2 rcs_id('$Id: InlineParser.php,v 1.65 2005-03-27 18:24:17 rurban Exp $');
3 /* Copyright (C) 2002 Geoffrey T. Dairiki <dairiki@dairiki.org>
4 * Copyright (C) 2004,2005 Reini Urban
6 * This file is part of PhpWiki.
8 * PhpWiki is free software; you can redistribute it and/or modify
9 * it under the terms of the GNU General Public License as published by
10 * the Free Software Foundation; either version 2 of the License, or
11 * (at your option) any later version.
13 * PhpWiki is distributed in the hope that it will be useful,
14 * but WITHOUT ANY WARRANTY; without even the implied warranty of
15 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
16 * GNU General Public License for more details.
18 * You should have received a copy of the GNU General Public License
19 * along with PhpWiki; if not, write to the Free Software
20 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
23 * This is the code which deals with the inline part of the (new-style)
27 * @author Geoffrey T. Dairiki
33 * This is the character used in wiki markup to escape characters with
36 define('ESCAPE_CHAR', '~');
38 require_once(dirname(__FILE__).'/HtmlElement.php');
39 require_once('lib/CachedMarkup.php');
40 require_once(dirname(__FILE__).'/stdlib.php');
43 function WikiEscape($text) {
44 return str_replace('#', ESCAPE_CHAR . '#', $text);
47 function UnWikiEscape($text) {
48 return preg_replace('/' . ESCAPE_CHAR . '(.)/', '\1', $text);
52 * Return type from RegexpSet::match and RegexpSet::nextMatch.
56 class RegexpSet_match {
58 * The text leading up the the next match.
66 * The text following the matched text.
70 * Index of the regular expression which matched.
76 * A set of regular expressions.
78 * This class is probably only useful for InlineTransformer.
84 * @param array $regexps A list of regular expressions. The
85 * regular expressions should not include any sub-pattern groups
86 * "(...)". (Anonymous groups, like "(?:...)", as well as
87 * look-ahead and look-behind assertions are okay.)
89 function RegexpSet ($regexps) {
91 $this->_regexps = array_unique($regexps);
92 if (!defined('_INLINE_OPTIMIZATION')) define('_INLINE_OPTIMIZATION',0);
96 * Search text for the next matching regexp from the Regexp Set.
98 * @param string $text The text to search.
100 * @return RegexpSet_match A RegexpSet_match object, or false if no match.
102 function match ($text) {
103 return $this->_match($text, $this->_regexps, '*?');
107 * Search for next matching regexp.
109 * Here, 'next' has two meanings:
111 * Match the next regexp(s) in the set, at the same position as the last match.
113 * If that fails, match the whole RegexpSet, starting after the position of the
116 * @param string $text Text to search.
118 * @param RegexpSet_match $prevMatch A RegexpSet_match object.
119 * $prevMatch should be a match object obtained by a previous
120 * match upon the same value of $text.
122 * @return RegexpSet_match A RegexpSet_match object, or false if no match.
124 function nextMatch ($text, $prevMatch) {
125 // Try to find match at same position.
126 $pos = strlen($prevMatch->prematch);
127 $regexps = array_slice($this->_regexps, $prevMatch->regexp_ind + 1);
129 $repeat = sprintf('{%d}', $pos);
130 if ( ($match = $this->_match($text, $regexps, $repeat)) ) {
131 $match->regexp_ind += $prevMatch->regexp_ind + 1;
137 // Failed. Look for match after current position.
138 $repeat = sprintf('{%d,}?', $pos + 1);
139 return $this->_match($text, $this->_regexps, $repeat);
142 // Syntax: http://www.pcre.org/pcre.txt
143 // x - EXTENDED, ignore whitespace
147 function _match ($text, $regexps, $repeat) {
148 // If one of the regexps is an empty string, php will crash here:
149 // sf.net: Fatal error: Allowed memory size of 8388608 bytes exhausted
150 // (tried to allocate 634 bytes)
151 if (_INLINE_OPTIMIZATION) { // disabled, wrong
152 // So we try to minize memory usage, by looping explicitly,
153 // and storing only those regexp which actually match.
