A-Za-z0-9_.,~%\\-+&;#*?!=()@\\x80-\\xFF]+)\\.((?i)gif|png|jpg|jpeg)$/Sx';
66
67 # State constants for the definition list colon extraction
68 const COLON_STATE_TEXT = 0;
69 const COLON_STATE_TAG = 1;
70 const COLON_STATE_TAGSTART = 2;
71 const COLON_STATE_CLOSETAG = 3;
72 const COLON_STATE_TAGSLASH = 4;
73 const COLON_STATE_COMMENT = 5;
74 const COLON_STATE_COMMENTDASH = 6;
75 const COLON_STATE_COMMENTDASHDASH = 7;
76
77 # Flags for preprocessToDom
78 const PTD_FOR_INCLUSION = 1;
79
80 # Allowed values for $this->mOutputType
81 # Parameter to startExternalParse().
82 const OT_HTML = 1; # like parse()
83 const OT_WIKI = 2; # like preSaveTransform()
84 const OT_PREPROCESS = 3; # like preprocess()
85 const OT_MSG = 3;
86 const OT_PLAIN = 4; # like extractSections() - portions of the original are returned unchanged.
87
88 # Marker Suffix needs to be accessible staticly.
89 const MARKER_SUFFIX = "-QINU\x7f";
90
91 # Persistent:
92 var $mTagHooks, $mTransparentTagHooks, $mFunctionHooks, $mFunctionSynonyms, $mVariables;
93 var $mSubstWords, $mImageParams, $mImageParamsMagicArray, $mStripList, $mMarkerIndex;
94 var $mPreprocessor, $mExtLinkBracketedRegex, $mUrlProtocols, $mDefaultStripList;
95 var $mVarCache, $mConf, $mFunctionTagHooks;
96
97
98 # Cleared with clearState():
99 var $mOutput, $mAutonumber, $mDTopen, $mStripState;
100 var $mIncludeCount, $mArgStack, $mLastSection, $mInPre;
101 var $mLinkHolders, $mLinkID;
102 var $mIncludeSizes, $mPPNodeCount, $mDefaultSort;
103 var $mTplExpandCache; # empty-frame expansion cache
104 var $mTplRedirCache, $mTplDomCache, $mHeadings, $mDoubleUnderscores;
105 var $mExpensiveFunctionCount; # number of expensive parser function calls
106
107 # Temporary
108 # These are variables reset at least once per parse regardless of $clearState
109 var $mOptions; # ParserOptions object
110 var $mTitle; # Title context, used for self-link rendering and similar things
111 var $mOutputType; # Output type, one of the OT_xxx constants
112 var $ot; # Shortcut alias, see setOutputType()
113 var $mRevisionId; # ID to display in {{REVISIONID}} tags
114 var $mRevisionTimestamp; # The timestamp of the specified revision ID
115 var $mRevIdForTs; # The revision ID which was used to fetch the timestamp
116
117 /**
118 * Constructor
119 *
120 * @public
121 */
122 function __construct( $conf = array() ) {
123 $this->mConf = $conf;
124 $this->mTagHooks = array();
125 $this->mTransparentTagHooks = array();
126 $this->mFunctionHooks = array();
127 $this->mFunctionTagHooks = array();
128 $this->mFunctionSynonyms = array( 0 => array(), 1 => array() );
129 $this->mDefaultStripList = $this->mStripList = array();
130 $this->mUrlProtocols = wfUrlProtocols();
131 $this->mExtLinkBracketedRegex = '/\[(\b(' . wfUrlProtocols() . ')'.
132 '[^][<>"\\x00-\\x20\\x7F]+) *([^\]\\x00-\\x08\\x0a-\\x1F]*?)\]/S';
133 $this->mVarCache = array();
134 if ( isset( $conf['preprocessorClass'] ) ) {
135 $this->mPreprocessorClass = $conf['preprocessorClass'];
136 } elseif ( extension_loaded( 'domxml' ) ) {
137 # PECL extension that conflicts with the core DOM extension (bug 13770)
138 wfDebug( "Warning: you have the obsolete domxml extension for PHP. Please remove it!\n" );
139 $this->mPreprocessorClass = 'Preprocessor_Hash';
140 } elseif ( extension_loaded( 'dom' ) ) {
141 $this->mPreprocessorClass = 'Preprocessor_DOM';
142 } else {
143 $this->mPreprocessorClass = 'Preprocessor_Hash';
144 }
145 $this->mMarkerIndex = 0;
146 $this->mFirstCall = true;
147 }
148
149 /**
150 * Reduce memory usage to reduce the impact of circular references
151 */
152 function __destruct() {
153 if ( isset( $this->mLinkHolders ) ) {
154 $this->mLinkHolders->__destruct();
155 }
156 foreach ( $this as $name => $value ) {
157 unset( $this->$name );
158 }
159 }
160
161 /**
162 * Do various kinds of initialisation on the first call of the parser
163 */
164 function firstCallInit() {
165 if ( !$this->mFirstCall ) {
166 return;
167 }
168 $this->mFirstCall = false;
169
170 wfProfileIn( __METHOD__ );
171
172 CoreParserFunctions::register( $this );
173 CoreTagHooks::register( $this );
174 $this->initialiseVariables();
175
176 wfRunHooks( 'ParserFirstCallInit', array( &$this ) );
177 wfProfileOut( __METHOD__ );
178 }
179
180 /**
181 * Clear Parser state
182 *
183 * @private
184 */
185 function clearState() {
186 wfProfileIn( __METHOD__ );
187 if ( $this->mFirstCall ) {
188 $this->firstCallInit();
189 }
190 $this->mOutput = new ParserOutput;
191 $this->mAutonumber = 0;
192 $this->mLastSection = '';
193 $this->mDTopen = false;
194 $this->mIncludeCount = array();
195 $this->mStripState = new StripState;
196 $this->mArgStack = false;
197 $this->mInPre = false;
198 $this->mLinkHolders = new LinkHolderArray( $this );
199 $this->mLinkID = 0;
200 $this->mRevisionTimestamp = $this->mRevisionId = null;
201 $this->mVarCache = array();
202
203 /**
204 * Prefix for temporary replacement strings for the multipass parser.
205 * \x07 should never appear in input as it's disallowed in XML.
206 * Using it at the front also gives us a little extra robustness
207 * since it shouldn't match when butted up against identifier-like
208 * string constructs.
209 *
210 * Must not consist of all title characters, or else it will change
211 * the behaviour of <nowiki> in a link.
212 */
213 # $this->mUniqPrefix = "\x07UNIQ" . Parser::getRandomString();
214 # Changed to \x7f to allow XML double-parsing -- TS
215 $this->mUniqPrefix = "\x7fUNIQ" . self::getRandomString();
216
217
218 # Clear these on every parse, bug 4549
219 $this->mTplExpandCache = $this->mTplRedirCache = $this->mTplDomCache = array();
220
221 $this->mShowToc = true;
222 $this->mForceTocPosition = false;
223 $this->mIncludeSizes = array(
224 'post-expand' => 0,
225 'arg' => 0,
226 );
227 $this->mPPNodeCount = 0;
228 $this->mDefaultSort = false;
229 $this->mHeadings = array();
230 $this->mDoubleUnderscores = array();
231 $this->mExpensiveFunctionCount = 0;
232
233 # Fix cloning
234 if ( isset( $this->mPreprocessor ) && $this->mPreprocessor->parser !== $this ) {
235 $this->mPreprocessor = null;
236 }
237
238 wfRunHooks( 'ParserClearState', array( &$this ) );
239 wfProfileOut( __METHOD__ );
240 }
241
242 /**
243 * Convert wikitext to HTML
244 * Do not call this function recursively.
245 *
246 * @param $text String: text we want to parse
247 * @param $title A title object
248 * @param $options ParserOptions
249 * @param $linestart boolean
250 * @param $clearState boolean
251 * @param $revid Int: number to pass in {{REVISIONID}}
252 * @return ParserOutput a ParserOutput
253 */
254 public function parse( $text, Title $title, ParserOptions $options, $linestart = true, $clearState = true, $revid = null ) {
255 /**
256 * First pass--just handle <nowiki> sections, pass the rest off
257 * to internalParse() which does all the real work.
258 */
259
260 global $wgUseTidy, $wgAlwaysUseTidy, $wgContLang, $wgDisableLangConversion, $wgDisableTitleConversion;
261 $fname = __METHOD__.'-' . wfGetCaller();
262 wfProfileIn( __METHOD__ );
263 wfProfileIn( $fname );
264
265 if ( $clearState ) {
266 $this->clearState();
267 }
268
269 $options->resetUsage();
270 $this->mOptions = $options;
271 $this->setTitle( $title ); # Page title has to be set for the pre-processor
272
273 $oldRevisionId = $this->mRevisionId;
274 $oldRevisionTimestamp = $this->mRevisionTimestamp;
275 if ( $revid !== null ) {
276 $this->mRevisionId = $revid;
277 $this->mRevisionTimestamp = null;
278 }
279 $this->setOutputType( self::OT_HTML );
280 wfRunHooks( 'ParserBeforeStrip', array( &$this, &$text, &$this->mStripState ) );
281 # No more strip!
282 wfRunHooks( 'ParserAfterStrip', array( &$this, &$text, &$this->mStripState ) );
283 $text = $this->internalParse( $text );
284
285 $text = $this->mStripState->unstripGeneral( $text );
286
287 # Clean up special characters, only run once, next-to-last before doBlockLevels
288 $fixtags = array(
289 # french spaces, last one Guillemet-left
290 # only if there is something before the space
291 '/(.) (?=\\?|:|;|!|%|\\302\\273)/' => '\\1&#160;\\2',
292 # french spaces, Guillemet-right
293 '/(\\302\\253) /' => '\\1&#160;',
294 '/&#160;(!\s*important)/' => ' \\1', # Beware of CSS magic word !important, bug #11874.
295 );
296 $text = preg_replace( array_keys( $fixtags ), array_values( $fixtags ), $text );
297
298 $text = $this->doBlockLevels( $text, $linestart );
299
300 $this->replaceLinkHolders( $text );
301
302 /**
303 * The page doesn't get language converted if
304 * a) It's disabled
305 * b) Content isn't converted
306 * c) It's a conversion table
307 */
308 if ( !( $wgDisableLangConversion
309 || isset( $this->mDoubleUnderscores['nocontentconvert'] )
310 || $this->mTitle->isConversionTable() ) ) {
311
312 # The position of the convert() call should not be changed. it
313 # assumes that the links are all replaced and the only thing left
314 # is the <nowiki> mark.
315
316 $text = $wgContLang->convert( $text );
317 }
318
319 /**
320 * A page get its title converted except:
321 * a) Language conversion is globally disabled
322 * b) Title convert is globally disabled
323 * c) The page is a redirect page
324 * d) User request with a "linkconvert" set to "no"
325 * e) A "nocontentconvert" magic word has been set
326 * f) A "notitleconvert" magic word has been set
327 * g) User sets "noconvertlink" in his/her preference
328 *
329 * Note that if a user tries to set a title in a conversion
330 * rule but content conversion was not done, then the parser
331 * won't pick it up. This is probably expected behavior.
332 */
333 if ( !( $wgDisableLangConversion
334 || $wgDisableTitleConversion
335 || isset( $this->mDoubleUnderscores['nocontentconvert'] )
336 || isset( $this->mDoubleUnderscores['notitleconvert'] )
337 || $this->mOutput->getDisplayTitle() !== false ) )
338 {
339 $convruletitle = $wgContLang->getConvRuleTitle();
340 if ( $convruletitle ) {
341 $this->mOutput->setTitleText( $convruletitle );
342 } else {
343 $titleText = $wgContLang->convertTitle( $title );
344 $this->mOutput->setTitleText( $titleText );
345 }
346 }
347
348 $text = $this->mStripState->unstripNoWiki( $text );
349
350 wfRunHooks( 'ParserBeforeTidy', array( &$this, &$text ) );
351
352 //!JF Move to its own function
353
354 $uniq_prefix = $this->mUniqPrefix;
355 $matches = array();
356 $elements = array_keys( $this->mTransparentTagHooks );
357 $text = self::extractTagsAndParams( $elements, $text, $matches, $uniq_prefix );
358
359 foreach ( $matches as $marker => $data ) {
360 list( $element, $content, $params, $tag ) = $data;
361 $tagName = strtolower( $element );
362 if ( isset( $this->mTransparentTagHooks[$tagName] ) ) {
363 $output = call_user_func_array( $this->mTransparentTagHooks[$tagName], array( $content, $params, $this ) );
364 } else {
365 $output = $tag;
366 }
367 $this->mStripState->general->setPair( $marker, $output );
368 }
369 $text = $this->mStripState->unstripGeneral( $text );
370
371 $text = Sanitizer::normalizeCharReferences( $text );
372
373 if ( ( $wgUseTidy && $this->mOptions->getTidy() ) || $wgAlwaysUseTidy ) {
374 $text = MWTidy::tidy( $text );
375 } else {
376 # attempt to sanitize at least some nesting problems
377 # (bug #2702 and quite a few others)
378 $tidyregs = array(
379 # ''Something [http://www.cool.com cool''] -->
380 # <i>Something</i><a href="http://www.cool.com"..><i>cool></i></a>
381 '/(<([bi])>)(<([bi])>)?([^<]*)(<\/?a[^<]*>)([^<]*)(<\/\\4>)?(<\/\\2>)/' =>
382 '\\1\\3\\5\\8\\9\\6\\1\\3\\7\\8\\9',
383 # fix up an anchor inside another anchor, only
384 # at least for a single single nested link (bug 3695)
385 '/(<a[^>]+>)([^<]*)(<a[^>]+>[^<]*)<\/a>(.*)<\/a>/' =>
386 '\\1\\2</a>\\3</a>\\1\\4</a>',
387 # fix div inside inline elements- doBlockLevels won't wrap a line which
388 # contains a div, so fix it up here; replace
389 # div with escaped text
390 '/(<([aib]) [^>]+>)([^<]*)(<div([^>]*)>)(.*)(<\/div>)([^<]*)(<\/\\2>)/' =>
391 '\\1\\3&lt;div\\5&gt;\\6&lt;/div&gt;\\8\\9',
392 # remove empty italic or bold tag pairs, some
393 # introduced by rules above
394 '/<([bi])><\/\\1>/' => '',
395 );
396
397 $text = preg_replace(
398 array_keys( $tidyregs ),
399 array_values( $tidyregs ),
400 $text );
401 }
402 global $wgExpensiveParserFunctionLimit;
403 if ( $this->mExpensiveFunctionCount > $wgExpensiveParserFunctionLimit ) {
404 $this->limitationWarn( 'expensive-parserfunction', $this->mExpensiveFunctionCount, $wgExpensiveParserFunctionLimit );
405 }
406
407 wfRunHooks( 'ParserAfterTidy', array( &$this, &$text ) );
408
409 # Information on include size limits, for the benefit of users who try to skirt them
410 if ( $this->mOptions->getEnableLimitReport() ) {
411 $max = $this->mOptions->getMaxIncludeSize();
412 $PFreport = "Expensive parser function count: {$this->mExpensiveFunctionCount}/$wgExpensiveParserFunctionLimit\n";
413 $limitReport =
414 "NewPP limit report\n" .
415 "Preprocessor node count: {$this->mPPNodeCount}/{$this->mOptions->getMaxPPNodeCount()}\n" .
416 "Post-expand include size: {$this->mIncludeSizes['post-expand']}/$max bytes\n" .
417 "Template argument size: {$this->mIncludeSizes['arg']}/$max bytes\n".
418 $PFreport;
419 wfRunHooks( 'ParserLimitReport', array( $this, &$limitReport ) );
420 $text .= "\n<!-- \n$limitReport-->\n";
421 }
422 $this->mOutput->setText( $text );
423
424 $this->mRevisionId = $oldRevisionId;
425 $this->mRevisionTimestamp = $oldRevisionTimestamp;
426 wfProfileOut( $fname );
427 wfProfileOut( __METHOD__ );
428
429 return $this->mOutput;
430 }
431
432 /**
433 * Recursive parser entry point that can be called from an extension tag
434 * hook.
435 *
436 * If $frame is not provided, then template variables (e.g., {{{1}}}) within $text are not expanded
437 *
438 * @param $text String: text extension wants to have parsed
439 * @param $frame PPFrame: The frame to use for expanding any template variables
440 */
441 function recursiveTagParse( $text, $frame=false ) {
442 wfProfileIn( __METHOD__ );
443 wfRunHooks( 'ParserBeforeStrip', array( &$this, &$text, &$this->mStripState ) );
444 wfRunHooks( 'ParserAfterStrip', array( &$this, &$text, &$this->mStripState ) );
445 $text = $this->internalParse( $text, false, $frame );
446 wfProfileOut( __METHOD__ );
447 return $text;
448 }
449
450 /**
451 * Expand templates and variables in the text, producing valid, static wikitext.
452 * Also removes comments.
453 */
454 function preprocess( $text, $title, $options, $revid = null ) {
455 wfProfileIn( __METHOD__ );
456 $this->clearState();
457 $this->setOutputType( self::OT_PREPROCESS );
458 $options->resetUsage();
459 $this->mOptions = $options;
460 $this->setTitle( $title );
461 if ( $revid !== null ) {
462 $this->mRevisionId = $revid;
463 }
464 wfRunHooks( 'ParserBeforeStrip', array( &$this, &$text, &$this->mStripState ) );
465 wfRunHooks( 'ParserAfterStrip', array( &$this, &$text, &$this->mStripState ) );
466 $text = $this->replaceVariables( $text );
467 $text = $this->mStripState->unstripBoth( $text );
468 wfProfileOut( __METHOD__ );
469 return $text;
470 }
471
472 /**
473 * Process the wikitext for the ?preload= feature. (bug 5210)
474 *
475 * <noinclude>, <includeonly> etc. are parsed as for template transclusion,
476 * comments, templates, arguments, tags hooks and parser functions are untouched.
477 */
478 public function getPreloadText( $text, $title, $options ) {
479 # Parser (re)initialisation
480 $this->clearState();
481 $this->setOutputType( self::OT_PLAIN );
482 $options->resetUsage();
483 $this->mOptions = $options;
484 $this->setTitle( $title );
485
486 $flags = PPFrame::NO_ARGS | PPFrame::NO_TEMPLATES;
487 $dom = $this->preprocessToDom( $text, self::PTD_FOR_INCLUSION );
488 return $this->getPreprocessor()->newFrame()->expand( $dom, $flags );
489 }
490
491 /**
492 * Get a random string
493 *
494 * @private
495 * @static
496 */
497 static private function getRandomString() {
498 return dechex( mt_rand( 0, 0x7fffffff ) ) . dechex( mt_rand( 0, 0x7fffffff ) );
499 }
500
501 /**
502 * Accessor for mUniqPrefix.
503 *
504 * @return String
505 */
506 public function uniqPrefix() {
507 if ( !isset( $this->mUniqPrefix ) ) {
508 # @todo Fixme: this is probably *horribly wrong*
509 # LanguageConverter seems to want $wgParser's uniqPrefix, however
510 # if this is called for a parser cache hit, the parser may not
511 # have ever been initialized in the first place.
512 # Not really sure what the heck is supposed to be going on here.
513 return '';
514 # throw new MWException( "Accessing uninitialized mUniqPrefix" );
515 }
516 return $this->mUniqPrefix;
517 }
518
519 /**
520 * Set the context title
521 */
522 function setTitle( $t ) {
523 if ( !$t || $t instanceof FakeTitle ) {
524 $t = Title::newFromText( 'NO TITLE' );
525 }
526
527 if ( strval( $t->getFragment() ) !== '' ) {
528 # Strip the fragment to avoid various odd effects
529 $this->mTitle = clone $t;
530 $this->mTitle->setFragment( '' );
531 } else {
532 $this->mTitle = $t;
533 }
534 }
535
536 /**
537 * Accessor for the Title object
538 *
539 * @return Title object
540 */
541 function getTitle() {
542 return $this->mTitle;
543 }
544
545 /**
546 * Accessor/mutator for the Title object
547 *
548 * @param $x New Title object or null to just get the current one
549 * @return Title object
550 */
551 function Title( $x = null ) {
552 return wfSetVar( $this->mTitle, $x );
553 }
554
555 /**
556 * Set the output type
557 *
558 * @param $ot Integer: new value
559 */
560 function setOutputType( $ot ) {
561 $this->mOutputType = $ot;
562 # Shortcut alias
563 $this->ot = array(
564 'html' => $ot == self::OT_HTML,
565 'wiki' => $ot == self::OT_WIKI,
566 'pre' => $ot == self::OT_PREPROCESS,
567 'plain' => $ot == self::OT_PLAIN,
568 );
569 }
570
571 /**
572 * Accessor/mutator for the output type
573 *
574 * @param $x New value or null to just get the current one
575 * @return Integer
576 */
577 function OutputType( $x = null ) {
578 return wfSetVar( $this->mOutputType, $x );
579 }
580
581 /**
582 * Get the ParserOutput object
583 *
584 * @return ParserOutput object
585 */
586 function getOutput() {
587 return $this->mOutput;
588 }
589
590 /**
591 * Get the ParserOptions object
592 *
593 * @return ParserOptions object
594 */
595 function getOptions() {
596 return $this->mOptions;
597 }
598
599 /**
600 * Accessor/mutator for the ParserOptions object
601 *
602 * @param $x New value or null to just get the current one
603 * @return Current ParserOptions object
604 */
605 function Options( $x = null ) {
606 return wfSetVar( $this->mOptions, $x );
607 }
608
609 function nextLinkID() {
610 return $this->mLinkID++;
611 }
612
613 function getFunctionLang() {
614 global $wgLang, $wgContLang;
615
616 $target = $this->mOptions->getTargetLanguage();
617 if ( $target !== null ) {
618 return $target;
619 } else {
620 return $this->mOptions->getInterfaceMessage() ? $wgLang : $wgContLang;
621 }
622 }
623
624 /**
625 * Get a preprocessor object
626 *
627 * @return Preprocessor instance
628 */
629 function getPreprocessor() {
630 if ( !isset( $this->mPreprocessor ) ) {
631 $class = $this->mPreprocessorClass;
632 $this->mPreprocessor = new $class( $this );
633 }
634 return $this->mPreprocessor;
635 }
636
637 /**
638 * Replaces all occurrences of HTML-style comments and the given tags
639 * in the text with a random marker and returns the next text. The output
640 * parameter $matches will be an associative array filled with data in
641 * the form:
642 * 'UNIQ-xxxxx' => array(
643 * 'element',
644 * 'tag content',
645 * array( 'param' => 'x' ),
646 * '<element param="x">tag content</element>' ) )
647 *
648 * @param $elements list of element names. Comments are always extracted.
649 * @param $text Source text string.
650 * @param $matches Out parameter, Array: extracted tags
651 * @param $uniq_prefix
652 * @return String: stripped text
653 *
654 * @static
655 */
656 public function extractTagsAndParams( $elements, $text, &$matches, $uniq_prefix = '' ) {
657 static $n = 1;
658 $stripped = '';
659 $matches = array();
660
661 $taglist = implode( '|', $elements );
662 $start = "/<($taglist)(\\s+[^>]*?|\\s*?)(\/?" . ">)|<(!--)/i";
663
664 while ( $text != '' ) {
665 $p = preg_split( $start, $text, 2, PREG_SPLIT_DELIM_CAPTURE );
666 $stripped .= $p[0];
667 if ( count( $p ) < 5 ) {
668 break;
669 }
670 if ( count( $p ) > 5 ) {
671 # comment
672 $element = $p[4];
673 $attributes = '';
674 $close = '';
675 $inside = $p[5];
676 } else {
677 # tag
678 $element = $p[1];
679 $attributes = $p[2];
680 $close = $p[3];
681 $inside = $p[4];
682 }
683
684 $marker = "$uniq_prefix-$element-" . sprintf( '%08X', $n++ ) . self::MARKER_SUFFIX;
685 $stripped .= $marker;
686
687 if ( $close === '/>' ) {
688 # Empty element tag, <tag />
689 $content = null;
690 $text = $inside;
691 $tail = null;
692 } else {
693 if ( $element === '!--' ) {
694 $end = '/(-->)/';
695 } else {
696 $end = "/(<\\/$element\\s*>)/i";
697 }
698 $q = preg_split( $end, $inside, 2, PREG_SPLIT_DELIM_CAPTURE );
699 $content = $q[0];
700 if ( count( $q ) < 3 ) {
701 # No end tag -- let it run out to the end of the text.
702 $tail = '';
703 $text = '';
704 } else {
705 $tail = $q[1];
706 $text = $q[2];
707 }
708 }
709
710 $matches[$marker] = array( $element,
711 $content,
712 Sanitizer::decodeTagAttributes( $attributes ),
713 "<$element$attributes$close$content$tail" );
714 }
715 return $stripped;
716 }
717
718 /**
719 * Get a list of strippable XML-like elements
720 */
721 function getStripList() {
722 return $this->mStripList;
723 }
724
725 /**
726 * @deprecated use replaceVariables
727 */
728 function strip( $text, $state, $stripcomments = false , $dontstrip = array() ) {
729 return $text;
730 }
731
732 /**
733 * Restores pre, math, and other extensions removed by strip()
734 *
735 * always call unstripNoWiki() after this one
736 * @private
737 * @deprecated use $this->mStripState->unstrip()
738 */
739 function unstrip( $text, $state ) {
740 return $state->unstripGeneral( $text );
741 }
742
743 /**
744 * Always call this after unstrip() to preserve the order
745 *
746 * @private
747 * @deprecated use $this->mStripState->unstrip()
748 */
749 function unstripNoWiki( $text, $state ) {
750 return $state->unstripNoWiki( $text );
751 }
752
753 /**
754 * @deprecated use $this->mStripState->unstripBoth()
755 */
756 function unstripForHTML( $text ) {
757 return $this->mStripState->unstripBoth( $text );
758 }
759
760 /**
761 * Add an item to the strip state
762 * Returns the unique tag which must be inserted into the stripped text
763 * The tag will be replaced with the original text in unstrip()
764 *
765 * @private
766 */
767 function insertStripItem( $text ) {
768 $rnd = "{$this->mUniqPrefix}-item-{$this->mMarkerIndex}-" . self::MARKER_SUFFIX;
769 $this->mMarkerIndex++;
770 $this->mStripState->general->setPair( $rnd, $text );
771 return $rnd;
772 }
773
774 /**
775 * Interface with html tidy
776 * @deprecated Use MWTidy::tidy()
777 */
778 public static function tidy( $text ) {
779 wfDeprecated( __METHOD__ );
780 return MWTidy::tidy( $text );
781 }
782
783 /**
784 * parse the wiki syntax used to render tables
785 *
786 * @private
787 */
788 function doTableStuff( $text ) {
789 wfProfileIn( __METHOD__ );
790
791 $lines = StringUtils::explode( "\n", $text );
792 $text = null;
793 $out = '';
794 $td_history = array(); # Is currently a td tag open?
