xref: /dokuwiki/inc/io.php (revision 64159a61e94d0ce680071c8890e144982c3a8cbe) 
1<?php
2/**
3 * File IO functions
4 *
5 * @license    GPL 2 (http://www.gnu.org/licenses/gpl.html)
6 * @author     Andreas Gohr <andi@splitbrain.org>
7 */
8
9/**
10 * Removes empty directories
11 *
12 * Sends IO_NAMESPACE_DELETED events for 'pages' and 'media' namespaces.
13 * Event data:
14 * $data[0]    ns: The colon separated namespace path minus the trailing page name.
15 * $data[1]    ns_type: 'pages' or 'media' namespace tree.
16 *
17 * @todo use safemode hack
18 * @param string $id      - a pageid, the namespace of that id will be tried to deleted
19 * @param string $basedir - the config name of the type to delete (datadir or mediadir usally)
20 * @return bool - true if at least one namespace was deleted
21 *
22 * @author  Andreas Gohr <andi@splitbrain.org>
23 * @author Ben Coburn <btcoburn@silicodon.net>
24 */
25function io_sweepNS($id,$basedir='datadir'){
26    global $conf;
27    $types = array ('datadir'=>'pages', 'mediadir'=>'media');
28    $ns_type = (isset($types[$basedir])?$types[$basedir]:false);
29
30    $delone = false;
31
32    //scan all namespaces
33    while(($id = getNS($id)) !== false){
34        $dir = $conf[$basedir].'/'.utf8_encodeFN(str_replace(':','/',$id));
35
36        //try to delete dir else return
37        if(@rmdir($dir)) {
38            if ($ns_type!==false) {
39                $data = array($id, $ns_type);
40                $delone = true; // we deleted at least one dir
41                trigger_event('IO_NAMESPACE_DELETED', $data);
42            }
43        } else { return $delone; }
44    }
45    return $delone;
46}
47
48/**
49 * Used to read in a DokuWiki page from file, and send IO_WIKIPAGE_READ events.
50 *
51 * Generates the action event which delegates to io_readFile().
52 * Action plugins are allowed to modify the page content in transit.
53 * The file path should not be changed.
54 *
55 * Event data:
56 * $data[0]    The raw arguments for io_readFile as an array.
57 * $data[1]    ns: The colon separated namespace path minus the trailing page name. (false if root ns)
58 * $data[2]    page_name: The wiki page name.
59 * $data[3]    rev: The page revision, false for current wiki pages.
60 *
61 * @author Ben Coburn <btcoburn@silicodon.net>
62 *
63 * @param string   $file filename
64 * @param string   $id page id
65 * @param bool|int $rev revision timestamp
66 * @return string
67 */
68function io_readWikiPage($file, $id, $rev=false) {
69    if (empty($rev)) { $rev = false; }
70    $data = array(array($file, true), getNS($id), noNS($id), $rev);
71    return trigger_event('IO_WIKIPAGE_READ', $data, '_io_readWikiPage_action', false);
72}
73
74/**
75 * Callback adapter for io_readFile().
76 *
77 * @author Ben Coburn <btcoburn@silicodon.net>
78 *
79 * @param array $data event data
80 * @return string
81 */
82function _io_readWikiPage_action($data) {
83    if (is_array($data) && is_array($data[0]) && count($data[0])===2) {
84        return call_user_func_array('io_readFile', $data[0]);
85    } else {
86        return ''; //callback error
87    }
88}
89
90/**
91 * Returns content of $file as cleaned string.
92 *
93 * Uses gzip if extension is .gz
94 *
95 * If you want to use the returned value in unserialize
96 * be sure to set $clean to false!
