At the moment we do not check the file header version in the Python impl.
We need to check that and bail out in case our reader is not compatible (or no longer compatible) with a file header version.
Ideally we would
- keep the file header fixed; it's already a fixed size in the spec
- always read the file header first and check its version
- based on the version dispatch to specific readers like
v1.reader and v2.reader and so on
- if we can not handle a specific version, bail out