| blob | 4385dd661cebe66caab3d1c63f06a25e757351e8 |
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 */
24 /**
25 * @defgroup Parser Parser
26 */
28 /**
29 * PHP Parser - Processes wiki markup (which uses a more user-friendly
30 * syntax, such as "[[link]]" for making links), and provides a one-way
31 * transformation of that wiki markup it into (X)HTML output / markup
32 * (which in turn the browser understands, and can display).
33 *
34 * There are seven main entry points into the Parser class:
35 *
36 * - Parser::parse()
37 * produces HTML output
38 * - Parser::preSaveTransform().
39 * produces altered wiki markup.
40 * - Parser::preprocess()
41 * removes HTML comments and expands templates
42 * - Parser::cleanSig() and Parser::cleanSigInSig()
43 * Cleans a signature before saving it to preferences
44 * - Parser::getSection()
45 * Return the content of a section from an article for section editing
46 * - Parser::replaceSection()
47 * Replaces a section by number inside an article
48 * - Parser::getPreloadText()
49 * Removes <noinclude> sections, and <includeonly> tags.
50 *
51 * Globals used:
52 * object: $wgContLang
53 *
54 * @warning $wgUser or $wgTitle or $wgRequest or $wgLang. Keep them away!
55 *
56 * @par Settings:
57 * $wgNamespacesWithSubpages
58 *
59 * @par Settings only within ParserOptions:
60 * $wgAllowExternalImages
61 * $wgAllowSpecialInclusion
62 * $wgInterwikiMagic
63 * $wgMaxArticleSize
64 *
65 * @ingroup Parser
66 */
68 /**
69 * Update this version number when the ParserOutput format
70 * changes in an incompatible way, so the parser cache
71 * can automatically discard old data.
72 */
75 /**
76 * Update this version number when the output of serialiseHalfParsedText()
77 * changes in an incompatible way
78 */
81 # Flags for Parser::setFunctionHook
82 # Also available as global constants from Defines.php
86 # Constants needed for external link processing
87 # Everything except bracket, space, or control characters
88 # \p{Zs} is unicode 'separator, space' category. It covers the space 0x20
89 # as well as U+3000 is IDEOGRAPHIC SPACE for bug 19052
94 # State constants for the definition list colon extraction
104 # Flags for preprocessToDom
107 # Allowed values for $this->mOutputType
108 # Parameter to startExternalParse().
115 # Marker Suffix needs to be accessible staticly.
118 # Markers used for wrapping the table of contents
122 # Persistent:
136 # Initialised by initialiseVariables()
138 /**
139 * @var MagicWordArray
140 */
143 /**
144 * @var MagicWordArray
145 */
147 var $mConf, $mPreprocessor, $mExtLinkBracketedRegex, $mUrlProtocols; # Initialised in constructor
149 # Cleared with clearState():
150 /**
151 * @var ParserOutput
152 */
156 /**
157 * @var StripState
158 */
162 /**
163 * @var LinkHolderArray
164 */
175 /**
176 * @var User
177 */
180 # Temporary
181 # These are variables reset at least once per parse regardless of $clearState
183 /**
184 * @var ParserOptions
185 */
188 /**
189 * @var Title
190 */
202 /**
203 * @var string
204 */
207 /**
208 * @var array Array with the language name of each language link (i.e. the
209 * interwiki prefix) in the key, value arbitrary. Used to avoid sending
210 * duplicate language links to the ParserOutput.
211 */
214 /**
215 * @var boolean Recursive call protection.
216 * This variable should be treated as if it were private.
217 */
220 /**
221 * @param array $conf
222 */
231 # Preprocessor_Hash is much faster than Preprocessor_DOM under HipHop
234 # PECL extension that conflicts with the core DOM extension (bug 13770)
241 }
243 }
245 /**
246 * Reduce memory usage to reduce the impact of circular references
247 */
251 }
254 }
255 }
257 /**
258 * Allow extensions to clean up when the parser is cloned
259 */
263 }
265 /**
266 * Do various kinds of initialisation on the first call of the parser
267 */
271 }
282 }
284 /**
285 * Clear Parser state
286 *
287 * @private
288 */
293 }
310 /**
311 * Prefix for temporary replacement strings for the multipass parser.
312 * \x07 should never appear in input as it's disallowed in XML.
313 * Using it at the front also gives us a little extra robustness
314 * since it shouldn't match when butted up against identifier-like
315 * string constructs.
316 *
317 * Must not consist of all title characters, or else it will change
318 * the behavior of <nowiki> in a link.
319 */
323 # Clear these on every parse, bug 4549
331 );
340 # Fix cloning
343 }
347 }
349 /**
350 * Convert wikitext to HTML
351 * Do not call this function recursively.
352 *
353 * @param string $text text we want to parse
354 * @param Title $title
355 * @param ParserOptions $options
356 * @param bool $linestart
357 * @param bool $clearState
358 * @param int $revid Number to pass in {{REVISIONID}}
359 * @return ParserOutput A ParserOutput
360 */
363 ) {
364 /**
365 * First pass--just handle <nowiki> sections, pass the rest off
366 * to internalParse() which does all the real work.
367 */
376 }
383 }
385 # Remove the strip marker tag prefix from the input, if present.
388 }
401 }
404 # No more strip!
411 # Clean up special characters, only run once, next-to-last before doBlockLevels
413 # french spaces, last one Guillemet-left
414 # only if there is something before the space
416 # french spaces, Guillemet-right
419 );
426 /**
427 * The input doesn't get language converted if
428 * a) It's disabled
429 * b) Content isn't converted
430 * c) It's a conversion table
431 * d) it is an interface message (which is in the user language)
432 */
435 ) {
437 # The position of the convert() call should not be changed. it
438 # assumes that the links are all replaced and the only thing left
439 # is the <nowiki> mark.
441 }
442 }
444 /**
445 * A converted title will be provided in the output object if title and
446 * content conversion are enabled, the article text does not contain
447 * a conversion-suppressing double-underscore tag, and no
448 * {{DISPLAYTITLE:...}} is present. DISPLAYTITLE takes precedence over
449 * automatic link conversion.
450 */
455 ) {
462 }
463 }
477 # attempt to sanitize at least some nesting problems
478 # (bug #2702 and quite a few others)
480 # ''Something [http://www.cool.com cool''] -->
481 # <i>Something</i><a href="http://www.cool.com"..><i>cool></i></a>
484 # fix up an anchor inside another anchor, only
485 # at least for a single single nested link (bug 3695)
488 # fix div inside inline elements- doBlockLevels won't wrap a line which
489 # contains a div, so fix it up here; replace
490 # div with escaped text
493 # remove empty italic or bold tag pairs, some
494 # introduced by rules above
496 );
502 }
508 );
509 }
513 # Information on include size limits, for the benefit of users who try to skirt them
521 );
522 }
527 );
531 );
534 );
537 );
540 );
543 );
546 );
552 }
556 ) ) {
562 }
566 }
567 }
568 }
569 // Since we're not really outputting HTML, decode the entities and
570 // then re-encode the things that need hiding inside HTML comments.
574 // Sanitize for comment. Note '‐' in the replacement is U+2010,
575 // which looks much like the problematic '-'.
582 }
583 }
596 }
598 /**
599 * Recursive parser entry point that can be called from an extension tag
600 * hook.
601 *
602 * If $frame is not provided, then template variables (e.g., {{{1}}}) within $text are not expanded
603 *
604 * @param string $text Text extension wants to have parsed
605 * @param bool|PPFrame $frame The frame to use for expanding any template variables
606 *
607 * @return string
608 */
616 }
618 /**
619 * Expand templates and variables in the text, producing valid, static wikitext.
620 * Also removes comments.
621 * @param string $text
622 * @param Title $title
623 * @param ParserOptions $options
624 * @param int|null $revid
625 * @return mixed|string
626 */
633 }
640 }
642 /**
643 * Recursive parser entry point that can be called from an extension tag
644 * hook.
645 *
646 * @param string $text Text to be expanded
647 * @param bool|PPFrame $frame The frame to use for expanding any template variables
648 * @return string
649 * @since 1.19
650 */
657 }
659 /**
660 * Process the wikitext for the "?preload=" feature. (bug 5210)
661 *
662 * "<noinclude>", "<includeonly>" etc. are parsed as for template
663 * transclusion, comments, templates, arguments, tags hooks and parser
664 * functions are untouched.
665 *
666 * @param string $text
667 * @param Title $title
668 * @param ParserOptions $options
669 * @param array $params
670 * @return string
671 */
672 public function getPreloadText( $text, Title $title, ParserOptions $options, $params = array() ) {
676 # Parser (re)initialisation
685 }
687 /**
688 * Get a random string
689 *
690 * @return string
691 */
694 }
696 /**
697 * Set the current user.
698 * Should only be used when doing pre-save transform.
699 *
700 * @param User|null $user User object or null (to reset)
701 */
704 }
706 /**
707 * Accessor for mUniqPrefix.
708 *
709 * @return string
710 */
713 # @todo FIXME: This is probably *horribly wrong*
714 # LanguageConverter seems to want $wgParser's uniqPrefix, however
715 # if this is called for a parser cache hit, the parser may not
716 # have ever been initialized in the first place.
717 # Not really sure what the heck is supposed to be going on here.
719 # throw new MWException( "Accessing uninitialized mUniqPrefix" );
720 }
722 }
724 /**
725 * Set the context title
726 *
727 * @param Title $t
728 */
732 }
735 # Strip the fragment to avoid various odd effects
740 }
741 }
743 /**
744 * Accessor for the Title object
745 *
746 * @return Title
747 */
750 }
752 /**
753 * Accessor/mutator for the Title object
754 *
755 * @param Title $x Title object or null to just get the current one
756 * @return Title
757 */
760 }
762 /**
763 * Set the output type
764 *
765 * @param int $ot New value
766 */
769 # Shortcut alias
775 );
776 }
778 /**
779 * Accessor/mutator for the output type
780 *
781 * @param int|null $x New value or null to just get the current one
782 * @return int
783 */
786 }
788 /**
789 * Get the ParserOutput object
790 *
791 * @return ParserOutput
792 */
795 }
797 /**
798 * Get the ParserOptions object
799 *
800 * @return ParserOptions object
801 */
804 }
806 /**
807 * Accessor/mutator for the ParserOptions object
808 *
809 * @param ParserOptions $x New value or null to just get the current one
810 * @return ParserOptions Current ParserOptions object
811 */
814 }
816 /**
817 * @return int
818 */
821 }
823 /**
824 * @param int $id
825 */
828 }
830 /**
831 * Get a language object for use in parser functions such as {{FORMATNUM:}}
832 * @return Language
833 */
836 }
838 /**
839 * Get the target language for the content being parsed. This is usually the
840 * language that the content is in.