795 $last_tag_history = array(); # Save history of last lag activated (td, th or caption)
796 $tr_history = array(); # Is currently a tr tag open?
797 $tr_attributes = array(); # history of tr attributes
798 $has_opened_tr = array(); # Did this table open a <tr> element?
799 $indent_level = 0; # indent level of the table
800
801 $table_tag = 'table';
802 $tr_tag = 'tr';
803 $th_tag = 'th';
804 $td_tag = 'td';
805 $caption_tag = 'caption';
806
807 $extra_table_attribs = null;
808 $extra_tr_attribs = null;
809 $extra_td_attribs = null;
810
811 $convert_style = false;
812
813 foreach ( $lines as $outLine ) {
814 $line = trim( $outLine );
815
816 if ( $line === '' ) { # empty line, go to next line
817 $out .= $outLine."\n";
818 continue;
819 }
820
821 $first_character = $line[0];
822 $matches = array();
823
824 if ( preg_match( '/^(:*)\{\|(.*)$/', $line , $matches ) ) {
825 # First check if we are starting a new table
826 $indent_level = strlen( $matches[1] );
827
828 $attributes = $this->mStripState->unstripBoth( $matches[2] );
829
830 $attr = Sanitizer::decodeTagAttributes( $attributes );
831
832 $mode = @$attr['mode'];
833 if ( !$mode ) $mode = 'data';
834
835 if ( $mode == 'grid' || $mode == 'layout' ) {
836 $table_tag = 'div';
837 $tr_tag = 'div';
838 $th_tag = 'div';
839 $td_tag = 'div';
840 $caption_tag = 'div';
841
842 $extra_table_attribs = array( 'class' => 'grid-table' );
843 $extra_tr_attribs = array( 'class' => 'grid-row' );
844 $extra_td_attribs = array( 'class' => 'grid-cell' );
845
846 $convert_style = true;
847 }
848
849 if ($convert_style) $attr['style'] = Sanitizer::styleFromAttributes( $attr );
850 $attr = Sanitizer::validateTagAttributes( $attr, $table_tag );
851 $attributes = Sanitizer::collapseTagAttributes( $attr, $extra_table_attribs );
852
853 $outLine = str_repeat( '<dl><dd>' , $indent_level ) . "<$table_tag{$attributes}>";
854 array_push( $td_history , false );
855 array_push( $last_tag_history , '' );
856 array_push( $tr_history , false );
857 array_push( $tr_attributes , '' );
858 array_push( $has_opened_tr , false );
859 } elseif ( count( $td_history ) == 0 ) {
860 # Don't do any of the following
861 $out .= $outLine."\n";
862 continue;
863 } elseif ( substr( $line , 0 , 2 ) === '|}' ) {
864 # We are ending a table
865 $line = "</$table_tag>" . substr( $line , 2 );
866 $last_tag = array_pop( $last_tag_history );
867
868 if ( !array_pop( $has_opened_tr ) ) {
869 $line = "<$tr_tag><$td_tag></$td_tag></$tr_tag>{$line}";
870 }
871
872 if ( array_pop( $tr_history ) ) {
873 $line = "</$tr_tag>{$line}";
874 }
875
876 if ( array_pop( $td_history ) ) {
877 $line = "</{$last_tag}>{$line}";
878 }
879 array_pop( $tr_attributes );
880 $outLine = $line . str_repeat( '</dd></dl>' , $indent_level );
881 } elseif ( substr( $line , 0 , 2 ) === '|-' ) {
882 # Now we have a table row
883 $line = preg_replace( '#^\|-+#', '', $line );
884
885 # Whats after the tag is now only attributes
886 $attributes = $this->mStripState->unstripBoth( $line );
887
888 $attr = Sanitizer::decodeTagAttributes( $attributes );
889 if ($convert_style) $attr['style'] = Sanitizer::styleFromAttributes( $attr );
890 $attr = Sanitizer::validateTagAttributes( $attr, $tr_tag );
891 $attributes = Sanitizer::collapseTagAttributes( $attr, $extra_tr_attribs );
892
893 array_pop( $tr_attributes );
894 array_push( $tr_attributes, $attributes );
895
896 $line = '';
897 $last_tag = array_pop( $last_tag_history );
898 array_pop( $has_opened_tr );
899 array_push( $has_opened_tr , true );
900
901 if ( array_pop( $tr_history ) ) {
902 $line = "</$tr_tag>";
903 }
904
905 if ( array_pop( $td_history ) ) {
906 $line = "</{$last_tag}>{$line}";
907 }
908
909 $outLine = $line;
910 array_push( $tr_history , false );
911 array_push( $td_history , false );
912 array_push( $last_tag_history , '' );
913 } elseif ( $first_character === '|' || $first_character === '!' || substr( $line , 0 , 2 ) === '|+' ) {
914 # This might be cell elements, td, th or captions
915 if ( substr( $line , 0 , 2 ) === '|+' ) {
916 $first_character = '+';
917 $line = substr( $line , 1 );
918 }
919
920 $line = substr( $line , 1 );
921
922 if ( $first_character === '!' ) {
923 $line = str_replace( '!!' , '||' , $line );
924 }
925
926 # Split up multiple cells on the same line.
927 # FIXME : This can result in improper nesting of tags processed
928 # by earlier parser steps, but should avoid splitting up eg
929 # attribute values containing literal "||".
930 $cells = StringUtils::explodeMarkup( '||' , $line );
931
932 $outLine = '';
933
934 # Loop through each table cell
935 foreach ( $cells as $cell ) {
936 $previous = '';
937 if ( $first_character !== '+' ) {
938 $tr_after = array_pop( $tr_attributes );
939 if ( !array_pop( $tr_history ) ) {
940 $previous = "<$tr_tag{$tr_after}>\n";
941 }
942 array_push( $tr_history , true );
943 array_push( $tr_attributes , '' );
944 array_pop( $has_opened_tr );
945 array_push( $has_opened_tr , true );
946 }
947
948 $last_tag = array_pop( $last_tag_history );
949
950 if ( array_pop( $td_history ) ) {
951 $previous = "</{$last_tag}>\n{$previous}";
952 }
953
954 if ( $first_character === '|' ) {
955 $last_tag = $td_tag;
956 } elseif ( $first_character === '!' ) {
957 $last_tag = $th_tag;
958 } elseif ( $first_character === '+' ) {
959 $last_tag = $caption_tag;
960 } else {
961 $last_tag = '';
962 }
963
964 array_push( $last_tag_history , $last_tag );
965
966 # A cell could contain both parameters and data
967 $cell_data = explode( '|' , $cell , 2 );
968
969 $attributes = '';
970
971 # Bug 553: Note that a '|' inside an invalid link should not
972 # be mistaken as delimiting cell parameters
973 if ( strpos( $cell_data[0], '[[' ) !== false ) {
974 if ($extra_td_attribs) $attributes = Sanitizer::collapseTagAttributes( $extra_td_attribs );
975 $cell = "{$previous}<{$last_tag}{$attributes}>{$cell}";
976 } elseif ( count( $cell_data ) == 1 ) {
977 if ($extra_td_attribs) $attributes = Sanitizer::collapseTagAttributes( $extra_td_attribs );
978 $cell = "{$previous}<{$last_tag}{$attributes}>{$cell_data[0]}";
979 } else {
980 $attributes = $this->mStripState->unstripBoth( $cell_data[0] );
981
982 $attr = Sanitizer::decodeTagAttributes( $attributes );
983 if ($convert_style) $attr['style'] = Sanitizer::styleFromAttributes( $attr );
984 $attr = Sanitizer::validateTagAttributes( $attr, $last_tag );
985 $attributes = Sanitizer::collapseTagAttributes( $attr, $extra_td_attribs );
986
987 $cell = "{$previous}<{$last_tag}{$attributes}>{$cell_data[1]}";
988 }
989
990 $outLine .= $cell;
991 array_push( $td_history , true );
992 }
993 }
994 $out .= $outLine . "\n";
995 }
996
997 # Closing open td, tr && table
998 while ( count( $td_history ) > 0 ) {
999 if ( array_pop( $td_history ) ) {
1000 $out .= "</$td_tag>\n";
1001 }
1002 if ( array_pop( $tr_history ) ) {
1003 $out .= "</$tr_tag>\n";
1004 }
1005 if ( !array_pop( $has_opened_tr ) ) {
1006 $out .= "<$tr_tag><$td_tag></$td_tag></$tr_tag>\n" ;
1007 }
1008
1009 $out .= "</$table_tag>\n";
1010 }
1011
1012 # Remove trailing line-ending (b/c)
1013 if ( substr( $out, -1 ) === "\n" ) {
1014 $out = substr( $out, 0, -1 );
1015 }
1016
1017 # special case: don't return empty table
1018 if ( $out === "<$table_tag>\n<$tr_tag><$td_tag></$td_tag></$tr_tag>\n</$table_tag>" ) {
1019 $out = '';
1020 }
1021
1022 wfProfileOut( __METHOD__ );
1023
1024 return $out;
1025 }
1026
1027 /**
1028 * Helper function for parse() that transforms wiki markup into
1029 * HTML. Only called for $mOutputType == self::OT_HTML.
1030 *
1031 * @private
1032 */
1033 function internalParse( $text, $isMain = true, $frame=false ) {
1034 wfProfileIn( __METHOD__ );
1035
1036 $origText = $text;
1037
1038 # Hook to suspend the parser in this state
1039 if ( !wfRunHooks( 'ParserBeforeInternalParse', array( &$this, &$text, &$this->mStripState ) ) ) {
1040 wfProfileOut( __METHOD__ );
1041 return $text ;
1042 }
1043
1044 # if $frame is provided, then use $frame for replacing any variables
1045 if ( $frame ) {
1046 # use frame depth to infer how include/noinclude tags should be handled
1047 # depth=0 means this is the top-level document; otherwise it's an included document
1048 if ( !$frame->depth ) {
1049 $flag = 0;
1050 } else {
1051 $flag = Parser::PTD_FOR_INCLUSION;
1052 }
1053 $dom = $this->preprocessToDom( $text, $flag );
1054 $text = $frame->expand( $dom );
1055 } else {
1056 # if $frame is not provided, then use old-style replaceVariables
1057 $text = $this->replaceVariables( $text );
1058 }
1059
1060 $text = Sanitizer::removeHTMLtags( $text, array( &$this, 'attributeStripCallback' ), false, array_keys( $this->mTransparentTagHooks ) );
1061 wfRunHooks( 'InternalParseBeforeLinks', array( &$this, &$text, &$this->mStripState ) );
1062
1063 # Tables need to come after variable replacement for things to work
1064 # properly; putting them before other transformations should keep
1065 # exciting things like link expansions from showing up in surprising
1066 # places.
1067 $text = $this->doTableStuff( $text );
1068
1069 $text = preg_replace( '/(^|\n)-----*/', '\\1<hr />', $text );
1070
1071 $text = $this->doDoubleUnderscore( $text );
1072
1073 $text = $this->doHeadings( $text );
1074 if ( $this->mOptions->getUseDynamicDates() ) {
1075 $df = DateFormatter::getInstance();
1076 $text = $df->reformat( $this->mOptions->getDateFormat(), $text );
1077 }
1078 $text = $this->replaceInternalLinks( $text );
1079 $text = $this->doAllQuotes( $text );
1080 $text = $this->replaceExternalLinks( $text );
1081
1082 # replaceInternalLinks may sometimes leave behind
1083 # absolute URLs, which have to be masked to hide them from replaceExternalLinks
1084 $text = str_replace( $this->mUniqPrefix.'NOPARSE', '', $text );
1085
1086 $text = $this->doMagicLinks( $text );
1087 $text = $this->formatHeadings( $text, $origText, $isMain );
1088
1089 wfProfileOut( __METHOD__ );
1090 return $text;
1091 }
1092
1093 /**
1094 * Replace special strings like "ISBN xxx" and "RFC xxx" with
1095 * magic external links.
1096 *
1097 * DML
1098 * @private
1099 */
1100 function doMagicLinks( $text ) {
1101 wfProfileIn( __METHOD__ );
1102 $prots = $this->mUrlProtocols;
1103 $urlChar = self::EXT_LINK_URL_CLASS;
1104 $text = preg_replace_callback(
1105 '!(?: # Start cases
1106 (<a[ \t\r\n>].*?</a>) | # m[1]: Skip link text
1107 (<.*?>) | # m[2]: Skip stuff inside HTML elements' . "
1108 (\\b(?:$prots)$urlChar+) | # m[3]: Free external links" . '
1109 (?:RFC|PMID)\s+([0-9]+) | # m[4]: RFC or PMID, capture number
1110 ISBN\s+(\b # m[5]: ISBN, capture number
1111 (?: 97[89] [\ \-]? )? # optional 13-digit ISBN prefix
1112 (?: [0-9] [\ \-]? ){9} # 9 digits with opt. delimiters
1113 [0-9Xx] # check digit
1114 \b)
1115 )!x', array( &$this, 'magicLinkCallback' ), $text );
1116 wfProfileOut( __METHOD__ );
1117 return $text;
1118 }
1119
1120 function magicLinkCallback( $m ) {
1121 if ( isset( $m[1] ) && $m[1] !== '' ) {
1122 # Skip anchor
1123 return $m[0];
1124 } elseif ( isset( $m[2] ) && $m[2] !== '' ) {
1125 # Skip HTML element
1126 return $m[0];
1127 } elseif ( isset( $m[3] ) && $m[3] !== '' ) {
1128 # Free external link
1129 return $this->makeFreeExternalLink( $m[0] );
1130 } elseif ( isset( $m[4] ) && $m[4] !== '' ) {
1131 # RFC or PMID
1132 if ( substr( $m[0], 0, 3 ) === 'RFC' ) {
1133 $keyword = 'RFC';
1134 $urlmsg = 'rfcurl';
1135 $CssClass = 'mw-magiclink-rfc';
1136 $id = $m[4];
1137 } elseif ( substr( $m[0], 0, 4 ) === 'PMID' ) {
1138 $keyword = 'PMID';
1139 $urlmsg = 'pubmedurl';
1140 $CssClass = 'mw-magiclink-pmid';
1141 $id = $m[4];
1142 } else {
1143 throw new MWException( __METHOD__.': unrecognised match type "' .
1144 substr( $m[0], 0, 20 ) . '"' );
1145 }
1146 $url = wfMsgForContent( $urlmsg, $id);
1147 $sk = $this->mOptions->getSkin( $this->mTitle );
1148 $la = $sk->getExternalLinkAttributes( "external $CssClass" );
1149 return "<a href=\"{$url}\"{$la}>{$keyword} {$id}</a>";
1150 } elseif ( isset( $m[5] ) && $m[5] !== '' ) {
1151 # ISBN
1152 $isbn = $m[5];
1153 $num = strtr( $isbn, array(
1154 '-' => '',
1155 ' ' => '',
1156 'x' => 'X',
1157 ));
1158 $titleObj = SpecialPage::getTitleFor( 'Booksources', $num );
1159 return'<a href="' .
1160 $titleObj->escapeLocalUrl() .
1161 "\" class=\"internal mw-magiclink-isbn\">ISBN $isbn</a>";
1162 } else {
1163 return $m[0];
1164 }
1165 }
1166
1167 /**
1168 * Make a free external link, given a user-supplied URL
1169 * @return HTML
1170 * @private
1171 */
1172 function makeFreeExternalLink( $url ) {
1173 global $wgContLang;
1174 wfProfileIn( __METHOD__ );
1175
1176 $sk = $this->mOptions->getSkin( $this->mTitle );
1177 $trail = '';
1178
1179 # The characters '<' and '>' (which were escaped by
1180 # removeHTMLtags()) should not be included in
1181 # URLs, per RFC 2396.
1182 $m2 = array();
1183 if ( preg_match( '/&(lt|gt);/', $url, $m2, PREG_OFFSET_CAPTURE ) ) {
1184 $trail = substr( $url, $m2[0][1] ) . $trail;
1185 $url = substr( $url, 0, $m2[0][1] );
1186 }
1187
1188 # Move trailing punctuation to $trail
1189 $sep = ',;\.:!?';
1190 # If there is no left bracket, then consider right brackets fair game too
1191 if ( strpos( $url, '(' ) === false ) {
1192 $sep .= ')';
1193 }
1194
1195 $numSepChars = strspn( strrev( $url ), $sep );
1196 if ( $numSepChars ) {
1197 $trail = substr( $url, -$numSepChars ) . $trail;
1198 $url = substr( $url, 0, -$numSepChars );
1199 }
1200
1201 $url = Sanitizer::cleanUrl( $url );
1202
1203 # Is this an external image?
1204 $text = $this->maybeMakeExternalImage( $url );
1205 if ( $text === false ) {
1206 # Not an image, make a link
1207 $text = $sk->makeExternalLink( $url, $wgContLang->markNoConversion($url), true, 'free',
1208 $this->getExternalLinkAttribs( $url ) );
1209 # Register it in the output object...
1210 # Replace unnecessary URL escape codes with their equivalent characters
1211 $pasteurized = self::replaceUnusualEscapes( $url );
1212 $this->mOutput->addExternalLink( $pasteurized );
1213 }
1214 wfProfileOut( __METHOD__ );
1215 return $text . $trail;
1216 }
1217
1218
1219 /**
1220 * Parse headers and return html
1221 *
1222 * @private
1223 */
1224 function doHeadings( $text ) {
1225 wfProfileIn( __METHOD__ );
1226 for ( $i = 6; $i >= 1; --$i ) {
1227 $h = str_repeat( '=', $i );
1228 $text = preg_replace( "/^$h(.+)$h\\s*$/m",
1229 "<h$i>\\1</h$i>", $text );
1230 }
1231 wfProfileOut( __METHOD__ );
1232 return $text;
1233 }
1234
1235 /**
1236 * Replace single quotes with HTML markup
1237 * @private
1238 * @return string the altered text
1239 */
1240 function doAllQuotes( $text ) {
1241 wfProfileIn( __METHOD__ );
1242 $outtext = '';
1243 $lines = StringUtils::explode( "\n", $text );
1244 foreach ( $lines as $line ) {
1245 $outtext .= $this->doQuotes( $line ) . "\n";
1246 }
1247 $outtext = substr( $outtext, 0,-1 );
1248 wfProfileOut( __METHOD__ );
1249 return $outtext;
1250 }
1251
1252 /**
1253 * Helper function for doAllQuotes()
1254 */
1255 public function doQuotes( $text ) {
1256 $arr = preg_split( "/(''+)/", $text, -1, PREG_SPLIT_DELIM_CAPTURE );
1257 if ( count( $arr ) == 1 ) {
1258 return $text;
1259 } else {
1260 # First, do some preliminary work. This may shift some apostrophes from
1261 # being mark-up to being text. It also counts the number of occurrences
1262 # of bold and italics mark-ups.
1263 $numbold = 0;
1264 $numitalics = 0;
1265 for ( $i = 0; $i < count( $arr ); $i++ ) {
1266 if ( ( $i % 2 ) == 1 ) {
1267 # If there are ever four apostrophes, assume the first is supposed to
1268 # be text, and the remaining three constitute mark-up for bold text.
1269 if ( strlen( $arr[$i] ) == 4 ) {
1270 $arr[$i-1] .= "'";
1271 $arr[$i] = "'''";
1272 } elseif ( strlen( $arr[$i] ) > 5 ) {
1273 # If there are more than 5 apostrophes in a row, assume they're all
1274 # text except for the last 5.
1275 $arr[$i-1] .= str_repeat( "'", strlen( $arr[$i] ) - 5 );
1276 $arr[$i] = "'''''";
1277 }
1278 # Count the number of occurrences of bold and italics mark-ups.
1279 # We are not counting sequences of five apostrophes.
1280 if ( strlen( $arr[$i] ) == 2 ) {
1281 $numitalics++;
1282 } elseif ( strlen( $arr[$i] ) == 3 ) {
1283 $numbold++;
1284 } elseif ( strlen( $arr[$i] ) == 5 ) {
1285 $numitalics++;
1286 $numbold++;
1287 }
1288 }
1289 }
1290
1291 # If there is an odd number of both bold and italics, it is likely
1292 # that one of the bold ones was meant to be an apostrophe followed
1293 # by italics. Which one we cannot know for certain, but it is more
1294 # likely to be one that has a single-letter word before it.
1295 if ( ( $numbold % 2 == 1 ) && ( $numitalics % 2 == 1 ) ) {
1296 $i = 0;
1297 $firstsingleletterword = -1;
1298 $firstmultiletterword = -1;
1299 $firstspace = -1;
1300 foreach ( $arr as $r ) {
1301 if ( ( $i % 2 == 1 ) and ( strlen( $r ) == 3 ) ) {
1302 $x1 = substr( $arr[$i-1], -1 );
1303 $x2 = substr( $arr[$i-1], -2, 1 );
1304 if ( $x1 === ' ' ) {
1305 if ( $firstspace == -1 ) {
1306 $firstspace = $i;
1307 }
1308 } elseif ( $x2 === ' ') {
1309 if ( $firstsingleletterword == -1 ) {
1310 $firstsingleletterword = $i;
1311 }
1312 } else {
1313 if ( $firstmultiletterword == -1 ) {
1314 $firstmultiletterword = $i;
1315 }
1316 }
1317 }
1318 $i++;
1319 }
1320
1321 # If there is a single-letter word, use it!
1322 if ( $firstsingleletterword > -1 ) {
1323 $arr[$firstsingleletterword] = "''";
1324 $arr[$firstsingleletterword-1] .= "'";
1325 } elseif ( $firstmultiletterword > -1 ) {
1326 # If not, but there's a multi-letter word, use that one.
1327 $arr[$firstmultiletterword] = "''";
1328 $arr[$firstmultiletterword-1] .= "'";
1329 } elseif ( $firstspace > -1 ) {
1330 # ... otherwise use the first one that has neither.
1331 # (notice that it is possible for all three to be -1 if, for example,
1332 # there is only one pentuple-apostrophe in the line)
1333 $arr[$firstspace] = "''";
1334 $arr[$firstspace-1] .= "'";
1335 }
1336 }
1337
1338 # Now let's actually convert our apostrophic mush to HTML!
1339 $output = '';
1340 $buffer = '';
1341 $state = '';
1342 $i = 0;
1343 foreach ( $arr as $r ) {
1344 if ( ( $i % 2 ) == 0 ) {
1345 if ( $state === 'both' ) {
1346 $buffer .= $r;
1347 } else {
1348 $output .= $r;
1349 }
1350 } else {
1351 if ( strlen( $r ) == 2 ) {
1352 if ( $state === 'i' ) {
1353 $output .= '</i>'; $state = '';
1354 } elseif ( $state === 'bi' ) {
1355 $output .= '</i>'; $state = 'b';
1356 } elseif ( $state === 'ib' ) {
1357 $output .= '</b></i><b>'; $state = 'b';
1358 } elseif ( $state === 'both' ) {
1359 $output .= '<b><i>'.$buffer.'</i>'; $state = 'b';
1360 } else { # $state can be 'b' or ''
1361 $output .= '<i>'; $state .= 'i';
1362 }
1363 } elseif ( strlen( $r ) == 3 ) {
1364 if ( $state === 'b' ) {
1365 $output .= '</b>'; $state = '';
1366 } elseif ( $state === 'bi' ) {
1367 $output .= '</i></b><i>'; $state = 'i';
1368 } elseif ( $state === 'ib' ) {
1369 $output .= '</b>'; $state = 'i';
1370 } elseif ( $state === 'both' ) {
1371 $output .= '<i><b>'.$buffer.'</b>'; $state = 'i';
1372 } else { # $state can be 'i' or ''
1373 $output .= '<b>'; $state .= 'b';
1374 }
1375 } elseif ( strlen( $r ) == 5 ) {
1376 if ( $state === 'b' ) {
1377 $output .= '</b><i>'; $state = 'i';
1378 } elseif ( $state === 'i' ) {
1379 $output .= '</i><b>'; $state = 'b';
1380 } elseif ( $state === 'bi' ) {
1381 $output .= '</i></b>'; $state = '';
1382 } elseif ( $state === 'ib' ) {
1383 $output .= '</b></i>'; $state = '';
1384 } elseif ( $state === 'both' ) {
1385 $output .= '<i><b>'.$buffer.'</b></i>'; $state = '';
1386 } else { # ($state == '')
1387 $buffer = ''; $state = 'both';
1388 }
1389 }
1390 }
1391 $i++;
1392 }
1393 # Now close all remaining tags. Notice that the order is important.
1394 if ( $state === 'b' || $state === 'ib' ) {
1395 $output .= '</b>';
1396 }
1397 if ( $state === 'i' || $state === 'bi' || $state === 'ib' ) {
1398 $output .= '</i>';
1399 }
1400 if ( $state === 'bi' ) {
1401 $output .= '</b>';
1402 }
1403 # There might be lonely ''''', so make sure we have a buffer
1404 if ( $state === 'both' && $buffer ) {
1405 $output .= '<b><i>'.$buffer.'</i></b>';
1406 }
1407 return $output;
1408 }
1409 }
1410
1411 /**
1412 * Replace external links (REL)
1413 *
1414 * Note: this is all very hackish and the order of execution matters a lot.
1415 * Make sure to run maintenance/parserTests.php if you change this code.
