xref: /dokuwiki/inc/parserutils.php (revision 6353ef5ad0cd97399815c17e59758611d1bc1f79) 
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(preg_split('/\s+/', $key, 2, PREG_SPLIT_NO_EMPTY) as $cur_key) {
248        if (!isset($val[$cur_key])) {
249            return null;
250        }
251        $val = $val[$cur_key];
252    }
253    return $val;
254}
255
256/**
257 * sets metadata elements of a page
258 *
259 * @author Esther Brunner <esther@kaffeehaus.ch>
260 */
261function p_set_metadata($id, $data, $render=false, $persistent=true){
262    if (!is_array($data)) return false;
263
264    global $ID;
265
266    // cache the current page
267    $cache = ($ID == $id);
268    $orig = p_read_metadata($id, $cache);
269
270    // render metadata first?
271    $meta = $render ? p_render_metadata($id, $orig) : $orig;
272
273    // now add the passed metadata
274    $protected = array('description', 'date', 'contributor');
275    foreach ($data as $key => $value){
276
277        // be careful with sub-arrays of $meta['relation']
278        if ($key == 'relation'){
279
280            foreach ($value as $subkey => $subvalue){
281                $meta['current'][$key][$subkey] = !empty($meta['current'][$key][$subkey]) ? array_merge($meta['current'][$key][$subkey], $subvalue) : $subvalue;
282                if ($persistent)
283                    $meta['persistent'][$key][$subkey] = !empty($meta['persistent'][$key][$subkey]) ? array_merge($meta['persistent'][$key][$subkey], $subvalue) : $subvalue;
284            }
285
286            // be careful with some senisitive arrays of $meta
287        } elseif (in_array($key, $protected)){
288
289            // these keys, must have subkeys - a legitimate value must be an array
290            if (is_array($value)) {
291                $meta['current'][$key] = !empty($meta['current'][$key]) ? array_merge($meta['current'][$key],$value) : $value;
292
293                if ($persistent) {
294                    $meta['persistent'][$key] = !empty($meta['persistent'][$key]) ? array_merge($meta['persistent'][$key],$value) : $value;
295                }
296            }
297
298            // no special treatment for the rest
299        } else {
300            $meta['current'][$key] = $value;
301            if ($persistent) $meta['persistent'][$key] = $value;
302        }
303    }
304
305    // save only if metadata changed
306    if ($meta == $orig) return true;
307
308    // sync cached copies, including $INFO metadata
309    global $cache_metadata, $INFO;
310
311    if (!empty($cache_metadata[$id])) $cache_metadata[$id] = $meta;
312    if (!empty($INFO) && ($id == $INFO['id'])) { $INFO['meta'] = $meta['current']; }
313
314    return io_saveFile(metaFN($id, '.meta'), serialize($meta));
315}
316
317/**
318 * Purges the non-persistant part of the meta data
319 * used on page deletion
320 *
321 * @author Michael Klier <chi@chimeric.de>
322 */
323function p_purge_metadata($id) {
324    $metafn = metaFN('id', '.meta');
325    $meta   = p_read_metadata($id);
326    foreach($meta['current'] as $key => $value) {
327        if(is_array($meta[$key])) {
328            $meta['current'][$key] = array();
329        } else {
330            $meta['current'][$key] = '';
331        }
332    }
333    return io_saveFile(metaFN($id, '.meta'), serialize($meta));
334}
335
336/**
337 * read the metadata from source/cache for $id
338 * (internal use only - called by p_get_metadata & p_set_metadata)
339 *
340 * this function also converts the metadata from the original format to
341 * the current format ('current' & 'persistent' arrays)
342 *
343 * @author   Christopher Smith <chris@jalakai.co.uk>
344 *
345 * @param    string   $id      absolute wiki page id
346 * @param    bool     $cache   whether or not to cache metadata in memory
347 *                             (only use for metadata likely to be accessed several times)
348 *
349 * @return   array             metadata
350 */
351function p_read_metadata($id,$cache=false) {
352    global $cache_metadata;
353
354    if (isset($cache_metadata[(string)$id])) return $cache_metadata[(string)$id];
