| blob | 10dfd2660cb3867957542ed0d64a1323562cfa10 |
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 bug 19052
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 (bug 13770)
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 // Bug 56226: 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 }
334 }
336 /**
337 * Clear Parser state
338 *
339 * @private
340 */
344 }
360 # Clear these on every parse, bug 4549
368 ];
377 # Fix cloning
380 }
385 }
387 /**
388 * Convert wikitext to HTML
389 * Do not call this function recursively.
390 *
391 * @param string $text Text we want to parse
392 * @param Title $title
393 * @param ParserOptions $options
394 * @param bool $linestart
395 * @param bool $clearState
396 * @param int $revid Number to pass in {{REVISIONID}}
397 * @return ParserOutput A ParserOutput
398 */
402 ) {
403 /**
404 * First pass--just handle <nowiki> sections, pass the rest off
405 * to internalParse() which does all the real work.
406 */
411 // We use U+007F DELETE to construct strip markers, so we have to make
412 // sure that this character does not occur in the input text.
415 }
423 }
436 }
439 # No more strip!
446 /**
447 * A converted title will be provided in the output object if title and
448 * content conversion are enabled, the article text does not contain
449 * a conversion-suppressing double-underscore tag, and no
450 * {{DISPLAYTITLE:...}} is present. DISPLAYTITLE takes precedence over
451 * automatic link conversion.
452 */
457 ) {
464 }
465 }
467 # Done parsing! Compute runtime adaptive expiry if set
470 # Warn if too many heavyweight parser functions were used
475 );
476 }
478 # Information on include size limits, for the benefit of users who try to skirt them
486 );
487 }
492 );
496 );
499 );
502 );
505 );
508 );
511 );
517 }
527 ) ) {
533 }
537 }
538 }
539 }
540 // Since we're not really outputting HTML, decode the entities and
541 // then re-encode the things that need hiding inside HTML comments.
545 // Sanitize for comment. Note '‐' in the replacement is U+2010,
546 // which looks much like the problematic '-'.
550 // Add on template profiling data
554 } );
560 }
566 }
567 }
579 }
581 /**
582 * Half-parse wikitext to half-parsed HTML. This recursive parser entry point
583 * can be called from an extension tag hook.
584 *
585 * The output of this function IS NOT SAFE PARSED HTML; it is "half-parsed"
586 * instead, which means that lists and links have not been fully parsed yet,
587 * and strip markers are still present.
588 *
589 * Use recursiveTagParseFully() to fully parse wikitext to output-safe HTML.
590 *
591 * Use this function if you're a parser tag hook and you want to parse
592 * wikitext before or after applying additional transformations, and you
593 * intend to *return the result as hook output*, which will cause it to go
594 * through the rest of parsing process automatically.
595 *
596 * If $frame is not provided, then template variables (e.g., {{{1}}}) within
597 * $text are not expanded
598 *
599 * @param string $text Text extension wants to have parsed
600 * @param bool|PPFrame $frame The frame to use for expanding any template variables
601 * @return string UNSAFE half-parsed HTML
602 */
608 }
610 /**
611 * Fully parse wikitext to fully parsed HTML. This recursive parser entry
612 * point can be called from an extension tag hook.
613 *
614 * The output of this function is fully-parsed HTML that is safe for output.
615 * If you're a parser tag hook, you might want to use recursiveTagParse()
616 * instead.
617 *
618 * If $frame is not provided, then template variables (e.g., {{{1}}}) within
619 * $text are not expanded
620 *
621 * @since 1.25
622 *
623 * @param string $text Text extension wants to have parsed
624 * @param bool|PPFrame $frame The frame to use for expanding any template variables
625 * @return string Fully parsed HTML
626 */
631 }
633 /**
634 * Expand templates and variables in the text, producing valid, static wikitext.
635 * Also removes comments.
636 * Do not call this function recursively.
637 * @param string $text
638 * @param Title $title
639 * @param ParserOptions $options
640 * @param int|null $revid
641 * @param bool|PPFrame $frame
642 * @return mixed|string
643 */
646 ) {
651 }
657 }
659 /**
660 * Recursive parser entry point that can be called from an extension tag
661 * hook.
662 *
663 * @param string $text Text to be expanded
664 * @param bool|PPFrame $frame The frame to use for expanding any template variables
665 * @return string
666 * @since 1.19
667 */
672 }
674 /**
675 * Process the wikitext for the "?preload=" feature. (bug 5210)
676 *
677 * "<noinclude>", "<includeonly>" etc. are parsed as for template
678 * transclusion, comments, templates, arguments, tags hooks and parser
679 * functions are untouched.
680 *
681 * @param string $text
682 * @param Title $title
683 * @param ParserOptions $options
684 * @param array $params
685 * @return string
686 */
691 # Parser (re)initialisation
700 }
702 /**
703 * Get a random string
704 *
705 * @return string
706 * @deprecated since 1.26; use wfRandomString() instead.
707 */
711 }
713 /**
714 * Set the current user.
715 * Should only be used when doing pre-save transform.
716 *
717 * @param User|null $user User object or null (to reset)
718 */
721 }
723 /**
724 * Accessor for mUniqPrefix.
725 *
726 * @return string
727 * @deprecated since 1.26; use Parser::MARKER_PREFIX instead.
728 */
732 }
734 /**
735 * Set the context title
736 *
737 * @param Title $t
738 */
742 }
745 # Strip the fragment to avoid various odd effects
749 }
750 }
752 /**
753 * Accessor for the Title object
754 *
755 * @return Title
756 */
759 }
761 /**
762 * Accessor/mutator for the Title object
763 *
764 * @param Title $x Title object or null to just get the current one
765 * @return Title
766 */
769 }
771 /**
772 * Set the output type
773 *
774 * @param int $ot New value
775 */
778 # Shortcut alias
784 ];
785 }
787 /**
788 * Accessor/mutator for the output type
789 *
790 * @param int|null $x New value or null to just get the current one
791 * @return int
792 */
795 }
797 /**
798 * Get the ParserOutput object
799 *
800 * @return ParserOutput
801 */
804 }
806 /**
807 * Get the ParserOptions object
808 *
809 * @return ParserOptions
810 */
813 }
815 /**
816 * Accessor/mutator for the ParserOptions object
817 *
818 * @param ParserOptions $x New value or null to just get the current one
819 * @return ParserOptions Current ParserOptions object
820 */
823 }
825 /**
826 * @return int
827 */
830 }
832 /**
833 * @param int $id
834 */
837 }
839 /**
840 * Get a language object for use in parser functions such as {{FORMATNUM:}}
841 * @return Language
842 */
845 }
847 /**
848 * Get the target language for the content being parsed. This is usually the
849 * language that the content is in.
850 *
851 * @since 1.19
852 *
853 * @throws MWException
854 * @return Language
855 */
865 }
868 }
870 /**
871 * Get the language object for language conversion
872 * @return Language|null
873 */
876 }
878 /**
879 * Get a User object either from $this->mUser, if set, or from the
880 * ParserOptions object otherwise
881 *
882 * @return User
883 */
887 }
889 }
891 /**
892 * Get a preprocessor object
893 *
894 * @return Preprocessor
895 */
900 }
902 }
904 /**
905 * Get a LinkRenderer instance to make links with
906 *
907 * @since 1.28
908 * @return LinkRenderer
909 */
916 );
917 }
920 }
922 /**
923 * Replaces all occurrences of HTML-style comments and the given tags
924 * in the text with a random marker and returns the next text. The output
925 * parameter $matches will be an associative array filled with data in
926 * the form:
927 *
928 * @code
929 * 'UNIQ-xxxxx' => [
930 * 'element',
931 * 'tag content',
932 * [ 'param' => 'x' ],
933 * '<element param="x">tag content</element>' ]
934 * @endcode
935 *
936 * @param array $elements List of element names. Comments are always extracted.
937 * @param string $text Source text string.
938 * @param array $matches Out parameter, Array: extracted tags
939 * @param string|null $uniq_prefix
940 * @return string Stripped text
941 * @since 1.26 The uniq_prefix argument is deprecated.
942 */
943 public static function extractTagsAndParams( $elements, $text, &$matches, $uniq_prefix = null ) {
946 }
959 }
961 # comment
967 # tag
972 }
978 # Empty element tag, <tag />
987 }
991 # No end tag -- let it run out to the end of the text.
997 }
998 }
1004 }
1006 }
1008 /**
1009 * Get a list of strippable XML-like elements
1010 *
1011 * @return array
1012 */
1015 }
1017 /**
1018 * Add an item to the strip state
1019 * Returns the unique tag which must be inserted into the stripped text
1020 * The tag will be replaced with the original text in unstrip()
1021 *
1022 * @param string $text
1023 *
1024 * @return string
1025 */
1031 }
1033 /**
1034 * parse the wiki syntax used to render tables
1035 *
1036 * @private
1037 * @param string $text
1038 * @return string
1039 */
1057 }
1064 # First check if we are starting a new table
1077 # Don't do any of the following
1081 # We are ending a table
1087 }
1091 }
1095 }
1099 # Now we have a table row
1102 # Whats after the tag is now only attributes
1115 }
1119 }
1128 ) {
1129 # This might be cell elements, td, th or captions
1135 }
1137 // Implies both are valid for table headings.
1140 }
1142 # Split up multiple cells on the same line.
1143 # FIXME : This can result in improper nesting of tags processed
1144 # by earlier parser steps.
1149 # Loop through each table cell
1156 }
1161 }
1167 }
1177 }
1181 # A cell could contain both parameters and data
1184 # Bug 553: Note that a '|' inside an invalid link should not
1185 # be mistaken as delimiting cell parameters
1194 }
1198 }
1199 }
1201 }
1203 # Closing open td, tr && table
1207 }
1210 }
1213 }
1216 }
1218 # Remove trailing line-ending (b/c)
1221 }
1223 # special case: don't return empty table
1226 }
1229 }
1231 /**
1232 * Helper function for parse() that transforms wiki markup into half-parsed
1233 * HTML. Only called for $mOutputType == self::OT_HTML.