97 *
98 * @author  Andreas Gohr <andi@splitbrain.org>
99 *
100 * @param string $file  filename
101 * @param bool   $clean
102 * @return string|bool the file contents or false on error
103 */
104function io_readFile($file,$clean=true){
105    $ret = '';
106    if(file_exists($file)){
107        if(substr($file,-3) == '.gz'){
108            if(!DOKU_HAS_GZIP) return false;
109            $ret = gzfile($file);
110            if(is_array($ret)) $ret = join('', $ret);
111        }else if(substr($file,-4) == '.bz2'){
112            if(!DOKU_HAS_BZIP) return false;
113            $ret = bzfile($file);
114        }else{
115            $ret = file_get_contents($file);
116        }
117    }
118    if($ret === null) return false;
119    if($ret !== false && $clean){
120        return cleanText($ret);
121    }else{
122        return $ret;
123    }
124}
125/**
126 * Returns the content of a .bz2 compressed file as string
127 *
128 * @author marcel senf <marcel@rucksackreinigung.de>
129 * @author  Andreas Gohr <andi@splitbrain.org>
130 *
131 * @param string $file filename
132 * @param bool   $array return array of lines
133 * @return string|array|bool content or false on error
134 */
135function bzfile($file, $array=false) {
136    $bz = bzopen($file,"r");
137    if($bz === false) return false;
138
139    if($array) $lines = array();
140    $str = '';
141    while (!feof($bz)) {
142        //8192 seems to be the maximum buffersize?
143        $buffer = bzread($bz,8192);
144        if(($buffer === false) || (bzerrno($bz) !== 0)) {
145            return false;
146        }
147        $str = $str . $buffer;
148        if($array) {
149            $pos = strpos($str, "\n");
150            while($pos !== false) {
151                $lines[] = substr($str, 0, $pos+1);
152                $str = substr($str, $pos+1);
153                $pos = strpos($str, "\n");
154            }
155        }
156    }
157    bzclose($bz);
158    if($array) {
159        if($str !== '') $lines[] = $str;
160        return $lines;
161    }
162    return $str;
163}
164
165/**
166 * Used to write out a DokuWiki page to file, and send IO_WIKIPAGE_WRITE events.
167 *
168 * This generates an action event and delegates to io_saveFile().
169 * Action plugins are allowed to modify the page content in transit.
170 * The file path should not be changed.
171 * (The append parameter is set to false.)
172 *
173 * Event data:
174 * $data[0]    The raw arguments for io_saveFile as an array.
175 * $data[1]    ns: The colon separated namespace path minus the trailing page name. (false if root ns)
176 * $data[2]    page_name: The wiki page name.
177 * $data[3]    rev: The page revision, false for current wiki pages.
178 *
179 * @author Ben Coburn <btcoburn@silicodon.net>
180 *
181 * @param string $file      filename
182 * @param string $content
183 * @param string $id        page id
184 * @param int|bool $rev timestamp of revision
185 * @return bool
186 */
187function io_writeWikiPage($file, $content, $id, $rev=false) {
188    if (empty($rev)) { $rev = false; }
189    if ($rev===false) { io_createNamespace($id); } // create namespaces as needed
190    $data = array(array($file, $content, false), getNS($id), noNS($id), $rev);
191    return trigger_event('IO_WIKIPAGE_WRITE', $data, '_io_writeWikiPage_action', false);
192}
193
194/**
195 * Callback adapter for io_saveFile().
196 * @author Ben Coburn <btcoburn@silicodon.net>
197 *
198 * @param array $data event data
199 * @return bool
200 */
201function _io_writeWikiPage_action($data) {
202    if (is_array($data) && is_array($data[0]) && count($data[0])===3) {
203        $ok = call_user_func_array('io_saveFile', $data[0]);
204        // for attic files make sure the file has the mtime of the revision
205        if($ok && is_int($data[3]) && $data[3] > 0) {
206            @touch($data[0][0], $data[3]);
207        }
208        return $ok;
209    } else {
210        return false; //callback error
211    }
212}
213
214/**
215 * Internal function to save contents to a file.
