xref: /dokuwiki/inc/io.php (revision dd61e06f8c51d4d3153dd89c66adc9bc4d9bc06a) 
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            $ret = gzfile($file);
111            if(is_array($ret)) $ret = join('', $ret);
112        }else if(substr($file,-4) == '.bz2'){
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        return call_user_func_array('io_saveFile', $data[0]);
204    } else {
205        return false; //callback error
206    }
207}
208
209/**
210 * Internal function to save contents to a file.
211 *
212 * @author  Andreas Gohr <andi@splitbrain.org>
213 *
214 * @param string $file filename path to file
215 * @param string $content
216 * @param bool   $append
217 * @return bool true on success, otherwise false
218 */
219function _io_saveFile($file, $content, $append) {
220    global $conf;
221    $mode = ($append) ? 'ab' : 'wb';
222    $fileexists = file_exists($file);
223
224    if(substr($file,-3) == '.gz'){
225        $fh = @gzopen($file,$mode.'9');
226        if(!$fh) return false;
227        gzwrite($fh, $content);
228        gzclose($fh);
229    }else if(substr($file,-4) == '.bz2'){
230        if($append) {
231            $bzcontent = bzfile($file);
232            if($bzcontent === false) return false;
233            $content = $bzcontent.$content;
234        }
235        $fh = @bzopen($file,'w');
236        if(!$fh) return false;
237        bzwrite($fh, $content);
238        bzclose($fh);
239    }else{
240        $fh = @fopen($file,$mode);
241        if(!$fh) return false;
242        fwrite($fh, $content);
243        fclose($fh);
244    }
245
246    if(!$fileexists and !empty($conf['fperm'])) chmod($file, $conf['fperm']);
247    return true;
248}
249
250/**
251 * Saves $content to $file.
252 *
253 * If the third parameter is set to true the given content
254 * will be appended.
255 *
256 * Uses gzip if extension is .gz
257 * and bz2 if extension is .bz2
258 *
259 * @author  Andreas Gohr <andi@splitbrain.org>
260 *
261 * @param string $file filename path to file
262 * @param string $content
263 * @param bool   $append
264 * @return bool true on success, otherwise false
265 */
266function io_saveFile($file, $content, $append=false) {
267    io_makeFileDir($file);
268    io_lock($file);
269    if(!_io_saveFile($file, $content, $append)) {
270        msg("Writing $file failed",-1);
271        io_unlock($file);
272        return false;
273    }
274    io_unlock($file);
275    return true;
276}
277
278/**
279 * Replace one or more occurrences of a line in a file.
280 *
281 * The default, when $maxlines is 0 is to delete all matching lines then append a single line.
282 * A regex that matches any part of the line will remove the entire line in this mode.
283 * Captures in $newline are not available.
284 *
285 * Otherwise each line is matched and replaced individually, up to the first $maxlines lines
286 * or all lines if $maxlines is -1. If $regex is true then captures can be used in $newline.
287 *
288 * Be sure to include the trailing newline in $oldline when replacing entire lines.
289 *
290 * Uses gzip if extension is .gz
291 * and bz2 if extension is .bz2
292 *
293 * @author Steven Danz <steven-danz@kc.rr.com>
294 * @author Christopher Smith <chris@jalakai.co.uk>
295 * @author Patrick Brown <ptbrown@whoopdedo.org>
296 *
297 * @param string $file     filename
298 * @param string $oldline  exact linematch to remove
299 * @param string $newline  new line to insert
300 * @param bool   $regex    use regexp?
301 * @param int    $maxlines number of occurrences of the line to replace
302 * @return bool true on success
303 */
304function io_replaceInFile($file, $oldline, $newline, $regex=false, $maxlines=0) {
305    if ((string)$oldline === '') {
306        trigger_error('$oldline parameter cannot be empty in io_replaceInFile()', E_USER_WARNING);
307        return false;
308    }
309
310    if (!file_exists($file)) return true;
311
312    io_lock($file);
313
314    // load into array
315    if(substr($file,-3) == '.gz'){
316        $lines = gzfile($file);
317    }else if(substr($file,-4) == '.bz2'){
318        $lines = bzfile($file, true);
319    }else{
320        $lines = file($file);
321    }
322
323    // make non-regexes into regexes
324    $pattern = $regex ? $oldline : '/^'.preg_quote($oldline,'/').'$/';
325    $replace = $regex ? $newline : addcslashes($newline, '\$');
326
327    // remove matching lines
328    if ($maxlines > 0) {
329        $count = 0;
330        $matched = 0;
331        while (($count < $maxlines) && (list($i,$line) = each($lines))) {
332            // $matched will be set to 0|1 depending on whether pattern is matched and line replaced
333            $lines[$i] = preg_replace($pattern, $replace, $line, -1, $matched);
334            if ($matched) $count++;
335        }
336    } else if ($maxlines == 0) {
337        $lines = preg_grep($pattern, $lines, PREG_GREP_INVERT);
338
339        if ((string)$newline !== ''){
340            $lines[] = $newline;
341        }
342    } else {
343        $lines = preg_replace($pattern, $replace, $lines);
344    }
345
346    if(count($lines)){
347        if(!_io_saveFile($file, join('',$lines), false)) {
348            msg("Removing content from $file failed",-1);
349            io_unlock($file);
350            return false;
351        }
352    }else{
353        @unlink($file);
354    }
355
356    io_unlock($file);
357    return true;
358}
359
360/**
361 * Delete lines that match $badline from $file.
