dépôts / lhc / web / wiklou.git / blob
? search:
summary | shortlog | log | commit | commitdiff | tree
history | raw | HEAD
Test ApiUserrights
[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 // phpcs:ignore Squiz.Classes.ValidClassName.NotCamelCaps
28 class Preprocessor_DOM extends Preprocessor {
29
30 /**
31 * @var Parser
32 */
33 public $parser;
34
35 public $memoryLimit;
36
37 const CACHE_PREFIX = 'preprocess-xml';
38
39 public function __construct( $parser ) {
40 $this->parser = $parser;
41 $mem = ini_get( 'memory_limit' );
42 $this->memoryLimit = false;
43 if ( strval( $mem ) !== '' && $mem != -1 ) {
44 if ( preg_match( '/^\d+$/', $mem ) ) {
45 $this->memoryLimit = $mem;
46 } elseif ( preg_match( '/^(\d+)M$/i', $mem, $m ) ) {
47 $this->memoryLimit = $m[1] * 1048576;
48 }
49 }
50 }
51
52 /**
53 * @return PPFrame_DOM
54 */
55 public function newFrame() {
56 return new PPFrame_DOM( $this );
57 }
58
59 /**
60 * @param array $args
61 * @return PPCustomFrame_DOM
62 */
63 public function newCustomFrame( $args ) {
64 return new PPCustomFrame_DOM( $this, $args );
65 }
66
67 /**
68 * @param array $values
69 * @return PPNode_DOM
70 * @throws MWException
71 */
72 public function newPartNodeArray( $values ) {
73 // NOTE: DOM manipulation is slower than building & parsing XML! (or so Tim sais)
74 $xml = "<list>";
75
76 foreach ( $values as $k => $val ) {
77 if ( is_int( $k ) ) {
78 $xml .= "<part><name index=\"$k\"/><value>"
79 . htmlspecialchars( $val ) . "</value></part>";
80 } else {
81 $xml .= "<part><name>" . htmlspecialchars( $k )
82 . "</name>=<value>" . htmlspecialchars( $val ) . "</value></part>";
83 }
84 }
85
86 $xml .= "</list>";
87
88 $dom = new DOMDocument();
89 Wikimedia\suppressWarnings();
90 $result = $dom->loadXML( $xml );
91 Wikimedia\restoreWarnings();
92 if ( !$result ) {
93 // Try running the XML through UtfNormal to get rid of invalid characters
94 $xml = UtfNormal\Validator::cleanUp( $xml );
95 // 1 << 19 == XML_PARSE_HUGE, needed so newer versions of libxml2
96 // don't barf when the XML is >256 levels deep
97 $result = $dom->loadXML( $xml, 1 << 19 );
98 }
99
100 if ( !$result ) {
101 throw new MWException( 'Parameters passed to ' . __METHOD__ . ' result in invalid XML' );
102 }
103
104 $root = $dom->documentElement;
105 $node = new PPNode_DOM( $root->childNodes );
106 return $node;
107 }
108
109 /**
110 * @throws MWException
111 * @return bool
112 */
113 public function memCheck() {
114 if ( $this->memoryLimit === false ) {
115 return true;
116 }
117 $usage = memory_get_usage();
118 if ( $usage > $this->memoryLimit * 0.9 ) {
119 $limit = intval( $this->memoryLimit * 0.9 / 1048576 + 0.5 );
120 throw new MWException( "Preprocessor hit 90% memory limit ($limit MB)" );
121 }
122 return $usage <= $this->memoryLimit * 0.8;
123 }
124
125 /**
126 * Preprocess some wikitext and return the document tree.
127 * This is the ghost of Parser::replace_variables().
128 *
129 * @param string $text The text to parse
130 * @param int $flags Bitwise combination of:
131 * Parser::PTD_FOR_INCLUSION Handle "<noinclude>" and "<includeonly>"
132 * as if the text is being included. Default
133 * is to assume a direct page view.
134 *
135 * The generated DOM tree must depend only on the input text and the flags.
136 * The DOM tree must be the same in OT_HTML and OT_WIKI mode, to avoid a regression of T6899.
137 *
138 * Any flag added to the $flags parameter here, or any other parameter liable to cause a
139 * change in the DOM tree for a given text, must be passed through the section identifier
140 * in the section edit link and thus back to extractSections().
141 *
142 * The output of this function is currently only cached in process memory, but a persistent
143 * cache may be implemented at a later date which takes further advantage of these strict
144 * dependency requirements.
145 *
146 * @throws MWException
147 * @return PPNode_DOM
148 */
149 public function preprocessToObj( $text, $flags = 0 ) {
150 $xml = $this->cacheGetTree( $text, $flags );
151 if ( $xml === false ) {
152 $xml = $this->preprocessToXml( $text, $flags );
153 $this->cacheSetTree( $text, $flags, $xml );
154 }
155
156 // Fail if the number of elements exceeds acceptable limits
157 // Do not attempt to generate the DOM
158 $this->parser->mGeneratedPPNodeCount += substr_count( $xml, '<' );
159 $max = $this->parser->mOptions->getMaxGeneratedPPNodeCount();
160 if ( $this->parser->mGeneratedPPNodeCount > $max ) {
161 // if ( $cacheable ) { ... }
162 throw new MWException( __METHOD__ . ': generated node count limit exceeded' );
163 }
164
165 $dom = new DOMDocument;
166 Wikimedia\suppressWarnings();
167 $result = $dom->loadXML( $xml );
168 Wikimedia\restoreWarnings();
169 if ( !$result ) {
170 // Try running the XML through UtfNormal to get rid of invalid characters
171 $xml = UtfNormal\Validator::cleanUp( $xml );
172 // 1 << 19 == XML_PARSE_HUGE, needed so newer versions of libxml2
173 // don't barf when the XML is >256 levels deep.
174 $result = $dom->loadXML( $xml, 1 << 19 );
175 }
176 if ( $result ) {
177 $obj = new PPNode_DOM( $dom->documentElement );
178 }
179
180 // if ( $cacheable ) { ... }
181
182 if ( !$result ) {
183 throw new MWException( __METHOD__ . ' generated invalid XML' );
184 }
185 return $obj;
186 }
187
188 /**
189 * @param string $text
190 * @param int $flags
191 * @return string
192 */
193 public function preprocessToXml( $text, $flags = 0 ) {
194 global $wgDisableLangConversion;
195
196 $forInclusion = $flags & Parser::PTD_FOR_INCLUSION;
197
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
207 ) {
208 $enableOnlyinclude = true;
209 }
210 } else {
211 $ignoredTags = [ 'noinclude', '/noinclude', 'onlyinclude', '/onlyinclude' ];
212 $ignoredElements = [ 'includeonly' ];
213 $xmlishElements[] = 'includeonly';
214 }
215 $xmlishRegex = implode( '|', array_merge( $xmlishElements, $ignoredTags ) );
216
217 // Use "A" modifier (anchored) instead of "^", because ^ doesn't work with an offset
218 $elementsRegex = "~($xmlishRegex)(?:\s|\/>|>)|(!--)~iA";
219
220 $stack = new PPDStack;
221
222 $searchBase = "[{<\n"; # }
223 if ( !$wgDisableLangConversion ) {
224 $searchBase .= '-';
225 }
226
227 // For fast reverse searches
228 $revText = strrev( $text );
229 $lengthText = strlen( $text );
230
231 // Input pointer, starts out pointing to a pseudo-newline before the start
232 $i = 0;
233 // Current accumulator
234 $accum =& $stack->getAccum();
235 $accum = '<root>';
236 // True to find equals signs in arguments
237 $findEquals = false;
238 // True to take notice of pipe characters
239 $findPipe = false;
240 $headingIndex = 1;
241 // True if $i is inside a possible heading
242 $inHeading = false;
243 // True if there are no more greater-than (>) signs right of $i
244 $noMoreGT = false;
245 // Map of tag name => true if there are no more closing tags of given type right of $i
246 $noMoreClosingTag = [];
247 // True to ignore all input up to the next <onlyinclude>
248 $findOnlyinclude = $enableOnlyinclude;
249 // Do a line-start run without outputting an LF character
250 $fakeLineStart = true;
251
252 while ( true ) {
253 // $this->memCheck();
254
255 if ( $findOnlyinclude ) {
256 // Ignore all input up to the next <onlyinclude>
257 $startPos = strpos( $text, '<onlyinclude>', $i );
258 if ( $startPos === false ) {
