Fix sequential event nonce check

This sneaky inline error caused the attestation to be processed based on
the correct validator address but the event nonce lookup was processed
on the delegate address instead of the validator address.

This didn't matter in the tests becuase events are submitted in batches
and we simply never ended up in a desync situation.

We should carefully inspect the codebase for more casts like this.

module/x/peggy/keeper/attestation.go

@@ -11,17 +11,17 @@ import (
 
 // TODO-JT: carefully look at atomicity of this function
 func (k Keeper) Attest(ctx sdk.Context, claim types.EthereumClaim, anyClaim *codectypes.Any) (*types.Attestation, error) {
+	valAddr := k.GetOrchestratorValidator(ctx, claim.GetClaimer())
+	if valAddr == nil {
+		panic("Could not find ValAddr for delegate key, should be checked by now")
+	}
 	// Check that the nonce of this event is exactly one higher than the last nonce stored by this validator.
 	// We check the event nonce in processAttestation as well, but checking it here gives individual eth signers a chance to retry,
 	// and prevents validators from submitting two claims with the same nonce
-	lastEventNonce := k.GetLastEventNonceByValidator(ctx, sdk.ValAddress(claim.GetClaimer()))
+	lastEventNonce := k.GetLastEventNonceByValidator(ctx, valAddr)
 	if claim.GetEventNonce() != lastEventNonce+1 {
 		return nil, types.ErrNonContiguousEventNonce
 	}
-	valAddr := k.GetOrchestratorValidator(ctx, claim.GetClaimer())
-	if valAddr == nil {
-		panic("Could not find ValAddr for delegate key, should be checked by now")
-	}
 
 	// Tries to get an attestation with the same eventNonce and claim as the claim that was submitted.
 	att := k.GetAttestation(ctx, claim.GetEventNonce(), claim.ClaimHash())