dépôts / lhc / web / wiklou.git / blob
? search:
summary | shortlog | log | commit | commitdiff | tree
history | raw | HEAD
Merge "Improve docs for Title::getInternalURL/getCanonicalURL"
[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 // FIXME: Don't use assert()
570 // phpcs:ignore MediaWiki.Usage.ForbiddenFunctions.assert
571 assert( $piece->open === "\n" );
572 $part = $piece->getCurrentPart();
573 // Search back through the input to see if it has a proper close.
574 // Do this using the reversed string since the other solutions
575 // (end anchor, etc.) are inefficient.
576 $wsLength = strspn( $revText, " \t", $lengthText - $i );
577 $searchStart = $i - $wsLength;
578 if ( isset( $part->commentEnd ) && $searchStart - 1 == $part->commentEnd ) {
579 // Comment found at line end
580 // Search for equals signs before the comment
581 $searchStart = $part->visualEnd;
582 $searchStart -= strspn( $revText, " \t", $lengthText - $searchStart );
583 }
584 $count = $piece->count;
585 $equalsLength = strspn( $revText, '=', $lengthText - $searchStart );
586 if ( $equalsLength > 0 ) {
587 if ( $searchStart - $equalsLength == $piece->startPos ) {
588 // This is just a single string of equals signs on its own line
589 // Replicate the doHeadings behavior /={count}(.+)={count}/
590 // First find out how many equals signs there really are (don't stop at 6)
591 $count = $equalsLength;
592 if ( $count < 3 ) {
593 $count = 0;
594 } else {
595 $count = min( 6, intval( ( $count - 1 ) / 2 ) );
596 }
597 } else {
598 $count = min( $equalsLength, $count );
599 }
600 if ( $count > 0 ) {
601 // Normal match, output <h>
602 $element = "<h level=\"$count\" i=\"$headingIndex\">$accum</h>";
603 $headingIndex++;
604 } else {
605 // Single equals sign on its own line, count=0
606 $element = $accum;
607 }
608 } else {
609 // No match, no <h>, just pass down the inner text
610 $element = $accum;
611 }
612 // Unwind the stack
613 $stack->pop();
614 $accum =& $stack->getAccum();
615 $stackFlags = $stack->getFlags();
616 if ( isset( $stackFlags['findEquals'] ) ) {
617 $findEquals = $stackFlags['findEquals'];
618 }
619 if ( isset( $stackFlags['findPipe'] ) ) {
620 $findPipe = $stackFlags['findPipe'];
621 }
622 if ( isset( $stackFlags['inHeading'] ) ) {
623 $inHeading = $stackFlags['inHeading'];
624 }
625
626 // Append the result to the enclosing accumulator
627 $accum .= $element;
628 // Note that we do NOT increment the input pointer.
629 // This is because the closing linebreak could be the opening linebreak of
630 // another heading. Infinite loops are avoided because the next iteration MUST
631 // hit the heading open case above, which unconditionally increments the
632 // input pointer.
633 } elseif ( $found == 'open' ) {
634 # count opening brace characters
635 $curLen = strlen( $curChar );
636 $count = ( $curLen > 1 ) ?
637 # allow the final character to repeat
638 strspn( $text, $curChar[$curLen - 1], $i + 1 ) + 1 :
639 strspn( $text, $curChar, $i );
640
641 $savedPrefix = '';
642 $lineStart = ( $i > 0 && $text[$i - 1] == "\n" );
643
644 if ( $curChar === "-{" && $count > $curLen ) {
645 // -{ => {{ transition because rightmost wins
646 $savedPrefix = '-';
647 $i++;
648 $curChar = '{';
649 $count--;
650 $rule = $this->rules[$curChar];
651 }
652
653 # we need to add to stack only if opening brace count is enough for one of the rules
654 if ( $count >= $rule['min'] ) {
655 # Add it to the stack
656 $piece = [
657 'open' => $curChar,
658 'close' => $rule['end'],
659 'savedPrefix' => $savedPrefix,
660 'count' => $count,
661 'lineStart' => $lineStart,
662 ];
663
664 $stack->push( $piece );
665 $accum =& $stack->getAccum();
666 $stackFlags = $stack->getFlags();
667 if ( isset( $stackFlags['findEquals'] ) ) {
668 $findEquals = $stackFlags['findEquals'];
669 }
670 if ( isset( $stackFlags['findPipe'] ) ) {
671 $findPipe = $stackFlags['findPipe'];
672 }
673 if ( isset( $stackFlags['inHeading'] ) ) {
674 $inHeading = $stackFlags['inHeading'];
675 }
676 } else {
677 # Add literal brace(s)
678 $accum .= htmlspecialchars( $savedPrefix . str_repeat( $curChar, $count ) );
679 }
680 $i += $count;
681 } elseif ( $found == 'close' ) {
682 $piece = $stack->top;
683 # lets check if there are enough characters for closing brace
684 $maxCount = $piece->count;
685 if ( $piece->close === '}-' && $curChar === '}' ) {
686 $maxCount--; # don't try to match closing '-' as a '}'
687 }
688 $curLen = strlen( $curChar );
689 $count = ( $curLen > 1 ) ? $curLen :
690 strspn( $text, $curChar, $i, $maxCount );
691
692 # check for maximum matching characters (if there are 5 closing
693 # characters, we will probably need only 3 - depending on the rules)
694 $rule = $this->rules[$piece->open];
695 if ( $count > $rule['max'] ) {
696 # The specified maximum exists in the callback array, unless the caller
697 # has made an error
698 $matchingCount = $rule['max'];
699 } else {
700 # Count is less than the maximum
701 # Skip any gaps in the callback array to find the true largest match
702 # Need to use array_key_exists not isset because the callback can be null
703 $matchingCount = $count;
704 while ( $matchingCount > 0 && !array_key_exists( $matchingCount, $rule['names'] ) ) {
705 --$matchingCount;
706 }
707 }
708
709 if ( $matchingCount <= 0 ) {
710 # No matching element found in callback array
711 # Output a literal closing brace and continue
712 $endText = substr( $text, $i, $count );
713 $accum .= htmlspecialchars( $endText );
714 $i += $count;
715 continue;
716 }
717 $name = $rule['names'][$matchingCount];
718 if ( $name === null ) {
719 // No element, just literal text
720 $endText = substr( $text, $i, $matchingCount );
721 $element = $piece->breakSyntax( $matchingCount ) . $endText;
722 } else {
723 # Create XML element
724 # Note: $parts is already XML, does not need to be encoded further
725 $parts = $piece->parts;
726 $title = $parts[0]->out;
727 unset( $parts[0] );
728
729 # The invocation is at the start of the line if lineStart is set in
730 # the stack, and all opening brackets are used up.
