xref: /dokuwiki/inc/io.php (revision 1f0ef30d2be7b76124353bb8e0e6ad6707762073) 
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
9if(!defined('DOKU_INC')) die('meh.');
10
11/**
12 * Removes empty directories
13 *
14 * Sends IO_NAMESPACE_DELETED events for 'pages' and 'media' namespaces.
15 * Event data:
16 * $data[0]    ns: The colon separated namespace path minus the trailing page name.
17 * $data[1]    ns_type: 'pages' or 'media' namespace tree.
18 *
19 * @todo use safemode hack
20 * @param string $id      - a pageid, the namespace of that id will be tried to deleted
21 * @param string $basedir - the config name of the type to delete (datadir or mediadir usally)
22 * @return bool - true if at least one namespace was deleted
23 *
24 * @author  Andreas Gohr <andi@splitbrain.org>
25 * @author Ben Coburn <btcoburn@silicodon.net>
26 */
27function io_sweepNS($id,$basedir='datadir'){
28    global $conf;
29    $types = array ('datadir'=>'pages', 'mediadir'=>'media');
30    $ns_type = (isset($types[$basedir])?$types[$basedir]:false);
31
32    $delone = false;
33
34    //scan all namespaces
35    while(($id = getNS($id)) !== false){
36        $dir = $conf[$basedir].'/'.utf8_encodeFN(str_replace(':','/',$id));
37
38        //try to delete dir else return
39        if(@rmdir($dir)) {
40            if ($ns_type!==false) {
41                $data = array($id, $ns_type);
42                $delone = true; // we deleted at least one dir
43                trigger_event('IO_NAMESPACE_DELETED', $data);
44            }
45        } else { return $delone; }
46    }
47    return $delone;
48}
49
50/**
51 * Used to read in a DokuWiki page from file, and send IO_WIKIPAGE_READ events.
52 *
53 * Generates the action event which delegates to io_readFile().
54 * Action plugins are allowed to modify the page content in transit.
55 * The file path should not be changed.
56 *
57 * Event data:
58 * $data[0]    The raw arguments for io_readFile as an array.
59 * $data[1]    ns: The colon separated namespace path minus the trailing page name. (false if root ns)
60 * $data[2]    page_name: The wiki page name.
61 * $data[3]    rev: The page revision, false for current wiki pages.
62 *
63 * @author Ben Coburn <btcoburn@silicodon.net>
64 *
65 * @param string   $file filename
66 * @param string   $id page id
67 * @param bool|int $rev revision timestamp
68 * @return string
69 */
70function io_readWikiPage($file, $id, $rev=false) {
71    if (empty($rev)) { $rev = false; }
72    $data = array(array($file, true), getNS($id), noNS($id), $rev);
73    return trigger_event('IO_WIKIPAGE_READ', $data, '_io_readWikiPage_action', false);
74}
75
76/**
77 * Callback adapter for io_readFile().
78 *
79 * @author Ben Coburn <btcoburn@silicodon.net>
80 *
81 * @param array $data event data
82 * @return string
83 */
84function _io_readWikiPage_action($data) {
85    if (is_array($data) && is_array($data[0]) && count($data[0])===2) {
86        return call_user_func_array('io_readFile', $data[0]);
87    } else {
88        return ''; //callback error
89    }
90}
91
92/**
93 * Returns content of $file as cleaned string.
94 *
95 * Uses gzip if extension is .gz
96 *
97 * If you want to use the returned value in unserialize
98 * be sure to set $clean to false!
