Expose $wgMaxArticleSize in siteinfo query api
[mediawiki.git] / includes / parser / Preprocessor_DOM.php
bloba28c0aa439d303162bb2c9be23044cd7b6b17aa7
1 <?php
2 /**
3 * Preprocessor using PHP's dom extension
5 * This program is free software; you can redistribute it and/or modify
6 * it under the terms of the GNU General Public License as published by
7 * the Free Software Foundation; either version 2 of the License, or
8 * (at your option) any later version.
10 * This program is distributed in the hope that it will be useful,
11 * but WITHOUT ANY WARRANTY; without even the implied warranty of
12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13 * GNU General Public License for more details.
15 * You should have received a copy of the GNU General Public License along
16 * with this program; if not, write to the Free Software Foundation, Inc.,
17 * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
18 * http://www.gnu.org/copyleft/gpl.html
20 * @file
21 * @ingroup Parser
24 /**
25 * @ingroup Parser
27 // @codingStandardsIgnoreStart Squiz.Classes.ValidClassName.NotCamelCaps
28 class Preprocessor_DOM extends Preprocessor {
29 // @codingStandardsIgnoreEnd
31 /**
32 * @var Parser
34 public $parser;
36 public $memoryLimit;
38 const CACHE_PREFIX = 'preprocess-xml';
40 public function __construct( $parser ) {
41 $this->parser = $parser;
42 $mem = ini_get( 'memory_limit' );
43 $this->memoryLimit = false;
44 if ( strval( $mem ) !== '' && $mem != -1 ) {
45 if ( preg_match( '/^\d+$/', $mem ) ) {
46 $this->memoryLimit = $mem;
47 } elseif ( preg_match( '/^(\d+)M$/i', $mem, $m ) ) {
48 $this->memoryLimit = $m[1] * 1048576;
53 /**
54 * @return PPFrame_DOM
56 public function newFrame() {
57 return new PPFrame_DOM( $this );
60 /**
61 * @param array $args
62 * @return PPCustomFrame_DOM
64 public function newCustomFrame( $args ) {
65 return new PPCustomFrame_DOM( $this, $args );
68 /**
69 * @param array $values
70 * @return PPNode_DOM
71 * @throws MWException
73 public function newPartNodeArray( $values ) {
74 // NOTE: DOM manipulation is slower than building & parsing XML! (or so Tim sais)
75 $xml = "<list>";
77 foreach ( $values as $k => $val ) {
78 if ( is_int( $k ) ) {
79 $xml .= "<part><name index=\"$k\"/><value>"
80 . htmlspecialchars( $val ) . "</value></part>";
81 } else {
82 $xml .= "<part><name>" . htmlspecialchars( $k )
83 . "</name>=<value>" . htmlspecialchars( $val ) . "</value></part>";
87 $xml .= "</list>";
89 $dom = new DOMDocument();
90 MediaWiki\suppressWarnings();
91 $result = $dom->loadXML( $xml );
92 MediaWiki\restoreWarnings();
93 if ( !$result ) {
94 // Try running the XML through UtfNormal to get rid of invalid characters
95 $xml = UtfNormal\Validator::cleanUp( $xml );
96 // 1 << 19 == XML_PARSE_HUGE, needed so newer versions of libxml2
97 // don't barf when the XML is >256 levels deep
98 $result = $dom->loadXML( $xml, 1 << 19 );
101 if ( !$result ) {
102 throw new MWException( 'Parameters passed to ' . __METHOD__ . ' result in invalid XML' );
105 $root = $dom->documentElement;
106 $node = new PPNode_DOM( $root->childNodes );
107 return $node;
111 * @throws MWException
112 * @return bool
114 public function memCheck() {
115 if ( $this->memoryLimit === false ) {
116 return true;
118 $usage = memory_get_usage();
119 if ( $usage > $this->memoryLimit * 0.9 ) {
120 $limit = intval( $this->memoryLimit * 0.9 / 1048576 + 0.5 );
121 throw new MWException( "Preprocessor hit 90% memory limit ($limit MB)" );
123 return $usage <= $this->memoryLimit * 0.8;
127 * Preprocess some wikitext and return the document tree.
128 * This is the ghost of Parser::replace_variables().
130 * @param string $text The text to parse
131 * @param int $flags Bitwise combination of:
132 * Parser::PTD_FOR_INCLUSION Handle "<noinclude>" and "<includeonly>"
133 * as if the text is being included. Default
134 * is to assume a direct page view.
136 * The generated DOM tree must depend only on the input text and the flags.
137 * The DOM tree must be the same in OT_HTML and OT_WIKI mode, to avoid a regression of bug 4899.
139 * Any flag added to the $flags parameter here, or any other parameter liable to cause a
140 * change in the DOM tree for a given text, must be passed through the section identifier
141 * in the section edit link and thus back to extractSections().
143 * The output of this function is currently only cached in process memory, but a persistent
144 * cache may be implemented at a later date which takes further advantage of these strict
145 * dependency requirements.
147 * @throws MWException
148 * @return PPNode_DOM
150 public function preprocessToObj( $text, $flags = 0 ) {
152 $xml = $this->cacheGetTree( $text, $flags );
153 if ( $xml === false ) {
154 $xml = $this->preprocessToXml( $text, $flags );
155 $this->cacheSetTree( $text, $flags, $xml );
158 // Fail if the number of elements exceeds acceptable limits
159 // Do not attempt to generate the DOM
160 $this->parser->mGeneratedPPNodeCount += substr_count( $xml, '<' );
161 $max = $this->parser->mOptions->getMaxGeneratedPPNodeCount();
162 if ( $this->parser->mGeneratedPPNodeCount > $max ) {
163 // if ( $cacheable ) { ... }
164 throw new MWException( __METHOD__ . ': generated node count limit exceeded' );
167 $dom = new DOMDocument;
168 MediaWiki\suppressWarnings();
169 $result = $dom->loadXML( $xml );
170 MediaWiki\restoreWarnings();
171 if ( !$result ) {
172 // Try running the XML through UtfNormal to get rid of invalid characters
173 $xml = UtfNormal\Validator::cleanUp( $xml );
174 // 1 << 19 == XML_PARSE_HUGE, needed so newer versions of libxml2
175 // don't barf when the XML is >256 levels deep.
