dépôts / lhc / web / wiklou.git / blob
? search:
summary | shortlog | log | commit | commitdiff | tree
history | raw | HEAD
tests: Complete test coverage of HtmlArmor
[lhc/web/wiklou.git] / includes / parser / Preprocessor_DOM.php
1 <?php
2 /**
3 * Preprocessor using PHP's dom extension
4 *
5 * This program is free software; you can redistribute it and/or modify
6 * it under the terms of the GNU General Public License as published by
7 * the Free Software Foundation; either version 2 of the License, or
8 * (at your option) any later version.
9 *
10 * This program is distributed in the hope that it will be useful,
11 * but WITHOUT ANY WARRANTY; without even the implied warranty of
12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13 * GNU General Public License for more details.
14 *
15 * You should have received a copy of the GNU General Public License along
16 * with this program; if not, write to the Free Software Foundation, Inc.,
17 * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
18 * http://www.gnu.org/copyleft/gpl.html
19 *
20 * @file
21 * @ingroup Parser
22 */
23
24 /**
25 * @ingroup Parser
26 */
27 // @codingStandardsIgnoreStart Squiz.Classes.ValidClassName.NotCamelCaps
28 class Preprocessor_DOM extends Preprocessor {
29 // @codingStandardsIgnoreEnd
30
31 /**
32 * @var Parser
33 */
34 public $parser;
35
36 public $memoryLimit;
37
38 const CACHE_PREFIX = 'preprocess-xml';
39
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;
49 }
50 }
51 }
52
53 /**
54 * @return PPFrame_DOM
55 */
56 public function newFrame() {
57 return new PPFrame_DOM( $this );
58 }
59
60 /**
61 * @param array $args
62 * @return PPCustomFrame_DOM
63 */
64 public function newCustomFrame( $args ) {
65 return new PPCustomFrame_DOM( $this, $args );
66 }
67
68 /**
69 * @param array $values
70 * @return PPNode_DOM
71 * @throws MWException
72 */
73 public function newPartNodeArray( $values ) {
74 // NOTE: DOM manipulation is slower than building & parsing XML! (or so Tim sais)
75 $xml = "<list>";
76
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>";
84 }
85 }
86
87 $xml .= "</list>";
88
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( $xml, 1 << 19 );
99 }
100
101 if ( !$result ) {
102 throw new MWException( 'Parameters passed to ' . __METHOD__ . ' result in invalid XML' );
103 }
104
105 $root = $dom->documentElement;
106 $node = new PPNode_DOM( $root->childNodes );
107 return $node;
108 }
109
110 /**
111 * @throws MWException
112 * @return bool
113 */
114 public function memCheck() {
115 if ( $this->memoryLimit === false ) {
116 return true;
117 }
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)" );
122 }
123 return $usage <= $this->memoryLimit * 0.8;
124 }
125
126 /**
127 * Preprocess some wikitext and return the document tree.
128 * This is the ghost of Parser::replace_variables().
129 *
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.
135 *
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 T6899.
138 *
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().
142 *
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.
146 *
147 * @throws MWException
148 * @return PPNode_DOM
149 */
150 public function preprocessToObj( $text, $flags = 0 ) {
151 $xml = $this->cacheGetTree( $text, $flags );
152 if ( $xml === false ) {
153 $xml = $this->preprocessToXml( $text, $flags );
154 $this->cacheSetTree( $text, $flags, $xml );
155 }
156
157 // Fail if the number of elements exceeds acceptable limits
158 // Do not attempt to generate the DOM
159 $this->parser->mGeneratedPPNodeCount += substr_count( $xml, '<' );
160 $max = $this->parser->mOptions->getMaxGeneratedPPNodeCount();
161 if ( $this->parser->mGeneratedPPNodeCount > $max ) {
162 // if ( $cacheable ) { ... }
163 throw new MWException( __METHOD__ . ': generated node count limit exceeded' );
164 }
165
166 $dom = new DOMDocument;
167 MediaWiki\suppressWarnings();
168 $result = $dom->loadXML( $xml );
169 MediaWiki\restoreWarnings();
170 if ( !$result ) {
171 // Try running the XML through UtfNormal to get rid of invalid characters
172 $xml = UtfNormal\Validator::cleanUp( $xml );
173 // 1 << 19 == XML_PARSE_HUGE, needed so newer versions of libxml2
174 // don't barf when the XML is >256 levels deep.
175 $result = $dom->loadXML( $xml, 1 << 19 );
176 }
177 if ( $result ) {
178 $obj = new PPNode_DOM( $dom->documentElement );
179 }
180
181 // if ( $cacheable ) { ... }
182
183 if ( !$result ) {
184 throw new MWException( __METHOD__ . ' generated invalid XML' );
185 }
186 return $obj;
187 }
188
189 /**
190 * @param string $text
191 * @param int $flags
192 * @return string
193 */
194 public function preprocessToXml( $text, $flags = 0 ) {
195 global $wgDisableLangConversion;
196
197 $forInclusion = $flags & Parser::PTD_FOR_INCLUSION;
198
199 $xmlishElements = $this->parser->getStripList();
200 $xmlishAllowMissingEndTag = [ 'includeonly', 'noinclude', 'onlyinclude' ];
201 $enableOnlyinclude = false;
202 if ( $forInclusion ) {
203 $ignoredTags = [ 'includeonly', '/includeonly' ];
204 $ignoredElements = [ 'noinclude' ];
205 $xmlishElements[] = 'noinclude';
206 if ( strpos( $text, '<onlyinclude>' ) !== false
207 && strpos( $text, '</onlyinclude>' ) !== false
208 ) {
209 $enableOnlyinclude = true;
210 }
211 } else {
212 $ignoredTags = [ 'noinclude', '/noinclude', 'onlyinclude', '/onlyinclude' ];
213 $ignoredElements = [ 'includeonly' ];
214 $xmlishElements[] = 'includeonly';
215 }
216 $xmlishRegex = implode( '|', array_merge( $xmlishElements, $ignoredTags ) );
217
218 // Use "A" modifier (anchored) instead of "^", because ^ doesn't work with an offset
219 $elementsRegex = "~($xmlishRegex)(?:\s|\/>|>)|(!--)~iA";
220
221 $stack = new PPDStack;
222
223 $searchBase = "[{<\n"; # }
224 if ( !$wgDisableLangConversion ) {
225 $searchBase .= '-';
226 }
227
228 // For fast reverse searches
229 $revText = strrev( $text );
230 $lengthText = strlen( $text );
231
232 // Input pointer, starts out pointing to a pseudo-newline before the start
233 $i = 0;
234 // Current accumulator
235 $accum =& $stack->getAccum();
236 $accum = '<root>';
237 // True to find equals signs in arguments
238 $findEquals = false;
239 // True to take notice of pipe characters
240 $findPipe = false;
241 $headingIndex = 1;
242 // True if $i is inside a possible heading
243 $inHeading = false;
244 // True if there are no more greater-than (>) signs right of $i
245 $noMoreGT = false;
246 // Map of tag name => true if there are no more closing tags of given type right of $i
247 $noMoreClosingTag = [];
248 // True to ignore all input up to the next <onlyinclude>
249 $findOnlyinclude = $enableOnlyinclude;
250 // Do a line-start run without outputting an LF character
251 $fakeLineStart = true;
252
253 while ( true ) {
254 // $this->memCheck();
255
256 if ( $findOnlyinclude ) {
257 // Ignore all input up to the next <onlyinclude>
258 $startPos = strpos( $text, '<onlyinclude>', $i );
259 if ( $startPos === false ) {
260 // Ignored section runs to the end
261 $accum .= '<ignore>' . htmlspecialchars( substr( $text, $i ) ) . '</ignore>';
262 break;
263 }
264 $tagEndPos = $startPos + strlen( '<onlyinclude>' ); // past-the-end
265 $accum .= '<ignore>' . htmlspecialchars( substr( $text, $i, $tagEndPos - $i ) ) . '</ignore>';
266 $i = $tagEndPos;
267 $findOnlyinclude = false;
268 }
269
270 if ( $fakeLineStart ) {
271 $found = 'line-start';
272 $curChar = '';
273 } else {
274 # Find next opening brace, closing brace or pipe
275 $search = $searchBase;
276 if ( $stack->top === false ) {
277 $currentClosing = '';
278 } elseif (
279 $stack->top->close === '}-' &&
280 $stack->top->count > 2
281 ) {
282 # adjust closing for -{{{...{{
283 $currentClosing = '}';
284 $search .= $currentClosing;
285 } else {
286 $currentClosing = $stack->top->close;
287 $search .= $currentClosing;
288 }
289 if ( $findPipe ) {
290 $search .= '|';
291 }
292 if ( $findEquals ) {
293 // First equals will be for the template
294 $search .= '=';
295 }
296 $rule = null;
297 # Output literal section, advance input counter
298 $literalLength = strcspn( $text, $search, $i );
299 if ( $literalLength > 0 ) {
300 $accum .= htmlspecialchars( substr( $text, $i, $literalLength ) );
301 $i += $literalLength;
302 }
303 if ( $i >= $lengthText ) {
304 if ( $currentClosing == "\n" ) {
305 // Do a past-the-end run to finish off the heading
306 $curChar = '';
307 $found = 'line-end';
308 } else {
309 # All done
310 break;
311 }
312 } else {
313 $curChar = $curTwoChar = $text[$i];
314 if ( ( $i + 1 ) < $lengthText ) {
315 $curTwoChar .= $text[$i + 1];
316 }
317 if ( $curChar == '|' ) {
318 $found = 'pipe';
319 } elseif ( $curChar == '=' ) {
320 $found = 'equals';
321 } elseif ( $curChar == '<' ) {
322 $found = 'angle';
323 } elseif ( $curChar == "\n" ) {
324 if ( $inHeading ) {
325 $found = 'line-end';
326 } else {
327 $found = 'line-start';
328 }
329 } elseif ( $curTwoChar == $currentClosing ) {
330 $found = 'close';
331 $curChar = $curTwoChar;
332 } elseif ( $curChar == $currentClosing ) {
333 $found = 'close';
334 } elseif ( isset( $this->rules[$curTwoChar] ) ) {
335 $curChar = $curTwoChar;
336 $found = 'open';
337 $rule = $this->rules[$curChar];
338 } elseif ( isset( $this->rules[$curChar] ) ) {
339 $found = 'open';
340 $rule = $this->rules[$curChar];
341 } else {
342 # Some versions of PHP have a strcspn which stops on
343 # null characters; ignore these and continue.
