| blob | 1d55c980a1ae8dc9967c41f26587bc1c42f3c561 |
1 <?php
2 /**
3 * PHP parser that converts wiki markup to HTML.
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 */
27 /**
28 * @defgroup Parser Parser
29 */
31 /**
32 * PHP Parser - Processes wiki markup (which uses a more user-friendly
33 * syntax, such as "[[link]]" for making links), and provides a one-way
34 * transformation of that wiki markup it into (X)HTML output / markup
35 * (which in turn the browser understands, and can display).
36 *
37 * There are seven main entry points into the Parser class:
38 *
39 * - Parser::parse()
40 * produces HTML output
41 * - Parser::preSaveTransform()
42 * produces altered wiki markup
43 * - Parser::preprocess()
44 * removes HTML comments and expands templates
45 * - Parser::cleanSig() and Parser::cleanSigInSig()
46 * cleans a signature before saving it to preferences
47 * - Parser::getSection()
48 * return the content of a section from an article for section editing
49 * - Parser::replaceSection()
50 * replaces a section by number inside an article
51 * - Parser::getPreloadText()
52 * removes <noinclude> sections and <includeonly> tags
53 *
54 * Globals used:
55 * object: $wgContLang
56 *
57 * @warning $wgUser or $wgTitle or $wgRequest or $wgLang. Keep them away!
58 *
59 * @par Settings:
60 * $wgNamespacesWithSubpages
61 *
62 * @par Settings only within ParserOptions:
63 * $wgAllowExternalImages
64 * $wgAllowSpecialInclusion
65 * $wgInterwikiMagic
66 * $wgMaxArticleSize
67 *
68 * @ingroup Parser
69 */
71 /**
72 * Update this version number when the ParserOutput format
73 * changes in an incompatible way, so the parser cache
74 * can automatically discard old data.
75 */
78 /**
79 * Update this version number when the output of serialiseHalfParsedText()
80 * changes in an incompatible way
81 */
84 # Flags for Parser::setFunctionHook
88 # Constants needed for external link processing
89 # Everything except bracket, space, or control characters
90 # \p{Zs} is unicode 'separator, space' category. It covers the space 0x20
91 # as well as U+3000 is IDEOGRAPHIC SPACE for T21052
93 # Simplified expression to match an IPv4 or IPv6 address, or
94 # at least one character of a host name (embeds EXT_LINK_URL_CLASS)
96 # RegExp to make image URLs (embeds IPv6 part of EXT_LINK_ADDR)
97 // @codingStandardsIgnoreStart Generic.Files.LineLength
98 const EXT_IMAGE_REGEX = '/^(http:\/\/|https:\/\/)((?:\\[(?i:[0-9a-f:.]+)\\])?[^][<>"\\x00-\\x20\\x7F\p{Zs}]+)
100 // @codingStandardsIgnoreEnd
102 # Regular expression for a non-newline space
105 # Flags for preprocessToDom
108 # Allowed values for $this->mOutputType
109 # Parameter to startExternalParse().
116 /**
117 * @var string Prefix and suffix for temporary replacement strings
118 * for the multipass parser.
119 *
120 * \x7f should never appear in input as it's disallowed in XML.
121 * Using it at the front also gives us a little extra robustness
122 * since it shouldn't match when butted up against identifier-like
123 * string constructs.
124 *
125 * Must not consist of all title characters, or else it will change
126 * the behavior of <nowiki> in a link.
127 *
128 * Must have a character that needs escaping in attributes, otherwise
129 * someone could put a strip marker in an attribute, to get around
130 * escaping quote marks, and break out of the attribute. Thus we add
131 * `'".
132 */
136 # Markers used for wrapping the table of contents
140 # Persistent:
154 # Initialised by initialiseVariables()
156 /**
157 * @var MagicWordArray
158 */
161 /**
162 * @var MagicWordArray
163 */
165 # Initialised in constructor
168 # Initialized in getPreprocessor()
169 /** @var Preprocessor */
172 # Cleared with clearState():
173 /**
174 * @var ParserOutput
175 */
179 /**
180 * @var StripState
181 */
185 /**
186 * @var LinkHolderArray
187 */
197 /**
198 * @var User
199 */
202 # Temporary
203 # These are variables reset at least once per parse regardless of $clearState
205 /**
206 * @var ParserOptions
207 */
210 /**
211 * @var Title
212 */
224 /**
225 * @var string Deprecated accessor for the strip marker prefix.
226 * @deprecated since 1.26; use Parser::MARKER_PREFIX instead.
227 */
230 /**
231 * @var array Array with the language name of each language link (i.e. the
232 * interwiki prefix) in the key, value arbitrary. Used to avoid sending
233 * duplicate language links to the ParserOutput.
234 */
237 /**
238 * @var MapCacheLRU|null
239 * @since 1.24
240 *
241 * A cache of the current revisions of titles. Keys are $title->getPrefixedDbKey()
242 */
245 /**
246 * @var bool Recursive call protection.
247 * This variable should be treated as if it were private.
248 */
251 /** @var SectionProfiler */
254 /**
255 * @var LinkRenderer
256 */
259 /**
260 * @param array $conf
261 */
271 # Preprocessor_Hash is much faster than Preprocessor_DOM under HipHop
274 # PECL extension that conflicts with the core DOM extension (T15770)
281 }
283 }
285 /**
286 * Reduce memory usage to reduce the impact of circular references
287 */
291 }
294 }
295 }
297 /**
298 * Allow extensions to clean up when the parser is cloned
299 */
303 // T58226: When you create a reference "to" an object field, that
304 // makes the object field itself be a reference too (until the other
305 // reference goes out of scope). When cloning, any field that's a
306 // reference is copied as a reference in the new object. Both of these
307 // are defined PHP5 behaviors, as inconvenient as it is for us when old
308 // hooks from PHP4 days are passing fields by reference.
310 // Make a non-reference copy of the field, then rebind the field to
311 // reference the new copy.
315 }
318 }
320 /**
321 * Do various kinds of initialisation on the first call of the parser
322 */
326 }
333 // Avoid PHP 7.1 warning from passing $this by reference
336 }
338 /**
339 * Clear Parser state
340 *
341 * @private
342 */
346 }
362 # Clear these on every parse, T6549
370 ];
379 # Fix cloning
382 }
386 // Avoid PHP 7.1 warning from passing $this by reference
389 }
391 /**
392 * Convert wikitext to HTML
393 * Do not call this function recursively.
394 *
395 * @param string $text Text we want to parse
396 * @param Title $title
397 * @param ParserOptions $options
398 * @param bool $linestart
399 * @param bool $clearState
400 * @param int $revid Number to pass in {{REVISIONID}}
401 * @return ParserOutput A ParserOutput
402 */
406 ) {
407 /**
408 * First pass--just handle <nowiki> sections, pass the rest off
409 * to internalParse() which does all the real work.
410 */
415 // We use U+007F DELETE to construct strip markers, so we have to make
416 // sure that this character does not occur in the input text.
419 }
427 }
440 }
442 // Avoid PHP 7.1 warning from passing $this by reference
445 # No more strip!
452 /**
453 * A converted title will be provided in the output object if title and
454 * content conversion are enabled, the article text does not contain
455 * a conversion-suppressing double-underscore tag, and no
456 * {{DISPLAYTITLE:...}} is present. DISPLAYTITLE takes precedence over
457 * automatic link conversion.
458 */
463 ) {
470 }
471 }
473 # Done parsing! Compute runtime adaptive expiry if set
476 # Warn if too many heavyweight parser functions were used
481 );
482 }
484 # Information on include size limits, for the benefit of users who try to skirt them
492 );
493 }
498 );
502 );
505 );
508 );
511 );
514 );
517 );
523 }
533 ) ) {
539 }
543 }
544 }
545 }
546 // Since we're not really outputting HTML, decode the entities and
547 // then re-encode the things that need hiding inside HTML comments.
551 // Sanitize for comment. Note '‐' in the replacement is U+2010,
552 // which looks much like the problematic '-'.
556 // Add on template profiling data in human/machine readable way
560 } );
566 }
572 // Add other cache related metadata
575 }
586 }
587 }
599 }
601 /**
602 * Half-parse wikitext to half-parsed HTML. This recursive parser entry point
603 * can be called from an extension tag hook.
604 *
605 * The output of this function IS NOT SAFE PARSED HTML; it is "half-parsed"
606 * instead, which means that lists and links have not been fully parsed yet,
607 * and strip markers are still present.
608 *
609 * Use recursiveTagParseFully() to fully parse wikitext to output-safe HTML.
610 *
611 * Use this function if you're a parser tag hook and you want to parse
612 * wikitext before or after applying additional transformations, and you
613 * intend to *return the result as hook output*, which will cause it to go
614 * through the rest of parsing process automatically.
615 *
616 * If $frame is not provided, then template variables (e.g., {{{1}}}) within
617 * $text are not expanded
618 *
619 * @param string $text Text extension wants to have parsed
620 * @param bool|PPFrame $frame The frame to use for expanding any template variables
621 * @return string UNSAFE half-parsed HTML
622 */
624 // Avoid PHP 7.1 warning from passing $this by reference
630 }
632 /**
633 * Fully parse wikitext to fully parsed HTML. This recursive parser entry
634 * point can be called from an extension tag hook.
635 *
636 * The output of this function is fully-parsed HTML that is safe for output.
637 * If you're a parser tag hook, you might want to use recursiveTagParse()
638 * instead.
639 *
640 * If $frame is not provided, then template variables (e.g., {{{1}}}) within
641 * $text are not expanded
642 *
643 * @since 1.25
644 *
645 * @param string $text Text extension wants to have parsed
646 * @param bool|PPFrame $frame The frame to use for expanding any template variables
647 * @return string Fully parsed HTML
648 */
653 }
655 /**
656 * Expand templates and variables in the text, producing valid, static wikitext.
657 * Also removes comments.
658 * Do not call this function recursively.
659 * @param string $text
660 * @param Title $title
661 * @param ParserOptions $options
662 * @param int|null $revid
663 * @param bool|PPFrame $frame
664 * @return mixed|string
665 */
668 ) {
673 }
674 // Avoid PHP 7.1 warning from passing $this by reference
681 }
683 /**
684 * Recursive parser entry point that can be called from an extension tag
685 * hook.
686 *
687 * @param string $text Text to be expanded
688 * @param bool|PPFrame $frame The frame to use for expanding any template variables
689 * @return string
690 * @since 1.19
691 */
696 }
698 /**
699 * Process the wikitext for the "?preload=" feature. (T7210)
700 *
701 * "<noinclude>", "<includeonly>" etc. are parsed as for template
702 * transclusion, comments, templates, arguments, tags hooks and parser
703 * functions are untouched.
704 *
705 * @param string $text
706 * @param Title $title
707 * @param ParserOptions $options
708 * @param array $params
709 * @return string
710 */
715 # Parser (re)initialisation
724 }
726 /**
727 * Get a random string
728 *
729 * @return string
730 * @deprecated since 1.26; use wfRandomString() instead.