99 *
100 * @author  Andreas Gohr <andi@splitbrain.org>
101 *
102 * @param string $file  filename
103 * @param bool   $clean
104 * @return string|bool the file contents or false on error
105 */
106function io_readFile($file,$clean=true){
107    $ret = '';
108    if(file_exists($file)){
109        if(substr($file,-3) == '.gz'){
110            if(!DOKU_HAS_GZIP) return false;
111            $ret = gzfile($file);
112            if(is_array($ret)) $ret = join('', $ret);
113        }else if(substr($file,-4) == '.bz2'){
114            if(!DOKU_HAS_BZIP) return false;
115            $ret = bzfile($file);
116        }else{
117            $ret = file_get_contents($file);
118        }
119    }
120    if($ret === null) return false;
121    if($ret !== false && $clean){
122        return cleanText($ret);
123    }else{
124        return $ret;
125    }
126}
127/**
128 * Returns the content of a .bz2 compressed file as string
129 *
130 * @author marcel senf <marcel@rucksackreinigung.de>
131 * @author  Andreas Gohr <andi@splitbrain.org>
132 *
133 * @param string $file filename
134 * @param bool   $array return array of lines
135 * @return string|array|bool content or false on error
136 */
137function bzfile($file, $array=false) {
138    $bz = bzopen($file,"r");
139    if($bz === false) return false;
140
141    if($array) $lines = array();
142    $str = '';
143    while (!feof($bz)) {
144        //8192 seems to be the maximum buffersize?
145        $buffer = bzread($bz,8192);
146        if(($buffer === false) || (bzerrno($bz) !== 0)) {
147            return false;
148        }
149        $str = $str . $buffer;
150        if($array) {
151            $pos = strpos($str, "\n");
152            while($pos !== false) {
153                $lines[] = substr($str, 0, $pos+1);
154                $str = substr($str, $pos+1);
155                $pos = strpos($str, "\n");
156            }
157        }
158    }
159    bzclose($bz);
160    if($array) {
161        if($str !== '') $lines[] = $str;
162        return $lines;
163    }
164    return $str;
165}
166
167/**
168 * Used to write out a DokuWiki page to file, and send IO_WIKIPAGE_WRITE events.
169 *
170 * This generates an action event and delegates to io_saveFile().
171 * Action plugins are allowed to modify the page content in transit.
172 * The file path should not be changed.
173 * (The append parameter is set to false.)
174 *
175 * Event data:
176 * $data[0]    The raw arguments for io_saveFile as an array.
177 * $data[1]    ns: The colon separated namespace path minus the trailing page name. (false if root ns)
178 * $data[2]    page_name: The wiki page name.
179 * $data[3]    rev: The page revision, false for current wiki pages.
180 *
181 * @author Ben Coburn <btcoburn@silicodon.net>
182 *
183 * @param string $file      filename
184 * @param string $content
185 * @param string $id        page id
186 * @param int|bool $rev timestamp of revision
187 * @return bool
188 */
189function io_writeWikiPage($file, $content, $id, $rev=false) {
190    if (empty($rev)) { $rev = false; }
191    if ($rev===false) { io_createNamespace($id); } // create namespaces as needed
192    $data = array(array($file, $content, false), getNS($id), noNS($id), $rev);
193    return trigger_event('IO_WIKIPAGE_WRITE', $data, '_io_writeWikiPage_action', false);
194}
195
196/**
197 * Callback adapter for io_saveFile().
198 * @author Ben Coburn <btcoburn@silicodon.net>
199 *
200 * @param array $data event data
201 * @return bool
202 */
203function _io_writeWikiPage_action($data) {
204    if (is_array($data) && is_array($data[0]) && count($data[0])===3) {
205        return call_user_func_array('io_saveFile', $data[0]);
206    } else {
207        return false; //callback error
208    }
209}
210
211/**
212 * Internal function to save contents to a file.
