xref: /dokuwiki/inc/parserutils.php (revision 178c8be8577764b678fc145787fb5a5fe5bd5bf7) 
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 $p;
218}
219
220/**
221 * returns the metadata of a page
222 *
223 * @author Esther Brunner <esther@kaffeehaus.ch>
224 * @author Michael Hamann <michael@content-space.de>
225 */
226function p_get_metadata($id, $key='', $render=false){
227    global $ID;
228
229    // cache the current page
230    // Benchmarking shows the current page's metadata is generally the only page metadata
231    // accessed several times. This may catch a few other pages, but that shouldn't be an issue.
232    $cache = ($ID == $id);
233    $meta = p_read_metadata($id, $cache);
234
235    // prevent recursive calls in the cache
236    static $recursion = false;
237    if (!$recursion){
238        $recursion = true;
239
240        $cachefile = new cache_renderer($id, wikiFN($id), 'metadata');
241
242        if (page_exists($id) && !$cachefile->useCache()){
243            $old_meta = $meta;
244            $meta = p_render_metadata($id, $meta);
245            // only update the file when the metadata has been changed
246            if ($meta == $old_meta || p_save_metadata($id, $meta)) {
247                // store a timestamp in order to make sure that the cachefile is touched
248                $cachefile->storeCache(time());
249            } else {
250                msg('Unable to save metadata file. Hint: disk full; file permissions; safe_mode setting.',-1);
251            }
252        }
253
254        $recursion = false;
255    }
256
257    $val = $meta['current'];
258
259    // filter by $key
260    foreach(preg_split('/\s+/', $key, 2, PREG_SPLIT_NO_EMPTY) as $cur_key) {
261        if (!isset($val[$cur_key])) {
262            return null;
263        }
264        $val = $val[$cur_key];
265    }
266    return $val;
267}
268
269/**
270 * sets metadata elements of a page
271 *
272 * @see http://www.dokuwiki.org/devel:metadata#functions_to_get_and_set_metadata
273 *
274 * @param String  $id         is the ID of a wiki page
275 * @param Array   $data       is an array with key ⇒ value pairs to be set in the metadata
276 * @param Boolean $render     whether or not the page metadata should be generated with the renderer
277 * @param Boolean $persistent indicates whether or not the particular metadata value will persist through
278 *                            the next metadata rendering.
279 * @return boolean true on success
280 *
281 * @author Esther Brunner <esther@kaffeehaus.ch>
282 */
283function p_set_metadata($id, $data, $render=false, $persistent=true){
284    if (!is_array($data)) return false;
285
286    global $ID;
287
288    // cache the current page
289    $cache = ($ID == $id);
290    $orig = p_read_metadata($id, $cache);
291
292    // render metadata first?
293    $meta = $render ? p_render_metadata($id, $orig) : $orig;
294
295    // now add the passed metadata
296    $protected = array('description', 'date', 'contributor');
297    foreach ($data as $key => $value){
298
299        // be careful with sub-arrays of $meta['relation']
300        if ($key == 'relation'){
301
302            foreach ($value as $subkey => $subvalue){
303                $meta['current'][$key][$subkey] = !empty($meta['current'][$key][$subkey]) ? array_merge($meta['current'][$key][$subkey], $subvalue) : $subvalue;
304                if ($persistent)
305                    $meta['persistent'][$key][$subkey] = !empty($meta['persistent'][$key][$subkey]) ? array_merge($meta['persistent'][$key][$subkey], $subvalue) : $subvalue;
306            }
307
308            // be careful with some senisitive arrays of $meta
309        } elseif (in_array($key, $protected)){
310
311            // these keys, must have subkeys - a legitimate value must be an array
312            if (is_array($value)) {
313                $meta['current'][$key] = !empty($meta['current'][$key]) ? array_merge($meta['current'][$key],$value) : $value;
314
315                if ($persistent) {
316                    $meta['persistent'][$key] = !empty($meta['persistent'][$key]) ? array_merge($meta['persistent'][$key],$value) : $value;
317                }
318            }
319
320            // no special treatment for the rest
321        } else {
322            $meta['current'][$key] = $value;
323            if ($persistent) $meta['persistent'][$key] = $value;
324        }
325    }
326
327    // save only if metadata changed
328    if ($meta == $orig) return true;
