Merge "Change 'editfont' default preference to 'monospace'"
[lhc/web/wiklou.git] / includes / parser / Preprocessor_Hash.php
1 <?php
2 /**
3 * Preprocessor using PHP arrays
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 * Differences from DOM schema:
26 * * attribute nodes are children
27 * * "<h>" nodes that aren't at the top are replaced with <possible-h>
28 *
29 * Nodes are stored in a recursive array data structure. A node store is an
30 * array where each element may be either a scalar (representing a text node)
31 * or a "descriptor", which is a two-element array where the first element is
32 * the node name and the second element is the node store for the children.
33 *
34 * Attributes are represented as children that have a node name starting with
35 * "@", and a single text node child.
36 *
37 * @todo: Consider replacing descriptor arrays with objects of a new class.
38 * Benchmark and measure resulting memory impact.
39 *
40 * @ingroup Parser
41 */
42 // @codingStandardsIgnoreStart Squiz.Classes.ValidClassName.NotCamelCaps
43 class Preprocessor_Hash extends Preprocessor {
44 // @codingStandardsIgnoreEnd
45
46 /**
47 * @var Parser
48 */
49 public $parser;
50
51 const CACHE_PREFIX = 'preprocess-hash';
52 const CACHE_VERSION = 2;
53
54 public function __construct( $parser ) {
55 $this->parser = $parser;
56 }
57
58 /**
59 * @return PPFrame_Hash
60 */
61 public function newFrame() {
62 return new PPFrame_Hash( $this );
63 }
64
65 /**
66 * @param array $args
67 * @return PPCustomFrame_Hash
68 */
69 public function newCustomFrame( $args ) {
70 return new PPCustomFrame_Hash( $this, $args );
71 }
72
73 /**
74 * @param array $values
75 * @return PPNode_Hash_Array
76 */
77 public function newPartNodeArray( $values ) {
78 $list = [];
79
80 foreach ( $values as $k => $val ) {
81 if ( is_int( $k ) ) {
82 $store = [ [ 'part', [
83 [ 'name', [ [ '@index', [ $k ] ] ] ],
84 [ 'value', [ strval( $val ) ] ],
85 ] ] ];
86 } else {
87 $store = [ [ 'part', [
88 [ 'name', [ strval( $k ) ] ],
89 '=',
90 [ 'value', [ strval( $val ) ] ],
91 ] ] ];
92 }
93
94 $list[] = new PPNode_Hash_Tree( $store, 0 );
95 }
96
97 $node = new PPNode_Hash_Array( $list );
98 return $node;
99 }
100
101 /**
102 * Preprocess some wikitext and return the document tree.
103 *
104 * @param string $text The text to parse
105 * @param int $flags Bitwise combination of:
106 * Parser::PTD_FOR_INCLUSION Handle "<noinclude>" and "<includeonly>" as if the text is being
107 * included. Default is to assume a direct page view.
108 *
109 * The generated DOM tree must depend only on the input text and the flags.
110 * The DOM tree must be the same in OT_HTML and OT_WIKI mode, to avoid a regression of T6899.
111 *
112 * Any flag added to the $flags parameter here, or any other parameter liable to cause a
113 * change in the DOM tree for a given text, must be passed through the section identifier
114 * in the section edit link and thus back to extractSections().
115 *
116 * @throws MWException
117 * @return PPNode_Hash_Tree
118 */
119 public function preprocessToObj( $text, $flags = 0 ) {
120 global $wgDisableLangConversion;
121
122 $tree = $this->cacheGetTree( $text, $flags );
123 if ( $tree !== false ) {
124 $store = json_decode( $tree );
125 if ( is_array( $store ) ) {
126 return new PPNode_Hash_Tree( $store, 0 );
127 }
128 }
129
130 $forInclusion = $flags & Parser::PTD_FOR_INCLUSION;
131
132 $xmlishElements = $this->parser->getStripList();
133 $xmlishAllowMissingEndTag = [ 'includeonly', 'noinclude', 'onlyinclude' ];
134 $enableOnlyinclude = false;
135 if ( $forInclusion ) {
136 $ignoredTags = [ 'includeonly', '/includeonly' ];
137 $ignoredElements = [ 'noinclude' ];
138 $xmlishElements[] = 'noinclude';
139 if ( strpos( $text, '<onlyinclude>' ) !== false
140 && strpos( $text, '</onlyinclude>' ) !== false
141 ) {
142 $enableOnlyinclude = true;
143 }
144 } else {
145 $ignoredTags = [ 'noinclude', '/noinclude', 'onlyinclude', '/onlyinclude' ];
146 $ignoredElements = [ 'includeonly' ];
147 $xmlishElements[] = 'includeonly';
148 }
149 $xmlishRegex = implode( '|', array_merge( $xmlishElements, $ignoredTags ) );
150
151 // Use "A" modifier (anchored) instead of "^", because ^ doesn't work with an offset
152 $elementsRegex = "~($xmlishRegex)(?:\s|\/>|>)|(!--)~iA";
153
154 $stack = new PPDStack_Hash;
155
156 $searchBase = "[{<\n";
157 if ( !$wgDisableLangConversion ) {
158 $searchBase .= '-';
159 }
160
161 // For fast reverse searches
162 $revText = strrev( $text );
163 $lengthText = strlen( $text );
164
165 // Input pointer, starts out pointing to a pseudo-newline before the start
166 $i = 0;
167 // Current accumulator. See the doc comment for Preprocessor_Hash for the format.
168 $accum =& $stack->getAccum();
169 // True to find equals signs in arguments
170 $findEquals = false;
171 // True to take notice of pipe characters
172 $findPipe = false;
173 $headingIndex = 1;
174 // True if $i is inside a possible heading
175 $inHeading = false;
176 // True if there are no more greater-than (>) signs right of $i
177 $noMoreGT = false;
178 // Map of tag name => true if there are no more closing tags of given type right of $i
179 $noMoreClosingTag = [];
180 // True to ignore all input up to the next <onlyinclude>
181 $findOnlyinclude = $enableOnlyinclude;
182 // Do a line-start run without outputting an LF character
183 $fakeLineStart = true;
184
185 while ( true ) {
186 // $this->memCheck();
187
188 if ( $findOnlyinclude ) {
189 // Ignore all input up to the next <onlyinclude>
190 $startPos = strpos( $text, '<onlyinclude>', $i );
191 if ( $startPos === false ) {
192 // Ignored section runs to the end
193 $accum[] = [ 'ignore', [ substr( $text, $i ) ] ];
194 break;
195 }
196 $tagEndPos = $startPos + strlen( '<onlyinclude>' ); // past-the-end
197 $accum[] = [ 'ignore', [ substr( $text, $i, $tagEndPos - $i ) ] ];
198 $i = $tagEndPos;
199 $findOnlyinclude = false;
200 }
201
202 if ( $fakeLineStart ) {
203 $found = 'line-start';
204 $curChar = '';
205 } else {
206 # Find next opening brace, closing brace or pipe
207 $search = $searchBase;
208 if ( $stack->top === false ) {
209 $currentClosing = '';
210 } elseif (
211 $stack->top->close === '}-' &&
212 $stack->top->count > 2
213 ) {
214 # adjust closing for -{{{...{{
215 $currentClosing = '}';
216 $search .= $currentClosing;
217 } else {
218 $currentClosing = $stack->top->close;
219 $search .= $currentClosing;
220 }
221 if ( $findPipe ) {
222 $search .= '|';
223 }
224 if ( $findEquals ) {
225 // First equals will be for the template
226 $search .= '=';
227 }
228 $rule = null;
229 # Output literal section, advance input counter
230 $literalLength = strcspn( $text, $search, $i );
231 if ( $literalLength > 0 ) {
232 self::addLiteral( $accum, substr( $text, $i, $literalLength ) );
233 $i += $literalLength;
234 }
235 if ( $i >= $lengthText ) {
236 if ( $currentClosing == "\n" ) {
237 // Do a past-the-end run to finish off the heading
238 $curChar = '';
239 $found = 'line-end';
240 } else {
241 # All done
242 break;
243 }
244 } else {
245 $curChar = $curTwoChar = $text[$i];
246 if ( ( $i + 1 ) < $lengthText ) {
247 $curTwoChar .= $text[$i + 1];
248 }
249 if ( $curChar == '|' ) {
250 $found = 'pipe';
251 } elseif ( $curChar == '=' ) {
252 $found = 'equals';
253 } elseif ( $curChar == '<' ) {
254 $found = 'angle';
255 } elseif ( $curChar == "\n" ) {
256 if ( $inHeading ) {
257 $found = 'line-end';
258 } else {
259 $found = 'line-start';
260 }
261 } elseif ( $curTwoChar == $currentClosing ) {
262 $found = 'close';
263 $curChar = $curTwoChar;
264 } elseif ( $curChar == $currentClosing ) {
265 $found = 'close';
266 } elseif ( isset( $this->rules[$curTwoChar] ) ) {
267 $curChar = $curTwoChar;
268 $found = 'open';
269 $rule = $this->rules[$curChar];
270 } elseif ( isset( $this->rules[$curChar] ) ) {
271 $found = 'open';
272 $rule = $this->rules[$curChar];
273 } else {
274 # Some versions of PHP have a strcspn which stops on
275 # null characters; ignore these and continue.
