sources

package
v0.0.0-...-23d24bf Latest Latest
Warning

This package is not in the latest version of its module.

 Go to latest Published: Feb 6, 2024 License: MIT, MIT Imports: 27 Imported by: 0

Documentation

Overview

Package sources exports a number of clients used to access ethereum chain data.

There are a number of these exported clients used by the pt-node: L1Client wraps an RPC client to retrieve L1 ethereum data. L2Client wraps an RPC client to retrieve L2 ethereum data. RollupClient wraps an RPC client to retrieve rollup data. EngineClient extends the L2Client providing engine API bindings.

Internally, the listed clients wrap an EthClient which itself wraps a specified RPC client.

Index

Constants

This section is empty.

Variables

View Source 
var ErrNoUnsafeL2PayloadChannel = errors.New("unsafeL2Payloads channel must not be nil")
View Source 
var RPCProviderKinds = []RPCProviderKind{
	RPCKindAlchemy,
	RPCKindQuickNode,
	RPCKindInfura,
	RPCKindParity,
	RPCKindNethermind,
	RPCKindDebugGeth,
	RPCKindErigon,
	RPCKindBasic,
	RPCKindAny,
}
View Source 
var RpcSyncPeer peer.ID = "ALT_RPC_SYNC"

RpcSyncPeer is a mock PeerID for the RPC sync client.

Functions

func LimitRPC 

func LimitRPC(c client.RPC, concurrentRequests int) client.RPC

LimitRPC limits concurrent RPC requests (excluding subscriptions) to a given number by wrapping the client with a semaphore.

func NewReceiptsFetchingJob 

func NewReceiptsFetchingJob(requester ReceiptsRequester, client rpcClient, maxBatchSize int, block eth.BlockID,
	receiptHash common.Hash, txHashes []common.Hash) *receiptsFetchingJob

func ValidRPCProviderKind 

func ValidRPCProviderKind(value RPCProviderKind) bool

Types

type BatchCallContextFn 

type BatchCallContextFn func(ctx context.Context, b []rpc.BatchElem) error

type CallContextFn 

type CallContextFn func(ctx context.Context, result any, method string, args ...any) error

type DebugClient 

type DebugClient struct {
	// contains filtered or unexported fields
}

func NewDebugClient 

func NewDebugClient(callContext CallContextFn) *DebugClient

func (*DebugClient) CodeByHash 

func (o *DebugClient) CodeByHash(ctx context.Context, hash common.Hash) ([]byte, error)

func (*DebugClient) NodeByHash 

func (o *DebugClient) NodeByHash(ctx context.Context, hash common.Hash) ([]byte, error)

type EngineClient 

type EngineClient struct {
	*L2Client
}

EngineClient extends L2Client with engine API bindings.

func NewEngineClient 

func NewEngineClient(client client.RPC, log log.Logger, metrics caching.Metrics, config *EngineClientConfig) (*EngineClient, error)

func (*EngineClient) ForkchoiceUpdate 

func (s *EngineClient) ForkchoiceUpdate(ctx context.Context, fc *eth.ForkchoiceState, attributes *eth.PayloadAttributes) (*eth.ForkchoiceUpdatedResult, error)

ForkchoiceUpdate updates the forkchoice on the execution client. If attributes is not nil, the engine client will also begin building a block based on attributes after the new head block and return the payload ID.

The RPC may return three types of errors: 1. Processing error: ForkchoiceUpdatedResult.PayloadStatusV1.ValidationError or other non-success PayloadStatusV1, 2. `error` as eth.InputError: the forkchoice state or attributes are not valid. 3. Other types of `error`: temporary RPC errors, like timeouts.

func (*EngineClient) GetPayload 

func (s *EngineClient) GetPayload(ctx context.Context, payloadId eth.PayloadID) (*eth.ExecutionPayload, error)

GetPayload gets the execution payload associated with the PayloadId. There may be two types of error: 1. `error` as eth.InputError: the payload ID may be unknown 2. Other types of `error`: temporary RPC errors, like timeouts.

func (*EngineClient) NewPayload 

func (s *EngineClient) NewPayload(ctx context.Context, payload *eth.ExecutionPayload) (*eth.PayloadStatusV1, error)

