xref: /dokuwiki/inc/parserutils.php (revision 26ee1edf9c1750df399baebcb70afb49778b4a92) 
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, $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    $val = $meta['current'];
245
246    // filter by $key
247    foreach(explode(' ', trim($key), 2) as $cur_key) {
248        $cur_key = trim($cur_key);
249        if (!isset($val[$cur_key])) {
250            return null;
251        }
252        $val = $val[$cur_key];
253    }
254    return $val;
255}
256
257/**
258 * sets metadata elements of a page
259 *
260 * @author Esther Brunner <esther@kaffeehaus.ch>
261 */
262function p_set_metadata($id, $data, $render=false, $persistent=true){
263    if (!is_array($data)) return false;
264
265    global $ID;
266
267    // cache the current page
268    $cache = ($ID == $id);
269    $orig = p_read_metadata($id, $cache);
270
271    // render metadata first?
272    $meta = $render ? p_render_metadata($id, $orig) : $orig;
273
274    // now add the passed metadata
275    $protected = array('description', 'date', 'contributor');
276    foreach ($data as $key => $value){
277
278        // be careful with sub-arrays of $meta['relation']
279        if ($key == 'relation'){
280
281            foreach ($value as $subkey => $subvalue){
282                $meta['current'][$key][$subkey] = !empty($meta['current'][$key][$subkey]) ? array_merge($meta['current'][$key][$subkey], $subvalue) : $subvalue;
283                if ($persistent)
284                    $meta['persistent'][$key][$subkey] = !empty($meta['persistent'][$key][$subkey]) ? array_merge($meta['persistent'][$key][$subkey], $subvalue) : $subvalue;
285            }
286
287            // be careful with some senisitive arrays of $meta
288        } elseif (in_array($key, $protected)){
289
290            // these keys, must have subkeys - a legitimate value must be an array
291            if (is_array($value)) {
292                $meta['current'][$key] = !empty($meta['current'][$key]) ? array_merge($meta['current'][$key],$value) : $value;
293
294                if ($persistent) {
295                    $meta['persistent'][$key] = !empty($meta['persistent'][$key]) ? array_merge($meta['persistent'][$key],$value) : $value;
296                }
297            }
298
299            // no special treatment for the rest
300        } else {
301            $meta['current'][$key] = $value;
302            if ($persistent) $meta['persistent'][$key] = $value;
303        }
304    }
305
306    // save only if metadata changed
307    if ($meta == $orig) return true;
308
309    // sync cached copies, including $INFO metadata
310    global $cache_metadata, $INFO;
311
312    if (!empty($cache_metadata[$id])) $cache_metadata[$id] = $meta;
313    if (!empty($INFO) && ($id == $INFO['id'])) { $INFO['meta'] = $meta['current']; }
314
315    return io_saveFile(metaFN($id, '.meta'), serialize($meta));
316}
317
318/**
319 * Purges the non-persistant part of the meta data
320 * used on page deletion
321 *
322 * @author Michael Klier <chi@chimeric.de>
323 */
324function p_purge_metadata($id) {
325    $metafn = metaFN('id', '.meta');
326    $meta   = p_read_metadata($id);
327    foreach($meta['current'] as $key => $value) {
328        if(is_array($meta[$key])) {
329            $meta['current'][$key] = array();
330        } else {
331            $meta['current'][$key] = '';
332        }
333    }
334    return io_saveFile(metaFN($id, '.meta'), serialize($meta));
335}
336
337/**
338 * read the metadata from source/cache for $id
339 * (internal use only - called by p_get_metadata & p_set_metadata)
340 *
341 * this function also converts the metadata from the original format to
342 * the current format ('current' & 'persistent' arrays)
343 *
344 * @author   Christopher Smith <chris@jalakai.co.uk>
345 *
346 * @param    string   $id      absolute wiki page id
347 * @param    bool     $cache   whether or not to cache metadata in memory
348 *                             (only use for metadata likely to be accessed several times)
349 *
350 * @return   array             metadata
351 */
352function p_read_metadata($id,$cache=false) {