344 # We also may get '-' and '}' characters here which
345 # don't match -{ or $currentClosing. Add these to
346 # output and continue.
347 if ( $curChar == '-' || $curChar == '}' ) {
348 $accum .= $curChar;
349 }
350 ++$i;
351 continue;
352 }
353 }
354 }
355
356 if ( $found == 'angle' ) {
357 $matches = false;
358 // Handle </onlyinclude>
359 if ( $enableOnlyinclude
360 && substr( $text, $i, strlen( '</onlyinclude>' ) ) == '</onlyinclude>'
361 ) {
362 $findOnlyinclude = true;
363 continue;
364 }
365
366 // Determine element name
367 if ( !preg_match( $elementsRegex, $text, $matches, 0, $i + 1 ) ) {
368 // Element name missing or not listed
369 $accum .= '&lt;';
370 ++$i;
371 continue;
372 }
373 // Handle comments
374 if ( isset( $matches[2] ) && $matches[2] == '!--' ) {
375 // To avoid leaving blank lines, when a sequence of
376 // space-separated comments is both preceded and followed by
377 // a newline (ignoring spaces), then
378 // trim leading and trailing spaces and the trailing newline.
379
380 // Find the end
381 $endPos = strpos( $text, '-->', $i + 4 );
382 if ( $endPos === false ) {
383 // Unclosed comment in input, runs to end
384 $inner = substr( $text, $i );
385 $accum .= '<comment>' . htmlspecialchars( $inner ) . '</comment>';
386 $i = $lengthText;
387 } else {
388 // Search backwards for leading whitespace
389 $wsStart = $i ? ( $i - strspn( $revText, " \t", $lengthText - $i ) ) : 0;
390
391 // Search forwards for trailing whitespace
392 // $wsEnd will be the position of the last space (or the '>' if there's none)
393 $wsEnd = $endPos + 2 + strspn( $text, " \t", $endPos + 3 );
394
395 // Keep looking forward as long as we're finding more
396 // comments.
397 $comments = [ [ $wsStart, $wsEnd ] ];
398 while ( substr( $text, $wsEnd + 1, 4 ) == '<!--' ) {
399 $c = strpos( $text, '-->', $wsEnd + 4 );
400 if ( $c === false ) {
401 break;
402 }
403 $c = $c + 2 + strspn( $text, " \t", $c + 3 );
404 $comments[] = [ $wsEnd + 1, $c ];
405 $wsEnd = $c;
406 }
407
408 // Eat the line if possible
409 // TODO: This could theoretically be done if $wsStart == 0, i.e. for comments at
410 // the overall start. That's not how Sanitizer::removeHTMLcomments() did it, but
411 // it's a possible beneficial b/c break.
412 if ( $wsStart > 0 && substr( $text, $wsStart - 1, 1 ) == "\n"
413 && substr( $text, $wsEnd + 1, 1 ) == "\n"
414 ) {
415 // Remove leading whitespace from the end of the accumulator
416 // Sanity check first though
417 $wsLength = $i - $wsStart;
418 if ( $wsLength > 0
419 && strspn( $accum, " \t", -$wsLength ) === $wsLength
420 ) {
421 $accum = substr( $accum, 0, -$wsLength );
422 }
423
424 // Dump all but the last comment to the accumulator
425 foreach ( $comments as $j => $com ) {
426 $startPos = $com[0];
427 $endPos = $com[1] + 1;
428 if ( $j == ( count( $comments ) - 1 ) ) {
429 break;
430 }
431 $inner = substr( $text, $startPos, $endPos - $startPos );
432 $accum .= '<comment>' . htmlspecialchars( $inner ) . '</comment>';
433 }
434
435 // Do a line-start run next time to look for headings after the comment
436 $fakeLineStart = true;
437 } else {
438 // No line to eat, just take the comment itself
439 $startPos = $i;
440 $endPos += 2;
441 }
442
443 if ( $stack->top ) {
444 $part = $stack->top->getCurrentPart();
445 if ( !( isset( $part->commentEnd ) && $part->commentEnd == $wsStart - 1 ) ) {
446 $part->visualEnd = $wsStart;
447 }
448 // Else comments abutting, no change in visual end
449 $part->commentEnd = $endPos;
450 }
451 $i = $endPos + 1;
452 $inner = substr( $text, $startPos, $endPos - $startPos + 1 );
453 $accum .= '<comment>' . htmlspecialchars( $inner ) . '</comment>';
454 }
455 continue;
456 }
457 $name = $matches[1];
458 $lowerName = strtolower( $name );
459 $attrStart = $i + strlen( $name ) + 1;
460
461 // Find end of tag
462 $tagEndPos = $noMoreGT ? false : strpos( $text, '>', $attrStart );
463 if ( $tagEndPos === false ) {
464 // Infinite backtrack
465 // Disable tag search to prevent worst-case O(N^2) performance
466 $noMoreGT = true;
467 $accum .= '&lt;';
468 ++$i;
469 continue;
470 }
471
472 // Handle ignored tags
473 if ( in_array( $lowerName, $ignoredTags ) ) {
474 $accum .= '<ignore>'
475 . htmlspecialchars( substr( $text, $i, $tagEndPos - $i + 1 ) )
476 . '</ignore>';
477 $i = $tagEndPos + 1;
478 continue;
479 }
480
481 $tagStartPos = $i;
482 if ( $text[$tagEndPos - 1] == '/' ) {
483 $attrEnd = $tagEndPos - 1;
484 $inner = null;
485 $i = $tagEndPos + 1;
486 $close = '';
487 } else {
488 $attrEnd = $tagEndPos;
489 // Find closing tag
490 if (
491 !isset( $noMoreClosingTag[$name] ) &&
492 preg_match( "/<\/" . preg_quote( $name, '/' ) . "\s*>/i",
493 $text, $matches, PREG_OFFSET_CAPTURE, $tagEndPos + 1 )
494 ) {
495 $inner = substr( $text, $tagEndPos + 1, $matches[0][1] - $tagEndPos - 1 );
496 $i = $matches[0][1] + strlen( $matches[0][0] );
497 $close = '<close>' . htmlspecialchars( $matches[0][0] ) . '</close>';
498 } else {
499 // No end tag
500 if ( in_array( $name, $xmlishAllowMissingEndTag ) ) {
501 // Let it run out to the end of the text.
502 $inner = substr( $text, $tagEndPos + 1 );
503 $i = $lengthText;
504 $close = '';
505 } else {
506 // Don't match the tag, treat opening tag as literal and resume parsing.
