xref: /dokuwiki/inc/parserutils.php (revision b080c3ddd9671fcd5a67865fe970ae0f8b5bdeed) 
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=false, $render=false){
226    global $ID, $INFO, $cache_metadata;
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, $meta);
237        io_saveFile(metaFN($id, '.meta'), serialize($meta));
238
239        // sync cached copies, including $INFO metadata
240        if (!empty($cache_metadata[$id])) $cache_metadata[$id] = $meta;
241        if (!empty($INFO) && ($id == $INFO['id'])) { $INFO['meta'] = $meta['current']; }
242    }
243
244    // filter by $key
245    if ($key){
246        list($key, $subkey) = explode(' ', $key, 2);
247        $subkey = trim($subkey);
248
249        if ($subkey) {
250            return isset($meta['current'][$key][$subkey]) ? $meta['current'][$key][$subkey] : null;
251        }
252
253        return isset($meta['current'][$key]) ? $meta['current'][$key] : null;
254    }
255
256    return $meta['current'];
257}
258
259/**
260 * sets metadata elements of a page
261 *
262 * @author Esther Brunner <esther@kaffeehaus.ch>
263 */
264function p_set_metadata($id, $data, $render=false, $persistent=true){
265    if (!is_array($data)) return false;
266
267    global $ID;
268
269    // cache the current page
270    $cache = ($ID == $id);
271    $orig = p_read_metadata($id, $cache);
272
273    // render metadata first?
274    $meta = $render ? p_render_metadata($id, $orig) : $orig;
275
276    // now add the passed metadata
277    $protected = array('description', 'date', 'contributor');
278    foreach ($data as $key => $value){
279
280        // be careful with sub-arrays of $meta['relation']
281        if ($key == 'relation'){
282
283            foreach ($value as $subkey => $subvalue){
284                $meta['current'][$key][$subkey] = !empty($meta['current'][$key][$subkey]) ? array_merge($meta['current'][$key][$subkey], $subvalue) : $subvalue;
285                if ($persistent)
286                    $meta['persistent'][$key][$subkey] = !empty($meta['persistent'][$key][$subkey]) ? array_merge($meta['persistent'][$key][$subkey], $subvalue) : $subvalue;
287            }
288
289            // be careful with some senisitive arrays of $meta
290        } elseif (in_array($key, $protected)){
291
292            // these keys, must have subkeys - a legitimate value must be an array
293            if (is_array($value)) {
294                $meta['current'][$key] = !empty($meta['current'][$key]) ? array_merge($meta['current'][$key],$value) : $value;
295
296                if ($persistent) {
297                    $meta['persistent'][$key] = !empty($meta['persistent'][$key]) ? array_merge($meta['persistent'][$key],$value) : $value;
298                }
299            }
300
301            // no special treatment for the rest
302        } else {
303            $meta['current'][$key] = $value;
304            if ($persistent) $meta['persistent'][$key] = $value;
305        }
306    }
307
308    // save only if metadata changed
309    if ($meta == $orig) return true;
310
311    // sync cached copies, including $INFO metadata
312    global $cache_metadata, $INFO;
313
314    if (!empty($cache_metadata[$id])) $cache_metadata[$id] = $meta;
315    if (!empty($INFO) && ($id == $INFO['id'])) { $INFO['meta'] = $meta['current']; }
316
317    return io_saveFile(metaFN($id, '.meta'), serialize($meta));
318}
319
320/**
321 * Purges the non-persistant part of the meta data
322 * used on page deletion
323 *
324 * @author Michael Klier <chi@chimeric.de>
325 */
326function p_purge_metadata($id) {
327    $metafn = metaFN('id', '.meta');
328    $meta   = p_read_metadata($id);
329    foreach($meta['current'] as $key => $value) {
330        if(is_array($meta[$key])) {
331            $meta['current'][$key] = array();
332        } else {
333            $meta['current'][$key] = '';
334        }
335    }
336    return io_saveFile(metaFN($id, '.meta'), serialize($meta));
337}
338
339/**
340 * read the metadata from source/cache for $id
341 * (internal use only - called by p_get_metadata & p_set_metadata)
342 *
343 * this function also converts the metadata from the original format to
344 * the current format ('current' & 'persistent' arrays)
345 *
346 * @author   Christopher Smith <chris@jalakai.co.uk>
347 *