353    global $cache_metadata;
354
355    if (isset($cache_metadata[(string)$id])) return $cache_metadata[(string)$id];
356
357    $file = metaFN($id, '.meta');
358    $meta = @file_exists($file) ? unserialize(io_readFile($file, false)) : array('current'=>array(),'persistent'=>array());
359
360    // convert $meta from old format to new (current+persistent) format
361    if (!isset($meta['current'])) {
362        $meta = array('current'=>$meta,'persistent'=>$meta);
363
364        // remove non-persistent keys
365        unset($meta['persistent']['title']);
366        unset($meta['persistent']['description']['abstract']);
367        unset($meta['persistent']['description']['tableofcontents']);
368        unset($meta['persistent']['relation']['haspart']);
369        unset($meta['persistent']['relation']['references']);
370        unset($meta['persistent']['date']['valid']);
371
372        if (empty($meta['persistent']['description'])) unset($meta['persistent']['description']);
373        if (empty($meta['persistent']['relation'])) unset($meta['persistent']['relation']);
374        if (empty($meta['persistent']['date'])) unset($meta['persistent']['date']);
375
376        // save converted metadata
377        io_saveFile($file, serialize($meta));
378    }
379
380    if ($cache) {
381        $cache_metadata[(string)$id] = $meta;
382    }
383
384    return $meta;
385}
386
387/**
388 * renders the metadata of a page
389 *
390 * @author Esther Brunner <esther@kaffeehaus.ch>
391 */
392function p_render_metadata($id, $orig){
393    // make sure the correct ID is in global ID
394    global $ID;
395    $keep = $ID;
396    $ID   = $id;
397
398    // add an extra key for the event - to tell event handlers the page whose metadata this is
399    $orig['page'] = $id;
400    $evt = new Doku_Event('PARSER_METADATA_RENDER', $orig);
401    if ($evt->advise_before()) {
402
403        require_once DOKU_INC."inc/parser/metadata.php";
404
405        // get instructions
406        $instructions = p_cached_instructions(wikiFN($id),false,$id);
407        if(is_null($instructions)){
408            $ID = $keep;
409            return null; // something went wrong with the instructions
410        }
411
412        // set up the renderer
413        $renderer = new Doku_Renderer_metadata();
414        $renderer->meta = $orig['current'];
415        $renderer->persistent = $orig['persistent'];
416
417        // loop through the instructions
418        foreach ($instructions as $instruction){
419            // execute the callback against the renderer
420            call_user_func_array(array(&$renderer, $instruction[0]), (array) $instruction[1]);
421        }
422
423        $evt->result = array('current'=>$renderer->meta,'persistent'=>$renderer->persistent);
424    }
425    $evt->advise_after();
426
427    $ID = $keep;
428    return $evt->result;
429}
430
431/**
432 * returns all available parser syntax modes in correct order
433 *
434 * @author Andreas Gohr <andi@splitbrain.org>
435 */
436function p_get_parsermodes(){
437    global $conf;
438
439    //reuse old data
440    static $modes = null;
441    if($modes != null){
442        return $modes;
443    }
444
445    //import parser classes and mode definitions
446    require_once DOKU_INC . 'inc/parser/parser.php';
447
448    // we now collect all syntax modes and their objects, then they will
449    // be sorted and added to the parser in correct order
450    $modes = array();
451
452    // add syntax plugins
453    $pluginlist = plugin_list('syntax');
454    if(count($pluginlist)){
455        global $PARSER_MODES;
456        $obj = null;
457        foreach($pluginlist as $p){
458            if(!$obj =& plugin_load('syntax',$p)) continue; //attempt to load plugin into $obj
459            $PARSER_MODES[$obj->getType()][] = "plugin_$p"; //register mode type
460            //add to modes
461            $modes[] = array(
462                    'sort' => $obj->getSort(),
463                    'mode' => "plugin_$p",
464                    'obj'  => $obj,
465                    );
466            unset($obj); //remove the reference
467        }
468    }