507 $i = $tagEndPos + 1;
508 $accum .= htmlspecialchars( substr( $text, $tagStartPos, $tagEndPos + 1 - $tagStartPos ) );
509 // Cache results, otherwise we have O(N^2) performance for input like <foo><foo><foo>...
510 $noMoreClosingTag[$name] = true;
511 continue;
512 }
513 }
514 }
515 // <includeonly> and <noinclude> just become <ignore> tags
516 if ( in_array( $lowerName, $ignoredElements ) ) {
517 $accum .= '<ignore>' . htmlspecialchars( substr( $text, $tagStartPos, $i - $tagStartPos ) )
518 . '</ignore>';
519 continue;
520 }
521
522 $accum .= '<ext>';
523 if ( $attrEnd <= $attrStart ) {
524 $attr = '';
525 } else {
526 $attr = substr( $text, $attrStart, $attrEnd - $attrStart );
527 }
528 $accum .= '<name>' . htmlspecialchars( $name ) . '</name>' .
529 // Note that the attr element contains the whitespace between name and attribute,
530 // this is necessary for precise reconstruction during pre-save transform.
531 '<attr>' . htmlspecialchars( $attr ) . '</attr>';
532 if ( $inner !== null ) {
533 $accum .= '<inner>' . htmlspecialchars( $inner ) . '</inner>';
534 }
535 $accum .= $close . '</ext>';
536 } elseif ( $found == 'line-start' ) {
537 // Is this the start of a heading?
538 // Line break belongs before the heading element in any case
539 if ( $fakeLineStart ) {
540 $fakeLineStart = false;
541 } else {
542 $accum .= $curChar;
543 $i++;
544 }
545
546 $count = strspn( $text, '=', $i, 6 );
547 if ( $count == 1 && $findEquals ) {
548 // DWIM: This looks kind of like a name/value separator.
549 // Let's let the equals handler have it and break the
550 // potential heading. This is heuristic, but AFAICT the
551 // methods for completely correct disambiguation are very
552 // complex.
553 } elseif ( $count > 0 ) {
554 $piece = [
555 'open' => "\n",
556 'close' => "\n",
557 'parts' => [ new PPDPart( str_repeat( '=', $count ) ) ],
558 'startPos' => $i,
559 'count' => $count ];
560 $stack->push( $piece );
561 $accum =& $stack->getAccum();
562 $flags = $stack->getFlags();
563 extract( $flags );
564 $i += $count;
565 }
566 } elseif ( $found == 'line-end' ) {
567 $piece = $stack->top;
568 // A heading must be open, otherwise \n wouldn't have been in the search list
569 assert( $piece->open === "\n" );
570 $part = $piece->getCurrentPart();
571 // Search back through the input to see if it has a proper close.
572 // Do this using the reversed string since the other solutions
573 // (end anchor, etc.) are inefficient.
574 $wsLength = strspn( $revText, " \t", $lengthText - $i );
575 $searchStart = $i - $wsLength;
576 if ( isset( $part->commentEnd ) && $searchStart - 1 == $part->commentEnd ) {
577 // Comment found at line end
578 // Search for equals signs before the comment
579 $searchStart = $part->visualEnd;
580 $searchStart -= strspn( $revText, " \t", $lengthText - $searchStart );
581 }
582 $count = $piece->count;
583 $equalsLength = strspn( $revText, '=', $lengthText - $searchStart );
584 if ( $equalsLength > 0 ) {
585 if ( $searchStart - $equalsLength == $piece->startPos ) {
586 // This is just a single string of equals signs on its own line
587 // Replicate the doHeadings behavior /={count}(.+)={count}/
588 // First find out how many equals signs there really are (don't stop at 6)
589 $count = $equalsLength;
590 if ( $count < 3 ) {
591 $count = 0;
592 } else {
593 $count = min( 6, intval( ( $count - 1 ) / 2 ) );
594 }
595 } else {
596 $count = min( $equalsLength, $count );
597 }
598 if ( $count > 0 ) {
599 // Normal match, output <h>
600 $element = "<h level=\"$count\" i=\"$headingIndex\">$accum</h>";
601 $headingIndex++;
602 } else {
603 // Single equals sign on its own line, count=0
604 $element = $accum;
605 }
606 } else {
607 // No match, no <h>, just pass down the inner text
608 $element = $accum;
609 }
610 // Unwind the stack
611 $stack->pop();
612 $accum =& $stack->getAccum();
613 $flags = $stack->getFlags();
614 extract( $flags );
615
616 // Append the result to the enclosing accumulator
617 $accum .= $element;
618 // Note that we do NOT increment the input pointer.
619 // This is because the closing linebreak could be the opening linebreak of
620 // another heading. Infinite loops are avoided because the next iteration MUST
621 // hit the heading open case above, which unconditionally increments the
622 // input pointer.
623 } elseif ( $found == 'open' ) {
624 # count opening brace characters
625 $curLen = strlen( $curChar );
626 $count = ( $curLen > 1 ) ?
627 # allow the final character to repeat
628 strspn( $text, $curChar[$curLen - 1], $i + 1 ) + 1 :
629 strspn( $text, $curChar, $i );
630
631 # we need to add to stack only if opening brace count is enough for one of the rules
632 if ( $count >= $rule['min'] ) {
633 # Add it to the stack
634 $piece = [
635 'open' => $curChar,
636 'close' => $rule['end'],
637 'count' => $count,
638 'lineStart' => ( $i > 0 && $text[$i - 1] == "\n" ),
639 ];
640
641 $stack->push( $piece );
642 $accum =& $stack->getAccum();
643 $flags = $stack->getFlags();
644 extract( $flags );
645 } else {
646 # Add literal brace(s)
647 $accum .= htmlspecialchars( str_repeat( $curChar, $count ) );
648 }
649 $i += $count;
650 } elseif ( $found == 'close' ) {
651 $piece = $stack->top;
652 # lets check if there are enough characters for closing brace
653 $maxCount = $piece->count;
654 if ( $piece->close === '}-' && $curChar === '}' ) {
655 $maxCount--; # don't try to match closing '-' as a '}'
656 }
657 $curLen = strlen( $curChar );
658 $count = ( $curLen > 1 ) ? $curLen :
659 strspn( $text, $curChar, $i, $maxCount );
660
661 # check for maximum matching characters (if there are 5 closing
662 # characters, we will probably need only 3 - depending on the rules)
663 $rule = $this->rules[$piece->open];
664 if ( $piece->close === '}-' && $piece->count > 2 ) {
665 # tweak for -{..{{ }}..}-
666 $rule = $this->rules['{'];
667 }
668 if ( $count > $rule['max'] ) {
669 # The specified maximum exists in the callback array, unless the caller
670 # has made an error
671 $matchingCount = $rule['max'];
672 } else {
673 # Count is less than the maximum
674 # Skip any gaps in the callback array to find the true largest match
675 # Need to use array_key_exists not isset because the callback can be null
676 $matchingCount = $count;
677 while ( $matchingCount > 0 && !array_key_exists( $matchingCount, $rule['names'] ) ) {
678 --$matchingCount;
679 }
680 }
681
682 if ( $matchingCount <= 0 ) {
683 # No matching element found in callback array
684 # Output a literal closing brace and continue
685 $endText = substr( $text, $i, $count );
686 $accum .= htmlspecialchars( $endText );
687 $i += $count;
688 continue;
689 }
690 $name = $rule['names'][$matchingCount];
691 if ( $name === null ) {
692 // No element, just literal text
693 $endText = substr( $text, $i, $matchingCount );
694 $element = $piece->breakSyntax( $matchingCount ) . $endText;
695 } else {
696 # Create XML element
697 # Note: $parts is already XML, does not need to be encoded further
698 $parts = $piece->parts;
699 $title = $parts[0]->out;
700 unset( $parts[0] );
701
702 # The invocation is at the start of the line if lineStart is set in
703 # the stack, and all opening brackets are used up.
