lbcd/rpcserver.go

4265 lines
136 KiB
Go
Raw Normal View History

blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
// Copyright (c) 2013-2016 The btcsuite developers
2013-08-06 23:55:22 +02:00
// Use of this source code is governed by an ISC
// license that can be found in the LICENSE file.
package main
import (
"bytes"
"crypto/subtle"
"crypto/tls"
2013-10-01 22:43:45 +02:00
"encoding/base64"
"encoding/binary"
"encoding/hex"
"encoding/json"
"errors"
2013-10-01 22:43:45 +02:00
"fmt"
"io"
"io/ioutil"
2013-08-06 23:55:22 +02:00
"math/big"
"math/rand"
2013-08-06 23:55:22 +02:00
"net"
"net/http"
"os"
2013-08-06 23:55:22 +02:00
"strconv"
"strings"
2013-08-06 23:55:22 +02:00
"sync"
"sync/atomic"
"time"
2014-07-02 15:50:08 +02:00
"github.com/btcsuite/btcd/blockchain"
"github.com/btcsuite/btcd/btcec"
"github.com/btcsuite/btcd/btcjson"
"github.com/btcsuite/btcd/chaincfg"
"github.com/btcsuite/btcd/chaincfg/chainhash"
"github.com/btcsuite/btcd/database"
"github.com/btcsuite/btcd/mining"
"github.com/btcsuite/btcd/txscript"
"github.com/btcsuite/btcd/wire"
"github.com/btcsuite/btcutil"
"github.com/btcsuite/fastsha256"
"github.com/btcsuite/websocket"
2013-08-06 23:55:22 +02:00
)
const (
// rpcAuthTimeoutSeconds is the number of seconds a connection to the
// RPC server is allowed to stay open without authenticating before it
// is closed.
rpcAuthTimeoutSeconds = 10
// uint256Size is the number of bytes needed to represent an unsigned
// 256-bit integer.
uint256Size = 32
// getworkDataLen is the length of the data field of the getwork RPC.
// It consists of the serialized block header plus the internal sha256
// padding. The internal sha256 padding consists of a single 1 bit
// followed by enough zeros to pad the message out to 56 bytes followed
// by length of the message in bits encoded as a big-endian uint64
// (8 bytes). Thus, the resulting length is a multiple of the sha256
// block size (64 bytes).
getworkDataLen = (1 + ((wire.MaxBlockHeaderPayload + 8) /
fastsha256.BlockSize)) * fastsha256.BlockSize
// hash1Len is the length of the hash1 field of the getwork RPC. It
// consists of a zero hash plus the internal sha256 padding. See
// the getworkDataLen comment for details about the internal sha256
// padding format.
hash1Len = (1 + ((chainhash.HashSize + 8) / fastsha256.BlockSize)) *
fastsha256.BlockSize
// gbtNonceRange is two 32-bit big-endian hexadecimal integers which
// represent the valid ranges of nonces returned by the getblocktemplate
// RPC.
gbtNonceRange = "00000000ffffffff"
// gbtRegenerateSeconds is the number of seconds that must pass before
// a new template is generated when the previous block hash has not
// changed and there have been changes to the available transactions
// in the memory pool.
gbtRegenerateSeconds = 60
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
// maxProtocolVersion is the max protocol version the server supports.
maxProtocolVersion = 70002
)
var (
// gbtMutableFields are the manipulations the server allows to be made
// to block templates generated by the getblocktemplate RPC. It is
// declared here to avoid the overhead of creating the slice on every
// invocation for constant data.
gbtMutableFields = []string{
"time", "transactions/add", "prevblock", "coinbase/append",
}
// gbtCoinbaseAux describes additional data that miners should include
// in the coinbase signature script. It is declared here to avoid the
// overhead of creating a new object on every invocation for constant
// data.
gbtCoinbaseAux = &btcjson.GetBlockTemplateResultAux{
Flags: hex.EncodeToString(builderScript(txscript.
NewScriptBuilder().AddData([]byte(coinbaseFlags)))),
}
// gbtCapabilities describes additional capabilities returned with a
// block template generated by the getblocktemplate RPC. It is
// declared here to avoid the overhead of creating the slice on every
// invocation for constant data.
gbtCapabilities = []string{"proposal"}
)
// Errors
var (
// ErrRPCUnimplemented is an error returned to RPC clients when the
// provided command is recognized, but not implemented.
ErrRPCUnimplemented = &btcjson.RPCError{
Code: btcjson.ErrRPCUnimplemented,
Message: "Command unimplemented",
}
// ErrRPCNoWallet is an error returned to RPC clients when the provided
// command is recognized as a wallet command.
ErrRPCNoWallet = &btcjson.RPCError{
Code: btcjson.ErrRPCNoWallet,
Message: "This implementation does not implement wallet commands",
}
)
type commandHandler func(*rpcServer, interface{}, <-chan struct{}) (interface{}, error)
// rpcHandlers maps RPC command strings to appropriate handler functions.
// This is set by init because help references rpcHandlers and thus causes
// a dependency loop.
var rpcHandlers map[string]commandHandler
var rpcHandlersBeforeInit = map[string]commandHandler{
indexers: Implement optional tx/address indexes. This introduces a new indexing infrastructure for supporting optional indexes using the new database and blockchain infrastructure along with two concrete indexer implementations which provide both a transaction-by-hash and a transaction-by-address index. The new infrastructure is mostly separated into a package named indexers which is housed under the blockchain package. In order to support this, a new interface named IndexManager has been introduced in the blockchain package which provides methods to be notified when the chain has been initialized and when blocks are connected and disconnected from the main chain. A concrete implementation of an index manager is provided by the new indexers package. The new indexers package also provides a new interface named Indexer which allows the index manager to manage concrete index implementations which conform to the interface. The following is high level overview of the main index infrastructure changes: - Define a new IndexManager interface in the blockchain package and modify the package to make use of the interface when specified - Create a new indexers package - Provides an Index interface which allows concrete indexes to plugin to an index manager - Provides a concrete IndexManager implementation - Handles the lifecycle of all indexes it manages - Tracks the index tips - Handles catching up disabled indexes that have been reenabled - Handles reorgs while the index was disabled - Invokes the appropriate methods for all managed indexes to allow them to index and deindex the blocks and transactions - Implement a transaction-by-hash index - Makes use of internal block IDs to save a significant amount of space and indexing costs over the old transaction index format - Implement a transaction-by-address index - Makes use of a leveling scheme in order to provide a good tradeoff between space required and indexing costs - Supports enabling and disabling indexes at will - Support the ability to drop indexes if they are no longer desired The following is an overview of the btcd changes: - Add a new index logging subsystem - Add new options --txindex and --addrindex in order to enable the optional indexes - NOTE: The transaction index will automatically be enabled when the address index is enabled because it depends on it - Add new options --droptxindex and --dropaddrindex to allow the indexes to be removed - NOTE: The address index will also be removed when the transaction index is dropped because it depends on it - Update getrawtransactions RPC to make use of the transaction index - Reimplement the searchrawtransaction RPC that makes use of the address index - Update sample-btcd.conf to include sample usage for the new optional index flags
2016-02-19 05:51:18 +01:00
"addnode": handleAddNode,
"createrawtransaction": handleCreateRawTransaction,
"debuglevel": handleDebugLevel,
"decoderawtransaction": handleDecodeRawTransaction,
"decodescript": handleDecodeScript,
"generate": handleGenerate,
"getaddednodeinfo": handleGetAddedNodeInfo,
"getbestblock": handleGetBestBlock,
"getbestblockhash": handleGetBestBlockHash,
"getblock": handleGetBlock,
"getblockcount": handleGetBlockCount,
"getblockhash": handleGetBlockHash,
"getblockheader": handleGetBlockHeader,
"getblocktemplate": handleGetBlockTemplate,
"getconnectioncount": handleGetConnectionCount,
"getcurrentnet": handleGetCurrentNet,
"getdifficulty": handleGetDifficulty,
"getgenerate": handleGetGenerate,
"gethashespersec": handleGetHashesPerSec,
"getinfo": handleGetInfo,
"getmempoolinfo": handleGetMempoolInfo,
"getmininginfo": handleGetMiningInfo,
"getnettotals": handleGetNetTotals,
"getnetworkhashps": handleGetNetworkHashPS,
"getpeerinfo": handleGetPeerInfo,
"getrawmempool": handleGetRawMempool,
"getrawtransaction": handleGetRawTransaction,
"gettxout": handleGetTxOut,
"getwork": handleGetWork,
"help": handleHelp,
"node": handleNode,
"ping": handlePing,
"searchrawtransactions": handleSearchRawTransactions,
"sendrawtransaction": handleSendRawTransaction,
"setgenerate": handleSetGenerate,
"stop": handleStop,
"submitblock": handleSubmitBlock,
"validateaddress": handleValidateAddress,
"verifychain": handleVerifyChain,
"verifymessage": handleVerifyMessage,
}
// list of commands that we recognize, but for which btcd has no support because
// it lacks support for wallet functionality. For these commands the user
// should ask a connected instance of btcwallet.
var rpcAskWallet = map[string]struct{}{
"addmultisigaddress": {},
"backupwallet": {},
"createencryptedwallet": {},
"createmultisig": {},
"dumpprivkey": {},
"dumpwallet": {},
"encryptwallet": {},
"getaccount": {},
"getaccountaddress": {},
"getaddressesbyaccount": {},
"getbalance": {},
"getnewaddress": {},
"getrawchangeaddress": {},
"getreceivedbyaccount": {},
"getreceivedbyaddress": {},
"gettransaction": {},
"gettxoutsetinfo": {},
"getunconfirmedbalance": {},
"getwalletinfo": {},
"importprivkey": {},
"importwallet": {},
"keypoolrefill": {},
"listaccounts": {},
"listaddressgroupings": {},
"listlockunspent": {},
"listreceivedbyaccount": {},
"listreceivedbyaddress": {},
"listsinceblock": {},
"listtransactions": {},
"listunspent": {},
"lockunspent": {},
"move": {},
"sendfrom": {},
"sendmany": {},
"sendtoaddress": {},
"setaccount": {},
"settxfee": {},
"signmessage": {},
"signrawtransaction": {},
"walletlock": {},
"walletpassphrase": {},
"walletpassphrasechange": {},
}
// Commands that are currently unimplemented, but should ultimately be.
var rpcUnimplemented = map[string]struct{}{
indexers: Implement optional tx/address indexes. This introduces a new indexing infrastructure for supporting optional indexes using the new database and blockchain infrastructure along with two concrete indexer implementations which provide both a transaction-by-hash and a transaction-by-address index. The new infrastructure is mostly separated into a package named indexers which is housed under the blockchain package. In order to support this, a new interface named IndexManager has been introduced in the blockchain package which provides methods to be notified when the chain has been initialized and when blocks are connected and disconnected from the main chain. A concrete implementation of an index manager is provided by the new indexers package. The new indexers package also provides a new interface named Indexer which allows the index manager to manage concrete index implementations which conform to the interface. The following is high level overview of the main index infrastructure changes: - Define a new IndexManager interface in the blockchain package and modify the package to make use of the interface when specified - Create a new indexers package - Provides an Index interface which allows concrete indexes to plugin to an index manager - Provides a concrete IndexManager implementation - Handles the lifecycle of all indexes it manages - Tracks the index tips - Handles catching up disabled indexes that have been reenabled - Handles reorgs while the index was disabled - Invokes the appropriate methods for all managed indexes to allow them to index and deindex the blocks and transactions - Implement a transaction-by-hash index - Makes use of internal block IDs to save a significant amount of space and indexing costs over the old transaction index format - Implement a transaction-by-address index - Makes use of a leveling scheme in order to provide a good tradeoff between space required and indexing costs - Supports enabling and disabling indexes at will - Support the ability to drop indexes if they are no longer desired The following is an overview of the btcd changes: - Add a new index logging subsystem - Add new options --txindex and --addrindex in order to enable the optional indexes - NOTE: The transaction index will automatically be enabled when the address index is enabled because it depends on it - Add new options --droptxindex and --dropaddrindex to allow the indexes to be removed - NOTE: The address index will also be removed when the transaction index is dropped because it depends on it - Update getrawtransactions RPC to make use of the transaction index - Reimplement the searchrawtransaction RPC that makes use of the address index - Update sample-btcd.conf to include sample usage for the new optional index flags
2016-02-19 05:51:18 +01:00
"estimatefee": {},
"estimatepriority": {},
"getblockchaininfo": {},
"getchaintips": {},
"getnetworkinfo": {},
}
// Commands that are available to a limited user
var rpcLimited = map[string]struct{}{
// Websockets commands
"notifyblocks": {},
"notifynewtransactions": {},
"notifyreceived": {},
"notifyspent": {},
"rescan": {},
"session": {},
// Websockets AND HTTP/S commands
"help": {},
// HTTP/S-only commands
"createrawtransaction": {},
"decoderawtransaction": {},
"decodescript": {},
"getbestblock": {},
"getbestblockhash": {},
"getblock": {},
"getblockcount": {},
"getblockhash": {},
"getcurrentnet": {},
"getdifficulty": {},
"getinfo": {},
"getnettotals": {},
"getnetworkhashps": {},
"getrawmempool": {},
"getrawtransaction": {},
"gettxout": {},
"searchrawtransactions": {},
"sendrawtransaction": {},
"submitblock": {},
"validateaddress": {},
"verifymessage": {},
}
// builderScript is a convenience function which is used for hard-coded scripts
// built with the script builder. Any errors are converted to a panic since it
// is only, and must only, be used with hard-coded, and therefore, known good,
// scripts.
func builderScript(builder *txscript.ScriptBuilder) []byte {
script, err := builder.Script()
if err != nil {
panic(err)
}
return script
}
// internalRPCError is a convenience function to convert an internal error to
// an RPC error with the appropriate code set. It also logs the error to the
// RPC server subsystem since internal errors really should not occur. The
// context parameter is only used in the log message and may be empty if it's
// not needed.
func internalRPCError(errStr, context string) *btcjson.RPCError {
logStr := errStr
if context != "" {
logStr = context + ": " + errStr
}
rpcsLog.Error(logStr)
return btcjson.NewRPCError(btcjson.ErrRPCInternal.Code, errStr)
}
// rpcDecodeHexError is a convenience function for returning a nicely formatted
// RPC error which indicates the provided hex string failed to decode.