731 */
735 }
737 /**
738 * Set the current user.
739 * Should only be used when doing pre-save transform.
740 *
741 * @param User|null $user User object or null (to reset)
742 */
745 }
747 /**
748 * Accessor for mUniqPrefix.
749 *
750 * @return string
751 * @deprecated since 1.26; use Parser::MARKER_PREFIX instead.
752 */
756 }
758 /**
759 * Set the context title
760 *
761 * @param Title $t
762 */
766 }
769 # Strip the fragment to avoid various odd effects
773 }
774 }
776 /**
777 * Accessor for the Title object
778 *
779 * @return Title
780 */
783 }
785 /**
786 * Accessor/mutator for the Title object
787 *
788 * @param Title $x Title object or null to just get the current one
789 * @return Title
790 */
793 }
795 /**
796 * Set the output type
797 *
798 * @param int $ot New value
799 */
802 # Shortcut alias
808 ];
809 }
811 /**
812 * Accessor/mutator for the output type
813 *
814 * @param int|null $x New value or null to just get the current one
815 * @return int
816 */
819 }
821 /**
822 * Get the ParserOutput object
823 *
824 * @return ParserOutput
825 */
828 }
830 /**
831 * Get the ParserOptions object
832 *
833 * @return ParserOptions
834 */
837 }
839 /**
840 * Accessor/mutator for the ParserOptions object
841 *
842 * @param ParserOptions $x New value or null to just get the current one
843 * @return ParserOptions Current ParserOptions object
844 */
847 }
849 /**
850 * @return int
851 */
854 }
856 /**
857 * @param int $id
858 */
861 }
863 /**
864 * Get a language object for use in parser functions such as {{FORMATNUM:}}
865 * @return Language
866 */
869 }
871 /**
872 * Get the target language for the content being parsed. This is usually the
873 * language that the content is in.
874 *
875 * @since 1.19
876 *
877 * @throws MWException
878 * @return Language
879 */
889 }
892 }
894 /**
895 * Get the language object for language conversion
896 * @return Language|null
897 */
900 }
902 /**
903 * Get a User object either from $this->mUser, if set, or from the
904 * ParserOptions object otherwise
905 *
906 * @return User
907 */
911 }
913 }
915 /**
916 * Get a preprocessor object
917 *
918 * @return Preprocessor
919 */
924 }
926 }
928 /**
929 * Get a LinkRenderer instance to make links with
930 *
931 * @since 1.28
932 * @return LinkRenderer
933 */
940 );
941 }
944 }
946 /**
947 * Replaces all occurrences of HTML-style comments and the given tags
948 * in the text with a random marker and returns the next text. The output
949 * parameter $matches will be an associative array filled with data in
950 * the form:
951 *
952 * @code
953 * 'UNIQ-xxxxx' => [
954 * 'element',
955 * 'tag content',
956 * [ 'param' => 'x' ],
957 * '<element param="x">tag content</element>' ]
958 * @endcode
959 *
960 * @param array $elements List of element names. Comments are always extracted.
961 * @param string $text Source text string.
962 * @param array $matches Out parameter, Array: extracted tags
963 * @param string|null $uniq_prefix
964 * @return string Stripped text
965 * @since 1.26 The uniq_prefix argument is deprecated.
966 */
967 public static function extractTagsAndParams( $elements, $text, &$matches, $uniq_prefix = null ) {
970 }
983 }
985 # comment
991 # tag
996 }
1002 # Empty element tag, <tag />
1011 }
1015 # No end tag -- let it run out to the end of the text.
1021 }
1022 }
1028 }
1030 }
1032 /**
1033 * Get a list of strippable XML-like elements
1034 *
1035 * @return array
1036 */
1039 }
1041 /**
1042 * Add an item to the strip state
1043 * Returns the unique tag which must be inserted into the stripped text
1044 * The tag will be replaced with the original text in unstrip()
1045 *
1046 * @param string $text
1047 *
1048 * @return string
1049 */
1055 }
1057 /**
1058 * parse the wiki syntax used to render tables
1059 *
1060 * @private
1061 * @param string $text
1062 * @return string
1063 */
1081 }
1088 # First check if we are starting a new table
1101 # Don't do any of the following
1105 # We are ending a table
1111 }
1115 }
1119 }
1123 # Now we have a table row
1126 # Whats after the tag is now only attributes
1139 }
1143 }
1152 ) {
1153 # This might be cell elements, td, th or captions
1159 }
1161 // Implies both are valid for table headings.
1164 }
1166 # Split up multiple cells on the same line.
1167 # FIXME : This can result in improper nesting of tags processed
1168 # by earlier parser steps.
1173 # Loop through each table cell
1180 }
1185 }
1191 }
1201 }
1205 # A cell could contain both parameters and data
1208 # T2553: Note that a '|' inside an invalid link should not
1209 # be mistaken as delimiting cell parameters
1210 # Bug T153140: Neither should language converter markup.
1219 }
1223 }
1224 }
1226 }
1228 # Closing open td, tr && table
1232 }
1235 }
1238 }
1241 }
1243 # Remove trailing line-ending (b/c)
1246 }
1248 # special case: don't return empty table
1251 }
1254 }
1256 /**
1257 * Helper function for parse() that transforms wiki markup into half-parsed
1258 * HTML. Only called for $mOutputType == self::OT_HTML.
1259 *
1260 * @private
1261 *
1262 * @param string $text The text to parse
1263 * @param bool $isMain Whether this is being called from the main parse() function
1264 * @param PPFrame|bool $frame A pre-processor frame
1265 *
1266 * @return string
1267 */
1272 // Avoid PHP 7.1 warning from passing $this by reference
1275 # Hook to suspend the parser in this state
1278 }
1280 # if $frame is provided, then use $frame for replacing any variables
1282 # use frame depth to infer how include/noinclude tags should be handled
1283 # depth=0 means this is the top-level document; otherwise it's an included document
1288 }
1292 # if $frame is not provided, then use old-style replaceVariables
1294 }
1302 [],
1304 );
1307 # Tables need to come after variable replacement for things to work
1308 # properly; putting them before other transformations should keep
1309 # exciting things like link expansions from showing up in surprising
1310 # places.
1322 # replaceInternalLinks may sometimes leave behind
1323 # absolute URLs, which have to be masked to hide them from replaceExternalLinks
1330 }
1332 /**
1333 * Helper function for parse() that transforms half-parsed HTML into fully
1334 * parsed HTML.
1335 *
1336 * @param string $text
1337 * @param bool $isMain
1338 * @param bool $linestart
1339 * @return string
1340 */
1344 // Avoid PHP 7.1 warning from passing $this by reference
1349 }
1351 # Clean up special characters, only run once, next-to-last before doBlockLevels
1353 # French spaces, last one Guillemet-left
1354 # only if there is something before the space
1356 # french spaces, Guillemet-right
1359 ];
1366 /**
1367 * The input doesn't get language converted if
1368 * a) It's disabled
1369 * b) Content isn't converted
1370 * c) It's a conversion table
1371 * d) it is an interface message (which is in the user language)
1372 */
1375 ) {
1377 # The position of the convert() call should not be changed. it
1378 # assumes that the links are all replaced and the only thing left
1379 # is the <nowiki> mark.
1381 }
1382 }
1388 }
1398 }
1400 # attempt to sanitize at least some nesting problems
1401 # (T4702 and quite a few others)
1403 # ''Something [http://www.cool.com cool''] -->
1404 # <i>Something</i><a href="http://www.cool.com"..><i>cool></i></a>
1407 # fix up an anchor inside another anchor, only
1408 # at least for a single single nested link (T5695)
1411 # fix div inside inline elements- doBlockLevels won't wrap a line which
1412 # contains a div, so fix it up here; replace
1413 # div with escaped text
1416 # remove empty italic or bold tag pairs, some
1417 # introduced by rules above
1419 ];
1425 }
1429 }
1432 }
1434 /**
1435 * Replace special strings like "ISBN xxx" and "RFC xxx" with
1436 * magic external links.
1437 *
1438 * DML
1439 * @private
1440 *
1441 * @param string $text
1442 *
1443 * @return string
1444 */
1453 '!(?: # Start cases
1455 (<.*?>) | # m[2]: Skip stuff inside
1458 # m[4]: Post-protocol path
1464 [0-9Xx] # check digit
1468 }
1470 /**
1471 * @throws MWException
1472 * @param array $m
1473 * @return HTML|string
1474 */
1477 # Skip anchor
1480 # Skip HTML element
1483 # Free external link
1486 # RFC or PMID
1490 }
1499 }
1508 }
1511 return Linker::makeExternalLink( $url, "{$keyword} {$id}", true, $cssClass, [], $this->mTitle );
1514 ) {
1515 # ISBN
1523 ] );
1528 [
1531 ]
1532 );
1535 }
1536 }
1538 /**
1539 * Make a free external link, given a user-supplied URL
1540 *
1541 * @param string $url
1542 * @param int $numPostProto
1543 * The number of characters after the protocol.
1544 * @return string HTML
1545 * @private
1546 */
1550 # The characters '<' and '>' (which were escaped by
1551 # removeHTMLtags()) should not be included in
1552 # URLs, per RFC 2396.
1553 # Make terminate a URL as well (bug T84937)
1559 PREG_OFFSET_CAPTURE
1560 ) ) {
1563 }
1565 # Move trailing punctuation to $trail
1567 # If there is no left bracket, then consider right brackets fair game too
1570 }
1574 # Don't break a trailing HTML entity by moving the ; into $trail
1575 # This is in hot code, so use substr_compare to avoid having to
1576 # create a new string object for the comparison
1578 # more optimization: instead of running preg_match with a $
1579 # anchor, which can be slow, do the match on the reversed
1580 # string starting at the desired offset.
1581 # un-reversed regexp is: /&([a-z]+|#x[\da-f]+|#\d+)$/i
1584 }
1585 }
1589 }
1591 # Verify that we still have a real URL after trail removal, and
1592 # not just lone protocol
1595 }
1599 # Is this an external image?
1602 # Not an image, make a link
1607 # Register it in the output object...
1608 # Replace unnecessary URL escape codes with their equivalent characters
1611 }
1613 }
1615 /**
1616 * Parse headers and return html
1617 *
1618 * @private
1619 *
1620 * @param string $text
1621 *
1622 * @return string
1623 */
1628 }
1630 }
1632 /**
1633 * Replace single quotes with HTML markup
1634 * @private
1635 *
1636 * @param string $text
1637 *
1638 * @return string The altered text
1639 */
1645 }
1648 }
1650 /**
1651 * Helper function for doAllQuotes()
1652 *
1653 * @param string $text
1654 *
1655 * @return string
1656 */
1662 }
1664 // First, do some preliminary work. This may shift some apostrophes from
1665 // being mark-up to being text. It also counts the number of occurrences
1666 // of bold and italics mark-ups.
