xref: /dokuwiki/lib/exe/indexer.php (revision 6cd2b1e4b7588c56b52418c91396c264538d5a54) 
1<?php
2/**
3 * DokuWiki indexer
4 *
5 * @license    GPL 2 (http://www.gnu.org/licenses/gpl.html)
6 * @author     Andreas Gohr <andi@splitbrain.org>
7 */
8if(!defined('DOKU_INC')) define('DOKU_INC',realpath(dirname(__FILE__).'/../../').'/');
9define('DOKU_DISABLE_GZIP_OUTPUT',1);
10require_once(DOKU_INC.'inc/init.php');
11require_once(DOKU_INC.'inc/auth.php');
12require_once(DOKU_INC.'inc/events.php');
13session_write_close();  //close session
14if(!defined('NL')) define('NL',"\n");
15
16// keep running after browser closes connection
17@ignore_user_abort(true);
18
19// check if user abort worked, if yes send output early
20if(@ignore_user_abort()){
21    sendGIF(); // send gif
22    $defer = false;
23}else{
24    $defer = true;
25}
26
27// Catch any possible output (e.g. errors)
28if(!$_REQUEST['debug']) ob_start();
29
30// run one of the jobs
31runIndexer() or metaUpdate() or runSitemapper() or runTrimRecentChanges();
32if($defer) sendGIF();
33
34if(!$_REQUEST['debug']) ob_end_clean();
35exit;
36
37// --------------------------------------------------------------------
38
39/**
40 * Trims the recent changes cache (or imports the old changelog) as needed.
41 *
42 * @author Ben Coburn <btcoburn@silicodon.net>
43 */
44function runTrimRecentChanges() {
45    global $conf;
46
47    // Import old changelog (if needed)
48    // Uses the imporoldchangelog plugin to upgrade the changelog automaticaly.
49    // FIXME: Remove this from runTrimRecentChanges when it is no longer needed.
50    if (isset($conf['changelog_old']) &&
51        @file_exists($conf['changelog_old']) && !@file_exists($conf['changelog']) &&
52        !@file_exists($conf['changelog'].'_importing') && !@file_exists($conf['changelog'].'_tmp')) {
53            $tmp = array(); // no event data
54            trigger_event('TEMPORARY_CHANGELOG_UPGRADE_EVENT', $tmp);
55            return true;
56    }
57
58    // Trim the Recent Changes
59    // Trims the recent changes cache to the last $conf['changes_days'] recent
60    // changes or $conf['recent'] items, which ever is larger.
61    // The trimming is only done once a day.
62    if (@file_exists($conf['changelog']) &&
63        (filectime($conf['changelog'])+86400)<time() &&
64        !@file_exists($conf['changelog'].'_tmp')) {
65            io_lock($conf['changelog']);
66            $lines = file($conf['changelog']);
67            if (count($lines)<$conf['recent']) {
68                // nothing to trim
69                io_unlock($conf['changelog']);
70                return true;
71            }
72
73            io_saveFile($conf['changelog'].'_tmp', '');          // presave tmp as 2nd lock
74            $trim_time = time() - $conf['recent_days']*86400;
75            $out_lines = array();
76
77            for ($i=0; $i<count($lines); $i++) {
78              $log = parseChangelogLine($lines[$i]);
79              if ($log === false) continue;                      // discard junk
80              if ($log['date'] < $trim_time) {
81                $old_lines[$log['date'].".$i"] = $lines[$i];     // keep old lines for now (append .$i to prevent key collisions)
82              } else {
83                $out_lines[$log['date'].".$i"] = $lines[$i];     // definitely keep these lines
84              }
85            }
86
87            // sort the final result, it shouldn't be necessary,
88            //   however the extra robustness in making the changelog cache self-correcting is worth it
89            ksort($out_lines);
90            $extra = $conf['recent'] - count($out_lines);        // do we need extra lines do bring us up to minimum
91            if ($extra > 0) {
92              ksort($old_lines);
93              $out_lines = array_merge(array_slice($old_lines,-$extra),$out_lines);
94            }
95
96            // save trimmed changelog
97            io_saveFile($conf['changelog'].'_tmp', implode('', $out_lines));
98            @unlink($conf['changelog']);
99            if (!rename($conf['changelog'].'_tmp', $conf['changelog'])) {
100                // rename failed so try another way...
