Possibility of auto unbonding / redelegation completion

[cwgoes]

Otherwise the UX is terrible - ref luniehq/lunie#808 (comment) - we previously elected not to do this due to concerns about iteration in BeginBlock/EndBlock functions, but we are iterating already in slashing / governance (vote counting) - maybe we can think of a safe way to do this, charging extra gas in the first transaction.

[faboweb]

I like the idea of extra gas cost for auto-termination.

[rigelrozanski]

Yeah I mean I don't think it really matters if we require a second transaction for unbonding so long as we are not depending on the that second transaction for redelegation

[faboweb]

Sorry for not getting it, but could you quickly explain why there is difference in the two?

[rigelrozanski]

As we've discussed previously - a person is explicitly incentivized to complete their unbonding transaction, because that's when they get the money - they are not incentived to complete their redelegation transaction because it doesn't really provide them any new explicit benefit

[faboweb]

Summing up:
We could add an option to autoterminate unbonding transactions at the cost of additional gas.
This should not be implemented for redelegation, as there it is not necessary.
Correct?

[rigelrozanski]

Yeah I mean it can be done, makes things more complicated on the back end - not sure if this is a problem from a "future-pricing-of-gas" issue

[cwgoes]

Per discussion in the slashing design meeting (@jaekwon) we are not going to do this, and leave it as two transactions.

There might be some lazy deletion during future iterations.

[faboweb]

Could you elaborate on the reasons why this was dropped?

[cwgoes]

Could you elaborate on the reasons why this was dropped?

That's mostly a question for @jaekwon - I think because it will simplify implementation.

[cwgoes]

Also ref #2082.

[rigelrozanski]

yeah due reasons explained in 2082 - I think we will still drop the redelegation completion transaction prelaunch.

[cwgoes]

Going to close this for now, since we aren't planning to implement it.