refactor: load transaction accounts #2442
Conversation
jstarry
force-pushed
the
refactor/load-tx-accounts
branch
from
748ddc0 to
829368a
Compare
|
|
||
| account | ||
| }; | ||
| accumulate_and_check_loaded_account_data_size( |
There was a problem hiding this comment.
This part looks a bit dangerous to me and could be consensus breaking if there is a slight difference in which accounts count towards the loaded_accounts_bytes limit.
There was a problem hiding this comment.
Agreed that it's a sensitive part of the code that needs extra eyes. This PR allows the instructions sysvar to hit this function call now but I have set the loaded size for that account to 0 so there should be no consensus change here.
|
@2501babe this changes some things in account loading, which we'll want to take into account for the SIMD83 changes. |
|
Any other feedback or concerns here? This refactor is not too important so I don't mind abandoning it but I think it's safe and makes things a bit easier to read. |
* refactor: load transaction accounts * feedback * fix account found * fix clippy
Problem
load_transaction_accountsdoes an unnecessary clone for the fee payer account and has an unnecessary additional level of nesting for the instructions sysvar.Summary of Changes
load_transaction_accountsto first handle the fee payer account without cloningFixes #