1234 *
1235 * @private
1236 *
1237 * @param string $text The text to parse
1238 * @param bool $isMain Whether this is being called from the main parse() function
1239 * @param PPFrame|bool $frame A pre-processor frame
1240 *
1241 * @return string
1242 */
1247 # Hook to suspend the parser in this state
1250 }
1252 # if $frame is provided, then use $frame for replacing any variables
1254 # use frame depth to infer how include/noinclude tags should be handled
1255 # depth=0 means this is the top-level document; otherwise it's an included document
1260 }
1264 # if $frame is not provided, then use old-style replaceVariables
1266 }
1274 [],
1276 );
1279 # Tables need to come after variable replacement for things to work
1280 # properly; putting them before other transformations should keep
1281 # exciting things like link expansions from showing up in surprising
1282 # places.
1294 # replaceInternalLinks may sometimes leave behind
1295 # absolute URLs, which have to be masked to hide them from replaceExternalLinks
1302 }
1304 /**
1305 * Helper function for parse() that transforms half-parsed HTML into fully
1306 * parsed HTML.
1307 *
1308 * @param string $text
1309 * @param bool $isMain
1310 * @param bool $linestart
1311 * @return string
1312 */
1318 }
1320 # Clean up special characters, only run once, next-to-last before doBlockLevels
1322 # french spaces, last one Guillemet-left
1323 # only if there is something before the space
1325 # french spaces, Guillemet-right
1328 ];
1335 /**
1336 * The input doesn't get language converted if
1337 * a) It's disabled
1338 * b) Content isn't converted
1339 * c) It's a conversion table
1340 * d) it is an interface message (which is in the user language)
1341 */
1344 ) {
1346 # The position of the convert() call should not be changed. it
1347 # assumes that the links are all replaced and the only thing left
1348 # is the <nowiki> mark.
1350 }
1351 }
1357 }
1367 }
1369 # attempt to sanitize at least some nesting problems
1370 # (bug #2702 and quite a few others)
1372 # ''Something [http://www.cool.com cool''] -->
1373 # <i>Something</i><a href="http://www.cool.com"..><i>cool></i></a>
1376 # fix up an anchor inside another anchor, only
1377 # at least for a single single nested link (bug 3695)
1380 # fix div inside inline elements- doBlockLevels won't wrap a line which
1381 # contains a div, so fix it up here; replace
1382 # div with escaped text
1385 # remove empty italic or bold tag pairs, some
1386 # introduced by rules above
1388 ];
1394 }
1398 }
1401 }
1403 /**
1404 * Replace special strings like "ISBN xxx" and "RFC xxx" with
1405 * magic external links.
1406 *
1407 * DML
1408 * @private
1409 *
1410 * @param string $text
1411 *
1412 * @return string
1413 */
1422 '!(?: # Start cases
1424 (<.*?>) | # m[2]: Skip stuff inside
1427 # m[4]: Post-protocol path
1433 [0-9Xx] # check digit
1437 }
1439 /**
1440 * @throws MWException
1441 * @param array $m
1442 * @return HTML|string
1443 */
1446 # Skip anchor
1449 # Skip HTML element
1452 # Free external link
1455 # RFC or PMID
1459 }
1468 }
1477 }
1480 return Linker::makeExternalLink( $url, "{$keyword} {$id}", true, $cssClass, [], $this->mTitle );
1483 ) {
1484 # ISBN
1492 ] );
1497 [
1500 ]
1501 );
1504 }
1505 }
1507 /**
1508 * Make a free external link, given a user-supplied URL
1509 *
1510 * @param string $url
1511 * @param int $numPostProto
1512 * The number of characters after the protocol.
1513 * @return string HTML
1514 * @private
1515 */
1519 # The characters '<' and '>' (which were escaped by
1520 # removeHTMLtags()) should not be included in
1521 # URLs, per RFC 2396.
1522 # Make terminate a URL as well (bug T84937)
1528 PREG_OFFSET_CAPTURE
1529 ) ) {
1532 }
1534 # Move trailing punctuation to $trail
1536 # If there is no left bracket, then consider right brackets fair game too
1539 }
1543 # Don't break a trailing HTML entity by moving the ; into $trail
1544 # This is in hot code, so use substr_compare to avoid having to
1545 # create a new string object for the comparison
1547 # more optimization: instead of running preg_match with a $
1548 # anchor, which can be slow, do the match on the reversed
1549 # string starting at the desired offset.
1550 # un-reversed regexp is: /&([a-z]+|#x[\da-f]+|#\d+)$/i
1553 }
1554 }
1558 }
1560 # Verify that we still have a real URL after trail removal, and
1561 # not just lone protocol
1564 }
1568 # Is this an external image?
1571 # Not an image, make a link
1576 # Register it in the output object...
1577 # Replace unnecessary URL escape codes with their equivalent characters
1580 }
1582 }
1584 /**
1585 * Parse headers and return html
1586 *
1587 * @private
1588 *
1589 * @param string $text
1590 *
1591 * @return string
1592 */
1597 }
1599 }
1601 /**
1602 * Replace single quotes with HTML markup
1603 * @private
1604 *
1605 * @param string $text
1606 *
1607 * @return string The altered text
1608 */
1614 }
1617 }
1619 /**
1620 * Helper function for doAllQuotes()
1621 *
1622 * @param string $text
1623 *
1624 * @return string
1625 */
1631 }
1633 // First, do some preliminary work. This may shift some apostrophes from
1634 // being mark-up to being text. It also counts the number of occurrences
1635 // of bold and italics mark-ups.
1640 // If there are ever four apostrophes, assume the first is supposed to
1641 // be text, and the remaining three constitute mark-up for bold text.
1642 // (bug 13227: ''''foo'''' turns into ' ''' foo ' ''')
1648 // If there are more than 5 apostrophes in a row, assume they're all
1649 // text except for the last 5.
1650 // (bug 13227: ''''''foo'''''' turns into ' ''''' foo ' ''''')
1654 }
1655 // Count the number of occurrences of bold and italics mark-ups.
1663 }
1664 }
1666 // If there is an odd number of both bold and italics, it is likely
1667 // that one of the bold ones was meant to be an apostrophe followed
1668 // by italics. Which one we cannot know for certain, but it is more
1669 // likely to be one that has a single-letter word before it.
1681 }
1684 // if $firstsingleletterword is set, we don't
1685 // look at the other options, so we can bail early.
1690 }
1691 }
1692 }
1693 }
1695 // If there is a single-letter word, use it!
1700 // If not, but there's a multi-letter word, use that one.
1704 // ... otherwise use the first one that has neither.
1705 // (notice that it is possible for all three to be -1 if, for example,
1706 // there is only one pentuple-apostrophe in the line)
1709 }
1710 }
1712 // Now let's actually convert our apostrophic mush to HTML!
1723 }
1742 }
1759 }
1779 }
1780 }
1781 }
1783 }
1784 // Now close all remaining tags. Notice that the order is important.
1787 }
1790 }
1793 }
1794 // There might be lonely ''''', so make sure we have a buffer
1797 }
1799 }
1801 /**
1802 * Replace external links (REL)
1803 *
1804 * Note: this is all very hackish and the order of execution matters a lot.
1805 * Make sure to run tests/parser/parserTests.php if you change this code.
1806 *
1807 * @private
1808 *
1809 * @param string $text
1810 *
1811 * @throws MWException
1812 * @return string
1813 */
1820 }
1830 # The characters '<' and '>' (which were escaped by
1831 # removeHTMLtags()) should not be included in
1832 # URLs, per RFC 2396.
1837 }
1839 # If the link text is an image URL, replace it with an <img> tag
1840 # This happened by accident in the original parser, but some people used it extensively
1844 }
1848 # Set linktype for CSS - if URL==text, link is essentially free
1851 # No link text, e.g. [http://domain.tld/some.link]
1853 # Autonumber
1858 # Have link text, e.g. [http://domain.tld/some.link text]s
1859 # Check for trail
1861 }
1867 # Use the encoded URL
1868 # This means that users can paste URLs directly into the text
1869 # Funny characters like ö aren't valid in URLs anyway
1870 # This was changed in August 2004
1874 # Register link in the output object.
1875 # Replace unnecessary URL escape codes with the referenced character
1876 # This prevents spammers from hiding links from the filters
1879 }
1882 }
1884 /**
1885 * Get the rel attribute for a particular external link.
1886 *
1887 * @since 1.21
1888 * @param string|bool $url Optional URL, to extract the domain from for rel =>
1889 * nofollow if appropriate
1890 * @param Title $title Optional Title, for wgNoFollowNsExceptions lookups
1891 * @return string|null Rel attribute for $url
1892 */
1898 ) {
1900 }
1902 }
1904 /**
1905 * Get an associative array of additional HTML attributes appropriate for a
1906 * particular external link. This currently may include rel => nofollow
1907 * (depending on configuration, namespace, and the URL's domain) and/or a
1908 * target attribute (depending on configuration).
1909 *
1910 * @param string $url URL to extract the domain from for rel =>
1911 * nofollow if appropriate
1912 * @return array Associative array of HTML attributes
1913 */
1922 // T133507. New windows can navigate parent cross-origin.
1923 // Including noreferrer due to lacking browser
1924 // support of noopener. Eventually noreferrer should be removed.
1927 }
1929 }
1930 }
1933 }
1935 /**
1936 * Replace unusual escape codes in a URL with their equivalent characters
1937 *
1938 * @deprecated since 1.24, use normalizeLinkUrl
1939 * @param string $url
1940 * @return string
1941 */
1945 }
1947 /**
1948 * Replace unusual escape codes in a URL with their equivalent characters
1949 *
1950 * This generally follows the syntax defined in RFC 3986, with special
1951 * consideration for HTTP query strings.