731 if ( $maxCount == $matchingCount &&
732 !empty( $piece->lineStart ) &&
733 strlen( $piece->savedPrefix ) == 0 ) {
734 $attr = ' lineStart="1"';
735 } else {
736 $attr = '';
737 }
738
739 $element = "<$name$attr>";
740 $element .= "<title>$title</title>";
741 $argIndex = 1;
742 foreach ( $parts as $part ) {
743 if ( isset( $part->eqpos ) ) {
744 $argName = substr( $part->out, 0, $part->eqpos );
745 $argValue = substr( $part->out, $part->eqpos + 1 );
746 $element .= "<part><name>$argName</name>=<value>$argValue</value></part>";
747 } else {
748 $element .= "<part><name index=\"$argIndex\" /><value>{$part->out}</value></part>";
749 $argIndex++;
750 }
751 }
752 $element .= "</$name>";
753 }
754
755 # Advance input pointer
756 $i += $matchingCount;
757
758 # Unwind the stack
759 $stack->pop();
760 $accum =& $stack->getAccum();
761
762 # Re-add the old stack element if it still has unmatched opening characters remaining
763 if ( $matchingCount < $piece->count ) {
764 $piece->parts = [ new PPDPart ];
765 $piece->count -= $matchingCount;
766 # do we still qualify for any callback with remaining count?
767 $min = $this->rules[$piece->open]['min'];
768 if ( $piece->count >= $min ) {
769 $stack->push( $piece );
770 $accum =& $stack->getAccum();
771 } elseif ( $piece->count == 1 && $piece->open === '{' && $piece->savedPrefix === '-' ) {
772 $piece->savedPrefix = '';
773 $piece->open = '-{';
774 $piece->count = 2;
775 $piece->close = $this->rules[$piece->open]['end'];
776 $stack->push( $piece );
777 $accum =& $stack->getAccum();
778 } else {
779 $s = substr( $piece->open, 0, -1 );
780 $s .= str_repeat(
781 substr( $piece->open, -1 ),
782 $piece->count - strlen( $s )
783 );
784 $accum .= $piece->savedPrefix . $s;
785 }
786 } elseif ( $piece->savedPrefix !== '' ) {
787 $accum .= $piece->savedPrefix;
788 }
789
790 $stackFlags = $stack->getFlags();
791 if ( isset( $stackFlags['findEquals'] ) ) {
792 $findEquals = $stackFlags['findEquals'];
793 }
794 if ( isset( $stackFlags['findPipe'] ) ) {
795 $findPipe = $stackFlags['findPipe'];
796 }
797 if ( isset( $stackFlags['inHeading'] ) ) {
798 $inHeading = $stackFlags['inHeading'];
799 }
800
801 # Add XML element to the enclosing accumulator
802 $accum .= $element;
803 } elseif ( $found == 'pipe' ) {
804 $findEquals = true; // shortcut for getFlags()
805 $stack->addPart();
806 $accum =& $stack->getAccum();
807 ++$i;
808 } elseif ( $found == 'equals' ) {
809 $findEquals = false; // shortcut for getFlags()
810 $stack->getCurrentPart()->eqpos = strlen( $accum );
811 $accum .= '=';
812 ++$i;
813 }
814 }
815
816 # Output any remaining unclosed brackets
817 foreach ( $stack->stack as $piece ) {
818 $stack->rootAccum .= $piece->breakSyntax();
819 }
820 $stack->rootAccum .= '</root>';
821 $xml = $stack->rootAccum;
822
823 return $xml;
824 }
825 }
826
827 /**
828 * Stack class to help Preprocessor::preprocessToObj()
829 * @ingroup Parser
830 */
831 class PPDStack {
832 public $stack, $rootAccum;
833
834 /**
835 * @var PPDStack
836 */
837 public $top;
838 public $out;
839 public $elementClass = PPDStackElement::class;
840
841 public static $false = false;
842
843 public function __construct() {
844 $this->stack = [];
845 $this->top = false;
846 $this->rootAccum = '';
847 $this->accum =& $this->rootAccum;
848 }
849
850 /**
851 * @return int
852 */
853 public function count() {
854 return count( $this->stack );
855 }
856
857 public function &getAccum() {
858 return $this->accum;
859 }
860
861 /**
862 * @return bool|PPDPart
863 */
864 public function getCurrentPart() {
865 if ( $this->top === false ) {
866 return false;
867 } else {
868 return $this->top->getCurrentPart();
869 }
870 }
871
872 public function push( $data ) {
873 if ( $data instanceof $this->elementClass ) {
874 $this->stack[] = $data;
875 } else {
876 $class = $this->elementClass;
877 $this->stack[] = new $class( $data );
878 }
879 $this->top = $this->stack[count( $this->stack ) - 1];
880 $this->accum =& $this->top->getAccum();
881 }
882
883 public function pop() {
884 if ( $this->stack === [] ) {
885 throw new MWException( __METHOD__ . ': no elements remaining' );
886 }
887 $temp = array_pop( $this->stack );
888
889 if ( count( $this->stack ) ) {
890 $this->top = $this->stack[count( $this->stack ) - 1];
891 $this->accum =& $this->top->getAccum();
892 } else {
893 $this->top = self::$false;
894 $this->accum =& $this->rootAccum;
895 }
896 return $temp;
897 }
898
899 public function addPart( $s = '' ) {
900 $this->top->addPart( $s );
901 $this->accum =& $this->top->getAccum();
902 }
903
904 /**
905 * @return array
906 */
907 public function getFlags() {
908 if ( $this->stack === [] ) {
909 return [
910 'findEquals' => false,
911 'findPipe' => false,
912 'inHeading' => false,
913 ];
914 } else {
915 return $this->top->getFlags();
916 }
917 }
918 }
919
920 /**
921 * @ingroup Parser
922 */
923 class PPDStackElement {
924 /**
925 * @var string Opening character (\n for heading)
926 */
927 public $open;
928
929 /**
930 * @var string Matching closing character
931 */
932 public $close;
933
934 /**
935 * @var string Saved prefix that may affect later processing,
936 * e.g. to differentiate `-{{{{` and `{{{{` after later seeing `}}}`.