NewPayload executes a full block on the execution engine. This returns a PayloadStatusV1 which encodes any validation/processing error, and this type of error is kept separate from the returned `error` used for RPC errors, like timeouts.

type EngineClientConfig 

type EngineClientConfig struct {
	L2ClientConfig
}

func EngineClientDefaultConfig 

func EngineClientDefaultConfig(config *rollup.Config) *EngineClientConfig

type EthClient 

type EthClient struct {
	// contains filtered or unexported fields
}

EthClient retrieves ethereum data with optimized batch requests, cached results, and flag to not trust the RPC.

func NewEthClient 

func NewEthClient(client client.RPC, log log.Logger, metrics caching.Metrics, config *EthClientConfig) (*EthClient, error)

NewEthClient returns an EthClient, wrapping an RPC with bindings to fetch ethereum data with added error logging, metric tracking, and caching. The EthClient uses a LimitRPC wrapper to limit the number of concurrent RPC requests.

func (*EthClient) ChainID 

func (s *EthClient) ChainID(ctx context.Context) (*big.Int, error)

ChainID fetches the chain id of the internal RPC.

func (*EthClient) Close 

func (s *EthClient) Close()

func (*EthClient) FetchReceipts 

func (s *EthClient) FetchReceipts(ctx context.Context, blockHash common.Hash) (eth.BlockInfo, types.Receipts, error)

FetchReceipts returns a block info and all of the receipts associated with transactions in the block. It verifies the receipt hash in the block header against the receipt hash of the fetched receipts to ensure that the execution engine did not fail to return any receipts.

func (*EthClient) GetProof 

func (s *EthClient) GetProof(ctx context.Context, address common.Address, storage []common.Hash, blockTag string) (*eth.AccountResult, error)

GetProof returns an account proof result, with any optional requested storage proofs. The retrieval does sanity-check that storage proofs for the expected keys are present in the response, but does not verify the result. Call accountResult.Verify(stateRoot) to verify the result.

func (*EthClient) GetStorageAt 

func (s *EthClient) GetStorageAt(ctx context.Context, address common.Address, storageSlot common.Hash, blockTag string) (common.Hash, error)

GetStorageAt returns the storage value at the given address and storage slot, **without verifying the correctness of the result**. This should only ever be used as alternative to GetProof when the user opts in. E.g. Erigon L1 node users may have to use this, since Erigon does not support eth_getProof, see https://github.com/ledgerwatch/erigon/issues/1349

func (*EthClient) InfoAndTxsByHash 

func (s *EthClient) InfoAndTxsByHash(ctx context.Context, hash common.Hash) (eth.BlockInfo, types.Transactions, error)

func (*EthClient) InfoAndTxsByLabel 

func (s *EthClient) InfoAndTxsByLabel(ctx context.Context, label eth.BlockLabel) (eth.BlockInfo, types.Transactions, error)

func (*EthClient) InfoAndTxsByNumber 

func (s *EthClient) InfoAndTxsByNumber(ctx context.Context, number uint64) (eth.BlockInfo, types.Transactions, error)

func (*EthClient) InfoByHash 

func (s *EthClient) InfoByHash(ctx context.Context, hash common.Hash) (eth.BlockInfo, error)

func (*EthClient) InfoByLabel 

func (s *EthClient) InfoByLabel(ctx context.Context, label eth.BlockLabel) (eth.BlockInfo, error)

func (*EthClient) InfoByNumber 

func (s *EthClient) InfoByNumber(ctx context.Context, number uint64) (eth.BlockInfo, error)

func (*EthClient) OnReceiptsMethodErr 

func (s *EthClient) OnReceiptsMethodErr(m ReceiptsFetchingMethod, err error)

func (*EthClient) PayloadByHash 

func (s *EthClient) PayloadByHash(ctx context.Context, hash common.Hash) (*eth.ExecutionPayload, error)

func (*EthClient) PayloadByLabel 

func (s *EthClient) PayloadByLabel(ctx context.Context, label eth.BlockLabel) (*eth.ExecutionPayload, error)

func (*EthClient) PayloadByNumber 

func (s *EthClient) PayloadByNumber(ctx context.Context, number uint64) (*eth.ExecutionPayload, error)

func (*EthClient) PickReceiptsMethod 

func (s *EthClient) PickReceiptsMethod(txCount uint64) ReceiptsFetchingMethod

func (*EthClient) ReadStorageAt 

func (s *EthClient) ReadStorageAt(ctx context.Context, address common.Address, storageSlot common.Hash, blockHash common.Hash) (common.Hash, error)