362 *
363 * Be sure to include the trailing newline in $badline
364 *
365 * @author Patrick Brown <ptbrown@whoopdedo.org>
366 *
367 * @param string $file    filename
368 * @param string $badline exact linematch to remove
369 * @param bool   $regex   use regexp?
370 * @return bool true on success
371 */
372function io_deleteFromFile($file,$badline,$regex=false){
373    return io_replaceInFile($file,$badline,null,$regex,0);
374}
375
376/**
377 * Tries to lock a file
378 *
379 * Locking is only done for io_savefile and uses directories
380 * inside $conf['lockdir']
381 *
382 * It waits maximal 3 seconds for the lock, after this time
383 * the lock is assumed to be stale and the function goes on
384 *
385 * @author Andreas Gohr <andi@splitbrain.org>
386 *
387 * @param string $file filename
388 */
389function io_lock($file){
390    global $conf;
391    // no locking if safemode hack
392    if($conf['safemodehack']) return;
393
394    $lockDir = $conf['lockdir'].'/'.md5($file);
395    @ignore_user_abort(1);
396
397    $timeStart = time();
398    do {
399        //waited longer than 3 seconds? -> stale lock
400        if ((time() - $timeStart) > 3) break;
401        $locked = @mkdir($lockDir, $conf['dmode']);
402        if($locked){
403            if(!empty($conf['dperm'])) chmod($lockDir, $conf['dperm']);
404            break;
405        }
406        usleep(50);
407    } while ($locked === false);
408}
409
410/**
411 * Unlocks a file
412 *
413 * @author Andreas Gohr <andi@splitbrain.org>
414 *
415 * @param string $file filename
416 */
417function io_unlock($file){
418    global $conf;
419    // no locking if safemode hack
420    if($conf['safemodehack']) return;
421
422    $lockDir = $conf['lockdir'].'/'.md5($file);
423    @rmdir($lockDir);
424    @ignore_user_abort(0);
425}
426
427/**
428 * Create missing namespace directories and send the IO_NAMESPACE_CREATED events
429 * in the order of directory creation. (Parent directories first.)
430 *
431 * Event data:
432 * $data[0]    ns: The colon separated namespace path minus the trailing page name.
433 * $data[1]    ns_type: 'pages' or 'media' namespace tree.
434 *
435 * @author Ben Coburn <btcoburn@silicodon.net>
436 *
437 * @param string $id page id
438 * @param string $ns_type 'pages' or 'media'
439 */
440function io_createNamespace($id, $ns_type='pages') {
441    // verify ns_type
442    $types = array('pages'=>'wikiFN', 'media'=>'mediaFN');
443    if (!isset($types[$ns_type])) {
444        trigger_error('Bad $ns_type parameter for io_createNamespace().');
445        return;
446    }
447    // make event list
448    $missing = array();
449    $ns_stack = explode(':', $id);
450    $ns = $id;
451    $tmp = dirname( $file = call_user_func($types[$ns_type], $ns) );
452    while (!@is_dir($tmp) && !(file_exists($tmp) && !is_dir($tmp))) {
453        array_pop($ns_stack);
454        $ns = implode(':', $ns_stack);
455        if (strlen($ns)==0) { break; }
456        $missing[] = $ns;
457        $tmp = dirname(call_user_func($types[$ns_type], $ns));
458    }
459    // make directories
460    io_makeFileDir($file);
461    // send the events
462    $missing = array_reverse($missing); // inside out
463    foreach ($missing as $ns) {
464        $data = array($ns, $ns_type);
465        trigger_event('IO_NAMESPACE_CREATED', $data);
466    }
467}
468
469/**
470 * Create the directory needed for the given file
471 *
472 * @author  Andreas Gohr <andi@splitbrain.org>
473 *
474 * @param string $file file name
475 */
476function io_makeFileDir($file){
477    $dir = dirname($file);
478    if(!@is_dir($dir)){
479        io_mkdir_p($dir) || msg("Creating directory $dir failed",-1);
480    }
481}
482
483/**
484 * Creates a directory hierachy.