1416 *
1417 * @private
1418 */
1419 function replaceExternalLinks( $text ) {
1420 global $wgContLang;
1421 wfProfileIn( __METHOD__ );
1422
1423 $sk = $this->mOptions->getSkin( $this->mTitle );
1424
1425 $bits = preg_split( $this->mExtLinkBracketedRegex, $text, -1, PREG_SPLIT_DELIM_CAPTURE );
1426 $s = array_shift( $bits );
1427
1428 $i = 0;
1429 while ( $i<count( $bits ) ) {
1430 $url = $bits[$i++];
1431 $protocol = $bits[$i++];
1432 $text = $bits[$i++];
1433 $trail = $bits[$i++];
1434
1435 # The characters '<' and '>' (which were escaped by
1436 # removeHTMLtags()) should not be included in
1437 # URLs, per RFC 2396.
1438 $m2 = array();
1439 if ( preg_match( '/&(lt|gt);/', $url, $m2, PREG_OFFSET_CAPTURE ) ) {
1440 $text = substr( $url, $m2[0][1] ) . ' ' . $text;
1441 $url = substr( $url, 0, $m2[0][1] );
1442 }
1443
1444 # If the link text is an image URL, replace it with an <img> tag
1445 # This happened by accident in the original parser, but some people used it extensively
1446 $img = $this->maybeMakeExternalImage( $text );
1447 if ( $img !== false ) {
1448 $text = $img;
1449 }
1450
1451 $dtrail = '';
1452
1453 # Set linktype for CSS - if URL==text, link is essentially free
1454 $linktype = ( $text === $url ) ? 'free' : 'text';
1455
1456 # No link text, e.g. [http://domain.tld/some.link]
1457 if ( $text == '' ) {
1458 # Autonumber if allowed. See bug #5918
1459 if ( strpos( wfUrlProtocols(), substr( $protocol, 0, strpos( $protocol, ':' ) ) ) !== false ) {
1460 $langObj = $this->getFunctionLang();
1461 $text = '[' . $langObj->formatNum( ++$this->mAutonumber ) . ']';
1462 $linktype = 'autonumber';
1463 } else {
1464 # Otherwise just use the URL
1465 $text = htmlspecialchars( $url );
1466 $linktype = 'free';
1467 }
1468 } else {
1469 # Have link text, e.g. [http://domain.tld/some.link text]s
1470 # Check for trail
1471 list( $dtrail, $trail ) = Linker::splitTrail( $trail );
1472 }
1473
1474 $text = $wgContLang->markNoConversion( $text );
1475
1476 $url = Sanitizer::cleanUrl( $url );
1477
1478 # Use the encoded URL
1479 # This means that users can paste URLs directly into the text
1480 # Funny characters like ö aren't valid in URLs anyway
1481 # This was changed in August 2004
1482 $s .= $sk->makeExternalLink( $url, $text, false, $linktype,
1483 $this->getExternalLinkAttribs( $url ) ) . $dtrail . $trail;
1484
1485 # Register link in the output object.
1486 # Replace unnecessary URL escape codes with the referenced character
1487 # This prevents spammers from hiding links from the filters
1488 $pasteurized = self::replaceUnusualEscapes( $url );
1489 $this->mOutput->addExternalLink( $pasteurized );
1490 }
1491
1492 wfProfileOut( __METHOD__ );
1493 return $s;
1494 }
1495
1496 /**
1497 * Get an associative array of additional HTML attributes appropriate for a
1498 * particular external link. This currently may include rel => nofollow
1499 * (depending on configuration, namespace, and the URL's domain) and/or a
1500 * target attribute (depending on configuration).
1501 *
1502 * @param $url String: optional URL, to extract the domain from for rel =>
1503 * nofollow if appropriate
1504 * @return Array: associative array of HTML attributes
1505 */
1506 function getExternalLinkAttribs( $url = false ) {
1507 $attribs = array();
1508 global $wgNoFollowLinks, $wgNoFollowNsExceptions;
1509 $ns = $this->mTitle->getNamespace();
1510 if ( $wgNoFollowLinks && !in_array( $ns, $wgNoFollowNsExceptions ) ) {
1511 $attribs['rel'] = 'nofollow';
1512
1513 global $wgNoFollowDomainExceptions;
1514 if ( $wgNoFollowDomainExceptions ) {
1515 $bits = wfParseUrl( $url );
1516 if ( is_array( $bits ) && isset( $bits['host'] ) ) {
1517 foreach ( $wgNoFollowDomainExceptions as $domain ) {
1518 if ( substr( $bits['host'], -strlen( $domain ) ) == $domain ) {
1519 unset( $attribs['rel'] );
1520 break;
1521 }
1522 }
1523 }
1524 }
1525 }
1526 if ( $this->mOptions->getExternalLinkTarget() ) {
1527 $attribs['target'] = $this->mOptions->getExternalLinkTarget();
1528 }
1529 return $attribs;
1530 }
1531
1532
1533 /**
1534 * Replace unusual URL escape codes with their equivalent characters
1535 *
1536 * @param $url String
1537 * @return String
1538 *
1539 * @todo This can merge genuinely required bits in the path or query string,
1540 * breaking legit URLs. A proper fix would treat the various parts of
1541 * the URL differently; as a workaround, just use the output for
1542 * statistical records, not for actual linking/output.
1543 */
1544 static function replaceUnusualEscapes( $url ) {
1545 return preg_replace_callback( '/%[0-9A-Fa-f]{2}/',
1546 array( __CLASS__, 'replaceUnusualEscapesCallback' ), $url );
1547 }
1548
1549 /**
1550 * Callback function used in replaceUnusualEscapes().
1551 * Replaces unusual URL escape codes with their equivalent character
1552 */
1553 private static function replaceUnusualEscapesCallback( $matches ) {
1554 $char = urldecode( $matches[0] );
1555 $ord = ord( $char );
1556 # Is it an unsafe or HTTP reserved character according to RFC 1738?
1557 if ( $ord > 32 && $ord < 127 && strpos( '<>"#{}|\^~[]`;/?', $char ) === false ) {
1558 # No, shouldn't be escaped
1559 return $char;
1560 } else {
1561 # Yes, leave it escaped
1562 return $matches[0];
1563 }
1564 }
1565
1566 /**
1567 * make an image if it's allowed, either through the global
1568 * option, through the exception, or through the on-wiki whitelist
1569 * @private
1570 */
1571 function maybeMakeExternalImage( $url ) {
1572 $sk = $this->mOptions->getSkin( $this->mTitle );
1573 $imagesfrom = $this->mOptions->getAllowExternalImagesFrom();
1574 $imagesexception = !empty( $imagesfrom );
1575 $text = false;
1576 # $imagesfrom could be either a single string or an array of strings, parse out the latter
1577 if ( $imagesexception && is_array( $imagesfrom ) ) {
1578 $imagematch = false;
1579 foreach ( $imagesfrom as $match ) {
1580 if ( strpos( $url, $match ) === 0 ) {
1581 $imagematch = true;
1582 break;
1583 }
1584 }
1585 } elseif ( $imagesexception ) {
1586 $imagematch = ( strpos( $url, $imagesfrom ) === 0 );
1587 } else {
1588 $imagematch = false;
1589 }
1590 if ( $this->mOptions->getAllowExternalImages()
1591 || ( $imagesexception && $imagematch ) ) {
1592 if ( preg_match( self::EXT_IMAGE_REGEX, $url ) ) {
1593 # Image found
1594 $text = $sk->makeExternalImage( $url );
1595 }
1596 }
1597 if ( !$text && $this->mOptions->getEnableImageWhitelist()
1598 && preg_match( self::EXT_IMAGE_REGEX, $url ) ) {
1599 $whitelist = explode( "\n", wfMsgForContent( 'external_image_whitelist' ) );
1600 foreach ( $whitelist as $entry ) {
1601 # Sanitize the regex fragment, make it case-insensitive, ignore blank entries/comments
1602 if ( strpos( $entry, '#' ) === 0 || $entry === '' ) {
1603 continue;
1604 }
1605 if ( preg_match( '/' . str_replace( '/', '\\/', $entry ) . '/i', $url ) ) {
1606 # Image matches a whitelist entry
1607 $text = $sk->makeExternalImage( $url );
1608 break;
1609 }
1610 }
1611 }
1612 return $text;
1613 }
1614
1615 /**
1616 * Process [[ ]] wikilinks
1617 * @return String: processed text
1618 *
1619 * @private
1620 */
1621 function replaceInternalLinks( $s ) {
1622 $this->mLinkHolders->merge( $this->replaceInternalLinks2( $s ) );
1623 return $s;
1624 }
1625
1626 /**
1627 * Process [[ ]] wikilinks (RIL)
1628 * @return LinkHolderArray
1629 *
1630 * @private
1631 */
1632 function replaceInternalLinks2( &$s ) {
1633 global $wgContLang;
1634
1635 wfProfileIn( __METHOD__ );
1636
1637 wfProfileIn( __METHOD__.'-setup' );
1638 static $tc = FALSE, $e1, $e1_img;
1639 # the % is needed to support urlencoded titles as well
1640 if ( !$tc ) {
1641 $tc = Title::legalChars() . '#%';
1642 # Match a link having the form [[namespace:link|alternate]]trail
1643 $e1 = "/^([{$tc}]+)(?:\\|(.+?))?]](.*)\$/sD";
1644 # Match cases where there is no "]]", which might still be images
1645 $e1_img = "/^([{$tc}]+)\\|(.*)\$/sD";
1646 }
1647
1648 $sk = $this->mOptions->getSkin( $this->mTitle );
1649 $holders = new LinkHolderArray( $this );
1650
1651 # split the entire text string on occurences of [[
1652 $a = StringUtils::explode( '[[', ' ' . $s );
1653 # get the first element (all text up to first [[), and remove the space we added
1654 $s = $a->current();
1655 $a->next();
1656 $line = $a->current(); # Workaround for broken ArrayIterator::next() that returns "void"
1657 $s = substr( $s, 1 );
1658
1659 $useLinkPrefixExtension = $wgContLang->linkPrefixExtension();
1660 $e2 = null;
1661 if ( $useLinkPrefixExtension ) {
1662 # Match the end of a line for a word that's not followed by whitespace,
1663 # e.g. in the case of 'The Arab al[[Razi]]', 'al' will be matched
1664 $e2 = wfMsgForContent( 'linkprefix' );
1665 }
1666
1667 if ( is_null( $this->mTitle ) ) {
1668 wfProfileOut( __METHOD__.'-setup' );
1669 wfProfileOut( __METHOD__ );
1670 throw new MWException( __METHOD__.": \$this->mTitle is null\n" );
1671 }
1672 $nottalk = !$this->mTitle->isTalkPage();
1673
1674 if ( $useLinkPrefixExtension ) {
1675 $m = array();
1676 if ( preg_match( $e2, $s, $m ) ) {
1677 $first_prefix = $m[2];
1678 } else {
1679 $first_prefix = false;
1680 }
1681 } else {
1682 $prefix = '';
1683 }
1684
1685 if ( $wgContLang->hasVariants() ) {
1686 $selflink = $wgContLang->autoConvertToAllVariants( $this->mTitle->getPrefixedText() );
1687 } else {
1688 $selflink = array( $this->mTitle->getPrefixedText() );
1689 }
1690 $useSubpages = $this->areSubpagesAllowed();
1691 wfProfileOut( __METHOD__.'-setup' );
1692
1693 # Loop for each link
1694 for ( ; $line !== false && $line !== null ; $a->next(), $line = $a->current() ) {
1695 # Check for excessive memory usage
1696 if ( $holders->isBig() ) {
1697 # Too big
1698 # Do the existence check, replace the link holders and clear the array
1699 $holders->replace( $s );
1700 $holders->clear();
1701 }
1702
1703 if ( $useLinkPrefixExtension ) {
1704 wfProfileIn( __METHOD__.'-prefixhandling' );
1705 if ( preg_match( $e2, $s, $m ) ) {
1706 $prefix = $m[2];
1707 $s = $m[1];
1708 } else {
1709 $prefix='';
1710 }
1711 # first link
1712 if ( $first_prefix ) {
1713 $prefix = $first_prefix;
1714 $first_prefix = false;
1715 }
1716 wfProfileOut( __METHOD__.'-prefixhandling' );
1717 }
1718
1719 $might_be_img = false;
1720
1721 wfProfileIn( __METHOD__."-e1" );
1722 if ( preg_match( $e1, $line, $m ) ) { # page with normal text or alt
1723 $text = $m[2];
1724 # If we get a ] at the beginning of $m[3] that means we have a link that's something like:
1725 # [[Image:Foo.jpg|[http://example.com desc]]] <- having three ] in a row fucks up,
1726 # the real problem is with the $e1 regex
1727 # See bug 1300.
1728 #
1729 # Still some problems for cases where the ] is meant to be outside punctuation,
1730 # and no image is in sight. See bug 2095.
1731 #
1732 if ( $text !== '' &&
1733 substr( $m[3], 0, 1 ) === ']' &&
1734 strpos( $text, '[' ) !== false
1735 )
1736 {
1737 $text .= ']'; # so that replaceExternalLinks($text) works later
1738 $m[3] = substr( $m[3], 1 );
1739 }
1740 # fix up urlencoded title texts
1741 if ( strpos( $m[1], '%' ) !== false ) {
1742 # Should anchors '#' also be rejected?
1743 $m[1] = str_replace( array('<', '>'), array('&lt;', '&gt;'), urldecode( $m[1] ) );
1744 }
1745 $trail = $m[3];
1746 } elseif ( preg_match( $e1_img, $line, $m ) ) { # Invalid, but might be an image with a link in its caption
1747 $might_be_img = true;
1748 $text = $m[2];
1749 if ( strpos( $m[1], '%' ) !== false ) {
1750 $m[1] = urldecode( $m[1] );
1751 }
1752 $trail = "";
1753 } else { # Invalid form; output directly
1754 $s .= $prefix . '[[' . $line ;
1755 wfProfileOut( __METHOD__."-e1" );
1756 continue;
1757 }
1758 wfProfileOut( __METHOD__."-e1" );
1759 wfProfileIn( __METHOD__."-misc" );
1760
1761 # Don't allow internal links to pages containing
1762 # PROTO: where PROTO is a valid URL protocol; these
1763 # should be external links.
1764 if ( preg_match( '/^\b(?:' . wfUrlProtocols() . ')/', $m[1] ) ) {
1765 $s .= $prefix . '[[' . $line ;
1766 wfProfileOut( __METHOD__."-misc" );
1767 continue;
1768 }
1769
1770 # Make subpage if necessary
1771 if ( $useSubpages ) {
1772 $link = $this->maybeDoSubpageLink( $m[1], $text );
1773 } else {
1774 $link = $m[1];
1775 }
1776
1777 $noforce = ( substr( $m[1], 0, 1 ) !== ':' );
1778 if ( !$noforce ) {
1779 # Strip off leading ':'
1780 $link = substr( $link, 1 );
1781 }
1782
1783 wfProfileOut( __METHOD__."-misc" );
1784 wfProfileIn( __METHOD__."-title" );
1785 $nt = Title::newFromText( $this->mStripState->unstripNoWiki( $link ) );
1786 if ( $nt === null ) {
1787 $s .= $prefix . '[[' . $line;
1788 wfProfileOut( __METHOD__."-title" );
1789 continue;
1790 }
1791
1792 $ns = $nt->getNamespace();
1793 $iw = $nt->getInterWiki();
1794 wfProfileOut( __METHOD__."-title" );
1795
1796 if ( $might_be_img ) { # if this is actually an invalid link
1797 wfProfileIn( __METHOD__."-might_be_img" );
1798 if ( $ns == NS_FILE && $noforce ) { # but might be an image
1799 $found = false;
1800 while ( true ) {
1801 # look at the next 'line' to see if we can close it there
1802 $a->next();
1803 $next_line = $a->current();
1804 if ( $next_line === false || $next_line === null ) {
1805 break;
1806 }
1807 $m = explode( ']]', $next_line, 3 );
1808 if ( count( $m ) == 3 ) {
1809 # the first ]] closes the inner link, the second the image
1810 $found = true;
1811 $text .= "[[{$m[0]}]]{$m[1]}";
1812 $trail = $m[2];
1813 break;
1814 } elseif ( count( $m ) == 2 ) {
1815 # if there's exactly one ]] that's fine, we'll keep looking
1816 $text .= "[[{$m[0]}]]{$m[1]}";
1817 } else {
1818 # if $next_line is invalid too, we need look no further
1819 $text .= '[[' . $next_line;
1820 break;
1821 }
1822 }
1823 if ( !$found ) {
1824 # we couldn't find the end of this imageLink, so output it raw
1825 # but don't ignore what might be perfectly normal links in the text we've examined
1826 $holders->merge( $this->replaceInternalLinks2( $text ) );
1827 $s .= "{$prefix}[[$link|$text";
1828 # note: no $trail, because without an end, there *is* no trail
1829 wfProfileOut( __METHOD__."-might_be_img" );
1830 continue;
1831 }
1832 } else { # it's not an image, so output it raw
1833 $s .= "{$prefix}[[$link|$text";
1834 # note: no $trail, because without an end, there *is* no trail
1835 wfProfileOut( __METHOD__."-might_be_img" );
1836 continue;
1837 }
1838 wfProfileOut( __METHOD__."-might_be_img" );
1839 }
1840
1841 $wasblank = ( $text == '' );
1842 if ( $wasblank ) {
1843 $text = $link;
1844 } else {
1845 # Bug 4598 madness. Handle the quotes only if they come from the alternate part
1846 # [[Lista d''e paise d''o munno]] -> <a href="">Lista d''e paise d''o munno</a>
1847 # [[Criticism of Harry Potter|Criticism of ''Harry Potter'']] -> <a href="Criticism of Harry Potter">Criticism of <i>Harry Potter</i></a>
1848 $text = $this->doQuotes($text);
1849 }
1850
1851 # Link not escaped by : , create the various objects
1852 if ( $noforce ) {
1853
1854 # Interwikis
1855 wfProfileIn( __METHOD__."-interwiki" );
1856 if ( $iw && $this->mOptions->getInterwikiMagic() && $nottalk && $wgContLang->getLanguageName( $iw ) ) {
1857 $this->mOutput->addLanguageLink( $nt->getFullText() );
1858 $s = rtrim( $s . $prefix );
1859 $s .= trim( $trail, "\n" ) == '' ? '': $prefix . $trail;
1860 wfProfileOut( __METHOD__."-interwiki" );
1861 continue;
1862 }
1863 wfProfileOut( __METHOD__."-interwiki" );
1864
1865 if ( $ns == NS_FILE ) {
1866 wfProfileIn( __METHOD__."-image" );
1867 if ( !wfIsBadImage( $nt->getDBkey(), $this->mTitle ) ) {
1868 if ( $wasblank ) {
1869 # if no parameters were passed, $text
1870 # becomes something like "File:Foo.png",
1871 # which we don't want to pass on to the
1872 # image generator
1873 $text = '';
1874 } else {
1875 # recursively parse links inside the image caption
1876 # actually, this will parse them in any other parameters, too,
1877 # but it might be hard to fix that, and it doesn't matter ATM
1878 $text = $this->replaceExternalLinks( $text );
1879 $holders->merge( $this->replaceInternalLinks2( $text ) );
1880 }
1881 # cloak any absolute URLs inside the image markup, so replaceExternalLinks() won't touch them
1882 $s .= $prefix . $this->armorLinks( $this->makeImage( $nt, $text, $holders ) ) . $trail;
1883 } else {
1884 $s .= $prefix . $trail;
1885 }
1886 $this->mOutput->addImage( $nt->getDBkey() );
1887 wfProfileOut( __METHOD__."-image" );
1888 continue;
1889
1890 }
1891
1892 if ( $ns == NS_CATEGORY ) {
1893 wfProfileIn( __METHOD__."-category" );
1894 $s = rtrim( $s . "\n" ); # bug 87
1895
1896 if ( $wasblank ) {
1897 $sortkey = $this->getDefaultSort();
1898 } else {
1899 $sortkey = $text;
1900 }
1901 $sortkey = Sanitizer::decodeCharReferences( $sortkey );
1902 $sortkey = str_replace( "\n", '', $sortkey );
1903 $sortkey = $wgContLang->convertCategoryKey( $sortkey );
1904 $this->mOutput->addCategory( $nt->getDBkey(), $sortkey );
1905
1906 /**
1907 * Strip the whitespace Category links produce, see bug 87
1908 * @todo We might want to use trim($tmp, "\n") here.
1909 */
1910 $s .= trim( $prefix . $trail, "\n" ) == '' ? '': $prefix . $trail;
1911
1912 wfProfileOut( __METHOD__."-category" );
1913 continue;
1914 }
1915 }
1916
1917 # Self-link checking
1918 if ( $nt->getFragment() === '' && $ns != NS_SPECIAL ) {
1919 if ( in_array( $nt->getPrefixedText(), $selflink, true ) ) {
1920 $s .= $prefix . $sk->makeSelfLinkObj( $nt, $text, '', $trail );
1921 continue;
1922 }
1923 }
1924
1925 # NS_MEDIA is a pseudo-namespace for linking directly to a file
1926 # FIXME: Should do batch file existence checks, see comment below
1927 if ( $ns == NS_MEDIA ) {
1928 wfProfileIn( __METHOD__."-media" );
1929 # Give extensions a chance to select the file revision for us
1930 $skip = $time = false;
1931 wfRunHooks( 'BeforeParserMakeImageLinkObj', array( &$this, &$nt, &$skip, &$time ) );
1932 if ( $skip ) {
1933 $link = $sk->link( $nt );
1934 } else {
1935 $link = $sk->makeMediaLinkObj( $nt, $text, $time );
1936 }
1937 # Cloak with NOPARSE to avoid replacement in replaceExternalLinks
1938 $s .= $prefix . $this->armorLinks( $link ) . $trail;
1939 $this->mOutput->addImage( $nt->getDBkey() );
1940 wfProfileOut( __METHOD__."-media" );
1941 continue;
1942 }
1943
1944 wfProfileIn( __METHOD__."-always_known" );
1945 # Some titles, such as valid special pages or files in foreign repos, should
1946 # be shown as bluelinks even though they're not included in the page table
1947 #
1948 # FIXME: isAlwaysKnown() can be expensive for file links; we should really do
1949 # batch file existence checks for NS_FILE and NS_MEDIA
1950 if ( $iw == '' && $nt->isAlwaysKnown() ) {
1951 $this->mOutput->addLink( $nt );
1952 $s .= $this->makeKnownLinkHolder( $nt, $text, '', $trail, $prefix );
1953 } else {
1954 # Links will be added to the output link list after checking
1955 $s .= $holders->makeHolder( $nt, $text, '', $trail, $prefix );
1956 }
1957 wfProfileOut( __METHOD__."-always_known" );
1958 }
1959 wfProfileOut( __METHOD__ );
1960 return $holders;
1961 }
1962
1963 /**
1964 * Make a link placeholder. The text returned can be later resolved to a real link with
1965 * replaceLinkHolders(). This is done for two reasons: firstly to avoid further
1966 * parsing of interwiki links, and secondly to allow all existence checks and
1967 * article length checks (for stub links) to be bundled into a single query.
1968 *
1969 * @deprecated
1970 */
1971 function makeLinkHolder( &$nt, $text = '', $query = '', $trail = '', $prefix = '' ) {
1972 return $this->mLinkHolders->makeHolder( $nt, $text, $query, $trail, $prefix );
1973 }
1974
1975 /**
1976 * Render a forced-blue link inline; protect against double expansion of
1977 * URLs if we're in a mode that prepends full URL prefixes to internal links.
1978 * Since this little disaster has to split off the trail text to avoid
1979 * breaking URLs in the following text without breaking trails on the
1980 * wiki links, it's been made into a horrible function.
1981 *
1982 * @param $nt Title
1983 * @param $text String
1984 * @param $query String
1985 * @param $trail String
1986 * @param $prefix String
1987 * @return String: HTML-wikitext mix oh yuck
1988 */
1989 function makeKnownLinkHolder( $nt, $text = '', $query = '', $trail = '', $prefix = '' ) {
1990 list( $inside, $trail ) = Linker::splitTrail( $trail );
1991 $sk = $this->mOptions->getSkin( $this->mTitle );
1992 # FIXME: use link() instead of deprecated makeKnownLinkObj()
1993 $link = $sk->makeKnownLinkObj( $nt, $text, $query, $inside, $prefix );
1994 return $this->armorLinks( $link ) . $trail;
1995 }
1996
1997 /**
1998 * Insert a NOPARSE hacky thing into any inline links in a chunk that's
1999 * going to go through further parsing steps before inline URL expansion.
2000 *
2001 * Not needed quite as much as it used to be since free links are a bit
2002 * more sensible these days. But bracketed links are still an issue.
2003 *
2004 * @param $text String: more-or-less HTML
2005 * @return String: less-or-more HTML with NOPARSE bits
2006 */
2007 function armorLinks( $text ) {
2008 return preg_replace( '/\b(' . wfUrlProtocols() . ')/',
2009 "{$this->mUniqPrefix}NOPARSE$1", $text );
2010 }
2011
2012 /**
2013 * Return true if subpage links should be expanded on this page.
2014 * @return Boolean
2015 */
2016 function areSubpagesAllowed() {
2017 # Some namespaces don't allow subpages
2018 return MWNamespace::hasSubpages( $this->mTitle->getNamespace() );
2019 }
2020
2021 /**
2022 * Handle link to subpage if necessary
2023 *
2024 * @param $target String: the source of the link
2025 * @param &$text String: the link text, modified as necessary
2026 * @return string the full name of the link
2027 * @private
2028 */
2029 function maybeDoSubpageLink( $target, &$text ) {
2030 return Linker::normalizeSubpageLink( $this->mTitle, $target, $text );
2031 }
2032
2033 /**#@+
2034 * Used by doBlockLevels()
2035 * @private
2036 */
2037 function closeParagraph() {
2038 $result = '';
2039 if ( $this->mLastSection != '' ) {
2040 $result = '</' . $this->mLastSection . ">\n";
2041 }
2042 $this->mInPre = false;
2043 $this->mLastSection = '';
2044 return $result;
2045 }
2046
2047 /**
2048 * getCommon() returns the length of the longest common substring
2049 * of both arguments, starting at the beginning of both.
2050 * @private
2051 */
2052 function getCommon( $st1, $st2 ) {
2053 $fl = strlen( $st1 );
2054 $shorter = strlen( $st2 );
2055 if ( $fl < $shorter ) {
2056 $shorter = $fl;
2057 }
2058
2059 for ( $i = 0; $i < $shorter; ++$i ) {
2060 if ( $st1{$i} != $st2{$i} ) {
2061 break;
2062 }
2063 }
2064 return $i;
2065 }
2066
2067 /**
2068 * These next three functions open, continue, and close the list
2069 * element appropriate to the prefix character passed into them.