348 * @param    string   $id      absolute wiki page id
349 * @param    bool     $cache   whether or not to cache metadata in memory
350 *                             (only use for metadata likely to be accessed several times)
351 *
352 * @return   array             metadata
353 */
354function p_read_metadata($id,$cache=false) {
355    global $cache_metadata;
356
357    if (isset($cache_metadata[(string)$id])) return $cache_metadata[(string)$id];
358
359    $file = metaFN($id, '.meta');
360    $meta = @file_exists($file) ? unserialize(io_readFile($file, false)) : array('current'=>array(),'persistent'=>array());
361
362    // convert $meta from old format to new (current+persistent) format
363    if (!isset($meta['current'])) {
364        $meta = array('current'=>$meta,'persistent'=>$meta);
365
366        // remove non-persistent keys
367        unset($meta['persistent']['title']);
368        unset($meta['persistent']['description']['abstract']);
369        unset($meta['persistent']['description']['tableofcontents']);
370        unset($meta['persistent']['relation']['haspart']);
371        unset($meta['persistent']['relation']['references']);
372        unset($meta['persistent']['date']['valid']);
373
374        if (empty($meta['persistent']['description'])) unset($meta['persistent']['description']);
375        if (empty($meta['persistent']['relation'])) unset($meta['persistent']['relation']);
376        if (empty($meta['persistent']['date'])) unset($meta['persistent']['date']);
377
378        // save converted metadata
379        io_saveFile($file, serialize($meta));
380    }
381
382    if ($cache) {
383        $cache_metadata[(string)$id] = $meta;
384    }
385
386    return $meta;
387}
388
389/**
390 * renders the metadata of a page
391 *
392 * @author Esther Brunner <esther@kaffeehaus.ch>
393 */
394function p_render_metadata($id, $orig){
395    // make sure the correct ID is in global ID
396    global $ID;
397    $keep = $ID;
398    $ID   = $id;
399
400    // add an extra key for the event - to tell event handlers the page whose metadata this is
401    $orig['page'] = $id;
402    $evt = new Doku_Event('PARSER_METADATA_RENDER', $orig);
403    if ($evt->advise_before()) {
404
405        require_once DOKU_INC."inc/parser/metadata.php";
406
407        // get instructions
408        $instructions = p_cached_instructions(wikiFN($id),false,$id);
409        if(is_null($instructions)){
410            $ID = $keep;
411            return null; // something went wrong with the instructions
412        }
413
414        // set up the renderer
415        $renderer = new Doku_Renderer_metadata();
416        $renderer->meta = $orig['current'];
417        $renderer->persistent = $orig['persistent'];
418
419        // loop through the instructions
420        foreach ($instructions as $instruction){
421            // execute the callback against the renderer
422            call_user_func_array(array(&$renderer, $instruction[0]), (array) $instruction[1]);
423        }
424
425        $evt->result = array('current'=>$renderer->meta,'persistent'=>$renderer->persistent);
426    }
427    $evt->advise_after();
428
429    $ID = $keep;
430    return $evt->result;
431}
432
433/**
434 * returns all available parser syntax modes in correct order
435 *
436 * @author Andreas Gohr <andi@splitbrain.org>
437 */
438function p_get_parsermodes(){
439    global $conf;
440
441    //reuse old data
442    static $modes = null;
443    if($modes != null){
444        return $modes;
445    }
446
447    //import parser classes and mode definitions
448    require_once DOKU_INC . 'inc/parser/parser.php';
449
450    // we now collect all syntax modes and their objects, then they will
451    // be sorted and added to the parser in correct order
452    $modes = array();
453
454    // add syntax plugins
455    $pluginlist = plugin_list('syntax');
456    if(count($pluginlist)){
457        global $PARSER_MODES;
458        $obj = null;
459        foreach($pluginlist as $p){
460            if(!$obj =& plugin_load('syntax',$p)) continue; //attempt to load plugin into $obj
461            $PARSER_MODES[$obj->getType()][] = "plugin_$p"; //register mode type
462            //add to modes
463            $modes[] = array(
464                    'sort' => $obj->getSort(),
