xref: /dokuwiki/inc/io.php (revision c0e0ec79d7123838c2f8b50236539247e4888830) 
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        $ok = call_user_func_array('io_saveFile', $data[0]);
206        // for attic files make sure the file has the mtime of the revision
207        if($ok && is_int($data[3]) && $data[3] > 0) {
208            @touch($data[0][0], $data[3]);
209        }
210        return $ok;
211    } else {
212        return false; //callback error
213    }
214}
215
216/**
217 * Internal function to save contents to a file.
218 *
219 * @author  Andreas Gohr <andi@splitbrain.org>
220 *
221 * @param string $file filename path to file
222 * @param string $content
223 * @param bool   $append
224 * @return bool true on success, otherwise false
225 */
226function _io_saveFile($file, $content, $append) {
227    global $conf;
228    $mode = ($append) ? 'ab' : 'wb';
229    $fileexists = file_exists($file);
230
231    if(substr($file,-3) == '.gz'){
232        if(!DOKU_HAS_GZIP) return false;
233        $fh = @gzopen($file,$mode.'9');
234        if(!$fh) return false;
235        gzwrite($fh, $content);
236        gzclose($fh);
237    }else if(substr($file,-4) == '.bz2'){
238        if(!DOKU_HAS_BZIP) return false;
239        if($append) {
240            $bzcontent = bzfile($file);
241            if($bzcontent === false) return false;
242            $content = $bzcontent.$content;
243        }
244        $fh = @bzopen($file,'w');
245        if(!$fh) return false;
246        bzwrite($fh, $content);
247        bzclose($fh);
248    }else{
249        $fh = @fopen($file,$mode);
250        if(!$fh) return false;
251        fwrite($fh, $content);
252        fclose($fh);
253    }
254
255    if(!$fileexists and !empty($conf['fperm'])) chmod($file, $conf['fperm']);
256    return true;
257}
258
259/**
260 * Saves $content to $file.
261 *
262 * If the third parameter is set to true the given content
263 * will be appended.
264 *
265 * Uses gzip if extension is .gz
266 * and bz2 if extension is .bz2
267 *
268 * @author  Andreas Gohr <andi@splitbrain.org>
269 *
270 * @param string $file filename path to file
271 * @param string $content
272 * @param bool   $append
273 * @return bool true on success, otherwise false
274 */
275function io_saveFile($file, $content, $append=false) {
276    io_makeFileDir($file);
277    io_lock($file);
278    if(!_io_saveFile($file, $content, $append)) {
279        msg("Writing $file failed",-1);
280        io_unlock($file);
281        return false;
282    }
283    io_unlock($file);
284    return true;
285}
286
287/**
288 * Replace one or more occurrences of a line in a file.
289 *
290 * The default, when $maxlines is 0 is to delete all matching lines then append a single line.
291 * A regex that matches any part of the line will remove the entire line in this mode.
292 * Captures in $newline are not available.
293 *
294 * Otherwise each line is matched and replaced individually, up to the first $maxlines lines
295 * or all lines if $maxlines is -1. If $regex is true then captures can be used in $newline.
296 *
297 * Be sure to include the trailing newline in $oldline when replacing entire lines.
298 *
299 * Uses gzip if extension is .gz
300 * and bz2 if extension is .bz2
301 *
302 * @author Steven Danz <steven-danz@kc.rr.com>
303 * @author Christopher Smith <chris@jalakai.co.uk>
304 * @author Patrick Brown <ptbrown@whoopdedo.org>
305 *
306 * @param string $file     filename
307 * @param string $oldline  exact linematch to remove
308 * @param string $newline  new line to insert
309 * @param bool   $regex    use regexp?
310 * @param int    $maxlines number of occurrences of the line to replace
311 * @return bool true on success
312 */
313function io_replaceInFile($file, $oldline, $newline, $regex=false, $maxlines=0) {
314    if ((string)$oldline === '') {
315        trigger_error('$oldline parameter cannot be empty in io_replaceInFile()', E_USER_WARNING);
316        return false;
317    }
318
319    if (!file_exists($file)) return true;
320
321    io_lock($file);
322
323    // load into array
324    if(substr($file,-3) == '.gz'){
325        if(!DOKU_HAS_GZIP) return false;
326        $lines = gzfile($file);
327    }else if(substr($file,-4) == '.bz2'){
328        if(!DOKU_HAS_BZIP) return false;
329        $lines = bzfile($file, true);
330    }else{
331        $lines = file($file);
332    }
333
334    // make non-regexes into regexes
335    $pattern = $regex ? $oldline : '/^'.preg_quote($oldline,'/').'$/';
336    $replace = $regex ? $newline : addcslashes($newline, '\$');
337
338    // remove matching lines
339    if ($maxlines > 0) {
340        $count = 0;
341        $matched = 0;
342        while (($count < $maxlines) && (list($i,$line) = each($lines))) {
343            // $matched will be set to 0|1 depending on whether pattern is matched and line replaced
344            $lines[$i] = preg_replace($pattern, $replace, $line, -1, $matched);
345            if ($matched) $count++;
346        }
347    } else if ($maxlines == 0) {
348        $lines = preg_grep($pattern, $lines, PREG_GREP_INVERT);
349
350        if ((string)$newline !== ''){
351            $lines[] = $newline;
352        }
353    } else {
354        $lines = preg_replace($pattern, $replace, $lines);
355    }
356
357    if(count($lines)){
358        if(!_io_saveFile($file, join('',$lines), false)) {
359            msg("Removing content from $file failed",-1);
360            io_unlock($file);
361            return false;
362        }
363    }else{
364        @unlink($file);
365    }
366
367    io_unlock($file);
368    return true;
369}
370
371/**
372 * Delete lines that match $badline from $file.
