Use offset tables for validation

[johannesvollmer]

What can be improved or is missing?

Invalid files can be detected without too much overhead (1 sort operation), which detects invalid files much (?) faster.

Implementation Approach

When sorted, the offset table will not only tell us the start of a chunk, but also the size of the chunk (except the last one). As each chunk also declares a size on its own, the size of the chunk can be double checked with the offset table.

Also, the distance between each offset table entry is at most header.max_tile_byte_size + c where c is the fixed size meta data in the chunk. This can be checked too.