1671 // If there are ever four apostrophes, assume the first is supposed to
1672 // be text, and the remaining three constitute mark-up for bold text.
1673 // (T15227: ''''foo'''' turns into ' ''' foo ' ''')
1679 // If there are more than 5 apostrophes in a row, assume they're all
1680 // text except for the last 5.
1681 // (T15227: ''''''foo'''''' turns into ' ''''' foo ' ''''')
1685 }
1686 // Count the number of occurrences of bold and italics mark-ups.
1694 }
1695 }
1697 // If there is an odd number of both bold and italics, it is likely
1698 // that one of the bold ones was meant to be an apostrophe followed
1699 // by italics. Which one we cannot know for certain, but it is more
1700 // likely to be one that has a single-letter word before it.
1712 }
1715 // if $firstsingleletterword is set, we don't
1716 // look at the other options, so we can bail early.
1721 }
1722 }
1723 }
1724 }
1726 // If there is a single-letter word, use it!
1731 // If not, but there's a multi-letter word, use that one.
1735 // ... otherwise use the first one that has neither.
1736 // (notice that it is possible for all three to be -1 if, for example,
1737 // there is only one pentuple-apostrophe in the line)
1740 }
1741 }
1743 // Now let's actually convert our apostrophic mush to HTML!
1754 }
1773 }
1790 }
1810 }
1811 }
1812 }
1814 }
1815 // Now close all remaining tags. Notice that the order is important.
1818 }
1821 }
1824 }
1825 // There might be lonely ''''', so make sure we have a buffer
1828 }
1830 }
1832 /**
1833 * Replace external links (REL)
1834 *
1835 * Note: this is all very hackish and the order of execution matters a lot.
1836 * Make sure to run tests/parser/parserTests.php if you change this code.
1837 *
1838 * @private
1839 *
1840 * @param string $text
1841 *
1842 * @throws MWException
1843 * @return string
1844 */
1851 }
1861 # The characters '<' and '>' (which were escaped by
1862 # removeHTMLtags()) should not be included in
1863 # URLs, per RFC 2396.
1868 }
1870 # If the link text is an image URL, replace it with an <img> tag
1871 # This happened by accident in the original parser, but some people used it extensively
1875 }
1879 # Set linktype for CSS - if URL==text, link is essentially free
1882 # No link text, e.g. [http://domain.tld/some.link]
1884 # Autonumber
1889 # Have link text, e.g. [http://domain.tld/some.link text]s
1890 # Check for trail
1892 }
1898 # Use the encoded URL
1899 # This means that users can paste URLs directly into the text
1900 # Funny characters like ö aren't valid in URLs anyway
1901 # This was changed in August 2004
1905 # Register link in the output object.
1906 # Replace unnecessary URL escape codes with the referenced character
1907 # This prevents spammers from hiding links from the filters
1910 }
1913 }
1915 /**
1916 * Get the rel attribute for a particular external link.
1917 *
1918 * @since 1.21
1919 * @param string|bool $url Optional URL, to extract the domain from for rel =>
1920 * nofollow if appropriate
1921 * @param Title $title Optional Title, for wgNoFollowNsExceptions lookups
1922 * @return string|null Rel attribute for $url
1923 */
1929 ) {
1931 }
1933 }
1935 /**
1936 * Get an associative array of additional HTML attributes appropriate for a
1937 * particular external link. This currently may include rel => nofollow
1938 * (depending on configuration, namespace, and the URL's domain) and/or a
1939 * target attribute (depending on configuration).
1940 *
1941 * @param string $url URL to extract the domain from for rel =>
1942 * nofollow if appropriate
1943 * @return array Associative array of HTML attributes
1944 */
1953 // T133507. New windows can navigate parent cross-origin.
1954 // Including noreferrer due to lacking browser
1955 // support of noopener. Eventually noreferrer should be removed.
1958 }
1960 }
1961 }
1964 }
1966 /**
1967 * Replace unusual escape codes in a URL with their equivalent characters
1968 *
1969 * This generally follows the syntax defined in RFC 3986, with special
1970 * consideration for HTTP query strings.
1971 *
1972 * @param string $url
1973 * @return string
1974 */
1976 # First, make sure unsafe characters are encoded
1980 },
1981 $url
1982 );
1987 # Fragment part - 'fragment'
1993 }
1995 # Query part - 'query' minus &=+;
2001 }
2003 # Scheme and path part - 'pchar'
2004 # (we assume no userinfo or encoded colons in the host)
2009 }
2016 # Unescape it
2019 # Leave it escaped, but use uppercase for a-f
2021 }
2022 };
2024 }
2026 /**
2027 * make an image if it's allowed, either through the global
2028 * option, through the exception, or through the on-wiki whitelist
2029 *
2030 * @param string $url
2031 *
2032 * @return string
2033 */
2038 # $imagesfrom could be either a single string or an array of strings, parse out the latter
2045 }
2046 }
2051 }
2055 ) {
2057 # Image found
2059 }
2060 }
2063 ) {
2067 );
2070 # Sanitize the regex fragment, make it case-insensitive, ignore blank entries/comments
2073 }
2075 # Image matches a whitelist entry
2078 }
2079 }
2080 }
2082 }
2084 /**
2085 * Process [[ ]] wikilinks
2086 *
2087 * @param string $s
2088 *
2089 * @return string Processed text
2090 *
2091 * @private
2092 */
2096 }
2098 /**
2099 * Process [[ ]] wikilinks (RIL)
2100 * @param string $s
2101 * @throws MWException
2102 * @return LinkHolderArray
2103 *
2104 * @private
2105 */
2110 # the % is needed to support urlencoded titles as well
2113 # Match a link having the form [[namespace:link|alternate]]trail
2115 # Match cases where there is no "]]", which might still be images
2117 }
2121 # split the entire text string on occurrences of [[
2123 # get the first element (all text up to first [[), and remove the space we added
2132 # Match the end of a line for a word that's not followed by whitespace,
2133 # e.g. in the case of 'The Arab al[[Razi]]', 'al' will be matched
2137 }
2141 }
2150 }
2153 }
2157 // @codingStandardsIgnoreStart Squiz.WhiteSpace.SemicolonSpacing.Incorrect
2158 # Loop for each link
2160 // @codingStandardsIgnoreEnd
2162 # Check for excessive memory usage
2164 # Too big
2165 # Do the existence check, replace the link holders and clear the array
2168 }
2176 }
2177 # first link
2181 }
2182 }
2188 # If we get a ] at the beginning of $m[3] that means we have a link that's something like:
2189 # [[Image:Foo.jpg|[http://example.com desc]]] <- having three ] in a row fucks up,
2190 # the real problem is with the $e1 regex
2191 # See T1500.
2192 # Still some problems for cases where the ] is meant to be outside punctuation,
2193 # and no image is in sight. See T4095.
2197 ) {
2200 }
2201 # fix up urlencoded title texts
2203 # Should anchors '#' also be rejected?
2205 }
2208 # Invalid, but might be an image with a link in its caption
2213 }
2218 }
2222 # Don't allow internal links to pages containing
2223 # PROTO: where PROTO is a valid URL protocol; these
2224 # should be external links.
2228 }
2230 # Make subpage if necessary
2235 }
2239 # Strip off leading ':'
2241 }
2248 }
2257 # look at the next 'line' to see if we can close it there
2262 }
2265 # the first ]] closes the inner link, the second the image
2271 # if there's exactly one ]] that's fine, we'll keep looking
2274 # if $next_line is invalid too, we need look no further
2277 }
2278 }
2280 # we couldn't find the end of this imageLink, so output it raw
2281 # but don't ignore what might be perfectly normal links in the text we've examined
2284 # note: no $trail, because without an end, there *is* no trail
2286 }
2289 # note: no $trail, because without an end, there *is* no trail
2291 }
2292 }
2298 # T6598 madness. Handle the quotes only if they come from the alternate part
2299 # [[Lista d''e paise d''o munno]] -> <a href="...">Lista d''e paise d''o munno</a>
2300 # [[Criticism of Harry Potter|Criticism of ''Harry Potter'']]
2301 # -> <a href="Criticism of Harry Potter">Criticism of <i>Harry Potter</i></a>
2303 }
2305 # Link not escaped by : , create the various objects
2307 # Interwikis
2312 )
2313 ) {
2314 # T26502: filter duplicates
2318 }
2323 }
2328 # if no parameters were passed, $text
2329 # becomes something like "File:Foo.png",
2330 # which we don't want to pass on to the
2331 # image generator
2334 # recursively parse links inside the image caption
2335 # actually, this will parse them in any other parameters, too,
2336 # but it might be hard to fix that, and it doesn't matter ATM
2339 }
2340 # cloak any absolute URLs inside the image markup, so replaceExternalLinks() won't touch them
2344 }
2352 }
2358 /**
2359 * Strip the whitespace Category links produce, see T2087
2360 */
2364 }
2365 }
2367 # Self-link checking. For some languages, variants of the title are checked in
2368 # LinkHolderArray::doVariants() to allow batching the existence checks necessary
2369 # for linking to a different variant.
2373 }
2375 # NS_MEDIA is a pseudo-namespace for linking directly to a file
2376 # @todo FIXME: Should do batch file existence checks, see comment below
2378 # Give extensions a chance to select the file revision for us
2383 # Fetch and register the file (file title may be different via hooks)
2385 # Cloak with NOPARSE to avoid replacement in replaceExternalLinks
2389 }
2391 # Some titles, such as valid special pages or files in foreign repos, should
2392 # be shown as bluelinks even though they're not included in the page table
2393 # @todo FIXME: isAlwaysKnown() can be expensive for file links; we should really do
2394 # batch file existence checks for NS_FILE and NS_MEDIA
2399 # Links will be added to the output link list after checking
2401 }
2402 }
2404 }
2406 /**
2407 * Render a forced-blue link inline; protect against double expansion of
2408 * URLs if we're in a mode that prepends full URL prefixes to internal links.
2409 * Since this little disaster has to split off the trail text to avoid
2410 * breaking URLs in the following text without breaking trails on the
2411 * wiki links, it's been made into a horrible function.
2412 *
2413 * @param Title $nt
2414 * @param string $text
2415 * @param string $trail
2416 * @param string $prefix
2417 * @return string HTML-wikitext mix oh yuck
2418 */
2424 }
2428 );
2431 }
2433 /**
2434 * Insert a NOPARSE hacky thing into any inline links in a chunk that's
2435 * going to go through further parsing steps before inline URL expansion.
2436 *
2437 * Not needed quite as much as it used to be since free links are a bit
2438 * more sensible these days. But bracketed links are still an issue.
2439 *
2440 * @param string $text More-or-less HTML
2441 * @return string Less-or-more HTML with NOPARSE bits
2442 */
2446 }
2448 /**
2449 * Return true if subpage links should be expanded on this page.