259 // Ignored section runs to the end
260 $accum .= '<ignore>' . htmlspecialchars( substr( $text, $i ) ) . '</ignore>';
261 break;
262 }
263 $tagEndPos = $startPos + strlen( '<onlyinclude>' ); // past-the-end
264 $accum .= '<ignore>' . htmlspecialchars( substr( $text, $i, $tagEndPos - $i ) ) . '</ignore>';
265 $i = $tagEndPos;
266 $findOnlyinclude = false;
267 }
268
269 if ( $fakeLineStart ) {
270 $found = 'line-start';
271 $curChar = '';
272 } else {
273 # Find next opening brace, closing brace or pipe
274 $search = $searchBase;
275 if ( $stack->top === false ) {
276 $currentClosing = '';
277 } else {
278 $currentClosing = $stack->top->close;
279 $search .= $currentClosing;
280 }
281 if ( $findPipe ) {
282 $search .= '|';
283 }
284 if ( $findEquals ) {
285 // First equals will be for the template
286 $search .= '=';
287 }
288 $rule = null;
289 # Output literal section, advance input counter
290 $literalLength = strcspn( $text, $search, $i );
291 if ( $literalLength > 0 ) {
292 $accum .= htmlspecialchars( substr( $text, $i, $literalLength ) );
293 $i += $literalLength;
294 }
295 if ( $i >= $lengthText ) {
296 if ( $currentClosing == "\n" ) {
297 // Do a past-the-end run to finish off the heading
298 $curChar = '';
299 $found = 'line-end';
300 } else {
301 # All done
302 break;
303 }
304 } else {
305 $curChar = $curTwoChar = $text[$i];
306 if ( ( $i + 1 ) < $lengthText ) {
307 $curTwoChar .= $text[$i + 1];
308 }
309 if ( $curChar == '|' ) {
310 $found = 'pipe';
311 } elseif ( $curChar == '=' ) {
312 $found = 'equals';
313 } elseif ( $curChar == '<' ) {
314 $found = 'angle';
315 } elseif ( $curChar == "\n" ) {
316 if ( $inHeading ) {
317 $found = 'line-end';
318 } else {
319 $found = 'line-start';
320 }
321 } elseif ( $curTwoChar == $currentClosing ) {
322 $found = 'close';
323 $curChar = $curTwoChar;
324 } elseif ( $curChar == $currentClosing ) {
325 $found = 'close';
326 } elseif ( isset( $this->rules[$curTwoChar] ) ) {
327 $curChar = $curTwoChar;
328 $found = 'open';
329 $rule = $this->rules[$curChar];
330 } elseif ( isset( $this->rules[$curChar] ) ) {
331 $found = 'open';
332 $rule = $this->rules[$curChar];
333 } else {
334 # Some versions of PHP have a strcspn which stops on
335 # null characters; ignore these and continue.
336 # We also may get '-' and '}' characters here which
337 # don't match -{ or $currentClosing. Add these to
338 # output and continue.
339 if ( $curChar == '-' || $curChar == '}' ) {
340 $accum .= $curChar;
341 }
342 ++$i;
343 continue;
344 }
345 }
346 }
347
348 if ( $found == 'angle' ) {
349 $matches = false;
350 // Handle </onlyinclude>
351 if ( $enableOnlyinclude
352 && substr( $text, $i, strlen( '</onlyinclude>' ) ) == '</onlyinclude>'
353 ) {
354 $findOnlyinclude = true;
355 continue;
356 }
357
358 // Determine element name
359 if ( !preg_match( $elementsRegex, $text, $matches, 0, $i + 1 ) ) {
360 // Element name missing or not listed
361 $accum .= '&lt;';
362 ++$i;
363 continue;
364 }
365 // Handle comments
366 if ( isset( $matches[2] ) && $matches[2] == '!--' ) {
367 // To avoid leaving blank lines, when a sequence of
368 // space-separated comments is both preceded and followed by
369 // a newline (ignoring spaces), then
370 // trim leading and trailing spaces and the trailing newline.
371
372 // Find the end
373 $endPos = strpos( $text, '-->', $i + 4 );
374 if ( $endPos === false ) {
375 // Unclosed comment in input, runs to end
376 $inner = substr( $text, $i );
377 $accum .= '<comment>' . htmlspecialchars( $inner ) . '</comment>';
378 $i = $lengthText;
379 } else {
380 // Search backwards for leading whitespace
381 $wsStart = $i ? ( $i - strspn( $revText, " \t", $lengthText - $i ) ) : 0;
382
383 // Search forwards for trailing whitespace
384 // $wsEnd will be the position of the last space (or the '>' if there's none)
385 $wsEnd = $endPos + 2 + strspn( $text, " \t", $endPos + 3 );
386
387 // Keep looking forward as long as we're finding more
388 // comments.
389 $comments = [ [ $wsStart, $wsEnd ] ];
390 while ( substr( $text, $wsEnd + 1, 4 ) == '<!--' ) {
391 $c = strpos( $text, '-->', $wsEnd + 4 );
392 if ( $c === false ) {
393 break;
394 }
395 $c = $c + 2 + strspn( $text, " \t", $c + 3 );
396 $comments[] = [ $wsEnd + 1, $c ];
397 $wsEnd = $c;
398 }
399
400 // Eat the line if possible
401 // TODO: This could theoretically be done if $wsStart == 0, i.e. for comments at
402 // the overall start. That's not how Sanitizer::removeHTMLcomments() did it, but
403 // it's a possible beneficial b/c break.
404 if ( $wsStart > 0 && substr( $text, $wsStart - 1, 1 ) == "\n"
405 && substr( $text, $wsEnd + 1, 1 ) == "\n"
406 ) {
407 // Remove leading whitespace from the end of the accumulator
408 // Sanity check first though
409 $wsLength = $i - $wsStart;
410 if ( $wsLength > 0
411 && strspn( $accum, " \t", -$wsLength ) === $wsLength
412 ) {
413 $accum = substr( $accum, 0, -$wsLength );
414 }
415
416 // Dump all but the last comment to the accumulator
417 foreach ( $comments as $j => $com ) {
418 $startPos = $com[0];
419 $endPos = $com[1] + 1;
420 if ( $j == ( count( $comments ) - 1 ) ) {
421 break;
422 }
423 $inner = substr( $text, $startPos, $endPos - $startPos );
424 $accum .= '<comment>' . htmlspecialchars( $inner ) . '</comment>';
425 }
426
427 // Do a line-start run next time to look for headings after the comment
428 $fakeLineStart = true;
429 } else {
430 // No line to eat, just take the comment itself
431 $startPos = $i;
432 $endPos += 2;
433 }
434
435 if ( $stack->top ) {
436 $part = $stack->top->getCurrentPart();
437 if ( !( isset( $part->commentEnd ) && $part->commentEnd == $wsStart - 1 ) ) {
438 $part->visualEnd = $wsStart;
439 }
440 // Else comments abutting, no change in visual end
441 $part->commentEnd = $endPos;
442 }
443 $i = $endPos + 1;
444 $inner = substr( $text, $startPos, $endPos - $startPos + 1 );
445 $accum .= '<comment>' . htmlspecialchars( $inner ) . '</comment>';
446 }
447 continue;
448 }
449 $name = $matches[1];
450 $lowerName = strtolower( $name );
451 $attrStart = $i + strlen( $name ) + 1;
452
453 // Find end of tag
454 $tagEndPos = $noMoreGT ? false : strpos( $text, '>', $attrStart );
455 if ( $tagEndPos === false ) {
456 // Infinite backtrack
457 // Disable tag search to prevent worst-case O(N^2) performance
458 $noMoreGT = true;
459 $accum .= '&lt;';
460 ++$i;
461 continue;
462 }
463
464 // Handle ignored tags
465 if ( in_array( $lowerName, $ignoredTags ) ) {
466 $accum .= '<ignore>'
467 . htmlspecialchars( substr( $text, $i, $tagEndPos - $i + 1 ) )
468 . '</ignore>';
469 $i = $tagEndPos + 1;
470 continue;
471 }
472
473 $tagStartPos = $i;
474 if ( $text[$tagEndPos - 1] == '/' ) {
475 $attrEnd = $tagEndPos - 1;
476 $inner = null;
477 $i = $tagEndPos + 1;
478 $close = '';
479 } else {
480 $attrEnd = $tagEndPos;
481 // Find closing tag
482 if (
483 !isset( $noMoreClosingTag[$name] ) &&
484 preg_match( "/<\/" . preg_quote( $name, '/' ) . "\s*>/i",
485 $text, $matches, PREG_OFFSET_CAPTURE, $tagEndPos + 1 )
486 ) {
487 $inner = substr( $text, $tagEndPos + 1, $matches[0][1] - $tagEndPos - 1 );
488 $i = $matches[0][1] + strlen( $matches[0][0] );
489 $close = '<close>' . htmlspecialchars( $matches[0][0] ) . '</close>';
490 } else {
491 // No end tag
492 if ( in_array( $name, $xmlishAllowMissingEndTag ) ) {
493 // Let it run out to the end of the text.