101                io_unlock($conf['changelog']);
102                io_saveFile($conf['changelog'], implode('', $out_lines));
103                @unlink($conf['changelog'].'_tmp');
104            } else {
105                io_unlock($conf['changelog']);
106            }
107            return true;
108    }
109
110    // nothing done
111    return false;
112}
113
114/**
115 * Runs the indexer for the current page
116 *
117 * @author Andreas Gohr <andi@splitbrain.org>
118 */
119function runIndexer(){
120    global $conf;
121    print "runIndexer(): started".NL;
122
123    $ID = cleanID($_REQUEST['id']);
124    if(!$ID) return false;
125
126    // check if indexing needed
127    $last = @filemtime(metaFN($ID,'.indexed'));
128    if($last > @filemtime(wikiFN($ID))){
129        print "runIndexer(): index for $ID up to date".NL;
130        return false;
131    }
132
133    // try to aquire a lock
134    $lock = $conf['lockdir'].'/_indexer.lock';
135    while(!@mkdir($lock,$conf['dmode'])){
136        usleep(50);
137        if(time()-@filemtime($lock) > 60*5){
138            // looks like a stale lock - remove it
139            @rmdir($lock);
140            print "runIndexer(): stale lock removed".NL;
141        }else{
142            print "runIndexer(): indexer locked".NL;
143            return false;
144        }
145    }
146    if($conf['dperm']) chmod($lock, $conf['dperm']);
147
148    require_once(DOKU_INC.'inc/indexer.php');
149
150    // do the work
151    idx_addPage($ID);
152
153    // we're finished - save and free lock
154    io_saveFile(metaFN($ID,'.indexed'),' ');
155    @rmdir($lock);
156    print "runIndexer(): finished".NL;
157    return true;
158}
159
160/**
161 * Will render the metadata for the page if not exists yet
162 *
163 * This makes sure pages which are created from outside DokuWiki will
164 * gain their data when viewed for the first time.
165 */
166function metaUpdate(){
167    print "metaUpdate(): started".NL;
168
169    $ID = cleanID($_REQUEST['id']);
170    if(!$ID) return false;
171    $file = metaFN($ID, '.meta');
172    echo "meta file: $file".NL;
173
174    // rendering needed?
175    if (@file_exists($file)) return false;
176    if (!@file_exists(wikiFN($ID))) return false;
177
178    require_once(DOKU_INC.'inc/common.php');
179    require_once(DOKU_INC.'inc/parserutils.php');
180    global $conf;
181
182
183    // gather some additional info from changelog
184    $info = io_grep($conf['changelog'],
185                    '/^(\d+)\t(\d+\.\d+\.\d+\.\d+)\t'.preg_quote($ID,'/').'\t([^\t]+)\t([^\t\n]+)/',
186                    0,true);
187
188    $meta = array();
189    if(count($info)){
190        $meta['date']['created'] = $info[0][1];
191        foreach($info as $item){
192            if($item[4] != '*'){
193                $meta['date']['modified'] = $item[1];
194                if($item[3]){
195                    $meta['contributor'][$item[3]] = $item[3];
196                }
197            }
198        }
199    }
200
201    $meta = p_render_metadata($ID, $meta);
202    io_saveFile($file, serialize($meta));
203
204    echo "metaUpdate(): finished".NL;
205    return true;
206}
207
208/**
209 * Builds a Google Sitemap of all public pages known to the indexer
210 *
211 * The map is placed in the root directory named sitemap.xml.gz - This
212 * file needs to be writable!