465                    'mode' => "plugin_$p",
466                    'obj'  => $obj,
467                    );
468            unset($obj); //remove the reference
469        }
470    }
471
472    // add default modes
473    $std_modes = array('listblock','preformatted','notoc','nocache',
474            'header','table','linebreak','footnote','hr',
475            'unformatted','php','html','code','file','quote',
476            'internallink','rss','media','externallink',
477            'emaillink','windowssharelink','eol');
478    if($conf['typography']){
479        $std_modes[] = 'quotes';
480        $std_modes[] = 'multiplyentity';
481    }
482    foreach($std_modes as $m){
483        $class = "Doku_Parser_Mode_$m";
484        $obj   = new $class();
485        $modes[] = array(
486                'sort' => $obj->getSort(),
487                'mode' => $m,
488                'obj'  => $obj
489                );
490    }
491
492    // add formatting modes
493    $fmt_modes = array('strong','emphasis','underline','monospace',
494            'subscript','superscript','deleted');
495    foreach($fmt_modes as $m){
496        $obj   = new Doku_Parser_Mode_formatting($m);
497        $modes[] = array(
498                'sort' => $obj->getSort(),
499                'mode' => $m,
500                'obj'  => $obj
501                );
502    }
503
504    // add modes which need files
505    $obj     = new Doku_Parser_Mode_smiley(array_keys(getSmileys()));
506    $modes[] = array('sort' => $obj->getSort(), 'mode' => 'smiley','obj'  => $obj );
507    $obj     = new Doku_Parser_Mode_acronym(array_keys(getAcronyms()));
508    $modes[] = array('sort' => $obj->getSort(), 'mode' => 'acronym','obj'  => $obj );
509    $obj     = new Doku_Parser_Mode_entity(array_keys(getEntities()));
510    $modes[] = array('sort' => $obj->getSort(), 'mode' => 'entity','obj'  => $obj );
511
512    // add optional camelcase mode
513    if($conf['camelcase']){
514        $obj     = new Doku_Parser_Mode_camelcaselink();
515        $modes[] = array('sort' => $obj->getSort(), 'mode' => 'camelcaselink','obj'  => $obj );
516    }
517
518    //sort modes
519    usort($modes,'p_sort_modes');
520
521    return $modes;
522}
523
524/**
525 * Callback function for usort
526 *
527 * @author Andreas Gohr <andi@splitbrain.org>
528 */
529function p_sort_modes($a, $b){
530    if($a['sort'] == $b['sort']) return 0;
531    return ($a['sort'] < $b['sort']) ? -1 : 1;
532}
533
534/**
535 * Renders a list of instruction to the specified output mode
536 *
537 * In the $info array are informations from the renderer returned
538 *
539 * @author Harry Fuecks <hfuecks@gmail.com>
540 * @author Andreas Gohr <andi@splitbrain.org>
541 */
542function p_render($mode,$instructions,&$info){
543    if(is_null($instructions)) return '';
544
545    $Renderer =& p_get_renderer($mode);
546    if (is_null($Renderer)) return null;
547
548    $Renderer->reset();
549
550    $Renderer->smileys = getSmileys();
551    $Renderer->entities = getEntities();
552    $Renderer->acronyms = getAcronyms();
553    $Renderer->interwiki = getInterwiki();
554
555    // Loop through the instructions
556    foreach ( $instructions as $instruction ) {
557        // Execute the callback against the Renderer
558        call_user_func_array(array(&$Renderer, $instruction[0]),$instruction[1]);
559    }
560
561    //set info array
562    $info = $Renderer->info;
563
564    // Post process and return the output
565    $data = array($mode,& $Renderer->doc);
566    trigger_event('RENDERER_CONTENT_POSTPROCESS',$data);
567    return $Renderer->doc;
568}
569
570function & p_get_renderer($mode) {
571    global $conf, $plugin_controller;
572
573    $rname = !empty($conf['renderer_'.$mode]) ? $conf['renderer_'.$mode] : $mode;
574
575    // try default renderer first:
576    $file = DOKU_INC."inc/parser/$rname.php";
577    if(@file_exists($file)){
578        require_once $file;
579        $rclass = "Doku_Renderer_$rname";
580
581        if ( !class_exists($rclass) ) {
582            trigger_error("Unable to resolve render class $rclass",E_USER_WARNING);
583            msg("Renderer '$rname' for $mode not valid",-1);
584            return null;
585        }
586        $Renderer = new $rclass();
587    }else{
588        // Maybe a plugin/component is available?