213 *
214 * @author  Andreas Gohr <andi@splitbrain.org>
215 *
216 * @param string $file filename path to file
217 * @param string $content
218 * @param bool   $append
219 * @return bool true on success, otherwise false
220 */
221function _io_saveFile($file, $content, $append) {
222    global $conf;
223    $mode = ($append) ? 'ab' : 'wb';
224    $fileexists = file_exists($file);
225
226    if(substr($file,-3) == '.gz'){
227        if(!DOKU_HAS_GZIP) return false;
228        $fh = @gzopen($file,$mode.'9');
229        if(!$fh) return false;
230        gzwrite($fh, $content);
231        gzclose($fh);
232    }else if(substr($file,-4) == '.bz2'){
233        if(!DOKU_HAS_BZIP) return false;
234        if($append) {
235            $bzcontent = bzfile($file);
236            if($bzcontent === false) return false;
237            $content = $bzcontent.$content;
238        }
239        $fh = @bzopen($file,'w');
240        if(!$fh) return false;
241        bzwrite($fh, $content);
242        bzclose($fh);
243    }else{
244        $fh = @fopen($file,$mode);
245        if(!$fh) return false;
246        fwrite($fh, $content);
247        fclose($fh);
248    }
249
250    if(!$fileexists and !empty($conf['fperm'])) chmod($file, $conf['fperm']);
251    return true;
252}
253
254/**
255 * Saves $content to $file.
256 *
257 * If the third parameter is set to true the given content
258 * will be appended.
259 *
260 * Uses gzip if extension is .gz
261 * and bz2 if extension is .bz2
262 *
263 * @author  Andreas Gohr <andi@splitbrain.org>
264 *
265 * @param string $file filename path to file
266 * @param string $content
267 * @param bool   $append
268 * @return bool true on success, otherwise false
269 */
270function io_saveFile($file, $content, $append=false) {
271    io_makeFileDir($file);
272    io_lock($file);
273    if(!_io_saveFile($file, $content, $append)) {
274        msg("Writing $file failed",-1);
275        io_unlock($file);
276        return false;
277    }
278    io_unlock($file);
279    return true;
280}
281
282/**
283 * Replace one or more occurrences of a line in a file.
284 *
285 * The default, when $maxlines is 0 is to delete all matching lines then append a single line.
286 * A regex that matches any part of the line will remove the entire line in this mode.
287 * Captures in $newline are not available.
288 *
289 * Otherwise each line is matched and replaced individually, up to the first $maxlines lines
290 * or all lines if $maxlines is -1. If $regex is true then captures can be used in $newline.
291 *
292 * Be sure to include the trailing newline in $oldline when replacing entire lines.
293 *
294 * Uses gzip if extension is .gz
295 * and bz2 if extension is .bz2
296 *
297 * @author Steven Danz <steven-danz@kc.rr.com>
298 * @author Christopher Smith <chris@jalakai.co.uk>
299 * @author Patrick Brown <ptbrown@whoopdedo.org>
300 *
301 * @param string $file     filename
302 * @param string $oldline  exact linematch to remove
303 * @param string $newline  new line to insert
304 * @param bool   $regex    use regexp?
305 * @param int    $maxlines number of occurrences of the line to replace
306 * @return bool true on success
307 */
308function io_replaceInFile($file, $oldline, $newline, $regex=false, $maxlines=0) {
309    if ((string)$oldline === '') {
310        trigger_error('$oldline parameter cannot be empty in io_replaceInFile()', E_USER_WARNING);
311        return false;
312    }
313
314    if (!file_exists($file)) return true;
315
316    io_lock($file);
317
318    // load into array
319    if(substr($file,-3) == '.gz'){
320        if(!DOKU_HAS_GZIP) return false;
321        $lines = gzfile($file);
322    }else if(substr($file,-4) == '.bz2'){
323        if(!DOKU_HAS_BZIP) return false;
324        $lines = bzfile($file, true);
325    }else{
326        $lines = file($file);
327    }
328
329    // make non-regexes into regexes
330    $pattern = $regex ? $oldline : '/^'.preg_quote($oldline,'/').'$/';
331    $replace = $regex ? $newline : addcslashes($newline, '\$');
332
333    // remove matching lines
334    if ($maxlines > 0) {
335        $count = 0;
336        $matched = 0;
337        while (($count < $maxlines) && (list($i,$line) = each($lines))) {
338            // $matched will be set to 0|1 depending on whether pattern is matched and line replaced
339            $lines[$i] = preg_replace($pattern, $replace, $line, -1, $matched);
340            if ($matched) $count++;
341        }
342    } else if ($maxlines == 0) {
343        $lines = preg_grep($pattern, $lines, PREG_GREP_INVERT);
344
345        if ((string)$newline !== ''){
346            $lines[] = $newline;
347        }
348    } else {
349        $lines = preg_replace($pattern, $replace, $lines);
350    }
351
352    if(count($lines)){
353        if(!_io_saveFile($file, join('',$lines), false)) {
354            msg("Removing content from $file failed",-1);
355            io_unlock($file);
356            return false;
357        }
358    }else{
359        @unlink($file);
360    }
361
362    io_unlock($file);
363    return true;
364}
365
366/**
367 * Delete lines that match $badline from $file.
