Add "check" parameter to action=email
[lhc/web/wiklou.git] / includes / parser / Preprocessor_DOM.php
1 <?php
2
3 /**
4 * @ingroup Parser
5 */
6 class Preprocessor_DOM implements Preprocessor {
7 var $parser, $memoryLimit;
8
9 function __construct( $parser ) {
10 $this->parser = $parser;
11 $mem = ini_get( 'memory_limit' );
12 $this->memoryLimit = false;
13 if ( strval( $mem ) !== '' && $mem != -1 ) {
14 if ( preg_match( '/^\d+$/', $mem ) ) {
15 $this->memoryLimit = $mem;
16 } elseif ( preg_match( '/^(\d+)M$/i', $mem, $m ) ) {
17 $this->memoryLimit = $m[1] * 1048576;
18 }
19 }
20 }
21
22 function newFrame() {
23 return new PPFrame_DOM( $this );
24 }
25
26 function newCustomFrame( $args ) {
27 return new PPCustomFrame_DOM( $this, $args );
28 }
29
30 function memCheck() {
31 if ( $this->memoryLimit === false ) {
32 return;
33 }
34 $usage = memory_get_usage();
35 if ( $usage > $this->memoryLimit * 0.9 ) {
36 $limit = intval( $this->memoryLimit * 0.9 / 1048576 + 0.5 );
37 throw new MWException( "Preprocessor hit 90% memory limit ($limit MB)" );
38 }
39 return $usage <= $this->memoryLimit * 0.8;
40 }
41
42 /**
43 * Preprocess some wikitext and return the document tree.
44 * This is the ghost of Parser::replace_variables().
45 *
46 * @param string $text The text to parse
47 * @param integer flags Bitwise combination of:
48 * Parser::PTD_FOR_INCLUSION Handle <noinclude>/<includeonly> as if the text is being
49 * included. Default is to assume a direct page view.
50 *
51 * The generated DOM tree must depend only on the input text and the flags.
52 * The DOM tree must be the same in OT_HTML and OT_WIKI mode, to avoid a regression of bug 4899.
53 *
54 * Any flag added to the $flags parameter here, or any other parameter liable to cause a
55 * change in the DOM tree for a given text, must be passed through the section identifier
56 * in the section edit link and thus back to extractSections().
57 *
58 * The output of this function is currently only cached in process memory, but a persistent
59 * cache may be implemented at a later date which takes further advantage of these strict
60 * dependency requirements.
61 *
62 * @private
63 */
64 function preprocessToObj( $text, $flags = 0 ) {
65 wfProfileIn( __METHOD__ );
66 wfProfileIn( __METHOD__.'-makexml' );
67
68 $rules = array(
69 '{' => array(
70 'end' => '}',
71 'names' => array(
72 2 => 'template',
73 3 => 'tplarg',
74 ),
75 'min' => 2,
76 'max' => 3,
77 ),
78 '[' => array(
79 'end' => ']',
80 'names' => array( 2 => null ),
81 'min' => 2,
82 'max' => 2,
83 )
84 );
85
86 $forInclusion = $flags & Parser::PTD_FOR_INCLUSION;
87
88 $xmlishElements = $this->parser->getStripList();
89 $enableOnlyinclude = false;
90 if ( $forInclusion ) {
91 $ignoredTags = array( 'includeonly', '/includeonly' );
92 $ignoredElements = array( 'noinclude' );
93 $xmlishElements[] = 'noinclude';
94 if ( strpos( $text, '<onlyinclude>' ) !== false && strpos( $text, '</onlyinclude>' ) !== false ) {
95 $enableOnlyinclude = true;
96 }
97 } else {
98 $ignoredTags = array( 'noinclude', '/noinclude', 'onlyinclude', '/onlyinclude' );
99 $ignoredElements = array( 'includeonly' );
100 $xmlishElements[] = 'includeonly';
101 }
102 $xmlishRegex = implode( '|', array_merge( $xmlishElements, $ignoredTags ) );
103
104 // Use "A" modifier (anchored) instead of "^", because ^ doesn't work with an offset
105 $elementsRegex = "~($xmlishRegex)(?:\s|\/>|>)|(!--)~iA";
106
107 $stack = new PPDStack;
108
109 $searchBase = "[{<\n"; #}
110 $revText = strrev( $text ); // For fast reverse searches
111
112 $i = 0; # Input pointer, starts out pointing to a pseudo-newline before the start
113 $accum =& $stack->getAccum(); # Current accumulator
114 $accum = '<root>';
115 $findEquals = false; # True to find equals signs in arguments
116 $findPipe = false; # True to take notice of pipe characters
117 $headingIndex = 1;
118 $inHeading = false; # True if $i is inside a possible heading
119 $noMoreGT = false; # True if there are no more greater-than (>) signs right of $i
120 $findOnlyinclude = $enableOnlyinclude; # True to ignore all input up to the next <onlyinclude>
121 $fakeLineStart = true; # Do a line-start run without outputting an LF character
122
123 while ( true ) {
124 //$this->memCheck();
125
126 if ( $findOnlyinclude ) {
127 // Ignore all input up to the next <onlyinclude>
128 $startPos = strpos( $text, '<onlyinclude>', $i );
129 if ( $startPos === false ) {
130 // Ignored section runs to the end
131 $accum .= '<ignore>' . htmlspecialchars( substr( $text, $i ) ) . '</ignore>';
132 break;
133 }
134 $tagEndPos = $startPos + strlen( '<onlyinclude>' ); // past-the-end
135 $accum .= '<ignore>' . htmlspecialchars( substr( $text, $i, $tagEndPos - $i ) ) . '</ignore>';
136 $i = $tagEndPos;
137 $findOnlyinclude = false;
138 }
139
140 if ( $fakeLineStart ) {
141 $found = 'line-start';
142 $curChar = '';
143 } else {
144 # Find next opening brace, closing brace or pipe
145 $search = $searchBase;
146 if ( $stack->top === false ) {
147 $currentClosing = '';
148 } else {
149 $currentClosing = $stack->top->close;
150 $search .= $currentClosing;
151 }
152 if ( $findPipe ) {
153 $search .= '|';
154 }
155 if ( $findEquals ) {
156 // First equals will be for the template
157 $search .= '=';
158 }
159 $rule = null;
160 # Output literal section, advance input counter
161 $literalLength = strcspn( $text, $search, $i );
162 if ( $literalLength > 0 ) {
163 $accum .= htmlspecialchars( substr( $text, $i, $literalLength ) );
164 $i += $literalLength;
165 }
166 if ( $i >= strlen( $text ) ) {
167 if ( $currentClosing == "\n" ) {
168 // Do a past-the-end run to finish off the heading
169 $curChar = '';
170 $found = 'line-end';
171 } else {
172 # All done
173 break;
174 }
175 } else {
176 $curChar = $text[$i];
177 if ( $curChar == '|' ) {
178 $found = 'pipe';
179 } elseif ( $curChar == '=' ) {
180 $found = 'equals';
181 } elseif ( $curChar == '<' ) {
182 $found = 'angle';
183 } elseif ( $curChar == "\n" ) {
184 if ( $inHeading ) {
185 $found = 'line-end';
186 } else {
187 $found = 'line-start';
188 }
189 } elseif ( $curChar == $currentClosing ) {
190 $found = 'close';
191 } elseif ( isset( $rules[$curChar] ) ) {
192 $found = 'open';
193 $rule = $rules[$curChar];
194 } else {
195 # Some versions of PHP have a strcspn which stops on null characters
196 # Ignore and continue
197 ++$i;
198 continue;
199 }
200 }
201 }
202
203 if ( $found == 'angle' ) {
204 $matches = false;
205 // Handle </onlyinclude>
206 if ( $enableOnlyinclude && substr( $text, $i, strlen( '</onlyinclude>' ) ) == '</onlyinclude>' ) {
207 $findOnlyinclude = true;
208 continue;
209 }
210
211 // Determine element name
212 if ( !preg_match( $elementsRegex, $text, $matches, 0, $i + 1 ) ) {
213 // Element name missing or not listed
214 $accum .= '&lt;';
215 ++$i;
216 continue;
217 }
218 // Handle comments
219 if ( isset( $matches[2] ) && $matches[2] == '!--' ) {
220 // To avoid leaving blank lines, when a comment is both preceded
221 // and followed by a newline (ignoring spaces), trim leading and
222 // trailing spaces and one of the newlines.