216 *
217 * @author  Andreas Gohr <andi@splitbrain.org>
218 *
219 * @param string $file filename path to file
220 * @param string $content
221 * @param bool   $append
222 * @return bool true on success, otherwise false
223 */
224function _io_saveFile($file, $content, $append) {
225    global $conf;
226    $mode = ($append) ? 'ab' : 'wb';
227    $fileexists = file_exists($file);
228
229    if(substr($file,-3) == '.gz'){
230        if(!DOKU_HAS_GZIP) return false;
231        $fh = @gzopen($file,$mode.'9');
232        if(!$fh) return false;
233        gzwrite($fh, $content);
234        gzclose($fh);
235    }else if(substr($file,-4) == '.bz2'){
236        if(!DOKU_HAS_BZIP) return false;
237        if($append) {
238            $bzcontent = bzfile($file);
239            if($bzcontent === false) return false;
240            $content = $bzcontent.$content;
241        }
242        $fh = @bzopen($file,'w');
243        if(!$fh) return false;
244        bzwrite($fh, $content);
245        bzclose($fh);
246    }else{
247        $fh = @fopen($file,$mode);
248        if(!$fh) return false;
249        fwrite($fh, $content);
250        fclose($fh);
251    }
252
253    if(!$fileexists and !empty($conf['fperm'])) chmod($file, $conf['fperm']);
254    return true;
255}
256
257/**
258 * Saves $content to $file.
259 *
260 * If the third parameter is set to true the given content
261 * will be appended.
262 *
263 * Uses gzip if extension is .gz
264 * and bz2 if extension is .bz2
265 *
266 * @author  Andreas Gohr <andi@splitbrain.org>
267 *
268 * @param string $file filename path to file
269 * @param string $content
270 * @param bool   $append
271 * @return bool true on success, otherwise false
272 */
273function io_saveFile($file, $content, $append=false) {
274    io_makeFileDir($file);
275    io_lock($file);
276    if(!_io_saveFile($file, $content, $append)) {
277        msg("Writing $file failed",-1);
278        io_unlock($file);
279        return false;
280    }
281    io_unlock($file);
282    return true;
283}
284
285/**
286 * Replace one or more occurrences of a line in a file.
287 *
288 * The default, when $maxlines is 0 is to delete all matching lines then append a single line.
289 * A regex that matches any part of the line will remove the entire line in this mode.
290 * Captures in $newline are not available.
291 *
292 * Otherwise each line is matched and replaced individually, up to the first $maxlines lines
293 * or all lines if $maxlines is -1. If $regex is true then captures can be used in $newline.
294 *
295 * Be sure to include the trailing newline in $oldline when replacing entire lines.
296 *
297 * Uses gzip if extension is .gz
298 * and bz2 if extension is .bz2
299 *
300 * @author Steven Danz <steven-danz@kc.rr.com>
301 * @author Christopher Smith <chris@jalakai.co.uk>
302 * @author Patrick Brown <ptbrown@whoopdedo.org>
303 *
304 * @param string $file     filename
305 * @param string $oldline  exact linematch to remove
306 * @param string $newline  new line to insert
307 * @param bool   $regex    use regexp?