368 *
369 * Be sure to include the trailing newline in $badline
370 *
371 * @author Patrick Brown <ptbrown@whoopdedo.org>
372 *
373 * @param string $file    filename
374 * @param string $badline exact linematch to remove
375 * @param bool   $regex   use regexp?
376 * @return bool true on success
377 */
378function io_deleteFromFile($file,$badline,$regex=false){
379    return io_replaceInFile($file,$badline,null,$regex,0);
380}
381
382/**
383 * Tries to lock a file
384 *
385 * Locking is only done for io_savefile and uses directories
386 * inside $conf['lockdir']
387 *
388 * It waits maximal 3 seconds for the lock, after this time
389 * the lock is assumed to be stale and the function goes on
390 *
391 * @author Andreas Gohr <andi@splitbrain.org>
392 *
393 * @param string $file filename
394 */
395function io_lock($file){
396    global $conf;
397    // no locking if safemode hack
398    if($conf['safemodehack']) return;
399
400    $lockDir = $conf['lockdir'].'/'.md5($file);
401    @ignore_user_abort(1);
402
403    $timeStart = time();
404    do {
405        //waited longer than 3 seconds? -> stale lock
406        if ((time() - $timeStart) > 3) break;
407        $locked = @mkdir($lockDir, $conf['dmode']);
408        if($locked){
409            if(!empty($conf['dperm'])) chmod($lockDir, $conf['dperm']);
410            break;
411        }
412        usleep(50);
413    } while ($locked === false);
414}
415
416/**
417 * Unlocks a file
418 *
419 * @author Andreas Gohr <andi@splitbrain.org>
420 *
421 * @param string $file filename
422 */
423function io_unlock($file){
424    global $conf;
425    // no locking if safemode hack
426    if($conf['safemodehack']) return;
427
428    $lockDir = $conf['lockdir'].'/'.md5($file);
429    @rmdir($lockDir);
430    @ignore_user_abort(0);
431}
432
433/**
434 * Create missing namespace directories and send the IO_NAMESPACE_CREATED events
435 * in the order of directory creation. (Parent directories first.)
436 *
437 * Event data:
438 * $data[0]    ns: The colon separated namespace path minus the trailing page name.
439 * $data[1]    ns_type: 'pages' or 'media' namespace tree.