ReadStorageAt is a convenience method to read a single storage value at the given slot in the given account. The storage slot value is verified against the state-root of the given block if we do not trust the RPC provider, or directly retrieved without proof if we do trust the RPC.

func (*EthClient) SubscribeNewHead 

func (s *EthClient) SubscribeNewHead(ctx context.Context, ch chan<- *types.Header) (ethereum.Subscription, error)

SubscribeNewHead subscribes to notifications about the current blockchain head on the given channel.

type EthClientConfig 

type EthClientConfig struct {
	// Maximum number of requests to make per batch
	MaxRequestsPerBatch int

	// limit concurrent requests, applies to the source as a whole
	MaxConcurrentRequests int

	// Number of blocks worth of receipts to cache
	ReceiptsCacheSize int
	// Number of blocks worth of transactions to cache
	TransactionsCacheSize int
	// Number of block headers to cache
	HeadersCacheSize int
	// Number of payloads to cache
	PayloadsCacheSize int

	// If the RPC is untrusted, then we should not use cached information from responses,
	// and instead verify against the block-hash.
	// Of real L1 blocks no deposits can be missed/faked, no batches can be missed/faked,
	// only the wrong L1 blocks can be retrieved.
	TrustRPC bool

	// If the RPC must ensure that the results fit the ExecutionPayload(Header) format.
	// If this is not checked, disabled header fields like the nonce or difficulty
	// may be used to get a different block-hash.
	MustBePostMerge bool

	// RPCProviderKind is a hint at what type of RPC provider we are dealing with
	RPCProviderKind RPCProviderKind

	// Method reset duration defines how long we stick to available RPC methods,
	// till we re-attempt the user-preferred methods.
	// If this is 0 then the client does not fall back to less optimal but available methods.
	MethodResetDuration time.Duration
}

func (*EthClientConfig) Check 

func (c *EthClientConfig) Check() error

type IterativeBatchCall 

type IterativeBatchCall[K any, V any] struct {
	// contains filtered or unexported fields
}

IterativeBatchCall batches many RPC requests with safe and easy parallelization. Request errors are handled and re-tried, and the batch size is configurable. Executing IterativeBatchCall is as simple as calling Fetch repeatedly until it returns io.EOF.

func NewIterativeBatchCall 

func NewIterativeBatchCall[K any, V any](
	requestsKeys []K,
	makeRequest func(K) (V, rpc.BatchElem),
	getBatch BatchCallContextFn,
	getSingle CallContextFn,
	batchSize int) *IterativeBatchCall[K, V]

NewIterativeBatchCall constructs a batch call, fetching the values with the given keys, and transforms them into a verified final result.

func (*IterativeBatchCall[K, V]) Complete 

func (ibc *IterativeBatchCall[K, V]) Complete() bool

Complete indicates if the batch call is done.

func (*IterativeBatchCall[K, V]) Fetch 

func (ibc *IterativeBatchCall[K, V]) Fetch(ctx context.Context) error

Fetch fetches more of the data, and returns io.EOF when all data has been fetched. This method is safe to call concurrently; it will parallelize the fetching work. If no work is available, but the fetching is not done yet, then Fetch will block until the next thing can be fetched, or until the context expires.

func (*IterativeBatchCall[K, V]) Reset 

func (ibc *IterativeBatchCall[K, V]) Reset()

Reset will clear the batch call, to start fetching all contents from scratch.

func (*IterativeBatchCall[K, V]) Result 

func (ibc *IterativeBatchCall[K, V]) Result() ([]V, error)