176 $result = $dom->loadXML( $xml, 1 << 19 );
178 if ( $result ) {
179 $obj = new PPNode_DOM( $dom->documentElement );
182 // if ( $cacheable ) { ... }
184 if ( !$result ) {
185 throw new MWException( __METHOD__ . ' generated invalid XML' );
187 return $obj;
191 * @param string $text
192 * @param int $flags
193 * @return string
195 public function preprocessToXml( $text, $flags = 0 ) {
196 $forInclusion = $flags & Parser::PTD_FOR_INCLUSION;
198 $xmlishElements = $this->parser->getStripList();
199 $xmlishAllowMissingEndTag = [ 'includeonly', 'noinclude', 'onlyinclude' ];
200 $enableOnlyinclude = false;
201 if ( $forInclusion ) {
202 $ignoredTags = [ 'includeonly', '/includeonly' ];
203 $ignoredElements = [ 'noinclude' ];
204 $xmlishElements[] = 'noinclude';
205 if ( strpos( $text, '<onlyinclude>' ) !== false
206 && strpos( $text, '</onlyinclude>' ) !== false
208 $enableOnlyinclude = true;
210 } else {
211 $ignoredTags = [ 'noinclude', '/noinclude', 'onlyinclude', '/onlyinclude' ];
212 $ignoredElements = [ 'includeonly' ];
213 $xmlishElements[] = 'includeonly';
215 $xmlishRegex = implode( '|', array_merge( $xmlishElements, $ignoredTags ) );
217 // Use "A" modifier (anchored) instead of "^", because ^ doesn't work with an offset
218 $elementsRegex = "~($xmlishRegex)(?:\s|\/>|>)|(!--)~iA";
220 $stack = new PPDStack;
222 $searchBase = "[{<\n"; # }
223 // For fast reverse searches
224 $revText = strrev( $text );
225 $lengthText = strlen( $text );
227 // Input pointer, starts out pointing to a pseudo-newline before the start
228 $i = 0;
229 // Current accumulator
230 $accum =& $stack->getAccum();
231 $accum = '<root>';
232 // True to find equals signs in arguments
233 $findEquals = false;
234 // True to take notice of pipe characters
235 $findPipe = false;
236 $headingIndex = 1;
237 // True if $i is inside a possible heading
238 $inHeading = false;
239 // True if there are no more greater-than (>) signs right of $i
240 $noMoreGT = false;
241 // Map of tag name => true if there are no more closing tags of given type right of $i
242 $noMoreClosingTag = [];
243 // True to ignore all input up to the next <onlyinclude>
244 $findOnlyinclude = $enableOnlyinclude;
245 // Do a line-start run without outputting an LF character
246 $fakeLineStart = true;
248 while ( true ) {
249 // $this->memCheck();
251 if ( $findOnlyinclude ) {
252 // Ignore all input up to the next <onlyinclude>
253 $startPos = strpos( $text, '<onlyinclude>', $i );
254 if ( $startPos === false ) {
255 // Ignored section runs to the end
256 $accum .= '<ignore>' . htmlspecialchars( substr( $text, $i ) ) . '</ignore>';
257 break;
259 $tagEndPos = $startPos + strlen( '<onlyinclude>' ); // past-the-end
260 $accum .= '<ignore>' . htmlspecialchars( substr( $text, $i, $tagEndPos - $i ) ) . '</ignore>';
261 $i = $tagEndPos;
262 $findOnlyinclude = false;
265 if ( $fakeLineStart ) {
266 $found = 'line-start';
267 $curChar = '';
268 } else {
269 # Find next opening brace, closing brace or pipe
270 $search = $searchBase;
271 if ( $stack->top === false ) {
272 $currentClosing = '';
273 } else {
274 $currentClosing = $stack->top->close;
275 $search .= $currentClosing;
277 if ( $findPipe ) {
278 $search .= '|';
280 if ( $findEquals ) {
281 // First equals will be for the template
282 $search .= '=';
284 $rule = null;
285 # Output literal section, advance input counter
286 $literalLength = strcspn( $text, $search, $i );
287 if ( $literalLength > 0 ) {
288 $accum .= htmlspecialchars( substr( $text, $i, $literalLength ) );
289 $i += $literalLength;
291 if ( $i >= $lengthText ) {
292 if ( $currentClosing == "\n" ) {
293 // Do a past-the-end run to finish off the heading
294 $curChar = '';
295 $found = 'line-end';
296 } else {
297 # All done
298 break;
300 } else {
301 $curChar = $text[$i];
302 if ( $curChar == '|' ) {
303 $found = 'pipe';
304 } elseif ( $curChar == '=' ) {
305 $found = 'equals';
306 } elseif ( $curChar == '<' ) {
307 $found = 'angle';
308 } elseif ( $curChar == "\n" ) {
309 if ( $inHeading ) {
310 $found = 'line-end';
311 } else {
312 $found = 'line-start';
314 } elseif ( $curChar == $currentClosing ) {
315 $found = 'close';
316 } elseif ( isset( $this->rules[$curChar] ) ) {
317 $found = 'open';
318 $rule = $this->rules[$curChar];
319 } else {
320 # Some versions of PHP have a strcspn which stops on null characters
321 # Ignore and continue
322 ++$i;
323 continue;
328 if ( $found == 'angle' ) {
329 $matches = false;
330 // Handle </onlyinclude>
331 if ( $enableOnlyinclude
332 && substr( $text, $i, strlen( '</onlyinclude>' ) ) == '</onlyinclude>'
334 $findOnlyinclude = true;
335 continue;
338 // Determine element name
339 if ( !preg_match( $elementsRegex, $text, $matches, 0, $i + 1 ) ) {
340 // Element name missing or not listed
341 $accum .= '&lt;';
342 ++$i;
343 continue;
345 // Handle comments
346 if ( isset( $matches[2] ) && $matches[2] == '!--' ) {
348 // To avoid leaving blank lines, when a sequence of
349 // space-separated comments is both preceded and followed by
350 // a newline (ignoring spaces), then
351 // trim leading and trailing spaces and the trailing newline.
353 // Find the end
354 $endPos = strpos( $text, '-->', $i + 4 );
355 if ( $endPos === false ) {
356 // Unclosed comment in input, runs to end
357 $inner = substr( $text, $i );
358 $accum .= '<comment>' . htmlspecialchars( $inner ) . '</comment>';
359 $i = $lengthText;
360 } else {
361 // Search backwards for leading whitespace
362 $wsStart = $i ? ( $i - strspn( $revText, " \t", $lengthText - $i ) ) : 0;
364 // Search forwards for trailing whitespace
365 // $wsEnd will be the position of the last space (or the '>' if there's none)
366 $wsEnd = $endPos + 2 + strspn( $text, " \t", $endPos + 3 );
368 // Keep looking forward as long as we're finding more
369 // comments.
370 $comments = [ [ $wsStart, $wsEnd ] ];
371 while ( substr( $text, $wsEnd + 1, 4 ) == '<!--' ) {
372 $c = strpos( $text, '-->', $wsEnd + 4 );
373 if ( $c === false ) {
374 break;
376 $c = $c + 2 + strspn( $text, " \t", $c + 3 );
377 $comments[] = [ $wsEnd + 1, $c ];
378 $wsEnd = $c;
381 // Eat the line if possible
382 // TODO: This could theoretically be done if $wsStart == 0, i.e. for comments at
383 // the overall start. That's not how Sanitizer::removeHTMLcomments() did it, but
384 // it's a possible beneficial b/c break.
385 if ( $wsStart > 0 && substr( $text, $wsStart - 1, 1 ) == "\n"
386 && substr( $text, $wsEnd + 1, 1 ) == "\n"
388 // Remove leading whitespace from the end of the accumulator
389 // Sanity check first though
390 $wsLength = $i - $wsStart;
391 if ( $wsLength > 0
392 && strspn( $accum, " \t", -$wsLength ) === $wsLength
394 $accum = substr( $accum, 0, -$wsLength );
397 // Dump all but the last comment to the accumulator
398 foreach ( $comments as $j => $com ) {
399 $startPos = $com[0];
400 $endPos = $com[1] + 1;
401 if ( $j == ( count( $comments ) - 1 ) ) {
402 break;
404 $inner = substr( $text, $startPos, $endPos - $startPos );
405 $accum .= '<comment>' . htmlspecialchars( $inner ) . '</comment>';
408 // Do a line-start run next time to look for headings after the comment
409 $fakeLineStart = true;
410 } else {
411 // No line to eat, just take the comment itself
412 $startPos = $i;
413 $endPos += 2;
416 if ( $stack->top ) {
417 $part = $stack->top->getCurrentPart();
418 if ( !( isset( $part->commentEnd ) && $part->commentEnd == $wsStart - 1 ) ) {
419 $part->visualEnd = $wsStart;
421 // Else comments abutting, no change in visual end
422 $part->commentEnd = $endPos;
424 $i = $endPos + 1;
425 $inner = substr( $text, $startPos, $endPos - $startPos + 1 );
426 $accum .= '<comment>' . htmlspecialchars( $inner ) . '</comment>';
428 continue;
430 $name = $matches[1];
431 $lowerName = strtolower( $name );
432 $attrStart = $i + strlen( $name ) + 1;
434 // Find end of tag
435 $tagEndPos = $noMoreGT ? false : strpos( $text, '>', $attrStart );
436 if ( $tagEndPos === false ) {
437 // Infinite backtrack
438 // Disable tag search to prevent worst-case O(N^2) performance
439 $noMoreGT = true;
440 $accum .= '&lt;';
441 ++$i;
442 continue;
445 // Handle ignored tags
446 if ( in_array( $lowerName, $ignoredTags ) ) {
447 $accum .= '<ignore>'
448 . htmlspecialchars( substr( $text, $i, $tagEndPos - $i + 1 ) )
449 . '</ignore>';
450 $i = $tagEndPos + 1;
451 continue;
454 $tagStartPos = $i;
455 if ( $text[$tagEndPos - 1] == '/' ) {
456 $attrEnd = $tagEndPos - 1;
457 $inner = null;
458 $i = $tagEndPos + 1;
459 $close = '';
460 } else {
461 $attrEnd = $tagEndPos;
462 // Find closing tag
463 if (
464 !isset( $noMoreClosingTag[$name] ) &&
465 preg_match( "/<\/" . preg_quote( $name, '/' ) . "\s*>/i",
466 $text, $matches, PREG_OFFSET_CAPTURE, $tagEndPos + 1 )
468 $inner = substr( $text, $tagEndPos + 1, $matches[0][1] - $tagEndPos - 1 );
469 $i = $matches[0][1] + strlen( $matches[0][0] );
470 $close = '<close>' . htmlspecialchars( $matches[0][0] ) . '</close>';
471 } else {
472 // No end tag
473 if ( in_array( $name, $xmlishAllowMissingEndTag ) ) {
474 // Let it run out to the end of the text.