937 */
938 public $savedPrefix = '';
939
940 /**
941 * @var int Number of opening characters found (number of "=" for heading)
942 */
943 public $count;
944
945 /**
946 * @var PPDPart[] Array of PPDPart objects describing pipe-separated parts.
947 */
948 public $parts;
949
950 /**
951 * @var bool True if the open char appeared at the start of the input line.
952 * Not set for headings.
953 */
954 public $lineStart;
955
956 public $partClass = PPDPart::class;
957
958 public function __construct( $data = [] ) {
959 $class = $this->partClass;
960 $this->parts = [ new $class ];
961
962 foreach ( $data as $name => $value ) {
963 $this->$name = $value;
964 }
965 }
966
967 public function &getAccum() {
968 return $this->parts[count( $this->parts ) - 1]->out;
969 }
970
971 public function addPart( $s = '' ) {
972 $class = $this->partClass;
973 $this->parts[] = new $class( $s );
974 }
975
976 /**
977 * @return PPDPart
978 */
979 public function getCurrentPart() {
980 return $this->parts[count( $this->parts ) - 1];
981 }
982
983 /**
984 * @return array
985 */
986 public function getFlags() {
987 $partCount = count( $this->parts );
988 $findPipe = $this->open != "\n" && $this->open != '[';
989 return [
990 'findPipe' => $findPipe,
991 'findEquals' => $findPipe && $partCount > 1 && !isset( $this->parts[$partCount - 1]->eqpos ),
992 'inHeading' => $this->open == "\n",
993 ];
994 }
995
996 /**
997 * Get the output string that would result if the close is not found.
998 *
999 * @param bool|int $openingCount
1000 * @return string
1001 */
1002 public function breakSyntax( $openingCount = false ) {
1003 if ( $this->open == "\n" ) {
1004 $s = $this->savedPrefix . $this->parts[0]->out;
1005 } else {
1006 if ( $openingCount === false ) {
1007 $openingCount = $this->count;
1008 }
1009 $s = substr( $this->open, 0, -1 );
1010 $s .= str_repeat(
1011 substr( $this->open, -1 ),
1012 $openingCount - strlen( $s )
1013 );
1014 $s = $this->savedPrefix . $s;
1015 $first = true;
1016 foreach ( $this->parts as $part ) {
1017 if ( $first ) {
1018 $first = false;
1019 } else {
1020 $s .= '|';
1021 }
1022 $s .= $part->out;
1023 }
1024 }
1025 return $s;
1026 }
1027 }
1028
1029 /**
1030 * @ingroup Parser
1031 */
1032 class PPDPart {
1033 /**
1034 * @var string Output accumulator string
1035 */
1036 public $out;
1037
1038 // Optional member variables:
1039 // eqpos Position of equals sign in output accumulator
1040 // commentEnd Past-the-end input pointer for the last comment encountered
1041 // visualEnd Past-the-end input pointer for the end of the accumulator minus comments
1042
1043 public function __construct( $out = '' ) {
1044 $this->out = $out;
1045 }
1046 }
1047
1048 /**
1049 * An expansion frame, used as a context to expand the result of preprocessToObj()
1050 * @ingroup Parser
1051 */
1052 // phpcs:ignore Squiz.Classes.ValidClassName.NotCamelCaps
1053 class PPFrame_DOM implements PPFrame {
1054
1055 /**
1056 * @var Preprocessor
1057 */
1058 public $preprocessor;
1059
1060 /**
1061 * @var Parser
1062 */
1063 public $parser;
1064
1065 /**
1066 * @var Title
1067 */
1068 public $title;
1069 public $titleCache;
1070
1071 /**
1072 * Hashtable listing templates which are disallowed for expansion in this frame,
1073 * having been encountered previously in parent frames.
1074 */
1075 public $loopCheckHash;
1076
1077 /**
1078 * Recursion depth of this frame, top = 0
1079 * Note that this is NOT the same as expansion depth in expand()
1080 */
1081 public $depth;
1082
1083 private $volatile = false;
1084 private $ttl = null;
1085
1086 /**
1087 * @var array
1088 */
1089 protected $childExpansionCache;
1090
1091 /**
1092 * Construct a new preprocessor frame.
1093 * @param Preprocessor $preprocessor The parent preprocessor
1094 */
1095 public function __construct( $preprocessor ) {
1096 $this->preprocessor = $preprocessor;
1097 $this->parser = $preprocessor->parser;
1098 $this->title = $this->parser->mTitle;
1099 $this->titleCache = [ $this->title ? $this->title->getPrefixedDBkey() : false ];
1100 $this->loopCheckHash = [];
1101 $this->depth = 0;
1102 $this->childExpansionCache = [];
1103 }
1104
1105 /**
1106 * Create a new child frame
1107 * $args is optionally a multi-root PPNode or array containing the template arguments
1108 *
1109 * @param bool|array $args
1110 * @param Title|bool $title
1111 * @param int $indexOffset
1112 * @return PPTemplateFrame_DOM
1113 */
1114 public function newChild( $args = false, $title = false, $indexOffset = 0 ) {
1115 $namedArgs = [];
1116 $numberedArgs = [];
1117 if ( $title === false ) {
1118 $title = $this->title;
1119 }
1120 if ( $args !== false ) {
1121 $xpath = false;
1122 if ( $args instanceof PPNode ) {
1123 $args = $args->node;
1124 }
1125 foreach ( $args as $arg ) {
1126 if ( $arg instanceof PPNode ) {
1127 $arg = $arg->node;
1128 }
1129 if ( !$xpath || $xpath->document !== $arg->ownerDocument ) {
1130 $xpath = new DOMXPath( $arg->ownerDocument );
1131 }
1132
1133 $nameNodes = $xpath->query( 'name', $arg );
1134 $value = $xpath->query( 'value', $arg );
1135 if ( $nameNodes->item( 0 )->hasAttributes() ) {
1136 // Numbered parameter
1137 $index = $nameNodes->item( 0 )->attributes->getNamedItem( 'index' )->textContent;
1138 $index = $index - $indexOffset;
1139 if ( isset( $namedArgs[$index] ) || isset( $numberedArgs[$index] ) ) {
1140 $this->parser->getOutput()->addWarning( wfMessage( 'duplicate-args-warning',
1141 wfEscapeWikiText( $this->title ),
1142 wfEscapeWikiText( $title ),
1143 wfEscapeWikiText( $index ) )->text() );
1144 $this->parser->addTrackingCategory( 'duplicate-args-category' );
1145 }
1146 $numberedArgs[$index] = $value->item( 0 );
1147 unset( $namedArgs[$index] );
1148 } else {
1149 // Named parameter
