Debt Auctions Can Be Blocked about dss

[gbalabasquer]

It doesn't matter the old flopper is still in use. It can be changed by the new one and the on going auctions will end without any issue.
@kmbarry1 do you want to send a PR for the flop fix or do you prefer I take care of it?

[brianmcmichael]

As this is not a critical bug I'd suggest a two-part approach.

First would be to deploy the instant access module. This solution is not the most elegant, but it can be bolted-on to the existing system without redeployment of a core contract, so it is the least disruptive.

Second, perform the modification to the Flopper contract in dss, so if another more serious bug is discovered or the system is redeployed for alternative currencies it will be included in that redeployment.

Other solutions may arise in the longer-term if we consider redesigning the auction systems.

[kmbarry1]

As far as having fixes in the system for if it gets redeployed--having Vow.heal revert if Ash != 0 is massively preferable in that case. We should probably bifurcate dss into "deployed system" and "version for clean redeployments" repos or branches.

Regarding how to fix the live system--I am not convinced that replacing at least the Flopper is an infeasible operation. Further, the instant access module is non-trivial and will require significant testing and verification to ensure that it operates safely under all conditions. That cost+risk should be weighed against the cost of replacing the Flopper (note that the Flopper fix is dead simple, almost impossible to mess up). And some thought should be given to a Vow upgrade as well, for completeness, even if just to prove that that option really is undesirable from a technical risk perspective.

[brianmcmichael]

I don't think I can argue against the Flopper upgrade as long as it is not in use at the time.

[godsflaw]

I am in support of the flopper fix. Happy to review once we have it.

[kmbarry1]

@gbalabasquer I don't care too much who does the work, so long as we are in agreement over the correct path to take. The reason I have been holding off is to see what feedback or discussion might occur. Since it seems strong support for fixing via the Flopper is emerging, I will implement it in a day or two if there are no strong objections.

[gbalabasquer]

Sounds good.

[gbalabasquer]

This was fixed in linked PR.