154 // There may be more than one, so we have to find the longest,
155 // and match inside until the shortest is empty.
156 $matched = array(); $matched_ind = array();
157 for ($i=0; $i<count($regexps); $i++) {
158 if (!trim($regexps[$i])) {
159 trigger_error("empty regexp $i", E_USER_WARNING);
162 $pat= "/ ( . $repeat ) ( " . $regexps[$i] . " ) /x";
163 if (preg_match($pat, $text, $_m)) {
164 $m = $_m; // FIXME: prematch, postmatch is wrong
165 $matched[] = $regexps[$i];
170 // To overcome ANCHORED:
171 // We could sort by longest match and iterate over these.
172 if (empty($matched)) return false;
174 $match = new RegexpSet_match;
176 // Optimization: if the matches are only "$" and another, then omit "$"
177 if (! _INLINE_OPTIMIZATION or count($matched) > 2) {
178 assert(!empty($repeat));
179 assert(!empty($regexps));
180 for ($i=0; $i<count($regexps); $i++) {
181 if (!trim($regexps[$i])) {
182 trigger_error("empty regexp $i", E_USER_WARNING);
183 $regexps[$i] = '\Wxxxx\w\W\w\W\w\W\w\W\w\W\w'; // some placeholder
186 // We could do much better, if we would know the matching markup for the
187 // longest regexp match:
188 $hugepat= "/ ( . $repeat ) ( (" . join(')|(', $regexps) . ") ) /Asx";
189 // Proposed premature optimization 1:
190 //$hugepat= "/ ( . $repeat ) ( (" . join(')|(', array_values($matched)) . ") ) /Asx";
191 if (! preg_match($hugepat, $text, $m)) {
194 // Proposed premature optimization 1:
195 //$match->regexp_ind = $matched_ind[count($m) - 4];
196 $match->regexp_ind = count($m) - 4;
198 $match->regexp_ind = $regexp_ind;
201 $match->postmatch = substr($text, strlen($m[0]));
202 $match->prematch = $m[1];
203 $match->match = $m[2];
208 var_dump($regexps); var_dump($matched); var_dump($matched_inc);
209 PrintXML(HTML::dl(HTML::dt("input"),
210 HTML::dd(HTML::pre($text)),
212 HTML::dd(HTML::pre($match->regexp_ind, ":", $regexps[$match->regexp_ind])),
213 HTML::dt("prematch"),
214 HTML::dd(HTML::pre($match->prematch)),
216 HTML::dd(HTML::pre($match->match)),
217 HTML::dt("postmatch"),
218 HTML::dd(HTML::pre($match->postmatch))
229 * A simple markup rule (i.e. terminal token).
231 * These are defined by a regexp.
233 * When a match is found for the regexp, the matching text is replaced.
234 * The replacement content is obtained by calling the SimpleMarkup::markup method.
242 * @return string Regexp which matches this token.
244 function getMatchRegexp () {
245 return $this->_match_regexp;
248 /** Markup matching text.
250 * @param string $match The text which matched the regexp
251 * (obtained from getMatchRegexp).
253 * @return mixed The expansion of the matched text.
255 function markup ($match /*, $body */) {
256 trigger_error("pure virtual", E_USER_ERROR);
261 * A balanced markup rule.
263 * These are defined by a start regexp, and an end regexp.
269 /** Get the starting regexp for this rule.
271 * @return string The starting regexp.
273 function getStartRegexp () {
274 return $this->_start_regexp;
277 /** Get the ending regexp for this rule.
279 * @param string $match The text which matched the starting regexp.
281 * @return string The ending regexp.
283 function getEndRegexp ($match) {
284 return $this->_end_regexp;
287 /** Get expansion for matching input.
289 * @param string $match The text which matched the starting regexp.
291 * @param mixed $body Transformed text found between the starting
292 * and ending regexps.
294 * @return mixed The expansion of the matched text.