355
356    $file = metaFN($id, '.meta');
357    $meta = @file_exists($file) ? unserialize(io_readFile($file, false)) : array('current'=>array(),'persistent'=>array());
358
359    // convert $meta from old format to new (current+persistent) format
360    if (!isset($meta['current'])) {
361        $meta = array('current'=>$meta,'persistent'=>$meta);
362
363        // remove non-persistent keys
364        unset($meta['persistent']['title']);
365        unset($meta['persistent']['description']['abstract']);
366        unset($meta['persistent']['description']['tableofcontents']);
367        unset($meta['persistent']['relation']['haspart']);
368        unset($meta['persistent']['relation']['references']);
369        unset($meta['persistent']['date']['valid']);
370
371        if (empty($meta['persistent']['description'])) unset($meta['persistent']['description']);
372        if (empty($meta['persistent']['relation'])) unset($meta['persistent']['relation']);
373        if (empty($meta['persistent']['date'])) unset($meta['persistent']['date']);
374
375        // save converted metadata
376        io_saveFile($file, serialize($meta));
377    }
378
379    if ($cache) {
380        $cache_metadata[(string)$id] = $meta;
381    }
382
383    return $meta;
384}
385
386/**
387 * renders the metadata of a page
388 *
389 * @author Esther Brunner <esther@kaffeehaus.ch>
390 */
391function p_render_metadata($id, $orig){
392    // make sure the correct ID is in global ID
393    global $ID;
394    $keep = $ID;
395    $ID   = $id;
396
397    // add an extra key for the event - to tell event handlers the page whose metadata this is
398    $orig['page'] = $id;
399    $evt = new Doku_Event('PARSER_METADATA_RENDER', $orig);
400    if ($evt->advise_before()) {
401
402        require_once DOKU_INC."inc/parser/metadata.php";
403
404        // get instructions
405        $instructions = p_cached_instructions(wikiFN($id),false,$id);
406        if(is_null($instructions)){
407            $ID = $keep;
408            return null; // something went wrong with the instructions
409        }
410
411        // set up the renderer
412        $renderer = new Doku_Renderer_metadata();
413        $renderer->meta = $orig['current'];
414        $renderer->persistent = $orig['persistent'];
415
416        // loop through the instructions
417        foreach ($instructions as $instruction){
418            // execute the callback against the renderer
419            call_user_func_array(array(&$renderer, $instruction[0]), (array) $instruction[1]);
420        }
421
422        $evt->result = array('current'=>$renderer->meta,'persistent'=>$renderer->persistent);
423    }
424    $evt->advise_after();
425
426    $ID = $keep;
427    return $evt->result;
428}
429
430/**
431 * returns all available parser syntax modes in correct order
432 *
433 * @author Andreas Gohr <andi@splitbrain.org>
434 */
435function p_get_parsermodes(){
436    global $conf;
437
438    //reuse old data
439    static $modes = null;
440    if($modes != null){
441        return $modes;
442    }
443
444    //import parser classes and mode definitions
445    require_once DOKU_INC . 'inc/parser/parser.php';
446
447    // we now collect all syntax modes and their objects, then they will
448    // be sorted and added to the parser in correct order
449    $modes = array();
450
451    // add syntax plugins
452    $pluginlist = plugin_list('syntax');
453    if(count($pluginlist)){
454        global $PARSER_MODES;
455        $obj = null;
456        foreach($pluginlist as $p){
457            if(!$obj =& plugin_load('syntax',$p)) continue; //attempt to load plugin into $obj
458            $PARSER_MODES[$obj->getType()][] = "plugin_$p"; //register mode type
459            //add to modes
460            $modes[] = array(
461                    'sort' => $obj->getSort(),
462                    'mode' => "plugin_$p",
463                    'obj'  => $obj,
464                    );
465            unset($obj); //remove the reference
466        }
467    }
468
469    // add default modes