2070 * @private
2071 */
2072 function openList( $char ) {
2073 $result = $this->closeParagraph();
2074
2075 if ( '*' === $char ) {
2076 $result .= '<ul><li>';
2077 } elseif ( '#' === $char ) {
2078 $result .= '<ol><li>';
2079 } elseif ( ':' === $char ) {
2080 $result .= '<dl><dd>';
2081 } elseif ( ';' === $char ) {
2082 $result .= '<dl><dt>';
2083 $this->mDTopen = true;
2084 } else {
2085 $result = '<!-- ERR 1 -->';
2086 }
2087
2088 return $result;
2089 }
2090
2091 /**
2092 * TODO: document
2093 * @param $char String
2094 * @private
2095 */
2096 function nextItem( $char ) {
2097 if ( '*' === $char || '#' === $char ) {
2098 return '</li><li>';
2099 } elseif ( ':' === $char || ';' === $char ) {
2100 $close = '</dd>';
2101 if ( $this->mDTopen ) {
2102 $close = '</dt>';
2103 }
2104 if ( ';' === $char ) {
2105 $this->mDTopen = true;
2106 return $close . '<dt>';
2107 } else {
2108 $this->mDTopen = false;
2109 return $close . '<dd>';
2110 }
2111 }
2112 return '<!-- ERR 2 -->';
2113 }
2114
2115 /**
2116 * TODO: document
2117 * @param $char String
2118 * @private
2119 */
2120 function closeList( $char ) {
2121 if ( '*' === $char ) {
2122 $text = '</li></ul>';
2123 } elseif ( '#' === $char ) {
2124 $text = '</li></ol>';
2125 } elseif ( ':' === $char ) {
2126 if ( $this->mDTopen ) {
2127 $this->mDTopen = false;
2128 $text = '</dt></dl>';
2129 } else {
2130 $text = '</dd></dl>';
2131 }
2132 } else {
2133 return '<!-- ERR 3 -->';
2134 }
2135 return $text."\n";
2136 }
2137 /**#@-*/
2138
2139 /**
2140 * Make lists from lines starting with ':', '*', '#', etc. (DBL)
2141 *
2142 * @param $text String
2143 * @param $linestart Boolean: whether or not this is at the start of a line.
2144 * @private
2145 * @return string the lists rendered as HTML
2146 */
2147 function doBlockLevels( $text, $linestart ) {
2148 wfProfileIn( __METHOD__ );
2149
2150 # Parsing through the text line by line. The main thing
2151 # happening here is handling of block-level elements p, pre,
2152 # and making lists from lines starting with * # : etc.
2153 #
2154 $textLines = StringUtils::explode( "\n", $text );
2155
2156 $lastPrefix = $output = '';
2157 $this->mDTopen = $inBlockElem = false;
2158 $prefixLength = 0;
2159 $paragraphStack = false;
2160
2161 foreach ( $textLines as $oLine ) {
2162 # Fix up $linestart
2163 if ( !$linestart ) {
2164 $output .= $oLine;
2165 $linestart = true;
2166 continue;
2167 }
2168 # * = ul
2169 # # = ol
2170 # ; = dt
2171 # : = dd
2172
2173 $lastPrefixLength = strlen( $lastPrefix );
2174 $preCloseMatch = preg_match( '/<\\/pre/i', $oLine );
2175 $preOpenMatch = preg_match( '/<pre/i', $oLine );
2176 # If not in a <pre> element, scan for and figure out what prefixes are there.
2177 if ( !$this->mInPre ) {
2178 # Multiple prefixes may abut each other for nested lists.
2179 $prefixLength = strspn( $oLine, '*#:;' );
2180 $prefix = substr( $oLine, 0, $prefixLength );
2181
2182 # eh?
2183 # ; and : are both from definition-lists, so they're equivalent
2184 # for the purposes of determining whether or not we need to open/close
2185 # elements.
2186 $prefix2 = str_replace( ';', ':', $prefix );
2187 $t = substr( $oLine, $prefixLength );
2188 $this->mInPre = (bool)$preOpenMatch;
2189 } else {
2190 # Don't interpret any other prefixes in preformatted text
2191 $prefixLength = 0;
2192 $prefix = $prefix2 = '';
2193 $t = $oLine;
2194 }
2195
2196 # List generation
2197 if ( $prefixLength && $lastPrefix === $prefix2 ) {
2198 # Same as the last item, so no need to deal with nesting or opening stuff
2199 $output .= $this->nextItem( substr( $prefix, -1 ) );
2200 $paragraphStack = false;
2201
2202 if ( substr( $prefix, -1 ) === ';') {
2203 # The one nasty exception: definition lists work like this:
2204 # ; title : definition text
2205 # So we check for : in the remainder text to split up the
2206 # title and definition, without b0rking links.
2207 $term = $t2 = '';
2208 if ( $this->findColonNoLinks( $t, $term, $t2 ) !== false ) {
2209 $t = $t2;
2210 $output .= $term . $this->nextItem( ':' );
2211 }
2212 }
2213 } elseif ( $prefixLength || $lastPrefixLength ) {
2214 # We need to open or close prefixes, or both.
2215
2216 # Either open or close a level...
2217 $commonPrefixLength = $this->getCommon( $prefix, $lastPrefix );
2218 $paragraphStack = false;
2219
2220 # Close all the prefixes which aren't shared.
2221 while ( $commonPrefixLength < $lastPrefixLength ) {
2222 $output .= $this->closeList( $lastPrefix[$lastPrefixLength-1] );
2223 --$lastPrefixLength;
2224 }
2225
2226 # Continue the current prefix if appropriate.
2227 if ( $prefixLength <= $commonPrefixLength && $commonPrefixLength > 0 ) {
2228 $output .= $this->nextItem( $prefix[$commonPrefixLength-1] );
2229 }
2230
2231 # Open prefixes where appropriate.
2232 while ( $prefixLength > $commonPrefixLength ) {
2233 $char = substr( $prefix, $commonPrefixLength, 1 );
2234 $output .= $this->openList( $char );
2235
2236 if ( ';' === $char ) {
2237 # FIXME: This is dupe of code above
2238 if ( $this->findColonNoLinks( $t, $term, $t2 ) !== false ) {
2239 $t = $t2;
2240 $output .= $term . $this->nextItem( ':' );
2241 }
2242 }
2243 ++$commonPrefixLength;
2244 }
2245 $lastPrefix = $prefix2;
2246 }
2247
2248 # If we have no prefixes, go to paragraph mode.
2249 if ( 0 == $prefixLength ) {
2250 wfProfileIn( __METHOD__."-paragraph" );
2251 # No prefix (not in list)--go to paragraph mode
2252 # XXX: use a stack for nestable elements like span, table and div
2253 $openmatch = preg_match('/(?:<table|<blockquote|<h1|<h2|<h3|<h4|<h5|<h6|<pre|<tr|<p|<ul|<ol|<li|<\\/tr|<\\/td|<\\/th)/iS', $t );
2254 $closematch = preg_match(
2255 '/(?:<\\/table|<\\/blockquote|<\\/h1|<\\/h2|<\\/h3|<\\/h4|<\\/h5|<\\/h6|'.
2256 '<td|<th|<\\/?div|<hr|<\\/pre|<\\/p|'.$this->mUniqPrefix.'-pre|<\\/li|<\\/ul|<\\/ol|<\\/?center)/iS', $t );
2257 if ( $openmatch or $closematch ) {
2258 $paragraphStack = false;
2259 # TODO bug 5718: paragraph closed
2260 $output .= $this->closeParagraph();
2261 if ( $preOpenMatch and !$preCloseMatch ) {
2262 $this->mInPre = true;
2263 }
2264 $inBlockElem = !$closematch;
2265 } elseif ( !$inBlockElem && !$this->mInPre ) {
2266 if ( ' ' == substr( $t, 0, 1 ) and ( $this->mLastSection === 'pre' || trim( $t ) != '' ) ) {
2267 # pre
2268 if ( $this->mLastSection !== 'pre' ) {
2269 $paragraphStack = false;
2270 $output .= $this->closeParagraph().'<pre>';
2271 $this->mLastSection = 'pre';
2272 }
2273 $t = substr( $t, 1 );
2274 } else {
2275 # paragraph
2276 if ( trim( $t ) === '' ) {
2277 if ( $paragraphStack ) {
2278 $output .= $paragraphStack.'<br />';
2279 $paragraphStack = false;
2280 $this->mLastSection = 'p';
2281 } else {
2282 if ( $this->mLastSection !== 'p' ) {
2283 $output .= $this->closeParagraph();
2284 $this->mLastSection = '';
2285 $paragraphStack = '<p>';
2286 } else {
2287 $paragraphStack = '</p><p>';
2288 }
2289 }
2290 } else {
2291 if ( $paragraphStack ) {
2292 $output .= $paragraphStack;
2293 $paragraphStack = false;
2294 $this->mLastSection = 'p';
2295 } elseif ( $this->mLastSection !== 'p' ) {
2296 $output .= $this->closeParagraph().'<p>';
2297 $this->mLastSection = 'p';
2298 }
2299 }
2300 }
2301 }
2302 wfProfileOut( __METHOD__."-paragraph" );
2303 }
2304 # somewhere above we forget to get out of pre block (bug 785)
2305 if ( $preCloseMatch && $this->mInPre ) {
2306 $this->mInPre = false;
2307 }
2308 if ( $paragraphStack === false ) {
2309 $output .= $t."\n";
2310 }
2311 }
2312 while ( $prefixLength ) {
2313 $output .= $this->closeList( $prefix2[$prefixLength-1] );
2314 --$prefixLength;
2315 }
2316 if ( $this->mLastSection != '' ) {
2317 $output .= '</' . $this->mLastSection . '>';
2318 $this->mLastSection = '';
2319 }
2320
2321 wfProfileOut( __METHOD__ );
2322 return $output;
2323 }
2324
2325 /**
2326 * Split up a string on ':', ignoring any occurences inside tags
2327 * to prevent illegal overlapping.
2328 *
2329 * @param $str String: the string to split
2330 * @param &$before String: set to everything before the ':'
2331 * @param &$after String: set to everything after the ':'
2332 * return String: the position of the ':', or false if none found
2333 */
2334 function findColonNoLinks( $str, &$before, &$after ) {
2335 wfProfileIn( __METHOD__ );
2336
2337 $pos = strpos( $str, ':' );
2338 if ( $pos === false ) {
2339 # Nothing to find!
2340 wfProfileOut( __METHOD__ );
2341 return false;
2342 }
2343
2344 $lt = strpos( $str, '<' );
2345 if ( $lt === false || $lt > $pos ) {
2346 # Easy; no tag nesting to worry about
2347 $before = substr( $str, 0, $pos );
2348 $after = substr( $str, $pos+1 );
2349 wfProfileOut( __METHOD__ );
2350 return $pos;
2351 }
2352
2353 # Ugly state machine to walk through avoiding tags.
2354 $state = self::COLON_STATE_TEXT;
2355 $stack = 0;
2356 $len = strlen( $str );
2357 for( $i = 0; $i < $len; $i++ ) {
2358 $c = $str{$i};
2359
2360 switch( $state ) {
2361 # (Using the number is a performance hack for common cases)
2362 case 0: # self::COLON_STATE_TEXT:
2363 switch( $c ) {
2364 case "<":
2365 # Could be either a <start> tag or an </end> tag
2366 $state = self::COLON_STATE_TAGSTART;
2367 break;
2368 case ":":
2369 if ( $stack == 0 ) {
2370 # We found it!
2371 $before = substr( $str, 0, $i );
2372 $after = substr( $str, $i + 1 );
2373 wfProfileOut( __METHOD__ );
2374 return $i;
2375 }
2376 # Embedded in a tag; don't break it.
2377 break;
2378 default:
2379 # Skip ahead looking for something interesting
2380 $colon = strpos( $str, ':', $i );
2381 if ( $colon === false ) {
2382 # Nothing else interesting
2383 wfProfileOut( __METHOD__ );
2384 return false;
2385 }
2386 $lt = strpos( $str, '<', $i );
2387 if ( $stack === 0 ) {
2388 if ( $lt === false || $colon < $lt ) {
2389 # We found it!
2390 $before = substr( $str, 0, $colon );
2391 $after = substr( $str, $colon + 1 );
2392 wfProfileOut( __METHOD__ );
2393 return $i;
2394 }
2395 }
2396 if ( $lt === false ) {
2397 # Nothing else interesting to find; abort!
2398 # We're nested, but there's no close tags left. Abort!
2399 break 2;
2400 }
2401 # Skip ahead to next tag start
2402 $i = $lt;
2403 $state = self::COLON_STATE_TAGSTART;
2404 }
2405 break;
2406 case 1: # self::COLON_STATE_TAG:
2407 # In a <tag>
2408 switch( $c ) {
2409 case ">":
2410 $stack++;
2411 $state = self::COLON_STATE_TEXT;
2412 break;
2413 case "/":
2414 # Slash may be followed by >?
2415 $state = self::COLON_STATE_TAGSLASH;
2416 break;
2417 default:
2418 # ignore
2419 }
2420 break;
2421 case 2: # self::COLON_STATE_TAGSTART:
2422 switch( $c ) {
2423 case "/":
2424 $state = self::COLON_STATE_CLOSETAG;
2425 break;
2426 case "!":
2427 $state = self::COLON_STATE_COMMENT;
2428 break;
2429 case ">":
2430 # Illegal early close? This shouldn't happen D:
2431 $state = self::COLON_STATE_TEXT;
2432 break;
2433 default:
2434 $state = self::COLON_STATE_TAG;
2435 }
2436 break;
2437 case 3: # self::COLON_STATE_CLOSETAG:
2438 # In a </tag>
2439 if ( $c === ">" ) {
2440 $stack--;
2441 if ( $stack < 0 ) {
2442 wfDebug( __METHOD__.": Invalid input; too many close tags\n" );
2443 wfProfileOut( __METHOD__ );
2444 return false;
2445 }
2446 $state = self::COLON_STATE_TEXT;
2447 }
2448 break;
2449 case self::COLON_STATE_TAGSLASH:
2450 if ( $c === ">" ) {
2451 # Yes, a self-closed tag <blah/>
2452 $state = self::COLON_STATE_TEXT;
2453 } else {
2454 # Probably we're jumping the gun, and this is an attribute
2455 $state = self::COLON_STATE_TAG;
2456 }
2457 break;
2458 case 5: # self::COLON_STATE_COMMENT:
2459 if ( $c === "-" ) {
2460 $state = self::COLON_STATE_COMMENTDASH;
2461 }
2462 break;
2463 case self::COLON_STATE_COMMENTDASH:
2464 if ( $c === "-" ) {
2465 $state = self::COLON_STATE_COMMENTDASHDASH;
2466 } else {
2467 $state = self::COLON_STATE_COMMENT;
2468 }
2469 break;
2470 case self::COLON_STATE_COMMENTDASHDASH:
2471 if ( $c === ">" ) {
2472 $state = self::COLON_STATE_TEXT;
2473 } else {
2474 $state = self::COLON_STATE_COMMENT;
2475 }
2476 break;
2477 default:
2478 throw new MWException( "State machine error in " . __METHOD__ );
2479 }
2480 }
2481 if ( $stack > 0 ) {
2482 wfDebug( __METHOD__.": Invalid input; not enough close tags (stack $stack, state $state)\n" );
2483 return false;
2484 }
2485 wfProfileOut( __METHOD__ );
2486 return false;
2487 }
2488
2489 /**
2490 * Return value of a magic variable (like PAGENAME)
2491 *
2492 * @private
2493 */
2494 function getVariableValue( $index, $frame=false ) {
2495 global $wgContLang, $wgSitename, $wgServer;
2496 global $wgArticlePath, $wgScriptPath, $wgStylePath;
2497
2498 /**
2499 * Some of these require message or data lookups and can be
2500 * expensive to check many times.
2501 */
2502 if ( wfRunHooks( 'ParserGetVariableValueVarCache', array( &$this, &$this->mVarCache ) ) ) {
2503 if ( isset( $this->mVarCache[$index] ) ) {
2504 return $this->mVarCache[$index];
2505 }
2506 }
2507
2508 $ts = wfTimestamp( TS_UNIX, $this->mOptions->getTimestamp() );
2509 wfRunHooks( 'ParserGetVariableValueTs', array( &$this, &$ts ) );
2510
2511 # Use the time zone
2512 global $wgLocaltimezone;
2513 if ( isset( $wgLocaltimezone ) ) {
2514 $oldtz = date_default_timezone_get();
2515 date_default_timezone_set( $wgLocaltimezone );
2516 }
2517
2518 $localTimestamp = date( 'YmdHis', $ts );
2519 $localMonth = date( 'm', $ts );
2520 $localMonth1 = date( 'n', $ts );
2521 $localMonthName = date( 'n', $ts );
2522 $localDay = date( 'j', $ts );
2523 $localDay2 = date( 'd', $ts );
2524 $localDayOfWeek = date( 'w', $ts );
2525 $localWeek = date( 'W', $ts );
2526 $localYear = date( 'Y', $ts );
2527 $localHour = date( 'H', $ts );
2528 if ( isset( $wgLocaltimezone ) ) {
2529 date_default_timezone_set( $oldtz );
2530 }
2531
2532 switch ( $index ) {
2533 case 'currentmonth':
2534 $value = $wgContLang->formatNum( gmdate( 'm', $ts ) );
2535 break;
2536 case 'currentmonth1':
2537 $value = $wgContLang->formatNum( gmdate( 'n', $ts ) );
2538 break;
2539 case 'currentmonthname':
2540 $value = $wgContLang->getMonthName( gmdate( 'n', $ts ) );
2541 break;
2542 case 'currentmonthnamegen':
2543 $value = $wgContLang->getMonthNameGen( gmdate( 'n', $ts ) );
2544 break;
2545 case 'currentmonthabbrev':
2546 $value = $wgContLang->getMonthAbbreviation( gmdate( 'n', $ts ) );
2547 break;
2548 case 'currentday':
2549 $value = $wgContLang->formatNum( gmdate( 'j', $ts ) );
2550 break;
2551 case 'currentday2':
2552 $value = $wgContLang->formatNum( gmdate( 'd', $ts ) );
2553 break;
2554 case 'localmonth':
2555 $value = $wgContLang->formatNum( $localMonth );
2556 break;
2557 case 'localmonth1':
2558 $value = $wgContLang->formatNum( $localMonth1 );
2559 break;
2560 case 'localmonthname':
2561 $value = $wgContLang->getMonthName( $localMonthName );
2562 break;
2563 case 'localmonthnamegen':
2564 $value = $wgContLang->getMonthNameGen( $localMonthName );
2565 break;
2566 case 'localmonthabbrev':
2567 $value = $wgContLang->getMonthAbbreviation( $localMonthName );
2568 break;
2569 case 'localday':
2570 $value = $wgContLang->formatNum( $localDay );
2571 break;
2572 case 'localday2':
2573 $value = $wgContLang->formatNum( $localDay2 );
2574 break;
2575 case 'pagename':
2576 $value = wfEscapeWikiText( $this->mTitle->getText() );
2577 break;
2578 case 'pagenamee':
2579 $value = $this->mTitle->getPartialURL();
2580 break;
2581 case 'fullpagename':
2582 $value = wfEscapeWikiText( $this->mTitle->getPrefixedText() );
2583 break;
2584 case 'fullpagenamee':
2585 $value = $this->mTitle->getPrefixedURL();
2586 break;
2587 case 'subpagename':
2588 $value = wfEscapeWikiText( $this->mTitle->getSubpageText() );
2589 break;
2590 case 'subpagenamee':
2591 $value = $this->mTitle->getSubpageUrlForm();
2592 break;
2593 case 'basepagename':
2594 $value = wfEscapeWikiText( $this->mTitle->getBaseText() );
2595 break;
2596 case 'basepagenamee':
2597 $value = wfUrlEncode( str_replace( ' ', '_', $this->mTitle->getBaseText() ) );
2598 break;
2599 case 'talkpagename':
2600 if ( $this->mTitle->canTalk() ) {
2601 $talkPage = $this->mTitle->getTalkPage();
2602 $value = wfEscapeWikiText( $talkPage->getPrefixedText() );
2603 } else {
2604 $value = '';
2605 }
2606 break;
2607 case 'talkpagenamee':
2608 if ( $this->mTitle->canTalk() ) {
2609 $talkPage = $this->mTitle->getTalkPage();
2610 $value = $talkPage->getPrefixedUrl();
2611 } else {
2612 $value = '';
2613 }
2614 break;
2615 case 'subjectpagename':
2616 $subjPage = $this->mTitle->getSubjectPage();
2617 $value = wfEscapeWikiText( $subjPage->getPrefixedText() );
2618 break;
2619 case 'subjectpagenamee':
2620 $subjPage = $this->mTitle->getSubjectPage();
2621 $value = $subjPage->getPrefixedUrl();
2622 break;
2623 case 'revisionid':
2624 # Let the edit saving system know we should parse the page
2625 # *after* a revision ID has been assigned.
2626 $this->mOutput->setFlag( 'vary-revision' );
2627 wfDebug( __METHOD__ . ": {{REVISIONID}} used, setting vary-revision...\n" );
2628 $value = $this->mRevisionId;
2629 break;
2630 case 'revisionday':
2631 # Let the edit saving system know we should parse the page
2632 # *after* a revision ID has been assigned. This is for null edits.
2633 $this->mOutput->setFlag( 'vary-revision' );
2634 wfDebug( __METHOD__ . ": {{REVISIONDAY}} used, setting vary-revision...\n" );
2635 $value = intval( substr( $this->getRevisionTimestamp(), 6, 2 ) );
2636 break;
2637 case 'revisionday2':
2638 # Let the edit saving system know we should parse the page
2639 # *after* a revision ID has been assigned. This is for null edits.
2640 $this->mOutput->setFlag( 'vary-revision' );
2641 wfDebug( __METHOD__ . ": {{REVISIONDAY2}} used, setting vary-revision...\n" );
2642 $value = substr( $this->getRevisionTimestamp(), 6, 2 );
2643 break;
2644 case 'revisionmonth':
2645 # Let the edit saving system know we should parse the page
2646 # *after* a revision ID has been assigned. This is for null edits.
2647 $this->mOutput->setFlag( 'vary-revision' );
2648 wfDebug( __METHOD__ . ": {{REVISIONMONTH}} used, setting vary-revision...\n" );
2649 $value = substr( $this->getRevisionTimestamp(), 4, 2 );
2650 break;
2651 case 'revisionmonth1':
2652 # Let the edit saving system know we should parse the page
2653 # *after* a revision ID has been assigned. This is for null edits.
2654 $this->mOutput->setFlag( 'vary-revision' );
2655 wfDebug( __METHOD__ . ": {{REVISIONMONTH1}} used, setting vary-revision...\n" );
2656 $value = intval( substr( $this->getRevisionTimestamp(), 4, 2 ) );
2657 break;
2658 case 'revisionyear':
2659 # Let the edit saving system know we should parse the page
2660 # *after* a revision ID has been assigned. This is for null edits.
2661 $this->mOutput->setFlag( 'vary-revision' );
2662 wfDebug( __METHOD__ . ": {{REVISIONYEAR}} used, setting vary-revision...\n" );
2663 $value = substr( $this->getRevisionTimestamp(), 0, 4 );
2664 break;
2665 case 'revisiontimestamp':
2666 # Let the edit saving system know we should parse the page
2667 # *after* a revision ID has been assigned. This is for null edits.
2668 $this->mOutput->setFlag( 'vary-revision' );
2669 wfDebug( __METHOD__ . ": {{REVISIONTIMESTAMP}} used, setting vary-revision...\n" );
2670 $value = $this->getRevisionTimestamp();
2671 break;
2672 case 'revisionuser':
2673 # Let the edit saving system know we should parse the page
2674 # *after* a revision ID has been assigned. This is for null edits.