2450 * @return bool
2451 */
2453 # Some namespaces don't allow subpages
2455 }
2457 /**
2458 * Handle link to subpage if necessary
2459 *
2460 * @param string $target The source of the link
2461 * @param string &$text The link text, modified as necessary
2462 * @return string The full name of the link
2463 * @private
2464 */
2467 }
2469 /**
2470 * Make lists from lines starting with ':', '*', '#', etc. (DBL)
2471 *
2472 * @param string $text
2473 * @param bool $linestart Whether or not this is at the start of a line.
2474 * @private
2475 * @return string The lists rendered as HTML
2476 */
2479 }
2481 /**
2482 * Return value of a magic variable (like PAGENAME)
2483 *
2484 * @private
2485 *
2486 * @param int $index
2487 * @param bool|PPFrame $frame
2488 *
2489 * @throws MWException
2490 * @return string
2491 */
2497 // If no title set, bad things are going to happen
2498 // later. Title should always be set since this
2499 // should only be called in the middle of a parse
2500 // operation (but the unit-tests do funky stuff)
2503 }
2505 // Avoid PHP 7.1 warning from passing $this by reference
2508 /**
2509 * Some of these require message or data lookups and can be
2510 * expensive to check many times.
2511 */
2515 }
2516 }
2561 $value = $pageLang->getMonthAbbreviation( MWTimestamp::getLocalInstance( $ts )->format( 'n' ) );
2595 ) ) );
2605 ) ) );
2613 }
2621 }
2634 # 0 means the page doesn't exist in the database,
2635 # which means the user is previewing a new page.
2636 # The vary-revision flag must be set, because the magic word
2637 # will have a different value once the page is saved.
2640 }
2644 # Let the edit saving system know we should parse the page
2645 # *after* a revision ID has been assigned.
2652 }
2655 # Let the edit saving system know we should parse the page
2656 # *after* a revision ID has been assigned. This is for null edits.
2662 # Let the edit saving system know we should parse the page
2663 # *after* a revision ID has been assigned. This is for null edits.
2669 # Let the edit saving system know we should parse the page
2670 # *after* a revision ID has been assigned. This is for null edits.
2676 # Let the edit saving system know we should parse the page
2677 # *after* a revision ID has been assigned. This is for null edits.
2683 # Let the edit saving system know we should parse the page
2684 # *after* a revision ID has been assigned. This is for null edits.
2690 # Let the edit saving system know we should parse the page
2691 # *after* a revision ID has been assigned. This is for null edits.
2697 # Let the edit saving system know we should parse the page
2698 # *after* a revision ID has been assigned for null edits.
2742 # @bug T6594 PHP5 has it zero padded, PHP4 does not, cast to
2743 # int to remove the padding
2752 );
2761 false
2762 );
2768 # @bug T6594 PHP5 has it zero padded, PHP4 does not, cast to
2769 # int to remove the padding
2833 );
2836 }
2840 }
2843 }
2845 /**
2846 * initialise the magic variables (like CURRENTMONTHNAME) and substitution modifiers
2847 *
2848 * @private
2849 */
2856 }
2858 /**
2859 * Preprocess some wikitext and return the document tree.
2860 * This is the ghost of replace_variables().
2861 *
2862 * @param string $text The text to parse
2863 * @param int $flags Bitwise combination of:
2864 * - self::PTD_FOR_INCLUSION: Handle "<noinclude>" and "<includeonly>" as if the text is being
2865 * included. Default is to assume a direct page view.
2866 *
2867 * The generated DOM tree must depend only on the input text and the flags.
2868 * The DOM tree must be the same in OT_HTML and OT_WIKI mode, to avoid a regression of T6899.
2869 *
2870 * Any flag added to the $flags parameter here, or any other parameter liable to cause a
2871 * change in the DOM tree for a given text, must be passed through the section identifier
2872 * in the section edit link and thus back to extractSections().
2873 *
2874 * The output of this function is currently only cached in process memory, but a persistent
2875 * cache may be implemented at a later date which takes further advantage of these strict
2876 * dependency requirements.
2877 *
2878 * @return PPNode
2879 */
2883 }
2885 /**
2886 * Return a three-element array: leading whitespace, string contents, trailing whitespace
2887 *
2888 * @param string $s
2889 *
2890 * @return array
2891 */
2901 }
2903 }
2905 /**
2906 * Replace magic variables, templates, and template arguments
2907 * with the appropriate text. Templates are substituted recursively,
2908 * taking care to avoid infinite loops.
2909 *
2910 * Note that the substitution depends on value of $mOutputType:
2911 * self::OT_WIKI: only {{subst:}} templates
2912 * self::OT_PREPROCESS: templates but not extension tags
2913 * self::OT_HTML: all templates and extension tags
2914 *
2915 * @param string $text The text to transform
2916 * @param bool|PPFrame $frame Object describing the arguments passed to the
2917 * template. Arguments may also be provided as an associative array, as
2918 * was the usual case before MW1.12. Providing arguments this way may be
2919 * useful for extensions wishing to perform variable replacement
2920 * explicitly.
2921 * @param bool $argsOnly Only do argument (triple-brace) expansion, not
2922 * double-brace expansion.
2923 * @return string
2924 */
2926 # Is there any text? Also, Prevent too big inclusions!
2930 }
2938 }
2945 }
2947 /**
2948 * Clean up argument array - refactored in 1.9 so parserfunctions can use it, too.
2949 *
2950 * @param array $args
2951 *
2952 * @return array
2953 */
2966 }
2969 }
2970 }
2971 }
2974 }
2976 /**
2977 * Warn the user when a parser limitation is reached
2978 * Will warn at most once the user per limitation type
2979 *
2980 * The results are shown during preview and run through the Parser (See EditPage.php)
2981 *
2982 * @param string $limitationType Should be one of:
2983 * 'expensive-parserfunction' (corresponding messages:
2984 * 'expensive-parserfunction-warning',
2985 * 'expensive-parserfunction-category')
2986 * 'post-expand-template-argument' (corresponding messages:
2987 * 'post-expand-template-argument-warning',
2988 * 'post-expand-template-argument-category')
2989 * 'post-expand-template-inclusion' (corresponding messages:
2990 * 'post-expand-template-inclusion-warning',
2991 * 'post-expand-template-inclusion-category')
2992 * 'node-count-exceeded' (corresponding messages:
2993 * 'node-count-exceeded-warning',
2994 * 'node-count-exceeded-category')
2995 * 'expansion-depth-exceeded' (corresponding messages:
2996 * 'expansion-depth-exceeded-warning',
2997 * 'expansion-depth-exceeded-category')
2998 * @param string|int|null $current Current value
2999 * @param string|int|null $max Maximum allowed, when an explicit limit has been
3000 * exceeded, provide the values (optional)
3001 */
3003 # does no harm if $current and $max are present but are unnecessary for the message
3004 # Not doing ->inLanguage( $this->mOptions->getUserLangObj() ), since this is shown
3005 # only during preview, and that would split the parser cache unnecessarily.
3010 }
3012 /**
3013 * Return the text of a template, after recursively
3014 * replacing any variables or templates within the template.
3015 *
3016 * @param array $piece The parts of the template
3017 * $piece['title']: the title, i.e. the part before the |
3018 * $piece['parts']: the parameter array
3019 * $piece['lineStart']: whether the brace was at the start of a line
3020 * @param PPFrame $frame The current frame, contains template arguments
3021 * @throws Exception
3022 * @return string The text of the template
3023 */
3026 // Flags
3028 // $text has been filled
3030 // wiki markup in $text should be escaped
3032 // $text is HTML, armour it against wikitext transformation
3034 // Force interwiki transclusion to be done in raw mode not rendered
3036 // $text is a DOM node needing expansion in a child frame
3038 // $text is a DOM node needing expansion in the current frame
3041 # Title object, where $text came from
3044 # $part1 is the bit before the first |, and must contain only title characters.
3045 # Various prefixes will be stripped from it later.
3050 # Original title text preserved for various purposes
3053 # $args is a list of argument nodes, starting from index 0, not including $part1
3054 # @todo FIXME: If piece['parts'] is null then the call to getLength()
3055 # below won't work b/c this $args isn't an object
3060 # SUBST
3064 # Possibilities for substMatch: "subst", "safesubst" or FALSE
3065 # Decide whether to expand template or keep wikitext as-is.
3071 }
3077 }
3078 }
3083 }
3084 }
3086 # Variables
3093 }
3095 }
3096 }
3098 # MSG, MSGNW and RAW
3100 # Check for MSGNW:
3105 # Remove obsolete MSG:
3108 }
3110 # Check for RAW:
3114 }
3115 }
3117 # Parser functions
3126 }
3131 }
3133 # The interface for parser functions allows for extracting
3134 # flags into the local scope. Extract any forwarded flags
3135 # here.
3137 }
3138 }
3140 # Finish mangling title and then check for loops.
3141 # Set $title to a Title object and $titleText to the PDBK
3144 # Split the title into page and subpage
3150 }
3154 # Check for language variants if the template is not found
3157 }
3158 # Do recursion depth check
3166 }
3167 }
3168 }
3170 # Load from database
3177 ) {
3179 // Pass the template arguments as URL parameters.
3180 // "uselang" will have no effect since the Language object
3181 // is forced to the one defined in ParserOptions.
3190 }
3191 }
3193 // Create a new context to execute the special page
3200 // If this page is cached, then we better not be per user.
3202 }
3214 );
3215 }
3216 }
3226 }
3227 }
3229 # If the title is valid but undisplayable, make a link to it
3233 }
3235 # Interwiki transclusion
3241 # Preprocess it like a template
3244 }
3246 }
3248 # Do infinite loop check
3249 # This has to be done after redirect resolution to avoid infinite loops via redirects
3256 }
3257 }
3259 # If we haven't found text to substitute by now, we're done
3260 # Recover the source wikitext and return it
3265 }
3267 }
3269 # Expand DOM-style return values in a child frame
3271 # Clean up argument array
3277 # Expansion is eligible for the empty-frame cache
3280 # Uncached expansion
3282 }
3283 }
3287 }
3291 }
3293 # Replace raw HTML by a placeholder
3297 # Escape nowiki-style return values
3302 ) {
3303 # T2529: if the template begins with a table or block-level
3304 # element, it should be treated as beginning a new line.
3305 # This behavior is somewhat controversial.
3307 }
3310 # Error, oversize inclusion
3312 # Make a working, properly escaped link if possible (T25588)
3315 # This will probably not be a working link, but at least it may
3316 # provide some hint of where the problem is
3319 }
3323 }
3329 }
3332 }
3334 /**
3335 * Call a parser function and return an array with text and flags.