296 function markup ($match, $body) {
297 trigger_error("pure virtual", E_USER_ERROR);
301 class Markup_escape extends SimpleMarkup
303 function getMatchRegexp () {
304 return ESCAPE_CHAR . '(?: [[:alnum:]]+ | .)';
307 function markup ($match) {
308 assert(strlen($match) >= 2);
309 return substr($match, 1);
314 * [image.jpg size=50% border=5], [image.jpg size=50x30]
315 * Support for the following attributes: see stdlib.php:LinkImage()
316 * size=<precent>%, size=<width>x<height>
317 * border=n, align=\w+, hspace=n, vspace=n
319 function isImageLink($link) {
320 if (!$link) return false;
321 assert(defined('INLINE_IMAGES'));
322 return preg_match("/\\.(" . INLINE_IMAGES . ")$/i", $link)
323 or preg_match("/\\.(" . INLINE_IMAGES . ")\s+(size|border|align|hspace|vspace)=/i", $link);
326 function LinkBracketLink($bracketlink) {
328 // $bracketlink will start and end with brackets; in between will
329 // be either a page name, a URL or both separated by a pipe.
331 // strip brackets and leading space
332 // FIXME: \n inside [] will lead to errors
333 preg_match('/(\#?) \[\s* (?: (.*?) \s* (?<!' . ESCAPE_CHAR . ')(\|) )? \s* (.+?) \s*\]/x',
334 $bracketlink, $matches);
335 if (count($matches) < 4) {
336 trigger_error(_("Invalid [] syntax ignored").": ".$bracketlink, E_USER_NOTICE);
337 return new Cached_Link;
339 list (, $hash, $label, $bar, $rawlink) = $matches;
341 $label = UnWikiEscape($label);
343 * Check if the user has typed a explicit URL. This solves the
344 * problem where the URLs have a ~ character, which would be stripped away.
345 * "[http:/server/~name/]" will work as expected
346 * "http:/server/~name/" will NOT work as expected, will remove the ~
348 if (strstr($rawlink, "http://") or strstr($rawlink, "https://")) {
350 // Mozilla Browser URI Obfuscation Weakness 2004-06-14
351 // http://www.securityfocus.com/bid/10532/
352 // goodurl+"%2F%20%20%20."+badurl
353 if (preg_match("/%2F(%20)+\./i", $rawlink)) {
354 $rawlink = preg_replace("/%2F(%20)+\./i","%2F.",$rawlink);
357 $link = UnWikiEscape($rawlink);
360 // if label looks like a url to an image, we want an image link.
361 if (isImageLink($label)) {
363 $intermap = getInterwikiMap();
364 if (preg_match("/^" . $intermap->getRegexp() . ":/", $label)) {
365 $imgurl = $intermap->link($label);
366 $imgurl = $imgurl->getAttr('href');
367 } elseif (! preg_match("#^(" . ALLOWED_PROTOCOLS . "):#", $imgurl)) {
368 // local theme linkname like 'images/next.gif'.
370 $imgurl = $WikiTheme->getImageURL($imgurl);
372 $label = LinkImage($imgurl, $link);
376 // It's an anchor, not a link...
377 $id = MangleXmlIdentifier($link);
378 return HTML::a(array('name' => $id, 'id' => $id),
379 $bar ? $label : $link);
382 if (preg_match("#^(" . ALLOWED_PROTOCOLS . "):#", $link)) {
383 // if it's an image, embed it; otherwise, it's a regular link
384 if (isImageLink($link))
385 return LinkImage($link, $label);
387 return new Cached_ExternalLink($link, $label);
389 elseif (preg_match("/^phpwiki:/", $link))
390 return new Cached_PhpwikiURL($link, $label);
392 * Inline images in Interwiki urls's:
393 * [File:my_image.gif] inlines the image,
394 * File:my_image.gif shows a plain inter-wiki link,
395 * [what a pic|File:my_image.gif] shows a named inter-wiki link to the gif
396 * [File:my_image.gif|what a pic] shows a inlimed image linked to the page "what a pic"
398 elseif (strstr($link,':') and
399 ($intermap = getInterwikiMap()) and
400 preg_match("/^" . $intermap->getRegexp() . ":/", $link)) {
401 if (empty($label) && isImageLink($link)) {
402 // if without label => inlined image [File:xx.gif]
403 $imgurl = $intermap->link($link);
404 return LinkImage($imgurl->getAttr('href'), $label);
406 return new Cached_InterwikiLink($link, $label);
408 // Split anchor off end of pagename.
