Fix status/execution bugs in flex-multisig

[uint]

Closes #630, ports tests/fixes from https://github.com/confio/poe-contracts.

confio/poe-contracts#12 was not an issue here, but I added a test for it.

confio/poe-contracts#14 is indeed a bug here, fixing in this PR.

confio/poe-contracts#16 was a false alarm in the first place.

[uint]

The failing test exposes the confio/poe-contracts#14 bug: https://app.circleci.com/pipelines/github/CosmWasm/cw-plus/2545/workflows/386501ae-a6f3-4174-a1a2-46ff9ee1d1e9/jobs/57318

Now for the fix...

[maurolacy]

Looks good.

Same comments than in confio/poe-contracts#60 apply:

• Why not use update_status() in execute (mutable) calls everywhere, and leave current_status() just for queries?

Perhaps take the opportunity and fix the other issues with state here as well. Or, in a follow-up. Your call.

[uint]

Looks good.

Same comments than in confio/poe-contracts#60 apply:

* Why not use `update_status()` in execute (mutable) calls everywhere, and leave `current_status()` just for queries?

As in here? https://github.com/CosmWasm/cw-plus/pull/643/files#diff-f395b2761fa612ba44398ada5f0a4d9a9fc7333ce4565281fafd99329c4dcb23R200-R206

I don't think there's a point using update_status() if it's changed to Executed in the next line anyway. But we could add something like mark_executed() to the Proposal type if you think that's clearer.

Perhaps take the opportunity and fix the other issues with state here as well. Or, in a follow-up. Your call.

Nah, I'd rather see another PR for that.

[maurolacy]

I don't think there's a point using update_status() if it's changed to Executed in the next line anyway.

I agree. In fact, update_status makes sense only in a limited number of cases, as the status changes will be discarded in case of error.

But we could add something like mark_executed() to the Proposal type if you think that's clearer.

I think that will be good, but that it will better to do it in (even) another iteration.

Perhaps take the opportunity and fix the other issues with state here as well. Or, in a follow-up. Your call.

Nah, I'd rather see another PR for that.

OK.