704 if ( $maxCount == $matchingCount && !empty( $piece->lineStart ) ) {
705 $attr = ' lineStart="1"';
706 } else {
707 $attr = '';
708 }
709
710 $element = "<$name$attr>";
711 $element .= "<title>$title</title>";
712 $argIndex = 1;
713 foreach ( $parts as $part ) {
714 if ( isset( $part->eqpos ) ) {
715 $argName = substr( $part->out, 0, $part->eqpos );
716 $argValue = substr( $part->out, $part->eqpos + 1 );
717 $element .= "<part><name>$argName</name>=<value>$argValue</value></part>";
718 } else {
719 $element .= "<part><name index=\"$argIndex\" /><value>{$part->out}</value></part>";
720 $argIndex++;
721 }
722 }
723 $element .= "</$name>";
724 }
725
726 # Advance input pointer
727 $i += $matchingCount;
728
729 # Unwind the stack
730 $stack->pop();
731 $accum =& $stack->getAccum();
732
733 # Re-add the old stack element if it still has unmatched opening characters remaining
734 if ( $matchingCount < $piece->count ) {
735 $piece->parts = [ new PPDPart ];
736 $piece->count -= $matchingCount;
737 # do we still qualify for any callback with remaining count?
738 $min = $this->rules[$piece->open]['min'];
739 if ( $piece->count >= $min ) {
740 $stack->push( $piece );
741 $accum =& $stack->getAccum();
742 } else {
743 $s = substr( $piece->open, 0, -1 );
744 $s .= str_repeat(
745 substr( $piece->open, -1 ),
746 $piece->count - strlen( $s )
747 );
748 $accum .= $s;
749 }
750 }
751 $flags = $stack->getFlags();
752 extract( $flags );
753
754 # Add XML element to the enclosing accumulator
755 $accum .= $element;
756 } elseif ( $found == 'pipe' ) {
757 $findEquals = true; // shortcut for getFlags()
758 $stack->addPart();
759 $accum =& $stack->getAccum();
760 ++$i;
761 } elseif ( $found == 'equals' ) {
762 $findEquals = false; // shortcut for getFlags()
763 $stack->getCurrentPart()->eqpos = strlen( $accum );
764 $accum .= '=';
765 ++$i;
766 } elseif ( $found == 'dash' ) {
767 $accum .= '-';
768 ++$i;
769 }
770 }
771
772 # Output any remaining unclosed brackets
773 foreach ( $stack->stack as $piece ) {
774 $stack->rootAccum .= $piece->breakSyntax();
775 }
776 $stack->rootAccum .= '</root>';
777 $xml = $stack->rootAccum;
778
779 return $xml;
780 }
781 }
782
783 /**
784 * Stack class to help Preprocessor::preprocessToObj()
785 * @ingroup Parser
786 */
787 class PPDStack {
788 public $stack, $rootAccum;
789
790 /**
791 * @var PPDStack
792 */
793 public $top;
794 public $out;
795 public $elementClass = 'PPDStackElement';
796
797 public static $false = false;
798
799 public function __construct() {
800 $this->stack = [];
801 $this->top = false;
802 $this->rootAccum = '';
803 $this->accum =& $this->rootAccum;
804 }
805
806 /**
807 * @return int
808 */
809 public function count() {
810 return count( $this->stack );
811 }
812
813 public function &getAccum() {
814 return $this->accum;
815 }
816
817 public function getCurrentPart() {
818 if ( $this->top === false ) {
819 return false;
820 } else {
821 return $this->top->getCurrentPart();
822 }
823 }
824
825 public function push( $data ) {
826 if ( $data instanceof $this->elementClass ) {
827 $this->stack[] = $data;
828 } else {
829 $class = $this->elementClass;
830 $this->stack[] = new $class( $data );
831 }
832 $this->top = $this->stack[count( $this->stack ) - 1];
833 $this->accum =& $this->top->getAccum();
834 }
835
836 public function pop() {
837 if ( !count( $this->stack ) ) {
838 throw new MWException( __METHOD__ . ': no elements remaining' );
839 }
840 $temp = array_pop( $this->stack );
841
842 if ( count( $this->stack ) ) {
843 $this->top = $this->stack[count( $this->stack ) - 1];
844 $this->accum =& $this->top->getAccum();
845 } else {
846 $this->top = self::$false;
847 $this->accum =& $this->rootAccum;
848 }
849 return $temp;
850 }
851
852 public function addPart( $s = '' ) {
853 $this->top->addPart( $s );
854 $this->accum =& $this->top->getAccum();
855 }
856
857 /**
858 * @return array
859 */
860 public function getFlags() {
861 if ( !count( $this->stack ) ) {
862 return [
863 'findEquals' => false,
864 'findPipe' => false,
865 'inHeading' => false,
866 ];
867 } else {
868 return $this->top->getFlags();
869 }
870 }
871 }
872
873 /**
874 * @ingroup Parser
875 */
876 class PPDStackElement {
877 /**
878 * @var string Opening character (\n for heading)
879 */
880 public $open;
881
882 /**
883 * @var string Matching closing character
884 */
885 public $close;
886
887 /**
888 * @var int Number of opening characters found (number of "=" for heading)
889 */
890 public $count;
891
892 /**
893 * @var PPDPart[] Array of PPDPart objects describing pipe-separated parts.
894 */
895 public $parts;
896
897 /**
898 * @var bool True if the open char appeared at the start of the input line.
899 * Not set for headings.
900 */
901 public $lineStart;
902
903 public $partClass = 'PPDPart';
904
905 public function __construct( $data = [] ) {
906 $class = $this->partClass;
907 $this->parts = [ new $class ];
908
909 foreach ( $data as $name => $value ) {
910 $this->$name = $value;
911 }
912 }
913
914 public function &getAccum() {
915 return $this->parts[count( $this->parts ) - 1]->out;
916 }
917
918 public function addPart( $s = '' ) {
919 $class = $this->partClass;
920 $this->parts[] = new $class( $s );
921 }
922
923 public function getCurrentPart() {
924 return $this->parts[count( $this->parts ) - 1];
925 }
926
927 /**
928 * @return array
929 */
930 public function getFlags() {
931 $partCount = count( $this->parts );
932 $findPipe = $this->open != "\n" && $this->open != '[';
933 return [
934 'findPipe' => $findPipe,
935 'findEquals' => $findPipe && $partCount > 1 && !isset( $this->parts[$partCount - 1]->eqpos ),
936 'inHeading' => $this->open == "\n",
937 ];
938 }
939
940 /**
941 * Get the output string that would result if the close is not found.
942 *
943 * @param bool|int $openingCount
944 * @return string
945 */
946 public function breakSyntax( $openingCount = false ) {
947 if ( $this->open == "\n" ) {
948 $s = $this->parts[0]->out;
949 } else {
950 if ( $openingCount === false ) {
951 $openingCount = $this->count;
952 }
953 $s = substr( $this->open, 0, -1 );
954 $s .= str_repeat(
955 substr( $this->open, -1 ),
956 $openingCount - strlen( $s )
957 );
958 $first = true;
959 foreach ( $this->parts as $part ) {
960 if ( $first ) {
961 $first = false;
962 } else {
963 $s .= '|';
964 }
965 $s .= $part->out;
966 }
967 }
968 return $s;
969 }
970 }
971
972 /**
973 * @ingroup Parser
974 */
975 class PPDPart {
976 /**
977 * @var string Output accumulator string
978 */
979 public $out;
980
981 // Optional member variables:
982 // eqpos Position of equals sign in output accumulator
983 // commentEnd Past-the-end input pointer for the last comment encountered
984 // visualEnd Past-the-end input pointer for the end of the accumulator minus comments
985
986 public function __construct( $out = '' ) {
987 $this->out = $out;
988 }
989 }
990
991 /**
992 * An expansion frame, used as a context to expand the result of preprocessToObj()
993 * @ingroup Parser
994 */
995 // @codingStandardsIgnoreStart Squiz.Classes.ValidClassName.NotCamelCaps
996 class PPFrame_DOM implements PPFrame {
997 // @codingStandardsIgnoreEnd
998
999 /**
1000 * @var Preprocessor
1001 */
1002 public $preprocessor;
1003
1004 /**
1005 * @var Parser
1006 */
1007 public $parser;
1008
1009 /**
1010 * @var Title
1011 */
1012 public $title;
1013 public $titleCache;
1014
1015 /**
1016 * Hashtable listing templates which are disallowed for expansion in this frame,
1017 * having been encountered previously in parent frames.
1018 */
1019 public $loopCheckHash;
1020
1021 /**
1022 * Recursion depth of this frame, top = 0
1023 * Note that this is NOT the same as expansion depth in expand()
1024 */
1025 public $depth;
1026
1027 private $volatile = false;
1028 private $ttl = null;
1029
1030 /**
1031 * @var array
1032 */
1033 protected $childExpansionCache;
1034
1035 /**
1036 * Construct a new preprocessor frame.