475 $inner = substr( $text, $tagEndPos + 1 );
476 $i = $lengthText;
477 $close = '';
478 } else {
479 // Don't match the tag, treat opening tag as literal and resume parsing.
480 $i = $tagEndPos + 1;
481 $accum .= htmlspecialchars( substr( $text, $tagStartPos, $tagEndPos + 1 - $tagStartPos ) );
482 // Cache results, otherwise we have O(N^2) performance for input like <foo><foo><foo>...
483 $noMoreClosingTag[$name] = true;
484 continue;
488 // <includeonly> and <noinclude> just become <ignore> tags
489 if ( in_array( $lowerName, $ignoredElements ) ) {
490 $accum .= '<ignore>' . htmlspecialchars( substr( $text, $tagStartPos, $i - $tagStartPos ) )
491 . '</ignore>';
492 continue;
495 $accum .= '<ext>';
496 if ( $attrEnd <= $attrStart ) {
497 $attr = '';
498 } else {
499 $attr = substr( $text, $attrStart, $attrEnd - $attrStart );
501 $accum .= '<name>' . htmlspecialchars( $name ) . '</name>' .
502 // Note that the attr element contains the whitespace between name and attribute,
503 // this is necessary for precise reconstruction during pre-save transform.
504 '<attr>' . htmlspecialchars( $attr ) . '</attr>';
505 if ( $inner !== null ) {
506 $accum .= '<inner>' . htmlspecialchars( $inner ) . '</inner>';
508 $accum .= $close . '</ext>';
509 } elseif ( $found == 'line-start' ) {
510 // Is this the start of a heading?
511 // Line break belongs before the heading element in any case
512 if ( $fakeLineStart ) {
513 $fakeLineStart = false;
514 } else {
515 $accum .= $curChar;
516 $i++;
519 $count = strspn( $text, '=', $i, 6 );
520 if ( $count == 1 && $findEquals ) {
521 // DWIM: This looks kind of like a name/value separator.
522 // Let's let the equals handler have it and break the
523 // potential heading. This is heuristic, but AFAICT the
524 // methods for completely correct disambiguation are very
525 // complex.
526 } elseif ( $count > 0 ) {
527 $piece = [
528 'open' => "\n",
529 'close' => "\n",
530 'parts' => [ new PPDPart( str_repeat( '=', $count ) ) ],
531 'startPos' => $i,
532 'count' => $count ];
533 $stack->push( $piece );
534 $accum =& $stack->getAccum();
535 $flags = $stack->getFlags();
536 extract( $flags );
537 $i += $count;
539 } elseif ( $found == 'line-end' ) {
540 $piece = $stack->top;
541 // A heading must be open, otherwise \n wouldn't have been in the search list
542 assert( '$piece->open == "\n"' );
543 $part = $piece->getCurrentPart();
544 // Search back through the input to see if it has a proper close.
545 // Do this using the reversed string since the other solutions
546 // (end anchor, etc.) are inefficient.
547 $wsLength = strspn( $revText, " \t", $lengthText - $i );
548 $searchStart = $i - $wsLength;
549 if ( isset( $part->commentEnd ) && $searchStart - 1 == $part->commentEnd ) {
550 // Comment found at line end
551 // Search for equals signs before the comment
552 $searchStart = $part->visualEnd;
553 $searchStart -= strspn( $revText, " \t", $lengthText - $searchStart );
555 $count = $piece->count;
556 $equalsLength = strspn( $revText, '=', $lengthText - $searchStart );
557 if ( $equalsLength > 0 ) {
558 if ( $searchStart - $equalsLength == $piece->startPos ) {
559 // This is just a single string of equals signs on its own line
560 // Replicate the doHeadings behavior /={count}(.+)={count}/
561 // First find out how many equals signs there really are (don't stop at 6)
562 $count = $equalsLength;
563 if ( $count < 3 ) {
564 $count = 0;
565 } else {
566 $count = min( 6, intval( ( $count - 1 ) / 2 ) );
568 } else {
569 $count = min( $equalsLength, $count );
571 if ( $count > 0 ) {
572 // Normal match, output <h>
573 $element = "<h level=\"$count\" i=\"$headingIndex\">$accum</h>";
574 $headingIndex++;
575 } else {
576 // Single equals sign on its own line, count=0
577 $element = $accum;
579 } else {
580 // No match, no <h>, just pass down the inner text
581 $element = $accum;
583 // Unwind the stack
584 $stack->pop();
585 $accum =& $stack->getAccum();
586 $flags = $stack->getFlags();
587 extract( $flags );
589 // Append the result to the enclosing accumulator
590 $accum .= $element;
591 // Note that we do NOT increment the input pointer.
592 // This is because the closing linebreak could be the opening linebreak of
593 // another heading. Infinite loops are avoided because the next iteration MUST
594 // hit the heading open case above, which unconditionally increments the
595 // input pointer.
