client: add getBlobsV1 to the client to support CL blob import #3711
Conversation
| pruneBlobsAndProofsCache() { | ||
| const pruneLength = this.blobsAndProofsByHash.size - BLOBS_AND_PROOFS_CACHE_LENGTH | ||
| let pruned = 0 | ||
| for (const versionedHash of this.blobsAndProofsByHash.keys()) { |
There was a problem hiding this comment.
Map.keys() returns in insertion order right? Can this get a comment that this deletes on insertion-order? This is not directly clear, it looks like it "randomly" deletes the order, but since keys() is sorted by insertion order this is not the case (it actually deletes based on first-seen).
There was a problem hiding this comment.
yes it deletes based on first seen
| @@ -34,6 +35,8 @@ const MIN_GAS_PRICE = BigInt(100000000) // .1 GWei | |||
| const TX_MAX_DATA_SIZE = 128 * 1024 // 128KB | |||
| const MAX_POOL_SIZE = 5000 | |||
| const MAX_TXS_PER_ACCOUNT = 100 | |||
| // keep for 1 epoch deep to handle reorgs | |||
| const BLOBS_AND_PROOFS_CACHE_LENGTH = 6 * 32 * 1 | |||
There was a problem hiding this comment.
The comment seems to imply we cache blobs based upon one epoch, but this is not the case, we cache the txpool blob pool based on 6 (max blobs per block) times 32 (epoch length) transactions. This can thus include both pending or included transactions (in the future).
Also slight nit: maybe define (if we don't have any) 32 as epoch length, and pull the 6 from max blobs per block param?
ref:
engine_getBlobsV1ethereum/execution-apis#559TODO: add to spec test