1037 * @param Preprocessor $preprocessor The parent preprocessor
1038 */
1039 public function __construct( $preprocessor ) {
1040 $this->preprocessor = $preprocessor;
1041 $this->parser = $preprocessor->parser;
1042 $this->title = $this->parser->mTitle;
1043 $this->titleCache = [ $this->title ? $this->title->getPrefixedDBkey() : false ];
1044 $this->loopCheckHash = [];
1045 $this->depth = 0;
1046 $this->childExpansionCache = [];
1047 }
1048
1049 /**
1050 * Create a new child frame
1051 * $args is optionally a multi-root PPNode or array containing the template arguments
1052 *
1053 * @param bool|array $args
1054 * @param Title|bool $title
1055 * @param int $indexOffset
1056 * @return PPTemplateFrame_DOM
1057 */
1058 public function newChild( $args = false, $title = false, $indexOffset = 0 ) {
1059 $namedArgs = [];
1060 $numberedArgs = [];
1061 if ( $title === false ) {
1062 $title = $this->title;
1063 }
1064 if ( $args !== false ) {
1065 $xpath = false;
1066 if ( $args instanceof PPNode ) {
1067 $args = $args->node;
1068 }
1069 foreach ( $args as $arg ) {
1070 if ( $arg instanceof PPNode ) {
1071 $arg = $arg->node;
1072 }
1073 if ( !$xpath || $xpath->document !== $arg->ownerDocument ) {
1074 $xpath = new DOMXPath( $arg->ownerDocument );
1075 }
1076
1077 $nameNodes = $xpath->query( 'name', $arg );
1078 $value = $xpath->query( 'value', $arg );
1079 if ( $nameNodes->item( 0 )->hasAttributes() ) {
1080 // Numbered parameter
1081 $index = $nameNodes->item( 0 )->attributes->getNamedItem( 'index' )->textContent;
1082 $index = $index - $indexOffset;
1083 if ( isset( $namedArgs[$index] ) || isset( $numberedArgs[$index] ) ) {
1084 $this->parser->getOutput()->addWarning( wfMessage( 'duplicate-args-warning',
1085 wfEscapeWikiText( $this->title ),
1086 wfEscapeWikiText( $title ),
1087 wfEscapeWikiText( $index ) )->text() );
1088 $this->parser->addTrackingCategory( 'duplicate-args-category' );
1089 }
1090 $numberedArgs[$index] = $value->item( 0 );
1091 unset( $namedArgs[$index] );
1092 } else {
1093 // Named parameter
1094 $name = trim( $this->expand( $nameNodes->item( 0 ), PPFrame::STRIP_COMMENTS ) );
1095 if ( isset( $namedArgs[$name] ) || isset( $numberedArgs[$name] ) ) {
1096 $this->parser->getOutput()->addWarning( wfMessage( 'duplicate-args-warning',
1097 wfEscapeWikiText( $this->title ),
1098 wfEscapeWikiText( $title ),
1099 wfEscapeWikiText( $name ) )->text() );
1100 $this->parser->addTrackingCategory( 'duplicate-args-category' );
1101 }
1102 $namedArgs[$name] = $value->item( 0 );
1103 unset( $numberedArgs[$name] );
1104 }
1105 }
1106 }
1107 return new PPTemplateFrame_DOM( $this->preprocessor, $this, $numberedArgs, $namedArgs, $title );
1108 }
1109
1110 /**
1111 * @throws MWException
1112 * @param string|int $key
1113 * @param string|PPNode_DOM|DOMDocument $root
1114 * @param int $flags
1115 * @return string
1116 */
1117 public function cachedExpand( $key, $root, $flags = 0 ) {
1118 // we don't have a parent, so we don't have a cache
1119 return $this->expand( $root, $flags );
1120 }
1121
1122 /**
1123 * @throws MWException
1124 * @param string|PPNode_DOM|DOMDocument $root
1125 * @param int $flags
1126 * @return string
1127 */
1128 public function expand( $root, $flags = 0 ) {
1129 static $expansionDepth = 0;
1130 if ( is_string( $root ) ) {
1131 return $root;
1132 }
1133
1134 if ( ++$this->parser->mPPNodeCount > $this->parser->mOptions->getMaxPPNodeCount() ) {
1135 $this->parser->limitationWarn( 'node-count-exceeded',
1136 $this->parser->mPPNodeCount,
1137 $this->parser->mOptions->getMaxPPNodeCount()
1138 );
1139 return '<span class="error">Node-count limit exceeded</span>';
1140 }
1141
1142 if ( $expansionDepth > $this->parser->mOptions->getMaxPPExpandDepth() ) {
1143 $this->parser->limitationWarn( 'expansion-depth-exceeded',
1144 $expansionDepth,
1145 $this->parser->mOptions->getMaxPPExpandDepth()
1146 );
1147 return '<span class="error">Expansion depth limit exceeded</span>';
1148 }
1149 ++$expansionDepth;
1150 if ( $expansionDepth > $this->parser->mHighestExpansionDepth ) {
1151 $this->parser->mHighestExpansionDepth = $expansionDepth;
1152 }
1153
1154 if ( $root instanceof PPNode_DOM ) {
1155 $root = $root->node;
1156 }
1157 if ( $root instanceof DOMDocument ) {
1158 $root = $root->documentElement;
1159 }
1160
1161 $outStack = [ '', '' ];
1162 $iteratorStack = [ false, $root ];
1163 $indexStack = [ 0, 0 ];
1164
1165 while ( count( $iteratorStack ) > 1 ) {
1166 $level = count( $outStack ) - 1;
1167 $iteratorNode =& $iteratorStack[$level];
1168 $out =& $outStack[$level];
1169 $index =& $indexStack[$level];
1170
1171 if ( $iteratorNode instanceof PPNode_DOM ) {
1172 $iteratorNode = $iteratorNode->node;
1173 }
1174
1175 if ( is_array( $iteratorNode ) ) {
1176 if ( $index >= count( $iteratorNode ) ) {
1177 // All done with this iterator
1178 $iteratorStack[$level] = false;
1179 $contextNode = false;
1180 } else {
1181 $contextNode = $iteratorNode[$index];
1182 $index++;
1183 }
1184 } elseif ( $iteratorNode instanceof DOMNodeList ) {
1185 if ( $index >= $iteratorNode->length ) {
1186 // All done with this iterator
1187 $iteratorStack[$level] = false;
1188 $contextNode = false;
1189 } else {
1190 $contextNode = $iteratorNode->item( $index );
1191 $index++;
1192 }
1193 } else {
1194 // Copy to $contextNode and then delete from iterator stack,
1195 // because this is not an iterator but we do have to execute it once
1196 $contextNode = $iteratorStack[$level];
1197 $iteratorStack[$level] = false;
1198 }
1199
1200 if ( $contextNode instanceof PPNode_DOM ) {
1201 $contextNode = $contextNode->node;
1202 }
1203
1204 $newIterator = false;
1205
1206 if ( $contextNode === false ) {
1207 // nothing to do
1208 } elseif ( is_string( $contextNode ) ) {
1209 $out .= $contextNode;
1210 } elseif ( is_array( $contextNode ) || $contextNode instanceof DOMNodeList ) {
1211 $newIterator = $contextNode;
1212 } elseif ( $contextNode instanceof DOMNode ) {
1213 if ( $contextNode->nodeType == XML_TEXT_NODE ) {
1214 $out .= $contextNode->nodeValue;
1215 } elseif ( $contextNode->nodeName == 'template' ) {
1216 # Double-brace expansion
1217 $xpath = new DOMXPath( $contextNode->ownerDocument );
1218 $titles = $xpath->query( 'title', $contextNode );
1219 $title = $titles->item( 0 );
1220 $parts = $xpath->query( 'part', $contextNode );
1221 if ( $flags & PPFrame::NO_TEMPLATES ) {
1222 $newIterator = $this->virtualBracketedImplode( '{{', '|', '}}', $title, $parts );
1223 } else {
1224 $lineStart = $contextNode->getAttribute( 'lineStart' );
1225 $params = [
1226 'title' => new PPNode_DOM( $title ),
1227 'parts' => new PPNode_DOM( $parts ),
1228 'lineStart' => $lineStart ];
1229 $ret = $this->parser->braceSubstitution( $params, $this );
1230 if ( isset( $ret['object'] ) ) {
1231 $newIterator = $ret['object'];
1232 } else {
1233 $out .= $ret['text'];
1234 }
1235 }
1236 } elseif ( $contextNode->nodeName == 'tplarg' ) {
1237 # Triple-brace expansion
1238 $xpath = new DOMXPath( $contextNode->ownerDocument );
1239 $titles = $xpath->query( 'title', $contextNode );
1240 $title = $titles->item( 0 );
1241 $parts = $xpath->query( 'part', $contextNode );
1242 if ( $flags & PPFrame::NO_ARGS ) {
1243 $newIterator = $this->virtualBracketedImplode( '{{{', '|', '}}}', $title, $parts );
1244 } else {
1245 $params = [
1246 'title' => new PPNode_DOM( $title ),
1247 'parts' => new PPNode_DOM( $parts ) ];
1248 $ret = $this->parser->argSubstitution( $params, $this );
1249 if ( isset( $ret['object'] ) ) {
1250 $newIterator = $ret['object'];
1251 } else {
1252 $out .= $ret['text'];
1253 }
1254 }
1255 } elseif ( $contextNode->nodeName == 'comment' ) {
1256 # HTML-style comment
1257 # Remove it in HTML, pre+remove and STRIP_COMMENTS modes
1258 # Not in RECOVER_COMMENTS mode (msgnw) though.
