[NIT-2570] Avoids fetching batch in execution layer, consensus layer fills all necessary information regarding a batch to the execution layer #2377
Conversation
cla-bot
bot
added
the
s
diegoximenes
changed the title
| @@ -288,8 +288,7 @@ func (s *ExecutionEngine) resequenceReorgedMessages(messages []*arbostypes.Messa | |||
| log.Warn("skipping non-standard sequencer message found from reorg", "header", header) | |||
There was a problem hiding this comment.
I want Lee's take on this,
But I think around line 275, when resequencing delayed messages, we should check if the delayed this is a BatchPostingReport message, and if so skip it and not resequence.
There was a problem hiding this comment.
As discussed in a meeting, this behavior will not be changed in this PR.
Maybe this will be revisited in a future PR.
| @@ -544,6 +544,18 @@ func (s *TransactionStreamer) GetMessage(seqNum arbutil.MessageIndex) (*arbostyp | |||
| return nil, err | |||
| } | |||
|
|
|||
| err = message.Message.FillInBatchGasCost(func(batchNum uint64) ([]byte, error) { | |||
There was a problem hiding this comment.
We want to make absolutley sure that in every way we can get a message to consensus client, it will always have the batchGasCost Filled.
I think there are other cases we want to cover, specifically:
legacyGetDelayedMessageAndAccumulator
GetDelayedMessageAccumulatorAndParentChainBlockNumber
There was a problem hiding this comment.
Covered those other cases.
We don't have tests covering L1MessageType_BatchPostingReport today, right?
…layedMessageAccumulatorAndParentChainBlockNumber
…yed messages" This reverts commit 17d04ec.
diegoximenes
force-pushed
the
rm_fetch_batch
branch
from
2d1aee9
to
4b71218
Compare
Avoids fetching batch in execution layer, consensus layer fills all necessary information regarding a batch to the execution layer