Drop Pending Stakers 2 - Replace txs.ScheduledStaker with txs.Staker #2305
Conversation
| @@ -545,3 +532,34 @@ func (e *StandardTxExecutor) BaseTx(tx *txs.BaseTx) error { | |||
| avax.Produce(e.State, e.Tx.ID(), tx.Outs) | |||
| return nil | |||
| } | |||
|
|
|||
| // addStakerFromStakerTx creates the staker and adds it to state. | |||
| func (e *StandardTxExecutor) addStakerFromStakerTx(stakerTx txs.Staker) error { | |||
There was a problem hiding this comment.
this is just a scaffolding. In next PRs we'll handle here (and here alone) the creation of PostDurango stakers
| func NewCurrentStaker( | ||
| txID ids.ID, | ||
| staker txs.PostDurangoStaker, | ||
| startTime time.Time, |
There was a problem hiding this comment.
This is the main point of this PR. Upon drop of pending stakers, we'll explicitly pass chainTime here to create the staker object. In this PR we anticipate the structural change by spreading use of txs.Staker instead of txs.ScheduledStaker wherever possible
vms/platformvm/state/state.go
Outdated
| staker, err := NewCurrentStaker(txID, stakerTx, metadata.PotentialReward) | ||
| staker, err := NewCurrentStaker( | ||
| txID, | ||
| stakerTx.(txs.PostDurangoStaker), |
There was a problem hiding this comment.
this is safe because all of our StakerTxs are also PostDurangoStakerTxs and viceversa. We type assert it for each transaction.
Still ain't pretty
abi87
changed the title
abi87
changed the title
vms/platformvm/state/state.go
Outdated
| stakerTx, ok := tx.Unsigned.(txs.Staker) | ||
| if !ok { | ||
| return fmt.Errorf("expected tx type txs.Staker but got %T", tx.Unsigned) | ||
| } |
There was a problem hiding this comment.
just moved up a fewl line. This should never happen, but always better to fail fast.
abi87
force-pushed
the
post_durango_staker_txs
branch
2 times, most recently
from
5cf9fb6 to
0deb60a
Compare
vms/platformvm/state/state.go
Outdated
| stakerTx, ok := tx.Unsigned.(txs.Staker) | ||
| if !ok { | ||
| return fmt.Errorf("expected tx type txs.Staker but got %T", tx.Unsigned) | ||
| } |
There was a problem hiding this comment.
just moved up a bit. Should never happen but always better to fail fast
abi87
force-pushed
the
post_durango_staker_txs
branch
from
0deb60a to
8e8ab33
Compare
abi87
force-pushed
the
post_durango_staker_txs
branch
from
8e8ab33 to
b213740
Compare
abi87
changed the title
vms/platformvm/state/state.go
Outdated
| staker, err := NewCurrentStaker( | ||
| txID, | ||
| stakerTx, | ||
| stakerTx.(txs.ScheduledStaker).StartTime(), |
There was a problem hiding this comment.
Can we properly check for the type cast here (and below)? I know this should never be able to fail... and that it will be removed in a followup PR once the startTime is persisted... But I'd prefer to keep the merged code "safe"
Why this should be merged
We chance NewCurrentStaker signature to allow specifying a staking start time that is not necessarily that indicated into the staker transaction.
How this works
Slicing #2175 up
How this was tested
CI