1259 if ( ( $this->parser->ot['html']
1260 || ( $this->parser->ot['pre'] && $this->parser->mOptions->getRemoveComments() )
1261 || ( $flags & PPFrame::STRIP_COMMENTS )
1262 ) && !( $flags & PPFrame::RECOVER_COMMENTS )
1263 ) {
1264 $out .= '';
1265 } elseif ( $this->parser->ot['wiki'] && !( $flags & PPFrame::RECOVER_COMMENTS ) ) {
1266 # Add a strip marker in PST mode so that pstPass2() can
1267 # run some old-fashioned regexes on the result.
1268 # Not in RECOVER_COMMENTS mode (extractSections) though.
1269 $out .= $this->parser->insertStripItem( $contextNode->textContent );
1270 } else {
1271 # Recover the literal comment in RECOVER_COMMENTS and pre+no-remove
1272 $out .= $contextNode->textContent;
1273 }
1274 } elseif ( $contextNode->nodeName == 'ignore' ) {
1275 # Output suppression used by <includeonly> etc.
1276 # OT_WIKI will only respect <ignore> in substed templates.
1277 # The other output types respect it unless NO_IGNORE is set.
1278 # extractSections() sets NO_IGNORE and so never respects it.
1279 if ( ( !isset( $this->parent ) && $this->parser->ot['wiki'] )
1280 || ( $flags & PPFrame::NO_IGNORE )
1281 ) {
1282 $out .= $contextNode->textContent;
1283 } else {
1284 $out .= '';
1285 }
1286 } elseif ( $contextNode->nodeName == 'ext' ) {
1287 # Extension tag
1288 $xpath = new DOMXPath( $contextNode->ownerDocument );
1289 $names = $xpath->query( 'name', $contextNode );
1290 $attrs = $xpath->query( 'attr', $contextNode );
1291 $inners = $xpath->query( 'inner', $contextNode );
1292 $closes = $xpath->query( 'close', $contextNode );
1293 if ( $flags & PPFrame::NO_TAGS ) {
1294 $s = '<' . $this->expand( $names->item( 0 ), $flags );
1295 if ( $attrs->length > 0 ) {
1296 $s .= $this->expand( $attrs->item( 0 ), $flags );
1297 }
1298 if ( $inners->length > 0 ) {
1299 $s .= '>' . $this->expand( $inners->item( 0 ), $flags );
1300 if ( $closes->length > 0 ) {
1301 $s .= $this->expand( $closes->item( 0 ), $flags );
1302 }
1303 } else {
1304 $s .= '/>';
1305 }
1306 $out .= $s;
1307 } else {
1308 $params = [
1309 'name' => new PPNode_DOM( $names->item( 0 ) ),
1310 'attr' => $attrs->length > 0 ? new PPNode_DOM( $attrs->item( 0 ) ) : null,
1311 'inner' => $inners->length > 0 ? new PPNode_DOM( $inners->item( 0 ) ) : null,
1312 'close' => $closes->length > 0 ? new PPNode_DOM( $closes->item( 0 ) ) : null,
1313 ];
1314 $out .= $this->parser->extensionSubstitution( $params, $this );
1315 }
1316 } elseif ( $contextNode->nodeName == 'h' ) {
1317 # Heading
1318 $s = $this->expand( $contextNode->childNodes, $flags );
1319
1320 # Insert a heading marker only for <h> children of <root>
1321 # This is to stop extractSections from going over multiple tree levels
1322 if ( $contextNode->parentNode->nodeName == 'root' && $this->parser->ot['html'] ) {
1323 # Insert heading index marker
1324 $headingIndex = $contextNode->getAttribute( 'i' );
1325 $titleText = $this->title->getPrefixedDBkey();
1326 $this->parser->mHeadings[] = [ $titleText, $headingIndex ];
1327 $serial = count( $this->parser->mHeadings ) - 1;
1328 $marker = Parser::MARKER_PREFIX . "-h-$serial-" . Parser::MARKER_SUFFIX;
1329 $count = $contextNode->getAttribute( 'level' );
1330 $s = substr( $s, 0, $count ) . $marker . substr( $s, $count );
1331 $this->parser->mStripState->addGeneral( $marker, '' );
1332 }
1333 $out .= $s;
1334 } else {
1335 # Generic recursive expansion
1336 $newIterator = $contextNode->childNodes;
1337 }
1338 } else {
1339 throw new MWException( __METHOD__ . ': Invalid parameter type' );
1340 }
1341
1342 if ( $newIterator !== false ) {
1343 if ( $newIterator instanceof PPNode_DOM ) {
1344 $newIterator = $newIterator->node;
1345 }
1346 $outStack[] = '';
1347 $iteratorStack[] = $newIterator;
1348 $indexStack[] = 0;
1349 } elseif ( $iteratorStack[$level] === false ) {
1350 // Return accumulated value to parent
1351 // With tail recursion
1352 while ( $iteratorStack[$level] === false && $level > 0 ) {
1353 $outStack[$level - 1] .= $out;
1354 array_pop( $outStack );
1355 array_pop( $iteratorStack );
1356 array_pop( $indexStack );
1357 $level--;
1358 }
1359 }
1360 }
1361 --$expansionDepth;
1362 return $outStack[0];
1363 }
1364
1365 /**
1366 * @param string $sep
1367 * @param int $flags
1368 * @param string|PPNode_DOM|DOMDocument $args,...
1369 * @return string
1370 */
1371 public function implodeWithFlags( $sep, $flags /*, ... */ ) {
1372 $args = array_slice( func_get_args(), 2 );
1373
1374 $first = true;
1375 $s = '';
1376 foreach ( $args as $root ) {
1377 if ( $root instanceof PPNode_DOM ) {
1378 $root = $root->node;
1379 }
1380 if ( !is_array( $root ) && !( $root instanceof DOMNodeList ) ) {
1381 $root = [ $root ];
1382 }
1383 foreach ( $root as $node ) {
1384 if ( $first ) {
1385 $first = false;
1386 } else {
1387 $s .= $sep;
1388 }
1389 $s .= $this->expand( $node, $flags );
1390 }
1391 }
1392 return $s;
1393 }
1394
1395 /**
1396 * Implode with no flags specified
1397 * This previously called implodeWithFlags but has now been inlined to reduce stack depth
1398 *
1399 * @param string $sep
1400 * @param string|PPNode_DOM|DOMDocument $args,...
1401 * @return string
1402 */
1403 public function implode( $sep /*, ... */ ) {
1404 $args = array_slice( func_get_args(), 1 );
1405
1406 $first = true;
1407 $s = '';
1408 foreach ( $args as $root ) {
1409 if ( $root instanceof PPNode_DOM ) {
1410 $root = $root->node;
1411 }
1412 if ( !is_array( $root ) && !( $root instanceof DOMNodeList ) ) {
1413 $root = [ $root ];
1414 }
1415 foreach ( $root as $node ) {
1416 if ( $first ) {
1417 $first = false;
1418 } else {
1419 $s .= $sep;
1420 }
1421 $s .= $this->expand( $node );
1422 }
1423 }
1424 return $s;
1425 }
1426
1427 /**
1428 * Makes an object that, when expand()ed, will be the same as one obtained
1429 * with implode()
1430 *
1431 * @param string $sep
1432 * @param string|PPNode_DOM|DOMDocument $args,...