3336 *
3337 * The returned array will always contain a boolean 'found', indicating
3338 * whether the parser function was found or not. It may also contain the
3339 * following:
3340 * text: string|object, resulting wikitext or PP DOM object
3341 * isHTML: bool, $text is HTML, armour it against wikitext transformation
3342 * isChildObj: bool, $text is a DOM node needing expansion in a child frame
3343 * isLocalObj: bool, $text is a DOM node needing expansion in the current frame
3344 * nowiki: bool, wiki markup in $text should be escaped
3345 *
3346 * @since 1.21
3347 * @param PPFrame $frame The current frame, contains template arguments
3348 * @param string $function Function name
3349 * @param array $args Arguments to the function
3350 * @throws MWException
3351 * @return array
3352 */
3356 # Case sensitive functions
3360 # Case insensitive functions
3366 }
3367 }
3371 # Workaround for PHP bug 35229 and similar
3374 }
3376 // Avoid PHP 7.1 warning from passing $this by reference
3381 # Convert arguments to PPNodes and collect for appending to $allArgs
3388 }
3389 }
3391 # Add a frame parameter, and pass the arguments as an array
3395 # Convert arguments to plain text and append to $allArgs
3403 }
3404 }
3405 }
3409 # The interface for function hooks allows them to return a wikitext
3410 # string or an array containing the string and any flags. This mungs
3411 # things around to match what this method should return.
3416 ];
3420 }
3424 ];
3425 }
3431 }
3434 }
3439 }
3442 }
3444 /**
3445 * Get the semi-parsed DOM representation of a template with a given title,
3446 * and its redirect destination title. Cached.
3447 *
3448 * @param Title $title
3449 *
3450 * @return array
3451 */
3460 }
3463 }
3465 # Cache miss, go to the database
3471 }
3479 }
3482 }
3484 /**
3485 * Fetch the current revision of a given title. Note that the revision
3486 * (and even the title) may not exist in the database, so everything
3487 * contributing to the output of the parser should use this method
3488 * where possible, rather than getting the revisions themselves. This
3489 * method also caches its results, so using it benefits performance.
3490 *
3491 * @since 1.24
3492 * @param Title $title
3493 * @return Revision
3494 */
3499 }
3502 // Defaults to Parser::statelessFetchRevision()
3504 );
3505 }
3507 }
3509 /**
3510 * Wrapper around Revision::newFromTitle to allow passing additional parameters
3511 * without passing them on to it.
3512 *
3513 * @since 1.24
3514 * @param Title $title
3515 * @param Parser|bool $parser
3516 * @return Revision|bool False if missing
3517 */
3525 }
3528 }
3530 /**
3531 * Fetch the unparsed text of a template and register a reference to it.
3532 * @param Title $title
3533 * @return array ( string or false, Title )
3534 */
3536 // Defaults to Parser::statelessFetchTemplate()
3539 // We use U+007F DELETE to distinguish strip markers from regular text.
3543 }
3549 // If we transclude ourselves, the final result
3550 // will change based on the new version of the page
3552 }
3553 }
3554 }
3556 }
3558 /**
3559 * Fetch the unparsed text of a template and register a reference to it.
3560 * @param Title $title
3561 * @return string|bool
3562 */
3565 }
3567 /**
3568 * Static function to get a template
3569 * Can be overridden via ParserOptions::setTemplateCallback().
3570 *
3571 * @param Title $title
3572 * @param bool|Parser $parser
3573 *
3574 * @return array
3575 */
3581 # Loop to fetch the article, with up to 1 redirect
3582 // @codingStandardsIgnoreStart Generic.CodeAnalysis.ForLoopWithTestFunctionCall.NotAllowed
3584 // @codingStandardsIgnoreEnd
3585 # Give extensions a chance to select the revision instead
3596 ];
3598 }
3599 # Get the revision
3606 }
3608 # If there is no current revision, there is no page
3612 }
3619 # We fetched a rev from a different title; register it too...
3624 }
3636 }
3643 }
3648 }
3651 }
3652 # Redirect?
3655 }
3660 }
3662 /**
3663 * Fetch a file and its title and register a reference to it.
3664 * If 'broken' is a key in $options then the file will appear as a broken thumbnail.
3665 * @param Title $title
3666 * @param array $options Array of options to RepoGroup::findFile
3667 * @return File|bool
3668 */
3671 }
3673 /**
3674 * Fetch a file and its title and register a reference to it.
3675 * If 'broken' is a key in $options then the file will appear as a broken thumbnail.
3676 * @param Title $title
3677 * @param array $options Array of options to RepoGroup::findFile
3678 * @return array ( File or false, Title of file )
3679 */
3685 # Register the file as a dependency...
3688 # Update fetched file title
3691 }
3693 }
3695 /**
3696 * Helper function for fetchFileAndTitle.
3697 *
3698 * Also useful if you need to fetch a file but not use it yet,
3699 * for example to get the file's handler.
3700 *
3701 * @param Title $title
3702 * @param array $options Array of options to RepoGroup::findFile
3703 * @return File|bool
3704 */
3712 }
3714 }
3716 /**
3717 * Transclude an interwiki link.
3718 *
3719 * @param Title $title
3720 * @param string $action
3721 *
3722 * @return string
3723 */
3729 }
3735 }
3737 }
3739 /**
3740 * @param string $url
3741 * @return mixed|string
3742 */
3751 }
3758 // Though we failed to fetch the content, this status is useless.
3763 }
3770 ] );
3772 }
3774 /**
3775 * Triple brace replacement -- used for template arguments
3776 * @private
3777 *
3778 * @param array $piece
3779 * @param PPFrame $frame
3780 *
3781 * @return array
3782 */
3795 )
3796 ) {
3797 # No match in frame, use the supplied default
3799 }
3803 }
3806 # No match anywhere
3808 }
3811 }
3816 }
3819 }
3821 /**
3822 * Return the text to be used for a given extension tag.
3823 * This is the ghost of strip().
3824 *
3825 * @param array $params Associative array of parameters:
3826 * name PPNode for the tag name
3827 * attr PPNode for unparsed text where tag attributes are thought to be
3828 * attributes Optional associative array of parsed attributes
3829 * inner Contents of extension element
3830 * noClose Original text did not have a close tag
3831 * @param PPFrame $frame
3832 *
3833 * @throws MWException
3834 * @return string
3835 */
3842 // Probably expansion depth or node count exceeded. Just punt the
3843 // error up.
3845 }
3849 // See above
3851 }
3853 // We can't safely check if the expansion for $content resulted in an
3854 // error, because the content could happen to be the error string
3855 // (T149622).
3867 }
3873 }
3876 # Workaround for PHP bug 35229 and similar
3879 }
3886 }
3888 // Avoid PHP 7.1 warning from passing $this by reference
3894 }
3897 # Extract flags to local scope (to override $markerType)
3902 }
3906 }
3911 }
3912 }
3918 // See above
3920 }
3922 }
3923 }
3933 }
3935 }
3937 /**
3938 * Increment an include size counter
3939 *
3940 * @param string $type The type of expansion
3941 * @param int $size The size of the text
3942 * @return bool False if this inclusion would take it over the maximum, true otherwise
3943 */
3950 }
3951 }
3953 /**
3954 * Increment the expensive function count
3955 *
3956 * @return bool False if the limit has been exceeded
3957 */
3961 }
3963 /**
3964 * Strip double-underscore items like __NOGALLERY__ and __NOTOC__
3965 * Fills $this->mDoubleUnderscores, returns the modified text
3966 *
3967 * @param string $text
3968 *
3969 * @return string
3970 */
3973 # The position of __TOC__ needs to be recorded
3979 # Set a placeholder. At the end we'll fill it in with the TOC.
3982 # Only keep the first one.
3984 }
3986 # Now match and remove the rest of them
3992 }
3995 }
3998 ) {
4000 }
4001 # (T10068) Allow control over whether robots index a page.
4002 # __INDEX__ always overrides __NOINDEX__, see T16899
4006 }
4010 }
4012 # Cache all double underscores in the database
4015 }
4018 }
4020 /**
4021 * @see ParserOutput::addTrackingCategory()
4022 * @param string $msg Message key
4023 * @return bool Whether the addition was successful
4024 */
4027 }
4029 /**
4030 * This function accomplishes several tasks:
4031 * 1) Auto-number headings if that option is enabled
4032 * 2) Add an [edit] link to sections for users who have enabled the option and can edit the page
4033 * 3) Add a Table of contents on the top for users who have enabled the option
4034 * 4) Auto-anchor headings
4035 *
4036 * It loops through all headlines, collects the necessary data, then splits up the
4037 * string and re-inserts the newly formatted headlines.
4038 *
4039 * @param string $text
4040 * @param string $origText Original, untouched wikitext
4041 * @param bool $isMain
4042 * @return mixed|string
4043 * @private
4044 */
4048 # Inhibit editsection links if requested in the page
4054 }
4057 }
4059 # Get all headlines for numbering them and adding funky stuff like [edit]
4060 # links - this is for later, but we need the number of headlines right now
4065 $matches
4066 );
4068 # if there are fewer than 4 headlines in the article, do not show TOC
4069 # unless it's been explicitly enabled.
4073 # Allow user to stipulate that a page should have a "new section"
4074 # link added via __NEWSECTIONLINK__
4077 }
4079 # Allow user to remove the "new section"
4080 # link via __NONEWSECTIONLINK__
4083 }
4085 # if the string __FORCETOC__ (not case-sensitive) occurs in the HTML,
4086 # override above conditions and always show TOC above first header
4090 }
4092 # headline counter
4096 # Ugh .. the TOC should have neat indentation levels which can be
4097 # passed to the skin functions. These are determined here
4131 }
4135 }
4139 # Increase TOC level
4146 }
4148 # Decrease TOC level, find level to jump to
4152 # Found last matching level
4156 # Found first matching level below current level
4159 }
4160 }
4163 }
4166 # Unindent only if the previous toc level was shown :p
4171 }
4172 }
4174 # No change in level, end TOC line
4177 }
4178 }
4182 # count number of headlines for each level
4189 }
4192 }
4193 }
4195 # The safe header is a version of the header text safe to use for links
4197 # Remove link placeholders by the link text.
4198 # <!--LINK number-->
4199 # turns into
4200 # link text with suffix
4201 # Do this before unstrip since link text can contain strip markers
4204 # Avoid insertion of weird stuff like <math> by expanding the relevant sections
4207 # Strip out HTML (first regex removes any tag not allowed)
4208 # Allowed tags are:
4209 # * <sup> and <sub> (T10393)
4210 # * <i> (T28375)
4211 # * <b> (r105284)
4212 # * <bdi> (T74884)
4213 # * <span dir="rtl"> and <span dir="ltr"> (T37167)
4214 # * <s> and <strike> (T35715)
4215 # We strip any parameter from accepted tags (second regex), except dir="rtl|ltr" from <span>,
4216 # to allow setting directionality in toc items.