223
224 // Find the end
225 $endPos = strpos( $text, '-->', $i + 4 );
226 if ( $endPos === false ) {
227 // Unclosed comment in input, runs to end
228 $inner = substr( $text, $i );
229 $accum .= '<comment>' . htmlspecialchars( $inner ) . '</comment>';
230 $i = strlen( $text );
231 } else {
232 // Search backwards for leading whitespace
233 $wsStart = $i ? ( $i - strspn( $revText, ' ', strlen( $text ) - $i ) ) : 0;
234 // Search forwards for trailing whitespace
235 // $wsEnd will be the position of the last space
236 $wsEnd = $endPos + 2 + strspn( $text, ' ', $endPos + 3 );
237 // Eat the line if possible
238 // TODO: This could theoretically be done if $wsStart == 0, i.e. for comments at
239 // the overall start. That's not how Sanitizer::removeHTMLcomments() did it, but
240 // it's a possible beneficial b/c break.
241 if ( $wsStart > 0 && substr( $text, $wsStart - 1, 1 ) == "\n"
242 && substr( $text, $wsEnd + 1, 1 ) == "\n" )
243 {
244 $startPos = $wsStart;
245 $endPos = $wsEnd + 1;
246 // Remove leading whitespace from the end of the accumulator
247 // Sanity check first though
248 $wsLength = $i - $wsStart;
249 if ( $wsLength > 0 && substr( $accum, -$wsLength ) === str_repeat( ' ', $wsLength ) ) {
250 $accum = substr( $accum, 0, -$wsLength );
251 }
252 // Do a line-start run next time to look for headings after the comment
253 $fakeLineStart = true;
254 } else {
255 // No line to eat, just take the comment itself
256 $startPos = $i;
257 $endPos += 2;
258 }
259
260 if ( $stack->top ) {
261 $part = $stack->top->getCurrentPart();
262 if ( isset( $part->commentEnd ) && $part->commentEnd == $wsStart - 1 ) {
263 // Comments abutting, no change in visual end
264 $part->commentEnd = $wsEnd;
265 } else {
266 $part->visualEnd = $wsStart;
267 $part->commentEnd = $endPos;
268 }
269 }
270 $i = $endPos + 1;
271 $inner = substr( $text, $startPos, $endPos - $startPos + 1 );
272 $accum .= '<comment>' . htmlspecialchars( $inner ) . '</comment>';
273 }
274 continue;
275 }
276 $name = $matches[1];
277 $lowerName = strtolower( $name );
278 $attrStart = $i + strlen( $name ) + 1;
279
280 // Find end of tag
281 $tagEndPos = $noMoreGT ? false : strpos( $text, '>', $attrStart );
282 if ( $tagEndPos === false ) {
283 // Infinite backtrack
284 // Disable tag search to prevent worst-case O(N^2) performance
285 $noMoreGT = true;
286 $accum .= '&lt;';
287 ++$i;
288 continue;
289 }
290
291 // Handle ignored tags
292 if ( in_array( $lowerName, $ignoredTags ) ) {
293 $accum .= '<ignore>' . htmlspecialchars( substr( $text, $i, $tagEndPos - $i + 1 ) ) . '</ignore>';
294 $i = $tagEndPos + 1;
295 continue;
296 }
297
298 $tagStartPos = $i;
299 if ( $text[$tagEndPos-1] == '/' ) {
300 $attrEnd = $tagEndPos - 1;
301 $inner = null;
302 $i = $tagEndPos + 1;
303 $close = '';
304 } else {
305 $attrEnd = $tagEndPos;
306 // Find closing tag
307 if ( preg_match( "/<\/$name\s*>/i", $text, $matches, PREG_OFFSET_CAPTURE, $tagEndPos + 1 ) ) {
308 $inner = substr( $text, $tagEndPos + 1, $matches[0][1] - $tagEndPos - 1 );
309 $i = $matches[0][1] + strlen( $matches[0][0] );
310 $close = '<close>' . htmlspecialchars( $matches[0][0] ) . '</close>';
311 } else {
312 // No end tag -- let it run out to the end of the text.
313 $inner = substr( $text, $tagEndPos + 1 );
314 $i = strlen( $text );
315 $close = '';
316 }
317 }
318 // <includeonly> and <noinclude> just become <ignore> tags
319 if ( in_array( $lowerName, $ignoredElements ) ) {
320 $accum .= '<ignore>' . htmlspecialchars( substr( $text, $tagStartPos, $i - $tagStartPos ) )
321 . '</ignore>';
322 continue;
323 }
324
325 $accum .= '<ext>';
326 if ( $attrEnd <= $attrStart ) {
327 $attr = '';
328 } else {
329 $attr = substr( $text, $attrStart, $attrEnd - $attrStart );
330 }
331 $accum .= '<name>' . htmlspecialchars( $name ) . '</name>' .
