xref: /dokuwiki/inc/parserutils.php (revision 1172f8dcef2c8198ddcdaffcdf65a735811d20a3) 
1<?php
2/**
3 * Utilities for accessing the parser
4 *
5 * @license    GPL 2 (http://www.gnu.org/licenses/gpl.html)
6 * @author     Harry Fuecks <hfuecks@gmail.com>
7 * @author     Andreas Gohr <andi@splitbrain.org>
8 */
9
10if(!defined('DOKU_INC')) die('meh.');
11
12/**
13 * Returns the parsed Wikitext in XHTML for the given id and revision.
14 *
15 * If $excuse is true an explanation is returned if the file
16 * wasn't found
17 *
18 * @author Andreas Gohr <andi@splitbrain.org>
19 */
20function p_wiki_xhtml($id, $rev='', $excuse=true){
21    $file = wikiFN($id,$rev);
22    $ret  = '';
23
24    //ensure $id is in global $ID (needed for parsing)
25    global $ID;
26    $keep = $ID;
27    $ID   = $id;
28
29    if($rev){
30        if(@file_exists($file)){
31            $ret = p_render('xhtml',p_get_instructions(io_readWikiPage($file,$id,$rev)),$info); //no caching on old revisions
32        }elseif($excuse){
33            $ret = p_locale_xhtml('norev');
34        }
35    }else{
36        if(@file_exists($file)){
37            $ret = p_cached_output($file,'xhtml',$id);
38        }elseif($excuse){
39            $ret = p_locale_xhtml('newpage');
40        }
41    }
42
43    //restore ID (just in case)
44    $ID = $keep;
45
46    return $ret;
47}
48
49/**
50 * Returns starting summary for a page (e.g. the first few
51 * paragraphs), marked up in XHTML.
52 *
53 * If $excuse is true an explanation is returned if the file
54 * wasn't found
55 *
56 * @param string wiki page id
57 * @param reference populated with page title from heading or page id
58 * @deprecated
59 * @author Harry Fuecks <hfuecks@gmail.com>
60 */
61function p_wiki_xhtml_summary($id, &$title, $rev='', $excuse=true){
62    $file = wikiFN($id,$rev);
63    $ret  = '';
64
65    //ensure $id is in global $ID (needed for parsing)
66    global $ID;
67    $keep = $ID;
68    $ID   = $id;
69
70    if($rev){
71        if(@file_exists($file)){
72            //no caching on old revisions
73            $ins = p_get_instructions(io_readWikiPage($file,$id,$rev));
74        }elseif($excuse){
75            $ret = p_locale_xhtml('norev');
76            //restore ID (just in case)
77            $ID = $keep;
78            return $ret;
79        }
80
81    }else{
82
83        if(@file_exists($file)){
84            // The XHTML for a summary is not cached so use the instruction cache
85            $ins = p_cached_instructions($file);
86        }elseif($excuse){
87            $ret = p_locale_xhtml('newpage');
88            //restore ID (just in case)
89            $ID = $keep;
90            return $ret;
91        }
92    }
93
94    $ret = p_render('xhtmlsummary',$ins,$info);
95
96    if ( $info['sum_pagetitle'] ) {
97        $title = $info['sum_pagetitle'];
98    } else {
99        $title = $id;
100    }
101
102    $ID = $keep;
103    return $ret;
104}
105
106/**
107 * Returns the specified local text in parsed format
108 *
109 * @author Andreas Gohr <andi@splitbrain.org>
110 */
111function p_locale_xhtml($id){
112    //fetch parsed locale
113    $html = p_cached_output(localeFN($id));
114    return $html;
115}
116
117/**
118 *     *** DEPRECATED ***
119 *
120 * use p_cached_output()
121 *
122 * Returns the given file parsed to XHTML
123 *
124 * Uses and creates a cachefile
125 *
126 * @deprecated
127 * @author Andreas Gohr <andi@splitbrain.org>
128 * @todo   rewrite to use mode instead of hardcoded XHTML
129 */
130function p_cached_xhtml($file){
131    return p_cached_output($file);
132}
133
134/**
135 * Returns the given file parsed into the requested output format
136 *
137 * @author Andreas Gohr <andi@splitbrain.org>
138 * @author Chris Smith <chris@jalakai.co.uk>
139 */
140function p_cached_output($file, $format='xhtml', $id='') {
141    global $conf;
142
143    $cache = new cache_renderer($id, $file, $format);
144    if ($cache->useCache()) {