4218 [
4220 '#<(/?(?:span(?: dir="(?:rtl|ltr)")?|sup|sub|bdi|i|b|s|strike))(?: .*?)?>#'
4221 ],
4223 $safeHeadline
4224 );
4226 # Strip '<span></span>', which is the result from the above if
4227 # <span id="foo"></span> is used to produce an additional anchor
4228 # for a section.
4233 # For the anchor, strip out HTML-y stuff period
4237 # Save headline for section edit hint before it's escaped
4241 # For reverse compatibility, provide an id that's
4242 # HTML4-compatible, like we used to.
4243 # It may be worth noting, academically, that it's possible for
4244 # the legacy anchor to conflict with a non-legacy headline
4245 # anchor on the page. In this case likely the "correct" thing
4246 # would be to either drop the legacy anchors or make sure
4247 # they're numbered first. However, this would require people
4248 # to type in section names like "abc_.D7.93.D7.90.D7.A4"
4249 # manually, so let's not bother worrying about it.
4255 # No reason to have both (in fact, we can't)
4257 }
4262 }
4264 # HTML names must be case-insensitively unique (T12721).
4265 # This does not apply to Unicode characters per
4266 # https://www.w3.org/TR/html5/infrastructure.html#case-sensitivity-and-string-comparison
4267 # @todo FIXME: We may be changing them depending on the current locale.
4273 }
4275 # Create the anchor for linking from the TOC to the section
4279 // @codingStandardsIgnoreStart
4281 // @codingStandardsIgnoreEnd
4286 }
4288 // @codingStandardsIgnoreStart
4290 // @codingStandardsIgnoreEnd
4295 }
4297 # Don't number the heading if it is the only one (looks silly)
4299 # the two are different if the line contains a link
4303 $numbering
4305 }
4310 }
4312 # Add the section to the section tree
4313 # Find the DOM node for this header
4320 }
4321 }
4325 }
4335 ];
4337 # give headline the correct <h#> tag
4339 // Output edit section links as markers with styles that can be customized by skins
4341 # Put a T flag in the section identifier, to indicate to extractSections()
4342 # that sections inside <includeonly> should be counted.
4350 }
4351 // We use a bit of pesudo-xml for editsection markers. The
4352 // language converter is run later on. Using a UNIQ style marker
4353 // leads to the converter screwing up the tokens when it
4354 // converts stuff. And trying to insert strip tags fails too. At
4355 // this point all real inputted tags have already been escaped,
4356 // so we don't have to worry about a user trying to input one of
4357 // these markers directly. We use a page and section attribute
4358 // to stop the language converter from converting these
4359 // important bits of data, but put the headline hint inside a
4360 // content block because the language converter is supposed to
4361 // be able to convert that piece of data.
4362 // Gets replaced with html in ParserOutput::getText
4369 }
4372 }
4378 }
4382 # Never ever show TOC if no headers
4385 }
4390 }
4395 }
4399 }
4401 # split up and insert constructed headlines
4405 // build an array of document sections
4408 // $head is zero-based, sections aren't.
4413 }
4415 /**
4416 * Send a hook, one per section.
4417 * The idea here is to be able to make section-level DIVs, but to do so in a
4418 * lower-impact, more correct way than r50769
4419 *
4420 * $this : caller
4421 * $section : the section number
4422 * &$sectionContent : ref to the content of the section
4423 * $showEditLinks : boolean describing whether this section has an edit link
4424 */
4428 }
4431 // append the TOC at the beginning
4432 // Top anchor now in skin
4434 }
4442 }
4443 }
4445 /**
4446 * Transform wiki markup when saving a page by doing "\r\n" -> "\n"
4447 * conversion, substituting signatures, {{subst:}} templates, etc.
4448 *
4449 * @param string $text The text to transform
4450 * @param Title $title The Title object for the current article
4451 * @param User $user The User object describing the current user
4452 * @param ParserOptions $options Parsing options
4453 * @param bool $clearState Whether to clear the parser state first
4454 * @return string The altered wiki markup
4455 */
4458 ) {
4461 }
4465 // We still normalize line endings for backwards-compatibility
4466 // with other code that just calls PST, but this should already
4467 // be handled in TextContent subclasses
4472 }
4478 }
4480 /**
4481 * Pre-save transform helper function
4482 *
4483 * @param string $text
4484 * @param User $user
4485 *
4486 * @return string
4487 */
4491 # Note: This is the timestamp saved as hardcoded wikitext to
4492 # the database, we use $wgContLang here in order to give
4493 # everyone the same signature and use the default one rather
4494 # than the one selected in each user's preferences.
4495 # (see also T14815)
4503 # Variable replacement
4504 # Because mOutputType is OT_WIKI, this will only process {{subst:xxx}} type tags
4507 # This works almost by chance, as the replaceVariables are done before the getUserSig(),
4508 # which may corrupt this parser instance via its wfMessage()->text() call-
4510 # Signatures
4516 ] );
4518 # Context links ("pipe tricks"): [[|name]] and [[name (context)|]]
4522 // [[ns:page (context)|]]
4524 // [[ns:page（context）|]] (double-width brackets, added in r40257)
4526 // [[ns:page (context), context|]] (using either single or double-width comma)
4528 // [[|page]] (reverse pipe trick: add context from page title)
4531 # try $p1 first, to turn "[[A, B (C)|]]" into "[[A, B (C)|A, B]]"
4543 # if there's no context, don't bother duplicating the title
4545 }
4548 }
4550 /**
4551 * Fetch the user's signature text, if any, and normalize to
4552 * validated, ready-to-insert wikitext.
4553 * If you have pre-fetched the nickname or the fancySig option, you can
4554 * specify them here to save a database query.
4555 * Do not reuse this parser instance after calling getUserSig(),
4556 * as it may have changed if it's the $wgParser.
4557 *
4558 * @param User $user
4559 * @param string|bool $nickname Nickname to use or false to use user's default nickname
4560 * @param bool|null $fancySig whether the nicknname is the complete signature
4561 * or null to use default value
4562 * @return string
4563 */
4569 # If not given, retrieve from the user object.
4572 }
4576 }
4584 # Sig. might contain markup; validate this
4586 # Validated; clean up (if needed) and return it
4589 # Failed to validate; fall back to the default
4592 }
4593 }
4595 # Make sure nickname doesnt get a sig in a sig
4598 # If we're still here, make it a link to the user page
4605 }
4607 /**
4608 * Check that the user's signature contains no bad XML
4609 *
4610 * @param string $text
4611 * @return string|bool An expanded string, or false if invalid.
4612 */
4615 }
4617 /**
4618 * Clean up signature text
4619 *
4620 * 1) Strip 3, 4 or 5 tildes out of signatures @see cleanSigInSig
4621 * 2) Substitute all transclusions
4622 *
4623 * @param string $text
4624 * @param bool $parsing Whether we're cleaning (preferences save) or parsing
4625 * @return string Signature text
4626 */
4632 }
4634 # Option to disable this feature
4637 }
4639 # @todo FIXME: Regex doesn't respect extension tags or nowiki
4640 # => Move this logic to braceSubstitution()
4642 $substRegex = '/\{\{(?!(?:' . $substWord->getBaseRegex() . '))/x' . $substWord->getRegexCase();
4653 }
4656 }
4658 /**
4659 * Strip 3, 4 or 5 tildes out of signatures.
4660 *
4661 * @param string $text
4662 * @return string Signature text with /~{3,5}/ removed
4663 */
4667 }
4669 /**
4670 * Set up some variables which are usually set up in parse()
4671 * so that an external function can call some class members with confidence
4672 *
4673 * @param Title|null $title
4674 * @param ParserOptions $options
4675 * @param int $outputType
4676 * @param bool $clearState
4677 */
4680 ) {
4682 }
4684 /**
4685 * @param Title|null $title
4686 * @param ParserOptions $options
4687 * @param int $outputType
4688 * @param bool $clearState
4689 */
4692 ) {
4698 }
4699 }
4701 /**
4702 * Wrapper for preprocess()
4703 *
4704 * @param string $text The text to preprocess
4705 * @param ParserOptions $options Options
4706 * @param Title|null $title Title object or null to use $wgTitle
4707 * @return string
4708 */
4712 # Guard against infinite recursion
4715 }
4721 }
4727 }
4729 /**
4730 * Create an HTML-style tag, e.g. "<yourtag>special text</yourtag>"
4731 * The callback should have the following form:
4732 * function myParserHook( $text, $params, $parser, $frame ) { ... }
4733 *
4734 * Transform and return $text. Use $parser for any required context, e.g. use
4735 * $parser->getTitle() and $parser->getOptions() not $wgTitle or $wgOut->mParserOptions
4736 *
4737 * Hooks may return extended information by returning an array, of which the
4738 * first numbered element (index 0) must be the return string, and all other
4739 * entries are extracted into local variables within an internal function
4740 * in the Parser class.
4741 *
4742 * This interface (introduced r61913) appears to be undocumented, but
4743 * 'markerType' is used by some core tag hooks to override which strip
4744 * array their results are placed in. **Use great caution if attempting
4745 * this interface, as it is not documented and injudicious use could smash
4746 * private variables.**
4747 *
4748 * @param string $tag The tag to use, e.g. 'hook' for "<hook>"
4749 * @param callable $callback The callback function (and object) to use for the tag
4750 * @throws MWException
4751 * @return callable|null The old value of the mTagHooks array associated with the hook
4752 */
4757 }
4762 }
4765 }
4767 /**
4768 * As setHook(), but letting the contents be parsed.
4769 *
4770 * Transparent tag hooks are like regular XML-style tag hooks, except they
4771 * operate late in the transformation sequence, on HTML instead of wikitext.
4772 *
4773 * This is probably obsoleted by things dealing with parser frames?
4774 * The only extension currently using it is geoserver.
4775 *
4776 * @since 1.10
4777 * @todo better document or deprecate this
4778 *
4779 * @param string $tag The tag to use, e.g. 'hook' for "<hook>"
4780 * @param callable $callback The callback function (and object) to use for the tag
4781 * @throws MWException
4782 * @return callable|null The old value of the mTagHooks array associated with the hook
4783 */
4788 }
4789 $oldVal = isset( $this->mTransparentTagHooks[$tag] ) ? $this->mTransparentTagHooks[$tag] : null;
4793 }
4795 /**
4796 * Remove all tag hooks
4797 */
4802 }
4804 /**
4805 * Create a function, e.g. {{sum:1|2|3}}
4806 * The callback function should have the form:
4807 * function myParserFunction( &$parser, $arg1, $arg2, $arg3 ) { ... }
4808 *
4809 * Or with Parser::SFH_OBJECT_ARGS:
4810 * function myParserFunction( $parser, $frame, $args ) { ... }
4811 *
4812 * The callback may either return the text result of the function, or an array with the text
4813 * in element 0, and a number of flags in the other elements. The names of the flags are
4814 * specified in the keys. Valid flags are:
4815 * found The text returned is valid, stop processing the template. This
4816 * is on by default.