841 *
842 * @since 1.19
843 *
844 * @throws MWException
845 * @return Language|null
846 */
856 }
859 }
861 /**
862 * Get the language object for language conversion
863 * @return Language|null
864 */
867 }
869 /**
870 * Get a User object either from $this->mUser, if set, or from the
871 * ParserOptions object otherwise
872 *
873 * @return User
874 */
878 }
880 }
882 /**
883 * Get a preprocessor object
884 *
885 * @return Preprocessor
886 */
891 }
893 }
895 /**
896 * Replaces all occurrences of HTML-style comments and the given tags
897 * in the text with a random marker and returns the next text. The output
898 * parameter $matches will be an associative array filled with data in
899 * the form:
900 *
901 * @code
902 * 'UNIQ-xxxxx' => array(
903 * 'element',
904 * 'tag content',
905 * array( 'param' => 'x' ),
906 * '<element param="x">tag content</element>' ) )
907 * @endcode
908 *
909 * @param array $elements List of element names. Comments are always extracted.
910 * @param string $text Source text string.
911 * @param array $matches Out parameter, Array: extracted tags
912 * @param string $uniq_prefix
913 * @return string Stripped text
914 */
928 }
930 # comment
936 # tag
941 }
947 # Empty element tag, <tag />
956 }
960 # No end tag -- let it run out to the end of the text.
966 }
967 }
973 }
975 }
977 /**
978 * Get a list of strippable XML-like elements
979 *
980 * @return array
981 */
984 }
986 /**
987 * Add an item to the strip state
988 * Returns the unique tag which must be inserted into the stripped text
989 * The tag will be replaced with the original text in unstrip()
990 *
991 * @param string $text
992 *
993 * @return string
994 */
1000 }
1002 /**
1003 * parse the wiki syntax used to render tables
1004 *
1005 * @private
1006 * @param string $text
1007 * @return string
1008 */
1027 }
1033 # First check if we are starting a new table
1046 # Don't do any of the following
1050 # We are ending a table
1056 }
1060 }
1064 }
1068 # Now we have a table row
1071 # Whats after the tag is now only attributes
1084 }
1088 }
1097 ) {
1098 # This might be cell elements, td, th or captions
1102 }
1108 }
1110 # Split up multiple cells on the same line.
1111 # FIXME : This can result in improper nesting of tags processed
1112 # by earlier parser steps, but should avoid splitting up eg
1113 # attribute values containing literal "||".
1118 # Loop through each table cell
1125 }
1130 }
1136 }
1146 }
1150 # A cell could contain both parameters and data
1153 # Bug 553: Note that a '|' inside an invalid link should not
1154 # be mistaken as delimiting cell parameters
1163 }
1167 }
1168 }
1170 }
1172 # Closing open td, tr && table
1176 }
1179 }
1182 }
1185 }
1187 # Remove trailing line-ending (b/c)
1190 }
1192 # special case: don't return empty table
1195 }
1200 }
1202 /**
1203 * Helper function for parse() that transforms wiki markup into
1204 * HTML. Only called for $mOutputType == self::OT_HTML.
1205 *
1206 * @private
1207 *
1208 * @param string $text
1209 * @param bool $isMain
1210 * @param bool $frame
1211 *
1212 * @return string
1213 */
1219 # Hook to suspend the parser in this state
1220 if ( !wfRunHooks( 'ParserBeforeInternalParse', array( &$this, &$text, &$this->mStripState ) ) ) {
1223 }
1225 # if $frame is provided, then use $frame for replacing any variables
1227 # use frame depth to infer how include/noinclude tags should be handled
1228 # depth=0 means this is the top-level document; otherwise it's an included document
1233 }
1237 # if $frame is not provided, then use old-style replaceVariables
1239 }
1247 );
1250 # Tables need to come after variable replacement for things to work
1251 # properly; putting them before other transformations should keep
1252 # exciting things like link expansions from showing up in surprising
1253 # places.
1265 # replaceInternalLinks may sometimes leave behind
1266 # absolute URLs, which have to be masked to hide them from replaceExternalLinks
1274 }
1276 /**
1277 * Replace special strings like "ISBN xxx" and "RFC xxx" with
1278 * magic external links.
1279 *
1280 * DML
1281 * @private
1282 *
1283 * @param string $text
1284 *
1285 * @return string
1286 */
1292 '!(?: # Start cases
1296 (?:RFC|PMID)\s+([0-9]+) | # m[4]: RFC or PMID, capture number
1298 (?: 97[89] [\ \-]? )? # optional 13-digit ISBN prefix
1299 (?: [0-9] [\ \-]? ){9} # 9 digits with opt. delimiters
1300 [0-9Xx] # check digit
1305 }
1307 /**
1308 * @throws MWException
1309 * @param array $m
1310 * @return HTML|string
1311 */
1314 # Skip anchor
1317 # Skip HTML element
1320 # Free external link
1323 # RFC or PMID
1337 }
1341 # ISBN
1347 ));
1354 }
1355 }
1357 /**
1358 * Make a free external link, given a user-supplied URL
1359 *
1360 * @param string $url
1361 *
1362 * @return string HTML
1363 * @private
1364 */
1370 # The characters '<' and '>' (which were escaped by
1371 # removeHTMLtags()) should not be included in
1372 # URLs, per RFC 2396.
1377 }
1379 # Move trailing punctuation to $trail
1381 # If there is no left bracket, then consider right brackets fair game too
1384 }
1390 }
1394 # Is this an external image?
1397 # Not an image, make a link
1402 # Register it in the output object...
1403 # Replace unnecessary URL escape codes with their equivalent characters
1406 }
1409 }
1411 /**
1412 * Parse headers and return html
1413 *
1414 * @private
1415 *
1416 * @param string $text
1417 *
1418 * @return string
1419 */
1425 }
1428 }
1430 /**
1431 * Replace single quotes with HTML markup
1432 * @private
1433 *
1434 * @param string $text
1435 *
1436 * @return string the altered text
1437 */
1444 }
1448 }
1450 /**
1451 * Helper function for doAllQuotes()
1452 *
1453 * @param string $text
1454 *
1455 * @return string
1456 */
1462 }
1464 // First, do some preliminary work. This may shift some apostrophes from
1465 // being mark-up to being text. It also counts the number of occurrences
1466 // of bold and italics mark-ups.
1471 // If there are ever four apostrophes, assume the first is supposed to
1472 // be text, and the remaining three constitute mark-up for bold text.
1473 // (bug 13227: ''''foo'''' turns into ' ''' foo ' ''')
1479 // If there are more than 5 apostrophes in a row, assume they're all
1480 // text except for the last 5.
1481 // (bug 13227: ''''''foo'''''' turns into ' ''''' foo ' ''''')
1485 }
1486 // Count the number of occurrences of bold and italics mark-ups.
1494 }
1495 }
1497 // If there is an odd number of both bold and italics, it is likely
1498 // that one of the bold ones was meant to be an apostrophe followed
1499 // by italics. Which one we cannot know for certain, but it is more
1500 // likely to be one that has a single-letter word before it.
1512 }
1516 // if $firstsingleletterword is set, we don't
1517 // look at the other options, so we can bail early.
1519 }
1523 }
1524 }
1525 }
1526 }
1528 // If there is a single-letter word, use it!
1533 // If not, but there's a multi-letter word, use that one.
1537 // ... otherwise use the first one that has neither.
1538 // (notice that it is possible for all three to be -1 if, for example,
1539 // there is only one pentuple-apostrophe in the line)
1542 }
1543 }
1545 // Now let's actually convert our apostrophic mush to HTML!
1556 }
1575 }
1592 }
1612 }
1613 }
1614 }
1616 }
1617 // Now close all remaining tags. Notice that the order is important.
1620 }
1623 }
1626 }
1627 // There might be lonely ''''', so make sure we have a buffer
1630 }
1632 }
1634 /**
1635 * Replace external links (REL)
1636 *
1637 * Note: this is all very hackish and the order of execution matters a lot.
1638 * Make sure to run tests/parserTests.php if you change this code.
1639 *
1640 * @private
1641 *
1642 * @param string $text
1643 *
1644 * @throws MWException
1645 * @return string
1646 */
1655 }
1665 # The characters '<' and '>' (which were escaped by
1666 # removeHTMLtags()) should not be included in
1667 # URLs, per RFC 2396.
1672 }
1674 # If the link text is an image URL, replace it with an <img> tag
1675 # This happened by accident in the original parser, but some people used it extensively
1679 }
1683 # Set linktype for CSS - if URL==text, link is essentially free
1686 # No link text, e.g. [http://domain.tld/some.link]
1688 # Autonumber
1693 # Have link text, e.g. [http://domain.tld/some.link text]s
1694 # Check for trail
1696 }
1702 # Use the encoded URL
1703 # This means that users can paste URLs directly into the text
1704 # Funny characters like ö aren't valid in URLs anyway
1705 # This was changed in August 2004
1709 # Register link in the output object.
1710 # Replace unnecessary URL escape codes with the referenced character
1711 # This prevents spammers from hiding links from the filters
1714 }
1718 }
1720 /**
1721 * Get the rel attribute for a particular external link.
1722 *
1723 * @since 1.21
1724 * @param string|bool $url Optional URL, to extract the domain from for rel =>
1725 * nofollow if appropriate
1726 * @param Title $title Optional Title, for wgNoFollowNsExceptions lookups
1727 * @return string|null Rel attribute for $url
1728 */
1734 ) {
1736 }
1738 }
1740 /**
1741 * Get an associative array of additional HTML attributes appropriate for a
1742 * particular external link. This currently may include rel => nofollow
1743 * (depending on configuration, namespace, and the URL's domain) and/or a
1744 * target attribute (depending on configuration).
1745 *
1746 * @param string|bool $url Optional URL, to extract the domain from for rel =>
1747 * nofollow if appropriate
1748 * @return array Associative array of HTML attributes
1749 */
1756 }
1758 }
1760 /**
1761 * Replace unusual URL escape codes with their equivalent characters
1762 *
1763 * @param string $url
1764 * @return string
1765 *
1766 * @todo This can merge genuinely required bits in the path or query string,
1767 * breaking legit URLs. A proper fix would treat the various parts of
1768 * the URL differently; as a workaround, just use the output for
1769 * statistical records, not for actual linking/output.
1770 */
1774 }
1776 /**
1777 * Callback function used in replaceUnusualEscapes().