2675 $this->mOutput->setFlag( 'vary-revision' );
2676 wfDebug( __METHOD__ . ": {{REVISIONUSER}} used, setting vary-revision...\n" );
2677 $value = $this->getRevisionUser();
2678 break;
2679 case 'namespace':
2680 $value = str_replace( '_',' ',$wgContLang->getNsText( $this->mTitle->getNamespace() ) );
2681 break;
2682 case 'namespacee':
2683 $value = wfUrlencode( $wgContLang->getNsText( $this->mTitle->getNamespace() ) );
2684 break;
2685 case 'talkspace':
2686 $value = $this->mTitle->canTalk() ? str_replace( '_',' ',$this->mTitle->getTalkNsText() ) : '';
2687 break;
2688 case 'talkspacee':
2689 $value = $this->mTitle->canTalk() ? wfUrlencode( $this->mTitle->getTalkNsText() ) : '';
2690 break;
2691 case 'subjectspace':
2692 $value = $this->mTitle->getSubjectNsText();
2693 break;
2694 case 'subjectspacee':
2695 $value = ( wfUrlencode( $this->mTitle->getSubjectNsText() ) );
2696 break;
2697 case 'currentdayname':
2698 $value = $wgContLang->getWeekdayName( gmdate( 'w', $ts ) + 1 );
2699 break;
2700 case 'currentyear':
2701 $value = $wgContLang->formatNum( gmdate( 'Y', $ts ), true );
2702 break;
2703 case 'currenttime':
2704 $value = $wgContLang->time( wfTimestamp( TS_MW, $ts ), false, false );
2705 break;
2706 case 'currenthour':
2707 $value = $wgContLang->formatNum( gmdate( 'H', $ts ), true );
2708 break;
2709 case 'currentweek':
2710 # @bug 4594 PHP5 has it zero padded, PHP4 does not, cast to
2711 # int to remove the padding
2712 $value = $wgContLang->formatNum( (int)gmdate( 'W', $ts ) );
2713 break;
2714 case 'currentdow':
2715 $value = $wgContLang->formatNum( gmdate( 'w', $ts ) );
2716 break;
2717 case 'localdayname':
2718 $value = $wgContLang->getWeekdayName( $localDayOfWeek + 1 );
2719 break;
2720 case 'localyear':
2721 $value = $wgContLang->formatNum( $localYear, true );
2722 break;
2723 case 'localtime':
2724 $value = $wgContLang->time( $localTimestamp, false, false );
2725 break;
2726 case 'localhour':
2727 $value = $wgContLang->formatNum( $localHour, true );
2728 break;
2729 case 'localweek':
2730 # @bug 4594 PHP5 has it zero padded, PHP4 does not, cast to
2731 # int to remove the padding
2732 $value = $wgContLang->formatNum( (int)$localWeek );
2733 break;
2734 case 'localdow':
2735 $value = $wgContLang->formatNum( $localDayOfWeek );
2736 break;
2737 case 'numberofarticles':
2738 $value = $wgContLang->formatNum( SiteStats::articles() );
2739 break;
2740 case 'numberoffiles':
2741 $value = $wgContLang->formatNum( SiteStats::images() );
2742 break;
2743 case 'numberofusers':
2744 $value = $wgContLang->formatNum( SiteStats::users() );
2745 break;
2746 case 'numberofactiveusers':
2747 $value = $wgContLang->formatNum( SiteStats::activeUsers() );
2748 break;
2749 case 'numberofpages':
2750 $value = $wgContLang->formatNum( SiteStats::pages() );
2751 break;
2752 case 'numberofadmins':
2753 $value = $wgContLang->formatNum( SiteStats::numberingroup( 'sysop' ) );
2754 break;
2755 case 'numberofedits':
2756 $value = $wgContLang->formatNum( SiteStats::edits() );
2757 break;
2758 case 'numberofviews':
2759 $value = $wgContLang->formatNum( SiteStats::views() );
2760 break;
2761 case 'currenttimestamp':
2762 $value = wfTimestamp( TS_MW, $ts );
2763 break;
2764 case 'localtimestamp':
2765 $value = $localTimestamp;
2766 break;
2767 case 'currentversion':
2768 $value = SpecialVersion::getVersion();
2769 break;
2770 case 'articlepath':
2771 return $wgArticlePath;
2772 case 'sitename':
2773 return $wgSitename;
2774 case 'server':
2775 return $wgServer;
2776 case 'servername':
2777 wfSuppressWarnings(); # May give an E_WARNING in PHP < 5.3.3
2778 $serverName = parse_url( $wgServer, PHP_URL_HOST );
2779 wfRestoreWarnings();
2780 return $serverName ? $serverName : $wgServer;
2781 case 'scriptpath':
2782 return $wgScriptPath;
2783 case 'stylepath':
2784 return $wgStylePath;
2785 case 'directionmark':
2786 return $wgContLang->getDirMark();
2787 case 'contentlanguage':
2788 global $wgLanguageCode;
2789 return $wgLanguageCode;
2790 default:
2791 $ret = null;
2792 if ( wfRunHooks( 'ParserGetVariableValueSwitch', array( &$this, &$this->mVarCache, &$index, &$ret, &$frame ) ) ) {
2793 return $ret;
2794 } else {
2795 return null;
2796 }
2797 }
2798
2799 if ( $index )
2800 $this->mVarCache[$index] = $value;
2801
2802 return $value;
2803 }
2804
2805 /**
2806 * initialise the magic variables (like CURRENTMONTHNAME) and substitution modifiers
2807 *
2808 * @private
2809 */
2810 function initialiseVariables() {
2811 wfProfileIn( __METHOD__ );
2812 $variableIDs = MagicWord::getVariableIDs();
2813 $substIDs = MagicWord::getSubstIDs();
2814
2815 $this->mVariables = new MagicWordArray( $variableIDs );
2816 $this->mSubstWords = new MagicWordArray( $substIDs );
2817 wfProfileOut( __METHOD__ );
2818 }
2819
2820 /**
2821 * Preprocess some wikitext and return the document tree.
2822 * This is the ghost of replace_variables().
2823 *
2824 * @param $text String: The text to parse
2825 * @param $flags Integer: bitwise combination of:
2826 * self::PTD_FOR_INCLUSION Handle <noinclude>/<includeonly> as if the text is being
2827 * included. Default is to assume a direct page view.
2828 *
2829 * The generated DOM tree must depend only on the input text and the flags.
2830 * The DOM tree must be the same in OT_HTML and OT_WIKI mode, to avoid a regression of bug 4899.
2831 *
2832 * Any flag added to the $flags parameter here, or any other parameter liable to cause a
2833 * change in the DOM tree for a given text, must be passed through the section identifier
2834 * in the section edit link and thus back to extractSections().
2835 *
2836 * The output of this function is currently only cached in process memory, but a persistent
2837 * cache may be implemented at a later date which takes further advantage of these strict
2838 * dependency requirements.
2839 *
2840 * @private
2841 */
2842 function preprocessToDom( $text, $flags = 0 ) {
2843 $dom = $this->getPreprocessor()->preprocessToObj( $text, $flags );
2844 return $dom;
2845 }
2846
2847 /**
2848 * Return a three-element array: leading whitespace, string contents, trailing whitespace
2849 */
2850 public static function splitWhitespace( $s ) {
2851 $ltrimmed = ltrim( $s );
2852 $w1 = substr( $s, 0, strlen( $s ) - strlen( $ltrimmed ) );
2853 $trimmed = rtrim( $ltrimmed );
2854 $diff = strlen( $ltrimmed ) - strlen( $trimmed );
2855 if ( $diff > 0 ) {
2856 $w2 = substr( $ltrimmed, -$diff );
2857 } else {
2858 $w2 = '';
2859 }
2860 return array( $w1, $trimmed, $w2 );
2861 }
2862
2863 /**
2864 * Replace magic variables, templates, and template arguments
2865 * with the appropriate text. Templates are substituted recursively,
2866 * taking care to avoid infinite loops.
2867 *
2868 * Note that the substitution depends on value of $mOutputType:
2869 * self::OT_WIKI: only {{subst:}} templates
2870 * self::OT_PREPROCESS: templates but not extension tags
2871 * self::OT_HTML: all templates and extension tags
2872 *
2873 * @param $text String: the text to transform
2874 * @param $frame PPFrame Object describing the arguments passed to the template.
2875 * Arguments may also be provided as an associative array, as was the usual case before MW1.12.
2876 * Providing arguments this way may be useful for extensions wishing to perform variable replacement explicitly.
2877 * @param $argsOnly Boolean: only do argument (triple-brace) expansion, not double-brace expansion
2878 * @private
2879 */
2880 function replaceVariables( $text, $frame = false, $argsOnly = false ) {
2881 # Is there any text? Also, Prevent too big inclusions!
2882 if ( strlen( $text ) < 1 || strlen( $text ) > $this->mOptions->getMaxIncludeSize() ) {
2883 return $text;
2884 }
2885 wfProfileIn( __METHOD__ );
2886
2887 if ( $frame === false ) {
2888 $frame = $this->getPreprocessor()->newFrame();
2889 } elseif ( !( $frame instanceof PPFrame ) ) {
2890 wfDebug( __METHOD__." called using plain parameters instead of a PPFrame instance. Creating custom frame.\n" );
2891 $frame = $this->getPreprocessor()->newCustomFrame( $frame );
2892 }
2893
2894 $dom = $this->preprocessToDom( $text );
2895 $flags = $argsOnly ? PPFrame::NO_TEMPLATES : 0;
2896 $text = $frame->expand( $dom, $flags );
2897
2898 wfProfileOut( __METHOD__ );
2899 return $text;
2900 }
2901
2902 # Clean up argument array - refactored in 1.9 so parserfunctions can use it, too.
2903 static function createAssocArgs( $args ) {
2904 $assocArgs = array();
2905 $index = 1;
2906 foreach ( $args as $arg ) {
2907 $eqpos = strpos( $arg, '=' );
2908 if ( $eqpos === false ) {
2909 $assocArgs[$index++] = $arg;
2910 } else {
2911 $name = trim( substr( $arg, 0, $eqpos ) );
2912 $value = trim( substr( $arg, $eqpos+1 ) );
2913 if ( $value === false ) {
2914 $value = '';
2915 }
2916 if ( $name !== false ) {
2917 $assocArgs[$name] = $value;
2918 }
2919 }
2920 }
2921
2922 return $assocArgs;
2923 }
2924
2925 /**
2926 * Warn the user when a parser limitation is reached
2927 * Will warn at most once the user per limitation type
2928 *
2929 * @param $limitationType String: should be one of:
2930 * 'expensive-parserfunction' (corresponding messages:
2931 * 'expensive-parserfunction-warning',
2932 * 'expensive-parserfunction-category')
2933 * 'post-expand-template-argument' (corresponding messages:
2934 * 'post-expand-template-argument-warning',
2935 * 'post-expand-template-argument-category')
2936 * 'post-expand-template-inclusion' (corresponding messages:
2937 * 'post-expand-template-inclusion-warning',
2938 * 'post-expand-template-inclusion-category')
2939 * @param $current Current value
2940 * @param $max Maximum allowed, when an explicit limit has been
2941 * exceeded, provide the values (optional)
2942 */
2943 function limitationWarn( $limitationType, $current=null, $max=null) {
2944 # does no harm if $current and $max are present but are unnecessary for the message
2945 $warning = wfMsgExt( "$limitationType-warning", array( 'parsemag', 'escape' ), $current, $max );
2946 $this->mOutput->addWarning( $warning );
2947 $this->addTrackingCategory( "$limitationType-category" );
2948 }
2949
2950 /**
2951 * Return the text of a template, after recursively
2952 * replacing any variables or templates within the template.
2953 *
2954 * @param $piece Array: the parts of the template
2955 * $piece['title']: the title, i.e. the part before the |
2956 * $piece['parts']: the parameter array
2957 * $piece['lineStart']: whether the brace was at the start of a line
2958 * @param $frame PPFrame The current frame, contains template arguments
2959 * @return String: the text of the template
2960 * @private
2961 */
2962 function braceSubstitution( $piece, $frame ) {
2963 global $wgContLang, $wgNonincludableNamespaces;
2964 wfProfileIn( __METHOD__ );
2965 wfProfileIn( __METHOD__.'-setup' );
2966
2967 # Flags
2968 $found = false; # $text has been filled
2969 $nowiki = false; # wiki markup in $text should be escaped
2970 $isHTML = false; # $text is HTML, armour it against wikitext transformation
2971 $forceRawInterwiki = false; # Force interwiki transclusion to be done in raw mode not rendered
2972 $isChildObj = false; # $text is a DOM node needing expansion in a child frame
2973 $isLocalObj = false; # $text is a DOM node needing expansion in the current frame
2974
2975 # Title object, where $text came from
2976 $title = null;
2977
2978 # $part1 is the bit before the first |, and must contain only title characters.
2979 # Various prefixes will be stripped from it later.
2980 $titleWithSpaces = $frame->expand( $piece['title'] );
2981 $part1 = trim( $titleWithSpaces );
2982 $titleText = false;
2983
2984 # Original title text preserved for various purposes
2985 $originalTitle = $part1;
2986
2987 # $args is a list of argument nodes, starting from index 0, not including $part1
2988 $args = ( null == $piece['parts'] ) ? array() : $piece['parts'];
2989 wfProfileOut( __METHOD__.'-setup' );
2990
2991 # SUBST
2992 wfProfileIn( __METHOD__.'-modifiers' );
2993 if ( !$found ) {
2994
2995 $substMatch = $this->mSubstWords->matchStartAndRemove( $part1 );
2996
2997 # Possibilities for substMatch: "subst", "safesubst" or FALSE
2998 # Decide whether to expand template or keep wikitext as-is.
2999 if ( $this->ot['wiki'] ) {
3000 if ( $substMatch === false ) {
3001 $literal = true; # literal when in PST with no prefix
3002 } else {
3003 $literal = false; # expand when in PST with subst: or safesubst:
3004 }
3005 } else {
3006 if ( $substMatch == 'subst' ) {
3007 $literal = true; # literal when not in PST with plain subst:
3008 } else {
3009 $literal = false; # expand when not in PST with safesubst: or no prefix
3010 }
3011 }
3012 if ( $literal ) {
3013 $text = $frame->virtualBracketedImplode( '{{', '|', '}}', $titleWithSpaces, $args );
3014 $isLocalObj = true;
3015 $found = true;
3016 }
3017 }
3018
3019 # Variables
3020 if ( !$found && $args->getLength() == 0 ) {
3021 $id = $this->mVariables->matchStartToEnd( $part1 );
3022 if ( $id !== false ) {
3023 $text = $this->getVariableValue( $id, $frame );
3024 if ( MagicWord::getCacheTTL( $id ) > -1 ) {
3025 $this->mOutput->updateCacheExpiry( MagicWord::getCacheTTL( $id ) );
3026 }
3027 $found = true;
3028 }
3029 }
3030
3031 # MSG, MSGNW and RAW
3032 if ( !$found ) {
3033 # Check for MSGNW:
3034 $mwMsgnw = MagicWord::get( 'msgnw' );
3035 if ( $mwMsgnw->matchStartAndRemove( $part1 ) ) {
3036 $nowiki = true;
3037 } else {
3038 # Remove obsolete MSG:
3039 $mwMsg = MagicWord::get( 'msg' );
3040 $mwMsg->matchStartAndRemove( $part1 );
3041 }
3042
3043 # Check for RAW:
3044 $mwRaw = MagicWord::get( 'raw' );
3045 if ( $mwRaw->matchStartAndRemove( $part1 ) ) {
3046 $forceRawInterwiki = true;
3047 }
3048 }
3049 wfProfileOut( __METHOD__.'-modifiers' );
3050
3051 # Parser functions
3052 if ( !$found ) {
3053 wfProfileIn( __METHOD__ . '-pfunc' );
3054
3055 $colonPos = strpos( $part1, ':' );
3056 if ( $colonPos !== false ) {
3057 # Case sensitive functions
3058 $function = substr( $part1, 0, $colonPos );
3059 if ( isset( $this->mFunctionSynonyms[1][$function] ) ) {
3060 $function = $this->mFunctionSynonyms[1][$function];
3061 } else {
3062 # Case insensitive functions
3063 $function = $wgContLang->lc( $function );
3064 if ( isset( $this->mFunctionSynonyms[0][$function] ) ) {
3065 $function = $this->mFunctionSynonyms[0][$function];
3066 } else {
3067 $function = false;
3068 }
3069 }
3070 if ( $function ) {
3071 list( $callback, $flags ) = $this->mFunctionHooks[$function];
3072 $initialArgs = array( &$this );
3073 $funcArgs = array( trim( substr( $part1, $colonPos + 1 ) ) );
3074 if ( $flags & SFH_OBJECT_ARGS ) {
3075 # Add a frame parameter, and pass the arguments as an array
3076 $allArgs = $initialArgs;
3077 $allArgs[] = $frame;
3078 for ( $i = 0; $i < $args->getLength(); $i++ ) {
3079 $funcArgs[] = $args->item( $i );
3080 }
3081 $allArgs[] = $funcArgs;
3082 } else {
3083 # Convert arguments to plain text
3084 for ( $i = 0; $i < $args->getLength(); $i++ ) {
3085 $funcArgs[] = trim( $frame->expand( $args->item( $i ) ) );
3086 }
3087 $allArgs = array_merge( $initialArgs, $funcArgs );
3088 }
3089
3090 # Workaround for PHP bug 35229 and similar
3091 if ( !is_callable( $callback ) ) {
3092 wfProfileOut( __METHOD__ . '-pfunc' );
3093 wfProfileOut( __METHOD__ );
3094 throw new MWException( "Tag hook for $function is not callable\n" );
3095 }
3096 $result = call_user_func_array( $callback, $allArgs );
3097 $found = true;
3098 $noparse = true;
3099 $preprocessFlags = 0;
3100
3101 if ( is_array( $result ) ) {
3102 if ( isset( $result[0] ) ) {
3103 $text = $result[0];
3104 unset( $result[0] );
3105 }
3106
3107 # Extract flags into the local scope
3108 # This allows callers to set flags such as nowiki, found, etc.
3109 extract( $result );
3110 } else {
3111 $text = $result;
3112 }
3113 if ( !$noparse ) {
3114 $text = $this->preprocessToDom( $text, $preprocessFlags );
3115 $isChildObj = true;
3116 }
3117 }
3118 }
3119 wfProfileOut( __METHOD__ . '-pfunc' );
3120 }
3121
3122 # Finish mangling title and then check for loops.
3123 # Set $title to a Title object and $titleText to the PDBK
3124 if ( !$found ) {
3125 $ns = NS_TEMPLATE;
3126 # Split the title into page and subpage
3127 $subpage = '';
3128 $part1 = $this->maybeDoSubpageLink( $part1, $subpage );
3129 if ( $subpage !== '' ) {
3130 $ns = $this->mTitle->getNamespace();
3131 }
3132 $title = Title::newFromText( $part1, $ns );
3133 if ( $title ) {
3134 $titleText = $title->getPrefixedText();
3135 # Check for language variants if the template is not found
3136 if ( $wgContLang->hasVariants() && $title->getArticleID() == 0 ) {
3137 $wgContLang->findVariantLink( $part1, $title, true );
3138 }
3139 # Do recursion depth check
3140 $limit = $this->mOptions->getMaxTemplateDepth();
3141 if ( $frame->depth >= $limit ) {
3142 $found = true;
3143 $text = '<span class="error">'
3144 . wfMsgForContent( 'parser-template-recursion-depth-warning', $limit )
3145 . '</span>';
3146 }
3147 }
3148 }
3149
3150 # Load from database
3151 if ( !$found && $title ) {
3152 wfProfileIn( __METHOD__ . '-loadtpl' );
3153 if ( !$title->isExternal() ) {
3154 if ( $title->getNamespace() == NS_SPECIAL
3155 && $this->mOptions->getAllowSpecialInclusion()
3156 && $this->ot['html'] )
3157 {
3158 $text = SpecialPage::capturePath( $title );
3159 if ( is_string( $text ) ) {
3160 $found = true;
3161 $isHTML = true;
3162 $this->disableCache();
3163 }
3164 } elseif ( $wgNonincludableNamespaces && in_array( $title->getNamespace(), $wgNonincludableNamespaces ) ) {
3165 $found = false; # access denied
3166 wfDebug( __METHOD__.": template inclusion denied for " . $title->getPrefixedDBkey() );
3167 } else {
3168 list( $text, $title ) = $this->getTemplateDom( $title );
3169 if ( $text !== false ) {
3170 $found = true;
3171 $isChildObj = true;
3172 }
3173 }
3174
3175 # If the title is valid but undisplayable, make a link to it
3176 if ( !$found && ( $this->ot['html'] || $this->ot['pre'] ) ) {
3177 $text = "[[:$titleText]]";
3178 $found = true;
3179 }
3180 } elseif ( $title->isTrans() ) {
3181 # Interwiki transclusion
3182 if ( $this->ot['html'] && !$forceRawInterwiki ) {
3183 $text = $this->interwikiTransclude( $title, 'render' );
3184 $isHTML = true;
3185 } else {
3186 $text = $this->interwikiTransclude( $title, 'raw' );
3187 # Preprocess it like a template
3188 $text = $this->preprocessToDom( $text, self::PTD_FOR_INCLUSION );
3189 $isChildObj = true;
3190 }
3191 $found = true;
3192 }
3193
3194 # Do infinite loop check
3195 # This has to be done after redirect resolution to avoid infinite loops via redirects
3196 if ( !$frame->loopCheck( $title ) ) {
3197 $found = true;
3198 $text = '<span class="error">' . wfMsgForContent( 'parser-template-loop-warning', $titleText ) . '</span>';
3199 wfDebug( __METHOD__.": template loop broken at '$titleText'\n" );
3200 }
3201 wfProfileOut( __METHOD__ . '-loadtpl' );
3202 }
3203
3204 # If we haven't found text to substitute by now, we're done
3205 # Recover the source wikitext and return it
3206 if ( !$found ) {
3207 $text = $frame->virtualBracketedImplode( '{{', '|', '}}', $titleWithSpaces, $args );
3208 wfProfileOut( __METHOD__ );
3209 return array( 'object' => $text );
3210 }
3211
3212 # Expand DOM-style return values in a child frame
3213 if ( $isChildObj ) {
3214 # Clean up argument array
3215 $newFrame = $frame->newChild( $args, $title );
3216
3217 if ( $nowiki ) {
3218 $text = $newFrame->expand( $text, PPFrame::RECOVER_ORIG );
3219 } elseif ( $titleText !== false && $newFrame->isEmpty() ) {
3220 # Expansion is eligible for the empty-frame cache
3221 if ( isset( $this->mTplExpandCache[$titleText] ) ) {
3222 $text = $this->mTplExpandCache[$titleText];
3223 } else {
3224 $text = $newFrame->expand( $text );
3225 $this->mTplExpandCache[$titleText] = $text;
3226 }
3227 } else {
3228 # Uncached expansion
3229 $text = $newFrame->expand( $text );
3230 }
3231 }
3232 if ( $isLocalObj && $nowiki ) {
3233 $text = $frame->expand( $text, PPFrame::RECOVER_ORIG );
3234 $isLocalObj = false;
3235 }
3236
3237 # Replace raw HTML by a placeholder
3238 # Add a blank line preceding, to prevent it from mucking up
3239 # immediately preceding headings
3240 if ( $isHTML ) {
3241 $text = "\n\n" . $this->insertStripItem( $text );
3242 } elseif ( $nowiki && ( $this->ot['html'] || $this->ot['pre'] ) ) {
3243 # Escape nowiki-style return values
3244 $text = wfEscapeWikiText( $text );
3245 } elseif ( is_string( $text )
3246 && !$piece['lineStart']
3247 && preg_match( '/^(?:{\\||:|;|#|\*)/', $text ) )
3248 {
3249 # Bug 529: if the template begins with a table or block-level
3250 # element, it should be treated as beginning a new line.
3251 # This behaviour is somewhat controversial.
3252 $text = "\n" . $text;
3253 }
3254
3255 if ( is_string( $text ) && !$this->incrementIncludeSize( 'post-expand', strlen( $text ) ) ) {
3256 # Error, oversize inclusion
3257 if ( $titleText !== false ) {
3258 # Make a working, properly escaped link if possible (bug 23588)
3259 $text = "[[:$titleText]]";
3260 } else {
3261 # This will probably not be a working link, but at least it may
3262 # provide some hint of where the problem is
3263 preg_replace( '/^:/', '', $originalTitle );
3264 $text = "[[:$originalTitle]]";
3265 }
3266 $text .= $this->insertStripItem( '<!-- WARNING: template omitted, post-expand include size too large -->' );
3267 $this->limitationWarn( 'post-expand-template-inclusion' );
3268 }
3269
3270 if ( $isLocalObj ) {
3271 $ret = array( 'object' => $text );
3272 } else {
3273 $ret = array( 'text' => $text );
3274 }
3275
3276 wfProfileOut( __METHOD__ );
3277 return $ret;
3278 }
3279
3280 /**
3281 * Get the semi-parsed DOM representation of a template with a given title,
3282 * and its redirect destination title. Cached.
3283 */
3284 function getTemplateDom( $title ) {
3285 $cacheTitle = $title;
3286 $titleText = $title->getPrefixedDBkey();
3287
3288 if ( isset( $this->mTplRedirCache[$titleText] ) ) {
3289 list( $ns, $dbk ) = $this->mTplRedirCache[$titleText];
3290 $title = Title::makeTitle( $ns, $dbk );
3291 $titleText = $title->getPrefixedDBkey();
3292 }
3293 if ( isset( $this->mTplDomCache[$titleText] ) ) {
3294 return array( $this->mTplDomCache[$titleText], $title );
3295 }
3296
3297 # Cache miss, go to the database
3298 list( $text, $title ) = $this->fetchTemplateAndTitle( $title );
3299
3300 if ( $text === false ) {
3301 $this->mTplDomCache[$titleText] = false;
3302 return array( false, $title );
3303 }
3304
3305 $dom = $this->preprocessToDom( $text, self::PTD_FOR_INCLUSION );
3306 $this->mTplDomCache[ $titleText ] = $dom;
3307
3308 if ( !$title->equals( $cacheTitle ) ) {
3309 $this->mTplRedirCache[$cacheTitle->getPrefixedDBkey()] =
3310 array( $title->getNamespace(), $cdb = $title->getDBkey() );
3311 }
3312
3313 return array( $dom, $title );
3314 }
3315
3316 /**
3317 * Fetch the unparsed text of a template and register a reference to it.
3318 */
3319 function fetchTemplateAndTitle( $title ) {
3320 $templateCb = $this->mOptions->getTemplateCallback(); # Defaults to Parser::statelessFetchTemplate()
3321 $stuff = call_user_func( $templateCb, $title, $this );
3322 $text = $stuff['text'];
3323 $finalTitle = isset( $stuff['finalTitle'] ) ? $stuff['finalTitle'] : $title;
3324 if ( isset( $stuff['deps'] ) ) {
3325 foreach ( $stuff['deps'] as $dep ) {
3326 $this->mOutput->addTemplate( $dep['title'], $dep['page_id'], $dep['rev_id'] );
3327 }
3328 }
3329 return array( $text, $finalTitle );
3330 }
3331
3332 function fetchTemplate( $title ) {
3333 $rv = $this->fetchTemplateAndTitle( $title );
3334 return $rv[0];
3335 }
3336
3337 /**
3338 * Static function to get a template
3339 * Can be overridden via ParserOptions::setTemplateCallback().
3340 */
3341 static function statelessFetchTemplate( $title, $parser=false ) {
3342 $text = $skip = false;
3343 $finalTitle = $title;
3344 $deps = array();
3345
3346 # Loop to fetch the article, with up to 1 redirect
3347 for ( $i = 0; $i < 2 && is_object( $title ); $i++ ) {
3348 # Give extensions a chance to select the revision instead
3349 $id = false; # Assume current
3350 wfRunHooks( 'BeforeParserFetchTemplateAndtitle', array( $parser, &$title, &$skip, &$id ) );
3351
3352 if ( $skip ) {
3353 $text = false;
3354 $deps[] = array(
3355 'title' => $title,
3356 'page_id' => $title->getArticleID(),
3357 'rev_id' => null );
3358 break;
3359 }
3360 $rev = $id ? Revision::newFromId( $id ) : Revision::newFromTitle( $title );
3361 $rev_id = $rev ? $rev->getId() : 0;
3362 # If there is no current revision, there is no page
3363 if ( $id === false && !$rev ) {
3364 $linkCache = LinkCache::singleton();
3365 $linkCache->addBadLinkObj( $title );
3366 }
3367
3368 $deps[] = array(
3369 'title' => $title,
3370 'page_id' => $title->getArticleID(),
3371 'rev_id' => $rev_id );
3372
3373 if ( $rev ) {
3374 $text = $rev->getText();
3375 } elseif ( $title->getNamespace() == NS_MEDIAWIKI ) {
3376 global $wgContLang;
3377 $message = $wgContLang->lcfirst( $title->getText() );
3378 $text = wfMsgForContentNoTrans( $message );
3379 if ( wfEmptyMsg( $message, $text ) ) {
3380 $text = false;
3381 break;
3382 }
3383 } else {
3384 break;
3385 }
3386 if ( $text === false ) {
3387 break;
3388 }
3389 # Redirect?