1150 $name = trim( $this->expand( $nameNodes->item( 0 ), PPFrame::STRIP_COMMENTS ) );
1151 if ( isset( $namedArgs[$name] ) || isset( $numberedArgs[$name] ) ) {
1152 $this->parser->getOutput()->addWarning( wfMessage( 'duplicate-args-warning',
1153 wfEscapeWikiText( $this->title ),
1154 wfEscapeWikiText( $title ),
1155 wfEscapeWikiText( $name ) )->text() );
1156 $this->parser->addTrackingCategory( 'duplicate-args-category' );
1157 }
1158 $namedArgs[$name] = $value->item( 0 );
1159 unset( $numberedArgs[$name] );
1160 }
1161 }
1162 }
1163 return new PPTemplateFrame_DOM( $this->preprocessor, $this, $numberedArgs, $namedArgs, $title );
1164 }
1165
1166 /**
1167 * @throws MWException
1168 * @param string|int $key
1169 * @param string|PPNode_DOM|DOMDocument $root
1170 * @param int $flags
1171 * @return string
1172 */
1173 public function cachedExpand( $key, $root, $flags = 0 ) {
1174 // we don't have a parent, so we don't have a cache
1175 return $this->expand( $root, $flags );
1176 }
1177
1178 /**
1179 * @throws MWException
1180 * @param string|PPNode_DOM|DOMDocument $root
1181 * @param int $flags
1182 * @return string
1183 */
1184 public function expand( $root, $flags = 0 ) {
1185 static $expansionDepth = 0;
1186 if ( is_string( $root ) ) {
1187 return $root;
1188 }
1189
1190 if ( ++$this->parser->mPPNodeCount > $this->parser->mOptions->getMaxPPNodeCount() ) {
1191 $this->parser->limitationWarn( 'node-count-exceeded',
1192 $this->parser->mPPNodeCount,
1193 $this->parser->mOptions->getMaxPPNodeCount()
1194 );
1195 return '<span class="error">Node-count limit exceeded</span>';
1196 }
1197
1198 if ( $expansionDepth > $this->parser->mOptions->getMaxPPExpandDepth() ) {
1199 $this->parser->limitationWarn( 'expansion-depth-exceeded',
1200 $expansionDepth,
1201 $this->parser->mOptions->getMaxPPExpandDepth()
1202 );
1203 return '<span class="error">Expansion depth limit exceeded</span>';
1204 }
1205 ++$expansionDepth;
1206 if ( $expansionDepth > $this->parser->mHighestExpansionDepth ) {
1207 $this->parser->mHighestExpansionDepth = $expansionDepth;
1208 }
1209
1210 if ( $root instanceof PPNode_DOM ) {
1211 $root = $root->node;
1212 }
1213 if ( $root instanceof DOMDocument ) {
1214 $root = $root->documentElement;
1215 }
1216
1217 $outStack = [ '', '' ];
1218 $iteratorStack = [ false, $root ];
1219 $indexStack = [ 0, 0 ];
1220
1221 while ( count( $iteratorStack ) > 1 ) {
1222 $level = count( $outStack ) - 1;
1223 $iteratorNode =& $iteratorStack[$level];
1224 $out =& $outStack[$level];
1225 $index =& $indexStack[$level];
1226
1227 if ( $iteratorNode instanceof PPNode_DOM ) {
1228 $iteratorNode = $iteratorNode->node;
1229 }
1230
1231 if ( is_array( $iteratorNode ) ) {
1232 if ( $index >= count( $iteratorNode ) ) {
1233 // All done with this iterator
1234 $iteratorStack[$level] = false;
1235 $contextNode = false;
1236 } else {
1237 $contextNode = $iteratorNode[$index];
1238 $index++;
1239 }
1240 } elseif ( $iteratorNode instanceof DOMNodeList ) {
1241 if ( $index >= $iteratorNode->length ) {
1242 // All done with this iterator
1243 $iteratorStack[$level] = false;
1244 $contextNode = false;
1245 } else {
1246 $contextNode = $iteratorNode->item( $index );
1247 $index++;
1248 }
1249 } else {
1250 // Copy to $contextNode and then delete from iterator stack,
1251 // because this is not an iterator but we do have to execute it once
1252 $contextNode = $iteratorStack[$level];
1253 $iteratorStack[$level] = false;
1254 }
1255
1256 if ( $contextNode instanceof PPNode_DOM ) {
1257 $contextNode = $contextNode->node;
1258 }
1259
1260 $newIterator = false;
1261
1262 if ( $contextNode === false ) {
1263 // nothing to do
1264 } elseif ( is_string( $contextNode ) ) {
1265 $out .= $contextNode;
1266 } elseif ( is_array( $contextNode ) || $contextNode instanceof DOMNodeList ) {
1267 $newIterator = $contextNode;
1268 } elseif ( $contextNode instanceof DOMNode ) {
1269 if ( $contextNode->nodeType == XML_TEXT_NODE ) {
1270 $out .= $contextNode->nodeValue;
1271 } elseif ( $contextNode->nodeName == 'template' ) {
1272 # Double-brace expansion
1273 $xpath = new DOMXPath( $contextNode->ownerDocument );
1274 $titles = $xpath->query( 'title', $contextNode );
1275 $title = $titles->item( 0 );
1276 $parts = $xpath->query( 'part', $contextNode );
1277 if ( $flags & PPFrame::NO_TEMPLATES ) {
1278 $newIterator = $this->virtualBracketedImplode( '{{', '|', '}}', $title, $parts );
1279 } else {
1280 $lineStart = $contextNode->getAttribute( 'lineStart' );
1281 $params = [
1282 'title' => new PPNode_DOM( $title ),
1283 'parts' => new PPNode_DOM( $parts ),
1284 'lineStart' => $lineStart ];
1285 $ret = $this->parser->braceSubstitution( $params, $this );
1286 if ( isset( $ret['object'] ) ) {
1287 $newIterator = $ret['object'];
1288 } else {
1289 $out .= $ret['text'];
1290 }
1291 }
1292 } elseif ( $contextNode->nodeName == 'tplarg' ) {
1293 # Triple-brace expansion
1294 $xpath = new DOMXPath( $contextNode->ownerDocument );
1295 $titles = $xpath->query( 'title', $contextNode );
1296 $title = $titles->item( 0 );
1297 $parts = $xpath->query( 'part', $contextNode );
1298 if ( $flags & PPFrame::NO_ARGS ) {
1299 $newIterator = $this->virtualBracketedImplode( '{{{', '|', '}}}', $title, $parts );
1300 } else {
1301 $params = [
1302 'title' => new PPNode_DOM( $title ),
1303 'parts' => new PPNode_DOM( $parts ) ];
1304 $ret = $this->parser->argSubstitution( $params, $this );
1305 if ( isset( $ret['object'] ) ) {
1306 $newIterator = $ret['object'];
1307 } else {
1308 $out .= $ret['text'];
1309 }
1310 }
1311 } elseif ( $contextNode->nodeName == 'comment' ) {
1312 # HTML-style comment
1313 # Remove it in HTML, pre+remove and STRIP_COMMENTS modes
1314 # Not in RECOVER_COMMENTS mode (msgnw) though.