1778 * Replaces unusual URL escape codes with their equivalent character
1779 *
1780 * @param array $matches
1781 *
1782 * @return string
1783 */
1787 # Is it an unsafe or HTTP reserved character according to RFC 1738?
1789 # No, shouldn't be escaped
1792 # Yes, leave it escaped
1794 }
1795 }
1797 /**
1798 * make an image if it's allowed, either through the global
1799 * option, through the exception, or through the on-wiki whitelist
1800 *
1801 * @param string $url
1802 *
1803 * @return string
1804 */
1809 # $imagesfrom could be either a single string or an array of strings, parse out the latter
1816 }
1817 }
1822 }
1826 ) {
1828 # Image found
1830 }
1831 }
1834 ) {
1838 );
1841 # Sanitize the regex fragment, make it case-insensitive, ignore blank entries/comments
1844 }
1846 # Image matches a whitelist entry
1849 }
1850 }
1851 }
1853 }
1855 /**
1856 * Process [[ ]] wikilinks
1857 *
1858 * @param string $s
1859 *
1860 * @return string Processed text
1861 *
1862 * @private
1863 */
1867 }
1869 /**
1870 * Process [[ ]] wikilinks (RIL)
1871 * @param string $s
1872 * @throws MWException
1873 * @return LinkHolderArray
1874 *
1875 * @private
1876 */
1882 # the % is needed to support urlencoded titles as well
1885 # Match a link having the form [[namespace:link|alternate]]trail
1887 # Match cases where there is no "]]", which might still be images
1889 }
1893 # split the entire text string on occurrences of [[
1895 # get the first element (all text up to first [[), and remove the space we added
1904 # Match the end of a line for a word that's not followed by whitespace,
1905 # e.g. in the case of 'The Arab al[[Razi]]', 'al' will be matched
1909 }
1915 }
1924 }
1927 }
1932 // @codingStandardsIgnoreStart Squiz.WhiteSpace.SemicolonSpacing.Incorrect
1933 # Loop for each link
1935 // @codingStandardsIgnoreStart
1937 # Check for excessive memory usage
1939 # Too big
1940 # Do the existence check, replace the link holders and clear the array
1943 }
1952 }
1953 # first link
1957 }
1959 }
1966 # If we get a ] at the beginning of $m[3] that means we have a link that's something like:
1967 # [[Image:Foo.jpg|[http://example.com desc]]] <- having three ] in a row fucks up,
1968 # the real problem is with the $e1 regex
1969 # See bug 1300.
1970 #
1971 # Still some problems for cases where the ] is meant to be outside punctuation,
1972 # and no image is in sight. See bug 2095.
1973 #
1977 ) {
1980 }
1981 # fix up urlencoded title texts
1983 # Should anchors '#' also be rejected?
1985 }
1988 # Invalid, but might be an image with a link in its caption
1993 }
1999 }
2003 # Don't allow internal links to pages containing
2004 # PROTO: where PROTO is a valid URL protocol; these
2005 # should be external links.
2010 }
2012 # Make subpage if necessary
2017 }
2021 # Strip off leading ':'
2023 }
2032 }
2043 # look at the next 'line' to see if we can close it there
2048 }
2051 # the first ]] closes the inner link, the second the image
2057 # if there's exactly one ]] that's fine, we'll keep looking
2060 # if $next_line is invalid too, we need look no further
2063 }
2064 }
2066 # we couldn't find the end of this imageLink, so output it raw
2067 # but don't ignore what might be perfectly normal links in the text we've examined
2070 # note: no $trail, because without an end, there *is* no trail
2073 }
2076 # note: no $trail, because without an end, there *is* no trail
2079 }
2081 }
2087 # Bug 4598 madness. Handle the quotes only if they come from the alternate part
2088 # [[Lista d''e paise d''o munno]] -> <a href="...">Lista d''e paise d''o munno</a>
2089 # [[Criticism of Harry Potter|Criticism of ''Harry Potter'']]
2090 # -> <a href="Criticism of Harry Potter">Criticism of <i>Harry Potter</i></a>
2092 }
2094 # Link not escaped by : , create the various objects
2096 # Interwikis
2100 ) {
2101 // XXX: the above check prevents links to sites with identifiers that are not language codes
2103 # Bug 24502: filter duplicates
2107 }
2113 }
2120 # if no parameters were passed, $text
2121 # becomes something like "File:Foo.png",
2122 # which we don't want to pass on to the
2123 # image generator
2126 # recursively parse links inside the image caption
2127 # actually, this will parse them in any other parameters, too,
2128 # but it might be hard to fix that, and it doesn't matter ATM
2131 }
2132 # cloak any absolute URLs inside the image markup, so replaceExternalLinks() won't touch them
2137 }
2140 }
2150 }
2156 /**
2157 * Strip the whitespace Category links produce, see bug 87
2158 */
2163 }
2164 }
2166 # Self-link checking. For some languages, variants of the title are checked in
2167 # LinkHolderArray::doVariants() to allow batching the existence checks necessary
2168 # for linking to a different variant.
2172 }
2174 # NS_MEDIA is a pseudo-namespace for linking directly to a file
2175 # @todo FIXME: Should do batch file existence checks, see comment below
2178 # Give extensions a chance to select the file revision for us
2183 # Fetch and register the file (file title may be different via hooks)
2185 # Cloak with NOPARSE to avoid replacement in replaceExternalLinks
2190 }
2193 # Some titles, such as valid special pages or files in foreign repos, should
2194 # be shown as bluelinks even though they're not included in the page table
2195 #
2196 # @todo FIXME: isAlwaysKnown() can be expensive for file links; we should really do
2197 # batch file existence checks for NS_FILE and NS_MEDIA
2202 # Links will be added to the output link list after checking
2204 }
2206 }
2209 }
2211 /**
2212 * Render a forced-blue link inline; protect against double expansion of
2213 * URLs if we're in a mode that prepends full URL prefixes to internal links.
2214 * Since this little disaster has to split off the trail text to avoid
2215 * breaking URLs in the following text without breaking trails on the
2216 * wiki links, it's been made into a horrible function.
2217 *
2218 * @param Title $nt
2219 * @param string $text
2220 * @param array|string $query
2221 * @param string $trail
2222 * @param string $prefix
2223 * @return string HTML-wikitext mix oh yuck
2224 */
2230 }
2233 }
2238 }
2240 /**
2241 * Insert a NOPARSE hacky thing into any inline links in a chunk that's
2242 * going to go through further parsing steps before inline URL expansion.
2243 *
2244 * Not needed quite as much as it used to be since free links are a bit
2245 * more sensible these days. But bracketed links are still an issue.
2246 *
2247 * @param string $text More-or-less HTML
2248 * @return string Less-or-more HTML with NOPARSE bits
2249 */
2253 }
2255 /**
2256 * Return true if subpage links should be expanded on this page.
2257 * @return bool
2258 */
2260 # Some namespaces don't allow subpages
2262 }
2264 /**
2265 * Handle link to subpage if necessary
2266 *
2267 * @param string $target The source of the link
2268 * @param string &$text The link text, modified as necessary
2269 * @return string The full name of the link
2270 * @private
2271 */
2274 }
2276 /**#@+
2277 * Used by doBlockLevels()
2278 * @private
2279 *
2280 * @return string
2281 */
2286 }
2290 }
2292 /**
2293 * getCommon() returns the length of the longest common substring
2294 * of both arguments, starting at the beginning of both.
2295 * @private
2296 *
2297 * @param string $st1
2298 * @param string $st2
2299 *
2300 * @return int
2301 */
2307 }
2312 }
2313 }
2315 }
2317 /**
2318 * These next three functions open, continue, and close the list
2319 * element appropriate to the prefix character passed into them.
2320 * @private
2321 *
2322 * @param string $char
2323 *
2324 * @return string
2325 */
2340 }
2343 }
2345 /**
2346 * TODO: document
2347 * @param string $char
2348 * @private
2349 *
2350 * @return string
2351 */
2359 }
2366 }
2367 }
2369 }
2371 /**
2372 * @todo Document
2373 * @param string $char
2374 * @private
2375 *
2376 * @return string
2377 */
2389 }
2392 }
2394 }
2395 /**#@-*/
2397 /**
2398 * Make lists from lines starting with ':', '*', '#', etc. (DBL)
2399 *
2400 * @param string $text
2401 * @param bool $linestart Whether or not this is at the start of a line.
2402 * @private
2403 * @return string The lists rendered as HTML
2404 */
2408 # Parsing through the text line by line. The main thing
2409 # happening here is handling of block-level elements p, pre,
2410 # and making lists from lines starting with * # : etc.
2411 #
2421 # Fix up $linestart
2426 }
2427 # * = ul
2428 # # = ol
2429 # ; = dt
2430 # : = dd
2435 # If not in a <pre> element, scan for and figure out what prefixes are there.
2437 # Multiple prefixes may abut each other for nested lists.
2441 # eh?
2442 # ; and : are both from definition-lists, so they're equivalent
2443 # for the purposes of determining whether or not we need to open/close
2444 # elements.
2449 # Don't interpret any other prefixes in preformatted text
2453 }
2455 # List generation
2457 # Same as the last item, so no need to deal with nesting or opening stuff
2462 # The one nasty exception: definition lists work like this:
2463 # ; title : definition text
2464 # So we check for : in the remainder text to split up the
2465 # title and definition, without b0rking links.
2470 }
2471 }
2473 # We need to open or close prefixes, or both.
2475 # Either open or close a level...
2479 # Close all the prefixes which aren't shared.
2483 }
2485 # Continue the current prefix if appropriate.
2488 }
2490 # Open prefixes where appropriate.
2493 }
2499 # @todo FIXME: This is dupe of code above
2503 }
2504 }
2506 }
2509 }
2511 }
2513 # If we have no prefixes, go to paragraph mode.
2516 # No prefix (not in list)--go to paragraph mode
2517 # XXX: use a stack for nestable elements like span, table and div
2519 '/(?:<table|<h1|<h2|<h3|<h4|<h5|<h6|<pre|<tr|'
2521 $t
2522 );
2528 $t
2529 );
2533 # @todo bug 5718: paragraph closed
2537 }
2539 while ( preg_match( '/<(\\/?)blockquote[\s>]/i', $t, $bqMatch, PREG_OFFSET_CAPTURE, $bqOffset ) ) {
2542 }
2548 ) {
2549 # pre
2554 }
2557 # paragraph
2570 }
2571 }
2580 }
2581 }
2582 }
2583 }
2585 }
2586 # somewhere above we forget to get out of pre block (bug 785)
2589 }
2594 }
2595 }
2596 }
2602 }
2603 }
2607 }
2611 }
2613 /**
2614 * Split up a string on ':', ignoring any occurrences inside tags
2615 * to prevent illegal overlapping.