3390 $finalTitle = $title;
3391 $title = Title::newFromRedirect( $text );
3392 }
3393 return array(
3394 'text' => $text,
3395 'finalTitle' => $finalTitle,
3396 'deps' => $deps );
3397 }
3398
3399 /**
3400 * Transclude an interwiki link.
3401 */
3402 function interwikiTransclude( $title, $action ) {
3403 global $wgEnableScaryTranscluding;
3404
3405 if ( !$wgEnableScaryTranscluding ) {
3406 return wfMsgForContent('scarytranscludedisabled');
3407 }
3408
3409 $url = $title->getFullUrl( "action=$action" );
3410
3411 if ( strlen( $url ) > 255 ) {
3412 return wfMsgForContent( 'scarytranscludetoolong' );
3413 }
3414 return $this->fetchScaryTemplateMaybeFromCache( $url );
3415 }
3416
3417 function fetchScaryTemplateMaybeFromCache( $url ) {
3418 global $wgTranscludeCacheExpiry;
3419 $dbr = wfGetDB( DB_SLAVE );
3420 $tsCond = $dbr->timestamp( time() - $wgTranscludeCacheExpiry );
3421 $obj = $dbr->selectRow( 'transcache', array('tc_time', 'tc_contents' ),
3422 array( 'tc_url' => $url, "tc_time >= " . $dbr->addQuotes( $tsCond ) ) );
3423 if ( $obj ) {
3424 return $obj->tc_contents;
3425 }
3426
3427 $text = Http::get( $url );
3428 if ( !$text ) {
3429 return wfMsgForContent( 'scarytranscludefailed', $url );
3430 }
3431
3432 $dbw = wfGetDB( DB_MASTER );
3433 $dbw->replace( 'transcache', array('tc_url'), array(
3434 'tc_url' => $url,
3435 'tc_time' => $dbw->timestamp( time() ),
3436 'tc_contents' => $text)
3437 );
3438 return $text;
3439 }
3440
3441
3442 /**
3443 * Triple brace replacement -- used for template arguments
3444 * @private
3445 */
3446 function argSubstitution( $piece, $frame ) {
3447 wfProfileIn( __METHOD__ );
3448
3449 $error = false;
3450 $parts = $piece['parts'];
3451 $nameWithSpaces = $frame->expand( $piece['title'] );
3452 $argName = trim( $nameWithSpaces );
3453 $object = false;
3454 $text = $frame->getArgument( $argName );
3455 if ( $text === false && $parts->getLength() > 0
3456 && (
3457 $this->ot['html']
3458 || $this->ot['pre']
3459 || ( $this->ot['wiki'] && $frame->isTemplate() )
3460 )
3461 ) {
3462 # No match in frame, use the supplied default
3463 $object = $parts->item( 0 )->getChildren();
3464 }
3465 if ( !$this->incrementIncludeSize( 'arg', strlen( $text ) ) ) {
3466 $error = '<!-- WARNING: argument omitted, expansion size too large -->';
3467 $this->limitationWarn( 'post-expand-template-argument' );
3468 }
3469
3470 if ( $text === false && $object === false ) {
3471 # No match anywhere
3472 $object = $frame->virtualBracketedImplode( '{{{', '|', '}}}', $nameWithSpaces, $parts );
3473 }
3474 if ( $error !== false ) {
3475 $text .= $error;
3476 }
3477 if ( $object !== false ) {
3478 $ret = array( 'object' => $object );
3479 } else {
3480 $ret = array( 'text' => $text );
3481 }
3482
3483 wfProfileOut( __METHOD__ );
3484 return $ret;
3485 }
3486
3487 /**
3488 * Return the text to be used for a given extension tag.
3489 * This is the ghost of strip().
3490 *
3491 * @param $params Associative array of parameters:
3492 * name PPNode for the tag name
3493 * attr PPNode for unparsed text where tag attributes are thought to be
3494 * attributes Optional associative array of parsed attributes
3495 * inner Contents of extension element
3496 * noClose Original text did not have a close tag
3497 * @param $frame PPFrame
3498 */
3499 function extensionSubstitution( $params, $frame ) {
3500 $name = $frame->expand( $params['name'] );
3501 $attrText = !isset( $params['attr'] ) ? null : $frame->expand( $params['attr'] );
3502 $content = !isset( $params['inner'] ) ? null : $frame->expand( $params['inner'] );
3503 $marker = "{$this->mUniqPrefix}-$name-" . sprintf( '%08X', $this->mMarkerIndex++ ) . self::MARKER_SUFFIX;
3504
3505 $isFunctionTag = isset( $this->mFunctionTagHooks[strtolower($name)] ) &&
3506 ( $this->ot['html'] || $this->ot['pre'] );
3507 if ( $isFunctionTag ) {
3508 $markerType = 'none';
3509 } else {
3510 $markerType = 'general';
3511 }
3512 if ( $this->ot['html'] || $isFunctionTag ) {
3513 $name = strtolower( $name );
3514 $attributes = Sanitizer::decodeTagAttributes( $attrText );
3515 if ( isset( $params['attributes'] ) ) {
3516 $attributes = $attributes + $params['attributes'];
3517 }
3518
3519 if ( isset( $this->mTagHooks[$name] ) ) {
3520 # Workaround for PHP bug 35229 and similar
3521 if ( !is_callable( $this->mTagHooks[$name] ) ) {
3522 throw new MWException( "Tag hook for $name is not callable\n" );
3523 }
3524 $output = call_user_func_array( $this->mTagHooks[$name],
3525 array( $content, $attributes, $this, $frame ) );
3526 } elseif ( isset( $this->mFunctionTagHooks[$name] ) ) {
3527 list( $callback, $flags ) = $this->mFunctionTagHooks[$name];
3528 if ( !is_callable( $callback ) ) {
3529 throw new MWException( "Tag hook for $name is not callable\n" );
3530 }
3531
3532 $output = call_user_func_array( $callback, array( &$this, $frame, $content, $attributes ) );
3533 } else {
3534 $output = '<span class="error">Invalid tag extension name: ' .
3535 htmlspecialchars( $name ) . '</span>';
3536 }
3537
3538 if ( is_array( $output ) ) {
3539 # Extract flags to local scope (to override $markerType)
3540 $flags = $output;
3541 $output = $flags[0];
3542 unset( $flags[0] );
3543 extract( $flags );
3544 }
3545 } else {
3546 if ( is_null( $attrText ) ) {
3547 $attrText = '';
3548 }
3549 if ( isset( $params['attributes'] ) ) {
3550 foreach ( $params['attributes'] as $attrName => $attrValue ) {
3551 $attrText .= ' ' . htmlspecialchars( $attrName ) . '="' .
3552 htmlspecialchars( $attrValue ) . '"';
3553 }
3554 }
3555 if ( $content === null ) {
3556 $output = "<$name$attrText/>";
3557 } else {
3558 $close = is_null( $params['close'] ) ? '' : $frame->expand( $params['close'] );
3559 $output = "<$name$attrText>$content$close";
3560 }
3561 }
3562
3563 if ( $markerType === 'none' ) {
3564 return $output;
3565 } elseif ( $markerType === 'nowiki' ) {
3566 $this->mStripState->nowiki->setPair( $marker, $output );
3567 } elseif ( $markerType === 'general' ) {
3568 $this->mStripState->general->setPair( $marker, $output );
3569 } else {
3570 throw new MWException( __METHOD__.': invalid marker type' );
3571 }
3572 return $marker;
3573 }
3574
3575 /**
3576 * Increment an include size counter
3577 *
3578 * @param $type String: the type of expansion
3579 * @param $size Integer: the size of the text
3580 * @return Boolean: false if this inclusion would take it over the maximum, true otherwise
3581 */
3582 function incrementIncludeSize( $type, $size ) {
3583 if ( $this->mIncludeSizes[$type] + $size > $this->mOptions->getMaxIncludeSize( $type ) ) {
3584 return false;
3585 } else {
3586 $this->mIncludeSizes[$type] += $size;
3587 return true;
3588 }
3589 }
3590
3591 /**
3592 * Increment the expensive function count
3593 *
3594 * @return Boolean: false if the limit has been exceeded
3595 */
3596 function incrementExpensiveFunctionCount() {
3597 global $wgExpensiveParserFunctionLimit;
3598 $this->mExpensiveFunctionCount++;
3599 if ( $this->mExpensiveFunctionCount <= $wgExpensiveParserFunctionLimit ) {
3600 return true;
3601 }
3602 return false;
3603 }
3604
3605 /**
3606 * Strip double-underscore items like __NOGALLERY__ and __NOTOC__
3607 * Fills $this->mDoubleUnderscores, returns the modified text
3608 */
3609 function doDoubleUnderscore( $text ) {
3610 wfProfileIn( __METHOD__ );
3611
3612 # The position of __TOC__ needs to be recorded
3613 $mw = MagicWord::get( 'toc' );
3614 if ( $mw->match( $text ) ) {
3615 $this->mShowToc = true;
3616 $this->mForceTocPosition = true;
3617
3618 # Set a placeholder. At the end we'll fill it in with the TOC.
3619 $text = $mw->replace( '<!--MWTOC-->', $text, 1 );
3620
3621 # Only keep the first one.
3622 $text = $mw->replace( '', $text );
3623 }
3624
3625 # Now match and remove the rest of them
3626 $mwa = MagicWord::getDoubleUnderscoreArray();
3627 $this->mDoubleUnderscores = $mwa->matchAndRemove( $text );
3628
3629 if ( isset( $this->mDoubleUnderscores['nogallery'] ) ) {
3630 $this->mOutput->mNoGallery = true;
3631 }
3632 if ( isset( $this->mDoubleUnderscores['notoc'] ) && !$this->mForceTocPosition ) {
3633 $this->mShowToc = false;
3634 }
3635 if ( isset( $this->mDoubleUnderscores['hiddencat'] ) && $this->mTitle->getNamespace() == NS_CATEGORY ) {
3636 $this->addTrackingCategory( 'hidden-category-category' );
3637 }
3638 # (bug 8068) Allow control over whether robots index a page.
3639 #
3640 # FIXME (bug 14899): __INDEX__ always overrides __NOINDEX__ here! This
3641 # is not desirable, the last one on the page should win.
3642 if ( isset( $this->mDoubleUnderscores['noindex'] ) && $this->mTitle->canUseNoindex() ) {
3643 $this->mOutput->setIndexPolicy( 'noindex' );
3644 $this->addTrackingCategory( 'noindex-category' );
3645 }
3646 if ( isset( $this->mDoubleUnderscores['index'] ) && $this->mTitle->canUseNoindex() ) {
3647 $this->mOutput->setIndexPolicy( 'index' );
3648 $this->addTrackingCategory( 'index-category' );
3649 }
3650
3651 # Cache all double underscores in the database
3652 foreach ( $this->mDoubleUnderscores as $key => $val ) {
3653 $this->mOutput->setProperty( $key, '' );
3654 }
3655
3656 wfProfileOut( __METHOD__ );
3657 return $text;
3658 }
3659
3660 /**
3661 * Add a tracking category, getting the title from a system message,
3662 * or print a debug message if the title is invalid.
3663 *
3664 * @param $msg String: message key
3665 * @return Boolean: whether the addition was successful
3666 */
3667 protected function addTrackingCategory( $msg ) {
3668 $cat = wfMsgForContent( $msg );
3669
3670 # Allow tracking categories to be disabled by setting them to "-"
3671 if ( $cat === '-' ) {
3672 return false;
3673 }
3674
3675 $containerCategory = Title::makeTitleSafe( NS_CATEGORY, $cat );
3676 if ( $containerCategory ) {
3677 $this->mOutput->addCategory( $containerCategory->getDBkey(), $this->getDefaultSort() );
3678 return true;
3679 } else {
3680 wfDebug( __METHOD__.": [[MediaWiki:$msg]] is not a valid title!\n" );
3681 return false;
3682 }
3683 }
3684
3685 /**
3686 * This function accomplishes several tasks:
3687 * 1) Auto-number headings if that option is enabled
3688 * 2) Add an [edit] link to sections for users who have enabled the option and can edit the page
3689 * 3) Add a Table of contents on the top for users who have enabled the option
3690 * 4) Auto-anchor headings
3691 *
3692 * It loops through all headlines, collects the necessary data, then splits up the
3693 * string and re-inserts the newly formatted headlines.
3694 *
3695 * @param $text String
3696 * @param $origText String: original, untouched wikitext
3697 * @param $isMain Boolean
3698 * @private
3699 */
3700 function formatHeadings( $text, $origText, $isMain=true ) {
3701 global $wgMaxTocLevel, $wgContLang, $wgHtml5, $wgExperimentalHtmlIds;
3702
3703 $doNumberHeadings = $this->mOptions->getNumberHeadings();
3704
3705 # Inhibit editsection links if requested in the page
3706 if ( isset( $this->mDoubleUnderscores['noeditsection'] ) ) {
3707 $showEditLink = 0;
3708 } else {
3709 $showEditLink = $this->mOptions->getEditSection();
3710 }
3711
3712 # Get all headlines for numbering them and adding funky stuff like [edit]
3713 # links - this is for later, but we need the number of headlines right now
3714 $matches = array();
3715 $numMatches = preg_match_all( '/<H(?P<level>[1-6])(?P<attrib>.*?'.'>)(?P<header>.*?)<\/H[1-6] *>/i', $text, $matches );
3716
3717 # if there are fewer than 4 headlines in the article, do not show TOC
3718 # unless it's been explicitly enabled.
3719 $enoughToc = $this->mShowToc &&
3720 ( ( $numMatches >= 4 ) || $this->mForceTocPosition );
3721
3722 # Allow user to stipulate that a page should have a "new section"
3723 # link added via __NEWSECTIONLINK__
3724 if ( isset( $this->mDoubleUnderscores['newsectionlink'] ) ) {
3725 $this->mOutput->setNewSection( true );
3726 }
3727
3728 # Allow user to remove the "new section"
3729 # link via __NONEWSECTIONLINK__
3730 if ( isset( $this->mDoubleUnderscores['nonewsectionlink'] ) ) {
3731 $this->mOutput->hideNewSection( true );
3732 }
3733
3734 # if the string __FORCETOC__ (not case-sensitive) occurs in the HTML,
3735 # override above conditions and always show TOC above first header
3736 if ( isset( $this->mDoubleUnderscores['forcetoc'] ) ) {
3737 $this->mShowToc = true;
3738 $enoughToc = true;
3739 }
3740
3741 # We need this to perform operations on the HTML
3742 $sk = $this->mOptions->getSkin( $this->mTitle );
3743
3744 # headline counter
3745 $headlineCount = 0;
3746 $numVisible = 0;
3747
3748 # Ugh .. the TOC should have neat indentation levels which can be
3749 # passed to the skin functions. These are determined here
3750 $toc = '';
3751 $full = '';
3752 $head = array();
3753 $sublevelCount = array();
3754 $levelCount = array();
3755 $level = 0;
3756 $prevlevel = 0;
3757 $toclevel = 0;
3758 $prevtoclevel = 0;
3759 $markerRegex = "{$this->mUniqPrefix}-h-(\d+)-" . self::MARKER_SUFFIX;
3760 $baseTitleText = $this->mTitle->getPrefixedDBkey();
3761 $oldType = $this->mOutputType;
3762 $this->setOutputType( self::OT_WIKI );
3763 $frame = $this->getPreprocessor()->newFrame();
3764 $root = $this->preprocessToDom( $origText );
3765 $node = $root->getFirstChild();
3766 $byteOffset = 0;
3767 $tocraw = array();
3768 $refers = array();
3769
3770 foreach ( $matches[3] as $headline ) {
3771 $isTemplate = false;
3772 $titleText = false;
3773 $sectionIndex = false;
3774 $numbering = '';
3775 $markerMatches = array();
3776 if ( preg_match("/^$markerRegex/", $headline, $markerMatches ) ) {
3777 $serial = $markerMatches[1];
3778 list( $titleText, $sectionIndex ) = $this->mHeadings[$serial];
3779 $isTemplate = ( $titleText != $baseTitleText );
3780 $headline = preg_replace( "/^$markerRegex/", "", $headline );
3781 }
3782
3783 if ( $toclevel ) {
3784 $prevlevel = $level;
3785 }
3786 $level = $matches[1][$headlineCount];
3787
3788 if ( $level > $prevlevel ) {
3789 # Increase TOC level
3790 $toclevel++;
3791 $sublevelCount[$toclevel] = 0;
3792 if ( $toclevel<$wgMaxTocLevel ) {
3793 $prevtoclevel = $toclevel;
3794 $toc .= $sk->tocIndent();
3795 $numVisible++;
3796 }
3797 } elseif ( $level < $prevlevel && $toclevel > 1 ) {
3798 # Decrease TOC level, find level to jump to
3799
3800 for ( $i = $toclevel; $i > 0; $i-- ) {
3801 if ( $levelCount[$i] == $level ) {
3802 # Found last matching level
3803 $toclevel = $i;
3804 break;
3805 } elseif ( $levelCount[$i] < $level ) {
3806 # Found first matching level below current level
3807 $toclevel = $i + 1;
3808 break;
3809 }
3810 }
3811 if ( $i == 0 ) {
3812 $toclevel = 1;
3813 }
3814 if ( $toclevel<$wgMaxTocLevel ) {
3815 if ( $prevtoclevel < $wgMaxTocLevel ) {
3816 # Unindent only if the previous toc level was shown :p
3817 $toc .= $sk->tocUnindent( $prevtoclevel - $toclevel );
3818 $prevtoclevel = $toclevel;
3819 } else {
3820 $toc .= $sk->tocLineEnd();
3821 }
3822 }
3823 } else {
3824 # No change in level, end TOC line
3825 if ( $toclevel<$wgMaxTocLevel ) {
3826 $toc .= $sk->tocLineEnd();
3827 }
3828 }
3829
3830 $levelCount[$toclevel] = $level;
3831
3832 # count number of headlines for each level
3833 @$sublevelCount[$toclevel]++;
3834 $dot = 0;
3835 for( $i = 1; $i <= $toclevel; $i++ ) {
3836 if ( !empty( $sublevelCount[$i] ) ) {
3837 if ( $dot ) {
3838 $numbering .= '.';
3839 }
3840 $numbering .= $wgContLang->formatNum( $sublevelCount[$i] );
3841 $dot = 1;
3842 }
3843 }
3844
3845 # The safe header is a version of the header text safe to use for links
3846 # Avoid insertion of weird stuff like <math> by expanding the relevant sections
3847 $safeHeadline = $this->mStripState->unstripBoth( $headline );
3848
3849 # Remove link placeholders by the link text.
3850 # <!--LINK number-->
3851 # turns into
3852 # link text with suffix
3853 $safeHeadline = $this->replaceLinkHoldersText( $safeHeadline );
3854
3855 # Strip out HTML (other than plain <sup> and <sub>: bug 8393)
3856 $tocline = preg_replace(
3857 array( '#<(?!/?(sup|sub)).*?'.'>#', '#<(/?(sup|sub)).*?'.'>#' ),
3858 array( '', '<$1>' ),
3859 $safeHeadline
3860 );
3861 $tocline = trim( $tocline );
3862
3863 # For the anchor, strip out HTML-y stuff period
3864 $safeHeadline = preg_replace( '/<.*?'.'>/', '', $safeHeadline );
3865 $safeHeadline = Sanitizer::normalizeSectionNameWhitespace( $safeHeadline );
3866
3867 # Save headline for section edit hint before it's escaped
3868 $headlineHint = $safeHeadline;
3869
3870 if ( $wgHtml5 && $wgExperimentalHtmlIds ) {
3871 # For reverse compatibility, provide an id that's
3872 # HTML4-compatible, like we used to.
3873 #
3874 # It may be worth noting, academically, that it's possible for
3875 # the legacy anchor to conflict with a non-legacy headline
3876 # anchor on the page. In this case likely the "correct" thing
3877 # would be to either drop the legacy anchors or make sure
3878 # they're numbered first. However, this would require people
3879 # to type in section names like "abc_.D7.93.D7.90.D7.A4"
3880 # manually, so let's not bother worrying about it.
3881 $legacyHeadline = Sanitizer::escapeId( $safeHeadline,
3882 array( 'noninitial', 'legacy' ) );
3883 $safeHeadline = Sanitizer::escapeId( $safeHeadline );
3884
3885 if ( $legacyHeadline == $safeHeadline ) {
3886 # No reason to have both (in fact, we can't)
3887 $legacyHeadline = false;
3888 }
3889 } else {
3890 $legacyHeadline = false;
3891 $safeHeadline = Sanitizer::escapeId( $safeHeadline,
3892 'noninitial' );
3893 }
3894
3895 # HTML names must be case-insensitively unique (bug 10721).
3896 # This does not apply to Unicode characters per
3897 # http://dev.w3.org/html5/spec/infrastructure.html#case-sensitivity-and-string-comparison
3898 # FIXME: We may be changing them depending on the current locale.
3899 $arrayKey = strtolower( $safeHeadline );
3900 if ( $legacyHeadline === false ) {
3901 $legacyArrayKey = false;
3902 } else {
3903 $legacyArrayKey = strtolower( $legacyHeadline );
3904 }
3905
3906 # count how many in assoc. array so we can track dupes in anchors
3907 if ( isset( $refers[$arrayKey] ) ) {
3908 $refers[$arrayKey]++;
3909 } else {
3910 $refers[$arrayKey] = 1;
3911 }
3912 if ( isset( $refers[$legacyArrayKey] ) ) {
3913 $refers[$legacyArrayKey]++;
3914 } else {
3915 $refers[$legacyArrayKey] = 1;
3916 }
3917
3918 # Don't number the heading if it is the only one (looks silly)
3919 if ( $doNumberHeadings && count( $matches[3] ) > 1) {
3920 # the two are different if the line contains a link
3921 $headline = $numbering . ' ' . $headline;
3922 }
3923
3924 # Create the anchor for linking from the TOC to the section
3925 $anchor = $safeHeadline;
3926 $legacyAnchor = $legacyHeadline;
3927 if ( $refers[$arrayKey] > 1 ) {
3928 $anchor .= '_' . $refers[$arrayKey];
3929 }
3930 if ( $legacyHeadline !== false && $refers[$legacyArrayKey] > 1 ) {
3931 $legacyAnchor .= '_' . $refers[$legacyArrayKey];
3932 }
3933 if ( $enoughToc && ( !isset( $wgMaxTocLevel ) || $toclevel < $wgMaxTocLevel ) ) {
3934 $toc .= $sk->tocLine( $anchor, $tocline,
3935 $numbering, $toclevel, ( $isTemplate ? false : $sectionIndex ) );
3936 }
3937
3938 # Add the section to the section tree
3939 # Find the DOM node for this header
3940 while ( $node && !$isTemplate ) {
3941 if ( $node->getName() === 'h' ) {
3942 $bits = $node->splitHeading();
3943 if ( $bits['i'] == $sectionIndex ) {
3944 break;
3945 }
3946 }
3947 $byteOffset += mb_strlen( $this->mStripState->unstripBoth(
3948 $frame->expand( $node, PPFrame::RECOVER_ORIG ) ) );
3949 $node = $node->getNextSibling();
3950 }
3951 $tocraw[] = array(
3952 'toclevel' => $toclevel,
3953 'level' => $level,
3954 'line' => $tocline,
3955 'number' => $numbering,
3956 'index' => ( $isTemplate ? 'T-' : '' ) . $sectionIndex,
3957 'fromtitle' => $titleText,
3958 'byteoffset' => ( $isTemplate ? null : $byteOffset ),
3959 'anchor' => $anchor,
3960 );
3961
3962 # give headline the correct <h#> tag
3963 if ( $showEditLink && $sectionIndex !== false ) {
3964 if ( $isTemplate ) {
3965 # Put a T flag in the section identifier, to indicate to extractSections()
3966 # that sections inside <includeonly> should be counted.
3967 $editlink = $sk->doEditSectionLink( Title::newFromText( $titleText ), "T-$sectionIndex", null, $this->mOptions->getUserLang() );
3968 } else {
3969 $editlink = $sk->doEditSectionLink( $this->mTitle, $sectionIndex, $headlineHint, $this->mOptions->getUserLang() );
3970 }
3971 } else {
3972 $editlink = '';
3973 }
3974 $head[$headlineCount] = $sk->makeHeadline( $level,
3975 $matches['attrib'][$headlineCount], $anchor, $headline,
3976 $editlink, $legacyAnchor );
3977
3978 $headlineCount++;
3979 }
3980
3981 $this->setOutputType( $oldType );
3982
3983 # Never ever show TOC if no headers
3984 if ( $numVisible < 1 ) {
3985 $enoughToc = false;
3986 }
3987
3988 if ( $enoughToc ) {
3989 if ( $prevtoclevel > 0 && $prevtoclevel < $wgMaxTocLevel ) {
3990 $toc .= $sk->tocUnindent( $prevtoclevel - 1 );
3991 }
3992 $toc = $sk->tocList( $toc );
3993 $this->mOutput->setTOCHTML( $toc );
3994 }
3995
3996 if ( $isMain ) {
3997 $this->mOutput->setSections( $tocraw );
3998 }
3999
4000 # split up and insert constructed headlines
4001
4002 $blocks = preg_split( '/<H[1-6].*?' . '>.*?<\/H[1-6]>/i', $text );
4003 $i = 0;
4004
4005 foreach ( $blocks as $block ) {
4006 if ( $showEditLink && $headlineCount > 0 && $i == 0 && $block !== "\n" ) {
4007 # This is the [edit] link that appears for the top block of text when
4008 # section editing is enabled
4009
4010 # Disabled because it broke block formatting
4011 # For example, a bullet point in the top line
4012 # $full .= $sk->editSectionLink(0);
4013 }
4014 $full .= $block;
4015 if ( $enoughToc && !$i && $isMain && !$this->mForceTocPosition ) {
4016 # Top anchor now in skin
4017 $full = $full.$toc;
4018 }
4019
4020 if ( !empty( $head[$i] ) ) {
4021 $full .= $head[$i];
4022 }
4023 $i++;
4024 }
4025 if ( $this->mForceTocPosition ) {
4026 return str_replace( '<!--MWTOC-->', $toc, $full );
4027 } else {
4028 return $full;
4029 }
4030 }
4031
4032 /**
4033 * Transform wiki markup when saving a page by doing \r\n -> \n
4034 * conversion, substitting signatures, {{subst:}} templates, etc.