494 $inner = substr( $text, $tagEndPos + 1 );
495 $i = $lengthText;
496 $close = '';
497 } else {
498 // Don't match the tag, treat opening tag as literal and resume parsing.
499 $i = $tagEndPos + 1;
500 $accum .= htmlspecialchars( substr( $text, $tagStartPos, $tagEndPos + 1 - $tagStartPos ) );
501 // Cache results, otherwise we have O(N^2) performance for input like <foo><foo><foo>...
502 $noMoreClosingTag[$name] = true;
503 continue;
504 }
505 }
506 }
507 // <includeonly> and <noinclude> just become <ignore> tags
508 if ( in_array( $lowerName, $ignoredElements ) ) {
509 $accum .= '<ignore>' . htmlspecialchars( substr( $text, $tagStartPos, $i - $tagStartPos ) )
510 . '</ignore>';
511 continue;
512 }
513
514 $accum .= '<ext>';
515 if ( $attrEnd <= $attrStart ) {
516 $attr = '';
517 } else {
518 $attr = substr( $text, $attrStart, $attrEnd - $attrStart );
519 }
520 $accum .= '<name>' . htmlspecialchars( $name ) . '</name>' .
521 // Note that the attr element contains the whitespace between name and attribute,
522 // this is necessary for precise reconstruction during pre-save transform.
523 '<attr>' . htmlspecialchars( $attr ) . '</attr>';
524 if ( $inner !== null ) {
525 $accum .= '<inner>' . htmlspecialchars( $inner ) . '</inner>';
526 }
527 $accum .= $close . '</ext>';
528 } elseif ( $found == 'line-start' ) {
529 // Is this the start of a heading?
530 // Line break belongs before the heading element in any case
531 if ( $fakeLineStart ) {
532 $fakeLineStart = false;
533 } else {
534 $accum .= $curChar;
535 $i++;
536 }
537
538 $count = strspn( $text, '=', $i, 6 );
539 if ( $count == 1 && $findEquals ) {
540 // DWIM: This looks kind of like a name/value separator.
541 // Let's let the equals handler have it and break the
542 // potential heading. This is heuristic, but AFAICT the
543 // methods for completely correct disambiguation are very
544 // complex.
545 } elseif ( $count > 0 ) {
546 $piece = [
547 'open' => "\n",
548 'close' => "\n",
549 'parts' => [ new PPDPart( str_repeat( '=', $count ) ) ],
550 'startPos' => $i,
551 'count' => $count ];
552 $stack->push( $piece );
553 $accum =& $stack->getAccum();
554 $stackFlags = $stack->getFlags();
555 if ( isset( $stackFlags['findEquals'] ) ) {
556 $findEquals = $stackFlags['findEquals'];
557 }
558 if ( isset( $stackFlags['findPipe'] ) ) {
559 $findPipe = $stackFlags['findPipe'];
560 }
561 if ( isset( $stackFlags['inHeading'] ) ) {
562 $inHeading = $stackFlags['inHeading'];
563 }
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 $stackFlags = $stack->getFlags();
614 if ( isset( $stackFlags['findEquals'] ) ) {
615 $findEquals = $stackFlags['findEquals'];
616 }
617 if ( isset( $stackFlags['findPipe'] ) ) {
618 $findPipe = $stackFlags['findPipe'];
619 }
620 if ( isset( $stackFlags['inHeading'] ) ) {
621 $inHeading = $stackFlags['inHeading'];
622 }
623
624 // Append the result to the enclosing accumulator
625 $accum .= $element;
626 // Note that we do NOT increment the input pointer.
627 // This is because the closing linebreak could be the opening linebreak of
628 // another heading. Infinite loops are avoided because the next iteration MUST
629 // hit the heading open case above, which unconditionally increments the
630 // input pointer.
631 } elseif ( $found == 'open' ) {
632 # count opening brace characters
633 $curLen = strlen( $curChar );
634 $count = ( $curLen > 1 ) ?
635 # allow the final character to repeat
636 strspn( $text, $curChar[$curLen - 1], $i + 1 ) + 1 :
637 strspn( $text, $curChar, $i );
638
639 $savedPrefix = '';
640 $lineStart = ( $i > 0 && $text[$i - 1] == "\n" );
641
642 if ( $curChar === "-{" && $count > $curLen ) {
643 // -{ => {{ transition because rightmost wins
644 $savedPrefix = '-';
645 $i++;
646 $curChar = '{';
647 $count--;
648 $rule = $this->rules[$curChar];
649 }
650
651 # we need to add to stack only if opening brace count is enough for one of the rules
652 if ( $count >= $rule['min'] ) {
653 # Add it to the stack
654 $piece = [
655 'open' => $curChar,
656 'close' => $rule['end'],
657 'savedPrefix' => $savedPrefix,
658 'count' => $count,
659 'lineStart' => $lineStart,
660 ];
661
662 $stack->push( $piece );
663 $accum =& $stack->getAccum();
664 $stackFlags = $stack->getFlags();
665 if ( isset( $stackFlags['findEquals'] ) ) {
666 $findEquals = $stackFlags['findEquals'];
667 }
668 if ( isset( $stackFlags['findPipe'] ) ) {
669 $findPipe = $stackFlags['findPipe'];
670 }
671 if ( isset( $stackFlags['inHeading'] ) ) {
672 $inHeading = $stackFlags['inHeading'];
673 }
674 } else {
675 # Add literal brace(s)
676 $accum .= htmlspecialchars( $savedPrefix . str_repeat( $curChar, $count ) );
677 }
678 $i += $count;
679 } elseif ( $found == 'close' ) {
680 $piece = $stack->top;
681 # lets check if there are enough characters for closing brace
682 $maxCount = $piece->count;
683 if ( $piece->close === '}-' && $curChar === '}' ) {
684 $maxCount--; # don't try to match closing '-' as a '}'
685 }
686 $curLen = strlen( $curChar );
687 $count = ( $curLen > 1 ) ? $curLen :
688 strspn( $text, $curChar, $i, $maxCount );
689
690 # check for maximum matching characters (if there are 5 closing
691 # characters, we will probably need only 3 - depending on the rules)
692 $rule = $this->rules[$piece->open];
693 if ( $count > $rule['max'] ) {
694 # The specified maximum exists in the callback array, unless the caller
695 # has made an error
696 $matchingCount = $rule['max'];
697 } else {
698 # Count is less than the maximum
699 # Skip any gaps in the callback array to find the true largest match
700 # Need to use array_key_exists not isset because the callback can be null
701 $matchingCount = $count;
702 while ( $matchingCount > 0 && !array_key_exists( $matchingCount, $rule['names'] ) ) {
703 --$matchingCount;
704 }
705 }
706
707 if ( $matchingCount <= 0 ) {
708 # No matching element found in callback array
709 # Output a literal closing brace and continue
710 $endText = substr( $text, $i, $count );
711 $accum .= htmlspecialchars( $endText );
712 $i += $count;
713 continue;
714 }
715 $name = $rule['names'][$matchingCount];
716 if ( $name === null ) {
717 // No element, just literal text
718 $endText = substr( $text, $i, $matchingCount );
719 $element = $piece->breakSyntax( $matchingCount ) . $endText;
720 } else {
721 # Create XML element
722 # Note: $parts is already XML, does not need to be encoded further
723 $parts = $piece->parts;
724 $title = $parts[0]->out;
725 unset( $parts[0] );
726
727 # The invocation is at the start of the line if lineStart is set in
728 # the stack, and all opening brackets are used up.
729 if ( $maxCount == $matchingCount &&
730 !empty( $piece->lineStart ) &&
731 strlen( $piece->savedPrefix ) == 0 ) {
732 $attr = ' lineStart="1"';
733 } else {
734 $attr = '';
735 }
736
737 $element = "<$name$attr>";
738 $element .= "<title>$title</title>";
739 $argIndex = 1;
740 foreach ( $parts as $part ) {
741 if ( isset( $part->eqpos ) ) {
742 $argName = substr( $part->out, 0, $part->eqpos );
743 $argValue = substr( $part->out, $part->eqpos + 1 );
744 $element .= "<part><name>$argName</name>=<value>$argValue</value></part>";
745 } else {
746 $element .= "<part><name index=\"$argIndex\" /><value>{$part->out}</value></part>";
747 $argIndex++;
748 }
749 }
750 $element .= "</$name>";
751 }
752
753 # Advance input pointer
754 $i += $matchingCount;
755
756 # Unwind the stack
757 $stack->pop();
758 $accum =& $stack->getAccum();
759
760 # Re-add the old stack element if it still has unmatched opening characters remaining
761 if ( $matchingCount < $piece->count ) {
762 $piece->parts = [ new PPDPart ];
763 $piece->count -= $matchingCount;
764 # do we still qualify for any callback with remaining count?