2616 *
2617 * @param string $str The string to split
2618 * @param string &$before Set to everything before the ':'
2619 * @param string &$after Set to everything after the ':'
2620 * @throws MWException
2621 * @return string The position of the ':', or false if none found
2622 */
2628 # Nothing to find!
2631 }
2635 # Easy; no tag nesting to worry about
2640 }
2642 # Ugly state machine to walk through avoiding tags.
2650 # (Using the number is a performance hack for common cases)
2654 # Could be either a <start> tag or an </end> tag
2659 # We found it!
2664 }
2665 # Embedded in a tag; don't break it.
2668 # Skip ahead looking for something interesting
2671 # Nothing else interesting
2674 }
2678 # We found it!
2683 }
2684 }
2686 # Nothing else interesting to find; abort!
2687 # We're nested, but there's no close tags left. Abort!
2689 }
2690 # Skip ahead to next tag start
2693 }
2696 # In a <tag>
2703 # Slash may be followed by >?
2707 # ignore
2708 }
2719 # Illegal early close? This shouldn't happen D:
2724 }
2727 # In a </tag>
2734 }
2736 }
2740 # Yes, a self-closed tag <blah/>
2743 # Probably we're jumping the gun, and this is an attribute
2745 }
2750 }
2757 }
2764 }
2769 }
2770 }
2772 wfDebug( __METHOD__ . ": Invalid input; not enough close tags (stack $stack, state $state)\n" );
2775 }
2778 }
2780 /**
2781 * Return value of a magic variable (like PAGENAME)
2782 *
2783 * @private
2784 *
2785 * @param int $index
2786 * @param bool|PPFrame $frame
2787 *
2788 * @throws MWException
2789 * @return string
2790 */
2796 // If no title set, bad things are going to happen
2797 // later. Title should always be set since this
2798 // should only be called in the middle of a parse
2799 // operation (but the unit-tests do funky stuff)
2802 }
2804 /**
2805 * Some of these require message or data lookups and can be
2806 * expensive to check many times.
2807 */
2811 }
2812 }
2854 $value = $pageLang->getMonthAbbreviation( MWTimestamp::getLocalInstance( $ts )->format( 'n' ) );
2888 ) ) );
2898 ) ) );
2906 }
2914 }
2927 # 0 means the page doesn't exist in the database,
2928 # which means the user is previewing a new page.
2929 # The vary-revision flag must be set, because the magic word
2930 # will have a different value once the page is saved.
2933 }
2937 # Let the edit saving system know we should parse the page
2938 # *after* a revision ID has been assigned.
2944 # Let the edit saving system know we should parse the page
2945 # *after* a revision ID has been assigned. This is for null edits.
2951 # Let the edit saving system know we should parse the page
2952 # *after* a revision ID has been assigned. This is for null edits.
2958 # Let the edit saving system know we should parse the page
2959 # *after* a revision ID has been assigned. This is for null edits.
2965 # Let the edit saving system know we should parse the page
2966 # *after* a revision ID has been assigned. This is for null edits.
2972 # Let the edit saving system know we should parse the page
2973 # *after* a revision ID has been assigned. This is for null edits.
2979 # Let the edit saving system know we should parse the page
2980 # *after* a revision ID has been assigned. This is for null edits.
2986 # Let the edit saving system know we should parse the page
2987 # *after* a revision ID has been assigned. This is for null edits.
2993 # Let the edit saving system know we should parse the page
2994 # *after* a revision ID has been assigned. This is for null edits.
3035 # @bug 4594 PHP5 has it zero padded, PHP4 does not, cast to
3036 # int to remove the padding
3045 );
3054 false
3055 );
3061 # @bug 4594 PHP5 has it zero padded, PHP4 does not, cast to
3062 # int to remove the padding
3127 );
3130 }
3134 }
3137 }
3139 /**
3140 * initialise the magic variables (like CURRENTMONTHNAME) and substitution modifiers
3141 *
3142 * @private
3143 */
3152 }
3154 /**
3155 * Preprocess some wikitext and return the document tree.
3156 * This is the ghost of replace_variables().
3157 *
3158 * @param string $text The text to parse
3159 * @param int $flags Bitwise combination of:
3160 * - self::PTD_FOR_INCLUSION: Handle "<noinclude>" and "<includeonly>" as if the text is being
3161 * included. Default is to assume a direct page view.
3162 *
3163 * The generated DOM tree must depend only on the input text and the flags.
3164 * The DOM tree must be the same in OT_HTML and OT_WIKI mode, to avoid a regression of bug 4899.
3165 *
3166 * Any flag added to the $flags parameter here, or any other parameter liable to cause a
3167 * change in the DOM tree for a given text, must be passed through the section identifier
3168 * in the section edit link and thus back to extractSections().
3169 *
3170 * The output of this function is currently only cached in process memory, but a persistent
3171 * cache may be implemented at a later date which takes further advantage of these strict
3172 * dependency requirements.
3173 *
3174 * @return PPNode
3175 */
3179 }
3181 /**
3182 * Return a three-element array: leading whitespace, string contents, trailing whitespace
3183 *
3184 * @param string $s
3185 *
3186 * @return array
3187 */
3197 }
3199 }
3201 /**
3202 * Replace magic variables, templates, and template arguments
3203 * with the appropriate text. Templates are substituted recursively,
3204 * taking care to avoid infinite loops.
3205 *
3206 * Note that the substitution depends on value of $mOutputType:
3207 * self::OT_WIKI: only {{subst:}} templates
3208 * self::OT_PREPROCESS: templates but not extension tags
3209 * self::OT_HTML: all templates and extension tags
3210 *
3211 * @param string $text The text to transform
3212 * @param bool|PPFrame $frame Object describing the arguments passed to the
3213 * template. Arguments may also be provided as an associative array, as
3214 * was the usual case before MW1.12. Providing arguments this way may be
3215 * useful for extensions wishing to perform variable replacement
3216 * explicitly.
3217 * @param bool $argsOnly Only do argument (triple-brace) expansion, not
3218 * double-brace expansion.
3219 * @return string
3220 */
3222 # Is there any text? Also, Prevent too big inclusions!
3225 }
3234 }
3242 }
3244 /**
3245 * Clean up argument array - refactored in 1.9 so parserfunctions can use it, too.
3246 *
3247 * @param array $args
3248 *
3249 * @return array
3250 */
3263 }
3266 }
3267 }
3268 }
3271 }
3273 /**
3274 * Warn the user when a parser limitation is reached
3275 * Will warn at most once the user per limitation type
3276 *
3277 * @param string $limitationType Should be one of:
3278 * 'expensive-parserfunction' (corresponding messages:
3279 * 'expensive-parserfunction-warning',
3280 * 'expensive-parserfunction-category')
3281 * 'post-expand-template-argument' (corresponding messages:
3282 * 'post-expand-template-argument-warning',
3283 * 'post-expand-template-argument-category')
3284 * 'post-expand-template-inclusion' (corresponding messages:
3285 * 'post-expand-template-inclusion-warning',
3286 * 'post-expand-template-inclusion-category')
3287 * 'node-count-exceeded' (corresponding messages:
3288 * 'node-count-exceeded-warning',
3289 * 'node-count-exceeded-category')
3290 * 'expansion-depth-exceeded' (corresponding messages:
3291 * 'expansion-depth-exceeded-warning',
3292 * 'expansion-depth-exceeded-category')
3293 * @param string|int|null $current Current value
3294 * @param string|int|null $max Maximum allowed, when an explicit limit has been
3295 * exceeded, provide the values (optional)
3296 */
3298 # does no harm if $current and $max are present but are unnecessary for the message
3303 }
3305 /**
3306 * Return the text of a template, after recursively
3307 * replacing any variables or templates within the template.
3308 *
3309 * @param array $piece The parts of the template
3310 * $piece['title']: the title, i.e. the part before the |
3311 * $piece['parts']: the parameter array
3312 * $piece['lineStart']: whether the brace was at the start of a line
3313 * @param PPFrame $frame The current frame, contains template arguments
3314 * @throws Exception
3315 * @return string The text of the template
3316 */
3321 // Flags
3323 // $text has been filled
3325 // wiki markup in $text should be escaped
3327 // $text is HTML, armour it against wikitext transformation
3329 // Force interwiki transclusion to be done in raw mode not rendered
3331 // $text is a DOM node needing expansion in a child frame
3333 // $text is a DOM node needing expansion in the current frame
3336 # Title object, where $text came from
3339 # $part1 is the bit before the first |, and must contain only title characters.
3340 # Various prefixes will be stripped from it later.
3345 # Original title text preserved for various purposes
3348 # $args is a list of argument nodes, starting from index 0, not including $part1
3349 # @todo FIXME: If piece['parts'] is null then the call to getLength()
3350 # below won't work b/c this $args isn't an object
3356 # SUBST
3362 # Possibilities for substMatch: "subst", "safesubst" or FALSE
3363 # Decide whether to expand template or keep wikitext as-is.
3369 }
3375 }
3376 }
3381 }
3382 }
3384 # Variables
3391 }
3393 }
3394 }
3396 # MSG, MSGNW and RAW
3398 # Check for MSGNW:
3403 # Remove obsolete MSG:
3406 }
3408 # Check for RAW:
3412 }
3413 }
3416 # Parser functions
3426 }
3433 }
3435 # The interface for parser functions allows for extracting
3436 # flags into the local scope. Extract any forwarded flags
3437 # here.
3439 }
3441 }
3443 # Finish mangling title and then check for loops.
3444 # Set $title to a Title object and $titleText to the PDBK
3447 # Split the title into page and subpage
3453 }
3457 # Check for language variants if the template is not found
3460 }
3461 # Do recursion depth check
3469 }
3470 }
3471 }
3473 # Load from database
3476 # Too many unique items can kill profiling DBs/collectors
3479 }
3485 ) {
3486 // Pass the template arguments as URL parameters.
3487 // "uselang" will have no effect since the Language object
3488 // is forced to the one defined in ParserOptions.