373 *
374 * Be sure to include the trailing newline in $badline
375 *
376 * @author Patrick Brown <ptbrown@whoopdedo.org>
377 *
378 * @param string $file    filename
379 * @param string $badline exact linematch to remove
380 * @param bool   $regex   use regexp?
381 * @return bool true on success
382 */
383function io_deleteFromFile($file,$badline,$regex=false){
384    return io_replaceInFile($file,$badline,null,$regex,0);
385}
386
387/**
388 * Tries to lock a file
389 *
390 * Locking is only done for io_savefile and uses directories
391 * inside $conf['lockdir']
392 *
393 * It waits maximal 3 seconds for the lock, after this time
394 * the lock is assumed to be stale and the function goes on
395 *
396 * @author Andreas Gohr <andi@splitbrain.org>
397 *
398 * @param string $file filename
399 */
400function io_lock($file){
401    global $conf;
402    // no locking if safemode hack
403    if($conf['safemodehack']) return;
404
405    $lockDir = $conf['lockdir'].'/'.md5($file);
406    @ignore_user_abort(1);
407
408    $timeStart = time();
409    do {
410        //waited longer than 3 seconds? -> stale lock
411        if ((time() - $timeStart) > 3) break;
412        $locked = @mkdir($lockDir, $conf['dmode']);
413        if($locked){
414            if(!empty($conf['dperm'])) chmod($lockDir, $conf['dperm']);
415            break;
416        }
417        usleep(50);
418    } while ($locked === false);
419}
420
421/**
422 * Unlocks a file
423 *
424 * @author Andreas Gohr <andi@splitbrain.org>
425 *
426 * @param string $file filename
427 */
428function io_unlock($file){
429    global $conf;
430    // no locking if safemode hack
431    if($conf['safemodehack']) return;
432
433    $lockDir = $conf['lockdir'].'/'.md5($file);
434    @rmdir($lockDir);
435    @ignore_user_abort(0);
436}
437
438/**
439 * Create missing namespace directories and send the IO_NAMESPACE_CREATED events
440 * in the order of directory creation. (Parent directories first.)
441 *
442 * Event data:
443 * $data[0]    ns: The colon separated namespace path minus the trailing page name.
444 * $data[1]    ns_type: 'pages' or 'media' namespace tree.
445 *
446 * @author Ben Coburn <btcoburn@silicodon.net>
447 *
448 * @param string $id page id
449 * @param string $ns_type 'pages' or 'media'
450 */
451function io_createNamespace($id, $ns_type='pages') {
452    // verify ns_type
453    $types = array('pages'=>'wikiFN', 'media'=>'mediaFN');
454    if (!isset($types[$ns_type])) {
455        trigger_error('Bad $ns_type parameter for io_createNamespace().');
456        return;
457    }
458    // make event list
459    $missing = array();
460    $ns_stack = explode(':', $id);
461    $ns = $id;
462    $tmp = dirname( $file = call_user_func($types[$ns_type], $ns) );
463    while (!@is_dir($tmp) && !(file_exists($tmp) && !is_dir($tmp))) {
464        array_pop($ns_stack);
465        $ns = implode(':', $ns_stack);
466        if (strlen($ns)==0) { break; }
467        $missing[] = $ns;
468        $tmp = dirname(call_user_func($types[$ns_type], $ns));
469    }
470    // make directories
471    io_makeFileDir($file);
472    // send the events
473    $missing = array_reverse($missing); // inside out
474    foreach ($missing as $ns) {
475        $data = array($ns, $ns_type);
476        trigger_event('IO_NAMESPACE_CREATED', $data);
477    }
478}
479
480/**
481 * Create the directory needed for the given file
482 *
483 * @author  Andreas Gohr <andi@splitbrain.org>
484 *
485 * @param string $file file name
486 */
487function io_makeFileDir($file){
488    $dir = dirname($file);
489    if(!@is_dir($dir)){
490        io_mkdir_p($dir) || msg("Creating directory $dir failed",-1);
491    }
492}
493
494/**
495 * Creates a directory hierachy.