596 } elseif ( $found == 'open' ) {
597 # count opening brace characters
598 $count = strspn( $text, $curChar, $i );
600 # we need to add to stack only if opening brace count is enough for one of the rules
601 if ( $count >= $rule['min'] ) {
602 # Add it to the stack
603 $piece = [
604 'open' => $curChar,
605 'close' => $rule['end'],
606 'count' => $count,
607 'lineStart' => ( $i > 0 && $text[$i - 1] == "\n" ),
610 $stack->push( $piece );
611 $accum =& $stack->getAccum();
612 $flags = $stack->getFlags();
613 extract( $flags );
614 } else {
615 # Add literal brace(s)
616 $accum .= htmlspecialchars( str_repeat( $curChar, $count ) );
618 $i += $count;
619 } elseif ( $found == 'close' ) {
620 $piece = $stack->top;
621 # lets check if there are enough characters for closing brace
622 $maxCount = $piece->count;
623 $count = strspn( $text, $curChar, $i, $maxCount );
625 # check for maximum matching characters (if there are 5 closing
626 # characters, we will probably need only 3 - depending on the rules)
627 $rule = $this->rules[$piece->open];
628 if ( $count > $rule['max'] ) {
629 # The specified maximum exists in the callback array, unless the caller
630 # has made an error
631 $matchingCount = $rule['max'];
632 } else {
633 # Count is less than the maximum
634 # Skip any gaps in the callback array to find the true largest match
635 # Need to use array_key_exists not isset because the callback can be null
636 $matchingCount = $count;
637 while ( $matchingCount > 0 && !array_key_exists( $matchingCount, $rule['names'] ) ) {
638 --$matchingCount;
642 if ( $matchingCount <= 0 ) {
643 # No matching element found in callback array
644 # Output a literal closing brace and continue
645 $accum .= htmlspecialchars( str_repeat( $curChar, $count ) );
646 $i += $count;
647 continue;
649 $name = $rule['names'][$matchingCount];
650 if ( $name === null ) {
651 // No element, just literal text
652 $element = $piece->breakSyntax( $matchingCount ) . str_repeat( $rule['end'], $matchingCount );
653 } else {
654 # Create XML element
655 # Note: $parts is already XML, does not need to be encoded further
656 $parts = $piece->parts;
657 $title = $parts[0]->out;
658 unset( $parts[0] );
660 # The invocation is at the start of the line if lineStart is set in
661 # the stack, and all opening brackets are used up.
662 if ( $maxCount == $matchingCount && !empty( $piece->lineStart ) ) {
663 $attr = ' lineStart="1"';
664 } else {
665 $attr = '';
668 $element = "<$name$attr>";
669 $element .= "<title>$title</title>";
670 $argIndex = 1;
671 foreach ( $parts as $part ) {
672 if ( isset( $part->eqpos ) ) {
673 $argName = substr( $part->out, 0, $part->eqpos );
674 $argValue = substr( $part->out, $part->eqpos + 1 );
675 $element .= "<part><name>$argName</name>=<value>$argValue</value></part>";
676 } else {
677 $element .= "<part><name index=\"$argIndex\" /><value>{$part->out}</value></part>";
678 $argIndex++;
681 $element .= "</$name>";
684 # Advance input pointer
685 $i += $matchingCount;
687 # Unwind the stack
688 $stack->pop();
689 $accum =& $stack->getAccum();
691 # Re-add the old stack element if it still has unmatched opening characters remaining
692 if ( $matchingCount < $piece->count ) {
693 $piece->parts = [ new PPDPart ];
694 $piece->count -= $matchingCount;
695 # do we still qualify for any callback with remaining count?
696 $min = $this->rules[$piece->open]['min'];
697 if ( $piece->count >= $min ) {
698 $stack->push( $piece );
699 $accum =& $stack->getAccum();
700 } else {
701 $accum .= str_repeat( $piece->open, $piece->count );
704 $flags = $stack->getFlags();
705 extract( $flags );
707 # Add XML element to the enclosing accumulator
708 $accum .= $element;
709 } elseif ( $found == 'pipe' ) {
710 $findEquals = true; // shortcut for getFlags()
711 $stack->addPart();
712 $accum =& $stack->getAccum();
713 ++$i;
714 } elseif ( $found == 'equals' ) {
715 $findEquals = false; // shortcut for getFlags()
716 $stack->getCurrentPart()->eqpos = strlen( $accum );
717 $accum .= '=';
718 ++$i;
722 # Output any remaining unclosed brackets
723 foreach ( $stack->stack as $piece ) {
724 $stack->rootAccum .= $piece->breakSyntax();
726 $stack->rootAccum .= '</root>';
727 $xml = $stack->rootAccum;
729 return $xml;
734 * Stack class to help Preprocessor::preprocessToObj()
735 * @ingroup Parser
737 class PPDStack {
738 public $stack, $rootAccum;
741 * @var PPDStack
743 public $top;
744 public $out;
745 public $elementClass = 'PPDStackElement';
747 public static $false = false;
749 public function __construct() {
750 $this->stack = [];
751 $this->top = false;
752 $this->rootAccum = '';
753 $this->accum =& $this->rootAccum;
757 * @return int
759 public function count() {
760 return count( $this->stack );
763 public function &getAccum() {
764 return $this->accum;
767 public function getCurrentPart() {
768 if ( $this->top === false ) {
769 return false;
770 } else {
771 return $this->top->getCurrentPart();
775 public function push( $data ) {
776 if ( $data instanceof $this->elementClass ) {
777 $this->stack[] = $data;
778 } else {
779 $class = $this->elementClass;
780 $this->stack[] = new $class( $data );
782 $this->top = $this->stack[count( $this->stack ) - 1];
783 $this->accum =& $this->top->getAccum();
786 public function pop() {
787 if ( !count( $this->stack ) ) {
788 throw new MWException( __METHOD__ . ': no elements remaining' );
790 $temp = array_pop( $this->stack );
792 if ( count( $this->stack ) ) {
793 $this->top = $this->stack[count( $this->stack ) - 1];
794 $this->accum =& $this->top->getAccum();
795 } else {
796 $this->top = self::$false;
797 $this->accum =& $this->rootAccum;
799 return $temp;
802 public function addPart( $s = '' ) {
803 $this->top->addPart( $s );
804 $this->accum =& $this->top->getAccum();
808 * @return array
810 public function getFlags() {
811 if ( !count( $this->stack ) ) {
812 return [
813 'findEquals' => false,
814 'findPipe' => false,
815 'inHeading' => false,
817 } else {
818 return $this->top->getFlags();
824 * @ingroup Parser
826 class PPDStackElement {
828 * @var string Opening character (\n for heading)
830 public $open;
833 * @var string Matching closing character
835 public $close;
838 * @var int Number of opening characters found (number of "=" for heading)
840 public $count;
843 * @var PPDPart[] Array of PPDPart objects describing pipe-separated parts.
845 public $parts;
848 * @var bool True if the open char appeared at the start of the input line.
849 * Not set for headings.
851 public $lineStart;
853 public $partClass = 'PPDPart';
855 public function __construct( $data = [] ) {
856 $class = $this->partClass;
857 $this->parts = [ new $class ];
859 foreach ( $data as $name => $value ) {
860 $this->$name = $value;
864 public function &getAccum() {
865 return $this->parts[count( $this->parts ) - 1]->out;
868 public function addPart( $s = '' ) {
869 $class = $this->partClass;
870 $this->parts[] = new $class( $s );
873 public function getCurrentPart() {
874 return $this->parts[count( $this->parts ) - 1];
878 * @return array
880 public function getFlags() {
881 $partCount = count( $this->parts );
882 $findPipe = $this->open != "\n" && $this->open != '[';
883 return [
884 'findPipe' => $findPipe,
885 'findEquals' => $findPipe && $partCount > 1 && !isset( $this->parts[$partCount - 1]->eqpos ),
886 'inHeading' => $this->open == "\n",
891 * Get the output string that would result if the close is not found.
893 * @param bool|int $openingCount
894 * @return string
896 public function breakSyntax( $openingCount = false ) {
897 if ( $this->open == "\n" ) {
898 $s = $this->parts[0]->out;
899 } else {
900 if ( $openingCount === false ) {
901 $openingCount = $this->count;
903 $s = str_repeat( $this->open, $openingCount );
904 $first = true;
905 foreach ( $this->parts as $part ) {
906 if ( $first ) {
907 $first = false;
908 } else {
909 $s .= '|';
911 $s .= $part->out;
914 return $s;
919 * @ingroup Parser
921 class PPDPart {
923 * @var string Output accumulator string
925 public $out;
927 // Optional member variables:
928 // eqpos Position of equals sign in output accumulator
929 // commentEnd Past-the-end input pointer for the last comment encountered
930 // visualEnd Past-the-end input pointer for the end of the accumulator minus comments
932 public function __construct( $out = '' ) {
933 $this->out = $out;
938 * An expansion frame, used as a context to expand the result of preprocessToObj()
939 * @ingroup Parser
941 // @codingStandardsIgnoreStart Squiz.Classes.ValidClassName.NotCamelCaps
942 class PPFrame_DOM implements PPFrame {
943 // @codingStandardsIgnoreEnd
946 * @var Preprocessor
948 public $preprocessor;
951 * @var Parser
953 public $parser;
956 * @var Title
958 public $title;
959 public $titleCache;
962 * Hashtable listing templates which are disallowed for expansion in this frame,
963 * having been encountered previously in parent frames.