213 *
214 * @author Andreas Gohr
215 * @link   https://www.google.com/webmasters/sitemaps/docs/en/about.html
216 */
217function runSitemapper(){
218    global $conf;
219    print "runSitemapper(): started".NL;
220    if(!$conf['sitemap']) return false;
221
222    if($conf['compression'] == 'bz2' || $conf['compression'] == 'gz'){
223        $sitemap = 'sitemap.xml.gz';
224    }else{
225        $sitemap = 'sitemap.xml';
226    }
227    print "runSitemapper(): using $sitemap".NL;
228
229    if(!is_writable(DOKU_INC.$sitemap)) return false;
230    if(@filesize(DOKU_INC.$sitemap) &&
231       @filemtime(DOKU_INC.$sitemap) > (time()-($conf['sitemap']*60*60*24))){
232       print 'runSitemapper(): Sitemap up to date'.NL;
233       return false;
234    }
235
236    $pages = file($conf['cachedir'].'/page.idx');
237    print 'runSitemapper(): creating sitemap using '.count($pages).' pages'.NL;
238
239    // build the sitemap
240    ob_start();
241    print '<?xml version="1.0" encoding="UTF-8"?>'.NL;
242    print '<urlset xmlns="http://www.google.com/schemas/sitemap/0.84">'.NL;
243    foreach($pages as $id){
244        $id = trim($id);
245        $file = wikiFN($id);
246
247        //skip hidden, non existing and restricted files
248        if(isHiddenPage($id)) continue;
249        $date = @filemtime($file);
250        if(!$date) continue;
251        if(auth_aclcheck($id,'','') < AUTH_READ) continue;
252
253        print '  <url>'.NL;
254        print '    <loc>'.wl($id,'',true).'</loc>'.NL;
255        print '    <lastmod>'.date_iso8601($date).'</lastmod>'.NL;
256        print '  </url>'.NL;
257    }
258    print '</urlset>'.NL;
259    $data = ob_get_contents();
260    ob_end_clean();
261
262    //save the new sitemap
263    io_saveFile(DOKU_INC.$sitemap,$data);
264
265    print 'runSitemapper(): pinging google'.NL;
266    //ping google
267    $url  = 'http://www.google.com/webmasters/sitemaps/ping?sitemap=';
268    $url .= urlencode(DOKU_URL.$sitemap);
269    $http = new DokuHTTPClient();
270    $http->get($url);
271    if($http->error) print 'runSitemapper(): '.$http->error.NL;
272
273    print 'runSitemapper(): finished'.NL;
274    return true;
275}
276
277/**
278 * Formats a timestamp as ISO 8601 date
279 *
280 * @author <ungu at terong dot com>
281 * @link http://www.php.net/manual/en/function.date.php#54072
282 */
283function date_iso8601($int_date) {
284   //$int_date: current date in UNIX timestamp
285   $date_mod = date('Y-m-d\TH:i:s', $int_date);
286   $pre_timezone = date('O', $int_date);
287   $time_zone = substr($pre_timezone, 0, 3).":".substr($pre_timezone, 3, 2);
288   $date_mod .= $time_zone;
289   return $date_mod;
290}
291
292/**
293 * Just send a 1x1 pixel blank gif to the browser
294 *
295 * @author Andreas Gohr <andi@splitbrain.org>
296 * @author Harry Fuecks <fuecks@gmail.com>
297 */
298function sendGIF(){
299    if($_REQUEST['debug']){
300        header('Content-Type: text/plain');
301        return;
302    }
303    $img = base64_decode('R0lGODlhAQABAIAAAAAAAP///yH5BAEAAAEALAAAAAABAAEAAAIBTAA7');
304    header('Content-Type: image/gif');
305    header('Content-Length: '.strlen($img));
306    header('Connection: Close');
307    print $img;
308    flush();
309    // Browser should drop connection after this
310    // Thinks it's got the whole image
311}
312
313//Setup VIM: ex: et ts=4 enc=utf-8 :
314// No trailing PHP closing tag - no output please!
315// See Note at http://www.php.net/manual/en/language.basic-syntax.instruction-separation.php
316