1952 *
1953 * @param string $url
1954 * @return string
1955 */
1957 # First, make sure unsafe characters are encoded
1961 },
1962 $url
1963 );
1968 # Fragment part - 'fragment'
1974 }
1976 # Query part - 'query' minus &=+;
1982 }
1984 # Scheme and path part - 'pchar'
1985 # (we assume no userinfo or encoded colons in the host)
1990 }
1997 # Unescape it
2000 # Leave it escaped, but use uppercase for a-f
2002 }
2003 };
2005 }
2007 /**
2008 * make an image if it's allowed, either through the global
2009 * option, through the exception, or through the on-wiki whitelist
2010 *
2011 * @param string $url
2012 *
2013 * @return string
2014 */
2019 # $imagesfrom could be either a single string or an array of strings, parse out the latter
2026 }
2027 }
2032 }
2036 ) {
2038 # Image found
2040 }
2041 }
2044 ) {
2048 );
2051 # Sanitize the regex fragment, make it case-insensitive, ignore blank entries/comments
2054 }
2056 # Image matches a whitelist entry
2059 }
2060 }
2061 }
2063 }
2065 /**
2066 * Process [[ ]] wikilinks
2067 *
2068 * @param string $s
2069 *
2070 * @return string Processed text
2071 *
2072 * @private
2073 */
2077 }
2079 /**
2080 * Process [[ ]] wikilinks (RIL)
2081 * @param string $s
2082 * @throws MWException
2083 * @return LinkHolderArray
2084 *
2085 * @private
2086 */
2091 # the % is needed to support urlencoded titles as well
2094 # Match a link having the form [[namespace:link|alternate]]trail
2096 # Match cases where there is no "]]", which might still be images
2098 }
2102 # split the entire text string on occurrences of [[
2104 # get the first element (all text up to first [[), and remove the space we added
2113 # Match the end of a line for a word that's not followed by whitespace,
2114 # e.g. in the case of 'The Arab al[[Razi]]', 'al' will be matched
2118 }
2122 }
2131 }
2134 }
2138 // @codingStandardsIgnoreStart Squiz.WhiteSpace.SemicolonSpacing.Incorrect
2139 # Loop for each link
2141 // @codingStandardsIgnoreEnd
2143 # Check for excessive memory usage
2145 # Too big
2146 # Do the existence check, replace the link holders and clear the array
2149 }
2157 }
2158 # first link
2162 }
2163 }
2169 # If we get a ] at the beginning of $m[3] that means we have a link that's something like:
2170 # [[Image:Foo.jpg|[http://example.com desc]]] <- having three ] in a row fucks up,
2171 # the real problem is with the $e1 regex
2172 # See bug 1300.
2173 # Still some problems for cases where the ] is meant to be outside punctuation,
2174 # and no image is in sight. See bug 2095.
2178 ) {
2181 }
2182 # fix up urlencoded title texts
2184 # Should anchors '#' also be rejected?
2186 }
2189 # Invalid, but might be an image with a link in its caption
2194 }
2199 }
2203 # Don't allow internal links to pages containing
2204 # PROTO: where PROTO is a valid URL protocol; these
2205 # should be external links.
2209 }
2211 # Make subpage if necessary
2216 }
2220 # Strip off leading ':'
2222 }
2229 }
2238 # look at the next 'line' to see if we can close it there
2243 }
2246 # the first ]] closes the inner link, the second the image
2252 # if there's exactly one ]] that's fine, we'll keep looking
2255 # if $next_line is invalid too, we need look no further
2258 }
2259 }
2261 # we couldn't find the end of this imageLink, so output it raw
2262 # but don't ignore what might be perfectly normal links in the text we've examined
2265 # note: no $trail, because without an end, there *is* no trail
2267 }
2270 # note: no $trail, because without an end, there *is* no trail
2272 }
2273 }
2279 # Bug 4598 madness. Handle the quotes only if they come from the alternate part
2280 # [[Lista d''e paise d''o munno]] -> <a href="...">Lista d''e paise d''o munno</a>
2281 # [[Criticism of Harry Potter|Criticism of ''Harry Potter'']]
2282 # -> <a href="Criticism of Harry Potter">Criticism of <i>Harry Potter</i></a>
2284 }
2286 # Link not escaped by : , create the various objects
2288 # Interwikis
2293 )
2294 ) {
2295 # Bug 24502: filter duplicates
2299 }
2304 }
2309 # if no parameters were passed, $text
2310 # becomes something like "File:Foo.png",
2311 # which we don't want to pass on to the
2312 # image generator
2315 # recursively parse links inside the image caption
2316 # actually, this will parse them in any other parameters, too,
2317 # but it might be hard to fix that, and it doesn't matter ATM
2320 }
2321 # cloak any absolute URLs inside the image markup, so replaceExternalLinks() won't touch them
2325 }
2333 }
2339 /**
2340 * Strip the whitespace Category links produce, see bug 87
2341 */
2345 }
2346 }
2348 # Self-link checking. For some languages, variants of the title are checked in
2349 # LinkHolderArray::doVariants() to allow batching the existence checks necessary
2350 # for linking to a different variant.
2354 }
2356 # NS_MEDIA is a pseudo-namespace for linking directly to a file
2357 # @todo FIXME: Should do batch file existence checks, see comment below
2359 # Give extensions a chance to select the file revision for us
2364 # Fetch and register the file (file title may be different via hooks)
2366 # Cloak with NOPARSE to avoid replacement in replaceExternalLinks
2370 }
2372 # Some titles, such as valid special pages or files in foreign repos, should
2373 # be shown as bluelinks even though they're not included in the page table
2374 # @todo FIXME: isAlwaysKnown() can be expensive for file links; we should really do
2375 # batch file existence checks for NS_FILE and NS_MEDIA
2380 # Links will be added to the output link list after checking
2382 }
2383 }
2385 }
2387 /**
2388 * Render a forced-blue link inline; protect against double expansion of
2389 * URLs if we're in a mode that prepends full URL prefixes to internal links.
2390 * Since this little disaster has to split off the trail text to avoid
2391 * breaking URLs in the following text without breaking trails on the
2392 * wiki links, it's been made into a horrible function.
2393 *
2394 * @param Title $nt
2395 * @param string $text
2396 * @param string $trail
2397 * @param string $prefix
2398 * @return string HTML-wikitext mix oh yuck
2399 */
2405 }
2409 );
2412 }
2414 /**
2415 * Insert a NOPARSE hacky thing into any inline links in a chunk that's
2416 * going to go through further parsing steps before inline URL expansion.
2417 *
2418 * Not needed quite as much as it used to be since free links are a bit
2419 * more sensible these days. But bracketed links are still an issue.
2420 *
2421 * @param string $text More-or-less HTML
2422 * @return string Less-or-more HTML with NOPARSE bits
2423 */
2427 }
2429 /**
2430 * Return true if subpage links should be expanded on this page.
2431 * @return bool
2432 */
2434 # Some namespaces don't allow subpages
2436 }
2438 /**
2439 * Handle link to subpage if necessary
2440 *
2441 * @param string $target The source of the link
2442 * @param string &$text The link text, modified as necessary
2443 * @return string The full name of the link
2444 * @private
2445 */
2448 }
2450 /**
2451 * Make lists from lines starting with ':', '*', '#', etc. (DBL)
2452 *
2453 * @param string $text
2454 * @param bool $linestart Whether or not this is at the start of a line.
2455 * @private
2456 * @return string The lists rendered as HTML
2457 */
2460 }
2462 /**
2463 * Return value of a magic variable (like PAGENAME)
2464 *
2465 * @private
2466 *
2467 * @param int $index
2468 * @param bool|PPFrame $frame
2469 *
2470 * @throws MWException
2471 * @return string
2472 */
2478 // If no title set, bad things are going to happen
2479 // later. Title should always be set since this
2480 // should only be called in the middle of a parse
2481 // operation (but the unit-tests do funky stuff)
2484 }
2486 /**
2487 * Some of these require message or data lookups and can be
2488 * expensive to check many times.
2489 */
2493 }
2494 }
2539 $value = $pageLang->getMonthAbbreviation( MWTimestamp::getLocalInstance( $ts )->format( 'n' ) );
2573 ) ) );
2583 ) ) );
2591 }
2599 }
2612 # 0 means the page doesn't exist in the database,
2613 # which means the user is previewing a new page.
2614 # The vary-revision flag must be set, because the magic word
2615 # will have a different value once the page is saved.
2618 }
2622 # Let the edit saving system know we should parse the page
2623 # *after* a revision ID has been assigned.
2630 }
2633 # Let the edit saving system know we should parse the page
2634 # *after* a revision ID has been assigned. This is for null edits.
2640 # Let the edit saving system know we should parse the page
2641 # *after* a revision ID has been assigned. This is for null edits.
2647 # Let the edit saving system know we should parse the page
2648 # *after* a revision ID has been assigned. This is for null edits.
2654 # Let the edit saving system know we should parse the page
2655 # *after* a revision ID has been assigned. This is for null edits.
2661 # Let the edit saving system know we should parse the page
2662 # *after* a revision ID has been assigned. This is for null edits.
2668 # Let the edit saving system know we should parse the page
2669 # *after* a revision ID has been assigned. This is for null edits.
2675 # Let the edit saving system know we should parse the page
2676 # *after* a revision ID has been assigned for null edits.
2720 # @bug 4594 PHP5 has it zero padded, PHP4 does not, cast to
2721 # int to remove the padding
2730 );
2739 false
2740 );
2746 # @bug 4594 PHP5 has it zero padded, PHP4 does not, cast to
2747 # int to remove the padding
2808 );
2811 }
2815 }
2818 }
2820 /**
2821 * initialise the magic variables (like CURRENTMONTHNAME) and substitution modifiers
2822 *
2823 * @private
2824 */
2831 }
2833 /**
2834 * Preprocess some wikitext and return the document tree.
2835 * This is the ghost of replace_variables().
2836 *
2837 * @param string $text The text to parse
2838 * @param int $flags Bitwise combination of:
2839 * - self::PTD_FOR_INCLUSION: Handle "<noinclude>" and "<includeonly>" as if the text is being
2840 * included. Default is to assume a direct page view.
2841 *
2842 * The generated DOM tree must depend only on the input text and the flags.
2843 * The DOM tree must be the same in OT_HTML and OT_WIKI mode, to avoid a regression of bug 4899.
2844 *
2845 * Any flag added to the $flags parameter here, or any other parameter liable to cause a
2846 * change in the DOM tree for a given text, must be passed through the section identifier
2847 * in the section edit link and thus back to extractSections().
