Conflux v1.1.6-fix

Bug fix

• Fix an issue that may cause the node to stop receiving network messages.

Improvement

• Optimize transaction pool to avoid replacing a transaction with a lower gas-price one after garbage collection.

Conflux v1.1.6

Improvements

• Improve transaction execution performance when a non-existent account is accessed.
• Improve txpool garbage collection behavior.
• Add txpool RPC group for querying transaction-pool-related information.
• Enable block number index by default.

Bug fixes

• Fix an issue that may prevent archive nodes to sync (#2249).
• Fix an issue that an archive may crash after several quick restarts (#2244).
• Fix duplicate log result for duplicate blockhash parameter (#2238).
• Fix an issue that block number index may be incorrect after pivot switch (#2236).

Conflux v1.1.5

Improvements

RPC Improvements

• Fix incorrect responses of cfx_getBlockRewardInfo. If the node needs to serve this RPC, it needs to clear all data and resync the chain.
• Simplify EVM revert reason in RPC responses. This may cause incompatible issues if the user was decoding the error messages manually.
• Add new RPC cfx_getBlockByBlockNumber.
• Add blockNumber to the returned JSON object in cfx_getBlockByHash, cfx_getBlockByEpochNumber, and cfx_getBlockByHashWithPivotAssumption.
• Accept pivot hash in cfx_getEpochReceipts.
• Raise error in cfx_getBlockByHashWithPivotAssumption if the provided block hash does not belong to the provided epoch.
• Fix returnData format of CallResult.
• Fix incorrect firstTxStatus response of cfx_getAccountPendingTransactions if start_nonce is provided.
• Update cfx_getCode so that it returns an empty hex 0x if the account does not exist instead of raising an error.

Configuration Improvement

• Add persist_block_number_index to persist block number indices. It allows responding to block-number-related RPC requests for old epochs.
• Add storage_max_open_mpt_count to configure maximal number of opened MPT.

Conflux v1.1.5-testnet

Improvements

RPC Improvements

• Fix incorrect responses of cfx_getBlockRewardInfo. If the node needs to serve this RPC, it needs to clear all data and resync the chain.
• Simplify EVM revert reason in RPC responses. This may cause incompatible issues if the user was decoding the error messages manually.
• Add new RPC cfx_getBlockByBlockNumber.
• Add blockNumber to the returned JSON object in cfx_getBlockByHash, cfx_getBlockByEpochNumber, and cfx_getBlockByHashWithPivotAssumption.
• Accept pivot hash in cfx_getEpochReceipts.
• Raise error in cfx_getBlockByHashWithPivotAssumption if the provided block hash does not belong to the provided epoch.
• Fix returnData format of CallResult.
• Fix incorrect firstTxStatus response of cfx_getAccountPendingTransactions if start_nonce is provided.
• Update cfx_getCode so that it returns an empty hex 0x if the account does not exist instead of raising an error.

Configuration Improvement

• Add persist_block_number_index to persist block number indices. It allows responding to block-number-related RPC requests for old epochs.
• Add storage_max_open_mpt_count to configure maximal number of opened MPT.

Conflux v1.1.4

Incompatible Changes: Some error messages stored in receipts/traces are changed. This will affect the return message format of some RPCs. Please read this for details: Conflux-Chain/CIPs#70

Improvements

RPC Improvements

• Add a new RPC cfx_getAccountPendingTransactions to get pending transactions of a given account.
• Make VM tracer records reasons for a fail execution.
• Make cfx_estimateGasAndCollateral return an error stack in case an error happens in sub-call.
• Use random miner address in virtual calls (cfx_call and cfx_estimateGasAndCollateral)
  instead of a special null address.

Configuration Improvements

• Allow setting maximum WebSocket response payload size using jsonrpc_ws_max_payload_bytes. The default is 30MB.

Bug fixes

• Fix a bug that causes repacking useless transactions.
• Fix a bug that causes the configuration sync_state_starting_epoch not effective on archive nodes.
• Fix a bug that may make cfx_getStatus return unexpected error.

Conflux v1.1.4-testnet

Incompatible Changes: Some error messages stored in receipts/traces are changed. This will affect the return message format of some RPCs. Please read this for details: Conflux-Chain/CIPs#70

Improvements

RPC Improvements

• Add a new RPC cfx_getAccountPendingTransactions to get pending transactions of a given account.
• Make VM tracer records reasons for a fail execution.
• Make cfx_estimateGasAndCollateral return an error stack in case an error happens in sub-call.
• Use random miner address in virtual calls (cfx_call and cfx_estimateGasAndCollateral)
  instead of a special null address.

Configuration Improvements

• Allow setting maximum WebSocket response payload size using jsonrpc_ws_max_payload_bytes. The default is 30MB.

