xref: /dokuwiki/inc/io.php (revision cbcf050340ec5c16d2d0aa9edce23ed980817dc9) 
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        $matched = 0;
342        foreach($lines as $i => $line) {
343            if($count >= $maxlines) break;
344            // $matched will be set to 0|1 depending on whether pattern is matched and line replaced
345            $lines[$i] = preg_replace($pattern, $replace, $line, -1, $matched);
346            if ($matched) $count++;
347        }
348    } else if ($maxlines == 0) {
349        $lines = preg_grep($pattern, $lines, PREG_GREP_INVERT);
350
351        if ((string)$newline !== ''){
352            $lines[] = $newline;
353        }
354    } else {
355        $lines = preg_replace($pattern, $replace, $lines);
356    }
357
358    if(count($lines)){
359        if(!_io_saveFile($file, join('',$lines), false)) {
360            msg("Removing content from $file failed",-1);
361            io_unlock($file);
362            return false;
363        }
364    }else{
365        @unlink($file);
366    }
367
368    io_unlock($file);
369    return true;
370}
371
372/**
373 * Delete lines that match $badline from $file.
374 *
375 * Be sure to include the trailing newline in $badline
376 *
377 * @author Patrick Brown <ptbrown@whoopdedo.org>
378 *
379 * @param string $file    filename
380 * @param string $badline exact linematch to remove
381 * @param bool   $regex   use regexp?
382 * @return bool true on success
383 */
384function io_deleteFromFile($file,$badline,$regex=false){
385    return io_replaceInFile($file,$badline,null,$regex,0);
386}
387
388/**
389 * Tries to lock a file
390 *
391 * Locking is only done for io_savefile and uses directories
392 * inside $conf['lockdir']
393 *
394 * It waits maximal 3 seconds for the lock, after this time
395 * the lock is assumed to be stale and the function goes on
396 *
397 * @author Andreas Gohr <andi@splitbrain.org>
398 *
399 * @param string $file filename
400 */
401function io_lock($file){
402    global $conf;
403    // no locking if safemode hack
404    if($conf['safemodehack']) return;
405
406    $lockDir = $conf['lockdir'].'/'.md5($file);
407    @ignore_user_abort(1);
408
409    $timeStart = time();
410    do {
411        //waited longer than 3 seconds? -> stale lock
412        if ((time() - $timeStart) > 3) break;
413        $locked = @mkdir($lockDir, $conf['dmode']);
414        if($locked){
415            if(!empty($conf['dperm'])) chmod($lockDir, $conf['dperm']);
416            break;
417        }
418        usleep(50);
419    } while ($locked === false);
420}
421
422/**
423 * Unlocks a file
424 *
425 * @author Andreas Gohr <andi@splitbrain.org>
426 *
427 * @param string $file filename
428 */
429function io_unlock($file){
430    global $conf;
431    // no locking if safemode hack
432    if($conf['safemodehack']) return;
433
434    $lockDir = $conf['lockdir'].'/'.md5($file);
435    @rmdir($lockDir);
436    @ignore_user_abort(0);
437}
438
439/**
440 * Create missing namespace directories and send the IO_NAMESPACE_CREATED events
441 * in the order of directory creation. (Parent directories first.)
442 *
443 * Event data:
444 * $data[0]    ns: The colon separated namespace path minus the trailing page name.
445 * $data[1]    ns_type: 'pages' or 'media' namespace tree.