308 * @param int    $maxlines number of occurrences of the line to replace
309 * @return bool true on success
310 */
311function io_replaceInFile($file, $oldline, $newline, $regex=false, $maxlines=0) {
312    if ((string)$oldline === '') {
313        trigger_error('$oldline parameter cannot be empty in io_replaceInFile()', E_USER_WARNING);
314        return false;
315    }
316
317    if (!file_exists($file)) return true;
318
319    io_lock($file);
320
321    // load into array
322    if(substr($file,-3) == '.gz'){
323        if(!DOKU_HAS_GZIP) return false;
324        $lines = gzfile($file);
325    }else if(substr($file,-4) == '.bz2'){
326        if(!DOKU_HAS_BZIP) return false;
327        $lines = bzfile($file, true);
328    }else{
329        $lines = file($file);
330    }
331
332    // make non-regexes into regexes
333    $pattern = $regex ? $oldline : '/^'.preg_quote($oldline,'/').'$/';
334    $replace = $regex ? $newline : addcslashes($newline, '\$');
335
336    // remove matching lines
337    if ($maxlines > 0) {
338        $count = 0;
339        $matched = 0;
340        foreach($lines as $i => $line) {
341            if($count >= $maxlines) break;
342            // $matched will be set to 0|1 depending on whether pattern is matched and line replaced
343            $lines[$i] = preg_replace($pattern, $replace, $line, -1, $matched);
344            if ($matched) $count++;
345        }
346    } else if ($maxlines == 0) {
347        $lines = preg_grep($pattern, $lines, PREG_GREP_INVERT);
348
349        if ((string)$newline !== ''){
350            $lines[] = $newline;
351        }
352    } else {
353        $lines = preg_replace($pattern, $replace, $lines);
354    }
355
356    if(count($lines)){
357        if(!_io_saveFile($file, join('',$lines), false)) {
358            msg("Removing content from $file failed",-1);
359            io_unlock($file);
360            return false;
361        }
362    }else{
363        @unlink($file);
364    }
365
366    io_unlock($file);
367    return true;
368}
369
370/**
371 * Delete lines that match $badline from $file.
372 *
373 * Be sure to include the trailing newline in $badline
374 *
375 * @author Patrick Brown <ptbrown@whoopdedo.org>
376 *
377 * @param string $file    filename
378 * @param string $badline exact linematch to remove
379 * @param bool   $regex   use regexp?
380 * @return bool true on success
381 */
382function io_deleteFromFile($file,$badline,$regex=false){
383    return io_replaceInFile($file,$badline,null,$regex,0);
384}
385
386/**
387 * Tries to lock a file
388 *
389 * Locking is only done for io_savefile and uses directories
390 * inside $conf['lockdir']
391 *
392 * It waits maximal 3 seconds for the lock, after this time
393 * the lock is assumed to be stale and the function goes on
394 *
395 * @author Andreas Gohr <andi@splitbrain.org>
396 *
397 * @param string $file filename
398 */
399function io_lock($file){
400    global $conf;
401    // no locking if safemode hack
402    if($conf['safemodehack']) return;
403
404    $lockDir = $conf['lockdir'].'/'.md5($file);
405    @ignore_user_abort(1);
406
407    $timeStart = time();
408    do {
409        //waited longer than 3 seconds? -> stale lock
410        if ((time() - $timeStart) > 3) break;
411        $locked = @mkdir($lockDir, $conf['dmode']);
412        if($locked){
413            if(!empty($conf['dperm'])) chmod($lockDir, $conf['dperm']);
414            break;
415        }
416        usleep(50);
417    } while ($locked === false);
418}
419
420/**
421 * Unlocks a file
422 *
423 * @author Andreas Gohr <andi@splitbrain.org>
424 *
425 * @param string $file filename
426 */
427function io_unlock($file){
428    global $conf;
429    // no locking if safemode hack
430    if($conf['safemodehack']) return;
431
432    $lockDir = $conf['lockdir'].'/'.md5($file);
433    @rmdir($lockDir);
434    @ignore_user_abort(0);
435}
436
437/**
438 * Create missing namespace directories and send the IO_NAMESPACE_CREATED events
439 * in the order of directory creation. (Parent directories first.)
440 *
441 * Event data:
442 * $data[0]    ns: The colon separated namespace path minus the trailing page name.
443 * $data[1]    ns_type: 'pages' or 'media' namespace tree.