276 # We also may get '-' and '}' characters here which
277 # don't match -{ or $currentClosing. Add these to
278 # output and continue.
279 if ( $curChar == '-' || $curChar == '}' ) {
280 self::addLiteral( $accum, $curChar );
281 }
282 ++$i;
283 continue;
284 }
285 }
286 }
287
288 if ( $found == 'angle' ) {
289 $matches = false;
290 // Handle </onlyinclude>
291 if ( $enableOnlyinclude
292 && substr( $text, $i, strlen( '</onlyinclude>' ) ) == '</onlyinclude>'
293 ) {
294 $findOnlyinclude = true;
295 continue;
296 }
297
298 // Determine element name
299 if ( !preg_match( $elementsRegex, $text, $matches, 0, $i + 1 ) ) {
300 // Element name missing or not listed
301 self::addLiteral( $accum, '<' );
302 ++$i;
303 continue;
304 }
305 // Handle comments
306 if ( isset( $matches[2] ) && $matches[2] == '!--' ) {
307 // To avoid leaving blank lines, when a sequence of
308 // space-separated comments is both preceded and followed by
309 // a newline (ignoring spaces), then
310 // trim leading and trailing spaces and the trailing newline.
311
312 // Find the end
313 $endPos = strpos( $text, '-->', $i + 4 );
314 if ( $endPos === false ) {
315 // Unclosed comment in input, runs to end
316 $inner = substr( $text, $i );
317 $accum[] = [ 'comment', [ $inner ] ];
318 $i = $lengthText;
319 } else {
320 // Search backwards for leading whitespace
321 $wsStart = $i ? ( $i - strspn( $revText, " \t", $lengthText - $i ) ) : 0;
322
323 // Search forwards for trailing whitespace
324 // $wsEnd will be the position of the last space (or the '>' if there's none)
325 $wsEnd = $endPos + 2 + strspn( $text, " \t", $endPos + 3 );
326
327 // Keep looking forward as long as we're finding more
328 // comments.
329 $comments = [ [ $wsStart, $wsEnd ] ];
330 while ( substr( $text, $wsEnd + 1, 4 ) == '<!--' ) {
331 $c = strpos( $text, '-->', $wsEnd + 4 );
332 if ( $c === false ) {
333 break;
334 }
335 $c = $c + 2 + strspn( $text, " \t", $c + 3 );
336 $comments[] = [ $wsEnd + 1, $c ];
337 $wsEnd = $c;
338 }
339
340 // Eat the line if possible
341 // TODO: This could theoretically be done if $wsStart == 0, i.e. for comments at
342 // the overall start. That's not how Sanitizer::removeHTMLcomments() did it, but
343 // it's a possible beneficial b/c break.
344 if ( $wsStart > 0 && substr( $text, $wsStart - 1, 1 ) == "\n"
345 && substr( $text, $wsEnd + 1, 1 ) == "\n"
346 ) {
347 // Remove leading whitespace from the end of the accumulator
348 $wsLength = $i - $wsStart;
349 $endIndex = count( $accum ) - 1;
350
351 // Sanity check
352 if ( $wsLength > 0
353 && $endIndex >= 0
354 && is_string( $accum[$endIndex] )
355 && strspn( $accum[$endIndex], " \t", -$wsLength ) === $wsLength
356 ) {
357 $accum[$endIndex] = substr( $accum[$endIndex], 0, -$wsLength );
358 }
359
360 // Dump all but the last comment to the accumulator
361 foreach ( $comments as $j => $com ) {
362 $startPos = $com[0];
363 $endPos = $com[1] + 1;
364 if ( $j == ( count( $comments ) - 1 ) ) {
365 break;
366 }
367 $inner = substr( $text, $startPos, $endPos - $startPos );
368 $accum[] = [ 'comment', [ $inner ] ];
369 }
370
371 // Do a line-start run next time to look for headings after the comment
372 $fakeLineStart = true;
373 } else {
374 // No line to eat, just take the comment itself
375 $startPos = $i;
376 $endPos += 2;
377 }
378
379 if ( $stack->top ) {
380 $part = $stack->top->getCurrentPart();
381 if ( !( isset( $part->commentEnd ) && $part->commentEnd == $wsStart - 1 ) ) {
382 $part->visualEnd = $wsStart;
383 }
384 // Else comments abutting, no change in visual end
385 $part->commentEnd = $endPos;
386 }
387 $i = $endPos + 1;
388 $inner = substr( $text, $startPos, $endPos - $startPos + 1 );
389 $accum[] = [ 'comment', [ $inner ] ];
390 }
391 continue;
392 }
393 $name = $matches[1];
394 $lowerName = strtolower( $name );
395 $attrStart = $i + strlen( $name ) + 1;
396
397 // Find end of tag
398 $tagEndPos = $noMoreGT ? false : strpos( $text, '>', $attrStart );
399 if ( $tagEndPos === false ) {
400 // Infinite backtrack
401 // Disable tag search to prevent worst-case O(N^2) performance
402 $noMoreGT = true;
403 self::addLiteral( $accum, '<' );
404 ++$i;
405 continue;
406 }
407
408 // Handle ignored tags
409 if ( in_array( $lowerName, $ignoredTags ) ) {
410 $accum[] = [ 'ignore', [ substr( $text, $i, $tagEndPos - $i + 1 ) ] ];
411 $i = $tagEndPos + 1;
412 continue;
413 }
414
415 $tagStartPos = $i;
416 if ( $text[$tagEndPos - 1] == '/' ) {
417 // Short end tag
418 $attrEnd = $tagEndPos - 1;
419 $inner = null;
420 $i = $tagEndPos + 1;
421 $close = null;
422 } else {
423 $attrEnd = $tagEndPos;
424 // Find closing tag
425 if (
426 !isset( $noMoreClosingTag[$name] ) &&
427 preg_match( "/<\/" . preg_quote( $name, '/' ) . "\s*>/i",
428 $text, $matches, PREG_OFFSET_CAPTURE, $tagEndPos + 1 )
429 ) {
430 $inner = substr( $text, $tagEndPos + 1, $matches[0][1] - $tagEndPos - 1 );
431 $i = $matches[0][1] + strlen( $matches[0][0] );
432 $close = $matches[0][0];
433 } else {
434 // No end tag
435 if ( in_array( $name, $xmlishAllowMissingEndTag ) ) {
436 // Let it run out to the end of the text.
437 $inner = substr( $text, $tagEndPos + 1 );
438 $i = $lengthText;
439 $close = null;
440 } else {
441 // Don't match the tag, treat opening tag as literal and resume parsing.
442 $i = $tagEndPos + 1;
443 self::addLiteral( $accum,
444 substr( $text, $tagStartPos, $tagEndPos + 1 - $tagStartPos ) );
445 // Cache results, otherwise we have O(N^2) performance for input like <foo><foo><foo>...
446 $noMoreClosingTag[$name] = true;
447 continue;
448 }
449 }
450 }
451 // <includeonly> and <noinclude> just become <ignore> tags
452 if ( in_array( $lowerName, $ignoredElements ) ) {
453 $accum[] = [ 'ignore', [ substr( $text, $tagStartPos, $i - $tagStartPos ) ] ];
454 continue;
455 }
456
457 if ( $attrEnd <= $attrStart ) {
458 $attr = '';
459 } else {
460 // Note that the attr element contains the whitespace between name and attribute,
461 // this is necessary for precise reconstruction during pre-save transform.
462 $attr = substr( $text, $attrStart, $attrEnd - $attrStart );
463 }
464
465 $children = [
466 [ 'name', [ $name ] ],
467 [ 'attr', [ $attr ] ] ];
468 if ( $inner !== null ) {
469 $children[] = [ 'inner', [ $inner ] ];
470 }
471 if ( $close !== null ) {
472 $children[] = [ 'close', [ $close ] ];
473 }
474 $accum[] = [ 'ext', $children ];
475 } elseif ( $found == 'line-start' ) {
476 // Is this the start of a heading?
477 // Line break belongs before the heading element in any case
478 if ( $fakeLineStart ) {
479 $fakeLineStart = false;
480 } else {
481 self::addLiteral( $accum, $curChar );
482 $i++;
483 }
484
485 $count = strspn( $text, '=', $i, 6 );
486 if ( $count == 1 && $findEquals ) {
487 // DWIM: This looks kind of like a name/value separator.
488 // Let's let the equals handler have it and break the potential
489 // heading. This is heuristic, but AFAICT the methods for
490 // completely correct disambiguation are very complex.
491 } elseif ( $count > 0 ) {
492 $piece = [
493 'open' => "\n",
494 'close' => "\n",
495 'parts' => [ new PPDPart_Hash( str_repeat( '=', $count ) ) ],
496 'startPos' => $i,
497 'count' => $count ];
498 $stack->push( $piece );
499 $accum =& $stack->getAccum();
500 extract( $stack->getFlags() );
501 $i += $count;
502 }
503 } elseif ( $found == 'line-end' ) {
504 $piece = $stack->top;
505 // A heading must be open, otherwise \n wouldn't have been in the search list
506 assert( $piece->open === "\n" );
507 $part = $piece->getCurrentPart();
508 // Search back through the input to see if it has a proper close.
509 // Do this using the reversed string since the other solutions
510 // (end anchor, etc.) are inefficient.