765 $min = $this->rules[$piece->open]['min'];
766 if ( $piece->count >= $min ) {
767 $stack->push( $piece );
768 $accum =& $stack->getAccum();
769 } elseif ( $piece->count == 1 && $piece->open === '{' && $piece->savedPrefix === '-' ) {
770 $piece->savedPrefix = '';
771 $piece->open = '-{';
772 $piece->count = 2;
773 $piece->close = $this->rules[$piece->open]['end'];
774 $stack->push( $piece );
775 $accum =& $stack->getAccum();
776 } else {
777 $s = substr( $piece->open, 0, -1 );
778 $s .= str_repeat(
779 substr( $piece->open, -1 ),
780 $piece->count - strlen( $s )
781 );
782 $accum .= $piece->savedPrefix . $s;
783 }
784 } elseif ( $piece->savedPrefix !== '' ) {
785 $accum .= $piece->savedPrefix;
786 }
787
788 $stackFlags = $stack->getFlags();
789 if ( isset( $stackFlags['findEquals'] ) ) {
790 $findEquals = $stackFlags['findEquals'];
791 }
792 if ( isset( $stackFlags['findPipe'] ) ) {
793 $findPipe = $stackFlags['findPipe'];
794 }
795 if ( isset( $stackFlags['inHeading'] ) ) {
796 $inHeading = $stackFlags['inHeading'];
797 }
798
799 # Add XML element to the enclosing accumulator
800 $accum .= $element;
801 } elseif ( $found == 'pipe' ) {
802 $findEquals = true; // shortcut for getFlags()
803 $stack->addPart();
804 $accum =& $stack->getAccum();
805 ++$i;
806 } elseif ( $found == 'equals' ) {
807 $findEquals = false; // shortcut for getFlags()
808 $stack->getCurrentPart()->eqpos = strlen( $accum );
809 $accum .= '=';
810 ++$i;
811 }
812 }
813
814 # Output any remaining unclosed brackets
815 foreach ( $stack->stack as $piece ) {
816 $stack->rootAccum .= $piece->breakSyntax();
817 }
818 $stack->rootAccum .= '</root>';
819 $xml = $stack->rootAccum;
820
821 return $xml;
822 }
823 }
824
825 /**
826 * Stack class to help Preprocessor::preprocessToObj()
827 * @ingroup Parser
828 */
829 class PPDStack {
830 public $stack, $rootAccum;
831
832 /**
833 * @var PPDStack
834 */
835 public $top;
836 public $out;
837 public $elementClass = PPDStackElement::class;
838
839 public static $false = false;
840
841 public function __construct() {
842 $this->stack = [];
843 $this->top = false;
844 $this->rootAccum = '';
845 $this->accum =& $this->rootAccum;
846 }
847
848 /**
849 * @return int
850 */
851 public function count() {
852 return count( $this->stack );
853 }
854
855 public function &getAccum() {
856 return $this->accum;
857 }
858
859 public function getCurrentPart() {
860 if ( $this->top === false ) {
861 return false;
862 } else {
863 return $this->top->getCurrentPart();
864 }
865 }
866
867 public function push( $data ) {
868 if ( $data instanceof $this->elementClass ) {
869 $this->stack[] = $data;
870 } else {
871 $class = $this->elementClass;
872 $this->stack[] = new $class( $data );
873 }
874 $this->top = $this->stack[count( $this->stack ) - 1];
875 $this->accum =& $this->top->getAccum();
876 }
877
878 public function pop() {
879 if ( !count( $this->stack ) ) {
880 throw new MWException( __METHOD__ . ': no elements remaining' );
881 }
882 $temp = array_pop( $this->stack );
883
884 if ( count( $this->stack ) ) {
885 $this->top = $this->stack[count( $this->stack ) - 1];
886 $this->accum =& $this->top->getAccum();
887 } else {
888 $this->top = self::$false;
889 $this->accum =& $this->rootAccum;
890 }
891 return $temp;
892 }
893
894 public function addPart( $s = '' ) {
895 $this->top->addPart( $s );
896 $this->accum =& $this->top->getAccum();
897 }
898
899 /**
900 * @return array
901 */
902 public function getFlags() {
903 if ( !count( $this->stack ) ) {
904 return [
905 'findEquals' => false,
906 'findPipe' => false,
907 'inHeading' => false,
908 ];
909 } else {
910 return $this->top->getFlags();
911 }
912 }
913 }
914
915 /**
916 * @ingroup Parser
917 */
918 class PPDStackElement {
919 /**
920 * @var string Opening character (\n for heading)
921 */
922 public $open;
923
924 /**
925 * @var string Matching closing character
926 */
927 public $close;
928
929 /**
930 * @var string Saved prefix that may affect later processing,
931 * e.g. to differentiate `-{{{{` and `{{{{` after later seeing `}}}`.
932 */
933 public $savedPrefix = '';
934
935 /**
936 * @var int Number of opening characters found (number of "=" for heading)
937 */
938 public $count;
939
940 /**
941 * @var PPDPart[] Array of PPDPart objects describing pipe-separated parts.
942 */
943 public $parts;
944
945 /**
946 * @var bool True if the open char appeared at the start of the input line.
947 * Not set for headings.
948 */
949 public $lineStart;
950
951 public $partClass = PPDPart::class;
952
953 public function __construct( $data = [] ) {
954 $class = $this->partClass;
955 $this->parts = [ new $class ];
956
957 foreach ( $data as $name => $value ) {
958 $this->$name = $value;
959 }
960 }
961
962 public function &getAccum() {
963 return $this->parts[count( $this->parts ) - 1]->out;
964 }
965
966 public function addPart( $s = '' ) {
967 $class = $this->partClass;
968 $this->parts[] = new $class( $s );
969 }
970
971 public function getCurrentPart() {
972 return $this->parts[count( $this->parts ) - 1];
973 }
974
975 /**
976 * @return array
977 */
978 public function getFlags() {
979 $partCount = count( $this->parts );
980 $findPipe = $this->open != "\n" && $this->open != '[';
981 return [
982 'findPipe' => $findPipe,
983 'findEquals' => $findPipe && $partCount > 1 && !isset( $this->parts[$partCount - 1]->eqpos ),
984 'inHeading' => $this->open == "\n",
985 ];
986 }
987
988 /**
989 * Get the output string that would result if the close is not found.
990 *
991 * @param bool|int $openingCount
992 * @return string
993 */
994 public function breakSyntax( $openingCount = false ) {
995 if ( $this->open == "\n" ) {
996 $s = $this->savedPrefix . $this->parts[0]->out;
997 } else {
998 if ( $openingCount === false ) {
999 $openingCount = $this->count;
1000 }
1001 $s = substr( $this->open, 0, -1 );
1002 $s .= str_repeat(
1003 substr( $this->open, -1 ),
1004 $openingCount - strlen( $s )
1005 );
1006 $s = $this->savedPrefix . $s;
1007 $first = true;
1008 foreach ( $this->parts as $part ) {
1009 if ( $first ) {
1010 $first = false;
1011 } else {
1012 $s .= '|';
1013 }
1014 $s .= $part->out;
1015 }
1016 }
1017 return $s;
1018 }
1019 }
1020
1021 /**
1022 * @ingroup Parser
1023 */
1024 class PPDPart {
1025 /**
1026 * @var string Output accumulator string
1027 */
1028 public $out;
1029
1030 // Optional member variables:
1031 // eqpos Position of equals sign in output accumulator
1032 // commentEnd Past-the-end input pointer for the last comment encountered
1033 // visualEnd Past-the-end input pointer for the end of the accumulator minus comments
1034
1035 public function __construct( $out = '' ) {
1036 $this->out = $out;
1037 }
1038 }
1039
1040 /**
1041 * An expansion frame, used as a context to expand the result of preprocessToObj()
1042 * @ingroup Parser
1043 */
1044 // phpcs:ignore Squiz.Classes.ValidClassName.NotCamelCaps
1045 class PPFrame_DOM implements PPFrame {
1046
1047 /**
1048 * @var Preprocessor
1049 */
1050 public $preprocessor;
1051
1052 /**
1053 * @var Parser
1054 */
1055 public $parser;
1056
1057 /**
1058 * @var Title
1059 */
1060 public $title;
1061 public $titleCache;
1062
1063 /**
1064 * Hashtable listing templates which are disallowed for expansion in this frame,
1065 * having been encountered previously in parent frames.