145        $parsed = $cache->retrieveCache(false);
146        if($conf['allowdebug'] && $format=='xhtml') $parsed .= "\n<!-- cachefile {$cache->cache} used -->\n";
147    } else {
148        $parsed = p_render($format, p_cached_instructions($file,false,$id), $info);
149
150        if ($info['cache']) {
151            $cache->storeCache($parsed);               //save cachefile
152            if($conf['allowdebug'] && $format=='xhtml') $parsed .= "\n<!-- no cachefile used, but created {$cache->cache} -->\n";
153        }else{
154            $cache->removeCache();                     //try to delete cachefile
155            if($conf['allowdebug'] && $format=='xhtml') $parsed .= "\n<!-- no cachefile used, caching forbidden -->\n";
156        }
157    }
158
159    return $parsed;
160}
161
162/**
163 * Returns the render instructions for a file
164 *
165 * Uses and creates a serialized cache file
166 *
167 * @author Andreas Gohr <andi@splitbrain.org>
168 */
169function p_cached_instructions($file,$cacheonly=false,$id='') {
170    global $conf;
171    static $run = null;
172    if(is_null($run)) $run = array();
173
174    $cache = new cache_instructions($id, $file);
175
176    if ($cacheonly || $cache->useCache() || isset($run[$file])) {
177        return $cache->retrieveCache();
178    } else if (@file_exists($file)) {
179        // no cache - do some work
180        $ins = p_get_instructions(io_readWikiPage($file,$id));
181        if ($cache->storeCache($ins)) {
182            $run[$file] = true; // we won't rebuild these instructions in the same run again
183        } else {
184            msg('Unable to save cache file. Hint: disk full; file permissions; safe_mode setting.',-1);
185        }
186        return $ins;
187    }
188
189    return null;
190}
191
192/**
193 * turns a page into a list of instructions
194 *
195 * @author Harry Fuecks <hfuecks@gmail.com>
196 * @author Andreas Gohr <andi@splitbrain.org>
197 */
198function p_get_instructions($text){
199
200    $modes = p_get_parsermodes();
201
202    // Create the parser
203    $Parser = new Doku_Parser();
204
205    // Add the Handler
206    $Parser->Handler = new Doku_Handler();
207
208    //add modes to parser
209    foreach($modes as $mode){
210        $Parser->addMode($mode['mode'],$mode['obj']);
211    }
212
213    // Do the parsing
214    trigger_event('PARSER_WIKITEXT_PREPROCESS', $text);
215    $p = $Parser->parse($text);
216    //  dbg($p);
217    return $p;
218}
219
220/**
221 * returns the metadata of a page
222 *
223 * @author Esther Brunner <esther@kaffeehaus.ch>
224 */
225function p_get_metadata($id, $key='', $render=false){
226    global $ID;
227
228    // cache the current page
229    // Benchmarking shows the current page's metadata is generally the only page metadata
230    // accessed several times. This may catch a few other pages, but that shouldn't be an issue.
231    $cache = ($ID == $id);
232    $meta = p_read_metadata($id, $cache);
233
234    // metadata has never been rendered before - do it! (but not for non-existent pages)
235    if ($render && !isset($meta['current']['description']['abstract']) && page_exists($id)){
236        $meta = p_render_metadata($id, $meta);
237        p_save_metadata($id, $meta);
238    }
239
240    $val = $meta['current'];
241
242    // filter by $key
243    foreach(preg_split('/\s+/', $key, 2, PREG_SPLIT_NO_EMPTY) as $cur_key) {
244        if (!isset($val[$cur_key])) {
245            return null;
246        }
247        $val = $val[$cur_key];
248    }
249    return $val;
250}
251
252/**
253 * sets metadata elements of a page
254 *
255 * @author Esther Brunner <esther@kaffeehaus.ch>
256 */
257function p_set_metadata($id, $data, $render=false, $persistent=true){
258    if (!is_array($data)) return false;
259
260    global $ID;
261
262    // cache the current page
263    $cache = ($ID == $id);
264    $orig = p_read_metadata($id, $cache);
265
266    // render metadata first?