2848 *
2849 * The output of this function is currently only cached in process memory, but a persistent
2850 * cache may be implemented at a later date which takes further advantage of these strict
2851 * dependency requirements.
2852 *
2853 * @return PPNode
2854 */
2858 }
2860 /**
2861 * Return a three-element array: leading whitespace, string contents, trailing whitespace
2862 *
2863 * @param string $s
2864 *
2865 * @return array
2866 */
2876 }
2878 }
2880 /**
2881 * Replace magic variables, templates, and template arguments
2882 * with the appropriate text. Templates are substituted recursively,
2883 * taking care to avoid infinite loops.
2884 *
2885 * Note that the substitution depends on value of $mOutputType:
2886 * self::OT_WIKI: only {{subst:}} templates
2887 * self::OT_PREPROCESS: templates but not extension tags
2888 * self::OT_HTML: all templates and extension tags
2889 *
2890 * @param string $text The text to transform
2891 * @param bool|PPFrame $frame Object describing the arguments passed to the
2892 * template. Arguments may also be provided as an associative array, as
2893 * was the usual case before MW1.12. Providing arguments this way may be
2894 * useful for extensions wishing to perform variable replacement
2895 * explicitly.
2896 * @param bool $argsOnly Only do argument (triple-brace) expansion, not
2897 * double-brace expansion.
2898 * @return string
2899 */
2901 # Is there any text? Also, Prevent too big inclusions!
2905 }
2913 }
2920 }
2922 /**
2923 * Clean up argument array - refactored in 1.9 so parserfunctions can use it, too.
2924 *
2925 * @param array $args
2926 *
2927 * @return array
2928 */
2941 }
2944 }
2945 }
2946 }
2949 }
2951 /**
2952 * Warn the user when a parser limitation is reached
2953 * Will warn at most once the user per limitation type
2954 *
2955 * The results are shown during preview and run through the Parser (See EditPage.php)
2956 *
2957 * @param string $limitationType Should be one of:
2958 * 'expensive-parserfunction' (corresponding messages:
2959 * 'expensive-parserfunction-warning',
2960 * 'expensive-parserfunction-category')
2961 * 'post-expand-template-argument' (corresponding messages:
2962 * 'post-expand-template-argument-warning',
2963 * 'post-expand-template-argument-category')
2964 * 'post-expand-template-inclusion' (corresponding messages:
2965 * 'post-expand-template-inclusion-warning',
2966 * 'post-expand-template-inclusion-category')
2967 * 'node-count-exceeded' (corresponding messages:
2968 * 'node-count-exceeded-warning',
2969 * 'node-count-exceeded-category')
2970 * 'expansion-depth-exceeded' (corresponding messages:
2971 * 'expansion-depth-exceeded-warning',
2972 * 'expansion-depth-exceeded-category')
2973 * @param string|int|null $current Current value
2974 * @param string|int|null $max Maximum allowed, when an explicit limit has been
2975 * exceeded, provide the values (optional)
2976 */
2978 # does no harm if $current and $max are present but are unnecessary for the message
2979 # Not doing ->inLanguage( $this->mOptions->getUserLangObj() ), since this is shown
2980 # only during preview, and that would split the parser cache unnecessarily.
2985 }
2987 /**
2988 * Return the text of a template, after recursively
2989 * replacing any variables or templates within the template.
2990 *
2991 * @param array $piece The parts of the template
2992 * $piece['title']: the title, i.e. the part before the |
2993 * $piece['parts']: the parameter array
2994 * $piece['lineStart']: whether the brace was at the start of a line
2995 * @param PPFrame $frame The current frame, contains template arguments
2996 * @throws Exception
2997 * @return string The text of the template
2998 */
3001 // Flags
3003 // $text has been filled
3005 // wiki markup in $text should be escaped
3007 // $text is HTML, armour it against wikitext transformation
3009 // Force interwiki transclusion to be done in raw mode not rendered
3011 // $text is a DOM node needing expansion in a child frame
3013 // $text is a DOM node needing expansion in the current frame
3016 # Title object, where $text came from
3019 # $part1 is the bit before the first |, and must contain only title characters.
3020 # Various prefixes will be stripped from it later.
3025 # Original title text preserved for various purposes
3028 # $args is a list of argument nodes, starting from index 0, not including $part1
3029 # @todo FIXME: If piece['parts'] is null then the call to getLength()
3030 # below won't work b/c this $args isn't an object
3035 # SUBST
3039 # Possibilities for substMatch: "subst", "safesubst" or FALSE
3040 # Decide whether to expand template or keep wikitext as-is.
3046 }
3052 }
3053 }
3058 }
3059 }
3061 # Variables
3068 }
3070 }
3071 }
3073 # MSG, MSGNW and RAW
3075 # Check for MSGNW:
3080 # Remove obsolete MSG:
3083 }
3085 # Check for RAW:
3089 }
3090 }
3092 # Parser functions
3101 }
3106 }
3108 # The interface for parser functions allows for extracting
3109 # flags into the local scope. Extract any forwarded flags
3110 # here.
3112 }
3113 }
3115 # Finish mangling title and then check for loops.
3116 # Set $title to a Title object and $titleText to the PDBK
3119 # Split the title into page and subpage
3125 }
3129 # Check for language variants if the template is not found
3132 }
3133 # Do recursion depth check
3141 }
3142 }
3143 }
3145 # Load from database
3152 ) {
3154 // Pass the template arguments as URL parameters.
3155 // "uselang" will have no effect since the Language object
3156 // is forced to the one defined in ParserOptions.
3165 }
3166 }
3168 // Create a new context to execute the special page
3175 // If this page is cached, then we better not be per user.
3177 }
3189 );
3190 }
3191 }
3201 }
3202 }
3204 # If the title is valid but undisplayable, make a link to it
3208 }
3210 # Interwiki transclusion
3216 # Preprocess it like a template
3219 }
3221 }
3223 # Do infinite loop check
3224 # This has to be done after redirect resolution to avoid infinite loops via redirects
3231 }
3232 }
3234 # If we haven't found text to substitute by now, we're done
3235 # Recover the source wikitext and return it
3240 }
3242 }
3244 # Expand DOM-style return values in a child frame
3246 # Clean up argument array
3252 # Expansion is eligible for the empty-frame cache
3255 # Uncached expansion
3257 }
3258 }
3262 }
3266 }
3268 # Replace raw HTML by a placeholder
3272 # Escape nowiki-style return values
3277 ) {
3278 # Bug 529: if the template begins with a table or block-level
3279 # element, it should be treated as beginning a new line.
3280 # This behavior is somewhat controversial.
3282 }
3285 # Error, oversize inclusion
3287 # Make a working, properly escaped link if possible (bug 23588)
3290 # This will probably not be a working link, but at least it may
3291 # provide some hint of where the problem is
3294 }
3298 }
3304 }
3307 }
3309 /**
3310 * Call a parser function and return an array with text and flags.
3311 *
3312 * The returned array will always contain a boolean 'found', indicating
3313 * whether the parser function was found or not. It may also contain the
3314 * following:
3315 * text: string|object, resulting wikitext or PP DOM object
3316 * isHTML: bool, $text is HTML, armour it against wikitext transformation
3317 * isChildObj: bool, $text is a DOM node needing expansion in a child frame
3318 * isLocalObj: bool, $text is a DOM node needing expansion in the current frame
3319 * nowiki: bool, wiki markup in $text should be escaped
3320 *
3321 * @since 1.21
3322 * @param PPFrame $frame The current frame, contains template arguments
3323 * @param string $function Function name
3324 * @param array $args Arguments to the function
3325 * @throws MWException
3326 * @return array
3327 */
3331 # Case sensitive functions
3335 # Case insensitive functions
3341 }
3342 }
3346 # Workaround for PHP bug 35229 and similar
3349 }
3353 # Convert arguments to PPNodes and collect for appending to $allArgs
3360 }
3361 }
3363 # Add a frame parameter, and pass the arguments as an array
3367 # Convert arguments to plain text and append to $allArgs
3375 }
3376 }
3377 }
3381 # The interface for function hooks allows them to return a wikitext
3382 # string or an array containing the string and any flags. This mungs
3383 # things around to match what this method should return.
3388 ];
3392 }
3396 ];
3397 }
3403 }
3406 }
3411 }
3414 }
3416 /**
3417 * Get the semi-parsed DOM representation of a template with a given title,
3418 * and its redirect destination title. Cached.
3419 *
3420 * @param Title $title
3421 *
3422 * @return array
3423 */
3432 }
3435 }
3437 # Cache miss, go to the database
3443 }
3451 }
3454 }
3456 /**
3457 * Fetch the current revision of a given title. Note that the revision
3458 * (and even the title) may not exist in the database, so everything
3459 * contributing to the output of the parser should use this method
3460 * where possible, rather than getting the revisions themselves. This
3461 * method also caches its results, so using it benefits performance.
3462 *
3463 * @since 1.24
3464 * @param Title $title
3465 * @return Revision
3466 */
3471 }
3474 // Defaults to Parser::statelessFetchRevision()
3476 );
3477 }
3479 }
3481 /**
3482 * Wrapper around Revision::newFromTitle to allow passing additional parameters
3483 * without passing them on to it.
3484 *
3485 * @since 1.24
3486 * @param Title $title
3487 * @param Parser|bool $parser
3488 * @return Revision|bool False if missing
3489 */
3497 }
3500 }
3502 /**
3503 * Fetch the unparsed text of a template and register a reference to it.
3504 * @param Title $title
3505 * @return array ( string or false, Title )
3506 */
3508 // Defaults to Parser::statelessFetchTemplate()
3511 // We use U+007F DELETE to distinguish strip markers from regular text.
3515 }
3521 // If we transclude ourselves, the final result
3522 // will change based on the new version of the page
3524 }
3525 }
3526 }
3528 }
3530 /**
3531 * Fetch the unparsed text of a template and register a reference to it.
3532 * @param Title $title
3533 * @return string|bool
3534 */
3537 }
3539 /**
3540 * Static function to get a template
3541 * Can be overridden via ParserOptions::setTemplateCallback().