1066 */
1067 public $loopCheckHash;
1068
1069 /**
1070 * Recursion depth of this frame, top = 0
1071 * Note that this is NOT the same as expansion depth in expand()
1072 */
1073 public $depth;
1074
1075 private $volatile = false;
1076 private $ttl = null;
1077
1078 /**
1079 * @var array
1080 */
1081 protected $childExpansionCache;
1082
1083 /**
1084 * Construct a new preprocessor frame.
1085 * @param Preprocessor $preprocessor The parent preprocessor
1086 */
1087 public function __construct( $preprocessor ) {
1088 $this->preprocessor = $preprocessor;
1089 $this->parser = $preprocessor->parser;
1090 $this->title = $this->parser->mTitle;
1091 $this->titleCache = [ $this->title ? $this->title->getPrefixedDBkey() : false ];
1092 $this->loopCheckHash = [];
1093 $this->depth = 0;
1094 $this->childExpansionCache = [];
1095 }
1096
1097 /**
1098 * Create a new child frame
1099 * $args is optionally a multi-root PPNode or array containing the template arguments
1100 *
1101 * @param bool|array $args
1102 * @param Title|bool $title
1103 * @param int $indexOffset
1104 * @return PPTemplateFrame_DOM
1105 */
1106 public function newChild( $args = false, $title = false, $indexOffset = 0 ) {
1107 $namedArgs = [];
1108 $numberedArgs = [];
1109 if ( $title === false ) {
1110 $title = $this->title;
1111 }
1112 if ( $args !== false ) {
1113 $xpath = false;
1114 if ( $args instanceof PPNode ) {
1115 $args = $args->node;
1116 }
1117 foreach ( $args as $arg ) {
1118 if ( $arg instanceof PPNode ) {
1119 $arg = $arg->node;
1120 }
1121 if ( !$xpath || $xpath->document !== $arg->ownerDocument ) {
1122 $xpath = new DOMXPath( $arg->ownerDocument );
1123 }
1124
1125 $nameNodes = $xpath->query( 'name', $arg );
1126 $value = $xpath->query( 'value', $arg );
1127 if ( $nameNodes->item( 0 )->hasAttributes() ) {
1128 // Numbered parameter
1129 $index = $nameNodes->item( 0 )->attributes->getNamedItem( 'index' )->textContent;
1130 $index = $index - $indexOffset;
1131 if ( isset( $namedArgs[$index] ) || isset( $numberedArgs[$index] ) ) {
1132 $this->parser->getOutput()->addWarning( wfMessage( 'duplicate-args-warning',
1133 wfEscapeWikiText( $this->title ),
1134 wfEscapeWikiText( $title ),
1135 wfEscapeWikiText( $index ) )->text() );
1136 $this->parser->addTrackingCategory( 'duplicate-args-category' );
1137 }
1138 $numberedArgs[$index] = $value->item( 0 );
1139 unset( $namedArgs[$index] );
1140 } else {
1141 // Named parameter
1142 $name = trim( $this->expand( $nameNodes->item( 0 ), PPFrame::STRIP_COMMENTS ) );
1143 if ( isset( $namedArgs[$name] ) || isset( $numberedArgs[$name] ) ) {
1144 $this->parser->getOutput()->addWarning( wfMessage( 'duplicate-args-warning',
1145 wfEscapeWikiText( $this->title ),
1146 wfEscapeWikiText( $title ),
1147 wfEscapeWikiText( $name ) )->text() );
1148 $this->parser->addTrackingCategory( 'duplicate-args-category' );
1149 }
1150 $namedArgs[$name] = $value->item( 0 );
1151 unset( $numberedArgs[$name] );
1152 }
1153 }
1154 }
1155 return new PPTemplateFrame_DOM( $this->preprocessor, $this, $numberedArgs, $namedArgs, $title );
1156 }
1157
1158 /**
1159 * @throws MWException
1160 * @param string|int $key
1161 * @param string|PPNode_DOM|DOMDocument $root
1162 * @param int $flags
1163 * @return string
1164 */
1165 public function cachedExpand( $key, $root, $flags = 0 ) {
1166 // we don't have a parent, so we don't have a cache
1167 return $this->expand( $root, $flags );
1168 }
1169
1170 /**
1171 * @throws MWException
1172 * @param string|PPNode_DOM|DOMDocument $root
1173 * @param int $flags
1174 * @return string
1175 */
1176 public function expand( $root, $flags = 0 ) {
1177 static $expansionDepth = 0;
1178 if ( is_string( $root ) ) {
1179 return $root;
1180 }
1181
1182 if ( ++$this->parser->mPPNodeCount > $this->parser->mOptions->getMaxPPNodeCount() ) {
1183 $this->parser->limitationWarn( 'node-count-exceeded',
1184 $this->parser->mPPNodeCount,
1185 $this->parser->mOptions->getMaxPPNodeCount()
1186 );
1187 return '<span class="error">Node-count limit exceeded</span>';
1188 }
1189
1190 if ( $expansionDepth > $this->parser->mOptions->getMaxPPExpandDepth() ) {
1191 $this->parser->limitationWarn( 'expansion-depth-exceeded',
1192 $expansionDepth,
1193 $this->parser->mOptions->getMaxPPExpandDepth()
1194 );
1195 return '<span class="error">Expansion depth limit exceeded</span>';
1196 }
1197 ++$expansionDepth;
1198 if ( $expansionDepth > $this->parser->mHighestExpansionDepth ) {
1199 $this->parser->mHighestExpansionDepth = $expansionDepth;
1200 }
1201
1202 if ( $root instanceof PPNode_DOM ) {
1203 $root = $root->node;
1204 }
1205 if ( $root instanceof DOMDocument ) {
1206 $root = $root->documentElement;
1207 }
1208
1209 $outStack = [ '', '' ];
1210 $iteratorStack = [ false, $root ];
1211 $indexStack = [ 0, 0 ];
1212
1213 while ( count( $iteratorStack ) > 1 ) {
1214 $level = count( $outStack ) - 1;
1215 $iteratorNode =& $iteratorStack[$level];
1216 $out =& $outStack[$level];
1217 $index =& $indexStack[$level];
1218
1219 if ( $iteratorNode instanceof PPNode_DOM ) {
1220 $iteratorNode = $iteratorNode->node;
1221 }
1222
1223 if ( is_array( $iteratorNode ) ) {
1224 if ( $index >= count( $iteratorNode ) ) {
1225 // All done with this iterator
1226 $iteratorStack[$level] = false;
1227 $contextNode = false;
1228 } else {
1229 $contextNode = $iteratorNode[$index];
1230 $index++;
1231 }
1232 } elseif ( $iteratorNode instanceof DOMNodeList ) {
1233 if ( $index >= $iteratorNode->length ) {
1234 // All done with this iterator
1235 $iteratorStack[$level] = false;
1236 $contextNode = false;
1237 } else {
1238 $contextNode = $iteratorNode->item( $index );
1239 $index++;
1240 }
1241 } else {
1242 // Copy to $contextNode and then delete from iterator stack,
1243 // because this is not an iterator but we do have to execute it once
1244 $contextNode = $iteratorStack[$level];
1245 $iteratorStack[$level] = false;
1246 }
1247
1248 if ( $contextNode instanceof PPNode_DOM ) {
1249 $contextNode = $contextNode->node;
1250 }
1251
1252 $newIterator = false;
1253
1254 if ( $contextNode === false ) {
1255 // nothing to do
1256 } elseif ( is_string( $contextNode ) ) {
1257 $out .= $contextNode;
1258 } elseif ( is_array( $contextNode ) || $contextNode instanceof DOMNodeList ) {
1259 $newIterator = $contextNode;
1260 } elseif ( $contextNode instanceof DOMNode ) {
1261 if ( $contextNode->nodeType == XML_TEXT_NODE ) {
1262 $out .= $contextNode->nodeValue;
1263 } elseif ( $contextNode->nodeName == 'template' ) {
1264 # Double-brace expansion
1265 $xpath = new DOMXPath( $contextNode->ownerDocument );
1266 $titles = $xpath->query( 'title', $contextNode );
1267 $title = $titles->item( 0 );
1268 $parts = $xpath->query( 'part', $contextNode );
1269 if ( $flags & PPFrame::NO_TEMPLATES ) {
1270 $newIterator = $this->virtualBracketedImplode( '{{', '|', '}}', $title, $parts );
1271 } else {
1272 $lineStart = $contextNode->getAttribute( 'lineStart' );
1273 $params = [
1274 'title' => new PPNode_DOM( $title ),
1275 'parts' => new PPNode_DOM( $parts ),
1276 'lineStart' => $lineStart ];
1277 $ret = $this->parser->braceSubstitution( $params, $this );
1278 if ( isset( $ret['object'] ) ) {
1279 $newIterator = $ret['object'];
1280 } else {
1281 $out .= $ret['text'];
1282 }
1283 }
1284 } elseif ( $contextNode->nodeName == 'tplarg' ) {
1285 # Triple-brace expansion
1286 $xpath = new DOMXPath( $contextNode->ownerDocument );
1287 $titles = $xpath->query( 'title', $contextNode );
1288 $title = $titles->item( 0 );
1289 $parts = $xpath->query( 'part', $contextNode );
1290 if ( $flags & PPFrame::NO_ARGS ) {
1291 $newIterator = $this->virtualBracketedImplode( '{{{', '|', '}}}', $title, $parts );
1292 } else {
1293 $params = [
1294 'title' => new PPNode_DOM( $title ),
1295 'parts' => new PPNode_DOM( $parts ) ];
1296 $ret = $this->parser->argSubstitution( $params, $this );
1297 if ( isset( $ret['object'] ) ) {
1298 $newIterator = $ret['object'];
1299 } else {
1300 $out .= $ret['text'];
1301 }
1302 }
1303 } elseif ( $contextNode->nodeName == 'comment' ) {
1304 # HTML-style comment
1305 # Remove it in HTML, pre+remove and STRIP_COMMENTS modes
1306 # Not in RECOVER_COMMENTS mode (msgnw) though.