446 *
447 * @author Ben Coburn <btcoburn@silicodon.net>
448 *
449 * @param string $id page id
450 * @param string $ns_type 'pages' or 'media'
451 */
452function io_createNamespace($id, $ns_type='pages') {
453    // verify ns_type
454    $types = array('pages'=>'wikiFN', 'media'=>'mediaFN');
455    if (!isset($types[$ns_type])) {
456        trigger_error('Bad $ns_type parameter for io_createNamespace().');
457        return;
458    }
459    // make event list
460    $missing = array();
461    $ns_stack = explode(':', $id);
462    $ns = $id;
463    $tmp = dirname( $file = call_user_func($types[$ns_type], $ns) );
464    while (!@is_dir($tmp) && !(file_exists($tmp) && !is_dir($tmp))) {
465        array_pop($ns_stack);
466        $ns = implode(':', $ns_stack);
467        if (strlen($ns)==0) { break; }
468        $missing[] = $ns;
469        $tmp = dirname(call_user_func($types[$ns_type], $ns));
470    }
471    // make directories
472    io_makeFileDir($file);
473    // send the events
474    $missing = array_reverse($missing); // inside out
475    foreach ($missing as $ns) {
476        $data = array($ns, $ns_type);
477        trigger_event('IO_NAMESPACE_CREATED', $data);
478    }
479}
480
481/**
482 * Create the directory needed for the given file
483 *
484 * @author  Andreas Gohr <andi@splitbrain.org>
485 *
486 * @param string $file file name
487 */
488function io_makeFileDir($file){
489    $dir = dirname($file);
490    if(!@is_dir($dir)){
491        io_mkdir_p($dir) || msg("Creating directory $dir failed",-1);
492    }
493}
494
495/**
496 * Creates a directory hierachy.
497 *
498 * @link    http://php.net/manual/en/function.mkdir.php
499 * @author  <saint@corenova.com>
500 * @author  Andreas Gohr <andi@splitbrain.org>
501 *
502 * @param string $target filename
503 * @return bool|int|string
504 */
505function io_mkdir_p($target){
506    global $conf;
507    if (@is_dir($target)||empty($target)) return 1; // best case check first
508    if (file_exists($target) && !is_dir($target)) return 0;
509    //recursion
510    if (io_mkdir_p(substr($target,0,strrpos($target,'/')))){
511        if($conf['safemodehack']){
512            $dir = preg_replace('/^'.preg_quote(fullpath($conf['ftp']['root']),'/').'/','', $target);
513            return io_mkdir_ftp($dir);
514        }else{
515            $ret = @mkdir($target,$conf['dmode']); // crawl back up & create dir tree
516            if($ret && !empty($conf['dperm'])) chmod($target, $conf['dperm']);
517            return $ret;
518        }
519    }
520    return 0;
521}
522
523/**
524 * Recursively delete a directory
525 *
526 * @author Andreas Gohr <andi@splitbrain.org>
527 * @param string $path
528 * @param bool   $removefiles defaults to false which will delete empty directories only
529 * @return bool
530 */
531function io_rmdir($path, $removefiles = false) {
532    if(!is_string($path) || $path == "") return false;
533    if(!file_exists($path)) return true; // it's already gone or was never there, count as success
534
535    if(is_dir($path) && !is_link($path)) {
536        $dirs  = array();
537        $files = array();
538
539        if(!$dh = @opendir($path)) return false;
540        while(false !== ($f = readdir($dh))) {
541            if($f == '..' || $f == '.') continue;
542
543            // collect dirs and files first
544            if(is_dir("$path/$f") && !is_link("$path/$f")) {
545                $dirs[] = "$path/$f";
546            } else if($removefiles) {
547                $files[] = "$path/$f";
548            } else {
549                return false; // abort when non empty
550            }
551
552        }
553        closedir($dh);
554
555        // now traverse into  directories first
556        foreach($dirs as $dir) {
557            if(!io_rmdir($dir, $removefiles)) return false; // abort on any error
558        }
559
560        // now delete files
561        foreach($files as $file) {
562            if(!@unlink($file)) return false; //abort on any error
563        }
564
565        // remove self
566        return @rmdir($path);
567    } else if($removefiles) {
568        return @unlink($path);
569    }
570    return false;
571}
572
573/**
574 * Creates a directory using FTP
575 *
576 * This is used when the safemode workaround is enabled
577 *
578 * @author <andi@splitbrain.org>
579 *
580 * @param string $dir name of the new directory
581 * @return false|string
582 */
583function io_mkdir_ftp($dir){
584    global $conf;
585
586    if(!function_exists('ftp_connect')){
587        msg("FTP support not found - safemode workaround not usable",-1);
588        return false;
589    }
590
591    $conn = @ftp_connect($conf['ftp']['host'],$conf['ftp']['port'],10);
592    if(!$conn){
593        msg("FTP connection failed",-1);
594        return false;
595    }
596
597    if(!@ftp_login($conn, $conf['ftp']['user'], conf_decodeString($conf['ftp']['pass']))){
598        msg("FTP login failed",-1);
599        return false;
600    }
601
602    //create directory
603    $ok = @ftp_mkdir($conn, $dir);
604    //set permissions
605    @ftp_site($conn,sprintf("CHMOD %04o %s",$conf['dmode'],$dir));
606
607    @ftp_close($conn);
608    return $ok;
609}
610
611/**
612 * Creates a unique temporary directory and returns
613 * its path.