func rpcDecodeHexError(gotHex string) *btcjson.RPCError {
return btcjson.NewRPCError(btcjson.ErrRPCDecodeHexString,
fmt.Sprintf("Argument must be hexadecimal string (not %q)",
gotHex))
}
indexers: Implement optional tx/address indexes. This introduces a new indexing infrastructure for supporting optional indexes using the new database and blockchain infrastructure along with two concrete indexer implementations which provide both a transaction-by-hash and a transaction-by-address index. The new infrastructure is mostly separated into a package named indexers which is housed under the blockchain package. In order to support this, a new interface named IndexManager has been introduced in the blockchain package which provides methods to be notified when the chain has been initialized and when blocks are connected and disconnected from the main chain. A concrete implementation of an index manager is provided by the new indexers package. The new indexers package also provides a new interface named Indexer which allows the index manager to manage concrete index implementations which conform to the interface. The following is high level overview of the main index infrastructure changes: - Define a new IndexManager interface in the blockchain package and modify the package to make use of the interface when specified - Create a new indexers package - Provides an Index interface which allows concrete indexes to plugin to an index manager - Provides a concrete IndexManager implementation - Handles the lifecycle of all indexes it manages - Tracks the index tips - Handles catching up disabled indexes that have been reenabled - Handles reorgs while the index was disabled - Invokes the appropriate methods for all managed indexes to allow them to index and deindex the blocks and transactions - Implement a transaction-by-hash index - Makes use of internal block IDs to save a significant amount of space and indexing costs over the old transaction index format - Implement a transaction-by-address index - Makes use of a leveling scheme in order to provide a good tradeoff between space required and indexing costs - Supports enabling and disabling indexes at will - Support the ability to drop indexes if they are no longer desired The following is an overview of the btcd changes: - Add a new index logging subsystem - Add new options --txindex and --addrindex in order to enable the optional indexes - NOTE: The transaction index will automatically be enabled when the address index is enabled because it depends on it - Add new options --droptxindex and --dropaddrindex to allow the indexes to be removed - NOTE: The address index will also be removed when the transaction index is dropped because it depends on it - Update getrawtransactions RPC to make use of the transaction index - Reimplement the searchrawtransaction RPC that makes use of the address index - Update sample-btcd.conf to include sample usage for the new optional index flags
2016-02-19 05:51:18 +01:00
// rpcNoTxInfoError is a convenience function for returning a nicely formatted
// RPC error which indiactes there is no information available for the provided
// transaction hash.
func rpcNoTxInfoError(txHash *chainhash.Hash) *btcjson.RPCError {
indexers: Implement optional tx/address indexes. This introduces a new indexing infrastructure for supporting optional indexes using the new database and blockchain infrastructure along with two concrete indexer implementations which provide both a transaction-by-hash and a transaction-by-address index. The new infrastructure is mostly separated into a package named indexers which is housed under the blockchain package. In order to support this, a new interface named IndexManager has been introduced in the blockchain package which provides methods to be notified when the chain has been initialized and when blocks are connected and disconnected from the main chain. A concrete implementation of an index manager is provided by the new indexers package. The new indexers package also provides a new interface named Indexer which allows the index manager to manage concrete index implementations which conform to the interface. The following is high level overview of the main index infrastructure changes: - Define a new IndexManager interface in the blockchain package and modify the package to make use of the interface when specified - Create a new indexers package - Provides an Index interface which allows concrete indexes to plugin to an index manager - Provides a concrete IndexManager implementation - Handles the lifecycle of all indexes it manages - Tracks the index tips - Handles catching up disabled indexes that have been reenabled - Handles reorgs while the index was disabled - Invokes the appropriate methods for all managed indexes to allow them to index and deindex the blocks and transactions - Implement a transaction-by-hash index - Makes use of internal block IDs to save a significant amount of space and indexing costs over the old transaction index format - Implement a transaction-by-address index - Makes use of a leveling scheme in order to provide a good tradeoff between space required and indexing costs - Supports enabling and disabling indexes at will - Support the ability to drop indexes if they are no longer desired The following is an overview of the btcd changes: - Add a new index logging subsystem - Add new options --txindex and --addrindex in order to enable the optional indexes - NOTE: The transaction index will automatically be enabled when the address index is enabled because it depends on it - Add new options --droptxindex and --dropaddrindex to allow the indexes to be removed - NOTE: The address index will also be removed when the transaction index is dropped because it depends on it - Update getrawtransactions RPC to make use of the transaction index - Reimplement the searchrawtransaction RPC that makes use of the address index - Update sample-btcd.conf to include sample usage for the new optional index flags
2016-02-19 05:51:18 +01:00
return btcjson.NewRPCError(btcjson.ErrRPCNoTxInfo,
fmt.Sprintf("No information available about transaction %v",
txHash))
}
// workStateBlockInfo houses information about how to reconstruct a block given
// its template and signature script.
type workStateBlockInfo struct {
msgBlock *wire.MsgBlock
signatureScript []byte
}
// workState houses state that is used in between multiple RPC invocations to
// getwork.
type workState struct {
sync.Mutex
lastTxUpdate time.Time
lastGenerated time.Time
prevHash *chainhash.Hash
msgBlock *wire.MsgBlock
extraNonce uint64
blockInfo map[chainhash.Hash]*workStateBlockInfo
}
// newWorkState returns a new instance of a workState with all internal fields
// initialized and ready to use.
func newWorkState() *workState {
return &workState{
blockInfo: make(map[chainhash.Hash]*workStateBlockInfo),
}
}
// gbtWorkState houses state that is used in between multiple RPC invocations to
// getblocktemplate.
type gbtWorkState struct {
sync.Mutex
lastTxUpdate time.Time
lastGenerated time.Time
prevHash *chainhash.Hash
minTimestamp time.Time
template *BlockTemplate
notifyMap map[chainhash.Hash]map[int64]chan struct{}
timeSource blockchain.MedianTimeSource
}
// newGbtWorkState returns a new instance of a gbtWorkState with all internal
// fields initialized and ready to use.
func newGbtWorkState(timeSource blockchain.MedianTimeSource) *gbtWorkState {
return &gbtWorkState{
notifyMap: make(map[chainhash.Hash]map[int64]chan struct{}),
timeSource: timeSource,
}
}
// handleUnimplemented is the handler for commands that should ultimately be
// supported but are not yet implemented.
func handleUnimplemented(s *rpcServer, cmd interface{}, closeChan <-chan struct{}) (interface{}, error) {
return nil, ErrRPCUnimplemented
}
// handleAskWallet is the handler for commands that are recognized as valid, but
// are unable to answer correctly since it involves wallet state.
// These commands will be implemented in btcwallet.
func handleAskWallet(s *rpcServer, cmd interface{}, closeChan <-chan struct{}) (interface{}, error) {
return nil, ErrRPCNoWallet
}
2013-08-06 23:55:22 +02:00
// handleAddNode handles addnode commands.
func handleAddNode(s *rpcServer, cmd interface{}, closeChan <-chan struct{}) (interface{}, error) {
c := cmd.(*btcjson.AddNodeCmd)
addr := normalizeAddress(c.Addr, activeNetParams.DefaultPort)
var err error
switch c.SubCmd {
case "add":
err = s.server.ConnectNode(addr, true)
case "remove":
err = s.server.RemoveNodeByAddr(addr)
case "onetry":
err = s.server.ConnectNode(addr, false)
default:
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCInvalidParameter,
Message: "invalid subcommand for addnode",
}
}
if err != nil {
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCInvalidParameter,
Message: err.Error(),
}
}
// no data returned unless an error.
return nil, nil
}
// handleNode handles node commands.
func handleNode(s *rpcServer, cmd interface{}, closeChan <-chan struct{}) (interface{}, error) {
c := cmd.(*btcjson.NodeCmd)
var addr string
var nodeID uint64
var errN, err error
switch c.SubCmd {
case "disconnect":
// If we have a valid uint disconnect by node id. Otherwise,
// attempt to disconnect by address, returning an error if a
// valid IP address is not supplied.
if nodeID, errN = strconv.ParseUint(c.Target, 10, 32); errN == nil {
err = s.server.DisconnectNodeByID(int32(nodeID))
} else {
if _, _, errP := net.SplitHostPort(c.Target); errP == nil || net.ParseIP(c.Target) != nil {
addr = normalizeAddress(c.Target, activeNetParams.DefaultPort)
err = s.server.DisconnectNodeByAddr(addr)
} else {
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCInvalidParameter,
Message: "invalid address or node ID",
}
}
}
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
if err != nil && peerExists(s.server.Peers(), addr, int32(nodeID)) {
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCMisc,
Message: "can't disconnect a permanent peer, use remove",
}
}
case "remove":
// If we have a valid uint disconnect by node id. Otherwise,
// attempt to disconnect by address, returning an error if a
// valid IP address is not supplied.
if nodeID, errN = strconv.ParseUint(c.Target, 10, 32); errN == nil {
err = s.server.RemoveNodeByID(int32(nodeID))
} else {
if _, _, errP := net.SplitHostPort(c.Target); errP == nil || net.ParseIP(c.Target) != nil {
addr = normalizeAddress(c.Target, activeNetParams.DefaultPort)
err = s.server.RemoveNodeByAddr(addr)
} else {
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCInvalidParameter,
Message: "invalid address or node ID",
}
}
}
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
if err != nil && peerExists(s.server.Peers(), addr, int32(nodeID)) {
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCMisc,
Message: "can't remove a temporary peer, use disconnect",
}
}
case "connect":
addr = normalizeAddress(c.Target, activeNetParams.DefaultPort)
// Default to temporary connections.
subCmd := "temp"
if c.ConnectSubCmd != nil {
subCmd = *c.ConnectSubCmd
}
switch subCmd {
case "perm", "temp":
err = s.server.ConnectNode(addr, subCmd == "perm")
default:
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCInvalidParameter,
Message: "invalid subcommand for node connect",
}
}
default:
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCInvalidParameter,
Message: "invalid subcommand for node",
}
}
if err != nil {
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCInvalidParameter,
Message: err.Error(),
}
}
// no data returned unless an error.
return nil, nil
}
// peerExists determines if a certain peer is currently connected given
// information about all currently connected peers. Peer existence is
// determined using either a target address or node id.
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
func peerExists(peers []*serverPeer, addr string, nodeID int32) bool {
for _, p := range peers {
if p.ID() == nodeID || p.Addr() == addr {
return true
}
}
return false
}
// messageToHex serializes a message to the wire protocol encoding using the
// latest protocol version and returns a hex-encoded string of the result.
func messageToHex(msg wire.Message) (string, error) {
var buf bytes.Buffer
if err := msg.BtcEncode(&buf, maxProtocolVersion); err != nil {
context := fmt.Sprintf("Failed to encode msg of type %T", msg)
return "", internalRPCError(err.Error(), context)
}
return hex.EncodeToString(buf.Bytes()), nil
}
// handleCreateRawTransaction handles createrawtransaction commands.
func handleCreateRawTransaction(s *rpcServer, cmd interface{}, closeChan <-chan struct{}) (interface{}, error) {
c := cmd.(*btcjson.CreateRawTransactionCmd)
// Validate the locktime, if given.
if c.LockTime != nil &&
(*c.LockTime < 0 || *c.LockTime > int64(wire.MaxTxInSequenceNum)) {
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCInvalidParameter,
Message: "Locktime out of range",
}
}
// Add all transaction inputs to a new transaction after performing
// some validity checks.
mtx := wire.NewMsgTx()
for _, input := range c.Inputs {
txHash, err := chainhash.NewHashFromStr(input.Txid)
if err != nil {
return nil, rpcDecodeHexError(input.Txid)
}
prevOut := wire.NewOutPoint(txHash, uint32(input.Vout))
txIn := wire.NewTxIn(prevOut, []byte{})
if c.LockTime != nil && *c.LockTime != 0 {
txIn.Sequence = wire.MaxTxInSequenceNum - 1
}
mtx.AddTxIn(txIn)
}
// Add all transaction outputs to the transaction after performing
// some validity checks.
for encodedAddr, amount := range c.Amounts {
// Ensure amount is in the valid range for monetary amounts.
if amount <= 0 || amount > btcutil.MaxSatoshi {
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCType,
Message: "Invalid amount",
}
}
2013-08-06 23:55:22 +02:00
// Decode the provided address.
addr, err := btcutil.DecodeAddress(encodedAddr,
activeNetParams.Params)
if err != nil {
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCInvalidAddressOrKey,
Message: "Invalid address or key: " + err.Error(),
}
}
// Ensure the address is one of the supported types and that
// the network encoded with the address matches the network the
// server is currently on.
switch addr.(type) {
case *btcutil.AddressPubKeyHash:
case *btcutil.AddressScriptHash:
default:
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCInvalidAddressOrKey,
Message: "Invalid address or key",
}
}
if !addr.IsForNet(s.server.chainParams) {
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCInvalidAddressOrKey,
Message: "Invalid address: " + encodedAddr +
" is for the wrong network",
}
}
// Create a new script which pays to the provided address.
pkScript, err := txscript.PayToAddrScript(addr)
if err != nil {
context := "Failed to generate pay-to-address script"
return nil, internalRPCError(err.Error(), context)
}
// Convert the amount to satoshi.
satoshi, err := btcutil.NewAmount(amount)
if err != nil {
context := "Failed to convert amount"
return nil, internalRPCError(err.Error(), context)
}
txOut := wire.NewTxOut(int64(satoshi), pkScript)
mtx.AddTxOut(txOut)
}
// Set the Locktime, if given.
if c.LockTime != nil {
mtx.LockTime = uint32(*c.LockTime)
}
// Return the serialized and hex-encoded transaction. Note that this
// is intentionally not directly returning because the first return
// value is a string and it would result in returning an empty string to
// the client instead of nothing (nil) in the case of an error.
mtxHex, err := messageToHex(mtx)
if err != nil {
return nil, err
}
return mtxHex, nil
}
// handleDebugLevel handles debuglevel commands.
func handleDebugLevel(s *rpcServer, cmd interface{}, closeChan <-chan struct{}) (interface{}, error) {
c := cmd.(*btcjson.DebugLevelCmd)
// Special show command to list supported subsystems.
if c.LevelSpec == "show" {
return fmt.Sprintf("Supported subsystems %v",
supportedSubsystems()), nil
}
err := parseAndSetDebugLevels(c.LevelSpec)
if err != nil {
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCInvalidParams.Code,
Message: err.Error(),
}
}
return "Done.", nil
}
// createVinList returns a slice of JSON objects for the inputs of the passed
// transaction.
func createVinList(mtx *wire.MsgTx) []btcjson.Vin {
// Coinbase transactions only have a single txin by definition.
vinList := make([]btcjson.Vin, len(mtx.TxIn))
if blockchain.IsCoinBaseTx(mtx) {
txIn := mtx.TxIn[0]
vinList[0].Coinbase = hex.EncodeToString(txIn.SignatureScript)
vinList[0].Sequence = txIn.Sequence
return vinList
}
for i, txIn := range mtx.TxIn {
// The disassembled string will contain [error] inline
// if the script doesn't fully parse, so ignore the
// error here.
