inc/io.php

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