329
330    return p_save_metadata($id, $meta);
331}
332
333/**
334 * Purges the non-persistant part of the meta data
335 * used on page deletion
336 *
337 * @author Michael Klier <chi@chimeric.de>
338 */
339function p_purge_metadata($id) {
340    $meta = p_read_metadata($id);
341    foreach($meta['current'] as $key => $value) {
342        if(is_array($meta[$key])) {
343            $meta['current'][$key] = array();
344        } else {
345            $meta['current'][$key] = '';
346        }
347
348    }
349    return p_save_metadata($id, $meta);
350}
351
352/**
353 * read the metadata from source/cache for $id
354 * (internal use only - called by p_get_metadata & p_set_metadata)
355 *
356 * @author   Christopher Smith <chris@jalakai.co.uk>
357 *
358 * @param    string   $id      absolute wiki page id
359 * @param    bool     $cache   whether or not to cache metadata in memory
360 *                             (only use for metadata likely to be accessed several times)
361 *
362 * @return   array             metadata
363 */
364function p_read_metadata($id,$cache=false) {
365    global $cache_metadata;
366
367    if (isset($cache_metadata[(string)$id])) return $cache_metadata[(string)$id];
368
369    $file = metaFN($id, '.meta');
370    $meta = @file_exists($file) ? unserialize(io_readFile($file, false)) : array('current'=>array(),'persistent'=>array());
371
372    if ($cache) {
373        $cache_metadata[(string)$id] = $meta;
374    }
375
376    return $meta;
377}
378
379/**
380 * This is the backend function to save a metadata array to a file
381 *
382 * @param    string   $id      absolute wiki page id
383 * @param    array    $meta    metadata
384 *
385 * @return   bool              success / fail
386 */
387function p_save_metadata($id, $meta) {
388    // sync cached copies, including $INFO metadata
389    global $cache_metadata, $INFO;
390
391    if (isset($cache_metadata[$id])) $cache_metadata[$id] = $meta;
392    if (!empty($INFO) && ($id == $INFO['id'])) { $INFO['meta'] = $meta['current']; }
393
394    return io_saveFile(metaFN($id, '.meta'), serialize($meta));
395}
396
397/**
398 * renders the metadata of a page
399 *
400 * @author Esther Brunner <esther@kaffeehaus.ch>
401 */
402function p_render_metadata($id, $orig){
403    // make sure the correct ID is in global ID
404    global $ID;
405    $keep = $ID;
406    $ID   = $id;
407
408    // add an extra key for the event - to tell event handlers the page whose metadata this is
409    $orig['page'] = $id;
410    $evt = new Doku_Event('PARSER_METADATA_RENDER', $orig);
411    if ($evt->advise_before()) {
412
413        require_once DOKU_INC."inc/parser/metadata.php";
414
415        // get instructions
416        $instructions = p_cached_instructions(wikiFN($id),false,$id);
417        if(is_null($instructions)){
418            $ID = $keep;
419            return null; // something went wrong with the instructions
420        }
421
422        // set up the renderer
423        $renderer = new Doku_Renderer_metadata();
424        $renderer->meta = $orig['current'];
425        $renderer->persistent = $orig['persistent'];
426
427        // loop through the instructions
428        foreach ($instructions as $instruction){
429            // execute the callback against the renderer
430            call_user_func_array(array(&$renderer, $instruction[0]), (array) $instruction[1]);
431        }
432
433        $evt->result = array('current'=>$renderer->meta,'persistent'=>$renderer->persistent);
434    }
435    $evt->advise_after();
436
437    $ID = $keep;
438    return $evt->result;
439}
440
441/**
442 * returns all available parser syntax modes in correct order
443 *
444 * @author Andreas Gohr <andi@splitbrain.org>
445 */
446function p_get_parsermodes(){
447    global $conf;
448
449    //reuse old data
450    static $modes = null;
451    if($modes != null){
452        return $modes;
453    }
454
455    //import parser classes and mode definitions
456    require_once DOKU_INC . 'inc/parser/parser.php';
457
458    // we now collect all syntax modes and their objects, then they will
459    // be sorted and added to the parser in correct order
460    $modes = array();
461
462    // add syntax plugins
463    $pluginlist = plugin_list('syntax');
464    if(count($pluginlist)){
465        global $PARSER_MODES;
466        $obj = null;
467        foreach($pluginlist as $p){
468            if(!$obj =& plugin_load('syntax',$p)) continue; //attempt to load plugin into $obj