3497 }
3498 }
3500 // Create a new context to execute the special page
3513 }
3523 }
3524 }
3526 # If the title is valid but undisplayable, make a link to it
3530 }
3532 # Interwiki transclusion
3538 # Preprocess it like a template
3541 }
3543 }
3545 # Do infinite loop check
3546 # This has to be done after redirect resolution to avoid infinite loops via redirects
3553 }
3555 }
3557 # If we haven't found text to substitute by now, we're done
3558 # Recover the source wikitext and return it
3563 }
3566 }
3568 # Expand DOM-style return values in a child frame
3570 # Clean up argument array
3576 # Expansion is eligible for the empty-frame cache
3582 }
3584 # Uncached expansion
3586 }
3587 }
3591 }
3595 }
3597 # Replace raw HTML by a placeholder
3601 # Escape nowiki-style return values
3606 ) {
3607 # Bug 529: if the template begins with a table or block-level
3608 # element, it should be treated as beginning a new line.
3609 # This behavior is somewhat controversial.
3611 }
3614 # Error, oversize inclusion
3616 # Make a working, properly escaped link if possible (bug 23588)
3619 # This will probably not be a working link, but at least it may
3620 # provide some hint of where the problem is
3623 }
3627 }
3633 }
3637 }
3639 /**
3640 * Call a parser function and return an array with text and flags.
3641 *
3642 * The returned array will always contain a boolean 'found', indicating
3643 * whether the parser function was found or not. It may also contain the
3644 * following:
3645 * text: string|object, resulting wikitext or PP DOM object
3646 * isHTML: bool, $text is HTML, armour it against wikitext transformation
3647 * isChildObj: bool, $text is a DOM node needing expansion in a child frame
3648 * isLocalObj: bool, $text is a DOM node needing expansion in the current frame
3649 * nowiki: bool, wiki markup in $text should be escaped
3650 *
3651 * @since 1.21
3652 * @param PPFrame $frame The current frame, contains template arguments
3653 * @param string $function Function name
3654 * @param array $args Arguments to the function
3655 * @throws MWException
3656 * @return array
3657 */
3663 # Case sensitive functions
3667 # Case insensitive functions
3674 }
3675 }
3680 # Workaround for PHP bug 35229 and similar
3685 }
3689 # Convert arguments to PPNodes and collect for appending to $allArgs
3696 }
3697 }
3699 # Add a frame parameter, and pass the arguments as an array
3703 # Convert arguments to plain text and append to $allArgs
3711 }
3712 }
3713 }
3717 # The interface for function hooks allows them to return a wikitext
3718 # string or an array containing the string and any flags. This mungs
3719 # things around to match what this method should return.
3724 );
3728 }
3732 );
3733 }
3739 }
3742 }
3747 }
3752 }
3754 /**
3755 * Get the semi-parsed DOM representation of a template with a given title,
3756 * and its redirect destination title. Cached.
3757 *
3758 * @param Title $title
3759 *
3760 * @return array
3761 */
3770 }
3773 }
3775 # Cache miss, go to the database
3781 }
3789 }
3792 }
3794 /**
3795 * Fetch the unparsed text of a template and register a reference to it.
3796 * @param Title $title
3797 * @return array ( string or false, Title )
3798 */
3800 // Defaults to Parser::statelessFetchTemplate()
3809 // If we transclude ourselves, the final result
3810 // will change based on the new version of the page
3812 }
3813 }
3814 }
3816 }
3818 /**
3819 * Fetch the unparsed text of a template and register a reference to it.
3820 * @param Title $title
3821 * @return string|bool
3822 */
3826 }
3828 /**
3829 * Static function to get a template
3830 * Can be overridden via ParserOptions::setTemplateCallback().
3831 *
3832 * @param Title $title
3833 * @param bool|Parser $parser
3834 *
3835 * @return array
3836 */
3842 # Loop to fetch the article, with up to 1 redirect
3844 # Give extensions a chance to select the revision instead
3855 );
3857 }
3858 # Get the revision
3863 # If there is no current revision, there is no page
3867 }
3874 # We fetched a rev from a different title; register it too...
3879 }
3888 }
3895 }
3900 }
3903 }
3904 # Redirect?
3907 }
3912 }
3914 /**
3915 * Fetch a file and its title and register a reference to it.
3916 * If 'broken' is a key in $options then the file will appear as a broken thumbnail.
3917 * @param Title $title
3918 * @param array $options Array of options to RepoGroup::findFile
3919 * @return File|bool
3920 */
3924 }
3926 /**
3927 * Fetch a file and its title and register a reference to it.
3928 * If 'broken' is a key in $options then the file will appear as a broken thumbnail.
3929 * @param Title $title
3930 * @param array $options Array of options to RepoGroup::findFile
3931 * @return array ( File or false, Title of file )
3932 */
3938 # Register the file as a dependency...
3941 # Update fetched file title
3944 }
3946 }
3948 /**
3949 * Helper function for fetchFileAndTitle.
3950 *
3951 * Also useful if you need to fetch a file but not use it yet,
3952 * for example to get the file's handler.
3953 *
3954 * @param Title $title
3955 * @param array $options Array of options to RepoGroup::findFile
3956 * @return File|bool
3957 */
3965 }
3967 }
3969 /**
3970 * Transclude an interwiki link.
3971 *
3972 * @param Title $title
3973 * @param string $action
3974 *
3975 * @return string
3976 */
3982 }
3988 }
3990 }
3992 /**
3993 * @param string $url
3994 * @return mixed|string
3995 */
4004 }
4011 // Though we failed to fetch the content, this status is useless.
4016 }
4023 ) );
4025 }
4027 /**
4028 * Triple brace replacement -- used for template arguments
4029 * @private
4030 *
4031 * @param array $piece
4032 * @param PPFrame $frame
4033 *
4034 * @return array
4035 */
4049 )
4050 ) {
4051 # No match in frame, use the supplied default
4053 }
4057 }
4060 # No match anywhere
4062 }
4065 }
4070 }
4074 }
4076 /**
4077 * Return the text to be used for a given extension tag.
4078 * This is the ghost of strip().
4079 *
4080 * @param array $params Associative array of parameters:
4081 * name PPNode for the tag name
4082 * attr PPNode for unparsed text where tag attributes are thought to be
4083 * attributes Optional associative array of parsed attributes
4084 * inner Contents of extension element
4085 * noClose Original text did not have a close tag
4086 * @param PPFrame $frame
4087 *
4088 * @throws MWException
4089 * @return string
4090 */
4104 }
4110 }
4113 # Workaround for PHP bug 35229 and similar
4116 }
4123 }
4129 }
4132 # Extract flags to local scope (to override $markerType)
4137 }
4141 }
4146 }
4147 }
4153 }
4154 }
4164 }
4166 }
4168 /**
4169 * Increment an include size counter
4170 *
4171 * @param string $type The type of expansion
4172 * @param int $size The size of the text
4173 * @return bool False if this inclusion would take it over the maximum, true otherwise
4174 */
4181 }
4182 }
4184 /**
4185 * Increment the expensive function count
4186 *
4187 * @return bool False if the limit has been exceeded
4188 */
4192 }
4194 /**
4195 * Strip double-underscore items like __NOGALLERY__ and __NOTOC__
4196 * Fills $this->mDoubleUnderscores, returns the modified text
4197 *
4198 * @param string $text
4199 *
4200 * @return string
4201 */
4205 # The position of __TOC__ needs to be recorded
4211 # Set a placeholder. At the end we'll fill it in with the TOC.
4214 # Only keep the first one.
4216 }
4218 # Now match and remove the rest of them
4224 }
4227 }
4230 ) {
4232 }
4233 # (bug 8068) Allow control over whether robots index a page.
4234 #
4235 # @todo FIXME: Bug 14899: __INDEX__ always overrides __NOINDEX__ here! This
4236 # is not desirable, the last one on the page should win.
4240 }
4244 }
4246 # Cache all double underscores in the database
4249 }
4253 }
4255 /**
4256 * Add a tracking category, getting the title from a system message,
4257 * or print a debug message if the title is invalid.
4258 *
4259 * Please add any message that you use with this function to
4260 * $wgTrackingCategories. That way they will be listed on
4261 * Special:TrackingCategories.
4262 *
4263 * @param string $msg Message key
4264 * @return bool Whether the addition was successful
4265 */
4270 }
4271 // Important to parse with correct title (bug 31469)
4277 # Allow tracking categories to be disabled by setting them to "-"
4280 }
4289 }
4290 }
4292 /**
4293 * This function accomplishes several tasks:
4294 * 1) Auto-number headings if that option is enabled
4295 * 2) Add an [edit] link to sections for users who have enabled the option and can edit the page
4296 * 3) Add a Table of contents on the top for users who have enabled the option
4297 * 4) Auto-anchor headings
4298 *
4299 * It loops through all headlines, collects the necessary data, then splits up the
4300 * string and re-inserts the newly formatted headlines.
4301 *
4302 * @param string $text
4303 * @param string $origText Original, untouched wikitext
4304 * @param bool $isMain
4305 * @return mixed|string
4306 * @private
4307 */
4311 # Inhibit editsection links if requested in the page
4317 }
4320 }
4322 # Get all headlines for numbering them and adding funky stuff like [edit]
4323 # links - this is for later, but we need the number of headlines right now
4328 $matches
4329 );
4331 # if there are fewer than 4 headlines in the article, do not show TOC
4332 # unless it's been explicitly enabled.
4336 # Allow user to stipulate that a page should have a "new section"
4337 # link added via __NEWSECTIONLINK__
4340 }
4342 # Allow user to remove the "new section"
4343 # link via __NONEWSECTIONLINK__
4346 }
4348 # if the string __FORCETOC__ (not case-sensitive) occurs in the HTML,
4349 # override above conditions and always show TOC above first header
4353 }
4355 # headline counter
4359 # Ugh .. the TOC should have neat indentation levels which can be
4360 # passed to the skin functions. These are determined here
4392 }
4396 }
4400 # Increase TOC level
4407 }
4409 # Decrease TOC level, find level to jump to
4413 # Found last matching level
4417 # Found first matching level below current level
4420 }
4421 }
4424 }
4427 # Unindent only if the previous toc level was shown :p
4432 }
4433 }
4435 # No change in level, end TOC line
4438 }
4439 }
4443 # count number of headlines for each level
4450 }
4453 }
4454 }
4456 # The safe header is a version of the header text safe to use for links
4458 # Remove link placeholders by the link text.
4459 # <!--LINK number-->
4460 # turns into
4461 # link text with suffix
4462 # Do this before unstrip since link text can contain strip markers
4465 # Avoid insertion of weird stuff like <math> by expanding the relevant sections
4468 # Strip out HTML (first regex removes any tag not allowed)
4469 # Allowed tags are:
4470 # * <sup> and <sub> (bug 8393)
4471 # * <i> (bug 26375)
4472 # * <b> (r105284)
4473 # * <span dir="rtl"> and <span dir="ltr"> (bug 35167)
4474 #
4475 # We strip any parameter from accepted tags (second regex), except dir="rtl|ltr" from <span>,
4476 # to allow setting directionality in toc items.
