Update to peer_keys_db_t to match changes in eos-system-contracts pr 185 #1316
Conversation
greg7mdp
changed the title
…ntegration tests. In that case the return value is an empty vector.
libraries/chain/controller.cpp
Outdated
| const auto& gpo = db.get<global_property_object>(); | ||
|
|
||
| auto trace = push_transaction(metadata, fc::time_point::maximum(), fc::microseconds::maximum(), | ||
| gpo.configuration.min_transaction_cpu_usage, true, 0); |
There was a problem hiding this comment.
No need to use gpo.configuration.min_transaction_cpu_usage for read-only, use 0 for billed_cpu_time_us. explicit_billed_cpu_time should be false not true. I think max_transaction_time can probably be left as maximum although we might consider using a smaller value. For block_deadline, if you are going to call this from producer_plugin then I think you should go ahead and use the actual block_deadline.
There was a problem hiding this comment.
@arhag any opinion on if we should set a limit on the read-only trx? Maybe something like 10ms?
|
|
||
| trx.actions.emplace_back(std::move(act)); | ||
| trx.set_reference_block(chain_head.id()); | ||
| set_trx_expiration(trx); |
There was a problem hiding this comment.
I don't think we require tapos & expiration on read only trx? You might be able to skip these couple lines
Either way, set_trx_expiration() does a set_reference_block() anyways so seems at least that line is redundant
| using getpeerkeys_res_t = std::vector<peerkeys_t>; | ||
|
|
||
| struct peer_info_t { | ||
| uint32_t rank; // rank by `total_votes` of all producers, active or not, may not match schedule rank |
There was a problem hiding this comment.
maybe consider default initializing this to something
Resolves #1183.
Updates for changes in VaultaFoundation/system-contracts#185.