469            $PARSER_MODES[$obj->getType()][] = "plugin_$p"; //register mode type
470            //add to modes
471            $modes[] = array(
472                    'sort' => $obj->getSort(),
473                    'mode' => "plugin_$p",
474                    'obj'  => $obj,
475                    );
476            unset($obj); //remove the reference
477        }
478    }
479
480    // add default modes
481    $std_modes = array('listblock','preformatted','notoc','nocache',
482            'header','table','linebreak','footnote','hr',
483            'unformatted','php','html','code','file','quote',
484            'internallink','rss','media','externallink',
485            'emaillink','windowssharelink','eol');
486    if($conf['typography']){
487        $std_modes[] = 'quotes';
488        $std_modes[] = 'multiplyentity';
489    }
490    foreach($std_modes as $m){
491        $class = "Doku_Parser_Mode_$m";
492        $obj   = new $class();
493        $modes[] = array(
494                'sort' => $obj->getSort(),
495                'mode' => $m,
496                'obj'  => $obj
497                );
498    }
499
500    // add formatting modes
501    $fmt_modes = array('strong','emphasis','underline','monospace',
502            'subscript','superscript','deleted');
503    foreach($fmt_modes as $m){
504        $obj   = new Doku_Parser_Mode_formatting($m);
505        $modes[] = array(
506                'sort' => $obj->getSort(),
507                'mode' => $m,
508                'obj'  => $obj
509                );
510    }
511
512    // add modes which need files
513    $obj     = new Doku_Parser_Mode_smiley(array_keys(getSmileys()));
514    $modes[] = array('sort' => $obj->getSort(), 'mode' => 'smiley','obj'  => $obj );
515    $obj     = new Doku_Parser_Mode_acronym(array_keys(getAcronyms()));
516    $modes[] = array('sort' => $obj->getSort(), 'mode' => 'acronym','obj'  => $obj );
517    $obj     = new Doku_Parser_Mode_entity(array_keys(getEntities()));
518    $modes[] = array('sort' => $obj->getSort(), 'mode' => 'entity','obj'  => $obj );
519
520    // add optional camelcase mode
521    if($conf['camelcase']){
522        $obj     = new Doku_Parser_Mode_camelcaselink();
523        $modes[] = array('sort' => $obj->getSort(), 'mode' => 'camelcaselink','obj'  => $obj );
524    }
525
526    //sort modes
527    usort($modes,'p_sort_modes');
528
529    return $modes;
530}
531
532/**
533 * Callback function for usort
534 *
535 * @author Andreas Gohr <andi@splitbrain.org>
536 */
537function p_sort_modes($a, $b){
538    if($a['sort'] == $b['sort']) return 0;
539    return ($a['sort'] < $b['sort']) ? -1 : 1;
540}
541
542/**
543 * Renders a list of instruction to the specified output mode
544 *
545 * In the $info array is information from the renderer returned
546 *
547 * @author Harry Fuecks <hfuecks@gmail.com>
548 * @author Andreas Gohr <andi@splitbrain.org>
549 */
550function p_render($mode,$instructions,&$info){
551    if(is_null($instructions)) return '';
552
553    $Renderer =& p_get_renderer($mode);
554    if (is_null($Renderer)) return null;
555
556    $Renderer->reset();
557
558    $Renderer->smileys = getSmileys();
559    $Renderer->entities = getEntities();
560    $Renderer->acronyms = getAcronyms();
561    $Renderer->interwiki = getInterwiki();
562
563    // Loop through the instructions
564    foreach ( $instructions as $instruction ) {
565        // Execute the callback against the Renderer
566        call_user_func_array(array(&$Renderer, $instruction[0]),$instruction[1]);
567    }
568
569    //set info array
570    $info = $Renderer->info;
571
572    // Post process and return the output
573    $data = array($mode,& $Renderer->doc);
574    trigger_event('RENDERER_CONTENT_POSTPROCESS',$data);
575    return $Renderer->doc;
576}
577
578function & p_get_renderer($mode) {
579    global $conf, $plugin_controller;
580
581    $rname = !empty($conf['renderer_'.$mode]) ? $conf['renderer_'.$mode] : $mode;
582
583    // try default renderer first:
584    $file = DOKU_INC."inc/parser/$rname.php";
585    if(@file_exists($file)){
586        require_once $file;
587        $rclass = "Doku_Renderer_$rname";
588
589        if ( !class_exists($rclass) ) {
590            trigger_error("Unable to resolve render class $rclass",E_USER_WARNING);
591            msg("Renderer '$rname' for $mode not valid",-1);
592            return null;
593        }
594        $Renderer = new $rclass();
595    }else{
596        // Maybe a plugin/component is available?