511 $wsLength = strspn( $revText, " \t", $lengthText - $i );
512 $searchStart = $i - $wsLength;
513 if ( isset( $part->commentEnd ) && $searchStart - 1 == $part->commentEnd ) {
514 // Comment found at line end
515 // Search for equals signs before the comment
516 $searchStart = $part->visualEnd;
517 $searchStart -= strspn( $revText, " \t", $lengthText - $searchStart );
518 }
519 $count = $piece->count;
520 $equalsLength = strspn( $revText, '=', $lengthText - $searchStart );
521 if ( $equalsLength > 0 ) {
522 if ( $searchStart - $equalsLength == $piece->startPos ) {
523 // This is just a single string of equals signs on its own line
524 // Replicate the doHeadings behavior /={count}(.+)={count}/
525 // First find out how many equals signs there really are (don't stop at 6)
526 $count = $equalsLength;
527 if ( $count < 3 ) {
528 $count = 0;
529 } else {
530 $count = min( 6, intval( ( $count - 1 ) / 2 ) );
531 }
532 } else {
533 $count = min( $equalsLength, $count );
534 }
535 if ( $count > 0 ) {
536 // Normal match, output <h>
537 $element = [ [ 'possible-h',
538 array_merge(
539 [
540 [ '@level', [ $count ] ],
541 [ '@i', [ $headingIndex++ ] ]
542 ],
543 $accum
544 )
545 ] ];
546 } else {
547 // Single equals sign on its own line, count=0
548 $element = $accum;
549 }
550 } else {
551 // No match, no <h>, just pass down the inner text
552 $element = $accum;
553 }
554 // Unwind the stack
555 $stack->pop();
556 $accum =& $stack->getAccum();
557 extract( $stack->getFlags() );
558
559 // Append the result to the enclosing accumulator
560 array_splice( $accum, count( $accum ), 0, $element );
561
562 // Note that we do NOT increment the input pointer.
563 // This is because the closing linebreak could be the opening linebreak of
564 // another heading. Infinite loops are avoided because the next iteration MUST
565 // hit the heading open case above, which unconditionally increments the
566 // input pointer.
567 } elseif ( $found == 'open' ) {
568 # count opening brace characters
569 $curLen = strlen( $curChar );
570 $count = ( $curLen > 1 ) ?
571 # allow the final character to repeat
572 strspn( $text, $curChar[$curLen - 1], $i + 1 ) + 1 :
573 strspn( $text, $curChar, $i );
574
575 # we need to add to stack only if opening brace count is enough for one of the rules
576 if ( $count >= $rule['min'] ) {
577 # Add it to the stack
578 $piece = [
579 'open' => $curChar,
580 'close' => $rule['end'],
581 'count' => $count,
582 'lineStart' => ( $i > 0 && $text[$i - 1] == "\n" ),
583 ];
584
585 $stack->push( $piece );
586 $accum =& $stack->getAccum();
587 extract( $stack->getFlags() );
588 } else {
589 # Add literal brace(s)
590 self::addLiteral( $accum, str_repeat( $curChar, $count ) );
591 }
592 $i += $count;
593 } elseif ( $found == 'close' ) {
594 $piece = $stack->top;
595 # lets check if there are enough characters for closing brace
596 $maxCount = $piece->count;
597 if ( $piece->close === '}-' && $curChar === '}' ) {
598 $maxCount--; # don't try to match closing '-' as a '}'
599 }
600 $curLen = strlen( $curChar );
601 $count = ( $curLen > 1 ) ? $curLen :
602 strspn( $text, $curChar, $i, $maxCount );
603
604 # check for maximum matching characters (if there are 5 closing
605 # characters, we will probably need only 3 - depending on the rules)
606 $rule = $this->rules[$piece->open];
607 if ( $piece->close === '}-' && $piece->count > 2 ) {
608 # tweak for -{..{{ }}..}-
609 $rule = $this->rules['{'];
610 }
611 if ( $count > $rule['max'] ) {
612 # The specified maximum exists in the callback array, unless the caller
613 # has made an error
614 $matchingCount = $rule['max'];
615 } else {
616 # Count is less than the maximum
617 # Skip any gaps in the callback array to find the true largest match
618 # Need to use array_key_exists not isset because the callback can be null
619 $matchingCount = $count;
620 while ( $matchingCount > 0 && !array_key_exists( $matchingCount, $rule['names'] ) ) {
621 --$matchingCount;
622 }
623 }
624
625 if ( $matchingCount <= 0 ) {
626 # No matching element found in callback array
627 # Output a literal closing brace and continue
628 $endText = substr( $text, $i, $count );
629 self::addLiteral( $accum, $endText );
630 $i += $count;
631 continue;
632 }
633 $name = $rule['names'][$matchingCount];
634 if ( $name === null ) {
635 // No element, just literal text
636 $endText = substr( $text, $i, $matchingCount );
637 $element = $piece->breakSyntax( $matchingCount );
638 self::addLiteral( $element, $endText );
639 } else {
640 # Create XML element
641 $parts = $piece->parts;
642 $titleAccum = $parts[0]->out;
643 unset( $parts[0] );
644
645 $children = [];
646
647 # The invocation is at the start of the line if lineStart is set in
648 # the stack, and all opening brackets are used up.
649 if ( $maxCount == $matchingCount && !empty( $piece->lineStart ) ) {
650 $children[] = [ '@lineStart', [ 1 ] ];
651 }
652 $titleNode = [ 'title', $titleAccum ];
653 $children[] = $titleNode;
654 $argIndex = 1;
655 foreach ( $parts as $part ) {
656 if ( isset( $part->eqpos ) ) {
657 $equalsNode = $part->out[$part->eqpos];
658 $nameNode = [ 'name', array_slice( $part->out, 0, $part->eqpos ) ];
659 $valueNode = [ 'value', array_slice( $part->out, $part->eqpos + 1 ) ];
660 $partNode = [ 'part', [ $nameNode, $equalsNode, $valueNode ] ];
661 $children[] = $partNode;
662 } else {
663 $nameNode = [ 'name', [ [ '@index', [ $argIndex++ ] ] ] ];
664 $valueNode = [ 'value', $part->out ];
665 $partNode = [ 'part', [ $nameNode, $valueNode ] ];
666 $children[] = $partNode;
667 }
668 }
669 $element = [ [ $name, $children ] ];
670 }
671
672 # Advance input pointer
673 $i += $matchingCount;
674
675 # Unwind the stack
676 $stack->pop();
677 $accum =& $stack->getAccum();
678
679 # Re-add the old stack element if it still has unmatched opening characters remaining
680 if ( $matchingCount < $piece->count ) {
681 $piece->parts = [ new PPDPart_Hash ];
682 $piece->count -= $matchingCount;
683 # do we still qualify for any callback with remaining count?
684 $min = $this->rules[$piece->open]['min'];
685 if ( $piece->count >= $min ) {
686 $stack->push( $piece );
687 $accum =& $stack->getAccum();
688 } else {
689 $s = substr( $piece->open, 0, -1 );
690 $s .= str_repeat(
691 substr( $piece->open, -1 ),
692 $piece->count - strlen( $s )
693 );
694 self::addLiteral( $accum, $s );
695 }
696 }
697
698 extract( $stack->getFlags() );
699
700 # Add XML element to the enclosing accumulator
701 array_splice( $accum, count( $accum ), 0, $element );
702 } elseif ( $found == 'pipe' ) {
703 $findEquals = true; // shortcut for getFlags()
704 $stack->addPart();
705 $accum =& $stack->getAccum();
706 ++$i;
707 } elseif ( $found == 'equals' ) {
708 $findEquals = false; // shortcut for getFlags()
709 $accum[] = [ 'equals', [ '=' ] ];
710 $stack->getCurrentPart()->eqpos = count( $accum ) - 1;
711 ++$i;
712 } elseif ( $found == 'dash' ) {
713 self::addLiteral( $accum, '-' );
714 ++$i;
715 }
716 }
717
718 # Output any remaining unclosed brackets
719 foreach ( $stack->stack as $piece ) {
720 array_splice( $stack->rootAccum, count( $stack->rootAccum ), 0, $piece->breakSyntax() );
721 }
722
723 # Enable top-level headings
724 foreach ( $stack->rootAccum as &$node ) {
725 if ( is_array( $node ) && $node[PPNode_Hash_Tree::NAME] === 'possible-h' ) {
726 $node[PPNode_Hash_Tree::NAME] = 'h';
727 }
728 }
729
730 $rootStore = [ [ 'root', $stack->rootAccum ] ];
731 $rootNode = new PPNode_Hash_Tree( $rootStore, 0 );
732
733 // Cache
734 $tree = json_encode( $rootStore, JSON_UNESCAPED_SLASHES | JSON_UNESCAPED_UNICODE );
735 if ( $tree !== false ) {
736 $this->cacheSetTree( $text, $flags, $tree );
737 }
738
739 return $rootNode;
740 }
741
742 private static function addLiteral( array &$accum, $text ) {
743 $n = count( $accum );
744 if ( $n && is_string( $accum[$n - 1] ) ) {
745 $accum[$n - 1] .= $text;
746 } else {
747 $accum[] = $text;
748 }
749 }
750 }
751
752 /**
753 * Stack class to help Preprocessor::preprocessToObj()
754 * @ingroup Parser
755 */
756 // @codingStandardsIgnoreStart Squiz.Classes.ValidClassName.NotCamelCaps
757 class PPDStack_Hash extends PPDStack {
758 // @codingStandardsIgnoreEnd
759
760 public function __construct() {
761 $this->elementClass = 'PPDStackElement_Hash';
762 parent::__construct();
763 $this->rootAccum = [];
764 }
765 }
766
767 /**
768 * @ingroup Parser
769 */
770 // @codingStandardsIgnoreStart Squiz.Classes.ValidClassName.NotCamelCaps
771 class PPDStackElement_Hash extends PPDStackElement {
772 // @codingStandardsIgnoreEnd
773
774 public function __construct( $data = [] ) {
775 $this->partClass = 'PPDPart_Hash';
776 parent::__construct( $data );
777 }
778
779 /**
780 * Get the accumulator that would result if the close is not found.