4817 * nowiki Wiki markup in the return value should be escaped
4818 * isHTML The returned text is HTML, armour it against wikitext transformation
4819 *
4820 * @param string $id The magic word ID
4821 * @param callable $callback The callback function (and object) to use
4822 * @param int $flags A combination of the following flags:
4823 * Parser::SFH_NO_HASH No leading hash, i.e. {{plural:...}} instead of {{#if:...}}
4824 *
4825 * Parser::SFH_OBJECT_ARGS Pass the template arguments as PPNode objects instead of text.
4826 * This allows for conditional expansion of the parse tree, allowing you to eliminate dead
4827 * branches and thus speed up parsing. It is also possible to analyse the parse tree of
4828 * the arguments, and to control the way they are expanded.
4829 *
4830 * The $frame parameter is a PPFrame. This can be used to produce expanded text from the
4831 * arguments, for instance:
4832 * $text = isset( $args[0] ) ? $frame->expand( $args[0] ) : '';
4833 *
4834 * For technical reasons, $args[0] is pre-expanded and will be a string. This may change in
4835 * future versions. Please call $frame->expand() on it anyway so that your code keeps
4836 * working if/when this is changed.
4837 *
4838 * If you want whitespace to be trimmed from $args, you need to do it yourself, post-
4839 * expansion.
4840 *
4841 * Please read the documentation in includes/parser/Preprocessor.php for more information
4842 * about the methods available in PPFrame and PPNode.
4843 *
4844 * @throws MWException
4845 * @return string|callable The old callback function for this name, if any
4846 */
4853 # Add to function cache
4857 }
4863 # Case
4866 }
4867 # Add leading hash
4870 }
4871 # Remove trailing colon
4874 }
4876 }
4878 }
4880 /**
4881 * Get all registered function hook identifiers
4882 *
4883 * @return array
4884 */
4887 }
4889 /**
4890 * Create a tag function, e.g. "<test>some stuff</test>".
4891 * Unlike tag hooks, tag functions are parsed at preprocessor level.
4892 * Unlike parser functions, their content is not preprocessed.
4893 * @param string $tag
4894 * @param callable $callback
4895 * @param int $flags
4896 * @throws MWException
4897 * @return null
4898 */
4903 }
4910 }
4913 }
4915 /**
4916 * Replace "<!--LINK-->" link placeholders with actual links, in the buffer
4917 * Placeholders created in Linker::link()
4918 *
4919 * @param string $text
4920 * @param int $options
4921 */
4924 }
4926 /**
4927 * Replace "<!--LINK-->" link placeholders with plain text of links
4928 * (not HTML-formatted).
4929 *
4930 * @param string $text
4931 * @return string
4932 */
4935 }
4937 /**
4938 * Renders an image gallery from a text with one line per image.
4939 * text labels may be given by using |-style alternative text. E.g.
4940 * Image:one.jpg|The number "1"
4941 * Image:tree.jpg|A tree
4942 * given as text will return the HTML of a gallery with two images,
4943 * labeled 'The number "1"' and
4944 * 'A tree'.
4945 *
4946 * @param string $text
4947 * @param array $params
4948 * @return string HTML
4949 */
4955 }
4960 // If invalid type set, fallback to default.
4962 }
4975 }
4981 }
4984 }
4987 }
4990 }
4993 // Avoid PHP 7.1 warning from passing $this by reference
4999 # match lines like these:
5000 # Image:someimage.jpg|This is some image
5003 # Skip empty lines
5006 }
5010 }
5013 # Bogus title. Ignore these so we don't bomb out later.
5015 }
5017 # We need to get what handler the file uses, to figure out parameters.
5018 # Note, a hook can overide the file name, and chose an entirely different
5019 # file (which potentially could be of a different type and have different handler).
5024 # Don't register it now, as TraditionalImageGallery does that later.
5031 ];
5034 // We don't want people to specify per-image widths.
5035 // Additionally the width parameter would need special casing anyhow.
5037 }
5046 // look for an |alt= definition while trying not to break existing
5047 // captions with multiple pipes (|) in it, until a more sensible grammar
5048 // is defined for images in galleries
5050 // FIXME: Doing recursiveTagParse at this stage, and the trim before
5051 // splitting on '|' is a bit odd, and different from makeImage.
5053 // Protect LanguageConverter markup
5056 );
5072 // check to see if link matches an absolute url, if not then it must be a wiki link.
5074 // Result of LanguageConverter::markNoConversion
5075 // invoked on an external link.
5077 }
5084 }
5085 }
5088 // Must be a handler specific parameter.
5092 // Guess not, consider it as caption.
5095 }
5096 }
5099 // Last pipe wins.
5101 }
5102 }
5103 // Remove the pipe.
5105 }
5108 }
5112 }
5114 /**
5115 * @param MediaHandler $handler
5116 * @return array
5117 */
5123 }
5125 # Initialise static lists
5132 ];
5140 }
5141 }
5142 }
5144 # Add handler params
5150 }
5151 }
5154 }
5156 }
5158 /**
5159 * Parse image options text and use it to make an image
5160 *
5161 * @param Title $title
5162 * @param string $options
5163 * @param LinkHolderArray|bool $holders
5164 * @return string HTML
5165 */
5167 # Check if the options text is of the form "options|alt text"
5168 # Options are:
5169 # * thumbnail make a thumbnail with enlarge-icon and caption, alignment depends on lang
5170 # * left no resizing, just left align. label is used for alt= only
5171 # * right same, but right aligned
5172 # * none same, but not aligned
5173 # * ___px scale to ___ pixels width, no aligning. e.g. use in taxobox
5174 # * center center the image
5175 # * frame Keep original image size, no magnify-button.
5176 # * framed Same as "frame"
5177 # * frameless like 'thumb' but without a frame. Keeps user preferences for width
5178 # * upright reduce width for upright images, rounded to full __0 px
5179 # * border draw a 1px border around the image
5180 # * alt Text for HTML alt attribute (defaults to empty)
5181 # * class Set a class for img node
5182 # * link Set the target of the image link. Can be external, interwiki, or local
5183 # vertical-align values (no % or length right now):
5184 # * baseline
5185 # * sub
5186 # * super
5187 # * top
5188 # * text-top
5189 # * middle
5190 # * bottom
5191 # * text-bottom
5193 # Protect LanguageConverter markup when splitting into parts
5196 );
5198 # Give extensions a chance to select the file revision for us
5203 # Fetch and register the file (file title may be different via hooks)
5206 # Get parameter map
5213 }
5215 # Process the input parameters
5227 # Special case; width and height come in one variable together
5235 }
5236 }
5242 }
5243 }
5244 # else no validation -- T15436
5247 # Validate handler parameter
5250 # Validate internal parameters
5255 # @todo FIXME: Possibly check validity here for
5256 # manualthumb? downstream behavior seems odd with
5257 # missing manual thumbs.
5275 }
5277 }
5285 }
5286 }
5291 // use first appearing option, discard others.
5296 # Most other things appear to be empty or numeric...
5298 }
5299 }
5303 }
5304 }
5305 }
5308 }
5309 }
5311 # Process alignment parameters
5314 }
5317 }
5321 # Will the image be presented in a frame, with the caption below?
5327 # In the old days, [[Image:Foo|text...]] would set alt text. Later it
5328 # came to also set the caption, ordinary text after the image -- which
5329 # makes no sense, because that just repeats the text multiple times in
5330 # screen readers. It *also* came to set the title attribute.
5331 # Now that we have an alt attribute, we should not set the alt text to
5332 # equal the caption: that's worse than useless, it just repeats the
5333 # text. This is the framed/thumbnail case. If there's no caption, we
5334 # use the unnamed parameter for alt text as well, just for the time be-
5335 # ing, if the unnamed param is set and the alt param is not.
5336 # For the future, we need to figure out if we want to tweak this more,
5337 # e.g., introducing a title= parameter for the title; ignoring the un-
5338 # named parameter entirely for images without a caption; adding an ex-
5339 # plicit caption= parameter and preserving the old magic unnamed para-
5340 # meter for BC; ...
5343 # No caption or alt text, add the filename as the alt text so
5344 # that screen readers at least get some description of the image
5346 }
5347 # Do not set $params['frame']['title'] because tooltips don't make sense
5348 # for framed images
5351 # No alt text, use the "caption" for the alt text
5355 # No caption, fall back to using the filename for the
5356 # alt text
5358 }
5359 }
5360 # Use the "caption" for the tooltip text
5362 }
5366 # Linker does the rest
5371 # Give the handler a chance to modify the parser object
5374 }
5377 }
5379 /**
5380 * @param string $caption
5381 * @param LinkHolderArray|bool $holders
5382 * @return mixed|string
5383 */
5385 # Strip bad stuff out of the title (tooltip). We can't just use
5386 # replaceLinkHoldersText() here, because if this function is called
5387 # from replaceInternalLinks2(), mLinkHolders won't be up-to-date.
5392 }
5394 # make sure there are no placeholders in thumbnail attributes
5395 # that are later expanded to html- so expand them now and
5396 # remove the tags
5401 }
5403 /**
5404 * Set a flag in the output object indicating that the content is dynamic and
5405 * shouldn't be cached.
5406 * @deprecated since 1.28; use getOutput()->updateCacheExpiry()
5407 */
5413 }
5415 }
5417 /**
5418 * Callback from the Sanitizer for expanding items found in HTML attribute
5419 * values, so they can be safely tested and escaped.
5420 *
5421 * @param string $text
5422 * @param bool|PPFrame $frame
5423 * @return string
5424 */
5429 }
5431 /**
5432 * Accessor
5433 *
5434 * @return array
5435 */
5441 );
5442 }
5444 /**
5445 * Replace transparent tags in $text with the values given by the callbacks.
5446 *
5447 * Transparent tag hooks are like regular XML-style tag hooks, except they
5448 * operate late in the transformation sequence, on HTML instead of wikitext.
5449 *
5450 * @param string $text
5451 *
5452 * @return string
5453 */
5467 );
5470 }
5472 }
5474 }
5476 /**
5477 * Break wikitext input into sections, and either pull or replace
5478 * some particular section's text.
5479 *
5480 * External callers should use the getSection and replaceSection methods.
5481 *
5482 * @param string $text Page wikitext
5483 * @param string|int $sectionId A section identifier string of the form:
5484 * "<flag1> - <flag2> - ... - <section number>"
5485 *
5486 * Currently the only recognised flag is "T", which means the target section number
5487 * was derived during a template inclusion parse, in other words this is a template
5488 * section edit link. If no flags are given, it was an ordinary section edit link.
5489 * This flag is required to avoid a section numbering mismatch when a section is
5490 * enclosed by "<includeonly>" (T8563).
5491 *
5492 * The section number 0 pulls the text before the first heading; other numbers will
5493 * pull the given section along with its lower-level subsections. If the section is
5494 * not found, $mode=get will return $newtext, and $mode=replace will return $text.