267    $meta = $render ? p_render_metadata($id, $orig) : $orig;
268
269    // now add the passed metadata
270    $protected = array('description', 'date', 'contributor');
271    foreach ($data as $key => $value){
272
273        // be careful with sub-arrays of $meta['relation']
274        if ($key == 'relation'){
275
276            foreach ($value as $subkey => $subvalue){
277                $meta['current'][$key][$subkey] = !empty($meta['current'][$key][$subkey]) ? array_merge($meta['current'][$key][$subkey], $subvalue) : $subvalue;
278                if ($persistent)
279                    $meta['persistent'][$key][$subkey] = !empty($meta['persistent'][$key][$subkey]) ? array_merge($meta['persistent'][$key][$subkey], $subvalue) : $subvalue;
280            }
281
282            // be careful with some senisitive arrays of $meta
283        } elseif (in_array($key, $protected)){
284
285            // these keys, must have subkeys - a legitimate value must be an array
286            if (is_array($value)) {
287                $meta['current'][$key] = !empty($meta['current'][$key]) ? array_merge($meta['current'][$key],$value) : $value;
288
289                if ($persistent) {
290                    $meta['persistent'][$key] = !empty($meta['persistent'][$key]) ? array_merge($meta['persistent'][$key],$value) : $value;
291                }
292            }
293
294            // no special treatment for the rest
295        } else {
296            $meta['current'][$key] = $value;
297            if ($persistent) $meta['persistent'][$key] = $value;
298        }
299    }
300
301    // save only if metadata changed
302    if ($meta == $orig) return true;
303
304    return p_save_metadata($id, $meta);
305}
306
307/**
308 * Purges the non-persistant part of the meta data
309 * used on page deletion
310 *
311 * @author Michael Klier <chi@chimeric.de>
312 */
313function p_purge_metadata($id) {
314    $meta = p_read_metadata($id);
315    foreach($meta['current'] as $key => $value) {
316        if(is_array($meta[$key])) {
317            $meta['current'][$key] = array();
318        } else {
319            $meta['current'][$key] = '';
320        }
321
322    }
323    return p_save_metadata($id, $meta);
324}
325
326/**
327 * read the metadata from source/cache for $id
328 * (internal use only - called by p_get_metadata & p_set_metadata)
329 *
330 * @author   Christopher Smith <chris@jalakai.co.uk>
331 *
332 * @param    string   $id      absolute wiki page id
333 * @param    bool     $cache   whether or not to cache metadata in memory
334 *                             (only use for metadata likely to be accessed several times)
335 *
336 * @return   array             metadata
337 */
338function p_read_metadata($id,$cache=false) {
339    global $cache_metadata;
340
341    if (isset($cache_metadata[(string)$id])) return $cache_metadata[(string)$id];
342
343    $file = metaFN($id, '.meta');
344    $meta = @file_exists($file) ? unserialize(io_readFile($file, false)) : array('current'=>array(),'persistent'=>array());
345
346    if ($cache) {
347        $cache_metadata[(string)$id] = $meta;
348    }
349
350    return $meta;
351}
352
353/**
354 * This is the backend function to save a metadata array to a file
355 *
356 * @param    string   $id      absolute wiki page id
357 * @param    array    $meta    metadata
358 *
359 * @return   bool              success / fail
360 */
361function p_save_metadata($id, $meta) {
362    // sync cached copies, including $INFO metadata
363    global $cache_metadata, $INFO;
364
365    if (isset($cache_metadata[$id])) $cache_metadata[$id] = $meta;
366    if (!empty($INFO) && ($id == $INFO['id'])) { $INFO['meta'] = $meta['current']; }
367
368    return io_saveFile(metaFN($id, '.meta'), serialize($meta));
369}
370
371/**
372 * renders the metadata of a page
373 *
374 * @author Esther Brunner <esther@kaffeehaus.ch>
375 */
376function p_render_metadata($id, $orig){
377    // make sure the correct ID is in global ID
378    global $ID;
379    $keep = $ID;
380    $ID   = $id;
381
382    // add an extra key for the event - to tell event handlers the page whose metadata this is
383    $orig['page'] = $id;
384    $evt = new Doku_Event('PARSER_METADATA_RENDER', $orig);
385    if ($evt->advise_before()) {
386
387        require_once DOKU_INC."inc/parser/metadata.php";
388
389        // get instructions
390        $instructions = p_cached_instructions(wikiFN($id),false,$id);
391        if(is_null($instructions)){
392            $ID = $keep;
393            return null; // something went wrong with the instructions
394        }
395
396        // set up the renderer
397        $renderer = new Doku_Renderer_metadata();
398        $renderer->meta = $orig['current'];
399        $renderer->persistent = $orig['persistent'];
400
401        // loop through the instructions
402        foreach ($instructions as $instruction){
403            // execute the callback against the renderer
404            call_user_func_array(array(&$renderer, $instruction[0]), (array) $instruction[1]);
405        }
406
407        $evt->result = array('current'=>$renderer->meta,'persistent'=>$renderer->persistent);
408    }
409    $evt->advise_after();
410
411    $ID = $keep;
412    return $evt->result;
413}
414
415/**
416 * returns all available parser syntax modes in correct order