781 *
782 * @param int|bool $openingCount
783 * @return array
784 */
785 public function breakSyntax( $openingCount = false ) {
786 if ( $this->open == "\n" ) {
787 $accum = $this->parts[0]->out;
788 } else {
789 if ( $openingCount === false ) {
790 $openingCount = $this->count;
791 }
792 $s = substr( $this->open, 0, -1 );
793 $s .= str_repeat(
794 substr( $this->open, -1 ),
795 $openingCount - strlen( $s )
796 );
797 $accum = [ $s ];
798 $lastIndex = 0;
799 $first = true;
800 foreach ( $this->parts as $part ) {
801 if ( $first ) {
802 $first = false;
803 } elseif ( is_string( $accum[$lastIndex] ) ) {
804 $accum[$lastIndex] .= '|';
805 } else {
806 $accum[++$lastIndex] = '|';
807 }
808 foreach ( $part->out as $node ) {
809 if ( is_string( $node ) && is_string( $accum[$lastIndex] ) ) {
810 $accum[$lastIndex] .= $node;
811 } else {
812 $accum[++$lastIndex] = $node;
813 }
814 }
815 }
816 }
817 return $accum;
818 }
819 }
820
821 /**
822 * @ingroup Parser
823 */
824 // @codingStandardsIgnoreStart Squiz.Classes.ValidClassName.NotCamelCaps
825 class PPDPart_Hash extends PPDPart {
826 // @codingStandardsIgnoreEnd
827
828 public function __construct( $out = '' ) {
829 if ( $out !== '' ) {
830 $accum = [ $out ];
831 } else {
832 $accum = [];
833 }
834 parent::__construct( $accum );
835 }
836 }
837
838 /**
839 * An expansion frame, used as a context to expand the result of preprocessToObj()
840 * @ingroup Parser
841 */
842 // @codingStandardsIgnoreStart Squiz.Classes.ValidClassName.NotCamelCaps
843 class PPFrame_Hash implements PPFrame {
844 // @codingStandardsIgnoreEnd
845
846 /**
847 * @var Parser
848 */
849 public $parser;
850
851 /**
852 * @var Preprocessor
853 */
854 public $preprocessor;
855
856 /**
857 * @var Title
858 */
859 public $title;
860 public $titleCache;
861
862 /**
863 * Hashtable listing templates which are disallowed for expansion in this frame,
864 * having been encountered previously in parent frames.
865 */
866 public $loopCheckHash;
867
868 /**
869 * Recursion depth of this frame, top = 0
870 * Note that this is NOT the same as expansion depth in expand()
871 */
872 public $depth;
873
874 private $volatile = false;
875 private $ttl = null;
876
877 /**
878 * @var array
879 */
880 protected $childExpansionCache;
881
882 /**
883 * Construct a new preprocessor frame.
884 * @param Preprocessor $preprocessor The parent preprocessor
885 */
886 public function __construct( $preprocessor ) {
887 $this->preprocessor = $preprocessor;
888 $this->parser = $preprocessor->parser;
889 $this->title = $this->parser->mTitle;
890 $this->titleCache = [ $this->title ? $this->title->getPrefixedDBkey() : false ];
891 $this->loopCheckHash = [];
892 $this->depth = 0;
893 $this->childExpansionCache = [];
894 }
895
896 /**
897 * Create a new child frame
898 * $args is optionally a multi-root PPNode or array containing the template arguments
899 *
900 * @param array|bool|PPNode_Hash_Array $args
901 * @param Title|bool $title
902 * @param int $indexOffset
903 * @throws MWException
904 * @return PPTemplateFrame_Hash
905 */
906 public function newChild( $args = false, $title = false, $indexOffset = 0 ) {
907 $namedArgs = [];
908 $numberedArgs = [];
909 if ( $title === false ) {
910 $title = $this->title;
911 }
912 if ( $args !== false ) {
913 if ( $args instanceof PPNode_Hash_Array ) {
914 $args = $args->value;
915 } elseif ( !is_array( $args ) ) {
916 throw new MWException( __METHOD__ . ': $args must be array or PPNode_Hash_Array' );
917 }
918 foreach ( $args as $arg ) {
919 $bits = $arg->splitArg();
920 if ( $bits['index'] !== '' ) {
921 // Numbered parameter
922 $index = $bits['index'] - $indexOffset;
923 if ( isset( $namedArgs[$index] ) || isset( $numberedArgs[$index] ) ) {
924 $this->parser->getOutput()->addWarning( wfMessage( 'duplicate-args-warning',
925 wfEscapeWikiText( $this->title ),
926 wfEscapeWikiText( $title ),
927 wfEscapeWikiText( $index ) )->text() );
928 $this->parser->addTrackingCategory( 'duplicate-args-category' );
929 }
930 $numberedArgs[$index] = $bits['value'];
931 unset( $namedArgs[$index] );
932 } else {
933 // Named parameter
934 $name = trim( $this->expand( $bits['name'], PPFrame::STRIP_COMMENTS ) );
935 if ( isset( $namedArgs[$name] ) || isset( $numberedArgs[$name] ) ) {
936 $this->parser->getOutput()->addWarning( wfMessage( 'duplicate-args-warning',
937 wfEscapeWikiText( $this->title ),
938 wfEscapeWikiText( $title ),
939 wfEscapeWikiText( $name ) )->text() );
940 $this->parser->addTrackingCategory( 'duplicate-args-category' );
941 }
942 $namedArgs[$name] = $bits['value'];
943 unset( $numberedArgs[$name] );
944 }
945 }
946 }
947 return new PPTemplateFrame_Hash( $this->preprocessor, $this, $numberedArgs, $namedArgs, $title );
948 }
949
950 /**
951 * @throws MWException
952 * @param string|int $key
953 * @param string|PPNode $root
954 * @param int $flags
955 * @return string
956 */
957 public function cachedExpand( $key, $root, $flags = 0 ) {
958 // we don't have a parent, so we don't have a cache
959 return $this->expand( $root, $flags );
960 }
961
962 /**
963 * @throws MWException
964 * @param string|PPNode $root
965 * @param int $flags
966 * @return string
967 */
968 public function expand( $root, $flags = 0 ) {
969 static $expansionDepth = 0;
970 if ( is_string( $root ) ) {
971 return $root;
972 }
973
974 if ( ++$this->parser->mPPNodeCount > $this->parser->mOptions->getMaxPPNodeCount() ) {
975 $this->parser->limitationWarn( 'node-count-exceeded',
976 $this->parser->mPPNodeCount,
977 $this->parser->mOptions->getMaxPPNodeCount()
978 );
979 return '<span class="error">Node-count limit exceeded</span>';
980 }
981 if ( $expansionDepth > $this->parser->mOptions->getMaxPPExpandDepth() ) {
982 $this->parser->limitationWarn( 'expansion-depth-exceeded',
983 $expansionDepth,
984 $this->parser->mOptions->getMaxPPExpandDepth()
985 );
986 return '<span class="error">Expansion depth limit exceeded</span>';
987 }
988 ++$expansionDepth;
989 if ( $expansionDepth > $this->parser->mHighestExpansionDepth ) {
990 $this->parser->mHighestExpansionDepth = $expansionDepth;
991 }
992
993 $outStack = [ '', '' ];
994 $iteratorStack = [ false, $root ];
995 $indexStack = [ 0, 0 ];
996
997 while ( count( $iteratorStack ) > 1 ) {
998 $level = count( $outStack ) - 1;
999 $iteratorNode =& $iteratorStack[$level];
1000 $out =& $outStack[$level];
1001 $index =& $indexStack[$level];
1002
1003 if ( is_array( $iteratorNode ) ) {
1004 if ( $index >= count( $iteratorNode ) ) {
1005 // All done with this iterator
1006 $iteratorStack[$level] = false;
1007 $contextNode = false;
1008 } else {
1009 $contextNode = $iteratorNode[$index];
1010 $index++;
1011 }
1012 } elseif ( $iteratorNode instanceof PPNode_Hash_Array ) {
1013 if ( $index >= $iteratorNode->getLength() ) {
1014 // All done with this iterator
1015 $iteratorStack[$level] = false;
1016 $contextNode = false;
1017 } else {
1018 $contextNode = $iteratorNode->item( $index );
1019 $index++;
1020 }
1021 } else {
1022 // Copy to $contextNode and then delete from iterator stack,
1023 // because this is not an iterator but we do have to execute it once
1024 $contextNode = $iteratorStack[$level];
1025 $iteratorStack[$level] = false;
1026 }
1027
1028 $newIterator = false;
1029 $contextName = false;
1030 $contextChildren = false;
1031
1032 if ( $contextNode === false ) {
1033 // nothing to do
1034 } elseif ( is_string( $contextNode ) ) {
1035 $out .= $contextNode;
1036 } elseif ( $contextNode instanceof PPNode_Hash_Array ) {
1037 $newIterator = $contextNode;
1038 } elseif ( $contextNode instanceof PPNode_Hash_Attr ) {
1039 // No output
1040 } elseif ( $contextNode instanceof PPNode_Hash_Text ) {
1041 $out .= $contextNode->value;
1042 } elseif ( $contextNode instanceof PPNode_Hash_Tree ) {
1043 $contextName = $contextNode->name;
1044 $contextChildren = $contextNode->getRawChildren();
1045 } elseif ( is_array( $contextNode ) ) {
1046 // Node descriptor array
1047 if ( count( $contextNode ) !== 2 ) {
1048 throw new MWException( __METHOD__.