469
470    // add default modes
471    $std_modes = array('listblock','preformatted','notoc','nocache',
472            'header','table','linebreak','footnote','hr',
473            'unformatted','php','html','code','file','quote',
474            'internallink','rss','media','externallink',
475            'emaillink','windowssharelink','eol');
476    if($conf['typography']){
477        $std_modes[] = 'quotes';
478        $std_modes[] = 'multiplyentity';
479    }
480    foreach($std_modes as $m){
481        $class = "Doku_Parser_Mode_$m";
482        $obj   = new $class();
483        $modes[] = array(
484                'sort' => $obj->getSort(),
485                'mode' => $m,
486                'obj'  => $obj
487                );
488    }
489
490    // add formatting modes
491    $fmt_modes = array('strong','emphasis','underline','monospace',
492            'subscript','superscript','deleted');
493    foreach($fmt_modes as $m){
494        $obj   = new Doku_Parser_Mode_formatting($m);
495        $modes[] = array(
496                'sort' => $obj->getSort(),
497                'mode' => $m,
498                'obj'  => $obj
499                );
500    }
501
502    // add modes which need files
503    $obj     = new Doku_Parser_Mode_smiley(array_keys(getSmileys()));
504    $modes[] = array('sort' => $obj->getSort(), 'mode' => 'smiley','obj'  => $obj );
505    $obj     = new Doku_Parser_Mode_acronym(array_keys(getAcronyms()));
506    $modes[] = array('sort' => $obj->getSort(), 'mode' => 'acronym','obj'  => $obj );
507    $obj     = new Doku_Parser_Mode_entity(array_keys(getEntities()));
508    $modes[] = array('sort' => $obj->getSort(), 'mode' => 'entity','obj'  => $obj );
509
510    // add optional camelcase mode
511    if($conf['camelcase']){
512        $obj     = new Doku_Parser_Mode_camelcaselink();
513        $modes[] = array('sort' => $obj->getSort(), 'mode' => 'camelcaselink','obj'  => $obj );
514    }
515
516    //sort modes
517    usort($modes,'p_sort_modes');
518
519    return $modes;
520}
521
522/**
523 * Callback function for usort
524 *
525 * @author Andreas Gohr <andi@splitbrain.org>
526 */
527function p_sort_modes($a, $b){
528    if($a['sort'] == $b['sort']) return 0;
529    return ($a['sort'] < $b['sort']) ? -1 : 1;
530}
531
532/**
533 * Renders a list of instruction to the specified output mode
534 *
535 * In the $info array are informations from the renderer returned
536 *
537 * @author Harry Fuecks <hfuecks@gmail.com>
538 * @author Andreas Gohr <andi@splitbrain.org>
539 */
540function p_render($mode,$instructions,&$info){
541    if(is_null($instructions)) return '';
542
543    $Renderer =& p_get_renderer($mode);
544    if (is_null($Renderer)) return null;
545
546    $Renderer->reset();
547
548    $Renderer->smileys = getSmileys();
549    $Renderer->entities = getEntities();
550    $Renderer->acronyms = getAcronyms();
551    $Renderer->interwiki = getInterwiki();
552
553    // Loop through the instructions
554    foreach ( $instructions as $instruction ) {
555        // Execute the callback against the Renderer
556        call_user_func_array(array(&$Renderer, $instruction[0]),$instruction[1]);
557    }
558
559    //set info array
560    $info = $Renderer->info;
561
562    // Post process and return the output
563    $data = array($mode,& $Renderer->doc);
564    trigger_event('RENDERER_CONTENT_POSTPROCESS',$data);
565    return $Renderer->doc;
566}
567
568function & p_get_renderer($mode) {
569    global $conf, $plugin_controller;
570
571    $rname = !empty($conf['renderer_'.$mode]) ? $conf['renderer_'.$mode] : $mode;
572
573    // try default renderer first:
574    $file = DOKU_INC."inc/parser/$rname.php";
575    if(@file_exists($file)){
576        require_once $file;
577        $rclass = "Doku_Renderer_$rname";
578
579        if ( !class_exists($rclass) ) {
580            trigger_error("Unable to resolve render class $rclass",E_USER_WARNING);
581            msg("Renderer '$rname' for $mode not valid",-1);
582            return null;
583        }
584        $Renderer = new $rclass();
585    }else{
586        // Maybe a plugin/component is available?