1307 if ( ( $this->parser->ot['html']
1308 || ( $this->parser->ot['pre'] && $this->parser->mOptions->getRemoveComments() )
1309 || ( $flags & PPFrame::STRIP_COMMENTS )
1310 ) && !( $flags & PPFrame::RECOVER_COMMENTS )
1311 ) {
1312 $out .= '';
1313 } elseif ( $this->parser->ot['wiki'] && !( $flags & PPFrame::RECOVER_COMMENTS ) ) {
1314 # Add a strip marker in PST mode so that pstPass2() can
1315 # run some old-fashioned regexes on the result.
1316 # Not in RECOVER_COMMENTS mode (extractSections) though.
1317 $out .= $this->parser->insertStripItem( $contextNode->textContent );
1318 } else {
1319 # Recover the literal comment in RECOVER_COMMENTS and pre+no-remove
1320 $out .= $contextNode->textContent;
1321 }
1322 } elseif ( $contextNode->nodeName == 'ignore' ) {
1323 # Output suppression used by <includeonly> etc.
1324 # OT_WIKI will only respect <ignore> in substed templates.
1325 # The other output types respect it unless NO_IGNORE is set.
1326 # extractSections() sets NO_IGNORE and so never respects it.
1327 if ( ( !isset( $this->parent ) && $this->parser->ot['wiki'] )
1328 || ( $flags & PPFrame::NO_IGNORE )
1329 ) {
1330 $out .= $contextNode->textContent;
1331 } else {
1332 $out .= '';
1333 }
1334 } elseif ( $contextNode->nodeName == 'ext' ) {
1335 # Extension tag
1336 $xpath = new DOMXPath( $contextNode->ownerDocument );
1337 $names = $xpath->query( 'name', $contextNode );
1338 $attrs = $xpath->query( 'attr', $contextNode );
1339 $inners = $xpath->query( 'inner', $contextNode );
1340 $closes = $xpath->query( 'close', $contextNode );
1341 if ( $flags & PPFrame::NO_TAGS ) {
1342 $s = '<' . $this->expand( $names->item( 0 ), $flags );
1343 if ( $attrs->length > 0 ) {
1344 $s .= $this->expand( $attrs->item( 0 ), $flags );
1345 }
1346 if ( $inners->length > 0 ) {
1347 $s .= '>' . $this->expand( $inners->item( 0 ), $flags );
1348 if ( $closes->length > 0 ) {
1349 $s .= $this->expand( $closes->item( 0 ), $flags );
1350 }
1351 } else {
1352 $s .= '/>';
1353 }
1354 $out .= $s;
1355 } else {
1356 $params = [
1357 'name' => new PPNode_DOM( $names->item( 0 ) ),
1358 'attr' => $attrs->length > 0 ? new PPNode_DOM( $attrs->item( 0 ) ) : null,
1359 'inner' => $inners->length > 0 ? new PPNode_DOM( $inners->item( 0 ) ) : null,
1360 'close' => $closes->length > 0 ? new PPNode_DOM( $closes->item( 0 ) ) : null,
1361 ];
1362 $out .= $this->parser->extensionSubstitution( $params, $this );
1363 }
1364 } elseif ( $contextNode->nodeName == 'h' ) {
1365 # Heading
1366 $s = $this->expand( $contextNode->childNodes, $flags );
1367
1368 # Insert a heading marker only for <h> children of <root>
1369 # This is to stop extractSections from going over multiple tree levels
1370 if ( $contextNode->parentNode->nodeName == 'root' && $this->parser->ot['html'] ) {
1371 # Insert heading index marker
1372 $headingIndex = $contextNode->getAttribute( 'i' );
1373 $titleText = $this->title->getPrefixedDBkey();
1374 $this->parser->mHeadings[] = [ $titleText, $headingIndex ];
1375 $serial = count( $this->parser->mHeadings ) - 1;
1376 $marker = Parser::MARKER_PREFIX . "-h-$serial-" . Parser::MARKER_SUFFIX;
1377 $count = $contextNode->getAttribute( 'level' );
1378 $s = substr( $s, 0, $count ) . $marker . substr( $s, $count );
1379 $this->parser->mStripState->addGeneral( $marker, '' );
1380 }
1381 $out .= $s;
1382 } else {
1383 # Generic recursive expansion
1384 $newIterator = $contextNode->childNodes;
1385 }
1386 } else {
1387 throw new MWException( __METHOD__ . ': Invalid parameter type' );
1388 }
1389
1390 if ( $newIterator !== false ) {
1391 if ( $newIterator instanceof PPNode_DOM ) {
1392 $newIterator = $newIterator->node;
1393 }
1394 $outStack[] = '';
1395 $iteratorStack[] = $newIterator;
1396 $indexStack[] = 0;
1397 } elseif ( $iteratorStack[$level] === false ) {
1398 // Return accumulated value to parent
1399 // With tail recursion
1400 while ( $iteratorStack[$level] === false && $level > 0 ) {
1401 $outStack[$level - 1] .= $out;
1402 array_pop( $outStack );
1403 array_pop( $iteratorStack );
1404 array_pop( $indexStack );
1405 $level--;
1406 }
1407 }
1408 }
1409 --$expansionDepth;
1410 return $outStack[0];
1411 }
1412
1413 /**
1414 * @param string $sep
1415 * @param int $flags
1416 * @param string|PPNode_DOM|DOMDocument $args,...
1417 * @return string
1418 */
1419 public function implodeWithFlags( $sep, $flags /*, ... */ ) {
1420 $args = array_slice( func_get_args(), 2 );
1421
1422 $first = true;
1423 $s = '';
1424 foreach ( $args as $root ) {
1425 if ( $root instanceof PPNode_DOM ) {
1426 $root = $root->node;
1427 }
1428 if ( !is_array( $root ) && !( $root instanceof DOMNodeList ) ) {
1429 $root = [ $root ];
1430 }
1431 foreach ( $root as $node ) {
1432 if ( $first ) {
1433 $first = false;
1434 } else {
1435 $s .= $sep;
1436 }
1437 $s .= $this->expand( $node, $flags );
1438 }
1439 }
1440 return $s;
1441 }
1442
1443 /**
1444 * Implode with no flags specified
1445 * This previously called implodeWithFlags but has now been inlined to reduce stack depth
1446 *
1447 * @param string $sep
1448 * @param string|PPNode_DOM|DOMDocument $args,...
1449 * @return string
1450 */
1451 public function implode( $sep /*, ... */ ) {
1452 $args = array_slice( func_get_args(), 1 );
1453
1454 $first = true;
1455 $s = '';
1456 foreach ( $args as $root ) {
1457 if ( $root instanceof PPNode_DOM ) {
1458 $root = $root->node;
1459 }
1460 if ( !is_array( $root ) && !( $root instanceof DOMNodeList ) ) {
1461 $root = [ $root ];
1462 }
1463 foreach ( $root as $node ) {
1464 if ( $first ) {
1465 $first = false;
1466 } else {
1467 $s .= $sep;
1468 }
1469 $s .= $this->expand( $node );
1470 }
1471 }
1472 return $s;
1473 }
1474
1475 /**
1476 * Makes an object that, when expand()ed, will be the same as one obtained
1477 * with implode()
1478 *
1479 * @param string $sep
1480 * @param string|PPNode_DOM|DOMDocument $args,...
