[PATCH v3] friendlier handling of overflows in archive-tar

This is a replacement for the jk/big-and-future-archive-tar topic. It's sufficiently rewritten that I won't bother with an interdiff, but rather just describe the changes below. Thanks to René for talking through some of the issues, and to Robin for suggesting the double-compression trick for the tests.

The important changes here are:

  - the "size" fix is more localized to write_tar_entry(). This results
    in shorter code, and means that if one were to somehow write an
    extended header larger than 8GB, we would notice and die(), rather
    than just writing a "0" (though such a thing should not be
    possible).

  - the mtime is now written only in a single global header, rather than
    once per file. I checked POSIX and this is explicitly allowed, and
    of course I also checked that it works against GNU tar.

  - there are now tests! I tried hard to make them portable and
    efficient. See the first patch for details.

  - the final patch is just a cleanup I noticed in the area; it could be
    done independently, but I floated it to the end in case it's
    controversial.

The patches are:

  [1/4]: t5000: test tar files that overflow ustar headers
  [2/4]: archive-tar: write extended headers for file sizes >= 8GB
  [3/4]: archive-tar: write extended headers for far-future mtime
  [4/4]: archive-tar: drop return value

-Peff