Panic in MessagePool::select_messages_optimal

[hanabi1224]

Describe the bug

Logic in select_messages_optimal is quite out-of-sync with lotus. We need to revisit and update the logic to get parity.

To reproduce

1. Send the below payload to http://localhost:2345/rpc/v0

{"jsonrpc":"2.0","id":0,"method":"Filecoin.MpoolSelect","params":
[[{"/":"bafy2bzaceb6vn533luic5oo7tnvudxy2g7jusjc4m3egjb3ei3tu2lhmdlbvi"},{"/":"bafy2bzaceabqr4736hunl6wu36xs5yqwdjhwbopplwqxykf54ssgbaqaphsq4"},{"/":"bafy2bzacec7zaiykrnemfuqmed7pkiyranuhixkfbv6aaa6rvkwojya5opnuk"}],0.8]
}

2. get the below panic

thread 'tokio-runtime-worker' panicked at /home/me/git/forest/src/message_pool/msgpool/selection.rs:446:17:
attempt to subtract with overflow
stack backtrace:
note: Some details are omitted, run with `RUST_BACKTRACE=full` for a verbose backtrace.

forest code (panic in the last line):

                chains[i].merged = true;
                result.extend(chains[i].msgs.clone());
                random_count += chains[i].msgs.len();
                gas_limit -= chain_gas_limit;

Log output

Log Output

paste log output...

Expected behaviour

Screenshots

Environment (please complete the following information):

• OS:
• Branch/commit
• Hardware

Other information and links