disbuf, _ := txscript.DisasmString(txIn.SignatureScript)
vinEntry := &vinList[i]
vinEntry.Txid = txIn.PreviousOutPoint.Hash.String()
vinEntry.Vout = txIn.PreviousOutPoint.Index
vinEntry.Sequence = txIn.Sequence
vinEntry.ScriptSig = &btcjson.ScriptSig{
Asm: disbuf,
Hex: hex.EncodeToString(txIn.SignatureScript),
}
}
return vinList
}
// createVoutList returns a slice of JSON objects for the outputs of the passed
// transaction.
func createVoutList(mtx *wire.MsgTx, chainParams *chaincfg.Params, filterAddrMap map[string]struct{}) []btcjson.Vout {
voutList := make([]btcjson.Vout, 0, len(mtx.TxOut))
for i, v := range mtx.TxOut {
// The disassembled string will contain [error] inline if the
// script doesn't fully parse, so ignore the error here.
disbuf, _ := txscript.DisasmString(v.PkScript)
// Ignore the error here since an error means the script
// couldn't parse and there is no additional information about
// it anyways.
scriptClass, addrs, reqSigs, _ := txscript.ExtractPkScriptAddrs(
v.PkScript, chainParams)
// Encode the addresses while checking if the address passes the
// filter when needed.
passesFilter := len(filterAddrMap) == 0
encodedAddrs := make([]string, len(addrs))
for j, addr := range addrs {
encodedAddr := addr.EncodeAddress()
encodedAddrs[j] = encodedAddr
// No need to check the map again if the filter already
// passes.
if passesFilter {
continue
}
if _, exists := filterAddrMap[encodedAddr]; exists {
passesFilter = true
}
}
if !passesFilter {
continue
}
var vout btcjson.Vout
vout.N = uint32(i)
vout.Value = btcutil.Amount(v.Value).ToBTC()
vout.ScriptPubKey.Addresses = encodedAddrs
vout.ScriptPubKey.Asm = disbuf
vout.ScriptPubKey.Hex = hex.EncodeToString(v.PkScript)
vout.ScriptPubKey.Type = scriptClass.String()
vout.ScriptPubKey.ReqSigs = int32(reqSigs)
voutList = append(voutList, vout)
2013-08-06 23:55:22 +02:00
}
return voutList
2013-08-06 23:55:22 +02:00
}
// createTxRawResult converts the passed transaction and associated parameters
// to a raw transaction JSON object.
func createTxRawResult(chainParams *chaincfg.Params, mtx *wire.MsgTx,
txHash string, blkHeader *wire.BlockHeader, blkHash string,
blkHeight int32, chainHeight int32) (*btcjson.TxRawResult, error) {
mtxHex, err := messageToHex(mtx)
if err != nil {
return nil, err
}
txReply := &btcjson.TxRawResult{
Hex: mtxHex,
Txid: txHash,
Vin: createVinList(mtx),
indexers: Implement optional tx/address indexes. This introduces a new indexing infrastructure for supporting optional indexes using the new database and blockchain infrastructure along with two concrete indexer implementations which provide both a transaction-by-hash and a transaction-by-address index. The new infrastructure is mostly separated into a package named indexers which is housed under the blockchain package. In order to support this, a new interface named IndexManager has been introduced in the blockchain package which provides methods to be notified when the chain has been initialized and when blocks are connected and disconnected from the main chain. A concrete implementation of an index manager is provided by the new indexers package. The new indexers package also provides a new interface named Indexer which allows the index manager to manage concrete index implementations which conform to the interface. The following is high level overview of the main index infrastructure changes: - Define a new IndexManager interface in the blockchain package and modify the package to make use of the interface when specified - Create a new indexers package - Provides an Index interface which allows concrete indexes to plugin to an index manager - Provides a concrete IndexManager implementation - Handles the lifecycle of all indexes it manages - Tracks the index tips - Handles catching up disabled indexes that have been reenabled - Handles reorgs while the index was disabled - Invokes the appropriate methods for all managed indexes to allow them to index and deindex the blocks and transactions - Implement a transaction-by-hash index - Makes use of internal block IDs to save a significant amount of space and indexing costs over the old transaction index format - Implement a transaction-by-address index - Makes use of a leveling scheme in order to provide a good tradeoff between space required and indexing costs - Supports enabling and disabling indexes at will - Support the ability to drop indexes if they are no longer desired The following is an overview of the btcd changes: - Add a new index logging subsystem - Add new options --txindex and --addrindex in order to enable the optional indexes - NOTE: The transaction index will automatically be enabled when the address index is enabled because it depends on it - Add new options --droptxindex and --dropaddrindex to allow the indexes to be removed - NOTE: The address index will also be removed when the transaction index is dropped because it depends on it - Update getrawtransactions RPC to make use of the transaction index - Reimplement the searchrawtransaction RPC that makes use of the address index - Update sample-btcd.conf to include sample usage for the new optional index flags
2016-02-19 05:51:18 +01:00
Vout: createVoutList(mtx, chainParams, nil),
Version: mtx.Version,
LockTime: mtx.LockTime,
}
if blkHeader != nil {
// This is not a typo, they are identical in bitcoind as well.
txReply.Time = blkHeader.Timestamp.Unix()
txReply.Blocktime = blkHeader.Timestamp.Unix()
txReply.BlockHash = blkHash
txReply.Confirmations = uint64(1 + chainHeight - blkHeight)
2013-08-06 23:55:22 +02:00
}
return txReply, nil
}
// handleDecodeRawTransaction handles decoderawtransaction commands.
func handleDecodeRawTransaction(s *rpcServer, cmd interface{}, closeChan <-chan struct{}) (interface{}, error) {
c := cmd.(*btcjson.DecodeRawTransactionCmd)
// Deserialize the transaction.
hexStr := c.HexTx
if len(hexStr)%2 != 0 {
hexStr = "0" + hexStr
}
serializedTx, err := hex.DecodeString(hexStr)
if err != nil {
return nil, rpcDecodeHexError(hexStr)
}
var mtx wire.MsgTx
err = mtx.Deserialize(bytes.NewReader(serializedTx))
if err != nil {
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCDeserialization,
Message: "TX decode failed: " + err.Error(),
}
2013-08-06 23:55:22 +02:00
}
// Create and return the result.
txReply := btcjson.TxRawDecodeResult{
Txid: mtx.TxHash().String(),
Version: mtx.Version,
Locktime: mtx.LockTime,
Vin: createVinList(&mtx),
Vout: createVoutList(&mtx, s.server.chainParams, nil),
}
return txReply, nil
2013-08-06 23:55:22 +02:00
}
// handleDecodeScript handles decodescript commands.
func handleDecodeScript(s *rpcServer, cmd interface{}, closeChan <-chan struct{}) (interface{}, error) {
c := cmd.(*btcjson.DecodeScriptCmd)
2013-10-01 22:43:45 +02:00
// Convert the hex script to bytes.
hexStr := c.HexScript
if len(hexStr)%2 != 0 {
hexStr = "0" + hexStr
2013-08-06 23:55:22 +02:00
}
script, err := hex.DecodeString(hexStr)
2013-08-06 23:55:22 +02:00
if err != nil {
return nil, rpcDecodeHexError(hexStr)
}
// The disassembled string will contain [error] inline if the script
// doesn't fully parse, so ignore the error here.
disbuf, _ := txscript.DisasmString(script)
// Get information about the script.
// Ignore the error here since an error means the script couldn't parse
// and there is no additinal information about it anyways.
scriptClass, addrs, reqSigs, _ := txscript.ExtractPkScriptAddrs(script,
s.server.chainParams)
addresses := make([]string, len(addrs))
for i, addr := range addrs {
addresses[i] = addr.EncodeAddress()
}
// Convert the script itself to a pay-to-script-hash address.
p2sh, err := btcutil.NewAddressScriptHash(script, s.server.chainParams)
if err != nil {
context := "Failed to convert script to pay-to-script-hash"
return nil, internalRPCError(err.Error(), context)
}
// Generate and return the reply.
reply := btcjson.DecodeScriptResult{
Asm: disbuf,
ReqSigs: int32(reqSigs),
Type: scriptClass.String(),
Addresses: addresses,
P2sh: p2sh.EncodeAddress(),
}
return reply, nil
}
// handleGenerate handles generate commands.
func handleGenerate(s *rpcServer, cmd interface{}, closeChan <-chan struct{}) (interface{}, error) {
// Respond with an error if there are no addresses to pay the
// created blocks to.
if len(cfg.miningAddrs) == 0 {
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCInternal.Code,
Message: "No payment addresses specified " +
"via --miningaddr",
}
}
c := cmd.(*btcjson.GenerateCmd)
// Respond with an error if the client is requesting 0 blocks to be generated.
if c.NumBlocks == 0 {
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCInternal.Code,
Message: "Please request a nonzero number of blocks to generate.",
}
}
// Create a reply
reply := make([]string, c.NumBlocks)
blockHashes, err := s.server.cpuMiner.GenerateNBlocks(c.NumBlocks)
if err != nil {
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCInternal.Code,
Message: err.Error(),
}
}
// Mine the correct number of blocks, assigning the hex representation of the
// hash of each one to its place in the reply.
for i, hash := range blockHashes {
reply[i] = hash.String()
}
return reply, nil
}
// handleGetAddedNodeInfo handles getaddednodeinfo commands.
func handleGetAddedNodeInfo(s *rpcServer, cmd interface{}, closeChan <-chan struct{}) (interface{}, error) {
c := cmd.(*btcjson.GetAddedNodeInfoCmd)
// Retrieve a list of persistent (added) peers from the bitcoin server
2016-02-22 11:35:54 +01:00
// and filter the list of peers per the specified address (if any).
peers := s.server.AddedNodeInfo()
if c.Node != nil {
node := *c.Node
found := false
for i, peer := range peers {
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
if peer.Addr() == node {
peers = peers[i : i+1]
found = true
}
}
if !found {
return nil, &btcjson.RPCError{
Code: -24, // TODO: ErrRPCClientNodeNotAdded
Message: "Node has not been added",
}
}
}
// Without the dns flag, the result is just a slice of the addresses as
// strings.
if !c.DNS {
results := make([]string, 0, len(peers))
for _, peer := range peers {
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
results = append(results, peer.Addr())
}
return results, nil
}
// With the dns flag, the result is an array of JSON objects which
// include the result of DNS lookups for each peer.
results := make([]*btcjson.GetAddedNodeInfoResult, 0, len(peers))
for _, peer := range peers {
// Set the "address" of the peer which could be an ip address
// or a domain name.
var result btcjson.GetAddedNodeInfoResult
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
result.AddedNode = peer.Addr()
result.Connected = btcjson.Bool(peer.Connected())
// Split the address into host and port portions so we can do
// a DNS lookup against the host. When no port is specified in
// the address, just use the address as the host.
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
host, _, err := net.SplitHostPort(peer.Addr())
if err != nil {
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
host = peer.Addr()
}
// Do a DNS lookup for the address. If the lookup fails, just
// use the host.
var ipList []string
ips, err := btcdLookup(host)
if err == nil {
ipList = make([]string, 0, len(ips))
for _, ip := range ips {
ipList = append(ipList, ip.String())
}
} else {
ipList = make([]string, 1)
ipList[0] = host
}
// Add the addresses and connection info to the result.
addrs := make([]btcjson.GetAddedNodeInfoResultAddr, 0, len(ipList))
for _, ip := range ipList {
var addr btcjson.GetAddedNodeInfoResultAddr
addr.Address = ip
addr.Connected = "false"
if ip == host && peer.Connected() {
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
addr.Connected = directionString(peer.Inbound())
}
addrs = append(addrs, addr)
}
result.Addresses = &addrs
results = append(results, &result)
}
return results, nil
}
// handleGetBestBlock implements the getbestblock command.
func handleGetBestBlock(s *rpcServer, cmd interface{}, closeChan <-chan struct{}) (interface{}, error) {
// All other "get block" commands give either the height, the
// hash, or both but require the block SHA. This gets both for
// the best block.
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
best := s.chain.BestSnapshot()
result := &btcjson.GetBestBlockResult{
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
Hash: best.Hash.String(),
Height: best.Height,
}
return result, nil
}
// handleGetBestBlockHash implements the getbestblockhash command.
func handleGetBestBlockHash(s *rpcServer, cmd interface{}, closeChan <-chan struct{}) (interface{}, error) {
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
best := s.chain.BestSnapshot()
return best.Hash.String(), nil
}
// getDifficultyRatio returns the proof-of-work difficulty as a multiple of the
// minimum difficulty using the passed bits field from the header of a block.
func getDifficultyRatio(bits uint32) float64 {
// The minimum difficulty is the max possible proof-of-work limit bits
2016-02-22 11:35:54 +01:00
// converted back to a number. Note this is not the same as the proof of
// work limit directly because the block difficulty is encoded in a block
// with the compact form which loses precision.
max := blockchain.CompactToBig(activeNetParams.PowLimitBits)
target := blockchain.CompactToBig(bits)
difficulty := new(big.Rat).SetFrac(max, target)
outString := difficulty.FloatString(8)
diff, err := strconv.ParseFloat(outString, 64)
if err != nil {
rpcsLog.Errorf("Cannot get difficulty: %v", err)
return 0
}
return diff
}
// handleGetBlock implements the getblock command.
func handleGetBlock(s *rpcServer, cmd interface{}, closeChan <-chan struct{}) (interface{}, error) {
c := cmd.(*btcjson.GetBlockCmd)
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
// Load the raw block bytes from the database.
hash, err := chainhash.NewHashFromStr(c.Hash)
if err != nil {
return nil, rpcDecodeHexError(c.Hash)
}
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
var blkBytes []byte
err = s.server.db.View(func(dbTx database.Tx) error {
var err error
blkBytes, err = dbTx.FetchBlock(hash)
return err
})
if err != nil {
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCBlockNotFound,
Message: "Block not found",
}
}
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
// When the verbose flag isn't set, simply return the serialized block
// as a hex-encoded string.
if c.Verbose != nil && !*c.Verbose {
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
return hex.EncodeToString(blkBytes), nil
}
// The verbose flag is set, so generate the JSON object and return it.