409 if (preg_match('/\A(.*)(?<!'.ESCAPE_CHAR.')#(.*?)\Z/', $rawlink, $m)) {
410 list(,$rawlink,$anchor) = $m;
411 $pagename = UnWikiEscape($rawlink);
412 $anchor = UnWikiEscape($anchor);
420 return new Cached_WikiLink($pagename, $label, $anchor);
424 class Markup_bracketlink extends SimpleMarkup
426 var $_match_regexp = "\\#? \\[ .*? [^]\\s] .*? \\]";
428 function markup ($match) {
429 $link = LinkBracketLink($match);
430 assert($link->isInlineElement());
435 class Markup_url extends SimpleMarkup
437 function getMatchRegexp () {
438 return "(?<![[:alnum:]]) (?:" . ALLOWED_PROTOCOLS . ") : [^\s<>\"']+ (?<![ ,.?; \] \) ])";
441 function markup ($match) {
442 return new Cached_ExternalLink(UnWikiEscape($match));
447 class Markup_interwiki extends SimpleMarkup
449 function getMatchRegexp () {
451 $map = getInterwikiMap();
452 return "(?<! [[:alnum:]])" . $map->getRegexp(). ": \S+ (?<![ ,.?;! \] \) \" \' ])";
455 function markup ($match) {
456 //$map = getInterwikiMap();
457 return new Cached_InterwikiLink(UnWikiEscape($match));
461 class Markup_wikiword extends SimpleMarkup
463 function getMatchRegexp () {
464 global $WikiNameRegexp;
465 if (!trim($WikiNameRegexp)) return " " . WIKI_NAME_REGEXP;
466 return " $WikiNameRegexp";
469 function markup ($match) {
470 if (!$match) return false;
471 if ($this->_isWikiUserPage($match))
472 return new Cached_UserLink($match); //$this->_UserLink($match);
474 return new Cached_WikiLink($match);
477 // FIXME: there's probably a more useful place to put these two functions
478 function _isWikiUserPage ($page) {
480 $dbi = $request->getDbh();
481 $page_handle = $dbi->getPage($page);
482 if ($page_handle and $page_handle->get('pref'))
488 function _UserLink($PageName) {
489 $link = HTML::a(array('href' => $PageName));
490 $link->pushContent(PossiblyGlueIconToText('wikiuser', $PageName));
491 $link->setAttr('class', 'wikiuser');
496 class Markup_linebreak extends SimpleMarkup
498 //var $_match_regexp = "(?: (?<! %) %%% (?! %) | <(?:br|BR)> | <(?:br|BR) \/> )";
499 var $_match_regexp = "(?: (?<! %) %%% (?! %) | <(?:br|BR)> )";
501 function markup ($match) {
506 class Markup_old_emphasis extends BalancedMarkup
508 var $_start_regexp = "''|__";
510 function getEndRegexp ($match) {
514 function markup ($match, $body) {
515 $tag = $match == "''" ? 'em' : 'strong';
516 return new HtmlElement($tag, $body);
520 class Markup_nestled_emphasis extends BalancedMarkup
522 function getStartRegexp() {
523 static $start_regexp = false;
525 if (!$start_regexp) {
526 // The three possible delimiters
527 // (none of which can be followed by itself.)
532 $any = "(?: ${i}|${b}|${tt})"; // any of the three.