332 // Note that the attr element contains the whitespace between name and attribute,
333 // this is necessary for precise reconstruction during pre-save transform.
334 '<attr>' . htmlspecialchars( $attr ) . '</attr>';
335 if ( $inner !== null ) {
336 $accum .= '<inner>' . htmlspecialchars( $inner ) . '</inner>';
337 }
338 $accum .= $close . '</ext>';
339 }
340
341 elseif ( $found == 'line-start' ) {
342 // Is this the start of a heading?
343 // Line break belongs before the heading element in any case
344 if ( $fakeLineStart ) {
345 $fakeLineStart = false;
346 } else {
347 $accum .= $curChar;
348 $i++;
349 }
350
351 $count = strspn( $text, '=', $i, 6 );
352 if ( $count == 1 && $findEquals ) {
353 // DWIM: This looks kind of like a name/value separator
354 // Let's let the equals handler have it and break the potential heading
355 // This is heuristic, but AFAICT the methods for completely correct disambiguation are very complex.
356 } elseif ( $count > 0 ) {
357 $piece = array(
358 'open' => "\n",
359 'close' => "\n",
360 'parts' => array( new PPDPart( str_repeat( '=', $count ) ) ),
361 'startPos' => $i,
362 'count' => $count );
363 $stack->push( $piece );
364 $accum =& $stack->getAccum();
365 extract( $stack->getFlags() );
366 $i += $count;
367 }
368 }
369
370 elseif ( $found == 'line-end' ) {
371 $piece = $stack->top;
372 // A heading must be open, otherwise \n wouldn't have been in the search list
373 assert( $piece->open == "\n" );
374 $part = $piece->getCurrentPart();
375 // Search back through the input to see if it has a proper close
376 // Do this using the reversed string since the other solutions (end anchor, etc.) are inefficient
377 $wsLength = strspn( $revText, " \t", strlen( $text ) - $i );
378 $searchStart = $i - $wsLength;
379 if ( isset( $part->commentEnd ) && $searchStart - 1 == $part->commentEnd ) {
380 // Comment found at line end
381 // Search for equals signs before the comment
382 $searchStart = $part->visualEnd;
383 $searchStart -= strspn( $revText, " \t", strlen( $text ) - $searchStart );
384 }
385 $count = $piece->count;
386 $equalsLength = strspn( $revText, '=', strlen( $text ) - $searchStart );
387 if ( $equalsLength > 0 ) {
388 if ( $i - $equalsLength == $piece->startPos ) {
389 // This is just a single string of equals signs on its own line
390 // Replicate the doHeadings behaviour /={count}(.+)={count}/
391 // First find out how many equals signs there really are (don't stop at 6)
392 $count = $equalsLength;
393 if ( $count < 3 ) {
394 $count = 0;
395 } else {
396 $count = min( 6, intval( ( $count - 1 ) / 2 ) );
397 }
398 } else {
399 $count = min( $equalsLength, $count );
400 }
401 if ( $count > 0 ) {
402 // Normal match, output <h>
403 $element = "<h level=\"$count\" i=\"$headingIndex\">$accum</h>";
404 $headingIndex++;
405 } else {
406 // Single equals sign on its own line, count=0
407 $element = $accum;
408 }
409 } else {
410 // No match, no <h>, just pass down the inner text
411 $element = $accum;
412 }
413 // Unwind the stack
414 $stack->pop();
415 $accum =& $stack->getAccum();
416 extract( $stack->getFlags() );
417
418 // Append the result to the enclosing accumulator
419 $accum .= $element;
420 // Note that we do NOT increment the input pointer.
421 // This is because the closing linebreak could be the opening linebreak of
422 // another heading. Infinite loops are avoided because the next iteration MUST
423 // hit the heading open case above, which unconditionally increments the
424 // input pointer.
425 }
426
427 elseif ( $found == 'open' ) {
428 # count opening brace characters
429 $count = strspn( $text, $curChar, $i );
430
431 # we need to add to stack only if opening brace count is enough for one of the rules
432 if ( $count >= $rule['min'] ) {
433 # Add it to the stack
434 $piece = array(
435 'open' => $curChar,
436 'close' => $rule['end'],
437 'count' => $count,
438 'lineStart' => ($i > 0 && $text[$i-1] == "\n"),
439 );
440
441 $stack->push( $piece );
442 $accum =& $stack->getAccum();
443 extract( $stack->getFlags() );
444 } else {
445 # Add literal brace(s)
446 $accum .= htmlspecialchars( str_repeat( $curChar, $count ) );
447 }
448 $i += $count;
449 }
450
451 elseif ( $found == 'close' ) {
452 $piece = $stack->top;
453 # lets check if there are enough characters for closing brace
454 $maxCount = $piece->count;
455 $count = strspn( $text, $curChar, $i, $maxCount );
456
457 # check for maximum matching characters (if there are 5 closing
458 # characters, we will probably need only 3 - depending on the rules)
459 $matchingCount = 0;
460 $rule = $rules[$piece->open];
461 if ( $count > $rule['max'] ) {
462 # The specified maximum exists in the callback array, unless the caller
463 # has made an error
464 $matchingCount = $rule['max'];
465 } else {
466 # Count is less than the maximum
467 # Skip any gaps in the callback array to find the true largest match
468 # Need to use array_key_exists not isset because the callback can be null
469 $matchingCount = $count;
470 while ( $matchingCount > 0 && !array_key_exists( $matchingCount, $rule['names'] ) ) {
471 --$matchingCount;
472 }
473 }
474
475 if ($matchingCount <= 0) {
476 # No matching element found in callback array
477 # Output a literal closing brace and continue
478 $accum .= htmlspecialchars( str_repeat( $curChar, $count ) );
479 $i += $count;
480 continue;
481 }
482 $name = $rule['names'][$matchingCount];
483 if ( $name === null ) {
484 // No element, just literal text
485 $element = $piece->breakSyntax( $matchingCount ) . str_repeat( $rule['end'], $matchingCount );
486 } else {
487 # Create XML element
488 # Note: $parts is already XML, does not need to be encoded further
489 $parts = $piece->parts;
490 $title = $parts[0]->out;
491 unset( $parts[0] );
492
493 # The invocation is at the start of the line if lineStart is set in
494 # the stack, and all opening brackets are used up.