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
// Deserialize the block.
blk, err := btcutil.NewBlockFromBytes(blkBytes)
if err != nil {
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
context := "Failed to deserialize block"
return nil, internalRPCError(err.Error(), context)
}
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
// Get the block height from chain.
blockHeight, err := s.chain.BlockHeightByHash(hash)
if err != nil {
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
context := "Failed to obtain block height"
return nil, internalRPCError(err.Error(), context)
}
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
blk.SetHeight(blockHeight)
best := s.chain.BestSnapshot()
// Get next block hash unless there are none.
var nextHashString string
if blockHeight < best.Height {
nextHash, err := s.chain.BlockHashByHeight(blockHeight + 1)
if err != nil {
context := "No next block"
return nil, internalRPCError(err.Error(), context)
}
nextHashString = nextHash.String()
}
blockHeader := &blk.MsgBlock().Header
blockReply := btcjson.GetBlockVerboseResult{
Hash: c.Hash,
Version: blockHeader.Version,
MerkleRoot: blockHeader.MerkleRoot.String(),
PreviousHash: blockHeader.PrevBlock.String(),
Nonce: blockHeader.Nonce,
Time: blockHeader.Timestamp.Unix(),
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
Confirmations: uint64(1 + best.Height - blockHeight),
Height: int64(blockHeight),
Size: int32(len(blkBytes)),
Bits: strconv.FormatInt(int64(blockHeader.Bits), 16),
Difficulty: getDifficultyRatio(blockHeader.Bits),
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
NextHash: nextHashString,
}
2013-08-06 23:55:22 +02:00
if c.VerboseTx == nil || !*c.VerboseTx {
transactions := blk.Transactions()
txNames := make([]string, len(transactions))
for i, tx := range transactions {
txNames[i] = tx.Hash().String()
}
blockReply.Tx = txNames
} else {
txns := blk.Transactions()
rawTxns := make([]btcjson.TxRawResult, len(txns))
for i, tx := range txns {
rawTxn, err := createTxRawResult(s.server.chainParams,
tx.MsgTx(), tx.Hash().String(), blockHeader,
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
hash.String(), blockHeight, best.Height)
if err != nil {
return nil, err
}
rawTxns[i] = *rawTxn
}
blockReply.RawTx = rawTxns
}
return blockReply, nil
}
2013-08-06 23:55:22 +02:00
// handleGetBlockCount implements the getblockcount command.
func handleGetBlockCount(s *rpcServer, cmd interface{}, closeChan <-chan struct{}) (interface{}, error) {
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
best := s.chain.BestSnapshot()
return int64(best.Height), nil
}
2013-08-06 23:55:22 +02:00
// handleGetBlockHash implements the getblockhash command.
func handleGetBlockHash(s *rpcServer, cmd interface{}, closeChan <-chan struct{}) (interface{}, error) {
c := cmd.(*btcjson.GetBlockHashCmd)
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
hash, err := s.chain.BlockHashByHeight(int32(c.Index))
if err != nil {
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCOutOfRange,
Message: "Block number out of range",
}
}
2013-08-06 23:55:22 +02:00
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
return hash.String(), nil
}
2013-08-06 23:55:22 +02:00
// handleGetBlockHeader implements the getblockheader command.
func handleGetBlockHeader(s *rpcServer, cmd interface{}, closeChan <-chan struct{}) (interface{}, error) {
c := cmd.(*btcjson.GetBlockHeaderCmd)
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
// Load the raw header bytes from the database.
hash, err := chainhash.NewHashFromStr(c.Hash)
if err != nil {
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
return nil, rpcDecodeHexError(c.Hash)
}
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
var headerBytes []byte
err = s.server.db.View(func(dbTx database.Tx) error {
var err error
headerBytes, err = dbTx.FetchBlockHeader(hash)
return err
})
if err != nil {
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCBlockNotFound,
Message: "Block not found",
}
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
}
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
// When the verbose flag isn't set, simply return the serialized block
// header as a hex-encoded string.
if c.Verbose != nil && !*c.Verbose {
return hex.EncodeToString(headerBytes), nil
}
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
// The verbose flag is set, so generate the JSON object and return it.
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
// Deserialize the header.
var blockHeader wire.BlockHeader
err = blockHeader.Deserialize(bytes.NewReader(headerBytes))
if err != nil {
context := "Failed to deserialize block header"
return nil, internalRPCError(err.Error(), context)
}
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
// Get the block height from chain.
blockHeight, err := s.chain.BlockHeightByHash(hash)
if err != nil {
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
context := "Failed to obtain block height"
return nil, internalRPCError(err.Error(), context)
}
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
best := s.chain.BestSnapshot()
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
// Get next block hash unless there are none.
var nextHashString string
if blockHeight < best.Height {
nextHash, err := s.chain.BlockHashByHeight(blockHeight + 1)
if err != nil {
context := "No next block"
return nil, internalRPCError(err.Error(), context)
}
nextHashString = nextHash.String()
}
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
blockHeaderReply := btcjson.GetBlockHeaderVerboseResult{
Hash: c.Hash,
Confirmations: uint64(1 + best.Height - blockHeight),
Height: int32(blockHeight),
Version: blockHeader.Version,
MerkleRoot: blockHeader.MerkleRoot.String(),
NextHash: nextHashString,
PreviousHash: blockHeader.PrevBlock.String(),
Nonce: uint64(blockHeader.Nonce),
Time: blockHeader.Timestamp.Unix(),
Bits: strconv.FormatInt(int64(blockHeader.Bits), 16),
Difficulty: getDifficultyRatio(blockHeader.Bits),
}
return blockHeaderReply, nil
}
// encodeTemplateID encodes the passed details into an ID that can be used to
// uniquely identify a block template.
func encodeTemplateID(prevHash *chainhash.Hash, lastGenerated time.Time) string {
return fmt.Sprintf("%s-%d", prevHash.String(), lastGenerated.Unix())
}
// decodeTemplateID decodes an ID that is used to uniquely identify a block
// template. This is mainly used as a mechanism to track when to update clients
// that are using long polling for block templates. The ID consists of the
// previous block hash for the associated template and the time the associated
// template was generated.
func decodeTemplateID(templateID string) (*chainhash.Hash, int64, error) {
fields := strings.Split(templateID, "-")
if len(fields) != 2 {
return nil, 0, errors.New("invalid longpollid format")
}
prevHash, err := chainhash.NewHashFromStr(fields[0])
if err != nil {
return nil, 0, errors.New("invalid longpollid format")
}
2014-06-30 21:14:31 +02:00
lastGenerated, err := strconv.ParseInt(fields[1], 10, 64)
if err != nil {
return nil, 0, errors.New("invalid longpollid format")
}
2014-06-30 21:14:31 +02:00
return prevHash, lastGenerated, nil
}
// notifyLongPollers notifies any channels that have been registered to be
// notified when block templates are stale.
//
// This function MUST be called with the state locked.
func (state *gbtWorkState) notifyLongPollers(latestHash *chainhash.Hash, lastGenerated time.Time) {
// Notify anything that is waiting for a block template update from a
// hash which is not the hash of the tip of the best chain since their
// work is now invalid.
for hash, channels := range state.notifyMap {
if !hash.IsEqual(latestHash) {
for _, c := range channels {
close(c)
}
delete(state.notifyMap, hash)
}
}
// Return now if the provided last generated timestamp has not been
// initialized.
if lastGenerated.IsZero() {
return
}
// Return now if there is nothing registered for updates to the current
// best block hash.
channels, ok := state.notifyMap[*latestHash]
if !ok {
return
}
// Notify anything that is waiting for a block template update from a
// block template generated before the most recently generated block
// template.
lastGeneratedUnix := lastGenerated.Unix()
for lastGen, c := range channels {
if lastGen < lastGeneratedUnix {
close(c)
delete(channels, lastGen)
}
}
// Remove the entry altogether if there are no more registered
// channels.
if len(channels) == 0 {
delete(state.notifyMap, *latestHash)
}
}
// NotifyBlockConnected uses the newly-connected block to notify any long poll
// clients with a new block template when their existing block template is
// stale due to the newly connected block.
func (state *gbtWorkState) NotifyBlockConnected(blockHash *chainhash.Hash) {
go func() {
state.Lock()
defer state.Unlock()
state.notifyLongPollers(blockHash, state.lastTxUpdate)
}()
}
// NotifyMempoolTx uses the new last updated time for the transaction memory
// pool to notify any long poll clients with a new block template when their
// existing block template is stale due to enough time passing and the contents
// of the memory pool changing.
func (state *gbtWorkState) NotifyMempoolTx(lastUpdated time.Time) {
go func() {
state.Lock()
defer state.Unlock()
// No need to notify anything if no block templates have been generated
// yet.
if state.prevHash == nil || state.lastGenerated.IsZero() {
return
}
if time.Now().After(state.lastGenerated.Add(time.Second *
gbtRegenerateSeconds)) {
state.notifyLongPollers(state.prevHash, lastUpdated)
}
}()
}
// templateUpdateChan returns a channel that will be closed once the block
// template associated with the passed previous hash and last generated time
// is stale. The function will return existing channels for duplicate
// parameters which allows multiple clients to wait for the same block template
// without requiring a different channel for each client.
//
// This function MUST be called with the state locked.
func (state *gbtWorkState) templateUpdateChan(prevHash *chainhash.Hash, lastGenerated int64) chan struct{} {
// Either get the current list of channels waiting for updates about
// changes to block template for the previous hash or create a new one.
channels, ok := state.notifyMap[*prevHash]
if !ok {
m := make(map[int64]chan struct{})
state.notifyMap[*prevHash] = m
channels = m
}
// Get the current channel associated with the time the block template
// was last generated or create a new one.
c, ok := channels[lastGenerated]
if !ok {
c = make(chan struct{})
channels[lastGenerated] = c
}
return c
}
// updateBlockTemplate creates or updates a block template for the work state.
// A new block template will be generated when the current best block has
// changed or the transactions in the memory pool have been updated and it has
2016-02-22 11:35:54 +01:00
// been long enough since the last template was generated. Otherwise, the
// timestamp for the existing block template is updated (and possibly the
// difficulty on testnet per the consesus rules). Finally, if the
2016-02-22 11:35:54 +01:00
// useCoinbaseValue flag is false and the existing block template does not
// already contain a valid payment address, the block template will be updated
// with a randomly selected payment address from the list of configured
// addresses.
//
// This function MUST be called with the state locked.
func (state *gbtWorkState) updateBlockTemplate(s *rpcServer, useCoinbaseValue bool) error {
lastTxUpdate := s.server.txMemPool.LastUpdated()
if lastTxUpdate.IsZero() {
lastTxUpdate = time.Now()
}
// Generate a new block template when the current best block has
// changed or the transactions in the memory pool have been updated and
// it has been at least gbtRegenerateSecond since the last template was
// generated.
var msgBlock *wire.MsgBlock
var targetDifficulty string
latestHash, _ := s.server.blockManager.chainState.Best()
template := state.template
if template == nil || state.prevHash == nil ||
!state.prevHash.IsEqual(latestHash) ||
(state.lastTxUpdate != lastTxUpdate &&
time.Now().After(state.lastGenerated.Add(time.Second*
gbtRegenerateSeconds))) {
// Reset the previous best hash the block template was generated
// against so any errors below cause the next invocation to try
// again.
state.prevHash = nil
// Choose a payment address at random if the caller requests a
// full coinbase as opposed to only the pertinent details needed
// to create their own coinbase.
var payAddr btcutil.Address
if !useCoinbaseValue {
payAddr = cfg.miningAddrs[rand.Intn(len(cfg.miningAddrs))]
}
// Create a new block template that has a coinbase which anyone
// can redeem. This is only acceptable because the returned
// block template doesn't include the coinbase, so the caller
// will ultimately create their own coinbase which pays to the
// appropriate address(es).
blkTemplate, err := NewBlockTemplate(s.policy, s.server, payAddr)
if err != nil {
return internalRPCError("Failed to create new block "+
"template: "+err.Error(), "")
}
template = blkTemplate
msgBlock = template.Block
targetDifficulty = fmt.Sprintf("%064x",
blockchain.CompactToBig(msgBlock.Header.Bits))
// Find the minimum allowed timestamp for the block based on the
// median timestamp of the last several blocks per the chain
// consensus rules.
chainState := &s.server.blockManager.chainState
minTimestamp, err := minimumMedianTime(chainState)
if err != nil {
context := "Failed to get minimum median time"
return internalRPCError(err.Error(), context)
}
// Update work state to ensure another block template isn't
// generated until needed.
state.template = template
state.lastGenerated = time.Now()
state.lastTxUpdate = lastTxUpdate
state.prevHash = latestHash
state.minTimestamp = minTimestamp
rpcsLog.Debugf("Generated block template (timestamp %v, "+
"target %s, merkle root %s)",
msgBlock.Header.Timestamp, targetDifficulty,
msgBlock.Header.MerkleRoot)
// Notify any clients that are long polling about the new
// template.
state.notifyLongPollers(latestHash, lastTxUpdate)
} else {
// At this point, there is a saved block template and another
// request for a template was made, but either the available
// transactions haven't change or it hasn't been long enough to
// trigger a new block template to be generated. So, update the
// existing block template.
// When the caller requires a full coinbase as opposed to only
// the pertinent details needed to create their own coinbase,
// add a payment address to the output of the coinbase of the
// template if it doesn't already have one. Since this requires
// mining addresses to be specified via the config, an error is
// returned if none have been specified.
if !useCoinbaseValue && !template.ValidPayAddress {
// Choose a payment address at random.
payToAddr := cfg.miningAddrs[rand.Intn(len(cfg.miningAddrs))]
// Update the block coinbase output of the template to
// pay to the randomly selected payment address.
pkScript, err := txscript.PayToAddrScript(payToAddr)
if err != nil {
context := "Failed to create pay-to-addr script"
return internalRPCError(err.Error(), context)
}
template.Block.Transactions[0].TxOut[0].PkScript = pkScript
template.ValidPayAddress = true
// Update the merkle root.
block := btcutil.NewBlock(template.Block)
merkles := blockchain.BuildMerkleTreeStore(block.Transactions())
template.Block.Header.MerkleRoot = *merkles[len(merkles)-1]
}
// Set locals for convenience.
msgBlock = template.Block
targetDifficulty = fmt.Sprintf("%064x",
blockchain.CompactToBig(msgBlock.Header.Bits))
// Update the time of the block template to the current time
// while accounting for the median time of the past several
// blocks per the chain consensus rules.
UpdateBlockTime(msgBlock, s.server.blockManager)
msgBlock.Header.Nonce = 0
rpcsLog.Debugf("Updated block template (timestamp %v, "+
"target %s)", msgBlock.Header.Timestamp,
targetDifficulty)
}
return nil
}
// blockTemplateResult returns the current block template associated with the
// state as a btcjson.GetBlockTemplateResult that is ready to be encoded to JSON
// and returned to the caller.