444 *
445 * @author Ben Coburn <btcoburn@silicodon.net>
446 *
447 * @param string $id page id
448 * @param string $ns_type 'pages' or 'media'
449 */
450function io_createNamespace($id, $ns_type='pages') {
451    // verify ns_type
452    $types = array('pages'=>'wikiFN', 'media'=>'mediaFN');
453    if (!isset($types[$ns_type])) {
454        trigger_error('Bad $ns_type parameter for io_createNamespace().');
455        return;
456    }
457    // make event list
458    $missing = array();
459    $ns_stack = explode(':', $id);
460    $ns = $id;
461    $tmp = dirname( $file = call_user_func($types[$ns_type], $ns) );
462    while (!@is_dir($tmp) && !(file_exists($tmp) && !is_dir($tmp))) {
463        array_pop($ns_stack);
464        $ns = implode(':', $ns_stack);
465        if (strlen($ns)==0) { break; }
466        $missing[] = $ns;
467        $tmp = dirname(call_user_func($types[$ns_type], $ns));
468    }
469    // make directories
470    io_makeFileDir($file);
471    // send the events
472    $missing = array_reverse($missing); // inside out
473    foreach ($missing as $ns) {
474        $data = array($ns, $ns_type);
475        trigger_event('IO_NAMESPACE_CREATED', $data);
476    }
477}
478
479/**
480 * Create the directory needed for the given file
481 *
482 * @author  Andreas Gohr <andi@splitbrain.org>
483 *
484 * @param string $file file name
485 */
486function io_makeFileDir($file){
487    $dir = dirname($file);
488    if(!@is_dir($dir)){
489        io_mkdir_p($dir) || msg("Creating directory $dir failed",-1);
490    }
491}
492
493/**
494 * Creates a directory hierachy.
495 *
496 * @link    http://php.net/manual/en/function.mkdir.php
497 * @author  <saint@corenova.com>
498 * @author  Andreas Gohr <andi@splitbrain.org>
499 *
500 * @param string $target filename
501 * @return bool|int|string
502 */
503function io_mkdir_p($target){
504    global $conf;
505    if (@is_dir($target)||empty($target)) return 1; // best case check first
506    if (file_exists($target) && !is_dir($target)) return 0;
507    //recursion
508    if (io_mkdir_p(substr($target,0,strrpos($target,'/')))){
509        if($conf['safemodehack']){
510            $dir = preg_replace('/^'.preg_quote(fullpath($conf['ftp']['root']),'/').'/','', $target);
511            return io_mkdir_ftp($dir);
512        }else{
513            $ret = @mkdir($target,$conf['dmode']); // crawl back up & create dir tree
514            if($ret && !empty($conf['dperm'])) chmod($target, $conf['dperm']);
515            return $ret;
516        }
517    }
518    return 0;
519}
520
521/**
522 * Recursively delete a directory
523 *
524 * @author Andreas Gohr <andi@splitbrain.org>
525 * @param string $path
526 * @param bool   $removefiles defaults to false which will delete empty directories only
527 * @return bool
528 */
529function io_rmdir($path, $removefiles = false) {
530    if(!is_string($path) || $path == "") return false;
531    if(!file_exists($path)) return true; // it's already gone or was never there, count as success
532
533    if(is_dir($path) && !is_link($path)) {
534        $dirs  = array();
535        $files = array();
536
537        if(!$dh = @opendir($path)) return false;
538        while(false !== ($f = readdir($dh))) {
539            if($f == '..' || $f == '.') continue;
540
541            // collect dirs and files first
542            if(is_dir("$path/$f") && !is_link("$path/$f")) {
543                $dirs[] = "$path/$f";
544            } else if($removefiles) {
545                $files[] = "$path/$f";
546            } else {
547                return false; // abort when non empty
548            }
549
550        }
551        closedir($dh);
552
553        // now traverse into  directories first
554        foreach($dirs as $dir) {
555            if(!io_rmdir($dir, $removefiles)) return false; // abort on any error
556        }
557
558        // now delete files
559        foreach($files as $file) {
560            if(!@unlink($file)) return false; //abort on any error
561        }
562
563        // remove self
564        return @rmdir($path);
565    } else if($removefiles) {
566        return @unlink($path);
567    }
568    return false;
569}
570
571/**
572 * Creates a directory using FTP
573 *
574 * This is used when the safemode workaround is enabled
575 *
576 * @author <andi@splitbrain.org>
577 *
578 * @param string $dir name of the new directory
579 * @return false|string
580 */
581function io_mkdir_ftp($dir){
582    global $conf;
583
584    if(!function_exists('ftp_connect')){
585        msg("FTP support not found - safemode workaround not usable",-1);
586        return false;
587    }
588
589    $conn = @ftp_connect($conf['ftp']['host'],$conf['ftp']['port'],10);
590    if(!$conn){
591        msg("FTP connection failed",-1);
592        return false;
593    }
594
595    if(!@ftp_login($conn, $conf['ftp']['user'], conf_decodeString($conf['ftp']['pass']))){
596        msg("FTP login failed",-1);
597        return false;
598    }
599
600    //create directory
601    $ok = @ftp_mkdir($conn, $dir);
602    //set permissions
603    @ftp_site($conn,sprintf("CHMOD %04o %s",$conf['dmode'],$dir));
604
605    @ftp_close($conn);
606    return $ok;
607}
608
609/**
610 * Creates a unique temporary directory and returns
611 * its path.