417 *
418 * @author Andreas Gohr <andi@splitbrain.org>
419 */
420function p_get_parsermodes(){
421    global $conf;
422
423    //reuse old data
424    static $modes = null;
425    if($modes != null){
426        return $modes;
427    }
428
429    //import parser classes and mode definitions
430    require_once DOKU_INC . 'inc/parser/parser.php';
431
432    // we now collect all syntax modes and their objects, then they will
433    // be sorted and added to the parser in correct order
434    $modes = array();
435
436    // add syntax plugins
437    $pluginlist = plugin_list('syntax');
438    if(count($pluginlist)){
439        global $PARSER_MODES;
440        $obj = null;
441        foreach($pluginlist as $p){
442            if(!$obj =& plugin_load('syntax',$p)) continue; //attempt to load plugin into $obj
443            $PARSER_MODES[$obj->getType()][] = "plugin_$p"; //register mode type
444            //add to modes
445            $modes[] = array(
446                    'sort' => $obj->getSort(),
447                    'mode' => "plugin_$p",
448                    'obj'  => $obj,
449                    );
450            unset($obj); //remove the reference
451        }
452    }
453
454    // add default modes
455    $std_modes = array('listblock','preformatted','notoc','nocache',
456            'header','table','linebreak','footnote','hr',
457            'unformatted','php','html','code','file','quote',
458            'internallink','rss','media','externallink',
459            'emaillink','windowssharelink','eol');
460    if($conf['typography']){
461        $std_modes[] = 'quotes';
462        $std_modes[] = 'multiplyentity';
463    }
464    foreach($std_modes as $m){
465        $class = "Doku_Parser_Mode_$m";
466        $obj   = new $class();
467        $modes[] = array(
468                'sort' => $obj->getSort(),
469                'mode' => $m,
470                'obj'  => $obj
471                );
472    }
473
474    // add formatting modes
475    $fmt_modes = array('strong','emphasis','underline','monospace',
476            'subscript','superscript','deleted');
477    foreach($fmt_modes as $m){
478        $obj   = new Doku_Parser_Mode_formatting($m);
479        $modes[] = array(
480                'sort' => $obj->getSort(),
481                'mode' => $m,
482                'obj'  => $obj
483                );
484    }
485
486    // add modes which need files
487    $obj     = new Doku_Parser_Mode_smiley(array_keys(getSmileys()));
488    $modes[] = array('sort' => $obj->getSort(), 'mode' => 'smiley','obj'  => $obj );
489    $obj     = new Doku_Parser_Mode_acronym(array_keys(getAcronyms()));
490    $modes[] = array('sort' => $obj->getSort(), 'mode' => 'acronym','obj'  => $obj );
491    $obj     = new Doku_Parser_Mode_entity(array_keys(getEntities()));
492    $modes[] = array('sort' => $obj->getSort(), 'mode' => 'entity','obj'  => $obj );
493
494    // add optional camelcase mode
495    if($conf['camelcase']){
496        $obj     = new Doku_Parser_Mode_camelcaselink();
497        $modes[] = array('sort' => $obj->getSort(), 'mode' => 'camelcaselink','obj'  => $obj );
498    }
499
500    //sort modes
501    usort($modes,'p_sort_modes');
502
503    return $modes;
504}
505
506/**
507 * Callback function for usort
508 *
509 * @author Andreas Gohr <andi@splitbrain.org>
510 */
511function p_sort_modes($a, $b){
512    if($a['sort'] == $b['sort']) return 0;
513    return ($a['sort'] < $b['sort']) ? -1 : 1;
514}
515
516/**
517 * Renders a list of instruction to the specified output mode
518 *
519 * In the $info array is information from the renderer returned
520 *
521 * @author Harry Fuecks <hfuecks@gmail.com>
522 * @author Andreas Gohr <andi@splitbrain.org>
523 */
524function p_render($mode,$instructions,&$info){
525    if(is_null($instructions)) return '';
526
527    $Renderer =& p_get_renderer($mode);
528    if (is_null($Renderer)) return null;
529
530    $Renderer->reset();
531
532    $Renderer->smileys = getSmileys();
533    $Renderer->entities = getEntities();
534    $Renderer->acronyms = getAcronyms();
535    $Renderer->interwiki = getInterwiki();
536
537    // Loop through the instructions
538    foreach ( $instructions as $instruction ) {
539        // Execute the callback against the Renderer
540        call_user_func_array(array(&$Renderer, $instruction[0]),$instruction[1]);
541    }
542
543    //set info array
544    $info = $Renderer->info;
545
546    // Post process and return the output
547    $data = array($mode,& $Renderer->doc);
548    trigger_event('RENDERER_CONTENT_POSTPROCESS',$data);
549    return $Renderer->doc;
550}
551
552function & p_get_renderer($mode) {
553    global $conf, $plugin_controller;
554
555    $rname = !empty($conf['renderer_'.$mode]) ? $conf['renderer_'.$mode] : $mode;
556
557    // try default renderer first:
558    $file = DOKU_INC."inc/parser/$rname.php";
559    if(@file_exists($file)){
560        require_once $file;
561        $rclass = "Doku_Renderer_$rname";
562
563        if ( !class_exists($rclass) ) {
564            trigger_error("Unable to resolve render class $rclass",E_USER_WARNING);
565            msg("Renderer '$rname' for $mode not valid",-1);
566            return null;
567        }
568        $Renderer = new $rclass();
569    }else{
570        // Maybe a plugin/component is available?