1481 * @return array
1482 */
1483 public function virtualImplode( $sep /*, ... */ ) {
1484 $args = array_slice( func_get_args(), 1 );
1485 $out = [];
1486 $first = true;
1487
1488 foreach ( $args as $root ) {
1489 if ( $root instanceof PPNode_DOM ) {
1490 $root = $root->node;
1491 }
1492 if ( !is_array( $root ) && !( $root instanceof DOMNodeList ) ) {
1493 $root = [ $root ];
1494 }
1495 foreach ( $root as $node ) {
1496 if ( $first ) {
1497 $first = false;
1498 } else {
1499 $out[] = $sep;
1500 }
1501 $out[] = $node;
1502 }
1503 }
1504 return $out;
1505 }
1506
1507 /**
1508 * Virtual implode with brackets
1509 * @param string $start
1510 * @param string $sep
1511 * @param string $end
1512 * @param string|PPNode_DOM|DOMDocument $args,...
1513 * @return array
1514 */
1515 public function virtualBracketedImplode( $start, $sep, $end /*, ... */ ) {
1516 $args = array_slice( func_get_args(), 3 );
1517 $out = [ $start ];
1518 $first = true;
1519
1520 foreach ( $args as $root ) {
1521 if ( $root instanceof PPNode_DOM ) {
1522 $root = $root->node;
1523 }
1524 if ( !is_array( $root ) && !( $root instanceof DOMNodeList ) ) {
1525 $root = [ $root ];
1526 }
1527 foreach ( $root as $node ) {
1528 if ( $first ) {
1529 $first = false;
1530 } else {
1531 $out[] = $sep;
1532 }
1533 $out[] = $node;
1534 }
1535 }
1536 $out[] = $end;
1537 return $out;
1538 }
1539
1540 public function __toString() {
1541 return 'frame{}';
1542 }
1543
1544 public function getPDBK( $level = false ) {
1545 if ( $level === false ) {
1546 return $this->title->getPrefixedDBkey();
1547 } else {
1548 return isset( $this->titleCache[$level] ) ? $this->titleCache[$level] : false;
1549 }
1550 }
1551
1552 /**
1553 * @return array
1554 */
1555 public function getArguments() {
1556 return [];
1557 }
1558
1559 /**
1560 * @return array
1561 */
1562 public function getNumberedArguments() {
1563 return [];
1564 }
1565
1566 /**
1567 * @return array
1568 */
1569 public function getNamedArguments() {
1570 return [];
1571 }
1572
1573 /**
1574 * Returns true if there are no arguments in this frame
1575 *
1576 * @return bool
1577 */
1578 public function isEmpty() {
1579 return true;
1580 }
1581
1582 /**
1583 * @param int|string $name
1584 * @return bool Always false in this implementation.
1585 */
1586 public function getArgument( $name ) {
1587 return false;
1588 }
1589
1590 /**
1591 * Returns true if the infinite loop check is OK, false if a loop is detected
1592 *
1593 * @param Title $title
1594 * @return bool
1595 */
1596 public function loopCheck( $title ) {
1597 return !isset( $this->loopCheckHash[$title->getPrefixedDBkey()] );
1598 }
1599
1600 /**
1601 * Return true if the frame is a template frame
1602 *
1603 * @return bool
1604 */
1605 public function isTemplate() {
1606 return false;
1607 }
1608
1609 /**
1610 * Get a title of frame
1611 *
1612 * @return Title
1613 */
1614 public function getTitle() {
1615 return $this->title;
1616 }
1617
1618 /**
1619 * Set the volatile flag
1620 *
1621 * @param bool $flag
1622 */
1623 public function setVolatile( $flag = true ) {
1624 $this->volatile = $flag;
1625 }
1626
1627 /**
1628 * Get the volatile flag
1629 *
1630 * @return bool
1631 */
1632 public function isVolatile() {
1633 return $this->volatile;
1634 }
1635
1636 /**
1637 * Set the TTL
1638 *
1639 * @param int $ttl
1640 */
1641 public function setTTL( $ttl ) {
1642 if ( $ttl !== null && ( $this->ttl === null || $ttl < $this->ttl ) ) {
1643 $this->ttl = $ttl;
1644 }
1645 }
1646
1647 /**
1648 * Get the TTL
1649 *
1650 * @return int|null
1651 */
1652 public function getTTL() {
1653 return $this->ttl;
1654 }
1655 }
1656
1657 /**
1658 * Expansion frame with template arguments
1659 * @ingroup Parser
1660 */
1661 // phpcs:ignore Squiz.Classes.ValidClassName.NotCamelCaps
1662 class PPTemplateFrame_DOM extends PPFrame_DOM {
1663
1664 public $numberedArgs, $namedArgs;
1665
1666 /**
1667 * @var PPFrame_DOM
1668 */
1669 public $parent;
1670 public $numberedExpansionCache, $namedExpansionCache;
1671
1672 /**
1673 * @param Preprocessor $preprocessor
1674 * @param bool|PPFrame_DOM $parent
1675 * @param array $numberedArgs
1676 * @param array $namedArgs
1677 * @param bool|Title $title
1678 */
1679 public function __construct( $preprocessor, $parent = false, $numberedArgs = [],
1680 $namedArgs = [], $title = false
1681 ) {
1682 parent::__construct( $preprocessor );
1683
1684 $this->parent = $parent;
1685 $this->numberedArgs = $numberedArgs;
1686 $this->namedArgs = $namedArgs;
1687 $this->title = $title;
1688 $pdbk = $title ? $title->getPrefixedDBkey() : false;
1689 $this->titleCache = $parent->titleCache;
1690 $this->titleCache[] = $pdbk;
1691 $this->loopCheckHash = /*clone*/ $parent->loopCheckHash;
1692 if ( $pdbk !== false ) {
1693 $this->loopCheckHash[$pdbk] = true;
1694 }
1695 $this->depth = $parent->depth + 1;
1696 $this->numberedExpansionCache = $this->namedExpansionCache = [];
1697 }
1698
1699 public function __toString() {
1700 $s = 'tplframe{';
1701 $first = true;
1702 $args = $this->numberedArgs + $this->namedArgs;
1703 foreach ( $args as $name => $value ) {
1704 if ( $first ) {
1705 $first = false;
1706 } else {
1707 $s .= ', ';
1708 }
1709 $s .= "\"$name\":\"" .