965 public $loopCheckHash;
968 * Recursion depth of this frame, top = 0
969 * Note that this is NOT the same as expansion depth in expand()
971 public $depth;
973 private $volatile = false;
974 private $ttl = null;
977 * @var array
979 protected $childExpansionCache;
982 * Construct a new preprocessor frame.
983 * @param Preprocessor $preprocessor The parent preprocessor
985 public function __construct( $preprocessor ) {
986 $this->preprocessor = $preprocessor;
987 $this->parser = $preprocessor->parser;
988 $this->title = $this->parser->mTitle;
989 $this->titleCache = [ $this->title ? $this->title->getPrefixedDBkey() : false ];
990 $this->loopCheckHash = [];
991 $this->depth = 0;
992 $this->childExpansionCache = [];
996 * Create a new child frame
997 * $args is optionally a multi-root PPNode or array containing the template arguments
999 * @param bool|array $args
1000 * @param Title|bool $title
1001 * @param int $indexOffset
1002 * @return PPTemplateFrame_DOM
1004 public function newChild( $args = false, $title = false, $indexOffset = 0 ) {
1005 $namedArgs = [];
1006 $numberedArgs = [];
1007 if ( $title === false ) {
1008 $title = $this->title;
1010 if ( $args !== false ) {
1011 $xpath = false;
1012 if ( $args instanceof PPNode ) {
1013 $args = $args->node;
1015 foreach ( $args as $arg ) {
1016 if ( $arg instanceof PPNode ) {
1017 $arg = $arg->node;
1019 if ( !$xpath || $xpath->document !== $arg->ownerDocument ) {
1020 $xpath = new DOMXPath( $arg->ownerDocument );
1023 $nameNodes = $xpath->query( 'name', $arg );
1024 $value = $xpath->query( 'value', $arg );
1025 if ( $nameNodes->item( 0 )->hasAttributes() ) {
1026 // Numbered parameter
1027 $index = $nameNodes->item( 0 )->attributes->getNamedItem( 'index' )->textContent;
1028 $index = $index - $indexOffset;
1029 if ( isset( $namedArgs[$index] ) || isset( $numberedArgs[$index] ) ) {
1030 $this->parser->getOutput()->addWarning( wfMessage( 'duplicate-args-warning',
1031 wfEscapeWikiText( $this->title ),
1032 wfEscapeWikiText( $title ),
1033 wfEscapeWikiText( $index ) )->text() );
1034 $this->parser->addTrackingCategory( 'duplicate-args-category' );
1036 $numberedArgs[$index] = $value->item( 0 );
1037 unset( $namedArgs[$index] );
1038 } else {
1039 // Named parameter
1040 $name = trim( $this->expand( $nameNodes->item( 0 ), PPFrame::STRIP_COMMENTS ) );
1041 if ( isset( $namedArgs[$name] ) || isset( $numberedArgs[$name] ) ) {
1042 $this->parser->getOutput()->addWarning( wfMessage( 'duplicate-args-warning',
1043 wfEscapeWikiText( $this->title ),
1044 wfEscapeWikiText( $title ),
1045 wfEscapeWikiText( $name ) )->text() );
1046 $this->parser->addTrackingCategory( 'duplicate-args-category' );
1048 $namedArgs[$name] = $value->item( 0 );
1049 unset( $numberedArgs[$name] );
1053 return new PPTemplateFrame_DOM( $this->preprocessor, $this, $numberedArgs, $namedArgs, $title );
1057 * @throws MWException
1058 * @param string|int $key
1059 * @param string|PPNode_DOM|DOMDocument $root
1060 * @param int $flags
1061 * @return string
1063 public function cachedExpand( $key, $root, $flags = 0 ) {
1064 // we don't have a parent, so we don't have a cache
1065 return $this->expand( $root, $flags );
1069 * @throws MWException
1070 * @param string|PPNode_DOM|DOMDocument $root
1071 * @param int $flags
1072 * @return string
1074 public function expand( $root, $flags = 0 ) {
1075 static $expansionDepth = 0;
1076 if ( is_string( $root ) ) {
1077 return $root;
1080 if ( ++$this->parser->mPPNodeCount > $this->parser->mOptions->getMaxPPNodeCount() ) {
1081 $this->parser->limitationWarn( 'node-count-exceeded',
1082 $this->parser->mPPNodeCount,
1083 $this->parser->mOptions->getMaxPPNodeCount()
1085 return '<span class="error">Node-count limit exceeded</span>';
1088 if ( $expansionDepth > $this->parser->mOptions->getMaxPPExpandDepth() ) {
1089 $this->parser->limitationWarn( 'expansion-depth-exceeded',
1090 $expansionDepth,
1091 $this->parser->mOptions->getMaxPPExpandDepth()
1093 return '<span class="error">Expansion depth limit exceeded</span>';
1095 ++$expansionDepth;
1096 if ( $expansionDepth > $this->parser->mHighestExpansionDepth ) {
1097 $this->parser->mHighestExpansionDepth = $expansionDepth;
1100 if ( $root instanceof PPNode_DOM ) {
1101 $root = $root->node;
1103 if ( $root instanceof DOMDocument ) {
1104 $root = $root->documentElement;
1107 $outStack = [ '', '' ];
1108 $iteratorStack = [ false, $root ];
1109 $indexStack = [ 0, 0 ];
1111 while ( count( $iteratorStack ) > 1 ) {
1112 $level = count( $outStack ) - 1;
1113 $iteratorNode =& $iteratorStack[$level];
1114 $out =& $outStack[$level];
1115 $index =& $indexStack[$level];
1117 if ( $iteratorNode instanceof PPNode_DOM ) {
1118 $iteratorNode = $iteratorNode->node;
1121 if ( is_array( $iteratorNode ) ) {
1122 if ( $index >= count( $iteratorNode ) ) {
1123 // All done with this iterator
1124 $iteratorStack[$level] = false;
1125 $contextNode = false;
1126 } else {
1127 $contextNode = $iteratorNode[$index];
1128 $index++;
1130 } elseif ( $iteratorNode instanceof DOMNodeList ) {
1131 if ( $index >= $iteratorNode->length ) {
1132 // All done with this iterator
1133 $iteratorStack[$level] = false;
1134 $contextNode = false;
1135 } else {
1136 $contextNode = $iteratorNode->item( $index );
1137 $index++;
1139 } else {
1140 // Copy to $contextNode and then delete from iterator stack,
1141 // because this is not an iterator but we do have to execute it once
1142 $contextNode = $iteratorStack[$level];
1143 $iteratorStack[$level] = false;
1146 if ( $contextNode instanceof PPNode_DOM ) {
1147 $contextNode = $contextNode->node;
1150 $newIterator = false;
1152 if ( $contextNode === false ) {
1153 // nothing to do
1154 } elseif ( is_string( $contextNode ) ) {
1155 $out .= $contextNode;
1156 } elseif ( is_array( $contextNode ) || $contextNode instanceof DOMNodeList ) {
1157 $newIterator = $contextNode;
1158 } elseif ( $contextNode instanceof DOMNode ) {
1159 if ( $contextNode->nodeType == XML_TEXT_NODE ) {
1160 $out .= $contextNode->nodeValue;
1161 } elseif ( $contextNode->nodeName == 'template' ) {
1162 # Double-brace expansion
1163 $xpath = new DOMXPath( $contextNode->ownerDocument );
1164 $titles = $xpath->query( 'title', $contextNode );
1165 $title = $titles->item( 0 );
1166 $parts = $xpath->query( 'part', $contextNode );
1167 if ( $flags & PPFrame::NO_TEMPLATES ) {
1168 $newIterator = $this->virtualBracketedImplode( '{{', '|', '}}', $title, $parts );
1169 } else {
1170 $lineStart = $contextNode->getAttribute( 'lineStart' );
1171 $params = [
1172 'title' => new PPNode_DOM( $title ),
1173 'parts' => new PPNode_DOM( $parts ),
1174 'lineStart' => $lineStart ];
1175 $ret = $this->parser->braceSubstitution( $params, $this );
1176 if ( isset( $ret['object'] ) ) {
1177 $newIterator = $ret['object'];
1178 } else {
1179 $out .= $ret['text'];
1182 } elseif ( $contextNode->nodeName == 'tplarg' ) {
1183 # Triple-brace expansion
1184 $xpath = new DOMXPath( $contextNode->ownerDocument );
1185 $titles = $xpath->query( 'title', $contextNode );
1186 $title = $titles->item( 0 );
1187 $parts = $xpath->query( 'part', $contextNode );
1188 if ( $flags & PPFrame::NO_ARGS ) {
1189 $newIterator = $this->virtualBracketedImplode( '{{{', '|', '}}}', $title, $parts );
1190 } else {
1191 $params = [
1192 'title' => new PPNode_DOM( $title ),
1193 'parts' => new PPNode_DOM( $parts ) ];
1194 $ret = $this->parser->argSubstitution( $params, $this );
1195 if ( isset( $ret['object'] ) ) {
1196 $newIterator = $ret['object'];
1197 } else {
1198 $out .= $ret['text'];
1201 } elseif ( $contextNode->nodeName == 'comment' ) {
1202 # HTML-style comment
1203 # Remove it in HTML, pre+remove and STRIP_COMMENTS modes
1204 # Not in RECOVER_COMMENTS mode (msgnw) though.