571        list($plugin, $component) = $plugin_controller->_splitName($rname);
572        if (!$plugin_controller->isdisabled($plugin)){
573            $Renderer =& $plugin_controller->load('renderer',$rname);
574        }
575
576        if(is_null($Renderer)){
577            msg("No renderer '$rname' found for mode '$mode'",-1);
578            return null;
579        }
580    }
581
582    return $Renderer;
583}
584
585/**
586 * Gets the first heading from a file
587 *
588 * @param   string   $id       dokuwiki page id
589 * @param   bool     $render   rerender if first heading not known
590 *                             default: true  -- must be set to false for calls from the metadata renderer to
591 *                                               protects against loops and excessive resource usage when pages
592 *                                               for which only a first heading is required will attempt to
593 *                                               render metadata for all the pages for which they require first
594 *                                               headings ... and so on.
595 *
596 * @author Andreas Gohr <andi@splitbrain.org>
597 */
598function p_get_first_heading($id, $render=true){
599    return p_get_metadata($id,'title',$render);
600}
601
602/**
603 * Wrapper for GeSHi Code Highlighter, provides caching of its output
604 *
605 * @param  string   $code       source code to be highlighted
606 * @param  string   $language   language to provide highlighting
607 * @param  string   $wrapper    html element to wrap the returned highlighted text
608 *
609 * @author Christopher Smith <chris@jalakai.co.uk>
610 * @author Andreas Gohr <andi@splitbrain.org>
611 */
612function p_xhtml_cached_geshi($code, $language, $wrapper='pre') {
613    global $conf, $config_cascade;
614    $language = strtolower($language);
615
616    // remove any leading or trailing blank lines
617    $code = preg_replace('/^\s*?\n|\s*?\n$/','',$code);
618
619    $cache = getCacheName($language.$code,".code");
620    $ctime = @filemtime($cache);
621    if($ctime && !$_REQUEST['purge'] &&
622            $ctime > filemtime(DOKU_INC.'inc/geshi.php') &&                 // geshi changed
623            $ctime > @filemtime(DOKU_INC.'inc/geshi/'.$language.'.php') &&  // language syntax definition changed
624            $ctime > filemtime(reset($config_cascade['main']['default']))){ // dokuwiki changed
625        $highlighted_code = io_readFile($cache, false);
626
627    } else {
628
629        $geshi = new GeSHi($code, $language, DOKU_INC . 'inc/geshi');
630        $geshi->set_encoding('utf-8');
631        $geshi->enable_classes();
632        $geshi->set_header_type(GESHI_HEADER_PRE);
633        $geshi->set_link_target($conf['target']['extern']);
634
635        // remove GeSHi's wrapper element (we'll replace it with our own later)
636        // we need to use a GeSHi wrapper to avoid <BR> throughout the highlighted text
637        $highlighted_code = trim(preg_replace('!^<pre[^>]*>|</pre>$!','',$geshi->parse_code()),"\n\r");
638        io_saveFile($cache,$highlighted_code);
639    }
640
641    // add a wrapper element if required
642    if ($wrapper) {
643        return "<$wrapper class=\"code $language\">$highlighted_code</$wrapper>";
644    } else {
645        return $highlighted_code;
646    }
647}
648
649