1710 str_replace( '"', '\\"', $value->ownerDocument->saveXML( $value ) ) . '"';
1711 }
1712 $s .= '}';
1713 return $s;
1714 }
1715
1716 /**
1717 * @throws MWException
1718 * @param string|int $key
1719 * @param string|PPNode_DOM|DOMDocument $root
1720 * @param int $flags
1721 * @return string
1722 */
1723 public function cachedExpand( $key, $root, $flags = 0 ) {
1724 if ( isset( $this->parent->childExpansionCache[$key] ) ) {
1725 return $this->parent->childExpansionCache[$key];
1726 }
1727 $retval = $this->expand( $root, $flags );
1728 if ( !$this->isVolatile() ) {
1729 $this->parent->childExpansionCache[$key] = $retval;
1730 }
1731 return $retval;
1732 }
1733
1734 /**
1735 * Returns true if there are no arguments in this frame
1736 *
1737 * @return bool
1738 */
1739 public function isEmpty() {
1740 return !count( $this->numberedArgs ) && !count( $this->namedArgs );
1741 }
1742
1743 public function getArguments() {
1744 $arguments = [];
1745 foreach ( array_merge(
1746 array_keys( $this->numberedArgs ),
1747 array_keys( $this->namedArgs ) ) as $key ) {
1748 $arguments[$key] = $this->getArgument( $key );
1749 }
1750 return $arguments;
1751 }
1752
1753 public function getNumberedArguments() {
1754 $arguments = [];
1755 foreach ( array_keys( $this->numberedArgs ) as $key ) {
1756 $arguments[$key] = $this->getArgument( $key );
1757 }
1758 return $arguments;
1759 }
1760
1761 public function getNamedArguments() {
1762 $arguments = [];
1763 foreach ( array_keys( $this->namedArgs ) as $key ) {
1764 $arguments[$key] = $this->getArgument( $key );
1765 }
1766 return $arguments;
1767 }
1768
1769 /**
1770 * @param int $index
1771 * @return string|bool
1772 */
1773 public function getNumberedArgument( $index ) {
1774 if ( !isset( $this->numberedArgs[$index] ) ) {
1775 return false;
1776 }
1777 if ( !isset( $this->numberedExpansionCache[$index] ) ) {
1778 # No trimming for unnamed arguments
1779 $this->numberedExpansionCache[$index] = $this->parent->expand(
1780 $this->numberedArgs[$index],
1781 PPFrame::STRIP_COMMENTS
1782 );
1783 }
1784 return $this->numberedExpansionCache[$index];
1785 }
1786
1787 /**
1788 * @param string $name
1789 * @return string|bool
1790 */
1791 public function getNamedArgument( $name ) {
1792 if ( !isset( $this->namedArgs[$name] ) ) {
1793 return false;
1794 }
1795 if ( !isset( $this->namedExpansionCache[$name] ) ) {
1796 # Trim named arguments post-expand, for backwards compatibility
1797 $this->namedExpansionCache[$name] = trim(
1798 $this->parent->expand( $this->namedArgs[$name], PPFrame::STRIP_COMMENTS ) );
1799 }
1800 return $this->namedExpansionCache[$name];
1801 }
1802
1803 /**
1804 * @param int|string $name
1805 * @return string|bool
1806 */
1807 public function getArgument( $name ) {
1808 $text = $this->getNumberedArgument( $name );
1809 if ( $text === false ) {
1810 $text = $this->getNamedArgument( $name );
1811 }
1812 return $text;
1813 }
1814
1815 /**
1816 * Return true if the frame is a template frame
1817 *
1818 * @return bool
1819 */
1820 public function isTemplate() {
1821 return true;
1822 }
1823
1824 public function setVolatile( $flag = true ) {
1825 parent::setVolatile( $flag );
1826 $this->parent->setVolatile( $flag );
1827 }
1828
1829 public function setTTL( $ttl ) {
1830 parent::setTTL( $ttl );
1831 $this->parent->setTTL( $ttl );
1832 }
1833 }
1834
1835 /**
1836 * Expansion frame with custom arguments
1837 * @ingroup Parser
1838 */
1839 // phpcs:ignore Squiz.Classes.ValidClassName.NotCamelCaps
1840 class PPCustomFrame_DOM extends PPFrame_DOM {
1841
1842 public $args;
1843
1844 public function __construct( $preprocessor, $args ) {
1845 parent::__construct( $preprocessor );
1846 $this->args = $args;
1847 }
1848
1849 public function __toString() {
1850 $s = 'cstmframe{';
1851 $first = true;
1852 foreach ( $this->args as $name => $value ) {
1853 if ( $first ) {
1854 $first = false;
1855 } else {
1856 $s .= ', ';
1857 }
1858 $s .= "\"$name\":\"" .
1859 str_replace( '"', '\\"', $value->__toString() ) . '"';
1860 }
1861 $s .= '}';
1862 return $s;
1863 }
1864
1865 /**
1866 * @return bool
1867 */
1868 public function isEmpty() {
1869 return !count( $this->args );
1870 }
1871
1872 /**
1873 * @param int|string $index
1874 * @return string|bool
1875 */
1876 public function getArgument( $index ) {
1877 if ( !isset( $this->args[$index] ) ) {
1878 return false;
1879 }
1880 return $this->args[$index];
1881 }
1882
1883 public function getArguments() {
1884 return $this->args;
1885 }
1886 }
1887
1888 /**
1889 * @ingroup Parser
1890 */
1891 // phpcs:ignore Squiz.Classes.ValidClassName.NotCamelCaps
1892 class PPNode_DOM implements PPNode {
1893
1894 /**
1895 * @var DOMElement
1896 */
1897 public $node;
1898 public $xpath;
1899
1900 public function __construct( $node, $xpath = false ) {
1901 $this->node = $node;
1902 }
1903
1904 /**
1905 * @return DOMXPath
1906 */
1907 public function getXPath() {
1908 if ( $this->xpath === null ) {
1909 $this->xpath = new DOMXPath( $this->node->ownerDocument );
1910 }
1911 return $this->xpath;
1912 }
1913
1914 public function __toString() {
1915 if ( $this->node instanceof DOMNodeList ) {
1916 $s = '';
1917 foreach ( $this->node as $node ) {
1918 $s .= $node->ownerDocument->saveXML( $node );
1919 }
1920 } else {
1921 $s = $this->node->ownerDocument->saveXML( $this->node );
1922 }
1923 return $s;
1924 }
1925
1926 /**
1927 * @return bool|PPNode_DOM
1928 */
1929 public function getChildren() {
1930 return $this->node->childNodes ? new self( $this->node->childNodes ) : false;
1931 }
1932
1933 /**
1934 * @return bool|PPNode_DOM
1935 */
1936 public function getFirstChild() {
1937 return $this->node->firstChild ? new self( $this->node->firstChild ) : false;
1938 }
1939
1940 /**
1941 * @return bool|PPNode_DOM
1942 */
1943 public function getNextSibling() {
1944 return $this->node->nextSibling ? new self( $this->node->nextSibling ) : false;
1945 }
1946
1947 /**
1948 * @param string $type
1949 *
1950 * @return bool|PPNode_DOM
1951 */
1952 public function getChildrenOfType( $type ) {
1953 return new self( $this->getXPath()->query( $type, $this->node ) );
1954 }
1955
1956 /**
1957 * @return int
1958 */
1959 public function getLength() {
1960 if ( $this->node instanceof DOMNodeList ) {
1961 return $this->node->length;
1962 } else {
1963 return false;
1964 }
1965 }
1966
1967 /**
1968 * @param int $i
1969 * @return bool|PPNode_DOM
1970 */
1971 public function item( $i ) {
1972 $item = $this->node->item( $i );
1973 return $item ? new self( $item ) : false;
1974 }
1975
1976 /**
1977 * @return string
1978 */
1979 public function getName() {
1980 if ( $this->node instanceof DOMNodeList ) {
1981 return '#nodelist';
1982 } else {
1983 return $this->node->nodeName;
1984 }
1985 }
1986
1987 /**
1988 * Split a "<part>" node into an associative array containing:
1989 * - name PPNode name
1990 * - index String index
1991 * - value PPNode value
1992 *
1993 * @throws MWException
1994 * @return array
1995 */
1996 public function splitArg() {
1997 $xpath = $this->getXPath();
1998 $names = $xpath->query( 'name', $this->node );
1999 $values = $xpath->query( 'value', $this->node );
2000 if ( !$names->length || !$values->length ) {
2001 throw new MWException( 'Invalid brace node passed to ' . __METHOD__ );
2002 }
2003 $name = $names->item( 0 );
2004 $index = $name->getAttribute( 'index' );
2005 return [
2006 'name' => new self( $name ),
2007 'index' => $index,
2008 'value' => new self( $values->item( 0 ) ) ];
2009 }
2010
2011 /**
2012 * Split an "<ext>" node into an associative array containing name, attr, inner and close
2013 * All values in the resulting array are PPNodes. Inner and close are optional.
2014 *
2015 * @throws MWException
2016 * @return array
2017 */
2018 public function splitExt() {
2019 $xpath = $this->getXPath();
2020 $names = $xpath->query( 'name', $this->node );
2021 $attrs = $xpath->query( 'attr', $this->node );
2022 $inners = $xpath->query( 'inner', $this->node );
2023 $closes = $xpath->query( 'close', $this->node );
2024 if ( !$names->length || !$attrs->length ) {
2025 throw new MWException( 'Invalid ext node passed to ' . __METHOD__ );
2026 }
2027 $parts = [
2028 'name' => new self( $names->item( 0 ) ),
2029 'attr' => new self( $attrs->item( 0 ) ) ];
2030 if ( $inners->length ) {
2031 $parts['inner'] = new self( $inners->item( 0 ) );
2032 }
2033 if ( $closes->length ) {
2034 $parts['close'] = new self( $closes->item( 0 ) );
2035 }
2036 return $parts;
2037 }
2038
2039 /**
2040 * Split a "<h>" node
2041 * @throws MWException
2042 * @return array
2043 */
2044 public function splitHeading() {
2045 if ( $this->getName() !== 'h' ) {
2046 throw new MWException( 'Invalid h node passed to ' . __METHOD__ );
2047 }
2048 return [
2049 'i' => $this->node->getAttribute( 'i' ),
2050 'level' => $this->node->getAttribute( 'level' ),
2051 'contents' => $this->getChildren()
2052 ];
2053 }
2054 }
Wiklou, le Wiki du Wiklou