4035 *
4036 * @param $text String: the text to transform
4037 * @param $title Title: the Title object for the current article
4038 * @param $user User: the User object describing the current user
4039 * @param $options ParserOptions: parsing options
4040 * @param $clearState Boolean: whether to clear the parser state first
4041 * @return String: the altered wiki markup
4042 */
4043 public function preSaveTransform( $text, Title $title, $user, $options, $clearState = true ) {
4044 $options->resetUsage();
4045 $this->mOptions = $options;
4046 $this->setTitle( $title );
4047 $this->setOutputType( self::OT_WIKI );
4048
4049 if ( $clearState ) {
4050 $this->clearState();
4051 }
4052
4053 $pairs = array(
4054 "\r\n" => "\n",
4055 );
4056 $text = str_replace( array_keys( $pairs ), array_values( $pairs ), $text );
4057 $text = $this->pstPass2( $text, $user );
4058 $text = $this->mStripState->unstripBoth( $text );
4059 return $text;
4060 }
4061
4062 /**
4063 * Pre-save transform helper function
4064 * @private
4065 */
4066 function pstPass2( $text, $user ) {
4067 global $wgContLang, $wgLocaltimezone;
4068
4069 # Note: This is the timestamp saved as hardcoded wikitext to
4070 # the database, we use $wgContLang here in order to give
4071 # everyone the same signature and use the default one rather
4072 # than the one selected in each user's preferences.
4073 # (see also bug 12815)
4074 $ts = $this->mOptions->getTimestamp();
4075 if ( isset( $wgLocaltimezone ) ) {
4076 $tz = $wgLocaltimezone;
4077 } else {
4078 $tz = date_default_timezone_get();
4079 }
4080
4081 $unixts = wfTimestamp( TS_UNIX, $ts );
4082 $oldtz = date_default_timezone_get();
4083 date_default_timezone_set( $tz );
4084 $ts = date( 'YmdHis', $unixts );
4085 $tzMsg = date( 'T', $unixts ); # might vary on DST changeover!
4086
4087 # Allow translation of timezones through wiki. date() can return
4088 # whatever crap the system uses, localised or not, so we cannot
4089 # ship premade translations.
4090 $key = 'timezone-' . strtolower( trim( $tzMsg ) );
4091 $value = wfMsgForContent( $key );
4092 if ( !wfEmptyMsg( $key, $value ) ) {
4093 $tzMsg = $value;
4094 }
4095
4096 date_default_timezone_set( $oldtz );
4097
4098 $d = $wgContLang->timeanddate( $ts, false, false ) . " ($tzMsg)";
4099
4100 # Variable replacement
4101 # Because mOutputType is OT_WIKI, this will only process {{subst:xxx}} type tags
4102 $text = $this->replaceVariables( $text );
4103
4104 # Signatures
4105 $sigText = $this->getUserSig( $user );
4106 $text = strtr( $text, array(
4107 '~~~~~' => $d,
4108 '~~~~' => "$sigText $d",
4109 '~~~' => $sigText
4110 ) );
4111
4112 # Context links: [[|name]] and [[name (context)|]]
4113 global $wgLegalTitleChars;
4114 $tc = "[$wgLegalTitleChars]";
4115 $nc = '[ _0-9A-Za-z\x80-\xff-]'; # Namespaces can use non-ascii!
4116
4117 $p1 = "/\[\[(:?$nc+:|:|)($tc+?)( \\($tc+\\))\\|]]/"; # [[ns:page (context)|]]
4118 $p4 = "/\[\[(:?$nc+:|:|)($tc+?)(（$tc+）)\\|]]/"; # [[ns:page（context）|]]
4119 $p3 = "/\[\[(:?$nc+:|:|)($tc+?)( \\($tc+\\)|)(, $tc+|)\\|]]/"; # [[ns:page (context), context|]]
4120 $p2 = "/\[\[\\|($tc+)]]/"; # [[|page]]
4121
4122 # try $p1 first, to turn "[[A, B (C)|]]" into "[[A, B (C)|A, B]]"
4123 $text = preg_replace( $p1, '[[\\1\\2\\3|\\2]]', $text );
4124 $text = preg_replace( $p4, '[[\\1\\2\\3|\\2]]', $text );
4125 $text = preg_replace( $p3, '[[\\1\\2\\3\\4|\\2]]', $text );
4126
4127 $t = $this->mTitle->getText();
4128 $m = array();
4129 if ( preg_match( "/^($nc+:|)$tc+?( \\($tc+\\))$/", $t, $m ) ) {
4130 $text = preg_replace( $p2, "[[$m[1]\\1$m[2]|\\1]]", $text );
4131 } elseif ( preg_match( "/^($nc+:|)$tc+?(, $tc+|)$/", $t, $m ) && "$m[1]$m[2]" != '' ) {
4132 $text = preg_replace( $p2, "[[$m[1]\\1$m[2]|\\1]]", $text );
4133 } else {
4134 # if there's no context, don't bother duplicating the title
4135 $text = preg_replace( $p2, '[[\\1]]', $text );
4136 }
4137
4138 # Trim trailing whitespace
4139 $text = rtrim( $text );
4140
4141 return $text;
4142 }
4143
4144 /**
4145 * Fetch the user's signature text, if any, and normalize to
4146 * validated, ready-to-insert wikitext.
4147 * If you have pre-fetched the nickname or the fancySig option, you can
4148 * specify them here to save a database query.
4149 *
4150 * @param $user User
4151 * @param $nickname String: nickname to use or false to use user's default nickname
4152 * @param $fancySig Boolean: whether the nicknname is the complete signature
4153 * or null to use default value
4154 * @return string
4155 */
4156 function getUserSig( &$user, $nickname = false, $fancySig = null ) {
4157 global $wgMaxSigChars;
4158
4159 $username = $user->getName();
4160
4161 # If not given, retrieve from the user object.
4162 if ( $nickname === false )
4163 $nickname = $user->getOption( 'nickname' );
4164
4165 if ( is_null( $fancySig ) ) {
4166 $fancySig = $user->getBoolOption( 'fancysig' );
4167 }
4168
4169 $nickname = $nickname == null ? $username : $nickname;
4170
4171 if ( mb_strlen( $nickname ) > $wgMaxSigChars ) {
4172 $nickname = $username;
4173 wfDebug( __METHOD__ . ": $username has overlong signature.\n" );
4174 } elseif ( $fancySig !== false ) {
4175 # Sig. might contain markup; validate this
4176 if ( $this->validateSig( $nickname ) !== false ) {
4177 # Validated; clean up (if needed) and return it
4178 return $this->cleanSig( $nickname, true );
4179 } else {
4180 # Failed to validate; fall back to the default
4181 $nickname = $username;
4182 wfDebug( __METHOD__.": $username has bad XML tags in signature.\n" );
4183 }
4184 }
4185
4186 # Make sure nickname doesnt get a sig in a sig
4187 $nickname = $this->cleanSigInSig( $nickname );
4188
4189 # If we're still here, make it a link to the user page
4190 $userText = wfEscapeWikiText( $username );
4191 $nickText = wfEscapeWikiText( $nickname );
4192 if ( $user->isAnon() ) {
4193 return wfMsgExt( 'signature-anon', array( 'content', 'parsemag' ), $userText, $nickText );
4194 } else {
4195 return wfMsgExt( 'signature', array( 'content', 'parsemag' ), $userText, $nickText );
4196 }
4197 }
4198
4199 /**
4200 * Check that the user's signature contains no bad XML
4201 *
4202 * @param $text String
4203 * @return mixed An expanded string, or false if invalid.
4204 */
4205 function validateSig( $text ) {
4206 return( Xml::isWellFormedXmlFragment( $text ) ? $text : false );
4207 }
4208
4209 /**
4210 * Clean up signature text
4211 *
4212 * 1) Strip ~~~, ~~~~ and ~~~~~ out of signatures @see cleanSigInSig
4213 * 2) Substitute all transclusions
4214 *
4215 * @param $text String
4216 * @param $parsing Whether we're cleaning (preferences save) or parsing
4217 * @return String: signature text
4218 */
4219 function cleanSig( $text, $parsing = false ) {
4220 if ( !$parsing ) {
4221 global $wgTitle;
4222 $this->clearState();
4223 $this->setTitle( $wgTitle );
4224 $this->mOptions = new ParserOptions;
4225 $this->setOutputType = self::OT_PREPROCESS;
4226 }
4227
4228 # Option to disable this feature
4229 if ( !$this->mOptions->getCleanSignatures() ) {
4230 return $text;
4231 }
4232
4233 # FIXME: regex doesn't respect extension tags or nowiki
4234 # => Move this logic to braceSubstitution()
4235 $substWord = MagicWord::get( 'subst' );
4236 $substRegex = '/\{\{(?!(?:' . $substWord->getBaseRegex() . '))/x' . $substWord->getRegexCase();
4237 $substText = '{{' . $substWord->getSynonym( 0 );
4238
4239 $text = preg_replace( $substRegex, $substText, $text );
4240 $text = $this->cleanSigInSig( $text );
4241 $dom = $this->preprocessToDom( $text );
4242 $frame = $this->getPreprocessor()->newFrame();
4243 $text = $frame->expand( $dom );
4244
4245 if ( !$parsing ) {
4246 $text = $this->mStripState->unstripBoth( $text );
4247 }
4248
4249 return $text;
4250 }
4251
4252 /**
4253 * Strip ~~~, ~~~~ and ~~~~~ out of signatures
4254 *
4255 * @param $text String
4256 * @return String: signature text with /~{3,5}/ removed
4257 */
4258 function cleanSigInSig( $text ) {
4259 $text = preg_replace( '/~{3,5}/', '', $text );
4260 return $text;
4261 }
4262
4263 /**
4264 * Set up some variables which are usually set up in parse()
4265 * so that an external function can call some class members with confidence
4266 */
4267 public function startExternalParse( &$title, $options, $outputType, $clearState = true ) {
4268 $this->setTitle( $title );
4269 $options->resetUsage();
4270 $this->mOptions = $options;
4271 $this->setOutputType( $outputType );
4272 if ( $clearState ) {
4273 $this->clearState();
4274 }
4275 }
4276
4277 /**
4278 * Wrapper for preprocess()
4279 *
4280 * @param $text String: the text to preprocess
4281 * @param $options ParserOptions: options
4282 * @return String
4283 */
4284 public function transformMsg( $text, $options ) {
4285 global $wgTitle;
4286 static $executing = false;
4287
4288 # Guard against infinite recursion
4289 if ( $executing ) {
4290 return $text;
4291 }
4292 $executing = true;
4293
4294 wfProfileIn( __METHOD__ );
4295 $text = $this->preprocess( $text, $wgTitle, $options );
4296
4297 $executing = false;
4298 wfProfileOut( __METHOD__ );
4299 return $text;
4300 }
4301
4302 /**
4303 * Create an HTML-style tag, e.g. <yourtag>special text</yourtag>
4304 * The callback should have the following form:
4305 * function myParserHook( $text, $params, $parser ) { ... }
4306 *
4307 * Transform and return $text. Use $parser for any required context, e.g. use
4308 * $parser->getTitle() and $parser->getOptions() not $wgTitle or $wgOut->mParserOptions
4309 *
4310 * @param $tag Mixed: the tag to use, e.g. 'hook' for <hook>
4311 * @param $callback Mixed: the callback function (and object) to use for the tag
4312 * @return The old value of the mTagHooks array associated with the hook
4313 */
4314 public function setHook( $tag, $callback ) {
4315 $tag = strtolower( $tag );
4316 $oldVal = isset( $this->mTagHooks[$tag] ) ? $this->mTagHooks[$tag] : null;
4317 $this->mTagHooks[$tag] = $callback;
4318 if ( !in_array( $tag, $this->mStripList ) ) {
4319 $this->mStripList[] = $tag;
4320 }
4321
4322 return $oldVal;
4323 }
4324
4325 function setTransparentTagHook( $tag, $callback ) {
4326 $tag = strtolower( $tag );
4327 $oldVal = isset( $this->mTransparentTagHooks[$tag] ) ? $this->mTransparentTagHooks[$tag] : null;
4328 $this->mTransparentTagHooks[$tag] = $callback;
4329
4330 return $oldVal;
4331 }
4332
4333 /**
4334 * Remove all tag hooks
4335 */
4336 function clearTagHooks() {
4337 $this->mTagHooks = array();
4338 $this->mStripList = $this->mDefaultStripList;
4339 }
4340
4341 /**
4342 * Create a function, e.g. {{sum:1|2|3}}
4343 * The callback function should have the form:
4344 * function myParserFunction( &$parser, $arg1, $arg2, $arg3 ) { ... }
4345 *
4346 * Or with SFH_OBJECT_ARGS:
4347 * function myParserFunction( $parser, $frame, $args ) { ... }
4348 *
4349 * The callback may either return the text result of the function, or an array with the text
4350 * in element 0, and a number of flags in the other elements. The names of the flags are
4351 * specified in the keys. Valid flags are:
4352 * found The text returned is valid, stop processing the template. This
4353 * is on by default.
4354 * nowiki Wiki markup in the return value should be escaped
4355 * isHTML The returned text is HTML, armour it against wikitext transformation
4356 *
4357 * @param $id String: The magic word ID
4358 * @param $callback Mixed: the callback function (and object) to use
4359 * @param $flags Integer: a combination of the following flags:
4360 * SFH_NO_HASH No leading hash, i.e. {{plural:...}} instead of {{#if:...}}
4361 *
4362 * SFH_OBJECT_ARGS Pass the template arguments as PPNode objects instead of text. This
4363 * allows for conditional expansion of the parse tree, allowing you to eliminate dead
4364 * branches and thus speed up parsing. It is also possible to analyse the parse tree of
4365 * the arguments, and to control the way they are expanded.
4366 *
4367 * The $frame parameter is a PPFrame. This can be used to produce expanded text from the
4368 * arguments, for instance:
4369 * $text = isset( $args[0] ) ? $frame->expand( $args[0] ) : '';
4370 *
4371 * For technical reasons, $args[0] is pre-expanded and will be a string. This may change in
4372 * future versions. Please call $frame->expand() on it anyway so that your code keeps
4373 * working if/when this is changed.
4374 *
4375 * If you want whitespace to be trimmed from $args, you need to do it yourself, post-
4376 * expansion.
4377 *
4378 * Please read the documentation in includes/parser/Preprocessor.php for more information
4379 * about the methods available in PPFrame and PPNode.
4380 *
4381 * @return The old callback function for this name, if any
4382 */
4383 public function setFunctionHook( $id, $callback, $flags = 0 ) {
4384 global $wgContLang;
4385
4386 $oldVal = isset( $this->mFunctionHooks[$id] ) ? $this->mFunctionHooks[$id][0] : null;
4387 $this->mFunctionHooks[$id] = array( $callback, $flags );
4388
4389 # Add to function cache
4390 $mw = MagicWord::get( $id );
4391 if ( !$mw )
4392 throw new MWException( __METHOD__.'() expecting a magic word identifier.' );
4393
4394 $synonyms = $mw->getSynonyms();
4395 $sensitive = intval( $mw->isCaseSensitive() );
4396
4397 foreach ( $synonyms as $syn ) {
4398 # Case
4399 if ( !$sensitive ) {
4400 $syn = $wgContLang->lc( $syn );
4401 }
4402 # Add leading hash
4403 if ( !( $flags & SFH_NO_HASH ) ) {
4404 $syn = '#' . $syn;
4405 }
4406 # Remove trailing colon
4407 if ( substr( $syn, -1, 1 ) === ':' ) {
4408 $syn = substr( $syn, 0, -1 );
4409 }
4410 $this->mFunctionSynonyms[$sensitive][$syn] = $id;
4411 }
4412 return $oldVal;
4413 }
4414
4415 /**
4416 * Get all registered function hook identifiers
4417 *
4418 * @return Array
4419 */
4420 function getFunctionHooks() {
4421 return array_keys( $this->mFunctionHooks );
4422 }
4423
4424 /**
4425 * Create a tag function, e.g. <test>some stuff</test>.
4426 * Unlike tag hooks, tag functions are parsed at preprocessor level.
4427 * Unlike parser functions, their content is not preprocessed.
4428 */
4429 function setFunctionTagHook( $tag, $callback, $flags ) {
4430 $tag = strtolower( $tag );
4431 $old = isset( $this->mFunctionTagHooks[$tag] ) ?
4432 $this->mFunctionTagHooks[$tag] : null;
4433 $this->mFunctionTagHooks[$tag] = array( $callback, $flags );
4434
4435 if ( !in_array( $tag, $this->mStripList ) ) {
4436 $this->mStripList[] = $tag;
4437 }
4438
4439 return $old;
4440 }
4441
4442 /**
4443 * FIXME: update documentation. makeLinkObj() is deprecated.
4444 * Replace <!--LINK--> link placeholders with actual links, in the buffer
4445 * Placeholders created in Skin::makeLinkObj()
4446 * Returns an array of link CSS classes, indexed by PDBK.
4447 */
4448 function replaceLinkHolders( &$text, $options = 0 ) {
4449 return $this->mLinkHolders->replace( $text );
4450 }
4451
4452 /**
4453 * Replace <!--LINK--> link placeholders with plain text of links
4454 * (not HTML-formatted).
4455 *
4456 * @param $text String
4457 * @return String
4458 */
4459 function replaceLinkHoldersText( $text ) {
4460 return $this->mLinkHolders->replaceText( $text );
4461 }
4462
4463 /**
4464 * Renders an image gallery from a text with one line per image.
4465 * text labels may be given by using |-style alternative text. E.g.
4466 * Image:one.jpg|The number "1"
4467 * Image:tree.jpg|A tree
4468 * given as text will return the HTML of a gallery with two images,
4469 * labeled 'The number "1"' and
4470 * 'A tree'.
4471 */
4472 function renderImageGallery( $text, $params ) {
4473 $ig = new ImageGallery();
4474 $ig->setContextTitle( $this->mTitle );
4475 $ig->setShowBytes( false );
4476 $ig->setShowFilename( false );
4477 $ig->setParser( $this );
4478 $ig->setHideBadImages();
4479 $ig->setAttributes( Sanitizer::validateTagAttributes( $params, 'table' ) );
4480 $ig->useSkin( $this->mOptions->getSkin( $this->mTitle ) );
4481 $ig->mRevisionId = $this->mRevisionId;
4482
4483 if ( isset( $params['showfilename'] ) ) {
4484 $ig->setShowFilename( true );
4485 } else {
4486 $ig->setShowFilename( false );
4487 }
4488 if ( isset( $params['caption'] ) ) {
4489 $caption = $params['caption'];
4490 $caption = htmlspecialchars( $caption );
4491 $caption = $this->replaceInternalLinks( $caption );
4492 $ig->setCaptionHtml( $caption );
4493 }
4494 if ( isset( $params['perrow'] ) ) {
4495 $ig->setPerRow( $params['perrow'] );
4496 }
4497 if ( isset( $params['widths'] ) ) {
4498 $ig->setWidths( $params['widths'] );
4499 }
4500 if ( isset( $params['heights'] ) ) {
4501 $ig->setHeights( $params['heights'] );
4502 }
4503
4504 wfRunHooks( 'BeforeParserrenderImageGallery', array( &$this, &$ig ) );
4505
4506 $lines = StringUtils::explode( "\n", $text );
4507 foreach ( $lines as $line ) {
4508 # match lines like these:
4509 # Image:someimage.jpg|This is some image
4510 $matches = array();
4511 preg_match( "/^([^|]+)(\\|(.*))?$/", $line, $matches );
4512 # Skip empty lines
4513 if ( count( $matches ) == 0 ) {
4514 continue;
4515 }
4516
4517 if ( strpos( $matches[0], '%' ) !== false ) {
4518 $matches[1] = urldecode( $matches[1] );
4519 }
4520 $tp = Title::newFromText( $matches[1] );
4521 $nt =& $tp;
4522 if ( is_null( $nt ) ) {
4523 # Bogus title. Ignore these so we don't bomb out later.
4524 continue;
4525 }
4526 if ( isset( $matches[3] ) ) {
4527 $label = $matches[3];
4528 } else {
4529 $label = '';
4530 }
4531
4532 $html = $this->recursiveTagParse( trim( $label ) );
4533
4534 $ig->add( $nt, $html );
4535
4536 # Only add real images (bug #5586)
4537 if ( $nt->getNamespace() == NS_FILE ) {
4538 $this->mOutput->addImage( $nt->getDBkey() );
4539 }
4540 }
4541 return $ig->toHTML();
4542 }
4543
4544 function getImageParams( $handler ) {
4545 if ( $handler ) {
4546 $handlerClass = get_class( $handler );
4547 } else {
4548 $handlerClass = '';
4549 }
4550 if ( !isset( $this->mImageParams[$handlerClass] ) ) {
4551 # Initialise static lists
4552 static $internalParamNames = array(
4553 'horizAlign' => array( 'left', 'right', 'center', 'none' ),
4554 'vertAlign' => array( 'baseline', 'sub', 'super', 'top', 'text-top', 'middle',
4555 'bottom', 'text-bottom' ),
4556 'frame' => array( 'thumbnail', 'manualthumb', 'framed', 'frameless',
4557 'upright', 'border', 'link', 'alt' ),
4558 );
4559 static $internalParamMap;
4560 if ( !$internalParamMap ) {
4561 $internalParamMap = array();
4562 foreach ( $internalParamNames as $type => $names ) {
4563 foreach ( $names as $name ) {
4564 $magicName = str_replace( '-', '_', "img_$name" );
4565 $internalParamMap[$magicName] = array( $type, $name );
4566 }
4567 }
4568 }
4569
4570 # Add handler params
4571 $paramMap = $internalParamMap;
4572 if ( $handler ) {
4573 $handlerParamMap = $handler->getParamMap();
4574 foreach ( $handlerParamMap as $magic => $paramName ) {
4575 $paramMap[$magic] = array( 'handler', $paramName );
4576 }
4577 }
4578 $this->mImageParams[$handlerClass] = $paramMap;
4579 $this->mImageParamsMagicArray[$handlerClass] = new MagicWordArray( array_keys( $paramMap ) );
4580 }
4581 return array( $this->mImageParams[$handlerClass], $this->mImageParamsMagicArray[$handlerClass] );
4582 }
4583
4584 /**
4585 * Parse image options text and use it to make an image
4586 *
4587 * @param $title Title
4588 * @param $options String
4589 * @param $holders LinkHolderArray
4590 */
4591 function makeImage( $title, $options, $holders = false ) {
4592 # Check if the options text is of the form "options|alt text"
4593 # Options are:
4594 # * thumbnail make a thumbnail with enlarge-icon and caption, alignment depends on lang
4595 # * left no resizing, just left align. label is used for alt= only
4596 # * right same, but right aligned
4597 # * none same, but not aligned
4598 # * ___px scale to ___ pixels width, no aligning. e.g. use in taxobox
4599 # * center center the image
4600 # * frame Keep original image size, no magnify-button.
4601 # * framed Same as "frame"
4602 # * frameless like 'thumb' but without a frame. Keeps user preferences for width
4603 # * upright reduce width for upright images, rounded to full __0 px
4604 # * border draw a 1px border around the image
4605 # * alt Text for HTML alt attribute (defaults to empty)
4606 # * link Set the target of the image link. Can be external, interwiki, or local
4607 # vertical-align values (no % or length right now):
4608 # * baseline
4609 # * sub
4610 # * super
4611 # * top
4612 # * text-top
4613 # * middle
4614 # * bottom
4615 # * text-bottom
4616
4617 $parts = StringUtils::explode( "|", $options );
4618 $sk = $this->mOptions->getSkin( $this->mTitle );
4619
4620 # Give extensions a chance to select the file revision for us
4621 $skip = $time = $descQuery = false;
4622 wfRunHooks( 'BeforeParserMakeImageLinkObj', array( &$this, &$title, &$skip, &$time, &$descQuery ) );
4623
4624 if ( $skip ) {
4625 return $sk->link( $title );
4626 }
4627
4628 # Get the file
4629 $file = wfFindFile( $title, array( 'time' => $time ) );
4630 # Get parameter map
4631 $handler = $file ? $file->getHandler() : false;
4632
4633 list( $paramMap, $mwArray ) = $this->getImageParams( $handler );
4634
4635 # Process the input parameters
4636 $caption = '';
4637 $params = array( 'frame' => array(), 'handler' => array(),
4638 'horizAlign' => array(), 'vertAlign' => array() );
4639 foreach ( $parts as $part ) {
4640 $part = trim( $part );
4641 list( $magicName, $value ) = $mwArray->matchVariableStartToEnd( $part );
4642 $validated = false;
4643 if ( isset( $paramMap[$magicName] ) ) {
4644 list( $type, $paramName ) = $paramMap[$magicName];
4645
4646 # Special case; width and height come in one variable together
4647 if ( $type === 'handler' && $paramName === 'width' ) {
4648 $m = array();
4649 # (bug 13500) In both cases (width/height and width only),
4650 # permit trailing "px" for backward compatibility.