1315 if ( ( $this->parser->ot['html']
1316 || ( $this->parser->ot['pre'] && $this->parser->mOptions->getRemoveComments() )
1317 || ( $flags & PPFrame::STRIP_COMMENTS )
1318 ) && !( $flags & PPFrame::RECOVER_COMMENTS )
1319 ) {
1320 $out .= '';
1321 } elseif ( $this->parser->ot['wiki'] && !( $flags & PPFrame::RECOVER_COMMENTS ) ) {
1322 # Add a strip marker in PST mode so that pstPass2() can
1323 # run some old-fashioned regexes on the result.
1324 # Not in RECOVER_COMMENTS mode (extractSections) though.
1325 $out .= $this->parser->insertStripItem( $contextNode->textContent );
1326 } else {
1327 # Recover the literal comment in RECOVER_COMMENTS and pre+no-remove
1328 $out .= $contextNode->textContent;
1329 }
1330 } elseif ( $contextNode->nodeName == 'ignore' ) {
1331 # Output suppression used by <includeonly> etc.
1332 # OT_WIKI will only respect <ignore> in substed templates.
1333 # The other output types respect it unless NO_IGNORE is set.
1334 # extractSections() sets NO_IGNORE and so never respects it.
1335 if ( ( !isset( $this->parent ) && $this->parser->ot['wiki'] )
1336 || ( $flags & PPFrame::NO_IGNORE )
1337 ) {
1338 $out .= $contextNode->textContent;
1339 } else {
1340 $out .= '';
1341 }
1342 } elseif ( $contextNode->nodeName == 'ext' ) {
1343 # Extension tag
1344 $xpath = new DOMXPath( $contextNode->ownerDocument );
1345 $names = $xpath->query( 'name', $contextNode );
1346 $attrs = $xpath->query( 'attr', $contextNode );
1347 $inners = $xpath->query( 'inner', $contextNode );
1348 $closes = $xpath->query( 'close', $contextNode );
1349 if ( $flags & PPFrame::NO_TAGS ) {
1350 $s = '<' . $this->expand( $names->item( 0 ), $flags );
1351 if ( $attrs->length > 0 ) {
1352 $s .= $this->expand( $attrs->item( 0 ), $flags );
1353 }
1354 if ( $inners->length > 0 ) {
1355 $s .= '>' . $this->expand( $inners->item( 0 ), $flags );
1356 if ( $closes->length > 0 ) {
1357 $s .= $this->expand( $closes->item( 0 ), $flags );
1358 }
1359 } else {
1360 $s .= '/>';
1361 }
1362 $out .= $s;
1363 } else {
1364 $params = [
1365 'name' => new PPNode_DOM( $names->item( 0 ) ),
1366 'attr' => $attrs->length > 0 ? new PPNode_DOM( $attrs->item( 0 ) ) : null,
1367 'inner' => $inners->length > 0 ? new PPNode_DOM( $inners->item( 0 ) ) : null,
1368 'close' => $closes->length > 0 ? new PPNode_DOM( $closes->item( 0 ) ) : null,
1369 ];
1370 $out .= $this->parser->extensionSubstitution( $params, $this );
1371 }
1372 } elseif ( $contextNode->nodeName == 'h' ) {
1373 # Heading
1374 $s = $this->expand( $contextNode->childNodes, $flags );
1375
1376 # Insert a heading marker only for <h> children of <root>
1377 # This is to stop extractSections from going over multiple tree levels
1378 if ( $contextNode->parentNode->nodeName == 'root' && $this->parser->ot['html'] ) {
1379 # Insert heading index marker
1380 $headingIndex = $contextNode->getAttribute( 'i' );
1381 $titleText = $this->title->getPrefixedDBkey();
1382 $this->parser->mHeadings[] = [ $titleText, $headingIndex ];
1383 $serial = count( $this->parser->mHeadings ) - 1;
1384 $marker = Parser::MARKER_PREFIX . "-h-$serial-" . Parser::MARKER_SUFFIX;
1385 $count = $contextNode->getAttribute( 'level' );
1386 $s = substr( $s, 0, $count ) . $marker . substr( $s, $count );
1387 $this->parser->mStripState->addGeneral( $marker, '' );
1388 }
1389 $out .= $s;
1390 } else {
1391 # Generic recursive expansion
1392 $newIterator = $contextNode->childNodes;
1393 }
1394 } else {
1395 throw new MWException( __METHOD__ . ': Invalid parameter type' );
1396 }
1397
1398 if ( $newIterator !== false ) {
1399 if ( $newIterator instanceof PPNode_DOM ) {
1400 $newIterator = $newIterator->node;
1401 }
1402 $outStack[] = '';
1403 $iteratorStack[] = $newIterator;
1404 $indexStack[] = 0;
1405 } elseif ( $iteratorStack[$level] === false ) {
1406 // Return accumulated value to parent
1407 // With tail recursion
1408 while ( $iteratorStack[$level] === false && $level > 0 ) {
1409 $outStack[$level - 1] .= $out;
1410 array_pop( $outStack );
1411 array_pop( $iteratorStack );
1412 array_pop( $indexStack );
1413 $level--;
1414 }
1415 }
1416 }
1417 --$expansionDepth;
1418 return $outStack[0];
1419 }
1420
1421 /**
1422 * @param string $sep
1423 * @param int $flags
1424 * @param string|PPNode_DOM|DOMDocument $args,...