612 *
613 * @author Michael Klier <chi@chimeric.de>
614 *
615 * @return false|string path to new directory or false
616 */
617function io_mktmpdir() {
618    global $conf;
619
620    $base = $conf['tmpdir'];
621    $dir  = md5(uniqid(mt_rand(), true));
622    $tmpdir = $base.'/'.$dir;
623
624    if(io_mkdir_p($tmpdir)) {
625        return($tmpdir);
626    } else {
627        return false;
628    }
629}
630
631/**
632 * downloads a file from the net and saves it
633 *
634 * if $useAttachment is false,
635 * - $file is the full filename to save the file, incl. path
636 * - if successful will return true, false otherwise
637 *
638 * if $useAttachment is true,
639 * - $file is the directory where the file should be saved
640 * - if successful will return the name used for the saved file, false otherwise
641 *
642 * @author Andreas Gohr <andi@splitbrain.org>
643 * @author Chris Smith <chris@jalakai.co.uk>
644 *
645 * @param string $url           url to download
646 * @param string $file          path to file or directory where to save
647 * @param bool   $useAttachment true: try to use name of download, uses otherwise $defaultName
648 *                              false: uses $file as path to file
649 * @param string $defaultName   fallback for if using $useAttachment
650 * @param int    $maxSize       maximum file size
651 * @return bool|string          if failed false, otherwise true or the name of the file in the given dir
652 */
653function io_download($url,$file,$useAttachment=false,$defaultName='',$maxSize=2097152){
654    global $conf;
655    $http = new DokuHTTPClient();
656    $http->max_bodysize = $maxSize;
657    $http->timeout = 25; //max. 25 sec
658    $http->keep_alive = false; // we do single ops here, no need for keep-alive
659
660    $data = $http->get($url);
661    if(!$data) return false;
662
663    $name = '';
664    if ($useAttachment) {
665        if (isset($http->resp_headers['content-disposition'])) {
666            $content_disposition = $http->resp_headers['content-disposition'];
667            $match=array();
668            if (is_string($content_disposition) &&
669                    preg_match('/attachment;\s*filename\s*=\s*"([^"]*)"/i', $content_disposition, $match)) {
670
671                $name = utf8_basename($match[1]);
672            }
673
674        }
675
676        if (!$name) {
677            if (!$defaultName) return false;
678            $name = $defaultName;
679        }
680
681        $file = $file.$name;
682    }
683
684    $fileexists = file_exists($file);
685    $fp = @fopen($file,"w");
686    if(!$fp) return false;
687    fwrite($fp,$data);
688    fclose($fp);
689    if(!$fileexists and $conf['fperm']) chmod($file, $conf['fperm']);
690    if ($useAttachment) return $name;
691    return true;
692}
693
694/**
695 * Windows compatible rename
696 *
697 * rename() can not overwrite existing files on Windows
698 * this function will use copy/unlink instead
699 *
700 * @param string $from
701 * @param string $to
702 * @return bool succes or fail
703 */
704function io_rename($from,$to){
705    global $conf;
706    if(!@rename($from,$to)){
707        if(@copy($from,$to)){
708            if($conf['fperm']) chmod($to, $conf['fperm']);
709            @unlink($from);
710            return true;
711        }
712        return false;
713    }
714    return true;
715}
716
717/**
718 * Runs an external command with input and output pipes.