3542 *
3543 * @param Title $title
3544 * @param bool|Parser $parser
3545 *
3546 * @return array
3547 */
3553 # Loop to fetch the article, with up to 1 redirect
3554 // @codingStandardsIgnoreStart Generic.CodeAnalysis.ForLoopWithTestFunctionCall.NotAllowed
3556 // @codingStandardsIgnoreEnd
3557 # Give extensions a chance to select the revision instead
3568 ];
3570 }
3571 # Get the revision
3578 }
3580 # If there is no current revision, there is no page
3584 }
3591 # We fetched a rev from a different title; register it too...
3596 }
3608 }
3615 }
3620 }
3623 }
3624 # Redirect?
3627 }
3632 }
3634 /**
3635 * Fetch a file and its title and register a reference to it.
3636 * If 'broken' is a key in $options then the file will appear as a broken thumbnail.
3637 * @param Title $title
3638 * @param array $options Array of options to RepoGroup::findFile
3639 * @return File|bool
3640 */
3643 }
3645 /**
3646 * Fetch a file and its title and register a reference to it.
3647 * If 'broken' is a key in $options then the file will appear as a broken thumbnail.
3648 * @param Title $title
3649 * @param array $options Array of options to RepoGroup::findFile
3650 * @return array ( File or false, Title of file )
3651 */
3657 # Register the file as a dependency...
3660 # Update fetched file title
3663 }
3665 }
3667 /**
3668 * Helper function for fetchFileAndTitle.
3669 *
3670 * Also useful if you need to fetch a file but not use it yet,
3671 * for example to get the file's handler.
3672 *
3673 * @param Title $title
3674 * @param array $options Array of options to RepoGroup::findFile
3675 * @return File|bool
3676 */
3684 }
3686 }
3688 /**
3689 * Transclude an interwiki link.
3690 *
3691 * @param Title $title
3692 * @param string $action
3693 *
3694 * @return string
3695 */
3701 }
3707 }
3709 }
3711 /**
3712 * @param string $url
3713 * @return mixed|string
3714 */
3723 }
3730 // Though we failed to fetch the content, this status is useless.
3735 }
3742 ] );
3744 }
3746 /**
3747 * Triple brace replacement -- used for template arguments
3748 * @private
3749 *
3750 * @param array $piece
3751 * @param PPFrame $frame
3752 *
3753 * @return array
3754 */
3767 )
3768 ) {
3769 # No match in frame, use the supplied default
3771 }
3775 }
3778 # No match anywhere
3780 }
3783 }
3788 }
3791 }
3793 /**
3794 * Return the text to be used for a given extension tag.
3795 * This is the ghost of strip().
3796 *
3797 * @param array $params Associative array of parameters:
3798 * name PPNode for the tag name
3799 * attr PPNode for unparsed text where tag attributes are thought to be
3800 * attributes Optional associative array of parsed attributes
3801 * inner Contents of extension element
3802 * noClose Original text did not have a close tag
3803 * @param PPFrame $frame
3804 *
3805 * @throws MWException
3806 * @return string
3807 */
3814 // Probably expansion depth or node count exceeded. Just punt the
3815 // error up.
3817 }
3821 // See above
3823 }
3825 // We can't safely check if the expansion for $content resulted in an
3826 // error, because the content could happen to be the error string
3827 // (T149622).
3839 }
3845 }
3848 # Workaround for PHP bug 35229 and similar
3851 }
3858 }
3864 }
3867 # Extract flags to local scope (to override $markerType)
3872 }
3876 }
3881 }
3882 }
3888 // See above
3890 }
3892 }
3893 }
3903 }
3905 }
3907 /**
3908 * Increment an include size counter
3909 *
3910 * @param string $type The type of expansion
3911 * @param int $size The size of the text
3912 * @return bool False if this inclusion would take it over the maximum, true otherwise
3913 */
3920 }
3921 }
3923 /**
3924 * Increment the expensive function count
3925 *
3926 * @return bool False if the limit has been exceeded
3927 */
3931 }
3933 /**
3934 * Strip double-underscore items like __NOGALLERY__ and __NOTOC__
3935 * Fills $this->mDoubleUnderscores, returns the modified text
3936 *
3937 * @param string $text
3938 *
3939 * @return string
3940 */
3943 # The position of __TOC__ needs to be recorded
3949 # Set a placeholder. At the end we'll fill it in with the TOC.
3952 # Only keep the first one.
3954 }
3956 # Now match and remove the rest of them
3962 }
3965 }
3968 ) {
3970 }
3971 # (bug 8068) Allow control over whether robots index a page.
3972 # @todo FIXME: Bug 14899: __INDEX__ always overrides __NOINDEX__ here! This
3973 # is not desirable, the last one on the page should win.
3977 }
3981 }
3983 # Cache all double underscores in the database
3986 }
3989 }
3991 /**
3992 * @see ParserOutput::addTrackingCategory()
3993 * @param string $msg Message key
3994 * @return bool Whether the addition was successful
3995 */
3998 }
4000 /**
4001 * This function accomplishes several tasks:
4002 * 1) Auto-number headings if that option is enabled
4003 * 2) Add an [edit] link to sections for users who have enabled the option and can edit the page
4004 * 3) Add a Table of contents on the top for users who have enabled the option
4005 * 4) Auto-anchor headings
4006 *
4007 * It loops through all headlines, collects the necessary data, then splits up the
4008 * string and re-inserts the newly formatted headlines.
4009 *
4010 * @param string $text
4011 * @param string $origText Original, untouched wikitext
4012 * @param bool $isMain
4013 * @return mixed|string
4014 * @private
4015 */
4019 # Inhibit editsection links if requested in the page
4025 }
4028 }
4030 # Get all headlines for numbering them and adding funky stuff like [edit]
4031 # links - this is for later, but we need the number of headlines right now
4036 $matches
4037 );
4039 # if there are fewer than 4 headlines in the article, do not show TOC
4040 # unless it's been explicitly enabled.
4044 # Allow user to stipulate that a page should have a "new section"
4045 # link added via __NEWSECTIONLINK__
4048 }
4050 # Allow user to remove the "new section"
4051 # link via __NONEWSECTIONLINK__
4054 }
4056 # if the string __FORCETOC__ (not case-sensitive) occurs in the HTML,
4057 # override above conditions and always show TOC above first header
4061 }
4063 # headline counter
4067 # Ugh .. the TOC should have neat indentation levels which can be
4068 # passed to the skin functions. These are determined here
4102 }
4106 }
4110 # Increase TOC level
4117 }
4119 # Decrease TOC level, find level to jump to
4123 # Found last matching level
4127 # Found first matching level below current level
4130 }
4131 }
4134 }
4137 # Unindent only if the previous toc level was shown :p
4142 }
4143 }
4145 # No change in level, end TOC line
4148 }
4149 }
4153 # count number of headlines for each level
4160 }
4163 }
4164 }
4166 # The safe header is a version of the header text safe to use for links
4168 # Remove link placeholders by the link text.
4169 # <!--LINK number-->
4170 # turns into
4171 # link text with suffix
4172 # Do this before unstrip since link text can contain strip markers
4175 # Avoid insertion of weird stuff like <math> by expanding the relevant sections
4178 # Strip out HTML (first regex removes any tag not allowed)
4179 # Allowed tags are:
4180 # * <sup> and <sub> (bug 8393)
4181 # * <i> (bug 26375)
4182 # * <b> (r105284)
4183 # * <bdi> (bug 72884)
4184 # * <span dir="rtl"> and <span dir="ltr"> (bug 35167)
4185 # * <s> and <strike> (T35715)
4186 # We strip any parameter from accepted tags (second regex), except dir="rtl|ltr" from <span>,
4187 # to allow setting directionality in toc items.
4189 [
4191 '#<(/?(?:span(?: dir="(?:rtl|ltr)")?|sup|sub|bdi|i|b|s|strike))(?: .*?)?>#'
4192 ],
4194 $safeHeadline
4195 );
4197 # Strip '<span></span>', which is the result from the above if
4198 # <span id="foo"></span> is used to produce an additional anchor
4199 # for a section.
4204 # For the anchor, strip out HTML-y stuff period
4208 # Save headline for section edit hint before it's escaped
4212 # For reverse compatibility, provide an id that's
4213 # HTML4-compatible, like we used to.
4214 # It may be worth noting, academically, that it's possible for
4215 # the legacy anchor to conflict with a non-legacy headline
4216 # anchor on the page. In this case likely the "correct" thing
4217 # would be to either drop the legacy anchors or make sure
4218 # they're numbered first. However, this would require people
4219 # to type in section names like "abc_.D7.93.D7.90.D7.A4"
4220 # manually, so let's not bother worrying about it.
4226 # No reason to have both (in fact, we can't)
4228 }
4233 }
4235 # HTML names must be case-insensitively unique (bug 10721).
4236 # This does not apply to Unicode characters per
4237 # https://www.w3.org/TR/html5/infrastructure.html#case-sensitivity-and-string-comparison
4238 # @todo FIXME: We may be changing them depending on the current locale.
4244 }
4246 # Create the anchor for linking from the TOC to the section
4250 // @codingStandardsIgnoreStart
4252 // @codingStandardsIgnoreEnd
4257 }
4259 // @codingStandardsIgnoreStart
4261 // @codingStandardsIgnoreEnd
4266 }
4268 # Don't number the heading if it is the only one (looks silly)
4270 # the two are different if the line contains a link
4274 $numbering
4276 }
4281 }
4283 # Add the section to the section tree
4284 # Find the DOM node for this header
4291 }
4292 }
4296 }
4306 ];
4308 # give headline the correct <h#> tag
4310 // Output edit section links as markers with styles that can be customized by skins
4312 # Put a T flag in the section identifier, to indicate to extractSections()
4313 # that sections inside <includeonly> should be counted.