Result returns the fetched values, checked and transformed to the final output type, if available. If the check fails, the IterativeBatchCall will Reset itself, to be ready for a re-attempt in fetching new data.

type L1Client 

type L1Client struct {
	*EthClient
	// contains filtered or unexported fields
}

L1Client provides typed bindings to retrieve L1 data from an RPC source, with optimized batch requests, cached results, and flag to not trust the RPC (i.e. to verify all returned contents against corresponding block hashes).

func NewL1Client 

func NewL1Client(client client.RPC, log log.Logger, metrics caching.Metrics, config *L1ClientConfig) (*L1Client, error)

NewL1Client wraps a RPC with bindings to fetch L1 data, while logging errors, tracking metrics (optional), and caching.

func (*L1Client) L1BlockRefByHash 

func (s *L1Client) L1BlockRefByHash(ctx context.Context, hash common.Hash) (eth.L1BlockRef, error)

L1BlockRefByHash returns the eth.L1BlockRef for the given block hash. We cache the block reference by hash as it is safe to assume collision will not occur.

func (*L1Client) L1BlockRefByLabel 

func (s *L1Client) L1BlockRefByLabel(ctx context.Context, label eth.BlockLabel) (eth.L1BlockRef, error)

L1BlockRefByLabel returns the eth.L1BlockRef for the given block label. Notice, we cannot cache a block reference by label because labels are not guaranteed to be unique.

func (*L1Client) L1BlockRefByNumber 

func (s *L1Client) L1BlockRefByNumber(ctx context.Context, num uint64) (eth.L1BlockRef, error)

L1BlockRefByNumber returns an eth.L1BlockRef for the given block number. Notice, we cannot cache a block reference by number because L1 re-orgs can invalidate the cached block reference.

type L1ClientConfig 

type L1ClientConfig struct {
	EthClientConfig

	L1BlockRefsCacheSize int
}

func L1ClientDefaultConfig 

func L1ClientDefaultConfig(config *rollup.Config, trustRPC bool, kind RPCProviderKind) *L1ClientConfig

type L2Client 

type L2Client struct {
	*EthClient
	// contains filtered or unexported fields
}

L2Client extends EthClient with functions to fetch and cache eth.L2BlockRef values.

func NewL2Client 

func NewL2Client(client client.RPC, log log.Logger, metrics caching.Metrics, config *L2ClientConfig) (*L2Client, error)

NewL2Client constructs a new L2Client instance. The L2Client is a thin wrapper around the EthClient with added functions for fetching and caching eth.L2BlockRef values. This includes fetching an L2BlockRef by block number, label, or hash. See: [L2BlockRefByLabel], [L2BlockRefByNumber], [L2BlockRefByHash]

func (*L2Client) L2BlockRefByHash 

func (s *L2Client) L2BlockRefByHash(ctx context.Context, hash common.Hash) (eth.L2BlockRef, error)

L2BlockRefByHash returns the eth.L2BlockRef for the given block hash. The returned BlockRef may not be in the canonical chain.

func (*L2Client) L2BlockRefByLabel 

func (s *L2Client) L2BlockRefByLabel(ctx context.Context, label eth.BlockLabel) (eth.L2BlockRef, error)

L2BlockRefByLabel returns the eth.L2BlockRef for the given block label.

func (*L2Client) L2BlockRefByNumber 

func (s *L2Client) L2BlockRefByNumber(ctx context.Context, num uint64) (eth.L2BlockRef, error)

L2BlockRefByNumber returns the eth.L2BlockRef for the given block number.

func (*L2Client) SystemConfigByL2Hash 

func (s *L2Client) SystemConfigByL2Hash(ctx context.Context, hash common.Hash) (eth.SystemConfig, error)

SystemConfigByL2Hash returns the eth.SystemConfig (matching the config updates up to and including the L1 origin) for the given L2 block hash. The returned eth.SystemConfig may not be in the canonical chain when the hash is not canonical.

type L2ClientConfig 

type L2ClientConfig struct {
	EthClientConfig

	L2BlockRefsCacheSize int
	L1ConfigsCacheSize   int

	RollupCfg *rollup.Config
}

func L2ClientDefaultConfig 

func L2ClientDefaultConfig(config *rollup.Config, trustRPC bool) *L2ClientConfig

type RPCProviderKind 

type RPCProviderKind string

RPCProviderKind identifies an RPC provider, used to hint at the optimal receipt fetching approach. 