//
// This function MUST be called with the state locked.
func (state *gbtWorkState) blockTemplateResult(useCoinbaseValue bool, submitOld *bool) (*btcjson.GetBlockTemplateResult, error) {
// Ensure the timestamps are still in valid range for the template.
// This should really only ever happen if the local clock is changed
// after the template is generated, but it's important to avoid serving
// invalid block templates.
template := state.template
msgBlock := template.Block
header := &msgBlock.Header
adjustedTime := state.timeSource.AdjustedTime()
maxTime := adjustedTime.Add(time.Second * blockchain.MaxTimeOffsetSeconds)
if header.Timestamp.After(maxTime) {
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCOutOfRange,
Message: fmt.Sprintf("The template time is after the "+
"maximum allowed time for a block - template "+
"time %v, maximum time %v", adjustedTime,
maxTime),
}
}
// Convert each transaction in the block template to a template result
// transaction. The result does not include the coinbase, so notice
// the adjustments to the various lengths and indices.
numTx := len(msgBlock.Transactions)
transactions := make([]btcjson.GetBlockTemplateResultTx, 0, numTx-1)
txIndex := make(map[chainhash.Hash]int64, numTx)
for i, tx := range msgBlock.Transactions {
txHash := tx.TxHash()
txIndex[txHash] = int64(i)
// Skip the coinbase transaction.
if i == 0 {
continue
}
// Create an array of 1-based indices to transactions that come
// before this one in the transactions list which this one
// depends on. This is necessary since the created block must
// ensure proper ordering of the dependencies. A map is used
// before creating the final array to prevent duplicate entries
2016-02-22 11:35:54 +01:00
// when multiple inputs reference the same transaction.
dependsMap := make(map[int64]struct{})
for _, txIn := range tx.TxIn {
if idx, ok := txIndex[txIn.PreviousOutPoint.Hash]; ok {
dependsMap[idx] = struct{}{}
}
}
depends := make([]int64, 0, len(dependsMap))
for idx := range dependsMap {
depends = append(depends, idx)
}
// Serialize the transaction for later conversion to hex.
txBuf := bytes.NewBuffer(make([]byte, 0, tx.SerializeSize()))
if err := tx.Serialize(txBuf); err != nil {
context := "Failed to serialize transaction"
return nil, internalRPCError(err.Error(), context)
}
resultTx := btcjson.GetBlockTemplateResultTx{
Data: hex.EncodeToString(txBuf.Bytes()),
Hash: txHash.String(),
Depends: depends,
Fee: template.Fees[i],
SigOps: template.SigOpCounts[i],
}
transactions = append(transactions, resultTx)
}
// Generate the block template reply. Note that following mutations are
// implied by the included or omission of fields:
// Including MinTime -> time/decrement
// Omitting CoinbaseTxn -> coinbase, generation
targetDifficulty := fmt.Sprintf("%064x", blockchain.CompactToBig(header.Bits))
templateID := encodeTemplateID(state.prevHash, state.lastGenerated)
reply := btcjson.GetBlockTemplateResult{
Bits: strconv.FormatInt(int64(header.Bits), 16),
CurTime: header.Timestamp.Unix(),
Height: int64(template.Height),
PreviousHash: header.PrevBlock.String(),
SigOpLimit: blockchain.MaxSigOpsPerBlock,
SizeLimit: wire.MaxBlockPayload,
Transactions: transactions,
Version: header.Version,
LongPollID: templateID,
SubmitOld: submitOld,
Target: targetDifficulty,
MinTime: state.minTimestamp.Unix(),
MaxTime: maxTime.Unix(),
Mutable: gbtMutableFields,
NonceRange: gbtNonceRange,
Capabilities: gbtCapabilities,
}
if useCoinbaseValue {
reply.CoinbaseAux = gbtCoinbaseAux
reply.CoinbaseValue = &msgBlock.Transactions[0].TxOut[0].Value
} else {
// Ensure the template has a valid payment address associated
// with it when a full coinbase is requested.
if !template.ValidPayAddress {
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCInternal.Code,
Message: "A coinbase transaction has been " +
"requested, but the server has not " +
2014-06-30 21:14:31 +02:00
"been configured with any payment " +
"addresses via --miningaddr",
}
}
// Serialize the transaction for conversion to hex.
tx := msgBlock.Transactions[0]
txBuf := bytes.NewBuffer(make([]byte, 0, tx.SerializeSize()))
if err := tx.Serialize(txBuf); err != nil {
context := "Failed to serialize transaction"
return nil, internalRPCError(err.Error(), context)
}
resultTx := btcjson.GetBlockTemplateResultTx{
Data: hex.EncodeToString(txBuf.Bytes()),
Hash: tx.TxHash().String(),
Depends: []int64{},
Fee: template.Fees[0],
SigOps: template.SigOpCounts[0],
}
reply.CoinbaseTxn = &resultTx
}
return &reply, nil
}
2016-02-22 11:35:54 +01:00
// handleGetBlockTemplateLongPoll is a helper for handleGetBlockTemplateRequest
// which deals with handling long polling for block templates. When a caller
// sends a request with a long poll ID that was previously returned, a response
// is not sent until the caller should stop working on the previous block
// template in favor of the new one. In particular, this is the case when the
// old block template is no longer valid due to a solution already being found
// and added to the block chain, or new transactions have shown up and some time
// has passed without finding a solution.
//
// See https://en.bitcoin.it/wiki/BIP_0022 for more details.
func handleGetBlockTemplateLongPoll(s *rpcServer, longPollID string, useCoinbaseValue bool, closeChan <-chan struct{}) (interface{}, error) {
state := s.gbtWorkState
state.Lock()
// The state unlock is intentionally not deferred here since it needs to
// be manually unlocked before waiting for a notification about block
// template changes.
if err := state.updateBlockTemplate(s, useCoinbaseValue); err != nil {
state.Unlock()
return nil, err
}
2016-02-22 11:35:54 +01:00
// Just return the current block template if the long poll ID provided by
// the caller is invalid.
prevHash, lastGenerated, err := decodeTemplateID(longPollID)
if err != nil {
result, err := state.blockTemplateResult(useCoinbaseValue, nil)
if err != nil {
state.Unlock()
return nil, err
}
state.Unlock()
return result, nil
}
// Return the block template now if the specific block template
// identified by the long poll ID no longer matches the current block
// template as this means the provided template is stale.
prevTemplateHash := &state.template.Block.Header.PrevBlock
if !prevHash.IsEqual(prevTemplateHash) ||
lastGenerated != state.lastGenerated.Unix() {
// Include whether or not it is valid to submit work against the
// old block template depending on whether or not a solution has
// already been found and added to the block chain.
submitOld := prevHash.IsEqual(prevTemplateHash)
result, err := state.blockTemplateResult(useCoinbaseValue,
&submitOld)
if err != nil {
state.Unlock()
return nil, err
}
state.Unlock()
return result, nil
}
// Register the previous hash and last generated time for notifications
// Get a channel that will be notified when the template associated with
2016-02-22 11:35:54 +01:00
// the provided ID is stale and a new block template should be returned to
// the caller.
longPollChan := state.templateUpdateChan(prevHash, lastGenerated)
state.Unlock()
select {
// When the client closes before it's time to send a reply, just return
// now so the goroutine doesn't hang around.
case <-closeChan:
return nil, ErrClientQuit
// Wait until signal received to send the reply.
case <-longPollChan:
// Fallthrough
}
// Get the lastest block template
state.Lock()
defer state.Unlock()
if err := state.updateBlockTemplate(s, useCoinbaseValue); err != nil {
return nil, err
}
// Include whether or not it is valid to submit work against the old
// block template depending on whether or not a solution has already
// been found and added to the block chain.
submitOld := prevHash.IsEqual(&state.template.Block.Header.PrevBlock)
result, err := state.blockTemplateResult(useCoinbaseValue, &submitOld)
if err != nil {
return nil, err
}
return result, nil
}
// handleGetBlockTemplateRequest is a helper for handleGetBlockTemplate which
// deals with generating and returning block templates to the caller. It
// handles both long poll requests as specified by BIP 0022 as well as regular
// requests. In addition, it detects the capabilities reported by the caller
// in regards to whether or not it supports creating its own coinbase (the
// coinbasetxn and coinbasevalue capabilities) and modifies the returned block
// template accordingly.
func handleGetBlockTemplateRequest(s *rpcServer, request *btcjson.TemplateRequest, closeChan <-chan struct{}) (interface{}, error) {
// Extract the relevant passed capabilities and restrict the result to
// either a coinbase value or a coinbase transaction object depending on
// the request. Default to only providing a coinbase value.
useCoinbaseValue := true
if request != nil {
var hasCoinbaseValue, hasCoinbaseTxn bool
for _, capability := range request.Capabilities {
switch capability {
case "coinbasetxn":
hasCoinbaseTxn = true
case "coinbasevalue":
hasCoinbaseValue = true
}
}
if hasCoinbaseTxn && !hasCoinbaseValue {
useCoinbaseValue = false
}
}
// When a coinbase transaction has been requested, respond with an error
// if there are no addresses to pay the created block template to.
if !useCoinbaseValue && len(cfg.miningAddrs) == 0 {
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCInternal.Code,
Message: "A coinbase transaction has been requested, " +
"but the server has not been configured with " +
2014-06-30 21:14:31 +02:00
"any payment addresses via --miningaddr",
}
}
// Return an error if there are no peers connected since there is no
// way to relay a found block or receive transactions to work on.
// However, allow this state when running in the regression test or
// simulation test mode.
if !(cfg.RegressionTest || cfg.SimNet) && s.server.ConnectedCount() == 0 {
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCClientNotConnected,
Message: "Bitcoin is not connected",
}
}
// No point in generating or accepting work before the chain is synced.
_, currentHeight := s.server.blockManager.chainState.Best()
if currentHeight != 0 && !s.server.blockManager.IsCurrent() {
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCClientInInitialDownload,
Message: "Bitcoin is downloading blocks...",
}
}
// When a long poll ID was provided, this is a long poll request by the
// client to be notified when block template referenced by the ID should
// be replaced with a new one.
if request != nil && request.LongPollID != "" {
return handleGetBlockTemplateLongPoll(s, request.LongPollID,
useCoinbaseValue, closeChan)
}
// Protect concurrent access when updating block templates.
state := s.gbtWorkState
state.Lock()
defer state.Unlock()
// Get and return a block template. A new block template will be
// generated when the current best block has changed or the transactions
// in the memory pool have been updated and it has been at least five
// seconds since the last template was generated. Otherwise, the
// timestamp for the existing block template is updated (and possibly
// the difficulty on testnet per the consesus rules).
if err := state.updateBlockTemplate(s, useCoinbaseValue); err != nil {
return nil, err
}
return state.blockTemplateResult(useCoinbaseValue, nil)
}
// chainErrToGBTErrString converts an error returned from btcchain to a string
// which matches the reasons and format described in BIP0022 for rejection
// reasons.
func chainErrToGBTErrString(err error) string {
// When the passed error is not a RuleError, just return a generic
// rejected string with the error text.
ruleErr, ok := err.(blockchain.RuleError)
if !ok {
return "rejected: " + err.Error()
}
switch ruleErr.ErrorCode {
case blockchain.ErrDuplicateBlock:
return "duplicate"
case blockchain.ErrBlockTooBig:
return "bad-block-size"
case blockchain.ErrBlockVersionTooOld:
return "bad-version"
case blockchain.ErrInvalidTime:
return "bad-time"
case blockchain.ErrTimeTooOld:
return "time-too-old"
case blockchain.ErrTimeTooNew:
return "time-too-new"
case blockchain.ErrDifficultyTooLow:
return "bad-diffbits"
case blockchain.ErrUnexpectedDifficulty:
return "bad-diffbits"
case blockchain.ErrHighHash:
return "high-hash"
case blockchain.ErrBadMerkleRoot:
return "bad-txnmrklroot"
case blockchain.ErrBadCheckpoint:
return "bad-checkpoint"
case blockchain.ErrForkTooOld:
return "fork-too-old"
case blockchain.ErrCheckpointTimeTooOld:
return "checkpoint-time-too-old"
case blockchain.ErrNoTransactions:
return "bad-txns-none"
case blockchain.ErrTooManyTransactions:
return "bad-txns-toomany"
case blockchain.ErrNoTxInputs:
return "bad-txns-noinputs"
case blockchain.ErrNoTxOutputs:
return "bad-txns-nooutputs"
case blockchain.ErrTxTooBig:
return "bad-txns-size"
case blockchain.ErrBadTxOutValue:
return "bad-txns-outputvalue"
case blockchain.ErrDuplicateTxInputs:
return "bad-txns-dupinputs"
case blockchain.ErrBadTxInput:
return "bad-txns-badinput"
case blockchain.ErrMissingTx:
return "bad-txns-missinginput"
case blockchain.ErrUnfinalizedTx:
return "bad-txns-unfinalizedtx"
case blockchain.ErrDuplicateTx:
return "bad-txns-duplicate"
case blockchain.ErrOverwriteTx:
return "bad-txns-overwrite"
case blockchain.ErrImmatureSpend:
return "bad-txns-maturity"
case blockchain.ErrDoubleSpend:
return "bad-txns-dblspend"
case blockchain.ErrSpendTooHigh:
return "bad-txns-highspend"
case blockchain.ErrBadFees:
return "bad-txns-fees"
case blockchain.ErrTooManySigOps:
return "high-sigops"
case blockchain.ErrFirstTxNotCoinbase:
return "bad-txns-nocoinbase"
case blockchain.ErrMultipleCoinbases:
return "bad-txns-multicoinbase"
case blockchain.ErrBadCoinbaseScriptLen:
return "bad-cb-length"
case blockchain.ErrBadCoinbaseValue:
return "bad-cb-value"
case blockchain.ErrMissingCoinbaseHeight:
return "bad-cb-height"
case blockchain.ErrBadCoinbaseHeight:
return "bad-cb-height"
case blockchain.ErrScriptMalformed:
return "bad-script-malformed"
case blockchain.ErrScriptValidation:
return "bad-script-validate"
}
return "rejected: " + err.Error()
}
// handleGetBlockTemplateProposal is a helper for handleGetBlockTemplate which
// deals with block proposals.