1049 ': found an array where a node descriptor should be' );
1050 }
1051 list( $contextName, $contextChildren ) = $contextNode;
1052 } else {
1053 throw new MWException( __METHOD__ . ': Invalid parameter type' );
1054 }
1055
1056 // Handle node descriptor array or tree object
1057 if ( $contextName === false ) {
1058 // Not a node, already handled above
1059 } elseif ( $contextName[0] === '@' ) {
1060 // Attribute: no output
1061 } elseif ( $contextName === 'template' ) {
1062 # Double-brace expansion
1063 $bits = PPNode_Hash_Tree::splitRawTemplate( $contextChildren );
1064 if ( $flags & PPFrame::NO_TEMPLATES ) {
1065 $newIterator = $this->virtualBracketedImplode(
1066 '{{', '|', '}}',
1067 $bits['title'],
1068 $bits['parts']
1069 );
1070 } else {
1071 $ret = $this->parser->braceSubstitution( $bits, $this );
1072 if ( isset( $ret['object'] ) ) {
1073 $newIterator = $ret['object'];
1074 } else {
1075 $out .= $ret['text'];
1076 }
1077 }
1078 } elseif ( $contextName === 'tplarg' ) {
1079 # Triple-brace expansion
1080 $bits = PPNode_Hash_Tree::splitRawTemplate( $contextChildren );
1081 if ( $flags & PPFrame::NO_ARGS ) {
1082 $newIterator = $this->virtualBracketedImplode(
1083 '{{{', '|', '}}}',
1084 $bits['title'],
1085 $bits['parts']
1086 );
1087 } else {
1088 $ret = $this->parser->argSubstitution( $bits, $this );
1089 if ( isset( $ret['object'] ) ) {
1090 $newIterator = $ret['object'];
1091 } else {
1092 $out .= $ret['text'];
1093 }
1094 }
1095 } elseif ( $contextName === 'comment' ) {
1096 # HTML-style comment
1097 # Remove it in HTML, pre+remove and STRIP_COMMENTS modes
1098 # Not in RECOVER_COMMENTS mode (msgnw) though.
1099 if ( ( $this->parser->ot['html']
1100 || ( $this->parser->ot['pre'] && $this->parser->mOptions->getRemoveComments() )
1101 || ( $flags & PPFrame::STRIP_COMMENTS )
1102 ) && !( $flags & PPFrame::RECOVER_COMMENTS )
1103 ) {
1104 $out .= '';
1105 } elseif ( $this->parser->ot['wiki'] && !( $flags & PPFrame::RECOVER_COMMENTS ) ) {
1106 # Add a strip marker in PST mode so that pstPass2() can
1107 # run some old-fashioned regexes on the result.
1108 # Not in RECOVER_COMMENTS mode (extractSections) though.
1109 $out .= $this->parser->insertStripItem( $contextChildren[0] );
1110 } else {
1111 # Recover the literal comment in RECOVER_COMMENTS and pre+no-remove
1112 $out .= $contextChildren[0];
1113 }
1114 } elseif ( $contextName === 'ignore' ) {
1115 # Output suppression used by <includeonly> etc.
1116 # OT_WIKI will only respect <ignore> in substed templates.
1117 # The other output types respect it unless NO_IGNORE is set.
1118 # extractSections() sets NO_IGNORE and so never respects it.
1119 if ( ( !isset( $this->parent ) && $this->parser->ot['wiki'] )
1120 || ( $flags & PPFrame::NO_IGNORE )
1121 ) {
1122 $out .= $contextChildren[0];
1123 } else {
1124 // $out .= '';
1125 }
1126 } elseif ( $contextName === 'ext' ) {
1127 # Extension tag
1128 $bits = PPNode_Hash_Tree::splitRawExt( $contextChildren ) +
1129 [ 'attr' => null, 'inner' => null, 'close' => null ];
1130 if ( $flags & PPFrame::NO_TAGS ) {
1131 $s = '<' . $bits['name']->getFirstChild()->value;
1132 if ( $bits['attr'] ) {
1133 $s .= $bits['attr']->getFirstChild()->value;
1134 }
1135 if ( $bits['inner'] ) {
1136 $s .= '>' . $bits['inner']->getFirstChild()->value;
1137 if ( $bits['close'] ) {
1138 $s .= $bits['close']->getFirstChild()->value;
1139 }
1140 } else {
1141 $s .= '/>';
1142 }
1143 $out .= $s;
1144 } else {
1145 $out .= $this->parser->extensionSubstitution( $bits, $this );
1146 }
1147 } elseif ( $contextName === 'h' ) {
1148 # Heading
1149 if ( $this->parser->ot['html'] ) {
1150 # Expand immediately and insert heading index marker
1151 $s = $this->expand( $contextChildren, $flags );
1152 $bits = PPNode_Hash_Tree::splitRawHeading( $contextChildren );
1153 $titleText = $this->title->getPrefixedDBkey();
1154 $this->parser->mHeadings[] = [ $titleText, $bits['i'] ];
1155 $serial = count( $this->parser->mHeadings ) - 1;
1156 $marker = Parser::MARKER_PREFIX . "-h-$serial-" . Parser::MARKER_SUFFIX;
1157 $s = substr( $s, 0, $bits['level'] ) . $marker . substr( $s, $bits['level'] );
1158 $this->parser->mStripState->addGeneral( $marker, '' );
1159 $out .= $s;
1160 } else {
1161 # Expand in virtual stack
1162 $newIterator = $contextChildren;
1163 }
1164 } else {
1165 # Generic recursive expansion
1166 $newIterator = $contextChildren;
1167 }
1168
1169 if ( $newIterator !== false ) {
1170 $outStack[] = '';
1171 $iteratorStack[] = $newIterator;
1172 $indexStack[] = 0;
1173 } elseif ( $iteratorStack[$level] === false ) {
1174 // Return accumulated value to parent
1175 // With tail recursion
1176 while ( $iteratorStack[$level] === false && $level > 0 ) {
1177 $outStack[$level - 1] .= $out;
1178 array_pop( $outStack );
1179 array_pop( $iteratorStack );
1180 array_pop( $indexStack );
1181 $level--;
1182 }
1183 }
1184 }
1185 --$expansionDepth;
1186 return $outStack[0];
1187 }
1188
1189 /**
1190 * @param string $sep
1191 * @param int $flags
1192 * @param string|PPNode $args,...
1193 * @return string
1194 */
1195 public function implodeWithFlags( $sep, $flags /*, ... */ ) {
1196 $args = array_slice( func_get_args(), 2 );
1197
1198 $first = true;
1199 $s = '';
1200 foreach ( $args as $root ) {
1201 if ( $root instanceof PPNode_Hash_Array ) {
1202 $root = $root->value;
1203 }
1204 if ( !is_array( $root ) ) {
1205 $root = [ $root ];
1206 }
1207 foreach ( $root as $node ) {
1208 if ( $first ) {
1209 $first = false;
1210 } else {
1211 $s .= $sep;
1212 }
1213 $s .= $this->expand( $node, $flags );
1214 }
1215 }
1216 return $s;
1217 }
1218
1219 /**
1220 * Implode with no flags specified
1221 * This previously called implodeWithFlags but has now been inlined to reduce stack depth
1222 * @param string $sep
1223 * @param string|PPNode $args,...
1224 * @return string
1225 */
1226 public function implode( $sep /*, ... */ ) {
1227 $args = array_slice( func_get_args(), 1 );
1228
1229 $first = true;
1230 $s = '';
1231 foreach ( $args as $root ) {
1232 if ( $root instanceof PPNode_Hash_Array ) {
1233 $root = $root->value;
1234 }
1235 if ( !is_array( $root ) ) {
1236 $root = [ $root ];
1237 }
1238 foreach ( $root as $node ) {
1239 if ( $first ) {
1240 $first = false;
1241 } else {
1242 $s .= $sep;
1243 }
1244 $s .= $this->expand( $node );
1245 }
1246 }
1247 return $s;
1248 }
1249
1250 /**
1251 * Makes an object that, when expand()ed, will be the same as one obtained
1252 * with implode()
1253 *
1254 * @param string $sep
1255 * @param string|PPNode $args,...