4321 }
4322 // We use a bit of pesudo-xml for editsection markers. The
4323 // language converter is run later on. Using a UNIQ style marker
4324 // leads to the converter screwing up the tokens when it
4325 // converts stuff. And trying to insert strip tags fails too. At
4326 // this point all real inputted tags have already been escaped,
4327 // so we don't have to worry about a user trying to input one of
4328 // these markers directly. We use a page and section attribute
4329 // to stop the language converter from converting these
4330 // important bits of data, but put the headline hint inside a
4331 // content block because the language converter is supposed to
4332 // be able to convert that piece of data.
4333 // Gets replaced with html in ParserOutput::getText
4340 }
4343 }
4349 }
4353 # Never ever show TOC if no headers
4356 }
4361 }
4366 }
4370 }
4372 # split up and insert constructed headlines
4376 // build an array of document sections
4379 // $head is zero-based, sections aren't.
4384 }
4386 /**
4387 * Send a hook, one per section.
4388 * The idea here is to be able to make section-level DIVs, but to do so in a
4389 * lower-impact, more correct way than r50769
4390 *
4391 * $this : caller
4392 * $section : the section number
4393 * &$sectionContent : ref to the content of the section
4394 * $showEditLinks : boolean describing whether this section has an edit link
4395 */
4399 }
4402 // append the TOC at the beginning
4403 // Top anchor now in skin
4405 }
4413 }
4414 }
4416 /**
4417 * Transform wiki markup when saving a page by doing "\r\n" -> "\n"
4418 * conversion, substituting signatures, {{subst:}} templates, etc.
4419 *
4420 * @param string $text The text to transform
4421 * @param Title $title The Title object for the current article
4422 * @param User $user The User object describing the current user
4423 * @param ParserOptions $options Parsing options
4424 * @param bool $clearState Whether to clear the parser state first
4425 * @return string The altered wiki markup
4426 */
4429 ) {
4432 }
4436 // We still normalize line endings for backwards-compatibility
4437 // with other code that just calls PST, but this should already
4438 // be handled in TextContent subclasses
4443 }
4449 }
4451 /**
4452 * Pre-save transform helper function
4453 *
4454 * @param string $text
4455 * @param User $user
4456 *
4457 * @return string
4458 */
4462 # Note: This is the timestamp saved as hardcoded wikitext to
4463 # the database, we use $wgContLang here in order to give
4464 # everyone the same signature and use the default one rather
4465 # than the one selected in each user's preferences.
4466 # (see also bug 12815)
4474 # Variable replacement
4475 # Because mOutputType is OT_WIKI, this will only process {{subst:xxx}} type tags
4478 # This works almost by chance, as the replaceVariables are done before the getUserSig(),
4479 # which may corrupt this parser instance via its wfMessage()->text() call-
4481 # Signatures
4487 ] );
4489 # Context links ("pipe tricks"): [[|name]] and [[name (context)|]]
4493 // [[ns:page (context)|]]
4495 // [[ns:page（context）|]] (double-width brackets, added in r40257)
4497 // [[ns:page (context), context|]] (using either single or double-width comma)
4499 // [[|page]] (reverse pipe trick: add context from page title)
4502 # try $p1 first, to turn "[[A, B (C)|]]" into "[[A, B (C)|A, B]]"
4514 # if there's no context, don't bother duplicating the title
4516 }
4519 }
4521 /**
4522 * Fetch the user's signature text, if any, and normalize to
4523 * validated, ready-to-insert wikitext.
4524 * If you have pre-fetched the nickname or the fancySig option, you can
4525 * specify them here to save a database query.
4526 * Do not reuse this parser instance after calling getUserSig(),
4527 * as it may have changed if it's the $wgParser.
4528 *
4529 * @param User $user
4530 * @param string|bool $nickname Nickname to use or false to use user's default nickname
4531 * @param bool|null $fancySig whether the nicknname is the complete signature
4532 * or null to use default value
4533 * @return string
4534 */
4540 # If not given, retrieve from the user object.
4543 }
4547 }
4555 # Sig. might contain markup; validate this
4557 # Validated; clean up (if needed) and return it
4560 # Failed to validate; fall back to the default
4563 }
4564 }
4566 # Make sure nickname doesnt get a sig in a sig
4569 # If we're still here, make it a link to the user page
4576 }
4578 /**
4579 * Check that the user's signature contains no bad XML
4580 *
4581 * @param string $text
4582 * @return string|bool An expanded string, or false if invalid.
4583 */
4586 }
4588 /**
4589 * Clean up signature text
4590 *
4591 * 1) Strip 3, 4 or 5 tildes out of signatures @see cleanSigInSig
4592 * 2) Substitute all transclusions
4593 *
4594 * @param string $text
4595 * @param bool $parsing Whether we're cleaning (preferences save) or parsing
4596 * @return string Signature text
4597 */
4603 }
4605 # Option to disable this feature
4608 }
4610 # @todo FIXME: Regex doesn't respect extension tags or nowiki
4611 # => Move this logic to braceSubstitution()
4613 $substRegex = '/\{\{(?!(?:' . $substWord->getBaseRegex() . '))/x' . $substWord->getRegexCase();
4624 }
4627 }
4629 /**
4630 * Strip 3, 4 or 5 tildes out of signatures.
4631 *
4632 * @param string $text
4633 * @return string Signature text with /~{3,5}/ removed
4634 */
4638 }
4640 /**
4641 * Set up some variables which are usually set up in parse()
4642 * so that an external function can call some class members with confidence
4643 *
4644 * @param Title|null $title
4645 * @param ParserOptions $options
4646 * @param int $outputType
4647 * @param bool $clearState
4648 */
4651 ) {
4653 }
4655 /**
4656 * @param Title|null $title
4657 * @param ParserOptions $options
4658 * @param int $outputType
4659 * @param bool $clearState
4660 */
4663 ) {
4669 }
4670 }
4672 /**
4673 * Wrapper for preprocess()
4674 *
4675 * @param string $text The text to preprocess
4676 * @param ParserOptions $options Options
4677 * @param Title|null $title Title object or null to use $wgTitle
4678 * @return string
4679 */
4683 # Guard against infinite recursion
4686 }
4692 }
4698 }
4700 /**
4701 * Create an HTML-style tag, e.g. "<yourtag>special text</yourtag>"
4702 * The callback should have the following form:
4703 * function myParserHook( $text, $params, $parser, $frame ) { ... }
4704 *
4705 * Transform and return $text. Use $parser for any required context, e.g. use
4706 * $parser->getTitle() and $parser->getOptions() not $wgTitle or $wgOut->mParserOptions
4707 *
4708 * Hooks may return extended information by returning an array, of which the
4709 * first numbered element (index 0) must be the return string, and all other
4710 * entries are extracted into local variables within an internal function
4711 * in the Parser class.
4712 *
4713 * This interface (introduced r61913) appears to be undocumented, but
4714 * 'markerType' is used by some core tag hooks to override which strip
4715 * array their results are placed in. **Use great caution if attempting
4716 * this interface, as it is not documented and injudicious use could smash
4717 * private variables.**
4718 *
4719 * @param string $tag The tag to use, e.g. 'hook' for "<hook>"
4720 * @param callable $callback The callback function (and object) to use for the tag
4721 * @throws MWException
4722 * @return callable|null The old value of the mTagHooks array associated with the hook
4723 */
4728 }
4733 }
4736 }
4738 /**
4739 * As setHook(), but letting the contents be parsed.
4740 *
4741 * Transparent tag hooks are like regular XML-style tag hooks, except they
4742 * operate late in the transformation sequence, on HTML instead of wikitext.
4743 *
4744 * This is probably obsoleted by things dealing with parser frames?
4745 * The only extension currently using it is geoserver.
4746 *
4747 * @since 1.10
4748 * @todo better document or deprecate this
4749 *
4750 * @param string $tag The tag to use, e.g. 'hook' for "<hook>"
4751 * @param callable $callback The callback function (and object) to use for the tag
4752 * @throws MWException
4753 * @return callable|null The old value of the mTagHooks array associated with the hook
4754 */
4759 }
4760 $oldVal = isset( $this->mTransparentTagHooks[$tag] ) ? $this->mTransparentTagHooks[$tag] : null;
4764 }
4766 /**
4767 * Remove all tag hooks
4768 */
4773 }
4775 /**
4776 * Create a function, e.g. {{sum:1|2|3}}
4777 * The callback function should have the form:
4778 * function myParserFunction( &$parser, $arg1, $arg2, $arg3 ) { ... }
4779 *
4780 * Or with Parser::SFH_OBJECT_ARGS:
4781 * function myParserFunction( $parser, $frame, $args ) { ... }
4782 *
4783 * The callback may either return the text result of the function, or an array with the text
4784 * in element 0, and a number of flags in the other elements. The names of the flags are
4785 * specified in the keys. Valid flags are:
4786 * found The text returned is valid, stop processing the template. This
4787 * is on by default.
4788 * nowiki Wiki markup in the return value should be escaped
4789 * isHTML The returned text is HTML, armour it against wikitext transformation
4790 *
4791 * @param string $id The magic word ID
4792 * @param callable $callback The callback function (and object) to use
4793 * @param int $flags A combination of the following flags:
4794 * Parser::SFH_NO_HASH No leading hash, i.e. {{plural:...}} instead of {{#if:...}}
4795 *
4796 * Parser::SFH_OBJECT_ARGS Pass the template arguments as PPNode objects instead of text.
4797 * This allows for conditional expansion of the parse tree, allowing you to eliminate dead
4798 * branches and thus speed up parsing. It is also possible to analyse the parse tree of
4799 * the arguments, and to control the way they are expanded.
4800 *
4801 * The $frame parameter is a PPFrame. This can be used to produce expanded text from the
4802 * arguments, for instance:
4803 * $text = isset( $args[0] ) ? $frame->expand( $args[0] ) : '';
4804 *
4805 * For technical reasons, $args[0] is pre-expanded and will be a string. This may change in
4806 * future versions. Please call $frame->expand() on it anyway so that your code keeps
4807 * working if/when this is changed.
4808 *
4809 * If you want whitespace to be trimmed from $args, you need to do it yourself, post-
4810 * expansion.
4811 *
4812 * Please read the documentation in includes/parser/Preprocessor.php for more information
4813 * about the methods available in PPFrame and PPNode.