534 // Any of [_*=] is okay if preceded by space or one of [-"'/:]
535 $start[] = "(?<= \\s|^|[-\"'\\/:]) ${any}";
537 // _ or * is okay after = as long as not immediately followed by =
538 $start[] = "(?<= =) (?: ${i}|${b}) (?! =)";
540 $start[] = "(?<= _) (?: ${b}|${tt}) (?! _)";
541 $start[] = "(?<= \\*) (?: ${i}|${tt}) (?! \\*)";
544 // any delimiter okay after an opening brace ( [{<(] )
545 // as long as it's not immediately followed by the matching closing
547 $start[] = "(?<= { ) ${any} (?! } )";
548 $start[] = "(?<= < ) ${any} (?! > )";
549 $start[] = "(?<= \\( ) ${any} (?! \\) )";
551 $start = "(?:" . join('|', $start) . ")";
553 // Any of the above must be immediately followed by non-whitespace.
554 $start_regexp = $start . "(?= \S)";
557 return $start_regexp;
560 function getEndRegexp ($match) {
561 $chr = preg_quote($match);
562 return "(?<= \S | ^ ) (?<! $chr) $chr (?! $chr) (?= \s | [-)}>\"'\\/:.,;!? _*=] | $)";
565 function markup ($match, $body) {
567 case '*': return new HtmlElement('b', $body);
568 case '=': return new HtmlElement('tt', $body);
569 case '_': return new HtmlElement('i', $body);
574 class Markup_html_emphasis extends BalancedMarkup
577 "<(?: b|big|i|small|tt|em|strong|cite|code|dfn|kbd|samp|var|sup|sub )>";
579 function getEndRegexp ($match) {
580 return "<\\/" . substr($match, 1);
583 function markup ($match, $body) {
584 $tag = substr($match, 1, -1);
585 return new HtmlElement($tag, $body);
589 class Markup_html_abbr extends BalancedMarkup
591 //rurban: abbr|acronym need an optional title tag.
593 var $_start_regexp = "<(?: abbr|acronym )(?: \stitle=[^>]*)?>";
595 function getEndRegexp ($match) {
596 if (substr($match,1,4) == 'abbr')
600 return "<\\/" . $tag . '>';
603 function markup ($match, $body) {
604 if (substr($match,1,4) == 'abbr')
608 $rest = substr($match,1+strlen($tag),-1);
610 list($key,$val) = explode("=",$rest);
611 $args = array($key => $val);
612 } else $args = array();
613 return new HtmlElement($tag, $args, $body);
617 // Special version for single-line plugins formatting,
618 // like: '<small>< ?plugin PopularNearby ? ></small>'
619 class Markup_plugin extends SimpleMarkup
621 var $_match_regexp = '<\?plugin(?:-form)?\s[^\n]+?\?>';
623 function markup ($match) {
624 //$xml = new Cached_PluginInvocation($match);
625 //$xml->setTightness(true,true);
626 return new Cached_PluginInvocation($match);
631 // TODO: "..." => "…" browser specific display (not cached?)
632 // TODO: "--" => "&emdash;" browser specific display (not cached?)
633 // TODO: Support more HTML::Entities: (C) for copy, --- for mdash, -- for ndash
635 class Markup_html_entities extends SimpleMarkup {
636 var $_match_regexp = '(: \.\.\.|\-\-|\-\-\-|\(C\) )';
638 function markup ($match) {
639 static $entities = array('...' => '…',
644 return HTML::Raw($entities[$match]);
648 class Markup_isonumchars extends SimpleMarkup {
649 var $_match_regexp = '\&\#\d{2,5};';
651 function markup ($match) {
652 return HTML::Raw($match);
656 class Markup_isohexchars extends SimpleMarkup {
657 // hexnums, like ¤ <=> ¤
658 var $_match_regexp = '\&\#x[0-9a-fA-F]{2,4};';
660 function markup ($match) {
661 return HTML::Raw($match);
665 // FIXME: Do away with magic phpwiki forms. (Maybe phpwiki: links too?)
666 // FIXME: Do away with plugin-links. They seem not to be used.