587        list($plugin, $component) = $plugin_controller->_splitName($rname);
588        if (!$plugin_controller->isdisabled($plugin)){
589            $Renderer =& $plugin_controller->load('renderer',$rname, true);
590        }
591
592        if(is_null($Renderer)){
593            msg("No renderer '$rname' found for mode '$mode'",-1);
594            return null;
595        }
596    }
597
598    return $Renderer;
599}
600
601/**
602 * Gets the first heading from a file
603 *
604 * @param   string   $id       dokuwiki page id
605 * @param   bool     $render   rerender if first heading not known
606 *                             default: true  -- must be set to false for calls from the metadata renderer to
607 *                                               protects against loops and excessive resource usage when pages
608 *                                               for which only a first heading is required will attempt to
609 *                                               render metadata for all the pages for which they require first
610 *                                               headings ... and so on.
611 *
612 * @author Andreas Gohr <andi@splitbrain.org>
613 */
614function p_get_first_heading($id, $render=true){
615    return p_get_metadata($id,'title',$render);
616}
617
618/**
619 * Wrapper for GeSHi Code Highlighter, provides caching of its output
620 *
621 * @param  string   $code       source code to be highlighted
622 * @param  string   $language   language to provide highlighting
623 * @param  string   $wrapper    html element to wrap the returned highlighted text
624 *
625 * @author Christopher Smith <chris@jalakai.co.uk>
626 * @author Andreas Gohr <andi@splitbrain.org>
627 */
628function p_xhtml_cached_geshi($code, $language, $wrapper='pre') {
629    global $conf, $config_cascade;
630    $language = strtolower($language);
631
632    // remove any leading or trailing blank lines
633    $code = preg_replace('/^\s*?\n|\s*?\n$/','',$code);
634
635    $cache = getCacheName($language.$code,".code");
636    $ctime = @filemtime($cache);
637    if($ctime && !$_REQUEST['purge'] &&
638            $ctime > filemtime(DOKU_INC.'inc/geshi.php') &&                 // geshi changed
639            $ctime > @filemtime(DOKU_INC.'inc/geshi/'.$language.'.php') &&  // language syntax definition changed
640            $ctime > filemtime(reset($config_cascade['main']['default']))){ // dokuwiki changed
641        $highlighted_code = io_readFile($cache, false);
642
643    } else {
644
645        $geshi = new GeSHi($code, $language, DOKU_INC . 'inc/geshi');
646        $geshi->set_encoding('utf-8');
647        $geshi->enable_classes();
648        $geshi->set_header_type(GESHI_HEADER_PRE);
649        $geshi->set_link_target($conf['target']['extern']);
650
651        // remove GeSHi's wrapper element (we'll replace it with our own later)
652        // we need to use a GeSHi wrapper to avoid <BR> throughout the highlighted text
653        $highlighted_code = trim(preg_replace('!^<pre[^>]*>|</pre>$!','',$geshi->parse_code()),"\n\r");
654        io_saveFile($cache,$highlighted_code);
655    }
656
657    // add a wrapper element if required
658    if ($wrapper) {
659        return "<$wrapper class=\"code $language\">$highlighted_code</$wrapper>";
660    } else {
661        return $highlighted_code;
662    }
663}
664
665