1256 * @return PPNode_Hash_Array
1257 */
1258 public function virtualImplode( $sep /*, ... */ ) {
1259 $args = array_slice( func_get_args(), 1 );
1260 $out = [];
1261 $first = true;
1262
1263 foreach ( $args as $root ) {
1264 if ( $root instanceof PPNode_Hash_Array ) {
1265 $root = $root->value;
1266 }
1267 if ( !is_array( $root ) ) {
1268 $root = [ $root ];
1269 }
1270 foreach ( $root as $node ) {
1271 if ( $first ) {
1272 $first = false;
1273 } else {
1274 $out[] = $sep;
1275 }
1276 $out[] = $node;
1277 }
1278 }
1279 return new PPNode_Hash_Array( $out );
1280 }
1281
1282 /**
1283 * Virtual implode with brackets
1284 *
1285 * @param string $start
1286 * @param string $sep
1287 * @param string $end
1288 * @param string|PPNode $args,...
1289 * @return PPNode_Hash_Array
1290 */
1291 public function virtualBracketedImplode( $start, $sep, $end /*, ... */ ) {
1292 $args = array_slice( func_get_args(), 3 );
1293 $out = [ $start ];
1294 $first = true;
1295
1296 foreach ( $args as $root ) {
1297 if ( $root instanceof PPNode_Hash_Array ) {
1298 $root = $root->value;
1299 }
1300 if ( !is_array( $root ) ) {
1301 $root = [ $root ];
1302 }
1303 foreach ( $root as $node ) {
1304 if ( $first ) {
1305 $first = false;
1306 } else {
1307 $out[] = $sep;
1308 }
1309 $out[] = $node;
1310 }
1311 }
1312 $out[] = $end;
1313 return new PPNode_Hash_Array( $out );
1314 }
1315
1316 public function __toString() {
1317 return 'frame{}';
1318 }
1319
1320 /**
1321 * @param bool $level
1322 * @return array|bool|string
1323 */
1324 public function getPDBK( $level = false ) {
1325 if ( $level === false ) {
1326 return $this->title->getPrefixedDBkey();
1327 } else {
1328 return isset( $this->titleCache[$level] ) ? $this->titleCache[$level] : false;
1329 }
1330 }
1331
1332 /**
1333 * @return array
1334 */
1335 public function getArguments() {
1336 return [];
1337 }
1338
1339 /**
1340 * @return array
1341 */
1342 public function getNumberedArguments() {
1343 return [];
1344 }
1345
1346 /**
1347 * @return array
1348 */
1349 public function getNamedArguments() {
1350 return [];
1351 }
1352
1353 /**
1354 * Returns true if there are no arguments in this frame
1355 *
1356 * @return bool
1357 */
1358 public function isEmpty() {
1359 return true;
1360 }
1361
1362 /**
1363 * @param int|string $name
1364 * @return bool Always false in this implementation.
1365 */
1366 public function getArgument( $name ) {
1367 return false;
1368 }
1369
1370 /**
1371 * Returns true if the infinite loop check is OK, false if a loop is detected
1372 *
1373 * @param Title $title
1374 *
1375 * @return bool
1376 */
1377 public function loopCheck( $title ) {
1378 return !isset( $this->loopCheckHash[$title->getPrefixedDBkey()] );
1379 }
1380
1381 /**
1382 * Return true if the frame is a template frame
1383 *
1384 * @return bool
1385 */
1386 public function isTemplate() {
1387 return false;
1388 }
1389
1390 /**
1391 * Get a title of frame
1392 *
1393 * @return Title
1394 */
1395 public function getTitle() {
1396 return $this->title;
1397 }
1398
1399 /**
1400 * Set the volatile flag
1401 *
1402 * @param bool $flag
1403 */
1404 public function setVolatile( $flag = true ) {
1405 $this->volatile = $flag;
1406 }
1407
1408 /**
1409 * Get the volatile flag
1410 *
1411 * @return bool
1412 */
1413 public function isVolatile() {
1414 return $this->volatile;
1415 }
1416
1417 /**
1418 * Set the TTL
1419 *
1420 * @param int $ttl
1421 */
1422 public function setTTL( $ttl ) {
1423 if ( $ttl !== null && ( $this->ttl === null || $ttl < $this->ttl ) ) {
1424 $this->ttl = $ttl;
1425 }
1426 }
1427
1428 /**
1429 * Get the TTL
1430 *
1431 * @return int|null
1432 */
1433 public function getTTL() {
1434 return $this->ttl;
1435 }
1436 }
1437
1438 /**
1439 * Expansion frame with template arguments
1440 * @ingroup Parser
1441 */
1442 // @codingStandardsIgnoreStart Squiz.Classes.ValidClassName.NotCamelCaps
1443 class PPTemplateFrame_Hash extends PPFrame_Hash {
1444 // @codingStandardsIgnoreEnd
1445
1446 public $numberedArgs, $namedArgs, $parent;
1447 public $numberedExpansionCache, $namedExpansionCache;
1448
1449 /**
1450 * @param Preprocessor $preprocessor
1451 * @param bool|PPFrame $parent
1452 * @param array $numberedArgs
1453 * @param array $namedArgs
1454 * @param bool|Title $title
1455 */
1456 public function __construct( $preprocessor, $parent = false, $numberedArgs = [],
1457 $namedArgs = [], $title = false
1458 ) {
1459 parent::__construct( $preprocessor );
1460
1461 $this->parent = $parent;
1462 $this->numberedArgs = $numberedArgs;
1463 $this->namedArgs = $namedArgs;
1464 $this->title = $title;
1465 $pdbk = $title ? $title->getPrefixedDBkey() : false;
1466 $this->titleCache = $parent->titleCache;
1467 $this->titleCache[] = $pdbk;
1468 $this->loopCheckHash = /*clone*/ $parent->loopCheckHash;
1469 if ( $pdbk !== false ) {
1470 $this->loopCheckHash[$pdbk] = true;
1471 }
1472 $this->depth = $parent->depth + 1;
1473 $this->numberedExpansionCache = $this->namedExpansionCache = [];
1474 }
1475
1476 public function __toString() {
1477 $s = 'tplframe{';
1478 $first = true;
1479 $args = $this->numberedArgs + $this->namedArgs;
1480 foreach ( $args as $name => $value ) {
1481 if ( $first ) {
1482 $first = false;
1483 } else {
1484 $s .= ', ';
1485 }
1486 $s .= "\"$name\":\"" .
1487 str_replace( '"', '\\"', $value->__toString() ) . '"';
1488 }
1489 $s .= '}';
1490 return $s;
1491 }
1492
1493 /**
1494 * @throws MWException
1495 * @param string|int $key
1496 * @param string|PPNode $root
1497 * @param int $flags
1498 * @return string
1499 */
1500 public function cachedExpand( $key, $root, $flags = 0 ) {
1501 if ( isset( $this->parent->childExpansionCache[$key] ) ) {
1502 return $this->parent->childExpansionCache[$key];
1503 }
1504 $retval = $this->expand( $root, $flags );
1505 if ( !$this->isVolatile() ) {
1506 $this->parent->childExpansionCache[$key] = $retval;
1507 }
1508 return $retval;
1509 }
1510
1511 /**
1512 * Returns true if there are no arguments in this frame
1513 *
1514 * @return bool
1515 */
1516 public function isEmpty() {
1517 return !count( $this->numberedArgs ) && !count( $this->namedArgs );
1518 }
1519
1520 /**
1521 * @return array
1522 */
1523 public function getArguments() {
1524 $arguments = [];
1525 foreach ( array_merge(
1526 array_keys( $this->numberedArgs ),
1527 array_keys( $this->namedArgs ) ) as $key ) {
1528 $arguments[$key] = $this->getArgument( $key );
1529 }
1530 return $arguments;
1531 }
1532
1533 /**
1534 * @return array
1535 */
1536 public function getNumberedArguments() {
1537 $arguments = [];
1538 foreach ( array_keys( $this->numberedArgs ) as $key ) {
1539 $arguments[$key] = $this->getArgument( $key );
1540 }
1541 return $arguments;
1542 }
1543
1544 /**
1545 * @return array
1546 */
1547 public function getNamedArguments() {
1548 $arguments = [];
1549 foreach ( array_keys( $this->namedArgs ) as $key ) {
1550 $arguments[$key] = $this->getArgument( $key );
1551 }
1552 return $arguments;
1553 }
1554
1555 /**
1556 * @param int $index
1557 * @return string|bool
1558 */
1559 public function getNumberedArgument( $index ) {
1560 if ( !isset( $this->numberedArgs[$index] ) ) {
1561 return false;
1562 }
1563 if ( !isset( $this->numberedExpansionCache[$index] ) ) {
1564 # No trimming for unnamed arguments
1565 $this->numberedExpansionCache[$index] = $this->parent->expand(
1566 $this->numberedArgs[$index],
1567 PPFrame::STRIP_COMMENTS
1568 );
1569 }
1570 return $this->numberedExpansionCache[$index];
1571 }
1572
1573 /**
1574 * @param string $name
1575 * @return string|bool
1576 */
1577 public function getNamedArgument( $name ) {
1578 if ( !isset( $this->namedArgs[$name] ) ) {
1579 return false;
1580 }
1581 if ( !isset( $this->namedExpansionCache[$name] ) ) {
1582 # Trim named arguments post-expand, for backwards compatibility
1583 $this->namedExpansionCache[$name] = trim(
1584 $this->parent->expand( $this->namedArgs[$name], PPFrame::STRIP_COMMENTS ) );
1585 }
1586 return $this->namedExpansionCache[$name];
1587 }
1588
1589 /**
1590 * @param int|string $name
1591 * @return string|bool
1592 */
1593 public function getArgument( $name ) {
1594 $text = $this->getNumberedArgument( $name );
1595 if ( $text === false ) {
1596 $text = $this->getNamedArgument( $name );
1597 }
1598 return $text;
1599 }
1600
1601 /**
1602 * Return true if the frame is a template frame
1603 *
1604 * @return bool
1605 */
1606 public function isTemplate() {
1607 return true;
1608 }
1609
1610 public function setVolatile( $flag = true ) {
1611 parent::setVolatile( $flag );
1612 $this->parent->setVolatile( $flag );
1613 }
1614
1615 public function setTTL( $ttl ) {
1616 parent::setTTL( $ttl );
1617 $this->parent->setTTL( $ttl );
1618 }
1619 }
1620
1621 /**
1622 * Expansion frame with custom arguments
1623 * @ingroup Parser
1624 */
1625 // @codingStandardsIgnoreStart Squiz.Classes.ValidClassName.NotCamelCaps
1626 class PPCustomFrame_Hash extends PPFrame_Hash {
1627 // @codingStandardsIgnoreEnd
1628
1629 public $args;
1630
1631 public function __construct( $preprocessor, $args ) {
1632 parent::__construct( $preprocessor );
1633 $this->args = $args;
1634 }
1635
1636 public function __toString() {
1637 $s = 'cstmframe{';
1638 $first = true;
1639 foreach ( $this->args as $name => $value ) {
1640 if ( $first ) {
1641 $first = false;
1642 } else {
1643 $s .= ', ';
1644 }
1645 $s .= "\"$name\":\"" .