669 class InlineTransformer
671 var $_regexps = array();
672 var $_markup = array();
674 function InlineTransformer ($markup_types = false) {
676 $markup_types = array('escape', 'bracketlink', 'url',
677 'interwiki', 'wikiword', 'linebreak',
678 'old_emphasis', 'nestled_emphasis',
679 'html_emphasis', 'html_abbr', 'plugin',
680 'isonumchars', 'isohexchars', /*'html_entities',*/
682 foreach ($markup_types as $mtype) {
683 $class = "Markup_$mtype";
684 $this->_addMarkup(new $class);
688 function _addMarkup ($markup) {
689 if (isa($markup, 'SimpleMarkup'))
690 $regexp = $markup->getMatchRegexp();
692 $regexp = $markup->getStartRegexp();
694 assert(!isset($this->_markup[$regexp]));
695 $this->_regexps[] = $regexp;
696 $this->_markup[] = $markup;
699 function parse (&$text, $end_regexps = array('$')) {
700 $regexps = $this->_regexps;
702 // $end_re takes precedence: "favor reduce over shift"
703 array_unshift($regexps, $end_regexps[0]);
704 //array_push($regexps, $end_regexps[0]);
705 $regexps = new RegexpSet($regexps);
708 $output = new XmlContent;
710 $match = $regexps->match($input);
713 if ($match->regexp_ind == 0) {
714 // No start pattern found before end pattern.
716 if (isset($markup) and is_object($markup) and isa($markup,'Markup_plugin')) {
717 $current =& $output->_content[count($output->_content)-1];
718 $current->setTightness(true,true);
720 $output->pushContent($match->prematch);
721 $text = $match->postmatch;
725 $markup = $this->_markup[$match->regexp_ind - 1];
726 $body = $this->_parse_markup_body($markup, $match->match, $match->postmatch, $end_regexps);
728 // Couldn't match balanced expression.
729 // Ignore and look for next matching start regexp.
730 $match = $regexps->nextMatch($input, $match);
734 // Matched markup. Eat input, push output.
735 // FIXME: combine adjacent strings.
736 if (isa($markup, 'SimpleMarkup'))
737 $current = $markup->markup($match->match);
739 $current = $markup->markup($match->match, $body);
740 $input = $match->postmatch;
741 if (isset($markup) and is_object($markup) and isa($markup,'Markup_plugin')) {
742 $current->setTightness(true,true);
744 $output->pushContent($match->prematch, $current);
746 $match = $regexps->match($input);
749 // No pattern matched, not even the end pattern.
754 function _parse_markup_body ($markup, $match, &$text, $end_regexps) {
755 if (isa($markup, 'SimpleMarkup'))
756 return true; // Done. SimpleMarkup is simple.
758 if (!is_object($markup)) return false; // Some error: Should assert
759 array_unshift($end_regexps, $markup->getEndRegexp($match));
761 // Optimization: if no end pattern in text, we know the
762 // parse will fail. This is an important optimization,
763 // e.g. when text is "*lots *of *start *delims *with
764 // *no *matching *end *delims".
765 $ends_pat = "/(?:" . join(").*(?:", $end_regexps) . ")/xs";
766 if (!preg_match($ends_pat, $text))
768 return $this->parse($text, $end_regexps);
772 class LinkTransformer extends InlineTransformer
774 function LinkTransformer () {
775 $this->InlineTransformer(array('escape', 'bracketlink', 'url',
776 'interwiki', 'wikiword'));
780 function TransformInline($text, $markup = 2.0, $basepage=false) {
784 $trfm = new InlineTransformer;
788 $text = ConvertOldMarkup($text, 'inline');
792 return new CacheableMarkup($trfm->parse($text), $basepage);
794 return $trfm->parse($text);
797 function TransformLinks($text, $markup = 2.0, $basepage = false) {
801 $trfm = new LinkTransformer;
805 $text = ConvertOldMarkup($text, 'links');
809 return new CacheableMarkup($trfm->parse($text), $basepage);
811 return $trfm->parse($text);
814 // $Log: not supported by cvs2svn $
816 // (c-file-style: "gnu")
821 // c-hanging-comment-ender-p: nil
822 // indent-tabs-mode: nil