470    $std_modes = array('listblock','preformatted','notoc','nocache',
471            'header','table','linebreak','footnote','hr',
472            'unformatted','php','html','code','file','quote',
473            'internallink','rss','media','externallink',
474            'emaillink','windowssharelink','eol');
475    if($conf['typography']){
476        $std_modes[] = 'quotes';
477        $std_modes[] = 'multiplyentity';
478    }
479    foreach($std_modes as $m){
480        $class = "Doku_Parser_Mode_$m";
481        $obj   = new $class();
482        $modes[] = array(
483                'sort' => $obj->getSort(),
484                'mode' => $m,
485                'obj'  => $obj
486                );
487    }
488
489    // add formatting modes
490    $fmt_modes = array('strong','emphasis','underline','monospace',
491            'subscript','superscript','deleted');
492    foreach($fmt_modes as $m){
493        $obj   = new Doku_Parser_Mode_formatting($m);
494        $modes[] = array(
495                'sort' => $obj->getSort(),
496                'mode' => $m,
497                'obj'  => $obj
498                );
499    }
500
501    // add modes which need files
502    $obj     = new Doku_Parser_Mode_smiley(array_keys(getSmileys()));
503    $modes[] = array('sort' => $obj->getSort(), 'mode' => 'smiley','obj'  => $obj );
504    $obj     = new Doku_Parser_Mode_acronym(array_keys(getAcronyms()));
505    $modes[] = array('sort' => $obj->getSort(), 'mode' => 'acronym','obj'  => $obj );
506    $obj     = new Doku_Parser_Mode_entity(array_keys(getEntities()));
507    $modes[] = array('sort' => $obj->getSort(), 'mode' => 'entity','obj'  => $obj );
508
509    // add optional camelcase mode
510    if($conf['camelcase']){
511        $obj     = new Doku_Parser_Mode_camelcaselink();
512        $modes[] = array('sort' => $obj->getSort(), 'mode' => 'camelcaselink','obj'  => $obj );
513    }
514
515    //sort modes
516    usort($modes,'p_sort_modes');
517
518    return $modes;
519}
520
521/**
522 * Callback function for usort
523 *
524 * @author Andreas Gohr <andi@splitbrain.org>
525 */
526function p_sort_modes($a, $b){
527    if($a['sort'] == $b['sort']) return 0;
528    return ($a['sort'] < $b['sort']) ? -1 : 1;
529}
530
531/**
532 * Renders a list of instruction to the specified output mode
533 *
534 * In the $info array is information from the renderer returned
535 *
536 * @author Harry Fuecks <hfuecks@gmail.com>
537 * @author Andreas Gohr <andi@splitbrain.org>
538 */
539function p_render($mode,$instructions,&$info){
540    if(is_null($instructions)) return '';
541
542    $Renderer =& p_get_renderer($mode);
543    if (is_null($Renderer)) return null;
544
545    $Renderer->reset();
546
547    $Renderer->smileys = getSmileys();
548    $Renderer->entities = getEntities();
549    $Renderer->acronyms = getAcronyms();
550    $Renderer->interwiki = getInterwiki();
551
552    // Loop through the instructions
553    foreach ( $instructions as $instruction ) {
554        // Execute the callback against the Renderer
555        call_user_func_array(array(&$Renderer, $instruction[0]),$instruction[1]);
556    }
557
558    //set info array
559    $info = $Renderer->info;
560
561    // Post process and return the output
562    $data = array($mode,& $Renderer->doc);
563    trigger_event('RENDERER_CONTENT_POSTPROCESS',$data);
564    return $Renderer->doc;
565}
566
567function & p_get_renderer($mode) {
568    global $conf, $plugin_controller;
569
570    $rname = !empty($conf['renderer_'.$mode]) ? $conf['renderer_'.$mode] : $mode;
571
572    // try default renderer first:
573    $file = DOKU_INC."inc/parser/$rname.php";
574    if(@file_exists($file)){
575        require_once $file;
576        $rclass = "Doku_Renderer_$rname";
577
578        if ( !class_exists($rclass) ) {
579            trigger_error("Unable to resolve render class $rclass",E_USER_WARNING);
580            msg("Renderer '$rname' for $mode not valid",-1);
581            return null;
582        }
583        $Renderer = new $rclass();
584    }else{
585        // Maybe a plugin/component is available?