1433 * @return array
1434 */
1435 public function virtualImplode( $sep /*, ... */ ) {
1436 $args = array_slice( func_get_args(), 1 );
1437 $out = [];
1438 $first = true;
1439
1440 foreach ( $args as $root ) {
1441 if ( $root instanceof PPNode_DOM ) {
1442 $root = $root->node;
1443 }
1444 if ( !is_array( $root ) && !( $root instanceof DOMNodeList ) ) {
1445 $root = [ $root ];
1446 }
1447 foreach ( $root as $node ) {
1448 if ( $first ) {
1449 $first = false;
1450 } else {
1451 $out[] = $sep;
1452 }
1453 $out[] = $node;
1454 }
1455 }
1456 return $out;
1457 }
1458
1459 /**
1460 * Virtual implode with brackets
1461 * @param string $start
1462 * @param string $sep
1463 * @param string $end
1464 * @param string|PPNode_DOM|DOMDocument $args,...
1465 * @return array
1466 */
1467 public function virtualBracketedImplode( $start, $sep, $end /*, ... */ ) {
1468 $args = array_slice( func_get_args(), 3 );
1469 $out = [ $start ];
1470 $first = true;
1471
1472 foreach ( $args as $root ) {
1473 if ( $root instanceof PPNode_DOM ) {
1474 $root = $root->node;
1475 }
1476 if ( !is_array( $root ) && !( $root instanceof DOMNodeList ) ) {
1477 $root = [ $root ];
1478 }
1479 foreach ( $root as $node ) {
1480 if ( $first ) {
1481 $first = false;
1482 } else {
1483 $out[] = $sep;
1484 }
1485 $out[] = $node;
1486 }
1487 }
1488 $out[] = $end;
1489 return $out;
1490 }
1491
1492 public function __toString() {
1493 return 'frame{}';
1494 }
1495
1496 public function getPDBK( $level = false ) {
1497 if ( $level === false ) {
1498 return $this->title->getPrefixedDBkey();
1499 } else {
1500 return isset( $this->titleCache[$level] ) ? $this->titleCache[$level] : false;
1501 }
1502 }
1503
1504 /**
1505 * @return array
1506 */
1507 public function getArguments() {
1508 return [];
1509 }
1510
1511 /**
1512 * @return array
1513 */
1514 public function getNumberedArguments() {
1515 return [];
1516 }
1517
1518 /**
1519 * @return array
1520 */
1521 public function getNamedArguments() {
1522 return [];
1523 }
1524
1525 /**
1526 * Returns true if there are no arguments in this frame
1527 *
1528 * @return bool
1529 */
1530 public function isEmpty() {
1531 return true;
1532 }
1533
1534 /**
1535 * @param int|string $name
1536 * @return bool Always false in this implementation.
1537 */
1538 public function getArgument( $name ) {
1539 return false;
1540 }
1541
1542 /**
1543 * Returns true if the infinite loop check is OK, false if a loop is detected
1544 *
1545 * @param Title $title
1546 * @return bool
1547 */
1548 public function loopCheck( $title ) {
1549 return !isset( $this->loopCheckHash[$title->getPrefixedDBkey()] );
1550 }
1551
1552 /**
1553 * Return true if the frame is a template frame
1554 *
1555 * @return bool
1556 */
1557 public function isTemplate() {
1558 return false;
1559 }
1560
1561 /**
1562 * Get a title of frame
1563 *
1564 * @return Title
1565 */
1566 public function getTitle() {
1567 return $this->title;
1568 }
1569
1570 /**
1571 * Set the volatile flag
1572 *
1573 * @param bool $flag
1574 */
1575 public function setVolatile( $flag = true ) {
1576 $this->volatile = $flag;
1577 }
1578
1579 /**
1580 * Get the volatile flag
1581 *
1582 * @return bool
1583 */
1584 public function isVolatile() {
1585 return $this->volatile;
1586 }
1587
1588 /**
1589 * Set the TTL
1590 *
1591 * @param int $ttl
1592 */
1593 public function setTTL( $ttl ) {
1594 if ( $ttl !== null && ( $this->ttl === null || $ttl < $this->ttl ) ) {
1595 $this->ttl = $ttl;
1596 }
1597 }
1598
1599 /**
1600 * Get the TTL
1601 *
1602 * @return int|null
1603 */
1604 public function getTTL() {
1605 return $this->ttl;
1606 }
1607 }
1608
1609 /**
1610 * Expansion frame with template arguments
1611 * @ingroup Parser
1612 */
1613 // @codingStandardsIgnoreStart Squiz.Classes.ValidClassName.NotCamelCaps
1614 class PPTemplateFrame_DOM extends PPFrame_DOM {
1615 // @codingStandardsIgnoreEnd
1616
1617 public $numberedArgs, $namedArgs;
1618
1619 /**
1620 * @var PPFrame_DOM
1621 */
1622 public $parent;
1623 public $numberedExpansionCache, $namedExpansionCache;
1624
1625 /**
1626 * @param Preprocessor $preprocessor
1627 * @param bool|PPFrame_DOM $parent
1628 * @param array $numberedArgs
1629 * @param array $namedArgs
1630 * @param bool|Title $title
1631 */
1632 public function __construct( $preprocessor, $parent = false, $numberedArgs = [],
1633 $namedArgs = [], $title = false
1634 ) {
1635 parent::__construct( $preprocessor );
1636
1637 $this->parent = $parent;
1638 $this->numberedArgs = $numberedArgs;
1639 $this->namedArgs = $namedArgs;
1640 $this->title = $title;
1641 $pdbk = $title ? $title->getPrefixedDBkey() : false;
1642 $this->titleCache = $parent->titleCache;
1643 $this->titleCache[] = $pdbk;
1644 $this->loopCheckHash = /*clone*/ $parent->loopCheckHash;
1645 if ( $pdbk !== false ) {
1646 $this->loopCheckHash[$pdbk] = true;
1647 }
1648 $this->depth = $parent->depth + 1;
1649 $this->numberedExpansionCache = $this->namedExpansionCache = [];
1650 }
1651
1652 public function __toString() {
1653 $s = 'tplframe{';
1654 $first = true;
1655 $args = $this->numberedArgs + $this->namedArgs;
1656 foreach ( $args as $name => $value ) {
1657 if ( $first ) {
1658 $first = false;
1659 } else {
1660 $s .= ', ';
1661 }
1662 $s .= "\"$name\":\"" .