1646 str_replace( '"', '\\"', $value->__toString() ) . '"';
1647 }
1648 $s .= '}';
1649 return $s;
1650 }
1651
1652 /**
1653 * @return bool
1654 */
1655 public function isEmpty() {
1656 return !count( $this->args );
1657 }
1658
1659 /**
1660 * @param int|string $index
1661 * @return string|bool
1662 */
1663 public function getArgument( $index ) {
1664 if ( !isset( $this->args[$index] ) ) {
1665 return false;
1666 }
1667 return $this->args[$index];
1668 }
1669
1670 public function getArguments() {
1671 return $this->args;
1672 }
1673 }
1674
1675 /**
1676 * @ingroup Parser
1677 */
1678 // @codingStandardsIgnoreStart Squiz.Classes.ValidClassName.NotCamelCaps
1679 class PPNode_Hash_Tree implements PPNode {
1680 // @codingStandardsIgnoreEnd
1681
1682 public $name;
1683
1684 /**
1685 * The store array for children of this node. It is "raw" in the sense that
1686 * nodes are two-element arrays ("descriptors") rather than PPNode_Hash_*
1687 * objects.
1688 */
1689 private $rawChildren;
1690
1691 /**
1692 * The store array for the siblings of this node, including this node itself.
1693 */
1694 private $store;
1695
1696 /**
1697 * The index into $this->store which contains the descriptor of this node.
1698 */
1699 private $index;
1700
1701 /**
1702 * The offset of the name within descriptors, used in some places for
1703 * readability.
1704 */
1705 const NAME = 0;
1706
1707 /**
1708 * The offset of the child list within descriptors, used in some places for
1709 * readability.
1710 */
1711 const CHILDREN = 1;
1712
1713 /**
1714 * Construct an object using the data from $store[$index]. The rest of the
1715 * store array can be accessed via getNextSibling().
1716 *
1717 * @param array $store
1718 * @param int $index
1719 */
1720 public function __construct( array $store, $index ) {
1721 $this->store = $store;
1722 $this->index = $index;
1723 list( $this->name, $this->rawChildren ) = $this->store[$index];
1724 }
1725
1726 /**
1727 * Construct an appropriate PPNode_Hash_* object with a class that depends
1728 * on what is at the relevant store index.
1729 *
1730 * @param array $store
1731 * @param int $index
1732 * @return PPNode_Hash_Tree|PPNode_Hash_Attr|PPNode_Hash_Text
1733 */
1734 public static function factory( array $store, $index ) {
1735 if ( !isset( $store[$index] ) ) {
1736 return false;
1737 }
1738
1739 $descriptor = $store[$index];
1740 if ( is_string( $descriptor ) ) {
1741 $class = 'PPNode_Hash_Text';
1742 } elseif ( is_array( $descriptor ) ) {
1743 if ( $descriptor[self::NAME][0] === '@' ) {
1744 $class = 'PPNode_Hash_Attr';
1745 } else {
1746 $class = 'PPNode_Hash_Tree';
1747 }
1748 } else {
1749 throw new MWException( __METHOD__.': invalid node descriptor' );
1750 }
1751 return new $class( $store, $index );
1752 }
1753
1754 /**
1755 * Convert a node to XML, for debugging
1756 */
1757 public function __toString() {
1758 $inner = '';
1759 $attribs = '';
1760 for ( $node = $this->getFirstChild(); $node; $node = $node->getNextSibling() ) {
1761 if ( $node instanceof PPNode_Hash_Attr ) {
1762 $attribs .= ' ' . $node->name . '="' . htmlspecialchars( $node->value ) . '"';
1763 } else {
1764 $inner .= $node->__toString();
1765 }
1766 }
1767 if ( $inner === '' ) {
1768 return "<{$this->name}$attribs/>";
1769 } else {
1770 return "<{$this->name}$attribs>$inner</{$this->name}>";
1771 }
1772 }
1773
1774 /**
1775 * @return PPNode_Hash_Array
1776 */
1777 public function getChildren() {
1778 $children = [];
1779 foreach ( $this->rawChildren as $i => $child ) {
1780 $children[] = self::factory( $this->rawChildren, $i );
1781 }
1782 return new PPNode_Hash_Array( $children );
1783 }
1784
1785 /**
1786 * Get the first child, or false if there is none. Note that this will
1787 * return a temporary proxy object: different instances will be returned
1788 * if this is called more than once on the same node.
1789 *
1790 * @return PPNode_Hash_Tree|PPNode_Hash_Attr|PPNode_Hash_Text|bool
1791 */
1792 public function getFirstChild() {
1793 if ( !isset( $this->rawChildren[0] ) ) {
1794 return false;
1795 } else {
1796 return self::factory( $this->rawChildren, 0 );
1797 }
1798 }
1799
1800 /**
1801 * Get the next sibling, or false if there is none. Note that this will
1802 * return a temporary proxy object: different instances will be returned
1803 * if this is called more than once on the same node.
1804 *
1805 * @return PPNode_Hash_Tree|PPNode_Hash_Attr|PPNode_Hash_Text|bool
1806 */
1807 public function getNextSibling() {
1808 return self::factory( $this->store, $this->index + 1 );
1809 }
1810
1811 /**
1812 * Get an array of the children with a given node name
1813 *
1814 * @param string $name
1815 * @return PPNode_Hash_Array
1816 */
1817 public function getChildrenOfType( $name ) {
1818 $children = [];
1819 foreach ( $this->rawChildren as $i => $child ) {
1820 if ( is_array( $child ) && $child[self::NAME] === $name ) {
1821 $children[] = self::factory( $this->rawChildren, $i );
1822 }
1823 }
1824 return new PPNode_Hash_Array( $children );
1825 }
1826
1827 /**
1828 * Get the raw child array. For internal use.
1829 * @return array
1830 */
1831 public function getRawChildren() {
1832 return $this->rawChildren;
1833 }
1834
1835 /**
1836 * @return bool
1837 */
1838 public function getLength() {
1839 return false;
1840 }
1841
1842 /**
1843 * @param int $i
1844 * @return bool
1845 */
1846 public function item( $i ) {
1847 return false;
1848 }
1849
1850 /**
1851 * @return string
1852 */
1853 public function getName() {
1854 return $this->name;
1855 }
1856
1857 /**
1858 * Split a "<part>" node into an associative array containing:
1859 * - name PPNode name
1860 * - index String index
1861 * - value PPNode value
1862 *
1863 * @throws MWException
1864 * @return array
1865 */
1866 public function splitArg() {
1867 return self::splitRawArg( $this->rawChildren );
1868 }
1869
1870 /**
1871 * Like splitArg() but for a raw child array. For internal use only.
1872 */
1873 public static function splitRawArg( array $children ) {
1874 $bits = [];
1875 foreach ( $children as $i => $child ) {
1876 if ( !is_array( $child ) ) {
1877 continue;
1878 }
1879 if ( $child[self::NAME] === 'name' ) {
1880 $bits['name'] = new self( $children, $i );
1881 if ( isset( $child[self::CHILDREN][0][self::NAME] )
1882 && $child[self::CHILDREN][0][self::NAME] === '@index'
1883 ) {
1884 $bits['index'] = $child[self::CHILDREN][0][self::CHILDREN][0];
1885 }
1886 } elseif ( $child[self::NAME] === 'value' ) {
1887 $bits['value'] = new self( $children, $i );
1888 }
1889 }
1890
1891 if ( !isset( $bits['name'] ) ) {
1892 throw new MWException( 'Invalid brace node passed to ' . __METHOD__ );
1893 }
1894 if ( !isset( $bits['index'] ) ) {
1895 $bits['index'] = '';
1896 }
1897 return $bits;
1898 }
1899
1900 /**
1901 * Split an "<ext>" node into an associative array containing name, attr, inner and close
1902 * All values in the resulting array are PPNodes. Inner and close are optional.