586        list($plugin, $component) = $plugin_controller->_splitName($rname);
587        if (!$plugin_controller->isdisabled($plugin)){
588            $Renderer =& $plugin_controller->load('renderer',$rname);
589        }
590
591        if(is_null($Renderer)){
592            msg("No renderer '$rname' found for mode '$mode'",-1);
593            return null;
594        }
595    }
596
597    return $Renderer;
598}
599
600/**
601 * Gets the first heading from a file
602 *
603 * @param   string   $id       dokuwiki page id
604 * @param   bool     $render   rerender if first heading not known
605 *                             default: true  -- must be set to false for calls from the metadata renderer to
606 *                                               protects against loops and excessive resource usage when pages
607 *                                               for which only a first heading is required will attempt to
608 *                                               render metadata for all the pages for which they require first
609 *                                               headings ... and so on.
610 *
611 * @author Andreas Gohr <andi@splitbrain.org>
612 */
613function p_get_first_heading($id, $render=true){
614    return p_get_metadata($id,'title',$render);
615}
616
617/**
618 * Wrapper for GeSHi Code Highlighter, provides caching of its output
619 *
620 * @param  string   $code       source code to be highlighted
621 * @param  string   $language   language to provide highlighting
622 * @param  string   $wrapper    html element to wrap the returned highlighted text
623 *
624 * @author Christopher Smith <chris@jalakai.co.uk>
625 * @author Andreas Gohr <andi@splitbrain.org>
626 */
627function p_xhtml_cached_geshi($code, $language, $wrapper='pre') {
628    global $conf, $config_cascade;
629    $language = strtolower($language);
630
631    // remove any leading or trailing blank lines
632    $code = preg_replace('/^\s*?\n|\s*?\n$/','',$code);
633
634    $cache = getCacheName($language.$code,".code");
635    $ctime = @filemtime($cache);
636    if($ctime && !$_REQUEST['purge'] &&
637            $ctime > filemtime(DOKU_INC.'inc/geshi.php') &&                 // geshi changed
638            $ctime > @filemtime(DOKU_INC.'inc/geshi/'.$language.'.php') &&  // language syntax definition changed
639            $ctime > filemtime(reset($config_cascade['main']['default']))){ // dokuwiki changed
640        $highlighted_code = io_readFile($cache, false);
641
642    } else {
643
644        $geshi = new GeSHi($code, $language, DOKU_INC . 'inc/geshi');
645        $geshi->set_encoding('utf-8');
646        $geshi->enable_classes();
647        $geshi->set_header_type(GESHI_HEADER_PRE);
648        $geshi->set_link_target($conf['target']['extern']);
649
650        // remove GeSHi's wrapper element (we'll replace it with our own later)
651        // we need to use a GeSHi wrapper to avoid <BR> throughout the highlighted text
652        $highlighted_code = trim(preg_replace('!^<pre[^>]*>|</pre>$!','',$geshi->parse_code()),"\n\r");
653        io_saveFile($cache,$highlighted_code);
654    }
655
656    // add a wrapper element if required
657    if ($wrapper) {
658        return "<$wrapper class=\"code $language\">$highlighted_code</$wrapper>";
659    } else {
660        return $highlighted_code;
661    }
662}
663
664