Remediations for EIP-1283 reentrancy bug

For any potential solution we’ll need to remember that EIP-1283 has already rolled out on a number of test networks. As such, we’ll need a way to not break them - most likely by having clients support Constantinople-with-EIP-1283 and Constantinople-as-deployed-on-Mainnet. Test nets may potentially want to have an additional fork to move from EIP-1238 to whatever winds up being deployed on Mainnet.

I suspect from an implementation perspective just dropping EIP-1283 from Constantinople is simplest - clients could then treat it as a separate hard fork which on test nets happened at the same time and never happened on MainNet. How clean that solution is will vary from client to client though. This assumes that the “fixed” EIP-1283 or whatever replaces it would then be part of a later fork which is then applied to all nets.