495 if ( $maxCount == $matchingCount && !empty( $piece->lineStart ) ) {
496 $attr = ' lineStart="1"';
497 } else {
498 $attr = '';
499 }
500
501 $element = "<$name$attr>";
502 $element .= "<title>$title</title>";
503 $argIndex = 1;
504 foreach ( $parts as $partIndex => $part ) {
505 if ( isset( $part->eqpos ) ) {
506 $argName = substr( $part->out, 0, $part->eqpos );
507 $argValue = substr( $part->out, $part->eqpos + 1 );
508 $element .= "<part><name>$argName</name>=<value>$argValue</value></part>";
509 } else {
510 $element .= "<part><name index=\"$argIndex\" /><value>{$part->out}</value></part>";
511 $argIndex++;
512 }
513 }
514 $element .= "</$name>";
515 }
516
517 # Advance input pointer
518 $i += $matchingCount;
519
520 # Unwind the stack
521 $stack->pop();
522 $accum =& $stack->getAccum();
523
524 # Re-add the old stack element if it still has unmatched opening characters remaining
525 if ($matchingCount < $piece->count) {
526 $piece->parts = array( new PPDPart );
527 $piece->count -= $matchingCount;
528 # do we still qualify for any callback with remaining count?
529 $names = $rules[$piece->open]['names'];
530 $skippedBraces = 0;
531 $enclosingAccum =& $accum;
532 while ( $piece->count ) {
533 if ( array_key_exists( $piece->count, $names ) ) {
534 $stack->push( $piece );
535 $accum =& $stack->getAccum();
536 break;
537 }
538 --$piece->count;
539 $skippedBraces ++;
540 }
541 $enclosingAccum .= str_repeat( $piece->open, $skippedBraces );
542 }
543
544 extract( $stack->getFlags() );
545
546 # Add XML element to the enclosing accumulator
547 $accum .= $element;
548 }
549
550 elseif ( $found == 'pipe' ) {
551 $findEquals = true; // shortcut for getFlags()
552 $stack->addPart();
553 $accum =& $stack->getAccum();
554 ++$i;
555 }
556
557 elseif ( $found == 'equals' ) {
558 $findEquals = false; // shortcut for getFlags()
559 $stack->getCurrentPart()->eqpos = strlen( $accum );
560 $accum .= '=';
561 ++$i;
562 }
563 }
564
565 # Output any remaining unclosed brackets
566 foreach ( $stack->stack as $piece ) {
567 $stack->rootAccum .= $piece->breakSyntax();
568 }
569 $stack->rootAccum .= '</root>';
570 $xml = $stack->rootAccum;
571
572 wfProfileOut( __METHOD__.'-makexml' );
573 wfProfileIn( __METHOD__.'-loadXML' );
574 $dom = new DOMDocument;
575 wfSuppressWarnings();
576 $result = $dom->loadXML( $xml );
577 wfRestoreWarnings();
578 if ( !$result ) {
579 // Try running the XML through UtfNormal to get rid of invalid characters
580 $xml = UtfNormal::cleanUp( $xml );
581 $result = $dom->loadXML( $xml );
582 if ( !$result ) {
583 throw new MWException( __METHOD__.' generated invalid XML' );
584 }
585 }
586 $obj = new PPNode_DOM( $dom->documentElement );
587 wfProfileOut( __METHOD__.'-loadXML' );
588 wfProfileOut( __METHOD__ );
589 return $obj;
590 }
591 }
592
593 /**
594 * Stack class to help Preprocessor::preprocessToObj()
595 * @ingroup Parser
596 */
597 class PPDStack {
598 var $stack, $rootAccum, $top;
599 var $out;
600 var $elementClass = 'PPDStackElement';
601
602 static $false = false;
603
604 function __construct() {
605 $this->stack = array();
606 $this->top = false;
607 $this->rootAccum = '';
608 $this->accum =& $this->rootAccum;
609 }
610
611 function count() {
612 return count( $this->stack );
613 }
614
615 function &getAccum() {
616 return $this->accum;
617 }
618
619 function getCurrentPart() {
620 if ( $this->top === false ) {
621 return false;
622 } else {
623 return $this->top->getCurrentPart();
624 }
625 }
626
627 function push( $data ) {
628 if ( $data instanceof $this->elementClass ) {
629 $this->stack[] = $data;
630 } else {
631 $class = $this->elementClass;
632 $this->stack[] = new $class( $data );
633 }
634 $this->top = $this->stack[ count( $this->stack ) - 1 ];
635 $this->accum =& $this->top->getAccum();
636 }
637
638 function pop() {
639 if ( !count( $this->stack ) ) {
640 throw new MWException( __METHOD__.': no elements remaining' );
641 }
642 $temp = array_pop( $this->stack );
643
644 if ( count( $this->stack ) ) {
645 $this->top = $this->stack[ count( $this->stack ) - 1 ];
646 $this->accum =& $this->top->getAccum();
647 } else {
648 $this->top = self::$false;
649 $this->accum =& $this->rootAccum;
650 }
651 return $temp;
652 }
653
654 function addPart( $s = '' ) {
655 $this->top->addPart( $s );
656 $this->accum =& $this->top->getAccum();
657 }
658
659 function getFlags() {
660 if ( !count( $this->stack ) ) {
661 return array(
662 'findEquals' => false,
663 'findPipe' => false,
664 'inHeading' => false,
665 );
666 } else {
667 return $this->top->getFlags();
668 }
669 }
670 }
671
672 /**
673 * @ingroup Parser
674 */
675 class PPDStackElement {
676 var $open, // Opening character (\n for heading)
677 $close, // Matching closing character
678 $count, // Number of opening characters found (number of "=" for heading)
679 $parts, // Array of PPDPart objects describing pipe-separated parts.
680 $lineStart; // True if the open char appeared at the start of the input line. Not set for headings.
681
682 var $partClass = 'PPDPart';
683
684 function __construct( $data = array() ) {
685 $class = $this->partClass;
686 $this->parts = array( new $class );
687
688 foreach ( $data as $name => $value ) {
689 $this->$name = $value;
690 }
691 }
692
693 function &getAccum() {
694 return $this->parts[count($this->parts) - 1]->out;
695 }
696
697 function addPart( $s = '' ) {
698 $class = $this->partClass;
699 $this->parts[] = new $class( $s );
700 }
701
702 function getCurrentPart() {
703 return $this->parts[count($this->parts) - 1];
704 }
705
706 function getFlags() {
707 $partCount = count( $this->parts );
708 $findPipe = $this->open != "\n" && $this->open != '[';
709 return array(
710 'findPipe' => $findPipe,
711 'findEquals' => $findPipe && $partCount > 1 && !isset( $this->parts[$partCount - 1]->eqpos ),
712 'inHeading' => $this->open == "\n",
713 );
714 }
715
716 /**
717 * Get the output string that would result if the close is not found.
