Blobs are currently constructed in a naive manner: build a multi-proof for X number of accounts, generate Y number of transactions, serialize proof.
There is one stage that is missing though. After the transactions are randomly generated, the multi-proof should be analyzed and only accounts that are directly involved in a transaction should be included.
For example, say a multi-proof consists of 100 accounts and the blob has only 1 transaction between account A and B. The only leaf nodes in the tree should be related to accounts A and B.
Blobs are currently constructed in a naive manner: build a multi-proof for
Xnumber of accounts, generateYnumber of transactions, serialize proof.There is one stage that is missing though. After the transactions are randomly generated, the multi-proof should be analyzed and only accounts that are directly involved in a transaction should be included.
For example, say a multi-proof consists of 100 accounts and the blob has only 1 transaction between account
AandB. The only leaf nodes in the tree should be related to accountsAandB.