1205 if ( ( $this->parser->ot['html']
1206 || ( $this->parser->ot['pre'] && $this->parser->mOptions->getRemoveComments() )
1207 || ( $flags & PPFrame::STRIP_COMMENTS )
1208 ) && !( $flags & PPFrame::RECOVER_COMMENTS )
1210 $out .= '';
1211 } elseif ( $this->parser->ot['wiki'] && !( $flags & PPFrame::RECOVER_COMMENTS ) ) {
1212 # Add a strip marker in PST mode so that pstPass2() can
1213 # run some old-fashioned regexes on the result.
1214 # Not in RECOVER_COMMENTS mode (extractSections) though.
1215 $out .= $this->parser->insertStripItem( $contextNode->textContent );
1216 } else {
1217 # Recover the literal comment in RECOVER_COMMENTS and pre+no-remove
1218 $out .= $contextNode->textContent;
1220 } elseif ( $contextNode->nodeName == 'ignore' ) {
1221 # Output suppression used by <includeonly> etc.
1222 # OT_WIKI will only respect <ignore> in substed templates.
1223 # The other output types respect it unless NO_IGNORE is set.
1224 # extractSections() sets NO_IGNORE and so never respects it.
1225 if ( ( !isset( $this->parent ) && $this->parser->ot['wiki'] )
1226 || ( $flags & PPFrame::NO_IGNORE )
1228 $out .= $contextNode->textContent;
1229 } else {
1230 $out .= '';
1232 } elseif ( $contextNode->nodeName == 'ext' ) {
1233 # Extension tag
1234 $xpath = new DOMXPath( $contextNode->ownerDocument );
1235 $names = $xpath->query( 'name', $contextNode );
1236 $attrs = $xpath->query( 'attr', $contextNode );
1237 $inners = $xpath->query( 'inner', $contextNode );
1238 $closes = $xpath->query( 'close', $contextNode );
1239 if ( $flags & PPFrame::NO_TAGS ) {
1240 $s = '<' . $this->expand( $names->item( 0 ), $flags );
1241 if ( $attrs->length > 0 ) {
1242 $s .= $this->expand( $attrs->item( 0 ), $flags );
1244 if ( $inners->length > 0 ) {
1245 $s .= '>' . $this->expand( $inners->item( 0 ), $flags );
1246 if ( $closes->length > 0 ) {
1247 $s .= $this->expand( $closes->item( 0 ), $flags );
1249 } else {
1250 $s .= '/>';
1252 $out .= $s;
1253 } else {
1254 $params = [
1255 'name' => new PPNode_DOM( $names->item( 0 ) ),
1256 'attr' => $attrs->length > 0 ? new PPNode_DOM( $attrs->item( 0 ) ) : null,
1257 'inner' => $inners->length > 0 ? new PPNode_DOM( $inners->item( 0 ) ) : null,
1258 'close' => $closes->length > 0 ? new PPNode_DOM( $closes->item( 0 ) ) : null,
1260 $out .= $this->parser->extensionSubstitution( $params, $this );
1262 } elseif ( $contextNode->nodeName == 'h' ) {
1263 # Heading
1264 $s = $this->expand( $contextNode->childNodes, $flags );
1266 # Insert a heading marker only for <h> children of <root>
1267 # This is to stop extractSections from going over multiple tree levels
1268 if ( $contextNode->parentNode->nodeName == 'root' && $this->parser->ot['html'] ) {
1269 # Insert heading index marker
1270 $headingIndex = $contextNode->getAttribute( 'i' );
1271 $titleText = $this->title->getPrefixedDBkey();
1272 $this->parser->mHeadings[] = [ $titleText, $headingIndex ];
1273 $serial = count( $this->parser->mHeadings ) - 1;
1274 $marker = Parser::MARKER_PREFIX . "-h-$serial-" . Parser::MARKER_SUFFIX;
1275 $count = $contextNode->getAttribute( 'level' );
1276 $s = substr( $s, 0, $count ) . $marker . substr( $s, $count );
1277 $this->parser->mStripState->addGeneral( $marker, '' );
1279 $out .= $s;
1280 } else {
1281 # Generic recursive expansion
1282 $newIterator = $contextNode->childNodes;
1284 } else {
1285 throw new MWException( __METHOD__ . ': Invalid parameter type' );
1288 if ( $newIterator !== false ) {
1289 if ( $newIterator instanceof PPNode_DOM ) {
1290 $newIterator = $newIterator->node;
1292 $outStack[] = '';
1293 $iteratorStack[] = $newIterator;
1294 $indexStack[] = 0;
1295 } elseif ( $iteratorStack[$level] === false ) {
1296 // Return accumulated value to parent
1297 // With tail recursion
1298 while ( $iteratorStack[$level] === false && $level > 0 ) {
1299 $outStack[$level - 1] .= $out;
1300 array_pop( $outStack );
1301 array_pop( $iteratorStack );
1302 array_pop( $indexStack );
1303 $level--;
1307 --$expansionDepth;
1308 return $outStack[0];
1312 * @param string $sep
1313 * @param int $flags
1314 * @param string|PPNode_DOM|DOMDocument $args,...
1315 * @return string
1317 public function implodeWithFlags( $sep, $flags /*, ... */ ) {
1318 $args = array_slice( func_get_args(), 2 );
1320 $first = true;
1321 $s = '';
1322 foreach ( $args as $root ) {
1323 if ( $root instanceof PPNode_DOM ) {
1324 $root = $root->node;
1326 if ( !is_array( $root ) && !( $root instanceof DOMNodeList ) ) {
1327 $root = [ $root ];
1329 foreach ( $root as $node ) {
1330 if ( $first ) {
1331 $first = false;
1332 } else {
1333 $s .= $sep;
1335 $s .= $this->expand( $node, $flags );
1338 return $s;
1342 * Implode with no flags specified
1343 * This previously called implodeWithFlags but has now been inlined to reduce stack depth
1345 * @param string $sep
1346 * @param string|PPNode_DOM|DOMDocument $args,...