4481 ),
4483 $safeHeadline
4484 );
4487 # For the anchor, strip out HTML-y stuff period
4491 # Save headline for section edit hint before it's escaped
4495 # For reverse compatibility, provide an id that's
4496 # HTML4-compatible, like we used to.
4497 #
4498 # It may be worth noting, academically, that it's possible for
4499 # the legacy anchor to conflict with a non-legacy headline
4500 # anchor on the page. In this case likely the "correct" thing
4501 # would be to either drop the legacy anchors or make sure
4502 # they're numbered first. However, this would require people
4503 # to type in section names like "abc_.D7.93.D7.90.D7.A4"
4504 # manually, so let's not bother worrying about it.
4510 # No reason to have both (in fact, we can't)
4512 }
4517 }
4519 # HTML names must be case-insensitively unique (bug 10721).
4520 # This does not apply to Unicode characters per
4521 # http://dev.w3.org/html5/spec/infrastructure.html#case-sensitivity-and-string-comparison
4522 # @todo FIXME: We may be changing them depending on the current locale.
4528 }
4530 # count how many in assoc. array so we can track dupes in anchors
4535 }
4540 }
4542 # Don't number the heading if it is the only one (looks silly)
4544 # the two are different if the line contains a link
4548 $numbering
4550 }
4552 # Create the anchor for linking from the TOC to the section
4557 }
4560 }
4564 }
4566 # Add the section to the section tree
4567 # Find the DOM node for this header
4574 }
4575 }
4579 }
4589 );
4591 # give headline the correct <h#> tag
4593 // Output edit section links as markers with styles that can be customized by skins
4595 # Put a T flag in the section identifier, to indicate to extractSections()
4596 # that sections inside <includeonly> should be counted.
4602 $headlineHint
4603 );
4604 }
4605 // We use a bit of pesudo-xml for editsection markers. The
4606 // language converter is run later on. Using a UNIQ style marker
4607 // leads to the converter screwing up the tokens when it
4608 // converts stuff. And trying to insert strip tags fails too. At
4609 // this point all real inputted tags have already been escaped,
4610 // so we don't have to worry about a user trying to input one of
4611 // these markers directly. We use a page and section attribute
4612 // to stop the language converter from converting these
4613 // important bits of data, but put the headline hint inside a
4614 // content block because the language converter is supposed to
4615 // be able to convert that piece of data.
4622 }
4625 }
4631 }
4635 # Never ever show TOC if no headers
4638 }
4643 }
4648 }
4652 }
4654 # split up and insert constructed headlines
4658 // build an array of document sections
4661 // $head is zero-based, sections aren't.
4666 }
4668 /**
4669 * Send a hook, one per section.
4670 * The idea here is to be able to make section-level DIVs, but to do so in a
4671 * lower-impact, more correct way than r50769
4672 *
4673 * $this : caller
4674 * $section : the section number
4675 * &$sectionContent : ref to the content of the section
4676 * $showEditLinks : boolean describing whether this section has an edit link
4677 */
4681 }
4684 // append the TOC at the beginning
4685 // Top anchor now in skin
4687 }
4695 }
4696 }
4698 /**
4699 * Transform wiki markup when saving a page by doing "\r\n" -> "\n"
4700 * conversion, substitting signatures, {{subst:}} templates, etc.
4701 *
4702 * @param string $text The text to transform
4703 * @param Title $title The Title object for the current article
4704 * @param User $user The User object describing the current user
4705 * @param ParserOptions $options Parsing options
4706 * @param bool $clearState Whether to clear the parser state first
4707 * @return string The altered wiki markup
4708 */
4711 ) {
4714 }
4720 );
4724 }
4730 }
4732 /**
4733 * Pre-save transform helper function
4734 *
4735 * @param string $text
4736 * @param User $user
4737 *
4738 * @return string
4739 */
4743 # Note: This is the timestamp saved as hardcoded wikitext to
4744 # the database, we use $wgContLang here in order to give
4745 # everyone the same signature and use the default one rather
4746 # than the one selected in each user's preferences.
4747 # (see also bug 12815)
4753 # Allow translation of timezones through wiki. format() can return
4754 # whatever crap the system uses, localised or not, so we cannot
4755 # ship premade translations.
4760 }
4764 # Variable replacement
4765 # Because mOutputType is OT_WIKI, this will only process {{subst:xxx}} type tags
4768 # This works almost by chance, as the replaceVariables are done before the getUserSig(),
4769 # which may corrupt this parser instance via its wfMessage()->text() call-
4771 # Signatures
4777 ) );
4779 # Context links ("pipe tricks"): [[|name]] and [[name (context)|]]
4783 // [[ns:page (context)|]]
4785 // [[ns:page（context）|]] (double-width brackets, added in r40257)
4787 // [[ns:page (context), context|]] (using either single or double-width comma)
4789 // [[|page]] (reverse pipe trick: add context from page title)
4792 # try $p1 first, to turn "[[A, B (C)|]]" into "[[A, B (C)|A, B]]"
4804 # if there's no context, don't bother duplicating the title
4806 }
4808 # Trim trailing whitespace
4812 }
4814 /**
4815 * Fetch the user's signature text, if any, and normalize to
4816 * validated, ready-to-insert wikitext.
4817 * If you have pre-fetched the nickname or the fancySig option, you can
4818 * specify them here to save a database query.
4819 * Do not reuse this parser instance after calling getUserSig(),
4820 * as it may have changed if it's the $wgParser.
4821 *
4822 * @param User $user
4823 * @param string|bool $nickname Nickname to use or false to use user's default nickname
4824 * @param bool|null $fancySig whether the nicknname is the complete signature
4825 * or null to use default value
4826 * @return string
4827 */
4833 # If not given, retrieve from the user object.
4836 }
4840 }
4848 # Sig. might contain markup; validate this
4850 # Validated; clean up (if needed) and return it
4853 # Failed to validate; fall back to the default
4856 }
4857 }
4859 # Make sure nickname doesnt get a sig in a sig
4862 # If we're still here, make it a link to the user page
4869 }
4871 /**
4872 * Check that the user's signature contains no bad XML
4873 *
4874 * @param string $text
4875 * @return string|bool An expanded string, or false if invalid.
4876 */
4879 }
4881 /**
4882 * Clean up signature text
4883 *
4884 * 1) Strip ~~~, ~~~~ and ~~~~~ out of signatures @see cleanSigInSig
4885 * 2) Substitute all transclusions
4886 *
4887 * @param string $text
4888 * @param bool $parsing Whether we're cleaning (preferences save) or parsing
4889 * @return string Signature text
4890 */
4896 }
4898 # Option to disable this feature
4901 }
4903 # @todo FIXME: Regex doesn't respect extension tags or nowiki
4904 # => Move this logic to braceSubstitution()
4906 $substRegex = '/\{\{(?!(?:' . $substWord->getBaseRegex() . '))/x' . $substWord->getRegexCase();
4917 }
4920 }
4922 /**
4923 * Strip ~~~, ~~~~ and ~~~~~ out of signatures
4924 *
4925 * @param string $text
4926 * @return string Signature text with /~{3,5}/ removed
4927 */
4931 }
4933 /**
4934 * Set up some variables which are usually set up in parse()
4935 * so that an external function can call some class members with confidence
4936 *
4937 * @param Title|null $title
4938 * @param ParserOptions $options
4939 * @param int $outputType
4940 * @param bool $clearState
4941 */
4944 ) {
4946 }
4948 /**
4949 * @param Title|null $title
4950 * @param ParserOptions $options
4951 * @param int $outputType
4952 * @param bool $clearState
4953 */
4956 ) {
4962 }
4963 }
4965 /**
4966 * Wrapper for preprocess()
4967 *
4968 * @param string $text The text to preprocess
4969 * @param ParserOptions $options Options
4970 * @param Title|null $title Title object or null to use $wgTitle
4971 * @return string
4972 */
4976 # Guard against infinite recursion
4979 }
4986 }
4993 }
4995 /**
4996 * Create an HTML-style tag, e.g. "<yourtag>special text</yourtag>"
4997 * The callback should have the following form:
4998 * function myParserHook( $text, $params, $parser, $frame ) { ... }
4999 *
5000 * Transform and return $text. Use $parser for any required context, e.g. use
5001 * $parser->getTitle() and $parser->getOptions() not $wgTitle or $wgOut->mParserOptions
5002 *
5003 * Hooks may return extended information by returning an array, of which the
5004 * first numbered element (index 0) must be the return string, and all other
5005 * entries are extracted into local variables within an internal function
5006 * in the Parser class.
5007 *
5008 * This interface (introduced r61913) appears to be undocumented, but
5009 * 'markerName' is used by some core tag hooks to override which strip
5010 * array their results are placed in. **Use great caution if attempting
5011 * this interface, as it is not documented and injudicious use could smash
5012 * private variables.**
5013 *
5014 * @param string $tag The tag to use, e.g. 'hook' for "<hook>"
5015 * @param mixed $callback The callback function (and object) to use for the tag
5016 * @throws MWException
5017 * @return mixed|null The old value of the mTagHooks array associated with the hook
5018 */
5023 }
5028 }
5031 }
5033 /**
5034 * As setHook(), but letting the contents be parsed.
5035 *
5036 * Transparent tag hooks are like regular XML-style tag hooks, except they
5037 * operate late in the transformation sequence, on HTML instead of wikitext.
5038 *
5039 * This is probably obsoleted by things dealing with parser frames?
5040 * The only extension currently using it is geoserver.
5041 *
5042 * @since 1.10
5043 * @todo better document or deprecate this
5044 *
5045 * @param string $tag The tag to use, e.g. 'hook' for "<hook>"
5046 * @param mixed $callback The callback function (and object) to use for the tag
5047 * @throws MWException
5048 * @return mixed|null The old value of the mTagHooks array associated with the hook
5049 */
5054 }
5055 $oldVal = isset( $this->mTransparentTagHooks[$tag] ) ? $this->mTransparentTagHooks[$tag] : null;
5059 }
5061 /**
5062 * Remove all tag hooks
5063 */
5068 }
5070 /**
5071 * Create a function, e.g. {{sum:1|2|3}}
5072 * The callback function should have the form:
5073 * function myParserFunction( &$parser, $arg1, $arg2, $arg3 ) { ... }
5074 *
5075 * Or with SFH_OBJECT_ARGS:
5076 * function myParserFunction( $parser, $frame, $args ) { ... }
5077 *
5078 * The callback may either return the text result of the function, or an array with the text
5079 * in element 0, and a number of flags in the other elements. The names of the flags are
5080 * specified in the keys. Valid flags are:
5081 * found The text returned is valid, stop processing the template. This
5082 * is on by default.