718 */
719 function breakSyntax( $openingCount = false ) {
720 if ( $this->open == "\n" ) {
721 $s = $this->parts[0]->out;
722 } else {
723 if ( $openingCount === false ) {
724 $openingCount = $this->count;
725 }
726 $s = str_repeat( $this->open, $openingCount );
727 $first = true;
728 foreach ( $this->parts as $part ) {
729 if ( $first ) {
730 $first = false;
731 } else {
732 $s .= '|';
733 }
734 $s .= $part->out;
735 }
736 }
737 return $s;
738 }
739 }
740
741 /**
742 * @ingroup Parser
743 */
744 class PPDPart {
745 var $out; // Output accumulator string
746
747 // Optional member variables:
748 // eqpos Position of equals sign in output accumulator
749 // commentEnd Past-the-end input pointer for the last comment encountered
750 // visualEnd Past-the-end input pointer for the end of the accumulator minus comments
751
752 function __construct( $out = '' ) {
753 $this->out = $out;
754 }
755 }
756
757 /**
758 * An expansion frame, used as a context to expand the result of preprocessToObj()
759 * @ingroup Parser
760 */
761 class PPFrame_DOM implements PPFrame {
762 var $preprocessor, $parser, $title;
763 var $titleCache;
764
765 /**
766 * Hashtable listing templates which are disallowed for expansion in this frame,
767 * having been encountered previously in parent frames.
768 */
769 var $loopCheckHash;
770
771 /**
772 * Recursion depth of this frame, top = 0
773 * Note that this is NOT the same as expansion depth in expand()
774 */
775 var $depth;
776
777
778 /**
779 * Construct a new preprocessor frame.
780 * @param Preprocessor $preprocessor The parent preprocessor
781 */
782 function __construct( $preprocessor ) {
783 $this->preprocessor = $preprocessor;
784 $this->parser = $preprocessor->parser;
785 $this->title = $this->parser->mTitle;
786 $this->titleCache = array( $this->title ? $this->title->getPrefixedDBkey() : false );
787 $this->loopCheckHash = array();
788 $this->depth = 0;
789 }
790
791 /**
792 * Create a new child frame
793 * $args is optionally a multi-root PPNode or array containing the template arguments
794 */
795 function newChild( $args = false, $title = false ) {
796 $namedArgs = array();
797 $numberedArgs = array();
798 if ( $title === false ) {
799 $title = $this->title;
800 }
801 if ( $args !== false ) {
802 $xpath = false;
803 if ( $args instanceof PPNode ) {
804 $args = $args->node;
805 }
806 foreach ( $args as $arg ) {
807 if ( !$xpath ) {
808 $xpath = new DOMXPath( $arg->ownerDocument );
809 }
810
811 $nameNodes = $xpath->query( 'name', $arg );
812 $value = $xpath->query( 'value', $arg );
813 if ( $nameNodes->item( 0 )->hasAttributes() ) {
814 // Numbered parameter
815 $index = $nameNodes->item( 0 )->attributes->getNamedItem( 'index' )->textContent;
816 $numberedArgs[$index] = $value->item( 0 );
817 unset( $namedArgs[$index] );
818 } else {
819 // Named parameter
820 $name = trim( $this->expand( $nameNodes->item( 0 ), PPFrame::STRIP_COMMENTS ) );
821 $namedArgs[$name] = $value->item( 0 );
822 unset( $numberedArgs[$name] );
823 }
824 }
825 }
826 return new PPTemplateFrame_DOM( $this->preprocessor, $this, $numberedArgs, $namedArgs, $title );
827 }
828
829 function expand( $root, $flags = 0 ) {
830 static $expansionDepth = 0;
831 if ( is_string( $root ) ) {
832 return $root;
833 }
834 wfProfileIn( __METHOD__ );
835
836 if ( ++$this->parser->mPPNodeCount > $this->parser->mOptions->mMaxPPNodeCount )
837 {
838 return '<span class="error">Node-count limit exceeded</span>';
839 }
840
841 if ( $expansionDepth > $this->parser->mOptions->mMaxPPExpandDepth ) {
842 return '<span class="error">Expansion depth limit exceeded</span>';
843 }
844 ++$expansionDepth;
845
846 if ( $root instanceof PPNode_DOM ) {
847 $root = $root->node;
848 }
849 if ( $root instanceof DOMDocument ) {
850 $root = $root->documentElement;
851 }
852
853 $outStack = array( '', '' );
854 $iteratorStack = array( false, $root );
855 $indexStack = array( 0, 0 );
856
857 while ( count( $iteratorStack ) > 1 ) {
858 $level = count( $outStack ) - 1;
859 $iteratorNode =& $iteratorStack[ $level ];
860 $out =& $outStack[$level];
861 $index =& $indexStack[$level];
862
863 if ( $iteratorNode instanceof PPNode_DOM ) $iteratorNode = $iteratorNode->node;
864
865 if ( is_array( $iteratorNode ) ) {
866 if ( $index >= count( $iteratorNode ) ) {
867 // All done with this iterator
868 $iteratorStack[$level] = false;
869 $contextNode = false;
870 } else {
871 $contextNode = $iteratorNode[$index];
872 $index++;
873 }
874 } elseif ( $iteratorNode instanceof DOMNodeList ) {
875 if ( $index >= $iteratorNode->length ) {
876 // All done with this iterator
877 $iteratorStack[$level] = false;
878 $contextNode = false;
879 } else {
880 $contextNode = $iteratorNode->item( $index );
881 $index++;
882 }
883 } else {
884 // Copy to $contextNode and then delete from iterator stack,
885 // because this is not an iterator but we do have to execute it once
886 $contextNode = $iteratorStack[$level];
887 $iteratorStack[$level] = false;
888 }
889
890 if ( $contextNode instanceof PPNode_DOM ) $contextNode = $contextNode->node;
891
892 $newIterator = false;
893
894 if ( $contextNode === false ) {