1663 str_replace( '"', '\\"', $value->ownerDocument->saveXML( $value ) ) . '"';
1664 }
1665 $s .= '}';
1666 return $s;
1667 }
1668
1669 /**
1670 * @throws MWException
1671 * @param string|int $key
1672 * @param string|PPNode_DOM|DOMDocument $root
1673 * @param int $flags
1674 * @return string
1675 */
1676 public function cachedExpand( $key, $root, $flags = 0 ) {
1677 if ( isset( $this->parent->childExpansionCache[$key] ) ) {
1678 return $this->parent->childExpansionCache[$key];
1679 }
1680 $retval = $this->expand( $root, $flags );
1681 if ( !$this->isVolatile() ) {
1682 $this->parent->childExpansionCache[$key] = $retval;
1683 }
1684 return $retval;
1685 }
1686
1687 /**
1688 * Returns true if there are no arguments in this frame
1689 *
1690 * @return bool
1691 */
1692 public function isEmpty() {
1693 return !count( $this->numberedArgs ) && !count( $this->namedArgs );
1694 }
1695
1696 public function getArguments() {
1697 $arguments = [];
1698 foreach ( array_merge(
1699 array_keys( $this->numberedArgs ),
1700 array_keys( $this->namedArgs ) ) as $key ) {
1701 $arguments[$key] = $this->getArgument( $key );
1702 }
1703 return $arguments;
1704 }
1705
1706 public function getNumberedArguments() {
1707 $arguments = [];
1708 foreach ( array_keys( $this->numberedArgs ) as $key ) {
1709 $arguments[$key] = $this->getArgument( $key );
1710 }
1711 return $arguments;
1712 }
1713
1714 public function getNamedArguments() {
1715 $arguments = [];
1716 foreach ( array_keys( $this->namedArgs ) as $key ) {
1717 $arguments[$key] = $this->getArgument( $key );
1718 }
1719 return $arguments;
1720 }
1721
1722 /**
1723 * @param int $index
1724 * @return string|bool
1725 */
1726 public function getNumberedArgument( $index ) {
1727 if ( !isset( $this->numberedArgs[$index] ) ) {
1728 return false;
1729 }
1730 if ( !isset( $this->numberedExpansionCache[$index] ) ) {
1731 # No trimming for unnamed arguments
1732 $this->numberedExpansionCache[$index] = $this->parent->expand(
1733 $this->numberedArgs[$index],
1734 PPFrame::STRIP_COMMENTS
1735 );
1736 }
1737 return $this->numberedExpansionCache[$index];
1738 }
1739
1740 /**
1741 * @param string $name
1742 * @return string|bool
1743 */
1744 public function getNamedArgument( $name ) {
1745 if ( !isset( $this->namedArgs[$name] ) ) {
1746 return false;
1747 }
1748 if ( !isset( $this->namedExpansionCache[$name] ) ) {
1749 # Trim named arguments post-expand, for backwards compatibility
1750 $this->namedExpansionCache[$name] = trim(
1751 $this->parent->expand( $this->namedArgs[$name], PPFrame::STRIP_COMMENTS ) );
1752 }
1753 return $this->namedExpansionCache[$name];
1754 }
1755
1756 /**
1757 * @param int|string $name
1758 * @return string|bool
1759 */
1760 public function getArgument( $name ) {
1761 $text = $this->getNumberedArgument( $name );
1762 if ( $text === false ) {
1763 $text = $this->getNamedArgument( $name );
1764 }
1765 return $text;
1766 }
1767
1768 /**
1769 * Return true if the frame is a template frame
1770 *
1771 * @return bool
1772 */
1773 public function isTemplate() {
1774 return true;
1775 }
1776
1777 public function setVolatile( $flag = true ) {
1778 parent::setVolatile( $flag );
1779 $this->parent->setVolatile( $flag );
1780 }
1781
1782 public function setTTL( $ttl ) {
1783 parent::setTTL( $ttl );
1784 $this->parent->setTTL( $ttl );
1785 }
1786 }
1787
1788 /**
1789 * Expansion frame with custom arguments
1790 * @ingroup Parser
1791 */
1792 // @codingStandardsIgnoreStart Squiz.Classes.ValidClassName.NotCamelCaps
1793 class PPCustomFrame_DOM extends PPFrame_DOM {
1794 // @codingStandardsIgnoreEnd
1795
1796 public $args;
1797
1798 public function __construct( $preprocessor, $args ) {
1799 parent::__construct( $preprocessor );
1800 $this->args = $args;
1801 }
1802
1803 public function __toString() {
1804 $s = 'cstmframe{';
1805 $first = true;
1806 foreach ( $this->args as $name => $value ) {
1807 if ( $first ) {
1808 $first = false;
1809 } else {
1810 $s .= ', ';
1811 }
1812 $s .= "\"$name\":\"" .
1813 str_replace( '"', '\\"', $value->__toString() ) . '"';
1814 }
1815 $s .= '}';
1816 return $s;
1817 }
1818
1819 /**
1820 * @return bool
1821 */
1822 public function isEmpty() {
1823 return !count( $this->args );
1824 }
1825
1826 /**
1827 * @param int|string $index
1828 * @return string|bool
1829 */
1830 public function getArgument( $index ) {
1831 if ( !isset( $this->args[$index] ) ) {
1832 return false;
1833 }
1834 return $this->args[$index];
1835 }
1836
1837 public function getArguments() {
1838 return $this->args;
1839 }
1840 }
1841
1842 /**
1843 * @ingroup Parser
1844 */
1845 // @codingStandardsIgnoreStart Squiz.Classes.ValidClassName.NotCamelCaps
1846 class PPNode_DOM implements PPNode {
1847 // @codingStandardsIgnoreEnd
1848
1849 /**
1850 * @var DOMElement
1851 */
1852 public $node;
1853 public $xpath;
1854
1855 public function __construct( $node, $xpath = false ) {
1856 $this->node = $node;
1857 }
1858
1859 /**
1860 * @return DOMXPath
1861 */
1862 public function getXPath() {
1863 if ( $this->xpath === null ) {
1864 $this->xpath = new DOMXPath( $this->node->ownerDocument );
1865 }
1866 return $this->xpath;
1867 }
1868
1869 public function __toString() {
1870 if ( $this->node instanceof DOMNodeList ) {
1871 $s = '';
1872 foreach ( $this->node as $node ) {
1873 $s .= $node->ownerDocument->saveXML( $node );
1874 }
1875 } else {
1876 $s = $this->node->ownerDocument->saveXML( $this->node );
1877 }
1878 return $s;
1879 }
1880
1881 /**
1882 * @return bool|PPNode_DOM
1883 */
1884 public function getChildren() {
1885 return $this->node->childNodes ? new self( $this->node->childNodes ) : false;
1886 }
1887
1888 /**
1889 * @return bool|PPNode_DOM
1890 */
1891 public function getFirstChild() {
1892 return $this->node->firstChild ? new self( $this->node->firstChild ) : false;
1893 }
1894
1895 /**
1896 * @return bool|PPNode_DOM
1897 */
1898 public function getNextSibling() {
1899 return $this->node->nextSibling ? new self( $this->node->nextSibling ) : false;
1900 }
1901
1902 /**
1903 * @param string $type
1904 *
1905 * @return bool|PPNode_DOM
1906 */
1907 public function getChildrenOfType( $type ) {
1908 return new self( $this->getXPath()->query( $type, $this->node ) );
1909 }
1910
1911 /**
1912 * @return int
1913 */
1914 public function getLength() {
1915 if ( $this->node instanceof DOMNodeList ) {
1916 return $this->node->length;
1917 } else {
1918 return false;
1919 }
1920 }
1921
1922 /**
1923 * @param int $i
1924 * @return bool|PPNode_DOM
1925 */
1926 public function item( $i ) {
1927 $item = $this->node->item( $i );
1928 return $item ? new self( $item ) : false;
1929 }
1930
1931 /**
1932 * @return string
1933 */
1934 public function getName() {
1935 if ( $this->node instanceof DOMNodeList ) {
1936 return '#nodelist';
1937 } else {
1938 return $this->node->nodeName;
1939 }
1940 }
1941
1942 /**
1943 * Split a "<part>" node into an associative array containing:
1944 * - name PPNode name
1945 * - index String index
1946 * - value PPNode value
1947 *
1948 * @throws MWException
1949 * @return array
1950 */
1951 public function splitArg() {
1952 $xpath = $this->getXPath();
1953 $names = $xpath->query( 'name', $this->node );
1954 $values = $xpath->query( 'value', $this->node );
1955 if ( !$names->length || !$values->length ) {
1956 throw new MWException( 'Invalid brace node passed to ' . __METHOD__ );
1957 }
1958 $name = $names->item( 0 );
1959 $index = $name->getAttribute( 'index' );
1960 return [
1961 'name' => new self( $name ),
1962 'index' => $index,
1963 'value' => new self( $values->item( 0 ) ) ];
1964 }
1965
1966 /**
1967 * Split an "<ext>" node into an associative array containing name, attr, inner and close
1968 * All values in the resulting array are PPNodes. Inner and close are optional.
1969 *
1970 * @throws MWException
1971 * @return array
1972 */
1973 public function splitExt() {
1974 $xpath = $this->getXPath();
1975 $names = $xpath->query( 'name', $this->node );
1976 $attrs = $xpath->query( 'attr', $this->node );
1977 $inners = $xpath->query( 'inner', $this->node );
1978 $closes = $xpath->query( 'close', $this->node );
1979 if ( !$names->length || !$attrs->length ) {
1980 throw new MWException( 'Invalid ext node passed to ' . __METHOD__ );
1981 }
1982 $parts = [
1983 'name' => new self( $names->item( 0 ) ),
1984 'attr' => new self( $attrs->item( 0 ) ) ];
1985 if ( $inners->length ) {
1986 $parts['inner'] = new self( $inners->item( 0 ) );
1987 }
1988 if ( $closes->length ) {
1989 $parts['close'] = new self( $closes->item( 0 ) );
1990 }
1991 return $parts;
1992 }
1993
1994 /**
1995 * Split a "<h>" node
1996 * @throws MWException
1997 * @return array
1998 */
1999 public function splitHeading() {
2000 if ( $this->getName() !== 'h' ) {
2001 throw new MWException( 'Invalid h node passed to ' . __METHOD__ );
2002 }
2003 return [
2004 'i' => $this->node->getAttribute( 'i' ),
2005 'level' => $this->node->getAttribute( 'level' ),
2006 'contents' => $this->getChildren()
2007 ];
2008 }
2009 }
Wiklou, le Wiki du Wiklou