4814 *
4815 * @throws MWException
4816 * @return string|callable The old callback function for this name, if any
4817 */
4824 # Add to function cache
4828 }
4834 # Case
4837 }
4838 # Add leading hash
4841 }
4842 # Remove trailing colon
4845 }
4847 }
4849 }
4851 /**
4852 * Get all registered function hook identifiers
4853 *
4854 * @return array
4855 */
4858 }
4860 /**
4861 * Create a tag function, e.g. "<test>some stuff</test>".
4862 * Unlike tag hooks, tag functions are parsed at preprocessor level.
4863 * Unlike parser functions, their content is not preprocessed.
4864 * @param string $tag
4865 * @param callable $callback
4866 * @param int $flags
4867 * @throws MWException
4868 * @return null
4869 */
4874 }
4881 }
4884 }
4886 /**
4887 * Replace "<!--LINK-->" link placeholders with actual links, in the buffer
4888 * Placeholders created in Linker::link()
4889 *
4890 * @param string $text
4891 * @param int $options
4892 */
4895 }
4897 /**
4898 * Replace "<!--LINK-->" link placeholders with plain text of links
4899 * (not HTML-formatted).
4900 *
4901 * @param string $text
4902 * @return string
4903 */
4906 }
4908 /**
4909 * Renders an image gallery from a text with one line per image.
4910 * text labels may be given by using |-style alternative text. E.g.
4911 * Image:one.jpg|The number "1"
4912 * Image:tree.jpg|A tree
4913 * given as text will return the HTML of a gallery with two images,
4914 * labeled 'The number "1"' and
4915 * 'A tree'.
4916 *
4917 * @param string $text
4918 * @param array $params
4919 * @return string HTML
4920 */
4926 }
4931 // If invalid type set, fallback to default.
4933 }
4946 }
4952 }
4955 }
4958 }
4961 }
4968 # match lines like these:
4969 # Image:someimage.jpg|This is some image
4972 # Skip empty lines
4975 }
4979 }
4982 # Bogus title. Ignore these so we don't bomb out later.
4984 }
4986 # We need to get what handler the file uses, to figure out parameters.
4987 # Note, a hook can overide the file name, and chose an entirely different
4988 # file (which potentially could be of a different type and have different handler).
4993 # Don't register it now, as ImageGallery does that later.
5000 ];
5003 // We don't want people to specify per-image widths.
5004 // Additionally the width parameter would need special casing anyhow.
5006 }
5015 // look for an |alt= definition while trying not to break existing
5016 // captions with multiple pipes (|) in it, until a more sensible grammar
5017 // is defined for images in galleries
5019 // FIXME: Doing recursiveTagParse at this stage, and the trim before
5020 // splitting on '|' is a bit odd, and different from makeImage.
5038 // check to see if link matches an absolute url, if not then it must be a wiki link.
5045 }
5046 }
5049 // Must be a handler specific parameter.
5053 // Guess not, consider it as caption.
5056 }
5057 }
5060 // Last pipe wins.
5062 }
5063 }
5064 // Remove the pipe.
5066 }
5069 }
5073 }
5075 /**
5076 * @param MediaHandler $handler
5077 * @return array
5078 */
5084 }
5086 # Initialise static lists
5093 ];
5101 }
5102 }
5103 }
5105 # Add handler params
5111 }
5112 }
5115 }
5117 }
5119 /**
5120 * Parse image options text and use it to make an image
5121 *
5122 * @param Title $title
5123 * @param string $options
5124 * @param LinkHolderArray|bool $holders
5125 * @return string HTML
5126 */
5128 # Check if the options text is of the form "options|alt text"
5129 # Options are:
5130 # * thumbnail make a thumbnail with enlarge-icon and caption, alignment depends on lang
5131 # * left no resizing, just left align. label is used for alt= only
5132 # * right same, but right aligned
5133 # * none same, but not aligned
5134 # * ___px scale to ___ pixels width, no aligning. e.g. use in taxobox
5135 # * center center the image
5136 # * frame Keep original image size, no magnify-button.
5137 # * framed Same as "frame"
5138 # * frameless like 'thumb' but without a frame. Keeps user preferences for width
5139 # * upright reduce width for upright images, rounded to full __0 px
5140 # * border draw a 1px border around the image
5141 # * alt Text for HTML alt attribute (defaults to empty)
5142 # * class Set a class for img node
5143 # * link Set the target of the image link. Can be external, interwiki, or local
5144 # vertical-align values (no % or length right now):
5145 # * baseline
5146 # * sub
5147 # * super
5148 # * top
5149 # * text-top
5150 # * middle
5151 # * bottom
5152 # * text-bottom
5156 # Give extensions a chance to select the file revision for us
5161 # Fetch and register the file (file title may be different via hooks)
5164 # Get parameter map
5171 }
5173 # Process the input parameters
5185 # Special case; width and height come in one variable together
5193 }
5194 }
5200 }
5201 }
5202 # else no validation -- bug 13436
5205 # Validate handler parameter
5208 # Validate internal parameters
5213 # @todo FIXME: Possibly check validity here for
5214 # manualthumb? downstream behavior seems odd with
5215 # missing manual thumbs.
5233 }
5235 }
5243 }
5244 }
5249 // use first appearing option, discard others.
5254 # Most other things appear to be empty or numeric...
5256 }
5257 }
5261 }
5262 }
5263 }
5266 }
5267 }
5269 # Process alignment parameters
5272 }
5275 }
5279 # Will the image be presented in a frame, with the caption below?
5285 # In the old days, [[Image:Foo|text...]] would set alt text. Later it
5286 # came to also set the caption, ordinary text after the image -- which
5287 # makes no sense, because that just repeats the text multiple times in
5288 # screen readers. It *also* came to set the title attribute.
5289 # Now that we have an alt attribute, we should not set the alt text to
5290 # equal the caption: that's worse than useless, it just repeats the
5291 # text. This is the framed/thumbnail case. If there's no caption, we
5292 # use the unnamed parameter for alt text as well, just for the time be-
5293 # ing, if the unnamed param is set and the alt param is not.
5294 # For the future, we need to figure out if we want to tweak this more,
5295 # e.g., introducing a title= parameter for the title; ignoring the un-
5296 # named parameter entirely for images without a caption; adding an ex-
5297 # plicit caption= parameter and preserving the old magic unnamed para-
5298 # meter for BC; ...
5301 # No caption or alt text, add the filename as the alt text so
5302 # that screen readers at least get some description of the image
5304 }
5305 # Do not set $params['frame']['title'] because tooltips don't make sense
5306 # for framed images
5309 # No alt text, use the "caption" for the alt text
5313 # No caption, fall back to using the filename for the
5314 # alt text
5316 }
5317 }
5318 # Use the "caption" for the tooltip text
5320 }
5324 # Linker does the rest
5329 # Give the handler a chance to modify the parser object
5332 }
5335 }
5337 /**
5338 * @param string $caption
5339 * @param LinkHolderArray|bool $holders
5340 * @return mixed|string
5341 */
5343 # Strip bad stuff out of the title (tooltip). We can't just use
5344 # replaceLinkHoldersText() here, because if this function is called
5345 # from replaceInternalLinks2(), mLinkHolders won't be up-to-date.
5350 }
5352 # make sure there are no placeholders in thumbnail attributes
5353 # that are later expanded to html- so expand them now and
5354 # remove the tags
5359 }
5361 /**
5362 * Set a flag in the output object indicating that the content is dynamic and
5363 * shouldn't be cached.
5364 * @deprecated since 1.28; use getOutput()->updateCacheExpiry()
5365 */
5371 }
5373 }
5375 /**
5376 * Callback from the Sanitizer for expanding items found in HTML attribute
5377 * values, so they can be safely tested and escaped.
5378 *
5379 * @param string $text
5380 * @param bool|PPFrame $frame
5381 * @return string
5382 */
5387 }
5389 /**
5390 * Accessor
5391 *
5392 * @return array
5393 */
5399 );
5400 }
5402 /**
5403 * Replace transparent tags in $text with the values given by the callbacks.
5404 *
5405 * Transparent tag hooks are like regular XML-style tag hooks, except they
5406 * operate late in the transformation sequence, on HTML instead of wikitext.
5407 *
5408 * @param string $text
5409 *
5410 * @return string
5411 */
5425 );
5428 }
5430 }
5432 }
5434 /**
5435 * Break wikitext input into sections, and either pull or replace
5436 * some particular section's text.
5437 *
5438 * External callers should use the getSection and replaceSection methods.
5439 *
5440 * @param string $text Page wikitext
5441 * @param string|number $sectionId A section identifier string of the form:
5442 * "<flag1> - <flag2> - ... - <section number>"
5443 *
5444 * Currently the only recognised flag is "T", which means the target section number
5445 * was derived during a template inclusion parse, in other words this is a template
5446 * section edit link. If no flags are given, it was an ordinary section edit link.
5447 * This flag is required to avoid a section numbering mismatch when a section is
5448 * enclosed by "<includeonly>" (bug 6563).
5449 *
5450 * The section number 0 pulls the text before the first heading; other numbers will
5451 * pull the given section along with its lower-level subsections. If the section is
5452 * not found, $mode=get will return $newtext, and $mode=replace will return $text.
5453 *
5454 * Section 0 is always considered to exist, even if it only contains the empty
5455 * string. If $text is the empty string and section 0 is replaced, $newText is
5456 * returned.
5457 *
5458 * @param string $mode One of "get" or "replace"
5459 * @param string $newText Replacement text for section data.
5460 * @return string For "get", the extracted section text.
5461 * for "replace", the whole page with the section replaced.