//
// See https://en.bitcoin.it/wiki/BIP_0023 for more details.
func handleGetBlockTemplateProposal(s *rpcServer, request *btcjson.TemplateRequest) (interface{}, error) {
hexData := request.Data
if hexData == "" {
return false, &btcjson.RPCError{
Code: btcjson.ErrRPCType,
Message: fmt.Sprintf("Data must contain the " +
"hex-encoded serialized block that is being " +
"proposed"),
}
}
// Ensure the provided data is sane and deserialize the proposed block.
if len(hexData)%2 != 0 {
hexData = "0" + hexData
}
dataBytes, err := hex.DecodeString(hexData)
if err != nil {
return false, &btcjson.RPCError{
Code: btcjson.ErrRPCDeserialization,
Message: fmt.Sprintf("Data must be "+
"hexadecimal string (not %q)", hexData),
}
}
var msgBlock wire.MsgBlock
if err := msgBlock.Deserialize(bytes.NewReader(dataBytes)); err != nil {
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCDeserialization,
Message: "Block decode failed: " + err.Error(),
}
}
block := btcutil.NewBlock(&msgBlock)
// Ensure the block is building from the expected previous block.
expectedPrevHash, _ := s.server.blockManager.chainState.Best()
prevHash := &block.MsgBlock().Header.PrevBlock
if expectedPrevHash == nil || !expectedPrevHash.IsEqual(prevHash) {
return "bad-prevblk", nil
}
flags := blockchain.BFDryRun | blockchain.BFNoPoWCheck
isOrphan, err := s.server.blockManager.ProcessBlock(block, flags)
if err != nil {
if _, ok := err.(blockchain.RuleError); !ok {
err := rpcsLog.Errorf("Failed to process block "+
"proposal: %v", err)
return nil, &btcjson.RPCError{
Code: -25, // TODO: ErrRpcVerify
Message: err.Error(),
}
}
rpcsLog.Infof("Rejected block proposal: %v", err)
return chainErrToGBTErrString(err), nil
}
if isOrphan {
return "orphan", nil
}
return nil, nil
}
// handleGetBlockTemplate implements the getblocktemplate command.
//
// See https://en.bitcoin.it/wiki/BIP_0022 and
// https://en.bitcoin.it/wiki/BIP_0023 for more details.
func handleGetBlockTemplate(s *rpcServer, cmd interface{}, closeChan <-chan struct{}) (interface{}, error) {
c := cmd.(*btcjson.GetBlockTemplateCmd)
request := c.Request
// Set the default mode and override it if supplied.
mode := "template"
if request != nil && request.Mode != "" {
mode = request.Mode
}
switch mode {
case "template":
return handleGetBlockTemplateRequest(s, request, closeChan)
case "proposal":
return handleGetBlockTemplateProposal(s, request)
}
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCInvalidParameter,
Message: "Invalid mode",
}
}
// handleGetConnectionCount implements the getconnectioncount command.
func handleGetConnectionCount(s *rpcServer, cmd interface{}, closeChan <-chan struct{}) (interface{}, error) {
return s.server.ConnectedCount(), nil
}
// handleGetCurrentNet implements the getcurrentnet command.
func handleGetCurrentNet(s *rpcServer, cmd interface{}, closeChan <-chan struct{}) (interface{}, error) {
return s.server.chainParams.Net, nil
}
// handleGetDifficulty implements the getdifficulty command.
func handleGetDifficulty(s *rpcServer, cmd interface{}, closeChan <-chan struct{}) (interface{}, error) {
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
best := s.chain.BestSnapshot()
return getDifficultyRatio(best.Bits), nil
}
2013-10-29 01:43:09 +01:00
// handleGetGenerate implements the getgenerate command.
func handleGetGenerate(s *rpcServer, cmd interface{}, closeChan <-chan struct{}) (interface{}, error) {
return s.server.cpuMiner.IsMining(), nil
}
2013-10-29 01:43:09 +01:00
// handleGetHashesPerSec implements the gethashespersec command.
func handleGetHashesPerSec(s *rpcServer, cmd interface{}, closeChan <-chan struct{}) (interface{}, error) {
return int64(s.server.cpuMiner.HashesPerSecond()), nil
}
2013-10-29 01:43:09 +01:00
// handleGetInfo implements the getinfo command. We only return the fields
// that are not related to wallet functionality.
func handleGetInfo(s *rpcServer, cmd interface{}, closeChan <-chan struct{}) (interface{}, error) {
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
best := s.chain.BestSnapshot()
ret := &btcjson.InfoChainResult{
Version: int32(1000000*appMajor + 10000*appMinor + 100*appPatch),
ProtocolVersion: int32(maxProtocolVersion),
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
Blocks: best.Height,
TimeOffset: int64(s.server.timeSource.Offset().Seconds()),
Connections: s.server.ConnectedCount(),
Proxy: cfg.Proxy,
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
Difficulty: getDifficultyRatio(best.Bits),
TestNet: cfg.TestNet3,
RelayFee: cfg.minRelayTxFee.ToBTC(),
}
return ret, nil
}
2015-06-23 21:43:39 +02:00
// handleGetMempoolInfo implements the getmempoolinfo command.
func handleGetMempoolInfo(s *rpcServer, cmd interface{}, closeChan <-chan struct{}) (interface{}, error) {
mempoolTxns := s.server.txMemPool.TxDescs()
2015-06-23 21:43:39 +02:00
var numBytes int64
for _, txD := range mempoolTxns {
numBytes += int64(txD.Tx.MsgTx().SerializeSize())
2015-06-23 21:43:39 +02:00
}
ret := &btcjson.GetMempoolInfoResult{
Size: int64(len(mempoolTxns)),
2015-06-23 21:43:39 +02:00
Bytes: numBytes,
}
return ret, nil
}
// handleGetMiningInfo implements the getmininginfo command. We only return the
// fields that are not related to wallet functionality.
func handleGetMiningInfo(s *rpcServer, cmd interface{}, closeChan <-chan struct{}) (interface{}, error) {
// Create a default getnetworkhashps command to use defaults and make
// use of the existing getnetworkhashps handler.
gnhpsCmd := btcjson.NewGetNetworkHashPSCmd(nil, nil)
networkHashesPerSecIface, err := handleGetNetworkHashPS(s, gnhpsCmd,
closeChan)
if err != nil {
return nil, err
}
networkHashesPerSec, ok := networkHashesPerSecIface.(int64)
if !ok {
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCInternal.Code,
Message: "networkHashesPerSec is not an int64",
}
}
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
best := s.chain.BestSnapshot()
result := btcjson.GetMiningInfoResult{
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
Blocks: int64(best.Height),
CurrentBlockSize: best.BlockSize,
CurrentBlockTx: best.NumTxns,
Difficulty: getDifficultyRatio(best.Bits),
Generate: s.server.cpuMiner.IsMining(),
GenProcLimit: s.server.cpuMiner.NumWorkers(),
HashesPerSec: int64(s.server.cpuMiner.HashesPerSecond()),
NetworkHashPS: networkHashesPerSec,
PooledTx: uint64(s.server.txMemPool.Count()),
TestNet: cfg.TestNet3,
}
return &result, nil
}
// handleGetNetTotals implements the getnettotals command.
func handleGetNetTotals(s *rpcServer, cmd interface{}, closeChan <-chan struct{}) (interface{}, error) {
totalBytesRecv, totalBytesSent := s.server.NetTotals()
reply := &btcjson.GetNetTotalsResult{
TotalBytesRecv: totalBytesRecv,
TotalBytesSent: totalBytesSent,
TimeMillis: time.Now().UTC().UnixNano() / int64(time.Millisecond),
}
return reply, nil
}
// handleGetNetworkHashPS implements the getnetworkhashps command.
func handleGetNetworkHashPS(s *rpcServer, cmd interface{}, closeChan <-chan struct{}) (interface{}, error) {
// Note: All valid error return paths should return an int64.
// Literal zeros are inferred as int, and won't coerce to int64
// because the return value is an interface{}.
c := cmd.(*btcjson.GetNetworkHashPSCmd)
// When the passed height is too high or zero, just return 0 now
// since we can't reasonably calculate the number of network hashes
// per second from invalid values. When it's negative, use the current
// best block height.
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
best := s.chain.BestSnapshot()
endHeight := int32(-1)
if c.Height != nil {
endHeight = int32(*c.Height)
}
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
if endHeight > best.Height || endHeight == 0 {
return int64(0), nil
}
if endHeight < 0 {
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
endHeight = best.Height
}
// Calculate the number of blocks per retarget interval based on the
// chain parameters.
blocksPerRetarget := int32(s.server.chainParams.TargetTimespan /
s.server.chainParams.TargetTimePerBlock)
// Calculate the starting block height based on the passed number of
// blocks. When the passed value is negative, use the last block the
// difficulty changed as the starting height. Also make sure the
// starting height is not before the beginning of the chain.
numBlocks := int32(120)
if c.Blocks != nil {
numBlocks = int32(*c.Blocks)
}
var startHeight int32
if numBlocks <= 0 {
startHeight = endHeight - ((endHeight % blocksPerRetarget) + 1)
} else {
startHeight = endHeight - numBlocks
}
if startHeight < 0 {
startHeight = 0
}
rpcsLog.Debugf("Calculating network hashes per second from %d to %d",
startHeight, endHeight)
// Find the min and max block timestamps as well as calculate the total
// amount of work that happened between the start and end blocks.
var minTimestamp, maxTimestamp time.Time
totalWork := big.NewInt(0)
for curHeight := startHeight; curHeight <= endHeight; curHeight++ {
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
hash, err := s.chain.BlockHashByHeight(curHeight)
if err != nil {
context := "Failed to fetch block hash"
return nil, internalRPCError(err.Error(), context)
}
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
// Load the raw header bytes.
var headerBytes []byte
err = s.server.db.View(func(dbTx database.Tx) error {
var err error
headerBytes, err = dbTx.FetchBlockHeader(hash)
return err
})
if err != nil {
context := "Failed to fetch block header"
return nil, internalRPCError(err.Error(), context)
}
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
// Deserialize the header.
var header wire.BlockHeader
err = header.Deserialize(bytes.NewReader(headerBytes))
if err != nil {
context := "Failed to deserialize block header"
return nil, internalRPCError(err.Error(), context)
}
if curHeight == startHeight {
minTimestamp = header.Timestamp
maxTimestamp = minTimestamp
} else {
totalWork.Add(totalWork, blockchain.CalcWork(header.Bits))
if minTimestamp.After(header.Timestamp) {
minTimestamp = header.Timestamp
}
if maxTimestamp.Before(header.Timestamp) {
maxTimestamp = header.Timestamp
}
}
}
// Calculate the difference in seconds between the min and max block
// timestamps and avoid division by zero in the case where there is no
// time difference.
timeDiff := int64(maxTimestamp.Sub(minTimestamp) / time.Second)
if timeDiff == 0 {
return int64(0), nil
}
hashesPerSec := new(big.Int).Div(totalWork, big.NewInt(timeDiff))
return hashesPerSec.Int64(), nil
}
// handleGetPeerInfo implements the getpeerinfo command.
func handleGetPeerInfo(s *rpcServer, cmd interface{}, closeChan <-chan struct{}) (interface{}, error) {
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
peers := s.server.Peers()
syncPeer := s.server.blockManager.SyncPeer()
infos := make([]*btcjson.GetPeerInfoResult, 0, len(peers))
for _, p := range peers {
statsSnap := p.StatsSnapshot()
info := &btcjson.GetPeerInfoResult{
ID: statsSnap.ID,
Addr: statsSnap.Addr,
Services: fmt.Sprintf("%08d", uint64(statsSnap.Services)),
LastSend: statsSnap.LastSend.Unix(),
LastRecv: statsSnap.LastRecv.Unix(),
BytesSent: statsSnap.BytesSent,
BytesRecv: statsSnap.BytesRecv,
ConnTime: statsSnap.ConnTime.Unix(),
PingTime: float64(statsSnap.LastPingMicros),
TimeOffset: statsSnap.TimeOffset,
Version: statsSnap.Version,
SubVer: statsSnap.UserAgent,
Inbound: statsSnap.Inbound,
StartingHeight: statsSnap.StartingHeight,
CurrentHeight: statsSnap.LastBlock,
BanScore: int32(p.banScore.Int()),
peer: Refactor peer code into its own package. This commit introduces package peer which contains peer related features refactored from peer.go. The following is an overview of the features the package provides: - Provides a basic concurrent safe bitcoin peer for handling bitcoin communications via the peer-to-peer protocol - Full duplex reading and writing of bitcoin protocol messages - Automatic handling of the initial handshake process including protocol version negotiation - Automatic periodic keep-alive pinging and pong responses - Asynchronous message queueing of outbound messages with optional channel for notification when the message is actually sent - Inventory message batching and send trickling with known inventory detection and avoidance - Ability to wait for shutdown/disconnect - Flexible peer configuration - Caller is responsible for creating outgoing connections and listening for incoming connections so they have flexibility to establish connections as they see fit (proxies, etc.) - User agent name and version - Bitcoin network - Service support signalling (full nodes, bloom filters, etc.) - Maximum supported protocol version - Ability to register callbacks for handling bitcoin protocol messages - Proper handling of bloom filter related commands when the caller does not specify the related flag to signal support - Disconnects the peer when the protocol version is high enough - Does not invoke the related callbacks for older protocol versions - Snapshottable peer statistics such as the total number of bytes read and written, the remote address, user agent, and negotiated protocol version - Helper functions for pushing addresses, getblocks, getheaders, and reject messages - These could all be sent manually via the standard message output function, but the helpers provide additional nice functionality such as duplicate filtering and address randomization - Full documentation with example usage - Test coverage In addition to the addition of the new package, btcd has been refactored to make use of the new package by extending the basic peer it provides to work with the blockmanager and server to act as a full node. The following is a broad overview of the changes to integrate the package: - The server is responsible for all connection management including persistent peers and banning - Callbacks for all messages that are required to implement a full node are registered - Logic necessary to serve data and behave as a full node is now in the callback registered with the peer Finally, the following peer-related things have been improved as a part of this refactor: - Don't log or send reject message due to peer disconnects - Remove trace logs that aren't particularly helpful - Finish an old TODO to switch the queue WaitGroup over to a channel - Improve various comments and fix some code consistency cases - Improve a few logging bits - Implement a most-recently-used nonce tracking for detecting self connections and generate a unique nonce for each peer
2015-10-02 08:03:20 +02:00
SyncNode: p == syncPeer,
}
if p.LastPingNonce() != 0 {
wait := float64(time.Now().Sub(statsSnap.LastPingTime).Nanoseconds())
// We actually want microseconds.
info.PingWait = wait / 1000
}
infos = append(infos, info)
}
return infos, nil
}
// handleGetRawMempool implements the getrawmempool command.
func handleGetRawMempool(s *rpcServer, cmd interface{}, closeChan <-chan struct{}) (interface{}, error) {
c := cmd.(*btcjson.GetRawMempoolCmd)
mp := s.server.txMemPool
descs := mp.TxDescs()
if c.Verbose != nil && *c.Verbose {
result := make(map[string]*btcjson.GetRawMempoolVerboseResult,
len(descs))
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
best := s.chain.BestSnapshot()
mp.RLock()
defer mp.RUnlock()
for _, desc := range descs {
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
// Calculate the current priority based on the inputs to
// the transaction. Use zero if one or more of the
// input transactions can't be found for some reason.