440 *
441 * @author Ben Coburn <btcoburn@silicodon.net>
442 *
443 * @param string $id page id
444 * @param string $ns_type 'pages' or 'media'
445 */
446function io_createNamespace($id, $ns_type='pages') {
447    // verify ns_type
448    $types = array('pages'=>'wikiFN', 'media'=>'mediaFN');
449    if (!isset($types[$ns_type])) {
450        trigger_error('Bad $ns_type parameter for io_createNamespace().');
451        return;
452    }
453    // make event list
454    $missing = array();
455    $ns_stack = explode(':', $id);
456    $ns = $id;
457    $tmp = dirname( $file = call_user_func($types[$ns_type], $ns) );
458    while (!@is_dir($tmp) && !(file_exists($tmp) && !is_dir($tmp))) {
459        array_pop($ns_stack);
460        $ns = implode(':', $ns_stack);
461        if (strlen($ns)==0) { break; }
462        $missing[] = $ns;
463        $tmp = dirname(call_user_func($types[$ns_type], $ns));
464    }
465    // make directories
466    io_makeFileDir($file);
467    // send the events
468    $missing = array_reverse($missing); // inside out
469    foreach ($missing as $ns) {
470        $data = array($ns, $ns_type);
471        trigger_event('IO_NAMESPACE_CREATED', $data);
472    }
473}
474
475/**
476 * Create the directory needed for the given file
477 *
478 * @author  Andreas Gohr <andi@splitbrain.org>
479 *
480 * @param string $file file name
481 */
482function io_makeFileDir($file){
483    $dir = dirname($file);
484    if(!@is_dir($dir)){
485        io_mkdir_p($dir) || msg("Creating directory $dir failed",-1);
486    }
487}
488
489/**
490 * Creates a directory hierachy.
491 *
492 * @link    http://www.php.net/manual/en/function.mkdir.php
493 * @author  <saint@corenova.com>
494 * @author  Andreas Gohr <andi@splitbrain.org>
495 *
496 * @param string $target filename
497 * @return bool|int|string
498 */
499function io_mkdir_p($target){
500    global $conf;
501    if (@is_dir($target)||empty($target)) return 1; // best case check first
502    if (file_exists($target) && !is_dir($target)) return 0;
503    //recursion
504    if (io_mkdir_p(substr($target,0,strrpos($target,'/')))){
505        if($conf['safemodehack']){
506            $dir = preg_replace('/^'.preg_quote(fullpath($conf['ftp']['root']),'/').'/','', $target);
507            return io_mkdir_ftp($dir);
508        }else{
509            $ret = @mkdir($target,$conf['dmode']); // crawl back up & create dir tree
510            if($ret && !empty($conf['dperm'])) chmod($target, $conf['dperm']);
511            return $ret;
512        }
513    }
514    return 0;
515}
516
517/**
518 * Recursively delete a directory
519 *
520 * @author Andreas Gohr <andi@splitbrain.org>
521 * @param string $path
522 * @param bool   $removefiles defaults to false which will delete empty directories only
523 * @return bool
524 */
525function io_rmdir($path, $removefiles = false) {
526    if(!is_string($path) || $path == "") return false;
527    if(!file_exists($path)) return true; // it's already gone or was never there, count as success
528
529    if(is_dir($path) && !is_link($path)) {
530        $dirs  = array();
531        $files = array();
532
533        if(!$dh = @opendir($path)) return false;
534        while(false !== ($f = readdir($dh))) {
535            if($f == '..' || $f == '.') continue;
536
537            // collect dirs and files first
538            if(is_dir("$path/$f") && !is_link("$path/$f")) {
539                $dirs[] = "$path/$f";
540            } else if($removefiles) {
541                $files[] = "$path/$f";
542            } else {
543                return false; // abort when non empty
544            }
545
546        }
547        closedir($dh);
548
549        // now traverse into  directories first
550        foreach($dirs as $dir) {
551            if(!io_rmdir($dir, $removefiles)) return false; // abort on any error
552        }
553
554        // now delete files
555        foreach($files as $file) {
556            if(!@unlink($file)) return false; //abort on any error
557        }
558
559        // remove self
560        return @rmdir($path);
561    } else if($removefiles) {
562        return @unlink($path);
563    }
564    return false;
565}
566
567/**
568 * Creates a directory using FTP
569 *
570 * This is used when the safemode workaround is enabled
571 *
572 * @author <andi@splitbrain.org>
573 *
574 * @param string $dir name of the new directory
575 * @return false|string
576 */
577function io_mkdir_ftp($dir){
578    global $conf;
579
580    if(!function_exists('ftp_connect')){
581        msg("FTP support not found - safemode workaround not usable",-1);
582        return false;
583    }
584
585    $conn = @ftp_connect($conf['ftp']['host'],$conf['ftp']['port'],10);
586    if(!$conn){
587        msg("FTP connection failed",-1);
588        return false;
589    }
590
591    if(!@ftp_login($conn, $conf['ftp']['user'], conf_decodeString($conf['ftp']['pass']))){
592        msg("FTP login failed",-1);
593        return false;
594    }
595
596    //create directory
597    $ok = @ftp_mkdir($conn, $dir);
598    //set permissions
599    @ftp_site($conn,sprintf("CHMOD %04o %s",$conf['dmode'],$dir));
600
601    @ftp_close($conn);
602    return $ok;
603}
604
605/**
606 * Creates a unique temporary directory and returns
607 * its path.