const (
	RPCKindAlchemy    RPCProviderKind = "alchemy"
	RPCKindQuickNode  RPCProviderKind = "quicknode"
	RPCKindInfura     RPCProviderKind = "infura"
	RPCKindParity     RPCProviderKind = "parity"
	RPCKindNethermind RPCProviderKind = "nethermind"
	RPCKindDebugGeth  RPCProviderKind = "debug_geth"
	RPCKindErigon     RPCProviderKind = "erigon"
	RPCKindBasic      RPCProviderKind = "basic" // try only the standard most basic receipt fetching
	RPCKindAny        RPCProviderKind = "any"   // try any method available
)

func (*RPCProviderKind) Set 

func (kind *RPCProviderKind) Set(value string) error

func (RPCProviderKind) String 

func (kind RPCProviderKind) String() string

type RPCSync 

type RPCSync interface {
	io.Closer
	// Start starts an additional worker syncing job
	Start() error
	// RequestL2Range signals that the given range should be fetched, implementing the alt-sync interface.
	RequestL2Range(ctx context.Context, start uint64, end eth.L2BlockRef) error
}

type ReceiptsFetchingMethod 

type ReceiptsFetchingMethod uint64

ReceiptsFetchingMethod is a bitfield with 1 bit for each receipts fetching type. Depending on errors, tx counts and preferences the code may select different sets of fetching methods. 

const (
	// EthGetTransactionReceiptBatch is standard per-tx receipt fetching with JSON-RPC batches.
	// Available in: standard, everywhere.
	//   - Alchemy: 15 CU / tx
	//   - Quicknode: 2 credits / tx
	// Method: eth_getTransactionReceipt
	// See: https://ethereum.github.io/execution-apis/api-documentation/
	EthGetTransactionReceiptBatch ReceiptsFetchingMethod = 1 << iota
	// AlchemyGetTransactionReceipts is a special receipt fetching method provided by Alchemy.
	// Available in:
	//   - Alchemy: 250 CU total
	// Method: alchemy_getTransactionReceipts
	// Params:
	//   - object with "blockNumber" or "blockHash" field
	// Returns: "array of receipts" - docs lie, array is wrapped in a struct with single "receipts" field
	// See: https://docs.alchemy.com/reference/alchemy-gettransactionreceipts#alchemy_gettransactionreceipts
	AlchemyGetTransactionReceipts
	// DebugGetRawReceipts is a debug method from Geth, faster by avoiding serialization and metadata overhead.
	// Ideal for fast syncing from a local geth node.
	// Available in:
	//   - Geth: free
	//   - QuickNode: 22 credits maybe? Unknown price, undocumented ("debug_getblockreceipts" exists in table though?)
	// Method: debug_getRawReceipts
	// Params:
	//   - string presenting a block number or hash
	// Returns: list of strings, hex encoded RLP of receipts data. "consensus-encoding of all receipts in a single block"
	// See: https://geth.ethereum.org/docs/rpc/ns-debug#debug_getrawreceipts
	DebugGetRawReceipts
	// ParityGetBlockReceipts is an old parity method, which has been adopted by Nethermind and some RPC providers.
	// Available in:
	//   - Alchemy: 500 CU total
	//   - QuickNode: 59 credits - docs are wrong, not actually available anymore.
	//   - Any open-ethereum/parity legacy: free
	//   - Nethermind: free
	// Method: parity_getBlockReceipts
	// Params:
	//   Parity: "quantity or tag"
	//   Alchemy: string with block hash, number in hex, or block tag.
	//   Nethermind: very flexible: tag, number, hex or object with "requireCanonical"/"blockHash" fields.
	// Returns: array of receipts
	// See:
	//   - Parity: https://openethereum.github.io/JSONRPC-parity-module#parity_getblockreceipts
	//   - QuickNode: undocumented.
	//   - Alchemy: https://docs.alchemy.com/reference/eth-getblockreceipts
	//   - Nethermind: https://docs.nethermind.io/nethermind/ethereum-client/json-rpc/parity#parity_getblockreceipts
	ParityGetBlockReceipts
	// EthGetBlockReceipts is a non-standard receipt fetching method in the eth namespace,
	// supported by some RPC platforms and Erigon.
	// Available in:
	//   - Alchemy: 500 CU total  (and deprecated)
	//   - Erigon: free
	//   - QuickNode: 59 credits total       (does not seem to work with block hash arg, inaccurate docs)
	// Method: eth_getBlockReceipts
	// Params:
	//   - QuickNode: string, "quantity or tag", docs say incl. block hash, but API does not actually accept it.
	//   - Alchemy: string, block hash / num (hex) / block tag
	// Returns: array of receipts
	// See:
	//   - QuickNode: https://www.quicknode.com/docs/ethereum/eth_getBlockReceipts
	//   - Alchemy: https://docs.alchemy.com/reference/eth-getblockreceipts
	EthGetBlockReceipts
)