5083 * nowiki Wiki markup in the return value should be escaped
5084 * isHTML The returned text is HTML, armour it against wikitext transformation
5085 *
5086 * @param string $id The magic word ID
5087 * @param mixed $callback The callback function (and object) to use
5088 * @param int $flags A combination of the following flags:
5089 * SFH_NO_HASH No leading hash, i.e. {{plural:...}} instead of {{#if:...}}
5090 *
5091 * SFH_OBJECT_ARGS Pass the template arguments as PPNode objects instead of text. This
5092 * allows for conditional expansion of the parse tree, allowing you to eliminate dead
5093 * branches and thus speed up parsing. It is also possible to analyse the parse tree of
5094 * the arguments, and to control the way they are expanded.
5095 *
5096 * The $frame parameter is a PPFrame. This can be used to produce expanded text from the
5097 * arguments, for instance:
5098 * $text = isset( $args[0] ) ? $frame->expand( $args[0] ) : '';
5099 *
5100 * For technical reasons, $args[0] is pre-expanded and will be a string. This may change in
5101 * future versions. Please call $frame->expand() on it anyway so that your code keeps
5102 * working if/when this is changed.
5103 *
5104 * If you want whitespace to be trimmed from $args, you need to do it yourself, post-
5105 * expansion.
5106 *
5107 * Please read the documentation in includes/parser/Preprocessor.php for more information
5108 * about the methods available in PPFrame and PPNode.
5109 *
5110 * @throws MWException
5111 * @return string|callable The old callback function for this name, if any
5112 */
5119 # Add to function cache
5123 }
5129 # Case
5132 }
5133 # Add leading hash
5136 }
5137 # Remove trailing colon
5140 }
5142 }
5144 }
5146 /**
5147 * Get all registered function hook identifiers
5148 *
5149 * @return array
5150 */
5153 }
5155 /**
5156 * Create a tag function, e.g. "<test>some stuff</test>".
5157 * Unlike tag hooks, tag functions are parsed at preprocessor level.
5158 * Unlike parser functions, their content is not preprocessed.
5159 * @param string $tag
5160 * @param callable $callback
5161 * @param int $flags
5162 * @throws MWException
5163 * @return null
5164 */
5169 }
5176 }
5179 }
5181 /**
5182 * @todo FIXME: Update documentation. makeLinkObj() is deprecated.
5183 * Replace "<!--LINK-->" link placeholders with actual links, in the buffer
5184 * Placeholders created in Skin::makeLinkObj()
5185 *
5186 * @param string $text
5187 * @param int $options
5188 *
5189 * @return array Array of link CSS classes, indexed by PDBK.
5190 */
5193 }
5195 /**
5196 * Replace "<!--LINK-->" link placeholders with plain text of links
5197 * (not HTML-formatted).
5198 *
5199 * @param string $text
5200 * @return string
5201 */
5204 }
5206 /**
5207 * Renders an image gallery from a text with one line per image.
5208 * text labels may be given by using |-style alternative text. E.g.
5209 * Image:one.jpg|The number "1"
5210 * Image:tree.jpg|A tree
5211 * given as text will return the HTML of a gallery with two images,
5212 * labeled 'The number "1"' and
5213 * 'A tree'.
5214 *
5215 * @param string $text
5216 * @param array $params
5217 * @return string HTML
5218 */
5225 }
5230 // If invalid type set, fallback to default.
5232 }
5245 }
5251 }
5254 }
5257 }
5260 }
5267 # match lines like these:
5268 # Image:someimage.jpg|This is some image
5271 # Skip empty lines
5274 }
5278 }
5281 # Bogus title. Ignore these so we don't bomb out later.
5283 }
5285 # We need to get what handler the file uses, to figure out parameters.
5286 # Note, a hook can overide the file name, and chose an entirely different
5287 # file (which potentially could be of a different type and have different handler).
5292 # Don't register it now, as ImageGallery does that later.
5300 );
5303 // We don't want people to specify per-image widths.
5304 // Additionally the width parameter would need special casing anyhow.
5306 }
5316 // look for an |alt= definition while trying not to break existing
5317 // captions with multiple pipes (|) in it, until a more sensible grammar
5318 // is defined for images in galleries
5320 // FIXME: Doing recursiveTagParse at this stage, and the trim before
5321 // splitting on '|' is a bit odd, and different from makeImage.
5338 //check to see if link matches an absolute url, if not then it must be a wiki link.
5345 }
5346 }
5349 // Must be a handler specific parameter.
5353 // Guess not. Append it to the caption.
5356 }
5357 }
5360 // concatenate all other pipes
5362 }
5363 }
5364 // remove the first pipe
5366 }
5369 }
5373 }
5375 /**
5376 * @param string $handler
5377 * @return array
5378 */
5384 }
5386 # Initialise static lists
5393 );
5401 }
5402 }
5403 }
5405 # Add handler params
5411 }
5412 }
5415 }
5416 return array( $this->mImageParams[$handlerClass], $this->mImageParamsMagicArray[$handlerClass] );
5417 }
5419 /**
5420 * Parse image options text and use it to make an image
5421 *
5422 * @param Title $title
5423 * @param string $options
5424 * @param LinkHolderArray|bool $holders
5425 * @return string HTML
5426 */
5428 # Check if the options text is of the form "options|alt text"
5429 # Options are:
5430 # * thumbnail make a thumbnail with enlarge-icon and caption, alignment depends on lang
5431 # * left no resizing, just left align. label is used for alt= only
5432 # * right same, but right aligned
5433 # * none same, but not aligned
5434 # * ___px scale to ___ pixels width, no aligning. e.g. use in taxobox
5435 # * center center the image
5436 # * frame Keep original image size, no magnify-button.
5437 # * framed Same as "frame"
5438 # * frameless like 'thumb' but without a frame. Keeps user preferences for width
5439 # * upright reduce width for upright images, rounded to full __0 px
5440 # * border draw a 1px border around the image
5441 # * alt Text for HTML alt attribute (defaults to empty)
5442 # * class Set a class for img node
5443 # * link Set the target of the image link. Can be external, interwiki, or local
5444 # vertical-align values (no % or length right now):
5445 # * baseline
5446 # * sub
5447 # * super
5448 # * top
5449 # * text-top
5450 # * middle
5451 # * bottom
5452 # * text-bottom
5456 # Give extensions a chance to select the file revision for us
5461 # Fetch and register the file (file title may be different via hooks)
5464 # Get parameter map
5471 }
5473 # Process the input parameters
5485 # Special case; width and height come in one variable together
5493 }
5494 }
5500 }
5501 }
5502 # else no validation -- bug 13436
5505 # Validate handler parameter
5508 # Validate internal parameters
5513 # @todo FIXME: Possibly check validity here for
5514 # manualthumb? downstream behavior seems odd with
5515 # missing manual thumbs.
5532 }
5534 }
5542 }
5543 }
5548 // use first appearing option, discard others.
5553 # Most other things appear to be empty or numeric...
5555 }
5556 }
5560 }
5561 }
5562 }
5565 }
5566 }
5568 # Process alignment parameters
5571 }
5574 }
5578 # Will the image be presented in a frame, with the caption below?
5584 # In the old days, [[Image:Foo|text...]] would set alt text. Later it
5585 # came to also set the caption, ordinary text after the image -- which
5586 # makes no sense, because that just repeats the text multiple times in
5587 # screen readers. It *also* came to set the title attribute.
5588 #
5589 # Now that we have an alt attribute, we should not set the alt text to
5590 # equal the caption: that's worse than useless, it just repeats the
5591 # text. This is the framed/thumbnail case. If there's no caption, we
5592 # use the unnamed parameter for alt text as well, just for the time be-
5593 # ing, if the unnamed param is set and the alt param is not.
5594 #
5595 # For the future, we need to figure out if we want to tweak this more,
5596 # e.g., introducing a title= parameter for the title; ignoring the un-
5597 # named parameter entirely for images without a caption; adding an ex-
5598 # plicit caption= parameter and preserving the old magic unnamed para-
5599 # meter for BC; ...
5602 # No caption or alt text, add the filename as the alt text so
5603 # that screen readers at least get some description of the image
5605 }
5606 # Do not set $params['frame']['title'] because tooltips don't make sense
5607 # for framed images
5610 # No alt text, use the "caption" for the alt text
5614 # No caption, fall back to using the filename for the
5615 # alt text
5617 }
5618 }
5619 # Use the "caption" for the tooltip text
5621 }
5625 # Linker does the rest
5630 # Give the handler a chance to modify the parser object
5633 }
5636 }
5638 /**
5639 * @param string $caption
5640 * @param LinkHolderArray|bool $holders
5641 * @return mixed|string
5642 */
5644 # Strip bad stuff out of the title (tooltip). We can't just use
5645 # replaceLinkHoldersText() here, because if this function is called
5646 # from replaceInternalLinks2(), mLinkHolders won't be up-to-date.
5651 }
5653 # make sure there are no placeholders in thumbnail attributes
5654 # that are later expanded to html- so expand them now and
5655 # remove the tags
5660 }
5662 /**
5663 * Set a flag in the output object indicating that the content is dynamic and
5664 * shouldn't be cached.
5665 */
5671 }
5674 }
5676 /**
5677 * Callback from the Sanitizer for expanding items found in HTML attribute
5678 * values, so they can be safely tested and escaped.
5679 *
5680 * @param string $text
5681 * @param bool|PPFrame $frame
5682 * @return string
5683 */
5688 }
5690 /**
5691 * Accessor
5692 *
5693 * @return array
5694 */
5700 );
5701 }
5703 /**
5704 * Replace transparent tags in $text with the values given by the callbacks.
5705 *
5706 * Transparent tag hooks are like regular XML-style tag hooks, except they
5707 * operate late in the transformation sequence, on HTML instead of wikitext.
5708 *
5709 * @param string $text
5710 *
5711 * @return string
5712 */
5726 );
5729 }
5731 }
5733 }
5735 /**
5736 * Break wikitext input into sections, and either pull or replace
5737 * some particular section's text.
5738 *
5739 * External callers should use the getSection and replaceSection methods.
