Tar: Added extraction support for long file names

Supports POSIX ustar prefixes and GNU longlink entries
dokuwiki · Nov 4, 2012 · 421a270 · 421a270
1 parent fba11f6
commit 421a270
Show file tree

Hide file tree

Showing 4 changed files with 44 additions and 16 deletions.
diff --git a/_test/tests/inc/tar.test.php b/_test/tests/inc/tar.test.php
@@ -218,7 +218,6 @@ public function test_excludeextract(){
         }
     }
 
-
     /**
      * Check the extension to compression guesser
      */
@@ -234,4 +233,20 @@ public function test_filetype(){
         $this->assertEquals(Tar::COMPRESS_BZIP, $tar->filetype('foo.tar.BZ2'));
         $this->assertEquals(Tar::COMPRESS_BZIP, $tar->filetype('foo.tar.bz2'));
     }
+
+    public function test_longpathextract(){
+        $dir = dirname(__FILE__).'/tar';
+        $out = sys_get_temp_dir().'/dwtartest'.md5(time());
+
+        foreach(array('ustar','gnu') as $format){
+            $tar  = new Tar();
+            $tar->open("$dir/longpath-$format.tgz");
+            $tar->extract($out);
+
+            $this->assertFileExists($out.'/1234567890/1234567890/1234567890/1234567890/1234567890/1234567890/1234567890/1234567890/1234567890/1234567890/1234567890/1234567890/test.txt');
+
+            TestUtils::rdelete($out);
+        }
+    }
+
 }
diff --git a/_test/tests/inc/tar/longpath-gnu.tgz b/_test/tests/inc/tar/longpath-gnu.tgz
diff --git a/_test/tests/inc/tar/longpath-ustar.tgz b/_test/tests/inc/tar/longpath-ustar.tgz
diff --git a/inc/Tar.class.php b/inc/Tar.class.php
@@ -204,7 +204,7 @@ function extract($outdir, $strip='', $exclude='', $include='') {
 
                 // is this a file?
                 if(!$header['typeflag']){
-                    $fp = @fopen($output, "wb");
+                    $fp = fopen($output, "wb");
                     if(!$fp) throw(new TarIOException('Could not open file for writing: '.$output));
 
                     $size = floor($header['size'] / 512);
@@ -508,7 +508,6 @@ protected function writeFileHeader($name, $uid, $gid, $perm, $size, $mtime, $isd
     /**
      * Decode the given tar file header
      *
-     * @todo how to handle filenames >100 chars?
      * @param string $block a 512 byte block containign the header data
      * @return array|bool
      */
@@ -521,22 +520,36 @@ protected function parseHeader($block) {
         for($i = 156, $chks += 256; $i < 512; $i++)
             $chks += ord($block[$i]);
 
-        $headers = @unpack("a100filename/a8perm/a8uid/a8gid/a12size/a12mtime/a8checksum/a1typeflag/a100link/a6magic/a2version/a32uname/a32gname/a8devmajor/a8devminor", $block);
-        if(!$headers) return false;
+        $header = @unpack("a100filename/a8perm/a8uid/a8gid/a12size/a12mtime/a8checksum/a1typeflag/a100link/a6magic/a2version/a32uname/a32gname/a8devmajor/a8devminor/a155prefix", $block);
+        if(!$header) return false;
 
-        $return['checksum'] = OctDec(trim($headers['checksum']));
+        $return['checksum'] = OctDec(trim($header['checksum']));
         if($return['checksum'] != $chks) return false;
 
-        $return['filename'] = trim($headers['filename']);
-        $return['perm']     = OctDec(trim($headers['perm']));
-        $return['uid']      = OctDec(trim($headers['uid']));
-        $return['gid']      = OctDec(trim($headers['gid']));
-        $return['size']     = OctDec(trim($headers['size']));
-        $return['mtime']    = OctDec(trim($headers['mtime']));
-        $return['typeflag'] = $headers['typeflag'];
-        $return['link']     = trim($headers['link']);
-        $return['uname']    = trim($headers['uname']);
-        $return['gname']    = trim($headers['gname']);
+        $return['filename'] = trim($header['filename']);
+        $return['perm']     = OctDec(trim($header['perm']));
+        $return['uid']      = OctDec(trim($header['uid']));
+        $return['gid']      = OctDec(trim($header['gid']));
+        $return['size']     = OctDec(trim($header['size']));
+        $return['mtime']    = OctDec(trim($header['mtime']));
+        $return['typeflag'] = $header['typeflag'];
+        $return['link']     = trim($header['link']);
+        $return['uname']    = trim($header['uname']);
+        $return['gname']    = trim($header['gname']);
+
+        // Handle ustar Posix compliant path prefixes
+        if(trim($header['prefix'])) $return['filename'] = trim($header['prefix']).'/'.$return['filename'];
+
+        // Handle Long-Link entries from GNU Tar
+        if($return['typeflag'] == 'L'){
+            // following data block(s) is the filename
+            $filename = trim($this->readbytes(ceil($header['size'] / 512) * 512));
+            // next block is the real header
+            $block = $this->readbytes(512);
+            $return = $this->parseHeader($block);
+            // overwrite the filename
+            $return['filename'] = $filename;
+        }
 
         return $return;
     }