895 // nothing to do
896 } elseif ( is_string( $contextNode ) ) {
897 $out .= $contextNode;
898 } elseif ( is_array( $contextNode ) || $contextNode instanceof DOMNodeList ) {
899 $newIterator = $contextNode;
900 } elseif ( $contextNode instanceof DOMNode ) {
901 if ( $contextNode->nodeType == XML_TEXT_NODE ) {
902 $out .= $contextNode->nodeValue;
903 } elseif ( $contextNode->nodeName == 'template' ) {
904 # Double-brace expansion
905 $xpath = new DOMXPath( $contextNode->ownerDocument );
906 $titles = $xpath->query( 'title', $contextNode );
907 $title = $titles->item( 0 );
908 $parts = $xpath->query( 'part', $contextNode );
909 if ( $flags & self::NO_TEMPLATES ) {
910 $newIterator = $this->virtualBracketedImplode( '{{', '|', '}}', $title, $parts );
911 } else {
912 $lineStart = $contextNode->getAttribute( 'lineStart' );
913 $params = array(
914 'title' => new PPNode_DOM( $title ),
915 'parts' => new PPNode_DOM( $parts ),
916 'lineStart' => $lineStart );
917 $ret = $this->parser->braceSubstitution( $params, $this );
918 if ( isset( $ret['object'] ) ) {
919 $newIterator = $ret['object'];
920 } else {
921 $out .= $ret['text'];
922 }
923 }
924 } elseif ( $contextNode->nodeName == 'tplarg' ) {
925 # Triple-brace expansion
926 $xpath = new DOMXPath( $contextNode->ownerDocument );
927 $titles = $xpath->query( 'title', $contextNode );
928 $title = $titles->item( 0 );
929 $parts = $xpath->query( 'part', $contextNode );
930 if ( $flags & self::NO_ARGS ) {
931 $newIterator = $this->virtualBracketedImplode( '{{{', '|', '}}}', $title, $parts );
932 } else {
933 $params = array(
934 'title' => new PPNode_DOM( $title ),
935 'parts' => new PPNode_DOM( $parts ) );
936 $ret = $this->parser->argSubstitution( $params, $this );
937 if ( isset( $ret['object'] ) ) {
938 $newIterator = $ret['object'];
939 } else {
940 $out .= $ret['text'];
941 }
942 }
943 } elseif ( $contextNode->nodeName == 'comment' ) {
944 # HTML-style comment
945 # Remove it in HTML, pre+remove and STRIP_COMMENTS modes
946 if ( $this->parser->ot['html']
947 || ( $this->parser->ot['pre'] && $this->parser->mOptions->getRemoveComments() )
948 || ( $flags & self::STRIP_COMMENTS ) )
949 {
950 $out .= '';
951 }
952 # Add a strip marker in PST mode so that pstPass2() can run some old-fashioned regexes on the result
953 # Not in RECOVER_COMMENTS mode (extractSections) though
954 elseif ( $this->parser->ot['wiki'] && ! ( $flags & self::RECOVER_COMMENTS ) ) {
955 $out .= $this->parser->insertStripItem( $contextNode->textContent );
956 }
957 # Recover the literal comment in RECOVER_COMMENTS and pre+no-remove
958 else {
959 $out .= $contextNode->textContent;
960 }
961 } elseif ( $contextNode->nodeName == 'ignore' ) {
962 # Output suppression used by <includeonly> etc.
963 # OT_WIKI will only respect <ignore> in substed templates.
964 # The other output types respect it unless NO_IGNORE is set.
965 # extractSections() sets NO_IGNORE and so never respects it.
966 if ( ( !isset( $this->parent ) && $this->parser->ot['wiki'] ) || ( $flags & self::NO_IGNORE ) ) {
967 $out .= $contextNode->textContent;
968 } else {
969 $out .= '';
970 }
971 } elseif ( $contextNode->nodeName == 'ext' ) {
972 # Extension tag
973 $xpath = new DOMXPath( $contextNode->ownerDocument );
974 $names = $xpath->query( 'name', $contextNode );
975 $attrs = $xpath->query( 'attr', $contextNode );
976 $inners = $xpath->query( 'inner', $contextNode );
977 $closes = $xpath->query( 'close', $contextNode );
978 $params = array(
979 'name' => new PPNode_DOM( $names->item( 0 ) ),
980 'attr' => $attrs->length > 0 ? new PPNode_DOM( $attrs->item( 0 ) ) : null,
981 'inner' => $inners->length > 0 ? new PPNode_DOM( $inners->item( 0 ) ) : null,
982 'close' => $closes->length > 0 ? new PPNode_DOM( $closes->item( 0 ) ) : null,
983 );
984 $out .= $this->parser->extensionSubstitution( $params, $this );
985 } elseif ( $contextNode->nodeName == 'h' ) {
986 # Heading
987 $s = $this->expand( $contextNode->childNodes, $flags );
988
989 # Insert a heading marker only for <h> children of <root>
990 # This is to stop extractSections from going over multiple tree levels
991 if ( $contextNode->parentNode->nodeName == 'root'
992 && $this->parser->ot['html'] )
993 {
994 # Insert heading index marker
995 $headingIndex = $contextNode->getAttribute( 'i' );
996 $titleText = $this->title->getPrefixedDBkey();
997 $this->parser->mHeadings[] = array( $titleText, $headingIndex );
998 $serial = count( $this->parser->mHeadings ) - 1;
999 $marker = "{$this->parser->mUniqPrefix}-h-$serial-" . Parser::MARKER_SUFFIX;
1000 $count = $contextNode->getAttribute( 'level' );
1001 $s = substr( $s, 0, $count ) . $marker . substr( $s, $count );
1002 $this->parser->mStripState->general->setPair( $marker, '' );
1003 }
1004 $out .= $s;
1005 } else {
1006 # Generic recursive expansion
1007 $newIterator = $contextNode->childNodes;
1008 }
1009 } else {
1010 wfProfileOut( __METHOD__ );
1011 throw new MWException( __METHOD__.': Invalid parameter type' );
1012 }
1013
1014 if ( $newIterator !== false ) {
1015 if ( $newIterator instanceof PPNode_DOM ) {
1016 $newIterator = $newIterator->node;
1017 }
1018 $outStack[] = '';
