Get the archive's record size. Because of its history, tar supports the concept of buffered IO consisting of RECORDS of BLOCKS. This allowed tar to match the IO characteristics of the physical device being used. Of course, in the C# world, this makes no sense, WITH ONE EXCEPTION - archives are expected to be properly "blocked". Thus, all of the horrible TarBuffer support boils down to simply getting the "boundaries" correct.
public int RecordSize {get;}