614 *
615 * @author Michael Klier <chi@chimeric.de>
616 *
617 * @return false|string path to new directory or false
618 */
619function io_mktmpdir() {
620    global $conf;
621
622    $base = $conf['tmpdir'];
623    $dir  = md5(uniqid(mt_rand(), true));
624    $tmpdir = $base.'/'.$dir;
625
626    if(io_mkdir_p($tmpdir)) {
627        return($tmpdir);
628    } else {
629        return false;
630    }
631}
632
633/**
634 * downloads a file from the net and saves it
635 *
636 * if $useAttachment is false,
637 * - $file is the full filename to save the file, incl. path
638 * - if successful will return true, false otherwise
639 *
640 * if $useAttachment is true,
641 * - $file is the directory where the file should be saved
642 * - if successful will return the name used for the saved file, false otherwise
643 *
644 * @author Andreas Gohr <andi@splitbrain.org>
645 * @author Chris Smith <chris@jalakai.co.uk>
646 *
647 * @param string $url           url to download
648 * @param string $file          path to file or directory where to save
649 * @param bool   $useAttachment if true: try to use name of download, uses otherwise $defaultName, false: uses $file as path to file
650 * @param string $defaultName   fallback for if using $useAttachment
651 * @param int    $maxSize       maximum file size
652 * @return bool|string          if failed false, otherwise true or the name of the file in the given dir
653 */
654function io_download($url,$file,$useAttachment=false,$defaultName='',$maxSize=2097152){
655    global $conf;
656    $http = new DokuHTTPClient();
657    $http->max_bodysize = $maxSize;
658    $http->timeout = 25; //max. 25 sec
659    $http->keep_alive = false; // we do single ops here, no need for keep-alive
660
661    $data = $http->get($url);
662    if(!$data) return false;
663
664    $name = '';
665    if ($useAttachment) {
666        if (isset($http->resp_headers['content-disposition'])) {
667            $content_disposition = $http->resp_headers['content-disposition'];
668            $match=array();
669            if (is_string($content_disposition) &&
670                    preg_match('/attachment;\s*filename\s*=\s*"([^"]*)"/i', $content_disposition, $match)) {
671
672                $name = utf8_basename($match[1]);
673            }
674
675        }
676
677        if (!$name) {
678            if (!$defaultName) return false;
679            $name = $defaultName;
680        }
681
682        $file = $file.$name;
683    }
684
685    $fileexists = file_exists($file);
686    $fp = @fopen($file,"w");
687    if(!$fp) return false;
688    fwrite($fp,$data);
689    fclose($fp);
690    if(!$fileexists and $conf['fperm']) chmod($file, $conf['fperm']);
691    if ($useAttachment) return $name;
692    return true;
693}
694
695/**
696 * Windows compatible rename
697 *
698 * rename() can not overwrite existing files on Windows
699 * this function will use copy/unlink instead
700 *
701 * @param string $from
702 * @param string $to
703 * @return bool succes or fail
704 */
705function io_rename($from,$to){
706    global $conf;
707    if(!@rename($from,$to)){
708        if(@copy($from,$to)){
709            if($conf['fperm']) chmod($to, $conf['fperm']);
710            @unlink($from);
711            return true;
712        }
713        return false;
714    }
715    return true;
716}
717
718/**
719 * Runs an external command with input and output pipes.