1019 $iteratorStack[] = $newIterator;
1020 $indexStack[] = 0;
1021 } elseif ( $iteratorStack[$level] === false ) {
1022 // Return accumulated value to parent
1023 // With tail recursion
1024 while ( $iteratorStack[$level] === false && $level > 0 ) {
1025 $outStack[$level - 1] .= $out;
1026 array_pop( $outStack );
1027 array_pop( $iteratorStack );
1028 array_pop( $indexStack );
1029 $level--;
1030 }
1031 }
1032 }
1033 --$expansionDepth;
1034 wfProfileOut( __METHOD__ );
1035 return $outStack[0];
1036 }
1037
1038 function implodeWithFlags( $sep, $flags /*, ... */ ) {
1039 $args = array_slice( func_get_args(), 2 );
1040
1041 $first = true;
1042 $s = '';
1043 foreach ( $args as $root ) {
1044 if ( $root instanceof PPNode_DOM ) $root = $root->node;
1045 if ( !is_array( $root ) && !( $root instanceof DOMNodeList ) ) {
1046 $root = array( $root );
1047 }
1048 foreach ( $root as $node ) {
1049 if ( $first ) {
1050 $first = false;
1051 } else {
1052 $s .= $sep;
1053 }
1054 $s .= $this->expand( $node, $flags );
1055 }
1056 }
1057 return $s;
1058 }
1059
1060 /**
1061 * Implode with no flags specified
1062 * This previously called implodeWithFlags but has now been inlined to reduce stack depth
1063 */
1064 function implode( $sep /*, ... */ ) {
1065 $args = array_slice( func_get_args(), 1 );
1066
1067 $first = true;
1068 $s = '';
1069 foreach ( $args as $root ) {
1070 if ( $root instanceof PPNode_DOM ) $root = $root->node;
1071 if ( !is_array( $root ) && !( $root instanceof DOMNodeList ) ) {
1072 $root = array( $root );
1073 }
1074 foreach ( $root as $node ) {
1075 if ( $first ) {
1076 $first = false;
1077 } else {
1078 $s .= $sep;
1079 }
1080 $s .= $this->expand( $node );
1081 }
1082 }
1083 return $s;
1084 }
1085
1086 /**
1087 * Makes an object that, when expand()ed, will be the same as one obtained
1088 * with implode()
1089 */
1090 function virtualImplode( $sep /*, ... */ ) {
1091 $args = array_slice( func_get_args(), 1 );
1092 $out = array();
1093 $first = true;
1094 if ( $root instanceof PPNode_DOM ) $root = $root->node;
1095
1096 foreach ( $args as $root ) {
1097 if ( !is_array( $root ) && !( $root instanceof DOMNodeList ) ) {
1098 $root = array( $root );
1099 }
1100 foreach ( $root as $node ) {
1101 if ( $first ) {
1102 $first = false;
1103 } else {
1104 $out[] = $sep;
1105 }
1106 $out[] = $node;
1107 }
1108 }
1109 return $out;
1110 }
1111
1112 /**
1113 * Virtual implode with brackets
1114 */
1115 function virtualBracketedImplode( $start, $sep, $end /*, ... */ ) {
1116 $args = array_slice( func_get_args(), 3 );
1117 $out = array( $start );
1118 $first = true;
1119
1120 foreach ( $args as $root ) {
1121 if ( $root instanceof PPNode_DOM ) $root = $root->node;
1122 if ( !is_array( $root ) && !( $root instanceof DOMNodeList ) ) {
1123 $root = array( $root );
1124 }
1125 foreach ( $root as $node ) {
1126 if ( $first ) {
1127 $first = false;
1128 } else {
1129 $out[] = $sep;
1130 }
1131 $out[] = $node;
1132 }
1133 }
1134 $out[] = $end;
1135 return $out;
1136 }
1137
1138 function __toString() {
1139 return 'frame{}';
1140 }
1141
1142 function getPDBK( $level = false ) {
1143 if ( $level === false ) {
1144 return $this->title->getPrefixedDBkey();
1145 } else {
1146 return isset( $this->titleCache[$level] ) ? $this->titleCache[$level] : false;
1147 }
1148 }
1149
1150 /**
1151 * Returns true if there are no arguments in this frame
1152 */
1153 function isEmpty() {
1154 return true;
1155 }
1156
1157 function getArgument( $name ) {
1158 return false;
1159 }
1160
1161 /**
1162 * Returns true if the infinite loop check is OK, false if a loop is detected
1163 */
1164 function loopCheck( $title ) {
1165 return !isset( $this->loopCheckHash[$title->getPrefixedDBkey()] );
1166 }
1167
1168 /**
1169 * Return true if the frame is a template frame
1170 */
1171 function isTemplate() {
1172 return false;
1173 }
1174 }
1175
1176 /**
1177 * Expansion frame with template arguments
1178 * @ingroup Parser
1179 */
1180 class PPTemplateFrame_DOM extends PPFrame_DOM {
1181 var $numberedArgs, $namedArgs, $parent;
1182 var $numberedExpansionCache, $namedExpansionCache;
1183
1184 function __construct( $preprocessor, $parent = false, $numberedArgs = array(), $namedArgs = array(), $title = false ) {
1185 $this->preprocessor = $preprocessor;
1186 $this->parser = $preprocessor->parser;
1187 $this->parent = $parent;
1188 $this->numberedArgs = $numberedArgs;
1189 $this->namedArgs = $namedArgs;
1190 $this->title = $title;
1191 $pdbk = $title ? $title->getPrefixedDBkey() : false;
1192 $this->titleCache = $parent->titleCache;
1193 $this->titleCache[] = $pdbk;
1194 $this->loopCheckHash = /*clone*/ $parent->loopCheckHash;
1195 if ( $pdbk !== false ) {
1196 $this->loopCheckHash[$pdbk] = true;
1197 }
1198 $this->depth = $parent->depth + 1;
1199 $this->numberedExpansionCache = $this->namedExpansionCache = array();
1200 }
1201
1202 function __toString() {
1203 $s = 'tplframe{';
1204 $first = true;
1205 $args = $this->numberedArgs + $this->namedArgs;
1206 foreach ( $args as $name => $value ) {
1207 if ( $first ) {
1208 $first = false;
1209 } else {
1210 $s .= ', ';
1211 }
1212 $s .= "\"$name\":\"" .