597        list($plugin, $component) = $plugin_controller->_splitName($rname);
598        if (!$plugin_controller->isdisabled($plugin)){
599            $Renderer =& $plugin_controller->load('renderer',$rname);
600        }
601
602        if(is_null($Renderer)){
603            msg("No renderer '$rname' found for mode '$mode'",-1);
604            return null;
605        }
606    }
607
608    return $Renderer;
609}
610
611/**
612 * Gets the first heading from a file
613 *
614 * @param   string   $id       dokuwiki page id
615 * @param   bool     $render   rerender if first heading not known
616 *                             default: true  -- must be set to false for calls from the metadata renderer to
617 *                                               protects against loops and excessive resource usage when pages
618 *                                               for which only a first heading is required will attempt to
619 *                                               render metadata for all the pages for which they require first
620 *                                               headings ... and so on.
621 *
622 * @author Andreas Gohr <andi@splitbrain.org>
623 */
624function p_get_first_heading($id, $render=true){
625    return p_get_metadata($id,'title',$render);
626}
627
628/**
629 * Wrapper for GeSHi Code Highlighter, provides caching of its output
630 *
631 * @param  string   $code       source code to be highlighted
632 * @param  string   $language   language to provide highlighting
633 * @param  string   $wrapper    html element to wrap the returned highlighted text
634 *
635 * @author Christopher Smith <chris@jalakai.co.uk>
636 * @author Andreas Gohr <andi@splitbrain.org>
637 */
638function p_xhtml_cached_geshi($code, $language, $wrapper='pre') {
639    global $conf, $config_cascade;
640    $language = strtolower($language);
641
642    // remove any leading or trailing blank lines
643    $code = preg_replace('/^\s*?\n|\s*?\n$/','',$code);
644
645    $cache = getCacheName($language.$code,".code");
646    $ctime = @filemtime($cache);
647    if($ctime && !$_REQUEST['purge'] &&
648            $ctime > filemtime(DOKU_INC.'inc/geshi.php') &&                 // geshi changed
649            $ctime > @filemtime(DOKU_INC.'inc/geshi/'.$language.'.php') &&  // language syntax definition changed
650            $ctime > filemtime(reset($config_cascade['main']['default']))){ // dokuwiki changed
651        $highlighted_code = io_readFile($cache, false);
652
653    } else {
654
655        $geshi = new GeSHi($code, $language, DOKU_INC . 'inc/geshi');
656        $geshi->set_encoding('utf-8');
657        $geshi->enable_classes();
658        $geshi->set_header_type(GESHI_HEADER_PRE);
659        $geshi->set_link_target($conf['target']['extern']);
660
661        // remove GeSHi's wrapper element (we'll replace it with our own later)
662        // we need to use a GeSHi wrapper to avoid <BR> throughout the highlighted text
663        $highlighted_code = trim(preg_replace('!^<pre[^>]*>|</pre>$!','',$geshi->parse_code()),"\n\r");
664        io_saveFile($cache,$highlighted_code);
665    }
666
667    // add a wrapper element if required
668    if ($wrapper) {
669        return "<$wrapper class=\"code $language\">$highlighted_code</$wrapper>";
670    } else {
671        return $highlighted_code;
672    }
673}
674
675