5462 */
5471 # Process section extraction flags
5478 }
5479 }
5481 # Check for empty input
5483 # Only sections 0 and T-0 exist in an empty document
5489 }
5495 }
5496 }
5497 }
5499 # Preprocess the text
5502 # <h> nodes indicate section breaks
5503 # They can only occur at the top level, so we can find them by iterating the root's children
5506 # Find the target section
5508 # Section zero doesn't nest, level=big
5517 }
5518 }
5521 }
5523 }
5524 }
5527 # Not found
5532 }
5533 }
5535 # Find the end of the section, including nested sections
5542 }
5543 }
5546 }
5550 # Write out the remainder (in replace mode only)
5552 # Output the replacement text
5553 # Add two newlines on -- trailing whitespace in $newText is conventionally
5554 # stripped by the editor, so we need both newlines to restore the paragraph gap
5555 # Only add trailing whitespace if there is newText
5558 }
5563 }
5564 }
5567 # Re-insert stripped tags
5569 }
5572 }
5574 /**
5575 * This function returns the text of a section, specified by a number ($section).
5576 * A section is text under a heading like == Heading == or \<h1\>Heading\</h1\>, or
5577 * the first section before any such heading (section 0).
5578 *
5579 * If a section contains subsections, these are also returned.
5580 *
5581 * @param string $text Text to look in
5582 * @param string|number $sectionId Section identifier as a number or string
5583 * (e.g. 0, 1 or 'T-1').
5584 * @param string $defaultText Default to return if section is not found
5585 *
5586 * @return string Text of the requested section
5587 */
5590 }
5592 /**
5593 * This function returns $oldtext after the content of the section
5594 * specified by $section has been replaced with $text. If the target
5595 * section does not exist, $oldtext is returned unchanged.
5596 *
5597 * @param string $oldText Former text of the article
5598 * @param string|number $sectionId Section identifier as a number or string
5599 * (e.g. 0, 1 or 'T-1').
5600 * @param string $newText Replacing text
5601 *
5602 * @return string Modified text
5603 */
5606 }
5608 /**
5609 * Get the ID of the revision we are parsing
5610 *
5611 * @return int|null
5612 */
5615 }
5617 /**
5618 * Get the revision object for $this->mRevisionId
5619 *
5620 * @return Revision|null Either a Revision object or null
5621 * @since 1.23 (public since 1.23)
5622 */
5626 }
5629 }
5633 );
5635 # If the parse is for a new revision, then the callback should have
5636 # already been set to force the object and should match mRevisionId.
5637 # If not, try to fetch by mRevisionId for sanity.
5640 }
5645 }
5647 /**
5648 * Get the timestamp associated with the current revision, adjusted for
5649 * the default server-local timestamp
5650 * @return string
5651 */
5659 # The cryptic '' timezone parameter tells to use the site-default
5660 # timezone offset instead of the user settings.
5661 # Since this value will be saved into the parser cache, served
5662 # to other users, and potentially even used inside links and such,
5663 # it needs to be consistent for all visitors.
5666 }
5668 }
5670 /**
5671 * Get the name of the user that edited the last revision
5672 *
5673 * @return string User name
5674 */
5679 # if this template is subst: the revision id will be blank,
5680 # so just use the current user's name
5685 }
5686 }
5688 }
5690 /**
5691 * Get the size of the revision
5692 *
5693 * @return int|null Revision size
5694 */
5699 # if this variable is subst: the revision id will be blank,
5700 # so just use the parser input size, because the own substituation
5701 # will change the size.
5706 }
5707 }
5709 }
5711 /**
5712 * Mutator for $mDefaultSort
5713 *
5714 * @param string $sort New value
5715 */
5719 }
5721 /**
5722 * Accessor for $mDefaultSort
5723 * Will use the empty string if none is set.
5724 *
5725 * This value is treated as a prefix, so the
5726 * empty string is equivalent to sorting by
5727 * page name.
5728 *
5729 * @return string
5730 */
5736 }
5737 }
5739 /**
5740 * Accessor for $mDefaultSort
5741 * Unlike getDefaultSort(), will return false if none is set
5742 *
5743 * @return string|bool
5744 */
5747 }
5749 /**
5750 * Try to guess the section anchor name based on a wikitext fragment
5751 * presumably extracted from a heading, for example "Header" from
5752 * "== Header ==".
5753 *
5754 * @param string $text
5755 *
5756 * @return string
5757 */
5759 # Strip out wikitext links(they break the anchor)
5763 }
5765 /**
5766 * Same as guessSectionNameFromWikiText(), but produces legacy anchors
5767 * instead. For use in redirects, since IE6 interprets Redirect: headers
5768 * as something other than UTF-8 (apparently?), resulting in breakage.
5769 *
5770 * @param string $text The section name
5771 * @return string An anchor
5772 */
5774 # Strip out wikitext links(they break the anchor)
5778 }
5780 /**
5781 * Strips a text string of wikitext for use in a section anchor
5782 *
5783 * Accepts a text string and then removes all wikitext from the
5784 * string and leaves only the resultant text (i.e. the result of
5785 * [[User:WikiSysop|Sysop]] would be "Sysop" and the result of
5786 * [[User:WikiSysop]] would be "User:WikiSysop") - this is intended
5787 * to create valid section anchors by mimicing the output of the
5788 * parser when headings are parsed.
5789 *
5790 * @param string $text Text string to be stripped of wikitext
5791 * for use in a Section anchor
5792 * @return string Filtered text string
5793 */
5795 # Strip internal link markup
5799 # Strip external link markup
5800 # @todo FIXME: Not tolerant to blank link text
5801 # I.E. [https://www.mediawiki.org] will render as [1] or something depending
5802 # on how many empty links there are on the page - need to figure that out.
5803 $text = preg_replace( '/\[(?i:' . $this->mUrlProtocols . ')([^ ]+?) ([^[]+)\]/', '$2', $text );
5805 # Parse wikitext quotes (italics & bold)
5808 # Strip HTML tags
5811 }
5813 /**
5814 * strip/replaceVariables/unstrip for preprocessor regression testing
5815 *
5816 * @param string $text
5817 * @param Title $title
5818 * @param ParserOptions $options
5819 * @param int $outputType
5820 *
5821 * @return string
5822 */
5825 ) {
5833 }
5835 /**
5836 * @param string $text
5837 * @param Title $title
5838 * @param ParserOptions $options
5839 * @return string
5840 */
5843 }
5845 /**
5846 * @param string $text
5847 * @param Title $title
5848 * @param ParserOptions $options
5849 * @return string
5850 */
5853 }
5855 /**
5856 * Call a callback function on all regions of the given text that are not
5857 * inside strip markers, and replace those regions with the return value
5858 * of the callback. For example, with input:
5859 *
5860 * aaa<MARKER>bbb
5861 *
5862 * This will call the callback function twice, with 'aaa' and 'bbb'. Those
5863 * two strings will be replaced with the value returned by the callback in
5864 * each case.
5865 *
5866 * @param string $s
5867 * @param callable $callback
5868 *
5869 * @return string
5870 */
5889 }
5890 }
5891 }
5893 }
5895 /**
5896 * Remove any strip markers found in the given text.
5897 *
5898 * @param string $text Input string
5899 * @return string
5900 */
5903 }
5905 /**
5906 * Save the parser state required to convert the given half-parsed text to
5907 * HTML. "Half-parsed" in this context means the output of
5908 * recursiveTagParse() or internalParse(). This output has strip markers
5909 * from replaceVariables (extensionSubstitution() etc.), and link
5910 * placeholders from replaceLinkHolders().
5911 *
5912 * Returns an array which can be serialized and stored persistently. This
5913 * array can later be loaded into another parser instance with
5914 * unserializeHalfParsedText(). The text can then be safely incorporated into
5915 * the return value of a parser hook.
5916 *
5917 * @param string $text
5918 *
5919 * @return array
5920 */
5927 ];
5929 }
5931 /**
5932 * Load the parser state given in the $data array, which is assumed to
5933 * have been generated by serializeHalfParsedText(). The text contents is
5934 * extracted from the array, and its markers are transformed into markers
5935 * appropriate for the current Parser instance. This transformed text is
5936 * returned, and can be safely included in the return value of a parser
5937 * hook.
5938 *
5939 * If the $data array has been stored persistently, the caller should first
5940 * check whether it is still valid, by calling isValidHalfParsedText().
5941 *
5942 * @param array $data Serialized data
5943 * @throws MWException
5944 * @return string
5945 */
5949 }
5951 # First, extract the strip state.
5955 # Now renumber links
5958 # Should be good to go.
5960 }
5962 /**
5963 * Returns true if the given array, presumed to be generated by
5964 * serializeHalfParsedText(), is compatible with the current version of the
5965 * parser.
5966 *
5967 * @param array $data
5968 *
5969 * @return bool
5970 */
5973 }
5975 /**
5976 * Parsed a width param of imagelink like 300px or 200x300px
5977 *
5978 * @param string $value
5979 *
5980 * @return array
5981 * @since 1.20
5982 */
5987 }
5989 # (bug 13500) In both cases (width/height and width only),
5990 # permit trailing "px" for backward compatibility.
5999 }
6001 }
6003 /**
6004 * Lock the current instance of the parser.
6005 *
6006 * This is meant to stop someone from calling the parser
6007 * recursively and messing up all the strip state.
6008 *
6009 * @throws MWException If parser is in a parse
6010 * @return ScopedCallback The lock will be released once the return value goes out of scope.
6011 */
6016 }
6021 } );
6024 }
6026 /**
6027 * Strip outer <p></p> tag from the HTML source of a single paragraph.
6028 *
6029 * Returns original HTML if the <p/> tag has any attributes, if there's no wrapping <p/> tag,
6030 * or if there is more than one <p/> tag in the input HTML.
6031 *
6032 * @param string $html
6033 * @return string
6034 * @since 1.24
6035 */
6041 }
6042 }
6045 }
6047 /**
6048 * Return this parser if it is not doing anything, otherwise
6049 * get a fresh parser. You can use this method by doing
6050 * $myParser = $wgParser->getFreshParser(), or more simply
6051 * $wgParser->getFreshParser()->parse( ... );
6052 * if you're unsure if $wgParser is safe to use.
6053 *
6054 * @since 1.24
6055 * @return Parser A parser object that is not parsing anything
6056 */
6063 }
6064 }
6066 /**
6067 * Set's up the PHP implementation of OOUI for use in this request
6068 * and instructs OutputPage to enable OOUI for itself.
6069 *
6070 * @since 1.26
6071 */
6075 }
6076 }