tx := desc.Tx
var currentPriority float64
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
utxos, err := mp.fetchInputUtxos(tx)
if err == nil {
currentPriority = calcPriority(tx.MsgTx(),
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
utxos, best.Height+1)
}
mpd := &btcjson.GetRawMempoolVerboseResult{
Size: int32(tx.MsgTx().SerializeSize()),
Fee: btcutil.Amount(desc.Fee).ToBTC(),
2013-12-17 15:02:35 +01:00
Time: desc.Added.Unix(),
Height: int64(desc.Height),
StartingPriority: desc.StartingPriority,
CurrentPriority: currentPriority,
Depends: make([]string, 0),
}
for _, txIn := range tx.MsgTx().TxIn {
hash := &txIn.PreviousOutPoint.Hash
if s.server.txMemPool.haveTransaction(hash) {
mpd.Depends = append(mpd.Depends,
hash.String())
}
}
result[tx.Hash().String()] = mpd
}
2013-12-17 15:02:35 +01:00
return result, nil
}
// The response is simply an array of the transaction hashes if the
// verbose flag is not set.
hashStrings := make([]string, len(descs))
2013-12-17 15:02:35 +01:00
for i := range hashStrings {
hashStrings[i] = descs[i].Tx.Hash().String()
}
return hashStrings, nil
}
// handleGetRawTransaction implements the getrawtransaction command.
func handleGetRawTransaction(s *rpcServer, cmd interface{}, closeChan <-chan struct{}) (interface{}, error) {
c := cmd.(*btcjson.GetRawTransactionCmd)
// Convert the provided transaction hash hex to a Hash.
txHash, err := chainhash.NewHashFromStr(c.Txid)
if err != nil {
return nil, rpcDecodeHexError(c.Txid)
}
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
verbose := false
if c.Verbose != nil {
verbose = *c.Verbose != 0
}
indexers: Implement optional tx/address indexes. This introduces a new indexing infrastructure for supporting optional indexes using the new database and blockchain infrastructure along with two concrete indexer implementations which provide both a transaction-by-hash and a transaction-by-address index. The new infrastructure is mostly separated into a package named indexers which is housed under the blockchain package. In order to support this, a new interface named IndexManager has been introduced in the blockchain package which provides methods to be notified when the chain has been initialized and when blocks are connected and disconnected from the main chain. A concrete implementation of an index manager is provided by the new indexers package. The new indexers package also provides a new interface named Indexer which allows the index manager to manage concrete index implementations which conform to the interface. The following is high level overview of the main index infrastructure changes: - Define a new IndexManager interface in the blockchain package and modify the package to make use of the interface when specified - Create a new indexers package - Provides an Index interface which allows concrete indexes to plugin to an index manager - Provides a concrete IndexManager implementation - Handles the lifecycle of all indexes it manages - Tracks the index tips - Handles catching up disabled indexes that have been reenabled - Handles reorgs while the index was disabled - Invokes the appropriate methods for all managed indexes to allow them to index and deindex the blocks and transactions - Implement a transaction-by-hash index - Makes use of internal block IDs to save a significant amount of space and indexing costs over the old transaction index format - Implement a transaction-by-address index - Makes use of a leveling scheme in order to provide a good tradeoff between space required and indexing costs - Supports enabling and disabling indexes at will - Support the ability to drop indexes if they are no longer desired The following is an overview of the btcd changes: - Add a new index logging subsystem - Add new options --txindex and --addrindex in order to enable the optional indexes - NOTE: The transaction index will automatically be enabled when the address index is enabled because it depends on it - Add new options --droptxindex and --dropaddrindex to allow the indexes to be removed - NOTE: The address index will also be removed when the transaction index is dropped because it depends on it - Update getrawtransactions RPC to make use of the transaction index - Reimplement the searchrawtransaction RPC that makes use of the address index - Update sample-btcd.conf to include sample usage for the new optional index flags
2016-02-19 05:51:18 +01:00
// Try to fetch the transaction from the memory pool and if that fails,
// try the block database.
var mtx *wire.MsgTx
var blkHash *chainhash.Hash
indexers: Implement optional tx/address indexes. This introduces a new indexing infrastructure for supporting optional indexes using the new database and blockchain infrastructure along with two concrete indexer implementations which provide both a transaction-by-hash and a transaction-by-address index. The new infrastructure is mostly separated into a package named indexers which is housed under the blockchain package. In order to support this, a new interface named IndexManager has been introduced in the blockchain package which provides methods to be notified when the chain has been initialized and when blocks are connected and disconnected from the main chain. A concrete implementation of an index manager is provided by the new indexers package. The new indexers package also provides a new interface named Indexer which allows the index manager to manage concrete index implementations which conform to the interface. The following is high level overview of the main index infrastructure changes: - Define a new IndexManager interface in the blockchain package and modify the package to make use of the interface when specified - Create a new indexers package - Provides an Index interface which allows concrete indexes to plugin to an index manager - Provides a concrete IndexManager implementation - Handles the lifecycle of all indexes it manages - Tracks the index tips - Handles catching up disabled indexes that have been reenabled - Handles reorgs while the index was disabled - Invokes the appropriate methods for all managed indexes to allow them to index and deindex the blocks and transactions - Implement a transaction-by-hash index - Makes use of internal block IDs to save a significant amount of space and indexing costs over the old transaction index format - Implement a transaction-by-address index - Makes use of a leveling scheme in order to provide a good tradeoff between space required and indexing costs - Supports enabling and disabling indexes at will - Support the ability to drop indexes if they are no longer desired The following is an overview of the btcd changes: - Add a new index logging subsystem - Add new options --txindex and --addrindex in order to enable the optional indexes - NOTE: The transaction index will automatically be enabled when the address index is enabled because it depends on it - Add new options --droptxindex and --dropaddrindex to allow the indexes to be removed - NOTE: The address index will also be removed when the transaction index is dropped because it depends on it - Update getrawtransactions RPC to make use of the transaction index - Reimplement the searchrawtransaction RPC that makes use of the address index - Update sample-btcd.conf to include sample usage for the new optional index flags
2016-02-19 05:51:18 +01:00
var blkHeight int32
tx, err := s.server.txMemPool.FetchTransaction(txHash)
if err != nil {
indexers: Implement optional tx/address indexes. This introduces a new indexing infrastructure for supporting optional indexes using the new database and blockchain infrastructure along with two concrete indexer implementations which provide both a transaction-by-hash and a transaction-by-address index. The new infrastructure is mostly separated into a package named indexers which is housed under the blockchain package. In order to support this, a new interface named IndexManager has been introduced in the blockchain package which provides methods to be notified when the chain has been initialized and when blocks are connected and disconnected from the main chain. A concrete implementation of an index manager is provided by the new indexers package. The new indexers package also provides a new interface named Indexer which allows the index manager to manage concrete index implementations which conform to the interface. The following is high level overview of the main index infrastructure changes: - Define a new IndexManager interface in the blockchain package and modify the package to make use of the interface when specified - Create a new indexers package - Provides an Index interface which allows concrete indexes to plugin to an index manager - Provides a concrete IndexManager implementation - Handles the lifecycle of all indexes it manages - Tracks the index tips - Handles catching up disabled indexes that have been reenabled - Handles reorgs while the index was disabled - Invokes the appropriate methods for all managed indexes to allow them to index and deindex the blocks and transactions - Implement a transaction-by-hash index - Makes use of internal block IDs to save a significant amount of space and indexing costs over the old transaction index format - Implement a transaction-by-address index - Makes use of a leveling scheme in order to provide a good tradeoff between space required and indexing costs - Supports enabling and disabling indexes at will - Support the ability to drop indexes if they are no longer desired The following is an overview of the btcd changes: - Add a new index logging subsystem - Add new options --txindex and --addrindex in order to enable the optional indexes - NOTE: The transaction index will automatically be enabled when the address index is enabled because it depends on it - Add new options --droptxindex and --dropaddrindex to allow the indexes to be removed - NOTE: The address index will also be removed when the transaction index is dropped because it depends on it - Update getrawtransactions RPC to make use of the transaction index - Reimplement the searchrawtransaction RPC that makes use of the address index - Update sample-btcd.conf to include sample usage for the new optional index flags
2016-02-19 05:51:18 +01:00
txIndex := s.server.txIndex
if txIndex == nil {
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCNoTxInfo,
Message: "The transaction index must be " +
"enabled to query the blockchain " +
"(specify --txindex)",
}
}
2013-10-29 01:43:09 +01:00
indexers: Implement optional tx/address indexes. This introduces a new indexing infrastructure for supporting optional indexes using the new database and blockchain infrastructure along with two concrete indexer implementations which provide both a transaction-by-hash and a transaction-by-address index. The new infrastructure is mostly separated into a package named indexers which is housed under the blockchain package. In order to support this, a new interface named IndexManager has been introduced in the blockchain package which provides methods to be notified when the chain has been initialized and when blocks are connected and disconnected from the main chain. A concrete implementation of an index manager is provided by the new indexers package. The new indexers package also provides a new interface named Indexer which allows the index manager to manage concrete index implementations which conform to the interface. The following is high level overview of the main index infrastructure changes: - Define a new IndexManager interface in the blockchain package and modify the package to make use of the interface when specified - Create a new indexers package - Provides an Index interface which allows concrete indexes to plugin to an index manager - Provides a concrete IndexManager implementation - Handles the lifecycle of all indexes it manages - Tracks the index tips - Handles catching up disabled indexes that have been reenabled - Handles reorgs while the index was disabled - Invokes the appropriate methods for all managed indexes to allow them to index and deindex the blocks and transactions - Implement a transaction-by-hash index - Makes use of internal block IDs to save a significant amount of space and indexing costs over the old transaction index format - Implement a transaction-by-address index - Makes use of a leveling scheme in order to provide a good tradeoff between space required and indexing costs - Supports enabling and disabling indexes at will - Support the ability to drop indexes if they are no longer desired The following is an overview of the btcd changes: - Add a new index logging subsystem - Add new options --txindex and --addrindex in order to enable the optional indexes - NOTE: The transaction index will automatically be enabled when the address index is enabled because it depends on it - Add new options --droptxindex and --dropaddrindex to allow the indexes to be removed - NOTE: The address index will also be removed when the transaction index is dropped because it depends on it - Update getrawtransactions RPC to make use of the transaction index - Reimplement the searchrawtransaction RPC that makes use of the address index - Update sample-btcd.conf to include sample usage for the new optional index flags
2016-02-19 05:51:18 +01:00
// Look up the location of the transaction.
blockRegion, err := txIndex.TxBlockRegion(txHash)
if err != nil {
indexers: Implement optional tx/address indexes. This introduces a new indexing infrastructure for supporting optional indexes using the new database and blockchain infrastructure along with two concrete indexer implementations which provide both a transaction-by-hash and a transaction-by-address index. The new infrastructure is mostly separated into a package named indexers which is housed under the blockchain package. In order to support this, a new interface named IndexManager has been introduced in the blockchain package which provides methods to be notified when the chain has been initialized and when blocks are connected and disconnected from the main chain. A concrete implementation of an index manager is provided by the new indexers package. The new indexers package also provides a new interface named Indexer which allows the index manager to manage concrete index implementations which conform to the interface. The following is high level overview of the main index infrastructure changes: - Define a new IndexManager interface in the blockchain package and modify the package to make use of the interface when specified - Create a new indexers package - Provides an Index interface which allows concrete indexes to plugin to an index manager - Provides a concrete IndexManager implementation - Handles the lifecycle of all indexes it manages - Tracks the index tips - Handles catching up disabled indexes that have been reenabled - Handles reorgs while the index was disabled - Invokes the appropriate methods for all managed indexes to allow them to index and deindex the blocks and transactions - Implement a transaction-by-hash index - Makes use of internal block IDs to save a significant amount of space and indexing costs over the old transaction index format - Implement a transaction-by-address index - Makes use of a leveling scheme in order to provide a good tradeoff between space required and indexing costs - Supports enabling and disabling indexes at will - Support the ability to drop indexes if they are no longer desired The following is an overview of the btcd changes: - Add a new index logging subsystem - Add new options --txindex and --addrindex in order to enable the optional indexes - NOTE: The transaction index will automatically be enabled when the address index is enabled because it depends on it - Add new options --droptxindex and --dropaddrindex to allow the indexes to be removed - NOTE: The address index will also be removed when the transaction index is dropped because it depends on it - Update getrawtransactions RPC to make use of the transaction index - Reimplement the searchrawtransaction RPC that makes use of the address index - Update sample-btcd.conf to include sample usage for the new optional index flags
2016-02-19 05:51:18 +01:00
context := "Failed to retrieve transaction location"
return nil, internalRPCError(err.Error(), context)
}
if blockRegion == nil {
return nil, rpcNoTxInfoError(txHash)
}
// Load the raw transaction bytes from the database.
var txBytes []byte
err = s.server.db.View(func(dbTx database.Tx) error {
var err error
txBytes, err = dbTx.FetchBlockRegion(blockRegion)
return err
})
if err != nil {
return nil, rpcNoTxInfoError(txHash)
}
// When the verbose flag isn't set, simply return the serialized
// transaction as a hex-encoded string. This is done here to
// avoid deserializing it only to reserialize it again later.
if !verbose {
return hex.EncodeToString(txBytes), nil
}
// Grab the block height.
blkHash = blockRegion.Hash
blkHeight, err = s.chain.BlockHeightByHash(blkHash)
if err != nil {
context := "Failed to retrieve block height"
return nil, internalRPCError(err.Error(), context)
}
// Deserialize the transaction
var msgTx wire.MsgTx
err = msgTx.Deserialize(bytes.NewReader(txBytes))
if err != nil {
context := "Failed to deserialize transaction"
return nil, internalRPCError(err.Error(), context)
}
mtx = &msgTx
} else {
// When the verbose flag isn't set, simply return the
// network-serialized transaction as a hex-encoded string.
if !verbose {
// Note that this is intentionally not directly
// returning because the first return value is a
// string and it would result in returning an empty
// string to the client instead of nothing (nil) in the
// case of an error.