1425 * @return string
1426 */
1427 public function implodeWithFlags( $sep, $flags /*, ... */ ) {
1428 $args = array_slice( func_get_args(), 2 );
1429
1430 $first = true;
1431 $s = '';
1432 foreach ( $args as $root ) {
1433 if ( $root instanceof PPNode_DOM ) {
1434 $root = $root->node;
1435 }
1436 if ( !is_array( $root ) && !( $root instanceof DOMNodeList ) ) {
1437 $root = [ $root ];
1438 }
1439 foreach ( $root as $node ) {
1440 if ( $first ) {
1441 $first = false;
1442 } else {
1443 $s .= $sep;
1444 }
1445 $s .= $this->expand( $node, $flags );
1446 }
1447 }
1448 return $s;
1449 }
1450
1451 /**
1452 * Implode with no flags specified
1453 * This previously called implodeWithFlags but has now been inlined to reduce stack depth
1454 *
1455 * @param string $sep
1456 * @param string|PPNode_DOM|DOMDocument $args,...
1457 * @return string
1458 */
1459 public function implode( $sep /*, ... */ ) {
1460 $args = array_slice( func_get_args(), 1 );
1461
1462 $first = true;
1463 $s = '';
1464 foreach ( $args as $root ) {
1465 if ( $root instanceof PPNode_DOM ) {
1466 $root = $root->node;
1467 }
1468 if ( !is_array( $root ) && !( $root instanceof DOMNodeList ) ) {
1469 $root = [ $root ];
1470 }
1471 foreach ( $root as $node ) {
1472 if ( $first ) {
1473 $first = false;
1474 } else {
1475 $s .= $sep;
1476 }
1477 $s .= $this->expand( $node );
1478 }
1479 }
1480 return $s;
1481 }
1482
1483 /**
1484 * Makes an object that, when expand()ed, will be the same as one obtained
1485 * with implode()
1486 *
1487 * @param string $sep
1488 * @param string|PPNode_DOM|DOMDocument $args,...
1489 * @return array
1490 */
1491 public function virtualImplode( $sep /*, ... */ ) {
1492 $args = array_slice( func_get_args(), 1 );
1493 $out = [];
1494 $first = true;
1495
1496 foreach ( $args as $root ) {
1497 if ( $root instanceof PPNode_DOM ) {
1498 $root = $root->node;
1499 }
1500 if ( !is_array( $root ) && !( $root instanceof DOMNodeList ) ) {
1501 $root = [ $root ];
1502 }
1503 foreach ( $root as $node ) {
1504 if ( $first ) {
1505 $first = false;
1506 } else {
1507 $out[] = $sep;
1508 }
1509 $out[] = $node;
1510 }
1511 }
1512 return $out;
1513 }
1514
1515 /**
1516 * Virtual implode with brackets
1517 * @param string $start
1518 * @param string $sep
1519 * @param string $end
1520 * @param string|PPNode_DOM|DOMDocument $args,...
1521 * @return array
1522 */
1523 public function virtualBracketedImplode( $start, $sep, $end /*, ... */ ) {
1524 $args = array_slice( func_get_args(), 3 );
1525 $out = [ $start ];
1526 $first = true;
1527
1528 foreach ( $args as $root ) {
1529 if ( $root instanceof PPNode_DOM ) {
1530 $root = $root->node;
1531 }
1532 if ( !is_array( $root ) && !( $root instanceof DOMNodeList ) ) {
1533 $root = [ $root ];
1534 }
1535 foreach ( $root as $node ) {
1536 if ( $first ) {
1537 $first = false;
1538 } else {
1539 $out[] = $sep;
1540 }
1541 $out[] = $node;
1542 }
1543 }
1544 $out[] = $end;
1545 return $out;
1546 }
1547
1548 public function __toString() {
1549 return 'frame{}';
1550 }
1551
1552 public function getPDBK( $level = false ) {
1553 if ( $level === false ) {
1554 return $this->title->getPrefixedDBkey();
1555 } else {
1556 return $this->titleCache[$level] ?? false;
1557 }
1558 }
1559
1560 /**
1561 * @return array
1562 */
1563 public function getArguments() {
1564 return [];
1565 }
1566
1567 /**
1568 * @return array
1569 */
1570 public function getNumberedArguments() {
1571 return [];
1572 }
1573
1574 /**
1575 * @return array
1576 */
1577 public function getNamedArguments() {
1578 return [];
1579 }
1580
1581 /**
1582 * Returns true if there are no arguments in this frame
1583 *
1584 * @return bool
1585 */
1586 public function isEmpty() {
1587 return true;
1588 }
1589
1590 /**
1591 * @param int|string $name
1592 * @return bool Always false in this implementation.