4651 if ( preg_match( '/^([0-9]*)x([0-9]*)\s*(?:px)?\s*$/', $value, $m ) ) {
4652 $width = intval( $m[1] );
4653 $height = intval( $m[2] );
4654 if ( $handler->validateParam( 'width', $width ) ) {
4655 $params[$type]['width'] = $width;
4656 $validated = true;
4657 }
4658 if ( $handler->validateParam( 'height', $height ) ) {
4659 $params[$type]['height'] = $height;
4660 $validated = true;
4661 }
4662 } elseif ( preg_match( '/^[0-9]*\s*(?:px)?\s*$/', $value ) ) {
4663 $width = intval( $value );
4664 if ( $handler->validateParam( 'width', $width ) ) {
4665 $params[$type]['width'] = $width;
4666 $validated = true;
4667 }
4668 } # else no validation -- bug 13436
4669 } else {
4670 if ( $type === 'handler' ) {
4671 # Validate handler parameter
4672 $validated = $handler->validateParam( $paramName, $value );
4673 } else {
4674 # Validate internal parameters
4675 switch( $paramName ) {
4676 case 'manualthumb':
4677 case 'alt':
4678 # @todo Fixme: possibly check validity here for
4679 # manualthumb? downstream behavior seems odd with
4680 # missing manual thumbs.
4681 $validated = true;
4682 $value = $this->stripAltText( $value, $holders );
4683 break;
4684 case 'link':
4685 $chars = self::EXT_LINK_URL_CLASS;
4686 $prots = $this->mUrlProtocols;
4687 if ( $value === '' ) {
4688 $paramName = 'no-link';
4689 $value = true;
4690 $validated = true;
4691 } elseif ( preg_match( "/^$prots/", $value ) ) {
4692 if ( preg_match( "/^($prots)$chars+$/", $value, $m ) ) {
4693 $paramName = 'link-url';
4694 $this->mOutput->addExternalLink( $value );
4695 if ( $this->mOptions->getExternalLinkTarget() ) {
4696 $params[$type]['link-target'] = $this->mOptions->getExternalLinkTarget();
4697 }
4698 $validated = true;
4699 }
4700 } else {
4701 $linkTitle = Title::newFromText( $value );
4702 if ( $linkTitle ) {
4703 $paramName = 'link-title';
4704 $value = $linkTitle;
4705 $this->mOutput->addLink( $linkTitle );
4706 $validated = true;
4707 }
4708 }
4709 break;
4710 default:
4711 # Most other things appear to be empty or numeric...
4712 $validated = ( $value === false || is_numeric( trim( $value ) ) );
4713 }
4714 }
4715
4716 if ( $validated ) {
4717 $params[$type][$paramName] = $value;
4718 }
4719 }
4720 }
4721 if ( !$validated ) {
4722 $caption = $part;
4723 }
4724 }
4725
4726 # Process alignment parameters
4727 if ( $params['horizAlign'] ) {
4728 $params['frame']['align'] = key( $params['horizAlign'] );
4729 }
4730 if ( $params['vertAlign'] ) {
4731 $params['frame']['valign'] = key( $params['vertAlign'] );
4732 }
4733
4734 $params['frame']['caption'] = $caption;
4735
4736 # Will the image be presented in a frame, with the caption below?
4737 $imageIsFramed = isset( $params['frame']['frame'] ) ||
4738 isset( $params['frame']['framed'] ) ||
4739 isset( $params['frame']['thumbnail'] ) ||
4740 isset( $params['frame']['manualthumb'] );
4741
4742 # In the old days, [[Image:Foo|text...]] would set alt text. Later it
4743 # came to also set the caption, ordinary text after the image -- which
4744 # makes no sense, because that just repeats the text multiple times in
4745 # screen readers. It *also* came to set the title attribute.
4746 #
4747 # Now that we have an alt attribute, we should not set the alt text to
4748 # equal the caption: that's worse than useless, it just repeats the
4749 # text. This is the framed/thumbnail case. If there's no caption, we
4750 # use the unnamed parameter for alt text as well, just for the time be-
4751 # ing, if the unnamed param is set and the alt param is not.
4752 #
4753 # For the future, we need to figure out if we want to tweak this more,
4754 # e.g., introducing a title= parameter for the title; ignoring the un-
4755 # named parameter entirely for images without a caption; adding an ex-
4756 # plicit caption= parameter and preserving the old magic unnamed para-
4757 # meter for BC; ...
4758 if ( $imageIsFramed ) { # Framed image
4759 if ( $caption === '' && !isset( $params['frame']['alt'] ) ) {
4760 # No caption or alt text, add the filename as the alt text so
4761 # that screen readers at least get some description of the image
4762 $params['frame']['alt'] = $title->getText();
4763 }
4764 # Do not set $params['frame']['title'] because tooltips don't make sense
4765 # for framed images
4766 } else { # Inline image
4767 if ( !isset( $params['frame']['alt'] ) ) {
4768 # No alt text, use the "caption" for the alt text
4769 if ( $caption !== '') {
4770 $params['frame']['alt'] = $this->stripAltText( $caption, $holders );
4771 } else {
4772 # No caption, fall back to using the filename for the
4773 # alt text
4774 $params['frame']['alt'] = $title->getText();
4775 }
4776 }
4777 # Use the "caption" for the tooltip text
4778 $params['frame']['title'] = $this->stripAltText( $caption, $holders );
4779 }
4780
4781 wfRunHooks( 'ParserMakeImageParams', array( $title, $file, &$params ) );
4782
4783 # Linker does the rest
4784 $ret = $sk->makeImageLink2( $title, $file, $params['frame'], $params['handler'], $time, $descQuery, $this->mOptions->getThumbSize() );
4785
4786 # Give the handler a chance to modify the parser object
4787 if ( $handler ) {
4788 $handler->parserTransformHook( $this, $file );
4789 }
4790
4791 return $ret;
4792 }
4793
4794 protected function stripAltText( $caption, $holders ) {
4795 # Strip bad stuff out of the title (tooltip). We can't just use
4796 # replaceLinkHoldersText() here, because if this function is called
4797 # from replaceInternalLinks2(), mLinkHolders won't be up-to-date.
4798 if ( $holders ) {
4799 $tooltip = $holders->replaceText( $caption );
4800 } else {
4801 $tooltip = $this->replaceLinkHoldersText( $caption );
4802 }
4803
4804 # make sure there are no placeholders in thumbnail attributes
4805 # that are later expanded to html- so expand them now and
4806 # remove the tags
4807 $tooltip = $this->mStripState->unstripBoth( $tooltip );
4808 $tooltip = Sanitizer::stripAllTags( $tooltip );
4809
4810 return $tooltip;
4811 }
4812
4813 /**
4814 * Set a flag in the output object indicating that the content is dynamic and
4815 * shouldn't be cached.
4816 */
4817 function disableCache() {
4818 wfDebug( "Parser output marked as uncacheable.\n" );
4819 $this->mOutput->setCacheTime( -1 ); // old style, for compatibility
4820 $this->mOutput->updateCacheExpiry( 0 ); // new style, for consistency
4821 }
4822
4823 /**
4824 * Callback from the Sanitizer for expanding items found in HTML attribute
4825 * values, so they can be safely tested and escaped.
4826 *
4827 * @param $text String
4828 * @param $frame PPFrame
4829 * @return String
4830 * @private
4831 */
4832 function attributeStripCallback( &$text, $frame = false ) {
4833 $text = $this->replaceVariables( $text, $frame );
4834 $text = $this->mStripState->unstripBoth( $text );
4835 return $text;
4836 }
4837
4838 /**
4839 * Accessor
4840 */
4841 function getTags() {
4842 return array_merge( array_keys( $this->mTransparentTagHooks ), array_keys( $this->mTagHooks ) );
4843 }
4844
4845 /**
4846 * Break wikitext input into sections, and either pull or replace
4847 * some particular section's text.
4848 *
4849 * External callers should use the getSection and replaceSection methods.
4850 *
4851 * @param $text String: Page wikitext
4852 * @param $section String: a section identifier string of the form:
4853 * <flag1> - <flag2> - ... - <section number>
4854 *
4855 * Currently the only recognised flag is "T", which means the target section number
4856 * was derived during a template inclusion parse, in other words this is a template
4857 * section edit link. If no flags are given, it was an ordinary section edit link.
4858 * This flag is required to avoid a section numbering mismatch when a section is
4859 * enclosed by <includeonly> (bug 6563).
4860 *
4861 * The section number 0 pulls the text before the first heading; other numbers will
4862 * pull the given section along with its lower-level subsections. If the section is
4863 * not found, $mode=get will return $newtext, and $mode=replace will return $text.
4864 *
4865 * @param $mode String: one of "get" or "replace"
4866 * @param $newText String: replacement text for section data.
4867 * @return String: for "get", the extracted section text.
4868 * for "replace", the whole page with the section replaced.
4869 */
4870 private function extractSections( $text, $section, $mode, $newText='' ) {
4871 global $wgTitle;
4872 $this->clearState();
4873 $this->setTitle( $wgTitle ); # not generally used but removes an ugly failure mode
4874 $this->mOptions = new ParserOptions;
4875 $this->setOutputType( self::OT_PLAIN );
4876 $outText = '';
4877 $frame = $this->getPreprocessor()->newFrame();
4878
4879 # Process section extraction flags
4880 $flags = 0;
4881 $sectionParts = explode( '-', $section );
4882 $sectionIndex = array_pop( $sectionParts );
4883 foreach ( $sectionParts as $part ) {
4884 if ( $part === 'T' ) {
4885 $flags |= self::PTD_FOR_INCLUSION;
4886 }
4887 }
4888 # Preprocess the text
4889 $root = $this->preprocessToDom( $text, $flags );
4890
4891 # <h> nodes indicate section breaks
4892 # They can only occur at the top level, so we can find them by iterating the root's children
4893 $node = $root->getFirstChild();
4894
4895 # Find the target section
4896 if ( $sectionIndex == 0 ) {
4897 # Section zero doesn't nest, level=big
4898 $targetLevel = 1000;
4899 } else {
4900 while ( $node ) {
4901 if ( $node->getName() === 'h' ) {
4902 $bits = $node->splitHeading();
4903 if ( $bits['i'] == $sectionIndex ) {
4904 $targetLevel = $bits['level'];
4905 break;
4906 }
4907 }
4908 if ( $mode === 'replace' ) {
4909 $outText .= $frame->expand( $node, PPFrame::RECOVER_ORIG );
4910 }
4911 $node = $node->getNextSibling();
4912 }
4913 }
4914
4915 if ( !$node ) {
4916 # Not found
4917 if ( $mode === 'get' ) {
4918 return $newText;
4919 } else {
4920 return $text;
4921 }
4922 }
4923
4924 # Find the end of the section, including nested sections
4925 do {
4926 if ( $node->getName() === 'h' ) {
4927 $bits = $node->splitHeading();
4928 $curLevel = $bits['level'];
4929 if ( $bits['i'] != $sectionIndex && $curLevel <= $targetLevel ) {
4930 break;
4931 }
4932 }
4933 if ( $mode === 'get' ) {
4934 $outText .= $frame->expand( $node, PPFrame::RECOVER_ORIG );
4935 }
4936 $node = $node->getNextSibling();
4937 } while ( $node );
4938
4939 # Write out the remainder (in replace mode only)
4940 if ( $mode === 'replace' ) {
4941 # Output the replacement text
4942 # Add two newlines on -- trailing whitespace in $newText is conventionally
4943 # stripped by the editor, so we need both newlines to restore the paragraph gap
4944 # Only add trailing whitespace if there is newText
4945 if ( $newText != "" ) {
4946 $outText .= $newText . "\n\n";
4947 }
4948
4949 while ( $node ) {
4950 $outText .= $frame->expand( $node, PPFrame::RECOVER_ORIG );
4951 $node = $node->getNextSibling();
4952 }
4953 }
4954
4955 if ( is_string( $outText ) ) {
4956 # Re-insert stripped tags
4957 $outText = rtrim( $this->mStripState->unstripBoth( $outText ) );
4958 }
4959
4960 return $outText;
4961 }
4962
4963 /**
4964 * This function returns the text of a section, specified by a number ($section).
4965 * A section is text under a heading like == Heading == or \<h1\>Heading\</h1\>, or
4966 * the first section before any such heading (section 0).
4967 *
4968 * If a section contains subsections, these are also returned.
4969 *
4970 * @param $text String: text to look in
4971 * @param $section String: section identifier
4972 * @param $deftext String: default to return if section is not found
4973 * @return string text of the requested section
4974 */
4975 public function getSection( $text, $section, $deftext='' ) {
4976 return $this->extractSections( $text, $section, "get", $deftext );
4977 }
4978
4979 public function replaceSection( $oldtext, $section, $text ) {
4980 return $this->extractSections( $oldtext, $section, "replace", $text );
4981 }
4982
4983 /**
4984 * Get the ID of the revision we are parsing
4985 *
4986 * @return Mixed: integer or null
4987 */
4988 function getRevisionId() {
4989 return $this->mRevisionId;
4990 }
4991
4992 /**
4993 * Get the timestamp associated with the current revision, adjusted for
4994 * the default server-local timestamp
4995 */
4996 function getRevisionTimestamp() {
4997 if ( is_null( $this->mRevisionTimestamp ) ) {
4998 wfProfileIn( __METHOD__ );
4999 global $wgContLang;
5000 $dbr = wfGetDB( DB_SLAVE );
5001 $timestamp = $dbr->selectField( 'revision', 'rev_timestamp',
5002 array( 'rev_id' => $this->mRevisionId ), __METHOD__ );
5003
5004 # Normalize timestamp to internal MW format for timezone processing.
5005 # This has the added side-effect of replacing a null value with
5006 # the current time, which gives us more sensible behavior for
5007 # previews.
5008 $timestamp = wfTimestamp( TS_MW, $timestamp );
5009
5010 # The cryptic '' timezone parameter tells to use the site-default
5011 # timezone offset instead of the user settings.
5012 #
5013 # Since this value will be saved into the parser cache, served
5014 # to other users, and potentially even used inside links and such,
5015 # it needs to be consistent for all visitors.
5016 $this->mRevisionTimestamp = $wgContLang->userAdjust( $timestamp, '' );
5017
5018 wfProfileOut( __METHOD__ );
5019 }
5020 return $this->mRevisionTimestamp;
5021 }
5022
5023 /**
5024 * Get the name of the user that edited the last revision
5025 *
5026 * @return String: user name
5027 */
5028 function getRevisionUser() {
5029 # if this template is subst: the revision id will be blank,
5030 # so just use the current user's name
5031 if ( $this->mRevisionId ) {
5032 $revision = Revision::newFromId( $this->mRevisionId );
5033 $revuser = $revision->getUserText();
5034 } else {
5035 global $wgUser;
5036 $revuser = $wgUser->getName();
5037 }
5038 return $revuser;
5039 }
5040
5041 /**
5042 * Mutator for $mDefaultSort
5043 *
5044 * @param $sort New value
5045 */
5046 public function setDefaultSort( $sort ) {
5047 $this->mDefaultSort = $sort;
5048 $this->mOutput->setProperty( 'defaultsort', $sort );
5049 }
5050
5051 /**
5052 * Accessor for $mDefaultSort
5053 * Will use the title/prefixed title if none is set
5054 *
5055 * @return string
5056 */
5057 public function getDefaultSort() {
5058 if ( $this->mDefaultSort !== false ) {
5059 return $this->mDefaultSort;
5060 } else {
5061 return $this->mTitle->getCategorySortkey();
5062 }
5063 }
5064
5065 /**
5066 * Accessor for $mDefaultSort
5067 * Unlike getDefaultSort(), will return false if none is set
5068 *
5069 * @return string or false
5070 */
5071 public function getCustomDefaultSort() {
5072 return $this->mDefaultSort;
5073 }
5074
5075 /**
5076 * Try to guess the section anchor name based on a wikitext fragment
5077 * presumably extracted from a heading, for example "Header" from
5078 * "== Header ==".
5079 */
5080 public function guessSectionNameFromWikiText( $text ) {
5081 # Strip out wikitext links(they break the anchor)
5082 $text = $this->stripSectionName( $text );
5083 $text = Sanitizer::normalizeSectionNameWhitespace( $text );
5084 return '#' . Sanitizer::escapeId( $text, 'noninitial' );
5085 }
5086
5087 /**
5088 * Same as guessSectionNameFromWikiText(), but produces legacy anchors
5089 * instead. For use in redirects, since IE6 interprets Redirect: headers
5090 * as something other than UTF-8 (apparently?), resulting in breakage.
5091 *
5092 * @param $text String: The section name
5093 * @return string An anchor
5094 */
5095 public function guessLegacySectionNameFromWikiText( $text ) {
5096 # Strip out wikitext links(they break the anchor)
5097 $text = $this->stripSectionName( $text );
5098 $text = Sanitizer::normalizeSectionNameWhitespace( $text );
5099 return '#' . Sanitizer::escapeId( $text, array( 'noninitial', 'legacy' ) );
5100 }
5101
5102 /**
5103 * Strips a text string of wikitext for use in a section anchor
5104 *
5105 * Accepts a text string and then removes all wikitext from the
5106 * string and leaves only the resultant text (i.e. the result of
5107 * [[User:WikiSysop|Sysop]] would be "Sysop" and the result of
5108 * [[User:WikiSysop]] would be "User:WikiSysop") - this is intended
5109 * to create valid section anchors by mimicing the output of the
5110 * parser when headings are parsed.
5111 *
5112 * @param $text String: text string to be stripped of wikitext
5113 * for use in a Section anchor
5114 * @return Filtered text string
5115 */
5116 public function stripSectionName( $text ) {
5117 # Strip internal link markup
5118 $text = preg_replace( '/\[\[:?([^[|]+)\|([^[]+)\]\]/', '$2', $text );
5119 $text = preg_replace( '/\[\[:?([^[]+)\|?\]\]/', '$1', $text );
5120
5121 # Strip external link markup (FIXME: Not Tolerant to blank link text
5122 # I.E. [http://www.mediawiki.org] will render as [1] or something depending
5123 # on how many empty links there are on the page - need to figure that out.
5124 $text = preg_replace( '/\[(?:' . wfUrlProtocols() . ')([^ ]+?) ([^[]+)\]/', '$2', $text );
5125
5126 # Parse wikitext quotes (italics & bold)
5127 $text = $this->doQuotes( $text );
5128
5129 # Strip HTML tags
5130 $text = StringUtils::delimiterReplace( '<', '>', '', $text );
5131 return $text;
5132 }
5133
5134 /**
5135 * strip/replaceVariables/unstrip for preprocessor regression testing
5136 */
5137 function testSrvus( $text, $title, $options, $outputType = self::OT_HTML ) {
5138 $this->clearState();
5139 if ( !$title instanceof Title ) {
5140 $title = Title::newFromText( $title );
5141 }
5142 $this->mTitle = $title;
5143 $options->resetUsage();
5144 $this->mOptions = $options;
5145 $this->setOutputType( $outputType );
5146 $text = $this->replaceVariables( $text );
5147 $text = $this->mStripState->unstripBoth( $text );
5148 $text = Sanitizer::removeHTMLtags( $text );
5149 return $text;
5150 }
5151
5152 function testPst( $text, $title, $options ) {
5153 global $wgUser;
5154 if ( !$title instanceof Title ) {
5155 $title = Title::newFromText( $title );
5156 }
5157 return $this->preSaveTransform( $text, $title, $wgUser, $options );
5158 }
5159
5160 function testPreprocess( $text, $title, $options ) {
5161 if ( !$title instanceof Title ) {
5162 $title = Title::newFromText( $title );
5163 }
5164 return $this->testSrvus( $text, $title, $options, self::OT_PREPROCESS );
5165 }
5166
5167 function markerSkipCallback( $s, $callback ) {
5168 $i = 0;
5169 $out = '';
5170 while ( $i < strlen( $s ) ) {
5171 $markerStart = strpos( $s, $this->mUniqPrefix, $i );
5172 if ( $markerStart === false ) {
5173 $out .= call_user_func( $callback, substr( $s, $i ) );
5174 break;
5175 } else {
5176 $out .= call_user_func( $callback, substr( $s, $i, $markerStart - $i ) );
5177 $markerEnd = strpos( $s, self::MARKER_SUFFIX, $markerStart );
5178 if ( $markerEnd === false ) {
5179 $out .= substr( $s, $markerStart );
5180 break;
5181 } else {
5182 $markerEnd += strlen( self::MARKER_SUFFIX );
5183 $out .= substr( $s, $markerStart, $markerEnd - $markerStart );
5184 $i = $markerEnd;
5185 }
5186 }
5187 }
5188 return $out;
5189 }
5190
5191 function serialiseHalfParsedText( $text ) {
5192 $data = array();
5193 $data['text'] = $text;
5194
5195 # First, find all strip markers, and store their
5196 # data in an array.
5197 $stripState = new StripState;
5198 $pos = 0;
5199 while ( ( $start_pos = strpos( $text, $this->mUniqPrefix, $pos ) )
5200 && ( $end_pos = strpos( $text, self::MARKER_SUFFIX, $pos ) ) )
5201 {
5202 $end_pos += strlen( self::MARKER_SUFFIX );
5203 $marker = substr( $text, $start_pos, $end_pos-$start_pos );
5204
5205 if ( !empty( $this->mStripState->general->data[$marker] ) ) {
5206 $replaceArray = $stripState->general;
5207 $stripText = $this->mStripState->general->data[$marker];
5208 } elseif ( !empty( $this->mStripState->nowiki->data[$marker] ) ) {
5209 $replaceArray = $stripState->nowiki;
5210 $stripText = $this->mStripState->nowiki->data[$marker];
5211 } else {
5212 throw new MWException( "Hanging strip marker: '$marker'." );
5213 }
5214
5215 $replaceArray->setPair( $marker, $stripText );
5216 $pos = $end_pos;
5217 }
5218 $data['stripstate'] = $stripState;
5219
5220 # Now, find all of our links, and store THEIR
5221 # data in an array! :)
5222 $links = array( 'internal' => array(), 'interwiki' => array() );
5223 $pos = 0;
5224
5225 # Internal links
5226 while ( ( $start_pos = strpos( $text, '<!--LINK ', $pos ) ) ) {
5227 list( $ns, $trail ) = explode( ':', substr( $text, $start_pos + strlen( '<!--LINK ' ) ), 2 );
5228
5229 $ns = trim( $ns );
5230 if ( empty( $links['internal'][$ns] ) ) {
5231 $links['internal'][$ns] = array();
5232 }
5233
5234 $key = trim( substr( $trail, 0, strpos( $trail, '-->' ) ) );
5235 $links['internal'][$ns][] = $this->mLinkHolders->internals[$ns][$key];
5236 $pos = $start_pos + strlen( "<!--LINK $ns:$key-->" );
5237 }
5238
5239 $pos = 0;
5240
5241 # Interwiki links
5242 while ( ( $start_pos = strpos( $text, '<!--IWLINK ', $pos ) ) ) {
5243 $data = substr( $text, $start_pos );
5244 $key = trim( substr( $data, 0, strpos( $data, '-->' ) ) );
5245 $links['interwiki'][] = $this->mLinkHolders->interwiki[$key];
5246 $pos = $start_pos + strlen( "<!--IWLINK $key-->" );
5247 }
5248
5249 $data['linkholder'] = $links;
5250
5251 return $data;
5252 }
5253
5254 /**
5255 * TODO: document
5256 * @param $data Array
5257 * @param $intPrefix String unique identifying prefix
5258 * @return String
5259 */
5260 function unserialiseHalfParsedText( $data, $intPrefix = null ) {
5261 if ( !$intPrefix ) {
5262 $intPrefix = self::getRandomString();
5263 }
5264
5265 # First, extract the strip state.
5266 $stripState = $data['stripstate'];
5267 $this->mStripState->general->merge( $stripState->general );
5268 $this->mStripState->nowiki->merge( $stripState->nowiki );
5269
5270 # Now, extract the text, and renumber links
5271 $text = $data['text'];
5272 $links = $data['linkholder'];
5273
5274 # Internal...
5275 foreach ( $links['internal'] as $ns => $nsLinks ) {
5276 foreach ( $nsLinks as $key => $entry ) {
5277 $newKey = $intPrefix . '-' . $key;
5278 $this->mLinkHolders->internals[$ns][$newKey] = $entry;
5279
5280 $text = str_replace( "<!--LINK $ns:$key-->", "<!--LINK $ns:$newKey-->", $text );
5281 }
5282 }
5283
5284 # Interwiki...
5285 foreach ( $links['interwiki'] as $key => $entry ) {
5286 $newKey = "$intPrefix-$key";
5287 $this->mLinkHolders->interwikis[$newKey] = $entry;
5288
5289 $text = str_replace( "<!--IWLINK $key-->", "<!--IWLINK $newKey-->", $text );
5290 }
5291
5292 # Should be good to go.
5293 return $text;
5294 }
5295 }
5296
5297 /**
5298 * @todo document, briefly.
5299 * @ingroup Parser
5300 */
5301 class StripState {
5302 var $general, $nowiki;
5303
5304 function __construct() {
5305 $this->general = new ReplacementArray;
5306 $this->nowiki = new ReplacementArray;
5307 }
5308
5309 function unstripGeneral( $text ) {
5310 wfProfileIn( __METHOD__ );
5311 do {
5312 $oldText = $text;
5313 $text = $this->general->replace( $text );
5314 } while ( $text !== $oldText );
5315 wfProfileOut( __METHOD__ );
5316 return $text;
5317 }
5318
5319 function unstripNoWiki( $text ) {
5320 wfProfileIn( __METHOD__ );
5321 do {
5322 $oldText = $text;
5323 $text = $this->nowiki->replace( $text );
5324 } while ( $text !== $oldText );
5325 wfProfileOut( __METHOD__ );
5326 return $text;
5327 }
5328
5329 function unstripBoth( $text ) {
5330 wfProfileIn( __METHOD__ );
5331 do {
5332 $oldText = $text;
5333 $text = $this->general->replace( $text );
5334 $text = $this->nowiki->replace( $text );
5335 } while ( $text !== $oldText );
5336 wfProfileOut( __METHOD__ );
5337 return $text;
5338 }
5339 }
5340
5341 /**
5342 * @todo document, briefly.
5343 * @ingroup Parser
5344 */
5345 class OnlyIncludeReplacer {
5346 var $output = '';
5347
5348 function replace( $matches ) {
5349 if ( substr( $matches[1], -1 ) === "\n" ) {
5350 $this->output .= substr( $matches[1], 0, -1 );
5351 } else {
5352 $this->output .= $matches[1];
5353 }
5354 }
5355 }
Wiklou, le Wiki du Wiklou