589        list($plugin, $component) = $plugin_controller->_splitName($rname);
590        if (!$plugin_controller->isdisabled($plugin)){
591            $Renderer =& $plugin_controller->load('renderer',$rname, true);
592        }
593
594        if(is_null($Renderer)){
595            msg("No renderer '$rname' found for mode '$mode'",-1);
596            return null;
597        }
598    }
599
600    return $Renderer;
601}
602
603/**
604 * Gets the first heading from a file
605 *
606 * @param   string   $id       dokuwiki page id
607 * @param   bool     $render   rerender if first heading not known
608 *                             default: true  -- must be set to false for calls from the metadata renderer to
609 *                                               protects against loops and excessive resource usage when pages
610 *                                               for which only a first heading is required will attempt to
611 *                                               render metadata for all the pages for which they require first
612 *                                               headings ... and so on.
613 *
614 * @author Andreas Gohr <andi@splitbrain.org>
615 */
616function p_get_first_heading($id, $render=true){
617    return p_get_metadata($id,'title',$render);
618}
619
620/**
621 * Wrapper for GeSHi Code Highlighter, provides caching of its output
622 *
623 * @param  string   $code       source code to be highlighted
624 * @param  string   $language   language to provide highlighting
625 * @param  string   $wrapper    html element to wrap the returned highlighted text
626 *
627 * @author Christopher Smith <chris@jalakai.co.uk>
628 * @author Andreas Gohr <andi@splitbrain.org>
629 */
630function p_xhtml_cached_geshi($code, $language, $wrapper='pre') {
631    global $conf, $config_cascade;
632    $language = strtolower($language);
633
634    // remove any leading or trailing blank lines
635    $code = preg_replace('/^\s*?\n|\s*?\n$/','',$code);
636
637    $cache = getCacheName($language.$code,".code");
638    $ctime = @filemtime($cache);
639    if($ctime && !$_REQUEST['purge'] &&
640            $ctime > filemtime(DOKU_INC.'inc/geshi.php') &&                 // geshi changed
641            $ctime > @filemtime(DOKU_INC.'inc/geshi/'.$language.'.php') &&  // language syntax definition changed
642            $ctime > filemtime(reset($config_cascade['main']['default']))){ // dokuwiki changed
643        $highlighted_code = io_readFile($cache, false);
644
645    } else {
646
647        $geshi = new GeSHi($code, $language, DOKU_INC . 'inc/geshi');
648        $geshi->set_encoding('utf-8');
649        $geshi->enable_classes();
650        $geshi->set_header_type(GESHI_HEADER_PRE);
651        $geshi->set_link_target($conf['target']['extern']);
652
653        // remove GeSHi's wrapper element (we'll replace it with our own later)
654        // we need to use a GeSHi wrapper to avoid <BR> throughout the highlighted text
655        $highlighted_code = trim(preg_replace('!^<pre[^>]*>|</pre>$!','',$geshi->parse_code()),"\n\r");
656        io_saveFile($cache,$highlighted_code);
657    }
658
659    // add a wrapper element if required
660    if ($wrapper) {
661        return "<$wrapper class=\"code $language\">$highlighted_code</$wrapper>";
662    } else {
663        return $highlighted_code;
664    }
665}
666
667