1347 * @return string
1349 public function implode( $sep /*, ... */ ) {
1350 $args = array_slice( func_get_args(), 1 );
1352 $first = true;
1353 $s = '';
1354 foreach ( $args as $root ) {
1355 if ( $root instanceof PPNode_DOM ) {
1356 $root = $root->node;
1358 if ( !is_array( $root ) && !( $root instanceof DOMNodeList ) ) {
1359 $root = [ $root ];
1361 foreach ( $root as $node ) {
1362 if ( $first ) {
1363 $first = false;
1364 } else {
1365 $s .= $sep;
1367 $s .= $this->expand( $node );
1370 return $s;
1374 * Makes an object that, when expand()ed, will be the same as one obtained
1375 * with implode()
1377 * @param string $sep
1378 * @param string|PPNode_DOM|DOMDocument $args,...
1379 * @return array
1381 public function virtualImplode( $sep /*, ... */ ) {
1382 $args = array_slice( func_get_args(), 1 );
1383 $out = [];
1384 $first = true;
1386 foreach ( $args as $root ) {
1387 if ( $root instanceof PPNode_DOM ) {
1388 $root = $root->node;
1390 if ( !is_array( $root ) && !( $root instanceof DOMNodeList ) ) {
1391 $root = [ $root ];
1393 foreach ( $root as $node ) {
1394 if ( $first ) {
1395 $first = false;
1396 } else {
1397 $out[] = $sep;
1399 $out[] = $node;
1402 return $out;
1406 * Virtual implode with brackets
1407 * @param string $start
1408 * @param string $sep
1409 * @param string $end
1410 * @param string|PPNode_DOM|DOMDocument $args,...
1411 * @return array
1413 public function virtualBracketedImplode( $start, $sep, $end /*, ... */ ) {
1414 $args = array_slice( func_get_args(), 3 );
1415 $out = [ $start ];
1416 $first = true;
1418 foreach ( $args as $root ) {
1419 if ( $root instanceof PPNode_DOM ) {
1420 $root = $root->node;
1422 if ( !is_array( $root ) && !( $root instanceof DOMNodeList ) ) {
1423 $root = [ $root ];
1425 foreach ( $root as $node ) {
1426 if ( $first ) {
1427 $first = false;
1428 } else {
1429 $out[] = $sep;
1431 $out[] = $node;
1434 $out[] = $end;
1435 return $out;
1438 public function __toString() {
1439 return 'frame{}';
1442 public function getPDBK( $level = false ) {
1443 if ( $level === false ) {
1444 return $this->title->getPrefixedDBkey();
1445 } else {
1446 return isset( $this->titleCache[$level] ) ? $this->titleCache[$level] : false;
1451 * @return array
1453 public function getArguments() {
1454 return [];
1458 * @return array
1460 public function getNumberedArguments() {
1461 return [];
1465 * @return array
1467 public function getNamedArguments() {
1468 return [];
1472 * Returns true if there are no arguments in this frame
1474 * @return bool
1476 public function isEmpty() {
1477 return true;
1481 * @param int|string $name
1482 * @return bool Always false in this implementation.
1484 public function getArgument( $name ) {
1485 return false;
1489 * Returns true if the infinite loop check is OK, false if a loop is detected
1491 * @param Title $title
1492 * @return bool
1494 public function loopCheck( $title ) {
1495 return !isset( $this->loopCheckHash[$title->getPrefixedDBkey()] );
1499 * Return true if the frame is a template frame
1501 * @return bool
1503 public function isTemplate() {
1504 return false;
1508 * Get a title of frame
1510 * @return Title
1512 public function getTitle() {
1513 return $this->title;
1517 * Set the volatile flag
1519 * @param bool $flag
1521 public function setVolatile( $flag = true ) {
1522 $this->volatile = $flag;
1526 * Get the volatile flag
1528 * @return bool
1530 public function isVolatile() {
1531 return $this->volatile;
1535 * Set the TTL
1537 * @param int $ttl
1539 public function setTTL( $ttl ) {
1540 if ( $ttl !== null && ( $this->ttl === null || $ttl < $this->ttl ) ) {
1541 $this->ttl = $ttl;
1546 * Get the TTL
1548 * @return int|null
1550 public function getTTL() {
1551 return $this->ttl;
1556 * Expansion frame with template arguments
1557 * @ingroup Parser
1559 // @codingStandardsIgnoreStart Squiz.Classes.ValidClassName.NotCamelCaps
1560 class PPTemplateFrame_DOM extends PPFrame_DOM {
1561 // @codingStandardsIgnoreEnd
1563 public $numberedArgs, $namedArgs;
1566 * @var PPFrame_DOM
1568 public $parent;
1569 public $numberedExpansionCache, $namedExpansionCache;
1572 * @param Preprocessor $preprocessor
1573 * @param bool|PPFrame_DOM $parent
1574 * @param array $numberedArgs
1575 * @param array $namedArgs
1576 * @param bool|Title $title
1578 public function __construct( $preprocessor, $parent = false, $numberedArgs = [],
1579 $namedArgs = [], $title = false
1581 parent::__construct( $preprocessor );
1583 $this->parent = $parent;
1584 $this->numberedArgs = $numberedArgs;
1585 $this->namedArgs = $namedArgs;
1586 $this->title = $title;
1587 $pdbk = $title ? $title->getPrefixedDBkey() : false;
1588 $this->titleCache = $parent->titleCache;
1589 $this->titleCache[] = $pdbk;
1590 $this->loopCheckHash = /*clone*/ $parent->loopCheckHash;
1591 if ( $pdbk !== false ) {
1592 $this->loopCheckHash[$pdbk] = true;
1594 $this->depth = $parent->depth + 1;
1595 $this->numberedExpansionCache = $this->namedExpansionCache = [];
1598 public function __toString() {
1599 $s = 'tplframe{';
1600 $first = true;
1601 $args = $this->numberedArgs + $this->namedArgs;
1602 foreach ( $args as $name => $value ) {
1603 if ( $first ) {
1604 $first = false;
1605 } else {
1606 $s .= ', ';
1608 $s .= "\"$name\":\"" .