1593 */
1594 public function getArgument( $name ) {
1595 return false;
1596 }
1597
1598 /**
1599 * Returns true if the infinite loop check is OK, false if a loop is detected
1600 *
1601 * @param Title $title
1602 * @return bool
1603 */
1604 public function loopCheck( $title ) {
1605 return !isset( $this->loopCheckHash[$title->getPrefixedDBkey()] );
1606 }
1607
1608 /**
1609 * Return true if the frame is a template frame
1610 *
1611 * @return bool
1612 */
1613 public function isTemplate() {
1614 return false;
1615 }
1616
1617 /**
1618 * Get a title of frame
1619 *
1620 * @return Title
1621 */
1622 public function getTitle() {
1623 return $this->title;
1624 }
1625
1626 /**
1627 * Set the volatile flag
1628 *
1629 * @param bool $flag
1630 */
1631 public function setVolatile( $flag = true ) {
1632 $this->volatile = $flag;
1633 }
1634
1635 /**
1636 * Get the volatile flag
1637 *
1638 * @return bool
1639 */
1640 public function isVolatile() {
1641 return $this->volatile;
1642 }
1643
1644 /**
1645 * Set the TTL
1646 *
1647 * @param int $ttl
1648 */
1649 public function setTTL( $ttl ) {
1650 if ( $ttl !== null && ( $this->ttl === null || $ttl < $this->ttl ) ) {
1651 $this->ttl = $ttl;
1652 }
1653 }
1654
1655 /**
1656 * Get the TTL
1657 *
1658 * @return int|null
1659 */
1660 public function getTTL() {
1661 return $this->ttl;
1662 }
1663 }
1664
1665 /**
1666 * Expansion frame with template arguments
1667 * @ingroup Parser
1668 */
1669 // phpcs:ignore Squiz.Classes.ValidClassName.NotCamelCaps
1670 class PPTemplateFrame_DOM extends PPFrame_DOM {
1671
1672 public $numberedArgs, $namedArgs;
1673
1674 /**
1675 * @var PPFrame_DOM
1676 */
1677 public $parent;
1678 public $numberedExpansionCache, $namedExpansionCache;
1679
1680 /**
1681 * @param Preprocessor $preprocessor
1682 * @param bool|PPFrame_DOM $parent
1683 * @param array $numberedArgs
1684 * @param array $namedArgs
1685 * @param bool|Title $title
1686 */
1687 public function __construct( $preprocessor, $parent = false, $numberedArgs = [],
1688 $namedArgs = [], $title = false
1689 ) {
1690 parent::__construct( $preprocessor );
1691
1692 $this->parent = $parent;
1693 $this->numberedArgs = $numberedArgs;
1694 $this->namedArgs = $namedArgs;
1695 $this->title = $title;
1696 $pdbk = $title ? $title->getPrefixedDBkey() : false;
1697 $this->titleCache = $parent->titleCache;
1698 $this->titleCache[] = $pdbk;
1699 $this->loopCheckHash = /*clone*/ $parent->loopCheckHash;
1700 if ( $pdbk !== false ) {
1701 $this->loopCheckHash[$pdbk] = true;
1702 }
1703 $this->depth = $parent->depth + 1;
1704 $this->numberedExpansionCache = $this->namedExpansionCache = [];
1705 }
1706
1707 public function __toString() {
1708 $s = 'tplframe{';
1709 $first = true;
1710 $args = $this->numberedArgs + $this->namedArgs;
1711 foreach ( $args as $name => $value ) {
1712 if ( $first ) {
1713 $first = false;
1714 } else {
1715 $s .= ', ';
1716 }
1717 $s .= "\"$name\":\"" .
1718 str_replace( '"', '\\"', $value->ownerDocument->saveXML( $value ) ) . '"';
1719 }
1720 $s .= '}';
1721 return $s;
1722 }
1723
1724 /**
1725 * @throws MWException
1726 * @param string|int $key
1727 * @param string|PPNode_DOM|DOMDocument $root
1728 * @param int $flags
1729 * @return string
1730 */
1731 public function cachedExpand( $key, $root, $flags = 0 ) {
1732 if ( isset( $this->parent->childExpansionCache[$key] ) ) {
1733 return $this->parent->childExpansionCache[$key];
1734 }
1735 $retval = $this->expand( $root, $flags );
1736 if ( !$this->isVolatile() ) {
1737 $this->parent->childExpansionCache[$key] = $retval;
1738 }
1739 return $retval;
1740 }
1741
1742 /**
1743 * Returns true if there are no arguments in this frame
1744 *
1745 * @return bool
1746 */
1747 public function isEmpty() {
1748 return !count( $this->numberedArgs ) && !count( $this->namedArgs );
1749 }
1750
1751 public function getArguments() {
1752 $arguments = [];
1753 foreach ( array_merge(
1754 array_keys( $this->numberedArgs ),
1755 array_keys( $this->namedArgs ) ) as $key ) {
1756 $arguments[$key] = $this->getArgument( $key );
1757 }
1758 return $arguments;
1759 }
1760
1761 public function getNumberedArguments() {
1762 $arguments = [];
1763 foreach ( array_keys( $this->numberedArgs ) as $key ) {
1764 $arguments[$key] = $this->getArgument( $key );
1765 }
1766 return $arguments;
1767 }
1768
1769 public function getNamedArguments() {
1770 $arguments = [];
1771 foreach ( array_keys( $this->namedArgs ) as $key ) {
1772 $arguments[$key] = $this->getArgument( $key );
1773 }
1774 return $arguments;
1775 }
1776
1777 /**
1778 * @param int $index
1779 * @return string|bool
1780 */
1781 public function getNumberedArgument( $index ) {
1782 if ( !isset( $this->numberedArgs[$index] ) ) {
1783 return false;
1784 }
1785 if ( !isset( $this->numberedExpansionCache[$index] ) ) {
1786 # No trimming for unnamed arguments
1787 $this->numberedExpansionCache[$index] = $this->parent->expand(
1788 $this->numberedArgs[$index],
1789 PPFrame::STRIP_COMMENTS
1790 );
1791 }
1792 return $this->numberedExpansionCache[$index];
1793 }
1794
1795 /**
1796 * @param string $name
1797 * @return string|bool
1798 */
1799 public function getNamedArgument( $name ) {
1800 if ( !isset( $this->namedArgs[$name] ) ) {
1801 return false;
1802 }
1803 if ( !isset( $this->namedExpansionCache[$name] ) ) {
1804 # Trim named arguments post-expand, for backwards compatibility
1805 $this->namedExpansionCache[$name] = trim(
1806 $this->parent->expand( $this->namedArgs[$name], PPFrame::STRIP_COMMENTS ) );
1807 }
1808 return $this->namedExpansionCache[$name];
1809 }
1810
1811 /**
1812 * @param int|string $name
1813 * @return string|bool
1814 */
1815 public function getArgument( $name ) {
1816 $text = $this->getNumberedArgument( $name );
1817 if ( $text === false ) {
1818 $text = $this->getNamedArgument( $name );
1819 }
1820 return $text;
1821 }
1822
1823 /**
1824 * Return true if the frame is a template frame
1825 *
1826 * @return bool
1827 */
1828 public function isTemplate() {
1829 return true;
1830 }
1831
1832 public function setVolatile( $flag = true ) {
1833 parent::setVolatile( $flag );
1834 $this->parent->setVolatile( $flag );
1835 }
1836
1837 public function setTTL( $ttl ) {
1838 parent::setTTL( $ttl );
1839 $this->parent->setTTL( $ttl );
1840 }
1841 }
1842
1843 /**
1844 * Expansion frame with custom arguments
1845 * @ingroup Parser
1846 */
1847 // phpcs:ignore Squiz.Classes.ValidClassName.NotCamelCaps
1848 class PPCustomFrame_DOM extends PPFrame_DOM {
1849
1850 public $args;
1851
1852 public function __construct( $preprocessor, $args ) {
1853 parent::__construct( $preprocessor );
1854 $this->args = $args;
1855 }
1856
1857 public function __toString() {
1858 $s = 'cstmframe{';
1859 $first = true;
1860 foreach ( $this->args as $name => $value ) {
1861 if ( $first ) {
1862 $first = false;
1863 } else {
1864 $s .= ', ';
1865 }
1866 $s .= "\"$name\":\"" .