Bug fixes

• Fix a bug that causes repacking useless transactions.
• Fix a bug that causes the configuration sync_state_starting_epoch not effective on archive nodes.
• Fix a bug that may make cfx_getStatus return unexpected error.

Conflux v1.1.3

Improvements

RPC Improvements

• Include blockHash, epochHash, epochNumber, transactionHash, and transactionPosition for trace RPCs.
  Note that the data format returned by trace_block is incompatible with old versions.
• Add new field offset in log filters used in cfx_getLogs.
  If specified, the response will skip the last offset logs.
  For instance, with 10 matching logs (0..9) and offset=0x1, limit=0x5, the response will contain logs 4..8.
  Note: Even if you specify offset, the corresponding logs still need to be processed by the node,
  so a filter with offset=10000, limit=10 has about the same performance as a filter with offset=0, limit=100010.
• Add a new parameter subscription_epoch to the epochs pubsub.
  The supported values are "latest_mined" (default) and "latest_state".
• Add cfx_getAccountPendingInfo to get pending transaction info for some account for better investigating pending tx problems.

Configuration Improvements

• Allow immediately packing sent transactions in dev mode by keeping dev_block_interval_ms unset.
  Note that setting dev_block_interval_ms will disable this immediate packing and generate blocks only periodically.

Performance Improvements

• Optimize the state implementation for better cache performance.

Bug fix

• Fix a bug that makes running nodes in dev mode not generate blocks automatically.

Conflux v1.1.3-testnet

Improvements

RPC Improvements

• Include blockHash, epochHash, epochNumber, transactionHash, and transactionPosition for trace RPCs.
  Note that the data format returned by trace_block is incompatible with old versions.
• Add new field offset in log filters used in cfx_getLogs.
  If specified, the response will skip the last offset logs.
  For instance, with 10 matching logs (0..9) and offset=0x1, limit=0x5, the response will contain logs 4..8.
  Note: Even if you specify offset, the corresponding logs still need to be processed by the node,
  so a filter with offset=10000, limit=10 has about the same performance as a filter with offset=0, limit=100010.
• Add a new parameter subscription_epoch to the epochs pubsub.
  The supported values are "latest_mined" (default) and "latest_state".
• Add cfx_getAccountPendingInfo to get pending transaction info for some account for better investigating pending tx problems.

Configuration Improvements

• Allow immediately packing sent transactions in dev mode by keeping dev_block_interval_ms unset.
  Note that setting dev_block_interval_ms will disable this immediate packing and generate blocks only periodically.

Performance Improvements

• Optimize the state implementation for better cache performance.

Bug fix

• Fix a bug that makes running nodes in dev mode not generate blocks automatically.

Conflux v1.1.2

IMPORTANT NOTE: The Trace data are incompatible with v1.1.1 or older versions, and the old data might be incorrect. If you were running a node with executive_trace=true, please clear the data and resync the chain.

Improvements

Configuration Improvements

• Change the default node type to full node instead of archive node. And allow setting the node type in the
  configuration file with the entry node_type.
• Add parameters to independently configure the garbage collection time of different kinds of data (like receipts,
  transactions, block traces, state, e.t.c.). Check the additional_maintained_* entries in run/tethys.toml.
• If block_db_dir or netconf_dir is not set, put the default directory in the one configured with conflux_data_dir.
  The old behavior is to be put in the hard-coded ./blockchain_data.
• Add a parameter public_rpc_apis to control the publicly available RPC interface sets.
  The access to test and debug RPCs is no longer related to mode.
• Remove the parameter enable_tracing because it has been included in the new public_rpc_apis.

RPC Improvements

• Add new local RPC cfx_getEpochReceipts to allow querying receipts based on an epoch number.
• Add new trace RPC trace_filter to allow querying traces based on epochs/types/offset.
• Add new trace RPC trace_transaction.
• Use hex encoding for the returned bytes in trace-related RPCs.
• Add new fields latestCheckpoint, latestConfirmed, and latestState in cfx_getStatus.
• Improve some RPC error reporting.

Performance Optimization

• Reduce the memory usage for maintaining more snapshots with the configuration additional_maintained_snapshot_count.

Bug Fixes

• Fix a possible OOM error when a full node is catching up.
• Fix a possible OOM error in transaction pool when an archive node is catching up.
• Return correct block_number in cfx_getStatus.
• Fix a bug that makes the configuration mining_author require extra quotes to use a CIP-37 base32 address.
• Fix a bug that the block traces may be incorrect if the pivot chain switches frequently.

Conflux v1.1.2-testnet-fix

IMPORTANT NOTE: The Trace data are incompatible with v1.1.1 or older versions, and the old data might be incorrect. If you were running a node with executive_trace=true, please clear the data and resync the chain.

Bug Fixes

• Fix a possible OOM error in transaction pool when an archive node is catching up.

Improvements

• Add new trace RPC trace_transaction.
• Use hex encoding for the returned bytes in trace-related RPCs.