719 * Returns the exit code from the process.
720 *
721 * @author Tom N Harris <tnharris@whoopdedo.org>
722 *
723 * @param string $cmd
724 * @param string $input  input pipe
725 * @param string $output output pipe
726 * @return int exit code from process
727 */
728function io_exec($cmd, $input, &$output){
729    $descspec = array(
730            0=>array("pipe","r"),
731            1=>array("pipe","w"),
732            2=>array("pipe","w"));
733    $ph = proc_open($cmd, $descspec, $pipes);
734    if(!$ph) return -1;
735    fclose($pipes[2]); // ignore stderr
736    fwrite($pipes[0], $input);
737    fclose($pipes[0]);
738    $output = stream_get_contents($pipes[1]);
739    fclose($pipes[1]);
740    return proc_close($ph);
741}
742
743/**
744 * Search a file for matching lines
745 *
746 * This is probably not faster than file()+preg_grep() but less
747 * memory intensive because not the whole file needs to be loaded
748 * at once.
749 *
750 * @author Andreas Gohr <andi@splitbrain.org>
751 * @param  string $file    The file to search
752 * @param  string $pattern PCRE pattern
753 * @param  int    $max     How many lines to return (0 for all)
754 * @param  bool   $backref When true returns array with backreferences instead of lines
755 * @return array matching lines or backref, false on error
756 */
757function io_grep($file,$pattern,$max=0,$backref=false){
758    $fh = @fopen($file,'r');
759    if(!$fh) return false;
760    $matches = array();
761
762    $cnt  = 0;
763    $line = '';
764    while (!feof($fh)) {
765        $line .= fgets($fh, 4096);  // read full line
766        if(substr($line,-1) != "\n") continue;
767
768        // check if line matches
769        if(preg_match($pattern,$line,$match)){
770            if($backref){
771                $matches[] = $match;
772            }else{
773                $matches[] = $line;
774            }
775            $cnt++;
776        }
777        if($max && $max == $cnt) break;
778        $line = '';
779    }
780    fclose($fh);
781    return $matches;
782}
783
784
785/**
786 * Get size of contents of a file, for a compressed file the uncompressed size
787 * Warning: reading uncompressed size of content of bz-files requires uncompressing
788 *
789 * @author  Gerrit Uitslag <klapinklapin@gmail.com>
790 *
791 * @param string $file filename path to file
792 * @return int size of file
793 */
794function io_getSizeFile($file) {
795    if (!file_exists($file)) return 0;
796
797    if(substr($file,-3) == '.gz'){
798        $fp = @fopen($file, "rb");
799        if($fp === false) return 0;
800
801        fseek($fp, -4, SEEK_END);
802        $buffer = fread($fp, 4);
803        fclose($fp);
804        $array = unpack("V", $buffer);
805        $uncompressedsize = end($array);
806    }else if(substr($file,-4) == '.bz2'){
807        if(!DOKU_HAS_BZIP) return 0;
808
809        $bz = bzopen($file,"r");
810        if($bz === false) return 0;
811
812        $uncompressedsize = 0;
813        while (!feof($bz)) {
814            //8192 seems to be the maximum buffersize?
815            $buffer = bzread($bz,8192);
816            if(($buffer === false) || (bzerrno($bz) !== 0)) {
817                return 0;
818            }
819            $uncompressedsize += strlen($buffer);
820        }
821    }else{
822        $uncompressedsize = filesize($file);
823    }
824
825    return $uncompressedsize;
826 }
827