1867 str_replace( '"', '\\"', $value->__toString() ) . '"';
1868 }
1869 $s .= '}';
1870 return $s;
1871 }
1872
1873 /**
1874 * @return bool
1875 */
1876 public function isEmpty() {
1877 return !count( $this->args );
1878 }
1879
1880 /**
1881 * @param int|string $index
1882 * @return string|bool
1883 */
1884 public function getArgument( $index ) {
1885 return $this->args[$index] ?? false;
1886 }
1887
1888 public function getArguments() {
1889 return $this->args;
1890 }
1891 }
1892
1893 /**
1894 * @ingroup Parser
1895 */
1896 // phpcs:ignore Squiz.Classes.ValidClassName.NotCamelCaps
1897 class PPNode_DOM implements PPNode {
1898
1899 /**
1900 * @var DOMElement
1901 */
1902 public $node;
1903 public $xpath;
1904
1905 public function __construct( $node, $xpath = false ) {
1906 $this->node = $node;
1907 }
1908
1909 /**
1910 * @return DOMXPath
1911 */
1912 public function getXPath() {
1913 if ( $this->xpath === null ) {
1914 $this->xpath = new DOMXPath( $this->node->ownerDocument );
1915 }
1916 return $this->xpath;
1917 }
1918
1919 public function __toString() {
1920 if ( $this->node instanceof DOMNodeList ) {
1921 $s = '';
1922 foreach ( $this->node as $node ) {
1923 $s .= $node->ownerDocument->saveXML( $node );
1924 }
1925 } else {
1926 $s = $this->node->ownerDocument->saveXML( $this->node );
1927 }
1928 return $s;
1929 }
1930
1931 /**
1932 * @return bool|PPNode_DOM
1933 */
1934 public function getChildren() {
1935 return $this->node->childNodes ? new self( $this->node->childNodes ) : false;
1936 }
1937
1938 /**
1939 * @return bool|PPNode_DOM
1940 */
1941 public function getFirstChild() {
1942 return $this->node->firstChild ? new self( $this->node->firstChild ) : false;
1943 }
1944
1945 /**
1946 * @return bool|PPNode_DOM
1947 */
1948 public function getNextSibling() {
1949 return $this->node->nextSibling ? new self( $this->node->nextSibling ) : false;
1950 }
1951
1952 /**
1953 * @param string $type
1954 *
1955 * @return bool|PPNode_DOM
1956 */
1957 public function getChildrenOfType( $type ) {
1958 return new self( $this->getXPath()->query( $type, $this->node ) );
1959 }
1960
1961 /**
1962 * @return int
1963 */
1964 public function getLength() {
1965 if ( $this->node instanceof DOMNodeList ) {
1966 return $this->node->length;
1967 } else {
1968 return false;
1969 }
1970 }
1971
1972 /**
1973 * @param int $i
1974 * @return bool|PPNode_DOM
1975 */
1976 public function item( $i ) {
1977 $item = $this->node->item( $i );
1978 return $item ? new self( $item ) : false;
1979 }
1980
1981 /**
1982 * @return string
1983 */
1984 public function getName() {
1985 if ( $this->node instanceof DOMNodeList ) {
1986 return '#nodelist';
1987 } else {
1988 return $this->node->nodeName;
1989 }
1990 }
1991
1992 /**
1993 * Split a "<part>" node into an associative array containing:
1994 * - name PPNode name
1995 * - index String index
1996 * - value PPNode value
1997 *
1998 * @throws MWException
1999 * @return array
2000 */
2001 public function splitArg() {
2002 $xpath = $this->getXPath();
2003 $names = $xpath->query( 'name', $this->node );
2004 $values = $xpath->query( 'value', $this->node );
2005 if ( !$names->length || !$values->length ) {
2006 throw new MWException( 'Invalid brace node passed to ' . __METHOD__ );
2007 }
2008 $name = $names->item( 0 );
2009 $index = $name->getAttribute( 'index' );
2010 return [
2011 'name' => new self( $name ),
2012 'index' => $index,
2013 'value' => new self( $values->item( 0 ) ) ];
2014 }
2015
2016 /**
2017 * Split an "<ext>" node into an associative array containing name, attr, inner and close
2018 * All values in the resulting array are PPNodes. Inner and close are optional.
2019 *
2020 * @throws MWException
2021 * @return array
2022 */
2023 public function splitExt() {
2024 $xpath = $this->getXPath();
2025 $names = $xpath->query( 'name', $this->node );
2026 $attrs = $xpath->query( 'attr', $this->node );
2027 $inners = $xpath->query( 'inner', $this->node );
2028 $closes = $xpath->query( 'close', $this->node );
2029 if ( !$names->length || !$attrs->length ) {
2030 throw new MWException( 'Invalid ext node passed to ' . __METHOD__ );
2031 }
2032 $parts = [
2033 'name' => new self( $names->item( 0 ) ),
2034 'attr' => new self( $attrs->item( 0 ) ) ];
2035 if ( $inners->length ) {
2036 $parts['inner'] = new self( $inners->item( 0 ) );
2037 }
2038 if ( $closes->length ) {
2039 $parts['close'] = new self( $closes->item( 0 ) );
2040 }
2041 return $parts;
2042 }
2043
2044 /**
2045 * Split a "<h>" node
2046 * @throws MWException
2047 * @return array
2048 */
2049 public function splitHeading() {
2050 if ( $this->getName() !== 'h' ) {
2051 throw new MWException( 'Invalid h node passed to ' . __METHOD__ );
2052 }
2053 return [
2054 'i' => $this->node->getAttribute( 'i' ),
2055 'level' => $this->node->getAttribute( 'level' ),
2056 'contents' => $this->getChildren()
2057 ];
2058 }
2059 }
Wiklou, le Wiki du Wiklou