1609 str_replace( '"', '\\"', $value->ownerDocument->saveXML( $value ) ) . '"';
1611 $s .= '}';
1612 return $s;
1616 * @throws MWException
1617 * @param string|int $key
1618 * @param string|PPNode_DOM|DOMDocument $root
1619 * @param int $flags
1620 * @return string
1622 public function cachedExpand( $key, $root, $flags = 0 ) {
1623 if ( isset( $this->parent->childExpansionCache[$key] ) ) {
1624 return $this->parent->childExpansionCache[$key];
1626 $retval = $this->expand( $root, $flags );
1627 if ( !$this->isVolatile() ) {
1628 $this->parent->childExpansionCache[$key] = $retval;
1630 return $retval;
1634 * Returns true if there are no arguments in this frame
1636 * @return bool
1638 public function isEmpty() {
1639 return !count( $this->numberedArgs ) && !count( $this->namedArgs );
1642 public function getArguments() {
1643 $arguments = [];
1644 foreach ( array_merge(
1645 array_keys( $this->numberedArgs ),
1646 array_keys( $this->namedArgs ) ) as $key ) {
1647 $arguments[$key] = $this->getArgument( $key );
1649 return $arguments;
1652 public function getNumberedArguments() {
1653 $arguments = [];
1654 foreach ( array_keys( $this->numberedArgs ) as $key ) {
1655 $arguments[$key] = $this->getArgument( $key );
1657 return $arguments;
1660 public function getNamedArguments() {
1661 $arguments = [];
1662 foreach ( array_keys( $this->namedArgs ) as $key ) {
1663 $arguments[$key] = $this->getArgument( $key );
1665 return $arguments;
1669 * @param int $index
1670 * @return string|bool
1672 public function getNumberedArgument( $index ) {
1673 if ( !isset( $this->numberedArgs[$index] ) ) {
1674 return false;
1676 if ( !isset( $this->numberedExpansionCache[$index] ) ) {
1677 # No trimming for unnamed arguments
1678 $this->numberedExpansionCache[$index] = $this->parent->expand(
1679 $this->numberedArgs[$index],
1680 PPFrame::STRIP_COMMENTS
1683 return $this->numberedExpansionCache[$index];
1687 * @param string $name
1688 * @return string|bool
1690 public function getNamedArgument( $name ) {
1691 if ( !isset( $this->namedArgs[$name] ) ) {
1692 return false;
1694 if ( !isset( $this->namedExpansionCache[$name] ) ) {
1695 # Trim named arguments post-expand, for backwards compatibility
1696 $this->namedExpansionCache[$name] = trim(
1697 $this->parent->expand( $this->namedArgs[$name], PPFrame::STRIP_COMMENTS ) );
1699 return $this->namedExpansionCache[$name];
1703 * @param int|string $name
1704 * @return string|bool
1706 public function getArgument( $name ) {
1707 $text = $this->getNumberedArgument( $name );
1708 if ( $text === false ) {
1709 $text = $this->getNamedArgument( $name );
1711 return $text;
1715 * Return true if the frame is a template frame
1717 * @return bool
1719 public function isTemplate() {
1720 return true;
1723 public function setVolatile( $flag = true ) {
1724 parent::setVolatile( $flag );
1725 $this->parent->setVolatile( $flag );
1728 public function setTTL( $ttl ) {
1729 parent::setTTL( $ttl );
1730 $this->parent->setTTL( $ttl );
1735 * Expansion frame with custom arguments
1736 * @ingroup Parser
1738 // @codingStandardsIgnoreStart Squiz.Classes.ValidClassName.NotCamelCaps
1739 class PPCustomFrame_DOM extends PPFrame_DOM {
1740 // @codingStandardsIgnoreEnd
1742 public $args;
1744 public function __construct( $preprocessor, $args ) {
1745 parent::__construct( $preprocessor );
1746 $this->args = $args;
1749 public function __toString() {
1750 $s = 'cstmframe{';
1751 $first = true;
1752 foreach ( $this->args as $name => $value ) {
1753 if ( $first ) {
1754 $first = false;
1755 } else {
1756 $s .= ', ';
1758 $s .= "\"$name\":\"" .
1759 str_replace( '"', '\\"', $value->__toString() ) . '"';
1761 $s .= '}';
1762 return $s;
1766 * @return bool
1768 public function isEmpty() {
1769 return !count( $this->args );
1773 * @param int|string $index
1774 * @return string|bool
1776 public function getArgument( $index ) {
1777 if ( !isset( $this->args[$index] ) ) {
1778 return false;
1780 return $this->args[$index];
1783 public function getArguments() {
1784 return $this->args;
1789 * @ingroup Parser
1791 // @codingStandardsIgnoreStart Squiz.Classes.ValidClassName.NotCamelCaps
1792 class PPNode_DOM implements PPNode {
1793 // @codingStandardsIgnoreEnd
1796 * @var DOMElement
1798 public $node;
1799 public $xpath;
1801 public function __construct( $node, $xpath = false ) {
1802 $this->node = $node;
1806 * @return DOMXPath
1808 public function getXPath() {
1809 if ( $this->xpath === null ) {
1810 $this->xpath = new DOMXPath( $this->node->ownerDocument );
1812 return $this->xpath;
1815 public function __toString() {
1816 if ( $this->node instanceof DOMNodeList ) {
1817 $s = '';
1818 foreach ( $this->node as $node ) {
1819 $s .= $node->ownerDocument->saveXML( $node );
1821 } else {
1822 $s = $this->node->ownerDocument->saveXML( $this->node );
1824 return $s;
1828 * @return bool|PPNode_DOM
1830 public function getChildren() {
1831 return $this->node->childNodes ? new self( $this->node->childNodes ) : false;
1835 * @return bool|PPNode_DOM
1837 public function getFirstChild() {
1838 return $this->node->firstChild ? new self( $this->node->firstChild ) : false;
1842 * @return bool|PPNode_DOM
1844 public function getNextSibling() {
1845 return $this->node->nextSibling ? new self( $this->node->nextSibling ) : false;
1849 * @param string $type
1851 * @return bool|PPNode_DOM
1853 public function getChildrenOfType( $type ) {
1854 return new self( $this->getXPath()->query( $type, $this->node ) );
1858 * @return int
1860 public function getLength() {
1861 if ( $this->node instanceof DOMNodeList ) {
1862 return $this->node->length;
1863 } else {
1864 return false;
1869 * @param int $i
1870 * @return bool|PPNode_DOM
1872 public function item( $i ) {
1873 $item = $this->node->item( $i );
1874 return $item ? new self( $item ) : false;
1878 * @return string
1880 public function getName() {
1881 if ( $this->node instanceof DOMNodeList ) {
1882 return '#nodelist';
1883 } else {
1884 return $this->node->nodeName;
1889 * Split a "<part>" node into an associative array containing:
1890 * - name PPNode name
1891 * - index String index
1892 * - value PPNode value
1894 * @throws MWException
1895 * @return array
1897 public function splitArg() {
1898 $xpath = $this->getXPath();
1899 $names = $xpath->query( 'name', $this->node );
1900 $values = $xpath->query( 'value', $this->node );
1901 if ( !$names->length || !$values->length ) {
1902 throw new MWException( 'Invalid brace node passed to ' . __METHOD__ );
1904 $name = $names->item( 0 );
1905 $index = $name->getAttribute( 'index' );
1906 return [
1907 'name' => new self( $name ),
1908 'index' => $index,
1909 'value' => new self( $values->item( 0 ) ) ];
1913 * Split an "<ext>" node into an associative array containing name, attr, inner and close
1914 * All values in the resulting array are PPNodes. Inner and close are optional.
1916 * @throws MWException
1917 * @return array
1919 public function splitExt() {
1920 $xpath = $this->getXPath();
1921 $names = $xpath->query( 'name', $this->node );
1922 $attrs = $xpath->query( 'attr', $this->node );
1923 $inners = $xpath->query( 'inner', $this->node );
1924 $closes = $xpath->query( 'close', $this->node );
1925 if ( !$names->length || !$attrs->length ) {
1926 throw new MWException( 'Invalid ext node passed to ' . __METHOD__ );
1928 $parts = [
1929 'name' => new self( $names->item( 0 ) ),
1930 'attr' => new self( $attrs->item( 0 ) ) ];
1931 if ( $inners->length ) {
1932 $parts['inner'] = new self( $inners->item( 0 ) );
1934 if ( $closes->length ) {
1935 $parts['close'] = new self( $closes->item( 0 ) );
1937 return $parts;
1941 * Split a "<h>" node
1942 * @throws MWException
1943 * @return array
1945 public function splitHeading() {
1946 if ( $this->getName() !== 'h' ) {
1947 throw new MWException( 'Invalid h node passed to ' . __METHOD__ );
1949 return [
1950 'i' => $this->node->getAttribute( 'i' ),
1951 'level' => $this->node->getAttribute( 'level' ),
1952 'contents' => $this->getChildren()