485 *
486 * @link    http://www.php.net/manual/en/function.mkdir.php
487 * @author  <saint@corenova.com>
488 * @author  Andreas Gohr <andi@splitbrain.org>
489 *
490 * @param string $target filename
491 * @return bool|int|string
492 */
493function io_mkdir_p($target){
494    global $conf;
495    if (@is_dir($target)||empty($target)) return 1; // best case check first
496    if (file_exists($target) && !is_dir($target)) return 0;
497    //recursion
498    if (io_mkdir_p(substr($target,0,strrpos($target,'/')))){
499        if($conf['safemodehack']){
500            $dir = preg_replace('/^'.preg_quote(fullpath($conf['ftp']['root']),'/').'/','', $target);
501            return io_mkdir_ftp($dir);
502        }else{
503            $ret = @mkdir($target,$conf['dmode']); // crawl back up & create dir tree
504            if($ret && !empty($conf['dperm'])) chmod($target, $conf['dperm']);
505            return $ret;
506        }
507    }
508    return 0;
509}
510
511/**
512 * Recursively delete a directory
513 *
514 * @author Andreas Gohr <andi@splitbrain.org>
515 * @param string $path
516 * @param bool   $removefiles defaults to false which will delete empty directories only
517 * @return bool
518 */
519function io_rmdir($path, $removefiles = false) {
520    if(!is_string($path) || $path == "") return false;
521    if(!file_exists($path)) return true; // it's already gone or was never there, count as success
522
523    if(is_dir($path) && !is_link($path)) {
524        $dirs  = array();
525        $files = array();
526
527        if(!$dh = @opendir($path)) return false;
528        while(false !== ($f = readdir($dh))) {
529            if($f == '..' || $f == '.') continue;
530
531            // collect dirs and files first
532            if(is_dir("$path/$f") && !is_link("$path/$f")) {
533                $dirs[] = "$path/$f";
534            } else if($removefiles) {
535                $files[] = "$path/$f";
536            } else {
537                return false; // abort when non empty
538            }
539
540        }
541        closedir($dh);
542
543        // now traverse into  directories first
544        foreach($dirs as $dir) {
545            if(!io_rmdir($dir, $removefiles)) return false; // abort on any error
546        }
547
548        // now delete files
549        foreach($files as $file) {
550            if(!@unlink($file)) return false; //abort on any error
551        }
552
553        // remove self
554        return @rmdir($path);
555    } else if($removefiles) {
556        return @unlink($path);
557    }
558    return false;
559}
560
561/**
562 * Creates a directory using FTP
563 *
564 * This is used when the safemode workaround is enabled
565 *
566 * @author <andi@splitbrain.org>
567 *
568 * @param string $dir name of the new directory
569 * @return false|string
570 */
571function io_mkdir_ftp($dir){
572    global $conf;
573
574    if(!function_exists('ftp_connect')){
575        msg("FTP support not found - safemode workaround not usable",-1);
576        return false;
577    }
578
579    $conn = @ftp_connect($conf['ftp']['host'],$conf['ftp']['port'],10);
580    if(!$conn){
581        msg("FTP connection failed",-1);
582        return false;
583    }
584
585    if(!@ftp_login($conn, $conf['ftp']['user'], conf_decodeString($conf['ftp']['pass']))){
586        msg("FTP login failed",-1);
587        return false;
588    }
589
590    //create directory
591    $ok = @ftp_mkdir($conn, $dir);
592    //set permissions
593    @ftp_site($conn,sprintf("CHMOD %04o %s",$conf['dmode'],$dir));
594
595    @ftp_close($conn);
596    return $ok;
597}
598
599/**
600 * Creates a unique temporary directory and returns
601 * its path.