496 *
497 * @link    http://php.net/manual/en/function.mkdir.php
498 * @author  <saint@corenova.com>
499 * @author  Andreas Gohr <andi@splitbrain.org>
500 *
501 * @param string $target filename
502 * @return bool|int|string
503 */
504function io_mkdir_p($target){
505    global $conf;
506    if (@is_dir($target)||empty($target)) return 1; // best case check first
507    if (file_exists($target) && !is_dir($target)) return 0;
508    //recursion
509    if (io_mkdir_p(substr($target,0,strrpos($target,'/')))){
510        if($conf['safemodehack']){
511            $dir = preg_replace('/^'.preg_quote(fullpath($conf['ftp']['root']),'/').'/','', $target);
512            return io_mkdir_ftp($dir);
513        }else{
514            $ret = @mkdir($target,$conf['dmode']); // crawl back up & create dir tree
515            if($ret && !empty($conf['dperm'])) chmod($target, $conf['dperm']);
516            return $ret;
517        }
518    }
519    return 0;
520}
521
522/**
523 * Recursively delete a directory
524 *
525 * @author Andreas Gohr <andi@splitbrain.org>
526 * @param string $path
527 * @param bool   $removefiles defaults to false which will delete empty directories only
528 * @return bool
529 */
530function io_rmdir($path, $removefiles = false) {
531    if(!is_string($path) || $path == "") return false;
532    if(!file_exists($path)) return true; // it's already gone or was never there, count as success
533
534    if(is_dir($path) && !is_link($path)) {
535        $dirs  = array();
536        $files = array();
537
538        if(!$dh = @opendir($path)) return false;
539        while(false !== ($f = readdir($dh))) {
540            if($f == '..' || $f == '.') continue;
541
542            // collect dirs and files first
543            if(is_dir("$path/$f") && !is_link("$path/$f")) {
544                $dirs[] = "$path/$f";
545            } else if($removefiles) {
546                $files[] = "$path/$f";
547            } else {
548                return false; // abort when non empty
549            }
550
551        }
552        closedir($dh);
553
554        // now traverse into  directories first
555        foreach($dirs as $dir) {
556            if(!io_rmdir($dir, $removefiles)) return false; // abort on any error
557        }
558
559        // now delete files
560        foreach($files as $file) {
561            if(!@unlink($file)) return false; //abort on any error
562        }
563
564        // remove self
565        return @rmdir($path);
566    } else if($removefiles) {
567        return @unlink($path);
568    }
569    return false;
570}
571
572/**
573 * Creates a directory using FTP
574 *
575 * This is used when the safemode workaround is enabled
576 *
577 * @author <andi@splitbrain.org>
578 *
579 * @param string $dir name of the new directory
580 * @return false|string
581 */
582function io_mkdir_ftp($dir){
583    global $conf;
584
585    if(!function_exists('ftp_connect')){
586        msg("FTP support not found - safemode workaround not usable",-1);
587        return false;
588    }
589
590    $conn = @ftp_connect($conf['ftp']['host'],$conf['ftp']['port'],10);
591    if(!$conn){
592        msg("FTP connection failed",-1);
593        return false;
594    }
595
596    if(!@ftp_login($conn, $conf['ftp']['user'], conf_decodeString($conf['ftp']['pass']))){
597        msg("FTP login failed",-1);
598        return false;
599    }
600
601    //create directory
602    $ok = @ftp_mkdir($conn, $dir);
603    //set permissions
604    @ftp_site($conn,sprintf("CHMOD %04o %s",$conf['dmode'],$dir));
605
606    @ftp_close($conn);
607    return $ok;
608}
609
610/**
611 * Creates a unique temporary directory and returns
612 * its path.
613 *
614 * @author Michael Klier <chi@chimeric.de>
615 *
616 * @return false|string path to new directory or false
617 */
618function io_mktmpdir() {
619    global $conf;
620
621    $base = $conf['tmpdir'];
622    $dir  = md5(uniqid(mt_rand(), true));
623    $tmpdir = $base.'/'.$dir;
624
625    if(io_mkdir_p($tmpdir)) {
626        return($tmpdir);
627    } else {
628        return false;
629    }
630}
631
632/**
633 * downloads a file from the net and saves it
634 *
635 * if $useAttachment is false,
636 * - $file is the full filename to save the file, incl. path
637 * - if successful will return true, false otherwise
638 *
639 * if $useAttachment is true,
640 * - $file is the directory where the file should be saved
641 * - if successful will return the name used for the saved file, false otherwise
642 *
643 * @author Andreas Gohr <andi@splitbrain.org>
644 * @author Chris Smith <chris@jalakai.co.uk>
645 *
646 * @param string $url           url to download
647 * @param string $file          path to file or directory where to save
648 * @param bool   $useAttachment if true: try to use name of download, uses otherwise $defaultName, 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