I didn't think blockchains had to be turing complete. What definition do you use that requires that? I can't see that the original bitcoin paper implies such a requirement, but maybe it can be derived somehow.

Sure it wasn’t in Satoshis paper but most of the largest block chains are now essentially virtual machines and transactions can be arbitrary code written for those VMs, where the fee is the computational cost for the nodes to run that code.

I still don't understand where this is going. If you're saying git is not a blockchain because it's not turing complete that means all blockchains must be turing complete. I don't think that is true.