602 *
603 * @author Michael Klier <chi@chimeric.de>
604 *
605 * @return false|string path to new directory or false
606 */
607function io_mktmpdir() {
608    global $conf;
609
610    $base = $conf['tmpdir'];
611    $dir  = md5(uniqid(mt_rand(), true));
612    $tmpdir = $base.'/'.$dir;
613
614    if(io_mkdir_p($tmpdir)) {
615        return($tmpdir);
616    } else {
617        return false;
618    }
619}
620
621/**
622 * downloads a file from the net and saves it
623 *
624 * if $useAttachment is false,
625 * - $file is the full filename to save the file, incl. path
626 * - if successful will return true, false otherwise
627 *
628 * if $useAttachment is true,
629 * - $file is the directory where the file should be saved
630 * - if successful will return the name used for the saved file, false otherwise
631 *
632 * @author Andreas Gohr <andi@splitbrain.org>
633 * @author Chris Smith <chris@jalakai.co.uk>
634 *
635 * @param string $url           url to download
636 * @param string $file          path to file or directory where to save
637 * @param bool   $useAttachment if true: try to use name of download, uses otherwise $defaultName, false: uses $file as path to file
638 * @param string $defaultName   fallback for if using $useAttachment
639 * @param int    $maxSize       maximum file size
640 * @return bool|string          if failed false, otherwise true or the name of the file in the given dir
641 */
642function io_download($url,$file,$useAttachment=false,$defaultName='',$maxSize=2097152){
643    global $conf;
644    $http = new DokuHTTPClient();
645    $http->max_bodysize = $maxSize;
646    $http->timeout = 25; //max. 25 sec
647    $http->keep_alive = false; // we do single ops here, no need for keep-alive
648
649    $data = $http->get($url);
650    if(!$data) return false;
651
652    $name = '';
653    if ($useAttachment) {
654        if (isset($http->resp_headers['content-disposition'])) {
655            $content_disposition = $http->resp_headers['content-disposition'];
656            $match=array();
657            if (is_string($content_disposition) &&
658                    preg_match('/attachment;\s*filename\s*=\s*"([^"]*)"/i', $content_disposition, $match)) {
659
660                $name = utf8_basename($match[1]);
661            }
662
663        }
664
665        if (!$name) {
666            if (!$defaultName) return false;
667            $name = $defaultName;
668        }
669
670        $file = $file.$name;
671    }
672
673    $fileexists = file_exists($file);
674    $fp = @fopen($file,"w");
675    if(!$fp) return false;
676    fwrite($fp,$data);
677    fclose($fp);
678    if(!$fileexists and $conf['fperm']) chmod($file, $conf['fperm']);
679    if ($useAttachment) return $name;
680    return true;
681}
682
683/**
684 * Windows compatible rename
685 *
686 * rename() can not overwrite existing files on Windows
687 * this function will use copy/unlink instead
688 *
689 * @param string $from
690 * @param string $to
691 * @return bool succes or fail
692 */
693function io_rename($from,$to){
694    global $conf;
695    if(!@rename($from,$to)){
696        if(@copy($from,$to)){
697            if($conf['fperm']) chmod($to, $conf['fperm']);
698            @unlink($from);
699            return true;
700        }
701        return false;
702    }
703    return true;
704}
705
706/**
707 * Runs an external command with input and output pipes.
708 * Returns the exit code from the process.
709 *
710 * @author Tom N Harris <tnharris@whoopdedo.org>
711 *
712 * @param string $cmd
713 * @param string $input  input pipe
714 * @param string $output output pipe
715 * @return int exit code from process
716 */
717function io_exec($cmd, $input, &$output){
718    $descspec = array(
719            0=>array("pipe","r"),
720            1=>array("pipe","w"),
721            2=>array("pipe","w"));
722    $ph = proc_open($cmd, $descspec, $pipes);
723    if(!$ph) return -1;
724    fclose($pipes[2]); // ignore stderr
725    fwrite($pipes[0], $input);
726    fclose($pipes[0]);
727    $output = stream_get_contents($pipes[1]);
728    fclose($pipes[1]);
729    return proc_close($ph);
730}
731
732/**
733 * Search a file for matching lines
734 *
735 * This is probably not faster than file()+preg_grep() but less
736 * memory intensive because not the whole file needs to be loaded
737 * at once.
738 *
739 * @author Andreas Gohr <andi@splitbrain.org>
740 * @param  string $file    The file to search
741 * @param  string $pattern PCRE pattern
742 * @param  int    $max     How many lines to return (0 for all)
743 * @param  bool   $backref When true returns array with backreferences instead of lines
744 * @return array matching lines or backref, false on error
745 */
746function io_grep($file,$pattern,$max=0,$backref=false){
747    $fh = @fopen($file,'r');
748    if(!$fh) return false;
749    $matches = array();
750
751    $cnt  = 0;
752    $line = '';
753    while (!feof($fh)) {
754        $line .= fgets($fh, 4096);  // read full line
755        if(substr($line,-1) != "\n") continue;
756
757        // check if line matches
758        if(preg_match($pattern,$line,$match)){
759            if($backref){
760                $matches[] = $match;
761            }else{
762                $matches[] = $line;
763            }
764            $cnt++;
765        }
766        if($max && $max == $cnt) break;
767        $line = '';
768    }
769    fclose($fh);
770    return $matches;
771}
772
773