1903 *
1904 * @throws MWException
1905 * @return array
1906 */
1907 public function splitExt() {
1908 return self::splitRawExt( $this->rawChildren );
1909 }
1910
1911 /**
1912 * Like splitExt() but for a raw child array. For internal use only.
1913 */
1914 public static function splitRawExt( array $children ) {
1915 $bits = [];
1916 foreach ( $children as $i => $child ) {
1917 if ( !is_array( $child ) ) {
1918 continue;
1919 }
1920 switch ( $child[self::NAME] ) {
1921 case 'name':
1922 $bits['name'] = new self( $children, $i );
1923 break;
1924 case 'attr':
1925 $bits['attr'] = new self( $children, $i );
1926 break;
1927 case 'inner':
1928 $bits['inner'] = new self( $children, $i );
1929 break;
1930 case 'close':
1931 $bits['close'] = new self( $children, $i );
1932 break;
1933 }
1934 }
1935 if ( !isset( $bits['name'] ) ) {
1936 throw new MWException( 'Invalid ext node passed to ' . __METHOD__ );
1937 }
1938 return $bits;
1939 }
1940
1941 /**
1942 * Split an "<h>" node
1943 *
1944 * @throws MWException
1945 * @return array
1946 */
1947 public function splitHeading() {
1948 if ( $this->name !== 'h' ) {
1949 throw new MWException( 'Invalid h node passed to ' . __METHOD__ );
1950 }
1951 return self::splitRawHeading( $this->rawChildren );
1952 }
1953
1954 /**
1955 * Like splitHeading() but for a raw child array. For internal use only.
1956 */
1957 public static function splitRawHeading( array $children ) {
1958 $bits = [];
1959 foreach ( $children as $i => $child ) {
1960 if ( !is_array( $child ) ) {
1961 continue;
1962 }
1963 if ( $child[self::NAME] === '@i' ) {
1964 $bits['i'] = $child[self::CHILDREN][0];
1965 } elseif ( $child[self::NAME] === '@level' ) {
1966 $bits['level'] = $child[self::CHILDREN][0];
1967 }
1968 }
1969 if ( !isset( $bits['i'] ) ) {
1970 throw new MWException( 'Invalid h node passed to ' . __METHOD__ );
1971 }
1972 return $bits;
1973 }
1974
1975 /**
1976 * Split a "<template>" or "<tplarg>" node
1977 *
1978 * @throws MWException
1979 * @return array
1980 */
1981 public function splitTemplate() {
1982 return self::splitRawTemplate( $this->rawChildren );
1983 }
1984
1985 /**
1986 * Like splitTemplate() but for a raw child array. For internal use only.
1987 */
1988 public static function splitRawTemplate( array $children ) {
1989 $parts = [];
1990 $bits = [ 'lineStart' => '' ];
1991 foreach ( $children as $i => $child ) {
1992 if ( !is_array( $child ) ) {
1993 continue;
1994 }
1995 switch ( $child[self::NAME] ) {
1996 case 'title':
1997 $bits['title'] = new self( $children, $i );
1998 break;
1999 case 'part':
2000 $parts[] = new self( $children, $i );
2001 break;
2002 case '@lineStart':
2003 $bits['lineStart'] = '1';
2004 break;
2005 }
2006 }
2007 if ( !isset( $bits['title'] ) ) {
2008 throw new MWException( 'Invalid node passed to ' . __METHOD__ );
2009 }
2010 $bits['parts'] = new PPNode_Hash_Array( $parts );
2011 return $bits;
2012 }
2013 }
2014
2015 /**
2016 * @ingroup Parser
2017 */
2018 // @codingStandardsIgnoreStart Squiz.Classes.ValidClassName.NotCamelCaps
2019 class PPNode_Hash_Text implements PPNode {
2020 // @codingStandardsIgnoreEnd
2021
2022 public $value;
2023 private $store, $index;
2024
2025 /**
2026 * Construct an object using the data from $store[$index]. The rest of the
2027 * store array can be accessed via getNextSibling().
2028 *
2029 * @param array $store
2030 * @param int $index
2031 */
2032 public function __construct( array $store, $index ) {
2033 $this->value = $store[$index];
2034 if ( !is_scalar( $this->value ) ) {
2035 throw new MWException( __CLASS__ . ' given object instead of string' );
2036 }
2037 $this->store = $store;
2038 $this->index = $index;
2039 }
2040
2041 public function __toString() {
2042 return htmlspecialchars( $this->value );
2043 }
2044
2045 public function getNextSibling() {
2046 return PPNode_Hash_Tree::factory( $this->store, $this->index + 1 );
2047 }
2048
2049 public function getChildren() {
2050 return false;
2051 }
2052
2053 public function getFirstChild() {
2054 return false;
2055 }
2056
2057 public function getChildrenOfType( $name ) {
2058 return false;
2059 }
2060
2061 public function getLength() {
2062 return false;
2063 }
2064
2065 public function item( $i ) {
2066 return false;
2067 }
2068
2069 public function getName() {
2070 return '#text';
2071 }
2072
2073 public function splitArg() {
2074 throw new MWException( __METHOD__ . ': not supported' );
2075 }
2076
2077 public function splitExt() {
2078 throw new MWException( __METHOD__ . ': not supported' );
2079 }
2080
2081 public function splitHeading() {
2082 throw new MWException( __METHOD__ . ': not supported' );
2083 }
2084 }
2085
2086 /**
2087 * @ingroup Parser
2088 */
2089 // @codingStandardsIgnoreStart Squiz.Classes.ValidClassName.NotCamelCaps
2090 class PPNode_Hash_Array implements PPNode {
2091 // @codingStandardsIgnoreEnd
2092
2093 public $value;
2094
2095 public function __construct( $value ) {
2096 $this->value = $value;
2097 }
2098
2099 public function __toString() {
2100 return var_export( $this, true );
2101 }
2102
2103 public function getLength() {
2104 return count( $this->value );
2105 }
2106
2107 public function item( $i ) {
2108 return $this->value[$i];
2109 }
2110
2111 public function getName() {
2112 return '#nodelist';
2113 }
2114
2115 public function getNextSibling() {
2116 return false;
2117 }
2118
2119 public function getChildren() {
2120 return false;
2121 }
2122
2123 public function getFirstChild() {
2124 return false;
2125 }
2126
2127 public function getChildrenOfType( $name ) {
2128 return false;
2129 }
2130
2131 public function splitArg() {
2132 throw new MWException( __METHOD__ . ': not supported' );
2133 }
2134
2135 public function splitExt() {
2136 throw new MWException( __METHOD__ . ': not supported' );
2137 }
2138
2139 public function splitHeading() {
2140 throw new MWException( __METHOD__ . ': not supported' );
2141 }
2142 }
2143
2144 /**
2145 * @ingroup Parser
2146 */
2147 // @codingStandardsIgnoreStart Squiz.Classes.ValidClassName.NotCamelCaps
2148 class PPNode_Hash_Attr implements PPNode {
2149 // @codingStandardsIgnoreEnd
2150
2151 public $name, $value;
2152 private $store, $index;
2153
2154 /**
2155 * Construct an object using the data from $store[$index]. The rest of the
2156 * store array can be accessed via getNextSibling().
2157 *
2158 * @param array $store
2159 * @param int $index
2160 */
2161 public function __construct( array $store, $index ) {
2162 $descriptor = $store[$index];
2163 if ( $descriptor[PPNode_Hash_Tree::NAME][0] !== '@' ) {
2164 throw new MWException( __METHOD__.': invalid name in attribute descriptor' );
2165 }
2166 $this->name = substr( $descriptor[PPNode_Hash_Tree::NAME], 1 );
2167 $this->value = $descriptor[PPNode_Hash_Tree::CHILDREN][0];
2168 $this->store = $store;
2169 $this->index = $index;
2170 }
2171
2172 public function __toString() {
2173 return "<@{$this->name}>" . htmlspecialchars( $this->value ) . "</@{$this->name}>";
2174 }
2175
2176 public function getName() {
2177 return $this->name;
2178 }
2179
2180 public function getNextSibling() {
2181 return PPNode_Hash_Tree::factory( $this->store, $this->index + 1 );
2182 }
2183
2184 public function getChildren() {
2185 return false;
2186 }
2187
2188 public function getFirstChild() {
2189 return false;
2190 }
2191
2192 public function getChildrenOfType( $name ) {
2193 return false;
2194 }
2195
2196 public function getLength() {
2197 return false;
2198 }
2199
2200 public function item( $i ) {
2201 return false;
2202 }
2203
2204 public function splitArg() {
2205 throw new MWException( __METHOD__ . ': not supported' );
2206 }
2207
2208 public function splitExt() {
2209 throw new MWException( __METHOD__ . ': not supported' );
2210 }
2211
2212 public function splitHeading() {
2213 throw new MWException( __METHOD__ . ': not supported' );
2214 }
2215 }