PoX-4 revoke-delegate-stx calls should fail if no state is changed

[zone117x]

There's a discussion in PR #4094 around the behavior of revoke-delegate-stx. Right now the function always returns (ok) even if no state was changed:

stacks-core/stackslib/src/chainstate/stacks/boot/pox-4.clar

Lines 636 to 641 in 2c1c777

	(define-public (revoke-delegate-stx)
	(begin
	;; must be called directly by the tx-sender or by an allowed contract-caller
	(asserts! (check-caller-allowed)
	(err ERR_STACKING_PERMISSION_DENIED))
	(ok (map-delete delegation-state { stacker: tx-sender }))))

This means A) the new synthetic event for revoke-delegate-stx can be triggered repeatedly for no valid/useful reason, and B) the event has empty data:

stacks-core/pox-locking/src/events.rs

Lines 340 to 346 in 789c235

	"revoke-delegate-stx" => {
	format!(
	r#"
	{{
	data: {{ }}
	}}
	"#,

Ideally the event would at least include the delegate-to address so that the event can be associated with the prior delegation. And ideally the call would return an error and not emit repeated events if called repeatedly.

[friedger]

Could you clarify the use case.
Revoking implies that the user does not delegate anymore after the tx was successful.

Users have to check the stacking state nevertheless to learn about the stacking state.

[zone117x]

If an event observer receives a revoke-delegate-stx event with no data and possibly repeated, then it has to:

1. Scan through past revoke-delegate-stx and delegate-stx events to check if this one actually changed state and is not a redundant no-op event.
2. Scan through past delegation events to find the associated delegate-to address and stx amounts.

Or, if you don't handle this during event ingestion, you just push off the complexity to read queries. E.g. a query for "as a delegate, what are the last X revocations that I should act on?".

Why require event observers to handle more stacking state complexity when It seems trivial to improve this on the pox/stacks-node side?

[friedger]

Should revoke-delegate-stx fail if the delegation expired?

[friedger]

yes, it should fail.

[friedger]

fixed in #4157