608 *
609 * @author Michael Klier <chi@chimeric.de>
610 *
611 * @return false|string path to new directory or false
612 */
613function io_mktmpdir() {
614    global $conf;
615
616    $base = $conf['tmpdir'];
617    $dir  = md5(uniqid(mt_rand(), true));
618    $tmpdir = $base.'/'.$dir;
619
620    if(io_mkdir_p($tmpdir)) {
621        return($tmpdir);
622    } else {
623        return false;
624    }
625}
626
627/**
628 * downloads a file from the net and saves it
629 *
630 * if $useAttachment is false,
631 * - $file is the full filename to save the file, incl. path
632 * - if successful will return true, false otherwise
633 *
634 * if $useAttachment is true,
635 * - $file is the directory where the file should be saved
636 * - if successful will return the name used for the saved file, false otherwise
637 *
638 * @author Andreas Gohr <andi@splitbrain.org>
639 * @author Chris Smith <chris@jalakai.co.uk>
640 *
641 * @param string $url           url to download
642 * @param string $file          path to file or directory where to save
643 * @param bool   $useAttachment if true: try to use name of download, uses otherwise $defaultName, false: uses $file as path to file
644 * @param string $defaultName   fallback for if using $useAttachment
645 * @param int    $maxSize       maximum file size
646 * @return bool|string          if failed false, otherwise true or the name of the file in the given dir
647 */
648function io_download($url,$file,$useAttachment=false,$defaultName='',$maxSize=2097152){
649    global $conf;
650    $http = new DokuHTTPClient();
651    $http->max_bodysize = $maxSize;
652    $http->timeout = 25; //max. 25 sec
653    $http->keep_alive = false; // we do single ops here, no need for keep-alive
654
655    $data = $http->get($url);
656    if(!$data) return false;
657
658    $name = '';
659    if ($useAttachment) {
660        if (isset($http->resp_headers['content-disposition'])) {
661            $content_disposition = $http->resp_headers['content-disposition'];
662            $match=array();
663            if (is_string($content_disposition) &&
664                    preg_match('/attachment;\s*filename\s*=\s*"([^"]*)"/i', $content_disposition, $match)) {
665
666                $name = utf8_basename($match[1]);
667            }
668
669        }
670
671        if (!$name) {
672            if (!$defaultName) return false;
673            $name = $defaultName;
674        }
675
676        $file = $file.$name;
677    }
678
679    $fileexists = file_exists($file);
680    $fp = @fopen($file,"w");
681    if(!$fp) return false;
682    fwrite($fp,$data);
683    fclose($fp);
684    if(!$fileexists and $conf['fperm']) chmod($file, $conf['fperm']);
685    if ($useAttachment) return $name;
686    return true;
687}
688
689/**
690 * Windows compatible rename
691 *
692 * rename() can not overwrite existing files on Windows
693 * this function will use copy/unlink instead
694 *
695 * @param string $from
696 * @param string $to
697 * @return bool succes or fail
698 */
699function io_rename($from,$to){
700    global $conf;
701    if(!@rename($from,$to)){
702        if(@copy($from,$to)){
703            if($conf['fperm']) chmod($to, $conf['fperm']);
704            @unlink($from);
705            return true;
706        }
707        return false;
708    }
709    return true;
710}
711
712/**
713 * Runs an external command with input and output pipes.