5495 *
5496 * Section 0 is always considered to exist, even if it only contains the empty
5497 * string. If $text is the empty string and section 0 is replaced, $newText is
5498 * returned.
5499 *
5500 * @param string $mode One of "get" or "replace"
5501 * @param string $newText Replacement text for section data.
5502 * @return string For "get", the extracted section text.
5503 * for "replace", the whole page with the section replaced.
5504 */
5513 # Process section extraction flags
5520 }
5521 }
5523 # Check for empty input
5525 # Only sections 0 and T-0 exist in an empty document
5531 }
5537 }
5538 }
5539 }
5541 # Preprocess the text
5544 # <h> nodes indicate section breaks
5545 # They can only occur at the top level, so we can find them by iterating the root's children
5548 # Find the target section
5550 # Section zero doesn't nest, level=big
5559 }
5560 }
5563 }
5565 }
5566 }
5569 # Not found
5574 }
5575 }
5577 # Find the end of the section, including nested sections
5584 }
5585 }
5588 }
5592 # Write out the remainder (in replace mode only)
5594 # Output the replacement text
5595 # Add two newlines on -- trailing whitespace in $newText is conventionally
5596 # stripped by the editor, so we need both newlines to restore the paragraph gap
5597 # Only add trailing whitespace if there is newText
5600 }
5605 }
5606 }
5609 # Re-insert stripped tags
5611 }
5614 }
5616 /**
5617 * This function returns the text of a section, specified by a number ($section).
5618 * A section is text under a heading like == Heading == or \<h1\>Heading\</h1\>, or
5619 * the first section before any such heading (section 0).
5620 *
5621 * If a section contains subsections, these are also returned.
5622 *
5623 * @param string $text Text to look in
5624 * @param string|int $sectionId Section identifier as a number or string
5625 * (e.g. 0, 1 or 'T-1').
5626 * @param string $defaultText Default to return if section is not found
5627 *
5628 * @return string Text of the requested section
5629 */
5632 }
5634 /**
5635 * This function returns $oldtext after the content of the section
5636 * specified by $section has been replaced with $text. If the target
5637 * section does not exist, $oldtext is returned unchanged.
5638 *
5639 * @param string $oldText Former text of the article
5640 * @param string|int $sectionId Section identifier as a number or string
5641 * (e.g. 0, 1 or 'T-1').
5642 * @param string $newText Replacing text
5643 *
5644 * @return string Modified text
5645 */
5648 }
5650 /**
5651 * Get the ID of the revision we are parsing
5652 *
5653 * @return int|null
5654 */
5657 }
5659 /**
5660 * Get the revision object for $this->mRevisionId
5661 *
5662 * @return Revision|null Either a Revision object or null
5663 * @since 1.23 (public since 1.23)
5664 */
5668 }
5671 }
5675 );
5677 # If the parse is for a new revision, then the callback should have
5678 # already been set to force the object and should match mRevisionId.
5679 # If not, try to fetch by mRevisionId for sanity.
5682 }
5687 }
5689 /**
5690 * Get the timestamp associated with the current revision, adjusted for
5691 * the default server-local timestamp
5692 * @return string
5693 */
5701 # The cryptic '' timezone parameter tells to use the site-default
5702 # timezone offset instead of the user settings.
5703 # Since this value will be saved into the parser cache, served
5704 # to other users, and potentially even used inside links and such,
5705 # it needs to be consistent for all visitors.
5708 }
5710 }
5712 /**
5713 * Get the name of the user that edited the last revision
5714 *
5715 * @return string User name
5716 */
5721 # if this template is subst: the revision id will be blank,
5722 # so just use the current user's name
5727 }
5728 }
5730 }
5732 /**
5733 * Get the size of the revision
5734 *
5735 * @return int|null Revision size
5736 */
5741 # if this variable is subst: the revision id will be blank,
5742 # so just use the parser input size, because the own substituation
5743 # will change the size.
5748 }
5749 }
5751 }
5753 /**
5754 * Mutator for $mDefaultSort
5755 *
5756 * @param string $sort New value
5757 */
5761 }
5763 /**
5764 * Accessor for $mDefaultSort
5765 * Will use the empty string if none is set.
5766 *
5767 * This value is treated as a prefix, so the
5768 * empty string is equivalent to sorting by
5769 * page name.
5770 *
5771 * @return string
5772 */
5778 }
5779 }
5781 /**
5782 * Accessor for $mDefaultSort
5783 * Unlike getDefaultSort(), will return false if none is set
5784 *
5785 * @return string|bool
5786 */
5789 }
5791 /**
5792 * Try to guess the section anchor name based on a wikitext fragment
5793 * presumably extracted from a heading, for example "Header" from
5794 * "== Header ==".
5795 *
5796 * @param string $text
5797 *
5798 * @return string
5799 */
5801 # Strip out wikitext links(they break the anchor)
5805 }
5807 /**
5808 * Same as guessSectionNameFromWikiText(), but produces legacy anchors
5809 * instead. For use in redirects, since IE6 interprets Redirect: headers
5810 * as something other than UTF-8 (apparently?), resulting in breakage.
5811 *
5812 * @param string $text The section name
5813 * @return string An anchor
5814 */
5816 # Strip out wikitext links(they break the anchor)
5820 }
5822 /**
5823 * Strips a text string of wikitext for use in a section anchor
5824 *
5825 * Accepts a text string and then removes all wikitext from the
5826 * string and leaves only the resultant text (i.e. the result of
5827 * [[User:WikiSysop|Sysop]] would be "Sysop" and the result of
5828 * [[User:WikiSysop]] would be "User:WikiSysop") - this is intended
5829 * to create valid section anchors by mimicing the output of the
5830 * parser when headings are parsed.
5831 *
5832 * @param string $text Text string to be stripped of wikitext
5833 * for use in a Section anchor
5834 * @return string Filtered text string
5835 */
5837 # Strip internal link markup
5841 # Strip external link markup
5842 # @todo FIXME: Not tolerant to blank link text
5843 # I.E. [https://www.mediawiki.org] will render as [1] or something depending
5844 # on how many empty links there are on the page - need to figure that out.
5845 $text = preg_replace( '/\[(?i:' . $this->mUrlProtocols . ')([^ ]+?) ([^[]+)\]/', '$2', $text );
5847 # Parse wikitext quotes (italics & bold)
5850 # Strip HTML tags
5853 }
5855 /**
5856 * strip/replaceVariables/unstrip for preprocessor regression testing
5857 *
5858 * @param string $text
5859 * @param Title $title
5860 * @param ParserOptions $options
5861 * @param int $outputType
5862 *
5863 * @return string
5864 */
5867 ) {
5875 }
5877 /**
5878 * @param string $text
5879 * @param Title $title
5880 * @param ParserOptions $options
5881 * @return string
5882 */
5885 }
5887 /**
5888 * @param string $text
5889 * @param Title $title
5890 * @param ParserOptions $options
5891 * @return string
5892 */
5895 }
5897 /**
5898 * Call a callback function on all regions of the given text that are not
5899 * inside strip markers, and replace those regions with the return value
5900 * of the callback. For example, with input:
5901 *
5902 * aaa<MARKER>bbb
5903 *
5904 * This will call the callback function twice, with 'aaa' and 'bbb'. Those
5905 * two strings will be replaced with the value returned by the callback in
5906 * each case.
5907 *
5908 * @param string $s
5909 * @param callable $callback
5910 *
5911 * @return string
5912 */
5931 }
5932 }
5933 }
5935 }
5937 /**
5938 * Remove any strip markers found in the given text.
5939 *
5940 * @param string $text Input string
5941 * @return string
5942 */
5945 }
5947 /**
5948 * Save the parser state required to convert the given half-parsed text to
5949 * HTML. "Half-parsed" in this context means the output of
5950 * recursiveTagParse() or internalParse(). This output has strip markers
5951 * from replaceVariables (extensionSubstitution() etc.), and link
5952 * placeholders from replaceLinkHolders().
5953 *
5954 * Returns an array which can be serialized and stored persistently. This
5955 * array can later be loaded into another parser instance with
5956 * unserializeHalfParsedText(). The text can then be safely incorporated into
5957 * the return value of a parser hook.
5958 *
5959 * @param string $text
5960 *
5961 * @return array
5962 */
5969 ];
5971 }
5973 /**
5974 * Load the parser state given in the $data array, which is assumed to
5975 * have been generated by serializeHalfParsedText(). The text contents is
5976 * extracted from the array, and its markers are transformed into markers
5977 * appropriate for the current Parser instance. This transformed text is
5978 * returned, and can be safely included in the return value of a parser
5979 * hook.
5980 *
5981 * If the $data array has been stored persistently, the caller should first
5982 * check whether it is still valid, by calling isValidHalfParsedText().
5983 *
5984 * @param array $data Serialized data
5985 * @throws MWException
5986 * @return string
5987 */
5991 }
5993 # First, extract the strip state.
5997 # Now renumber links
6000 # Should be good to go.
6002 }
6004 /**
6005 * Returns true if the given array, presumed to be generated by
6006 * serializeHalfParsedText(), is compatible with the current version of the
6007 * parser.
6008 *
6009 * @param array $data
6010 *
6011 * @return bool
6012 */
6015 }
6017 /**
6018 * Parsed a width param of imagelink like 300px or 200x300px
6019 *
6020 * @param string $value
6021 *
6022 * @return array
6023 * @since 1.20
6024 */
6029 }
6031 # (T15500) In both cases (width/height and width only),
6032 # permit trailing "px" for backward compatibility.
6041 }
6043 }
6045 /**
6046 * Lock the current instance of the parser.
6047 *
6048 * This is meant to stop someone from calling the parser
6049 * recursively and messing up all the strip state.
6050 *
6051 * @throws MWException If parser is in a parse
6052 * @return ScopedCallback The lock will be released once the return value goes out of scope.
6053 */
6058 }
6063 } );
6066 }
6068 /**
6069 * Strip outer <p></p> tag from the HTML source of a single paragraph.
6070 *
6071 * Returns original HTML if the <p/> tag has any attributes, if there's no wrapping <p/> tag,
6072 * or if there is more than one <p/> tag in the input HTML.
6073 *
6074 * @param string $html
6075 * @return string
6076 * @since 1.24
6077 */
6083 }
6084 }
6087 }
6089 /**
6090 * Return this parser if it is not doing anything, otherwise
6091 * get a fresh parser. You can use this method by doing
6092 * $myParser = $wgParser->getFreshParser(), or more simply
6093 * $wgParser->getFreshParser()->parse( ... );
6094 * if you're unsure if $wgParser is safe to use.
6095 *
6096 * @since 1.24
6097 * @return Parser A parser object that is not parsing anything
6098 */
6105 }
6106 }
6108 /**
6109 * Set's up the PHP implementation of OOUI for use in this request
6110 * and instructs OutputPage to enable OOUI for itself.
6111 *
6112 * @since 1.26
6113 */
6117 }
6118 }