TarUtils.parseName does not follow the spec

TarUtils.parseName does not stop at the first NUL byte, resulting in non-empty strings where they should be empty.

This manifests if the tar file contains a star_header struct instead of a posix_header as defined in https://www.gnu.org/software/tar/manual/html_node/Standard.html

The javadoc on parseName states "Parse an entry name from a buffer. Parsing stops when a NUL is found or the buffer length is reached."

However, the implementation starts at the end of the buffer and stops when the first non-NUL is found.

The solution is to replace:

  int len = length;
  for (; len > 0; len--) {
    if (buffer[offset + len - 1] != 0) {
      break;
    }
  }

by

  int len = 0;
  for (int i = offset; len < length && buffer[i] != 0; i++, len++);

This has been introduce in commit https://git-wip-us.apache.org/repos/asf?p=commons-compress.git;a=commitdiff;h=69ceb4e14feb6273c06c1e35ba116b6783bb3278