720 * Returns the exit code from the process.
721 *
722 * @author Tom N Harris <tnharris@whoopdedo.org>
723 *
724 * @param string $cmd
725 * @param string $input  input pipe
726 * @param string $output output pipe
727 * @return int exit code from process
728 */
729function io_exec($cmd, $input, &$output){
730    $descspec = array(
731            0=>array("pipe","r"),
732            1=>array("pipe","w"),
733            2=>array("pipe","w"));
734    $ph = proc_open($cmd, $descspec, $pipes);
735    if(!$ph) return -1;
736    fclose($pipes[2]); // ignore stderr
737    fwrite($pipes[0], $input);
738    fclose($pipes[0]);
739    $output = stream_get_contents($pipes[1]);
740    fclose($pipes[1]);
741    return proc_close($ph);
742}
743
744/**
745 * Search a file for matching lines
746 *
747 * This is probably not faster than file()+preg_grep() but less
748 * memory intensive because not the whole file needs to be loaded
749 * at once.
750 *
751 * @author Andreas Gohr <andi@splitbrain.org>
752 * @param  string $file    The file to search
753 * @param  string $pattern PCRE pattern
754 * @param  int    $max     How many lines to return (0 for all)
755 * @param  bool   $backref When true returns array with backreferences instead of lines
756 * @return array matching lines or backref, false on error
757 */
758function io_grep($file,$pattern,$max=0,$backref=false){
759    $fh = @fopen($file,'r');
760    if(!$fh) return false;
761    $matches = array();
762
763    $cnt  = 0;
764    $line = '';
765    while (!feof($fh)) {
766        $line .= fgets($fh, 4096);  // read full line
767        if(substr($line,-1) != "\n") continue;
768
769        // check if line matches
770        if(preg_match($pattern,$line,$match)){
771            if($backref){
772                $matches[] = $match;
773            }else{
774                $matches[] = $line;
775            }
776            $cnt++;
777        }
778        if($max && $max == $cnt) break;
779        $line = '';
780    }
781    fclose($fh);
782    return $matches;
783}
784
785
786/**
787 * Get size of contents of a file, for a compressed file the uncompressed size
788 * Warning: reading uncompressed size of content of bz-files requires uncompressing
789 *
790 * @author  Gerrit Uitslag <klapinklapin@gmail.com>
791 *
792 * @param string $file filename path to file
793 * @return int size of file
794 */
795function io_getSizeFile($file) {
796    if (!file_exists($file)) return 0;
797
798    if(substr($file,-3) == '.gz'){
799        $fp = @fopen($file, "rb");
800        if($fp === false) return 0;
801
802        fseek($fp, -4, SEEK_END);
803        $buffer = fread($fp, 4);
804        fclose($fp);
805        $array = unpack("V", $buffer);
806        $uncompressedsize = end($array);
807    }else if(substr($file,-4) == '.bz2'){
808        if(!DOKU_HAS_BZIP) return 0;
809
810        $bz = bzopen($file,"r");
811        if($bz === false) return 0;
812
813        $uncompressedsize = 0;
814        while (!feof($bz)) {
815            //8192 seems to be the maximum buffersize?
816            $buffer = bzread($bz,8192);
817            if(($buffer === false) || (bzerrno($bz) !== 0)) {
818                return 0;
819            }
820            $uncompressedsize += strlen($buffer);
821        }
822    }else{
823        $uncompressedsize = filesize($file);
824    }
825
826    return $uncompressedsize;
827 }
828