Note: do not submit before updating nns/sns proto messages.

UPD: it was decided not to do any nns/sns proto message changes.

Note: do not submit before updating nns/sns proto messages.

I think we can let the NNS team take care of updating their interface if they are interested to use the new log visibility and focus only on updating the replica to use the new type (canisters can still send the old type and we know what to do with it). There's no reason to add a dependency in our process on the NNS/SNS canisters to be upgraded before we can proceed with further steps.

I had to update rs/nns/cmc/cmc.did with new type so that their test passes and after that I suppose we have to update nns proto.

Note: do not submit before updating nns/sns proto messages.

I think we can let the NNS team take care of updating their interface if they are interested to use the new log visibility and focus only on updating the replica to use the new type (canisters can still send the old type and we know what to do with it). There's no reason to add a dependency in our process on the NNS/SNS canisters to be upgraded before we can proceed with further steps.

I had to update rs/nns/cmc/cmc.did with new type so that their test passes and after that I suppose we have to update nns proto.

What if you don't make the change in their tests to use V2? I doubt that you'd need to change any interface in that case. We can leave this update to after the replica fully uses V2 (the only thing blocked is completely removing V1 which is something that can wait a bit longer).

EDIT: Wait I see it now. This direct dependency on management types is killing us. We need to fix it. I'll try to push for a resolution because we've had similar issues in the past.

LGTM modulo CMC field optionality.

I removed cmc.did change, and added a copy-type to overcome type dependency from management canister. PTAL.