5740 *
5741 * @param string $text Page wikitext
5742 * @param string $section A section identifier string of the form:
5743 * "<flag1> - <flag2> - ... - <section number>"
5744 *
5745 * Currently the only recognised flag is "T", which means the target section number
5746 * was derived during a template inclusion parse, in other words this is a template
5747 * section edit link. If no flags are given, it was an ordinary section edit link.
5748 * This flag is required to avoid a section numbering mismatch when a section is
5749 * enclosed by "<includeonly>" (bug 6563).
5750 *
5751 * The section number 0 pulls the text before the first heading; other numbers will
5752 * pull the given section along with its lower-level subsections. If the section is
5753 * not found, $mode=get will return $newtext, and $mode=replace will return $text.
5754 *
5755 * Section 0 is always considered to exist, even if it only contains the empty
5756 * string. If $text is the empty string and section 0 is replaced, $newText is
5757 * returned.
5758 *
5759 * @param string $mode One of "get" or "replace"
5760 * @param string $newText Replacement text for section data.
5761 * @return string For "get", the extracted section text.
5762 * for "replace", the whole page with the section replaced.
5763 */
5772 # Process section extraction flags
5779 }
5780 }
5782 # Check for empty input
5784 # Only sections 0 and T-0 exist in an empty document
5790 }
5796 }
5797 }
5798 }
5800 # Preprocess the text
5803 # <h> nodes indicate section breaks
5804 # They can only occur at the top level, so we can find them by iterating the root's children
5807 # Find the target section
5809 # Section zero doesn't nest, level=big
5818 }
5819 }
5822 }
5824 }
5825 }
5828 # Not found
5833 }
5834 }
5836 # Find the end of the section, including nested sections
5843 }
5844 }
5847 }
5851 # Write out the remainder (in replace mode only)
5853 # Output the replacement text
5854 # Add two newlines on -- trailing whitespace in $newText is conventionally
5855 # stripped by the editor, so we need both newlines to restore the paragraph gap
5856 # Only add trailing whitespace if there is newText
5859 }
5864 }
5865 }
5868 # Re-insert stripped tags
5870 }
5873 }
5875 /**
5876 * This function returns the text of a section, specified by a number ($section).
5877 * A section is text under a heading like == Heading == or \<h1\>Heading\</h1\>, or
5878 * the first section before any such heading (section 0).
5879 *
5880 * If a section contains subsections, these are also returned.
5881 *
5882 * @param string $text Text to look in
5883 * @param string $section Section identifier
5884 * @param string $deftext Default to return if section is not found
5885 * @return string Text of the requested section
5886 */
5889 }
5891 /**
5892 * This function returns $oldtext after the content of the section
5893 * specified by $section has been replaced with $text. If the target
5894 * section does not exist, $oldtext is returned unchanged.
5895 *
5896 * @param string $oldtext Former text of the article
5897 * @param int $section Section identifier
5898 * @param string $text Replacing text
5899 * @return string Modified text
5900 */
5903 }
5905 /**
5906 * Get the ID of the revision we are parsing
5907 *
5908 * @return int|null
5909 */
5912 }
5914 /**
5915 * Get the revision object for $this->mRevisionId
5916 *
5917 * @return Revision|null Either a Revision object or null
5918 * @since 1.23 (public since 1.23)
5919 */
5923 }
5926 }
5930 }
5932 /**
5933 * Get the timestamp associated with the current revision, adjusted for
5934 * the default server-local timestamp
5935 * @return string
5936 */
5946 # The cryptic '' timezone parameter tells to use the site-default
5947 # timezone offset instead of the user settings.
5948 #
5949 # Since this value will be saved into the parser cache, served
5950 # to other users, and potentially even used inside links and such,
5951 # it needs to be consistent for all visitors.
5955 }
5957 }
5959 /**
5960 * Get the name of the user that edited the last revision
5961 *
5962 * @return string User name
5963 */
5968 # if this template is subst: the revision id will be blank,
5969 # so just use the current user's name
5974 }
5975 }
5977 }
5979 /**
5980 * Get the size of the revision
5981 *
5982 * @return int|null Revision size
5983 */
5988 # if this variable is subst: the revision id will be blank,
5989 # so just use the parser input size, because the own substituation
5990 # will change the size.
5995 }
5996 }
5998 }
6000 /**
6001 * Mutator for $mDefaultSort
6002 *
6003 * @param string $sort New value
6004 */
6008 }
6010 /**
6011 * Accessor for $mDefaultSort
6012 * Will use the empty string if none is set.
6013 *
6014 * This value is treated as a prefix, so the
6015 * empty string is equivalent to sorting by
6016 * page name.
6017 *
6018 * @return string
6019 */
6025 }
6026 }
6028 /**
6029 * Accessor for $mDefaultSort
6030 * Unlike getDefaultSort(), will return false if none is set
6031 *
6032 * @return string|bool
6033 */
6036 }
6038 /**
6039 * Try to guess the section anchor name based on a wikitext fragment
6040 * presumably extracted from a heading, for example "Header" from
6041 * "== Header ==".
6042 *
6043 * @param string $text
6044 *
6045 * @return string
6046 */
6048 # Strip out wikitext links(they break the anchor)
6052 }
6054 /**
6055 * Same as guessSectionNameFromWikiText(), but produces legacy anchors
6056 * instead. For use in redirects, since IE6 interprets Redirect: headers
6057 * as something other than UTF-8 (apparently?), resulting in breakage.
6058 *
6059 * @param string $text The section name
6060 * @return string An anchor
6061 */
6063 # Strip out wikitext links(they break the anchor)
6067 }
6069 /**
6070 * Strips a text string of wikitext for use in a section anchor
6071 *
6072 * Accepts a text string and then removes all wikitext from the
6073 * string and leaves only the resultant text (i.e. the result of
6074 * [[User:WikiSysop|Sysop]] would be "Sysop" and the result of
6075 * [[User:WikiSysop]] would be "User:WikiSysop") - this is intended
6076 * to create valid section anchors by mimicing the output of the
6077 * parser when headings are parsed.
6078 *
6079 * @param string $text Text string to be stripped of wikitext
6080 * for use in a Section anchor
6081 * @return string Filtered text string
6082 */
6084 # Strip internal link markup
6088 # Strip external link markup
6089 # @todo FIXME: Not tolerant to blank link text
6090 # I.E. [https://www.mediawiki.org] will render as [1] or something depending
6091 # on how many empty links there are on the page - need to figure that out.
6092 $text = preg_replace( '/\[(?i:' . $this->mUrlProtocols . ')([^ ]+?) ([^[]+)\]/', '$2', $text );
6094 # Parse wikitext quotes (italics & bold)
6097 # Strip HTML tags
6100 }
6102 /**
6103 * strip/replaceVariables/unstrip for preprocessor regression testing
6104 *
6105 * @param string $text
6106 * @param Title $title
6107 * @param ParserOptions $options
6108 * @param int $outputType
6109 *
6110 * @return string
6111 */
6112 function testSrvus( $text, Title $title, ParserOptions $options, $outputType = self::OT_HTML ) {
6120 }
6122 /**
6123 * @param string $text
6124 * @param Title $title
6125 * @param ParserOptions $options
6126 * @return string
6127 */
6130 }
6132 /**
6133 * @param string $text
6134 * @param Title $title
6135 * @param ParserOptions $options
6136 * @return string
6137 */
6140 }
6142 /**
6143 * Call a callback function on all regions of the given text that are not
6144 * inside strip markers, and replace those regions with the return value
6145 * of the callback. For example, with input:
6146 *
6147 * aaa<MARKER>bbb
6148 *
6149 * This will call the callback function twice, with 'aaa' and 'bbb'. Those
6150 * two strings will be replaced with the value returned by the callback in
6151 * each case.
6152 *
6153 * @param string $s
6154 * @param callable $callback
6155 *
6156 * @return string
6157 */
6176 }
6177 }
6178 }
6180 }
6182 /**
6183 * Remove any strip markers found in the given text.
6184 *
6185 * @param string $text Input string
6186 * @return string
6187 */
6190 }
6192 /**
6193 * Save the parser state required to convert the given half-parsed text to
6194 * HTML. "Half-parsed" in this context means the output of
6195 * recursiveTagParse() or internalParse(). This output has strip markers
6196 * from replaceVariables (extensionSubstitution() etc.), and link
6197 * placeholders from replaceLinkHolders().
6198 *
6199 * Returns an array which can be serialized and stored persistently. This
6200 * array can later be loaded into another parser instance with
6201 * unserializeHalfParsedText(). The text can then be safely incorporated into
6202 * the return value of a parser hook.
6203 *
6204 * @param string $text
6205 *
6206 * @return array
6207 */
6215 );
6218 }
6220 /**
6221 * Load the parser state given in the $data array, which is assumed to
6222 * have been generated by serializeHalfParsedText(). The text contents is
6223 * extracted from the array, and its markers are transformed into markers
6224 * appropriate for the current Parser instance. This transformed text is
6225 * returned, and can be safely included in the return value of a parser
6226 * hook.
6227 *
6228 * If the $data array has been stored persistently, the caller should first
6229 * check whether it is still valid, by calling isValidHalfParsedText().
6230 *
6231 * @param array $data Serialized data
6232 * @throws MWException
6233 * @return string
6234 */
6238 }
6240 # First, extract the strip state.
6244 # Now renumber links
6247 # Should be good to go.
6249 }
6251 /**
6252 * Returns true if the given array, presumed to be generated by
6253 * serializeHalfParsedText(), is compatible with the current version of the
6254 * parser.
6255 *
6256 * @param array $data
6257 *
6258 * @return bool
6259 */
6262 }
6264 /**
6265 * Parsed a width param of imagelink like 300px or 200x300px
6266 *
6267 * @param string $value
6268 *
6269 * @return array
6270 * @since 1.20
6271 */
6276 }
6278 # (bug 13500) In both cases (width/height and width only),
6279 # permit trailing "px" for backward compatibility.
6288 }
6290 }
6292 /**
6293 * Lock the current instance of the parser.
6294 *
6295 * This is meant to stop someone from calling the parser
6296 * recursively and messing up all the strip state.
6297 *
6298 * @throws MWException If parser is in a parse
6299 * @return ScopedCallback The lock will be released once the return value goes out of scope.
6300 */
6305 }
6311 } );
6314 }
6316 /**
6317 * Strip outer <p></p> tag from the HTML source of a single paragraph.
6318 *
6319 * Returns original HTML if the <p/> tag has any attributes, if there's no wrapping <p/> tag,
6320 * or if there is more than one <p/> tag in the input HTML.
6321 *
6322 * @param string $html
6323 * @return string
6324 * @since 1.24
6325 */
6331 }
6332 }
6335 }
6336 }