func AvailableReceiptsFetchingMethods 

func AvailableReceiptsFetchingMethods(kind RPCProviderKind) ReceiptsFetchingMethod

AvailableReceiptsFetchingMethods selects receipt fetching methods based on the RPC provider kind.

func PickBestReceiptsFetchingMethod 

func PickBestReceiptsFetchingMethod(kind RPCProviderKind, available ReceiptsFetchingMethod, txCount uint64) ReceiptsFetchingMethod

PickBestReceiptsFetchingMethod selects an RPC method that is still available, and optimal for fetching the given number of tx receipts from the specified provider kind.

func (ReceiptsFetchingMethod) String 

func (r ReceiptsFetchingMethod) String() string

type ReceiptsRequester 

type ReceiptsRequester interface {
	PickReceiptsMethod(txCount uint64) ReceiptsFetchingMethod
	OnReceiptsMethodErr(m ReceiptsFetchingMethod, err error)
}

ReceiptsRequester helps determine which receipts fetching method can be used, and is given feedback upon receipt fetching errors to adapt the choice of method.

type RollupClient 

type RollupClient struct {
	// contains filtered or unexported fields
}

func NewRollupClient 

func NewRollupClient(rpc client.RPC) *RollupClient

func (*RollupClient) OutputAtBlock 

func (r *RollupClient) OutputAtBlock(ctx context.Context, blockNum uint64) (*eth.OutputResponse, error)

func (*RollupClient) RollupConfig 

func (r *RollupClient) RollupConfig(ctx context.Context) (*rollup.Config, error)

func (*RollupClient) SyncStatus 

func (r *RollupClient) SyncStatus(ctx context.Context) (*eth.SyncStatus, error)

func (*RollupClient) Version 

func (r *RollupClient) Version(ctx context.Context) (string, error)

type SyncClient 

type SyncClient struct {
	*L2Client
	// contains filtered or unexported fields
}

SyncClient implements the driver AltSync interface, including support for fetching an open-ended chain of L2 blocks.

func NewSyncClient 

func NewSyncClient(receiver receivePayload, client client.RPC, log log.Logger, metrics caching.Metrics, config *SyncClientConfig) (*SyncClient, error)

func (*SyncClient) Close 

func (s *SyncClient) Close() error

Close sends a signal to close all concurrent syncing work.

func (*SyncClient) RequestL2Range 

func (s *SyncClient) RequestL2Range(ctx context.Context, start, end eth.L2BlockRef) error

func (*SyncClient) Start 

func (s *SyncClient) Start() error

Start starts the syncing background work. This may not be called after Close().

type SyncClientConfig 

type SyncClientConfig struct {
	L2ClientConfig
}

func SyncClientDefaultConfig 

func SyncClientDefaultConfig(config *rollup.Config, trustRPC bool) *SyncClientConfig

Source Files

View all Source files

Directories

Path Synopsis

Jump to

Keyboard shortcuts

? : This menu
/ : Search site
f or F : Jump to
y or Y : Canonical URL
go.dev uses cookies from Google to deliver and enhance the quality of its services and to analyze traffic. Learn more.