1213 str_replace( '"', '\\"', $value->ownerDocument->saveXML( $value ) ) . '"';
1214 }
1215 $s .= '}';
1216 return $s;
1217 }
1218 /**
1219 * Returns true if there are no arguments in this frame
1220 */
1221 function isEmpty() {
1222 return !count( $this->numberedArgs ) && !count( $this->namedArgs );
1223 }
1224
1225 function getArguments() {
1226 $arguments = array();
1227 foreach ( array_merge(
1228 array_keys($this->numberedArgs),
1229 array_keys($this->namedArgs)) as $key ) {
1230 $arguments[$key] = $this->getArgument($key);
1231 }
1232 return $arguments;
1233 }
1234
1235 function getNumberedArguments() {
1236 $arguments = array();
1237 foreach ( array_keys($this->numberedArgs) as $key ) {
1238 $arguments[$key] = $this->getArgument($key);
1239 }
1240 return $arguments;
1241 }
1242
1243 function getNamedArguments() {
1244 $arguments = array();
1245 foreach ( array_keys($this->namedArgs) as $key ) {
1246 $arguments[$key] = $this->getArgument($key);
1247 }
1248 return $arguments;
1249 }
1250
1251 function getNumberedArgument( $index ) {
1252 if ( !isset( $this->numberedArgs[$index] ) ) {
1253 return false;
1254 }
1255 if ( !isset( $this->numberedExpansionCache[$index] ) ) {
1256 # No trimming for unnamed arguments
1257 $this->numberedExpansionCache[$index] = $this->parent->expand( $this->numberedArgs[$index], self::STRIP_COMMENTS );
1258 }
1259 return $this->numberedExpansionCache[$index];
1260 }
1261
1262 function getNamedArgument( $name ) {
1263 if ( !isset( $this->namedArgs[$name] ) ) {
1264 return false;
1265 }
1266 if ( !isset( $this->namedExpansionCache[$name] ) ) {
1267 # Trim named arguments post-expand, for backwards compatibility
1268 $this->namedExpansionCache[$name] = trim(
1269 $this->parent->expand( $this->namedArgs[$name], self::STRIP_COMMENTS ) );
1270 }
1271 return $this->namedExpansionCache[$name];
1272 }
1273
1274 function getArgument( $name ) {
1275 $text = $this->getNumberedArgument( $name );
1276 if ( $text === false ) {
1277 $text = $this->getNamedArgument( $name );
1278 }
1279 return $text;
1280 }
1281
1282 /**
1283 * Return true if the frame is a template frame
1284 */
1285 function isTemplate() {
1286 return true;
1287 }
1288 }
1289
1290 /**
1291 * Expansion frame with custom arguments
1292 * @ingroup Parser
1293 */
1294 class PPCustomFrame_DOM extends PPFrame_DOM {
1295 var $args;
1296
1297 function __construct( $preprocessor, $args ) {
1298 $this->preprocessor = $preprocessor;
1299 $this->parser = $preprocessor->parser;
1300 $this->args = $args;
1301 }
1302
1303 function __toString() {
1304 $s = 'cstmframe{';
1305 $first = true;
1306 foreach ( $this->args as $name => $value ) {
1307 if ( $first ) {
1308 $first = false;
1309 } else {
1310 $s .= ', ';
1311 }
1312 $s .= "\"$name\":\"" .
1313 str_replace( '"', '\\"', $value->__toString() ) . '"';
1314 }
1315 $s .= '}';
1316 return $s;
1317 }
1318
1319 function isEmpty() {
1320 return !count( $this->args );
1321 }
1322
1323 function getArgument( $index ) {
1324 if ( !isset( $this->args[$index] ) ) {
1325 return false;
1326 }
1327 return $this->args[$index];
1328 }
1329 }
1330
1331 /**
1332 * @ingroup Parser
1333 */
1334 class PPNode_DOM implements PPNode {
1335 var $node;
1336
1337 function __construct( $node, $xpath = false ) {
1338 $this->node = $node;
1339 }
1340
1341 function __get( $name ) {
1342 if ( $name == 'xpath' ) {
1343 $this->xpath = new DOMXPath( $this->node->ownerDocument );
1344 }
1345 return $this->xpath;
1346 }
1347
1348 function __toString() {
1349 if ( $this->node instanceof DOMNodeList ) {
1350 $s = '';
1351 foreach ( $this->node as $node ) {
1352 $s .= $node->ownerDocument->saveXML( $node );
1353 }
1354 } else {
1355 $s = $this->node->ownerDocument->saveXML( $this->node );
1356 }
1357 return $s;
1358 }
1359
1360 function getChildren() {
1361 return $this->node->childNodes ? new self( $this->node->childNodes ) : false;
1362 }
1363
1364 function getFirstChild() {
1365 return $this->node->firstChild ? new self( $this->node->firstChild ) : false;
1366 }
1367
1368 function getNextSibling() {
1369 return $this->node->nextSibling ? new self( $this->node->nextSibling ) : false;
1370 }
1371
1372 function getChildrenOfType( $type ) {
1373 return new self( $this->xpath->query( $type, $this->node ) );
1374 }
1375
1376 function getLength() {
1377 if ( $this->node instanceof DOMNodeList ) {
1378 return $this->node->length;
1379 } else {
1380 return false;
1381 }
1382 }
1383
1384 function item( $i ) {
1385 $item = $this->node->item( $i );
1386 return $item ? new self( $item ) : false;
1387 }
1388
1389 function getName() {
1390 if ( $this->node instanceof DOMNodeList ) {
1391 return '#nodelist';
1392 } else {
1393 return $this->node->nodeName;
1394 }
1395 }
1396
1397 /**
1398 * Split a <part> node into an associative array containing:
1399 * name PPNode name
1400 * index String index
1401 * value PPNode value
1402 */
1403 function splitArg() {
1404 $names = $this->xpath->query( 'name', $this->node );
1405 $values = $this->xpath->query( 'value', $this->node );
1406 if ( !$names->length || !$values->length ) {
1407 throw new MWException( 'Invalid brace node passed to ' . __METHOD__ );
1408 }
1409 $name = $names->item( 0 );
1410 $index = $name->getAttribute( 'index' );
1411 return array(
1412 'name' => new self( $name ),
1413 'index' => $index,
1414 'value' => new self( $values->item( 0 ) ) );
1415 }
1416
1417 /**
1418 * Split an <ext> node into an associative array containing name, attr, inner and close
1419 * All values in the resulting array are PPNodes. Inner and close are optional.
1420 */
1421 function splitExt() {
1422 $names = $this->xpath->query( 'name', $this->node );
1423 $attrs = $this->xpath->query( 'attr', $this->node );
1424 $inners = $this->xpath->query( 'inner', $this->node );
1425 $closes = $this->xpath->query( 'close', $this->node );
1426 if ( !$names->length || !$attrs->length ) {
1427 throw new MWException( 'Invalid ext node passed to ' . __METHOD__ );
1428 }
1429 $parts = array(
1430 'name' => new self( $names->item( 0 ) ),
1431 'attr' => new self( $attrs->item( 0 ) ) );
1432 if ( $inners->length ) {
1433 $parts['inner'] = new self( $inners->item( 0 ) );
1434 }
1435 if ( $closes->length ) {
1436 $parts['close'] = new self( $closes->item( 0 ) );
1437 }
1438 return $parts;
1439 }
1440
1441 /**
1442 * Split a <h> node
1443 */
1444 function splitHeading() {
1445 if ( !$this->nodeName == 'h' ) {
1446 throw new MWException( 'Invalid h node passed to ' . __METHOD__ );
1447 }
1448 return array(
1449 'i' => $this->node->getAttribute( 'i' ),
1450 'level' => $this->node->getAttribute( 'level' ),
1451 'contents' => $this->getChildren()
1452 );
1453 }
1454 }