714 * Returns the exit code from the process.
715 *
716 * @author Tom N Harris <tnharris@whoopdedo.org>
717 *
718 * @param string $cmd
719 * @param string $input  input pipe
720 * @param string $output output pipe
721 * @return int exit code from process
722 */
723function io_exec($cmd, $input, &$output){
724    $descspec = array(
725            0=>array("pipe","r"),
726            1=>array("pipe","w"),
727            2=>array("pipe","w"));
728    $ph = proc_open($cmd, $descspec, $pipes);
729    if(!$ph) return -1;
730    fclose($pipes[2]); // ignore stderr
731    fwrite($pipes[0], $input);
732    fclose($pipes[0]);
733    $output = stream_get_contents($pipes[1]);
734    fclose($pipes[1]);
735    return proc_close($ph);
736}
737
738/**
739 * Search a file for matching lines
740 *
741 * This is probably not faster than file()+preg_grep() but less
742 * memory intensive because not the whole file needs to be loaded
743 * at once.
744 *
745 * @author Andreas Gohr <andi@splitbrain.org>
746 * @param  string $file    The file to search
747 * @param  string $pattern PCRE pattern
748 * @param  int    $max     How many lines to return (0 for all)
749 * @param  bool   $backref When true returns array with backreferences instead of lines
750 * @return array matching lines or backref, false on error
751 */
752function io_grep($file,$pattern,$max=0,$backref=false){
753    $fh = @fopen($file,'r');
754    if(!$fh) return false;
755    $matches = array();
756
757    $cnt  = 0;
758    $line = '';
759    while (!feof($fh)) {
760        $line .= fgets($fh, 4096);  // read full line
761        if(substr($line,-1) != "\n") continue;
762
763        // check if line matches
764        if(preg_match($pattern,$line,$match)){
765            if($backref){
766                $matches[] = $match;
767            }else{
768                $matches[] = $line;
769            }
770            $cnt++;
771        }
772        if($max && $max == $cnt) break;
773        $line = '';
774    }
775    fclose($fh);
776    return $matches;
777}
778
779
780/**
781 * Get size of contents of a file, for a compressed file the uncompressed size
782 * Warning: reading uncompressed size of content of bz-files requires uncompressing
783 *
784 * @author  Gerrit Uitslag <klapinklapin@gmail.com>
785 *
786 * @param string $file filename path to file
787 * @return int size of file
788 */
789function io_getSizeFile($file) {
790    if (!file_exists($file)) return 0;
791
792    if(substr($file,-3) == '.gz'){
793        $fp = @fopen($file, "rb");
794        if($fp === false) return 0;
795
796        fseek($fp, -4, SEEK_END);
797        $buffer = fread($fp, 4);
798        fclose($fp);
799        $array = unpack("V", $buffer);
800        $uncompressedsize = end($array);
801    }else if(substr($file,-4) == '.bz2'){
802        if(!DOKU_HAS_BZIP) return 0;
803
804        $bz = bzopen($file,"r");
805        if($bz === false) return 0;
806
807        $uncompressedsize = 0;
808        while (!feof($bz)) {
809            //8192 seems to be the maximum buffersize?
810            $buffer = bzread($bz,8192);
811            if(($buffer === false) || (bzerrno($bz) !== 0)) {
812                return 0;
813            }
814            if (UTF8_MBSTRING) {
815                $uncompressedsize += mb_strlen($buffer, '8bit');
816            } else {
817                $uncompressedsize += strlen($buffer);
818            }
819        }
820    }else{
821        $uncompressedsize = filesize($file);
822    }
823
824    return $uncompressedsize;
825 }
826