mtxHex, err := messageToHex(tx.MsgTx())
if err != nil {
return nil, err
}
return mtxHex, nil
}
indexers: Implement optional tx/address indexes. This introduces a new indexing infrastructure for supporting optional indexes using the new database and blockchain infrastructure along with two concrete indexer implementations which provide both a transaction-by-hash and a transaction-by-address index. The new infrastructure is mostly separated into a package named indexers which is housed under the blockchain package. In order to support this, a new interface named IndexManager has been introduced in the blockchain package which provides methods to be notified when the chain has been initialized and when blocks are connected and disconnected from the main chain. A concrete implementation of an index manager is provided by the new indexers package. The new indexers package also provides a new interface named Indexer which allows the index manager to manage concrete index implementations which conform to the interface. The following is high level overview of the main index infrastructure changes: - Define a new IndexManager interface in the blockchain package and modify the package to make use of the interface when specified - Create a new indexers package - Provides an Index interface which allows concrete indexes to plugin to an index manager - Provides a concrete IndexManager implementation - Handles the lifecycle of all indexes it manages - Tracks the index tips - Handles catching up disabled indexes that have been reenabled - Handles reorgs while the index was disabled - Invokes the appropriate methods for all managed indexes to allow them to index and deindex the blocks and transactions - Implement a transaction-by-hash index - Makes use of internal block IDs to save a significant amount of space and indexing costs over the old transaction index format - Implement a transaction-by-address index - Makes use of a leveling scheme in order to provide a good tradeoff between space required and indexing costs - Supports enabling and disabling indexes at will - Support the ability to drop indexes if they are no longer desired The following is an overview of the btcd changes: - Add a new index logging subsystem - Add new options --txindex and --addrindex in order to enable the optional indexes - NOTE: The transaction index will automatically be enabled when the address index is enabled because it depends on it - Add new options --droptxindex and --dropaddrindex to allow the indexes to be removed - NOTE: The address index will also be removed when the transaction index is dropped because it depends on it - Update getrawtransactions RPC to make use of the transaction index - Reimplement the searchrawtransaction RPC that makes use of the address index - Update sample-btcd.conf to include sample usage for the new optional index flags
2016-02-19 05:51:18 +01:00
mtx = tx.MsgTx()
}
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
// The verbose flag is set, so generate the JSON object and return it.
indexers: Implement optional tx/address indexes. This introduces a new indexing infrastructure for supporting optional indexes using the new database and blockchain infrastructure along with two concrete indexer implementations which provide both a transaction-by-hash and a transaction-by-address index. The new infrastructure is mostly separated into a package named indexers which is housed under the blockchain package. In order to support this, a new interface named IndexManager has been introduced in the blockchain package which provides methods to be notified when the chain has been initialized and when blocks are connected and disconnected from the main chain. A concrete implementation of an index manager is provided by the new indexers package. The new indexers package also provides a new interface named Indexer which allows the index manager to manage concrete index implementations which conform to the interface. The following is high level overview of the main index infrastructure changes: - Define a new IndexManager interface in the blockchain package and modify the package to make use of the interface when specified - Create a new indexers package - Provides an Index interface which allows concrete indexes to plugin to an index manager - Provides a concrete IndexManager implementation - Handles the lifecycle of all indexes it manages - Tracks the index tips - Handles catching up disabled indexes that have been reenabled - Handles reorgs while the index was disabled - Invokes the appropriate methods for all managed indexes to allow them to index and deindex the blocks and transactions - Implement a transaction-by-hash index - Makes use of internal block IDs to save a significant amount of space and indexing costs over the old transaction index format - Implement a transaction-by-address index - Makes use of a leveling scheme in order to provide a good tradeoff between space required and indexing costs - Supports enabling and disabling indexes at will - Support the ability to drop indexes if they are no longer desired The following is an overview of the btcd changes: - Add a new index logging subsystem - Add new options --txindex and --addrindex in order to enable the optional indexes - NOTE: The transaction index will automatically be enabled when the address index is enabled because it depends on it - Add new options --droptxindex and --dropaddrindex to allow the indexes to be removed - NOTE: The address index will also be removed when the transaction index is dropped because it depends on it - Update getrawtransactions RPC to make use of the transaction index - Reimplement the searchrawtransaction RPC that makes use of the address index - Update sample-btcd.conf to include sample usage for the new optional index flags
2016-02-19 05:51:18 +01:00
var blkHeader *wire.BlockHeader
var blkHashStr string
var chainHeight int32
if blkHash != nil {
// Load the raw header bytes.
var headerBytes []byte
err := s.server.db.View(func(dbTx database.Tx) error {
var err error
headerBytes, err = dbTx.FetchBlockHeader(blkHash)
return err
})
if err != nil {
context := "Failed to fetch block header"
return nil, internalRPCError(err.Error(), context)
}
// Deserialize the header.
var header wire.BlockHeader
err = header.Deserialize(bytes.NewReader(headerBytes))
if err != nil {
context := "Failed to deserialize block header"
return nil, internalRPCError(err.Error(), context)
}
blkHeader = &header
blkHashStr = blkHash.String()
chainHeight = s.chain.BestSnapshot().Height
}
rawTxn, err := createTxRawResult(s.server.chainParams, mtx,
txHash.String(), blkHeader, blkHashStr, blkHeight, chainHeight)
if err != nil {
return nil, err
}
return *rawTxn, nil
}
// bigToLEUint256 returns the passed big integer as an unsigned 256-bit integer
// encoded as little-endian bytes. Numbers which are larger than the max
// unsigned 256-bit integer are truncated.
func bigToLEUint256(n *big.Int) [uint256Size]byte {
// Pad or truncate the big-endian big int to correct number of bytes.
nBytes := n.Bytes()
nlen := len(nBytes)
pad := 0
start := 0
if nlen <= uint256Size {
pad = uint256Size - nlen
} else {
start = nlen - uint256Size
}
var buf [uint256Size]byte
copy(buf[pad:], nBytes[start:])
// Reverse the bytes to little endian and return them.
for i := 0; i < uint256Size/2; i++ {
buf[i], buf[uint256Size-1-i] = buf[uint256Size-1-i], buf[i]
}
return buf
}
// reverseUint32Array treats the passed bytes as a series of uint32s and
// reverses the byte order of each uint32. The passed byte slice must be a
// multiple of 4 for a correct result. The passed bytes slice is modified.
func reverseUint32Array(b []byte) {
blen := len(b)
for i := 0; i < blen; i += 4 {
b[i], b[i+3] = b[i+3], b[i]
b[i+1], b[i+2] = b[i+2], b[i+1]
}
}
// handleGetTxOut handles gettxout commands.
func handleGetTxOut(s *rpcServer, cmd interface{}, closeChan <-chan struct{}) (interface{}, error) {
c := cmd.(*btcjson.GetTxOutCmd)
// Convert the provided transaction hash hex to a Hash.
txHash, err := chainhash.NewHashFromStr(c.Txid)
if err != nil {
return nil, rpcDecodeHexError(c.Txid)
}
// If requested and the tx is available in the mempool try to fetch it
// from there, otherwise attempt to fetch from the block database.
var bestBlockHash string
var confirmations int32
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
var txVersion int32
var value int64
var pkScript []byte
var isCoinbase bool
includeMempool := true
if c.IncludeMempool != nil {
includeMempool = *c.IncludeMempool
}
// TODO: This is racy. It should attempt to fetch it directly and check
// the error.
if includeMempool && s.server.txMemPool.HaveTransaction(txHash) {
tx, err := s.server.txMemPool.FetchTransaction(txHash)
if err != nil {
indexers: Implement optional tx/address indexes. This introduces a new indexing infrastructure for supporting optional indexes using the new database and blockchain infrastructure along with two concrete indexer implementations which provide both a transaction-by-hash and a transaction-by-address index. The new infrastructure is mostly separated into a package named indexers which is housed under the blockchain package. In order to support this, a new interface named IndexManager has been introduced in the blockchain package which provides methods to be notified when the chain has been initialized and when blocks are connected and disconnected from the main chain. A concrete implementation of an index manager is provided by the new indexers package. The new indexers package also provides a new interface named Indexer which allows the index manager to manage concrete index implementations which conform to the interface. The following is high level overview of the main index infrastructure changes: - Define a new IndexManager interface in the blockchain package and modify the package to make use of the interface when specified - Create a new indexers package - Provides an Index interface which allows concrete indexes to plugin to an index manager - Provides a concrete IndexManager implementation - Handles the lifecycle of all indexes it manages - Tracks the index tips - Handles catching up disabled indexes that have been reenabled - Handles reorgs while the index was disabled - Invokes the appropriate methods for all managed indexes to allow them to index and deindex the blocks and transactions - Implement a transaction-by-hash index - Makes use of internal block IDs to save a significant amount of space and indexing costs over the old transaction index format - Implement a transaction-by-address index - Makes use of a leveling scheme in order to provide a good tradeoff between space required and indexing costs - Supports enabling and disabling indexes at will - Support the ability to drop indexes if they are no longer desired The following is an overview of the btcd changes: - Add a new index logging subsystem - Add new options --txindex and --addrindex in order to enable the optional indexes - NOTE: The transaction index will automatically be enabled when the address index is enabled because it depends on it - Add new options --droptxindex and --dropaddrindex to allow the indexes to be removed - NOTE: The address index will also be removed when the transaction index is dropped because it depends on it - Update getrawtransactions RPC to make use of the transaction index - Reimplement the searchrawtransaction RPC that makes use of the address index - Update sample-btcd.conf to include sample usage for the new optional index flags
2016-02-19 05:51:18 +01:00
return nil, rpcNoTxInfoError(txHash)
}
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
mtx := tx.MsgTx()
if c.Vout > uint32(len(mtx.TxOut)-1) {
return nil, &btcjson.RPCError{
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
Code: btcjson.ErrRPCInvalidTxVout,
Message: "Ouput index number (vout) does not " +
"exist for transaction.",
}
}
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
txOut := mtx.TxOut[c.Vout]
if txOut == nil {
errStr := fmt.Sprintf("Output index: %d for txid: %s "+
"does not exist", c.Vout, txHash)
return nil, internalRPCError(errStr, "")
}
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
best := s.chain.BestSnapshot()
bestBlockHash = best.Hash.String()
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
confirmations = 0
txVersion = mtx.Version
value = txOut.Value
pkScript = txOut.PkScript
isCoinbase = blockchain.IsCoinBaseTx(mtx)
} else {
entry, err := s.chain.FetchUtxoEntry(txHash)
if err != nil {
indexers: Implement optional tx/address indexes. This introduces a new indexing infrastructure for supporting optional indexes using the new database and blockchain infrastructure along with two concrete indexer implementations which provide both a transaction-by-hash and a transaction-by-address index. The new infrastructure is mostly separated into a package named indexers which is housed under the blockchain package. In order to support this, a new interface named IndexManager has been introduced in the blockchain package which provides methods to be notified when the chain has been initialized and when blocks are connected and disconnected from the main chain. A concrete implementation of an index manager is provided by the new indexers package. The new indexers package also provides a new interface named Indexer which allows the index manager to manage concrete index implementations which conform to the interface. The following is high level overview of the main index infrastructure changes: - Define a new IndexManager interface in the blockchain package and modify the package to make use of the interface when specified - Create a new indexers package - Provides an Index interface which allows concrete indexes to plugin to an index manager - Provides a concrete IndexManager implementation - Handles the lifecycle of all indexes it manages - Tracks the index tips - Handles catching up disabled indexes that have been reenabled - Handles reorgs while the index was disabled - Invokes the appropriate methods for all managed indexes to allow them to index and deindex the blocks and transactions - Implement a transaction-by-hash index - Makes use of internal block IDs to save a significant amount of space and indexing costs over the old transaction index format - Implement a transaction-by-address index - Makes use of a leveling scheme in order to provide a good tradeoff between space required and indexing costs - Supports enabling and disabling indexes at will - Support the ability to drop indexes if they are no longer desired The following is an overview of the btcd changes: - Add a new index logging subsystem - Add new options --txindex and --addrindex in order to enable the optional indexes - NOTE: The transaction index will automatically be enabled when the address index is enabled because it depends on it - Add new options --droptxindex and --dropaddrindex to allow the indexes to be removed - NOTE: The address index will also be removed when the transaction index is dropped because it depends on it - Update getrawtransactions RPC to make use of the transaction index - Reimplement the searchrawtransaction RPC that makes use of the address index - Update sample-btcd.conf to include sample usage for the new optional index flags
2016-02-19 05:51:18 +01:00
return nil, rpcNoTxInfoError(txHash)
}
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
// To match the behavior of the reference client, return nil
// (JSON null) if the transaction output is spent by another
// transaction already in the main chain. Mined transactions
// that are spent by a mempool transaction are not affected by
// this.
if entry == nil || entry.IsOutputSpent(c.Vout) {
return nil, nil
}
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
best := s.chain.BestSnapshot()
bestBlockHash = best.Hash.String()
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
confirmations = 1 + best.Height - entry.BlockHeight()
txVersion = entry.Version()
value = entry.AmountByIndex(c.Vout)
pkScript = entry.PkScriptByIndex(c.Vout)
isCoinbase = entry.IsCoinBase()
}
// Disassemble script into single line printable format.
// The disassembled string will contain [error] inline if the script
// doesn't fully parse, so ignore the error here.
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
disbuf, _ := txscript.DisasmString(pkScript)
// Get further info about the script.
// Ignore the error here since an error means the script couldn't parse
// and there is no additional information about it anyways.
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
scriptClass, addrs, reqSigs, _ := txscript.ExtractPkScriptAddrs(pkScript,
s.server.chainParams)
addresses := make([]string, len(addrs))
for i, addr := range addrs {
addresses[i] = addr.EncodeAddress()
}
txOutReply := &btcjson.GetTxOutResult{
BestBlock: bestBlockHash,
Confirmations: int64(confirmations),
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
Value: btcutil.Amount(value).ToBTC(),
Version: txVersion,
ScriptPubKey: btcjson.ScriptPubKeyResult{
Asm: disbuf,
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
Hex: hex.EncodeToString(pkScript),
ReqSigs: int32(reqSigs),
Type: scriptClass.String(),
Addresses: addresses,
},
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
Coinbase: isCoinbase,
}
return txOutReply, nil
}
// handleGetWorkRequest is a helper for handleGetWork which deals with
// generating and returning work to the caller.
//
// This function MUST be called with the RPC workstate locked.
func handleGetWorkRequest(s *rpcServer) (interface{}, error) {
state := s.workState
// Generate a new block template when the current best block has
// changed or the transactions in the memory pool have been updated
// and it has been at least one minute since the last template was
// generated.
lastTxUpdate := s.server.txMemPool.LastUpdated()
latestHash, latestHeight := s.server.blockManager.chainState.Best()
msgBlock := state.msgBlock
if msgBlock == nil || state.prevHash == nil ||
!state.prevHash.IsEqual(latestHash) ||
(state.lastTxUpdate != lastTxUpdate &&
time.Now().After(state.lastGenerated.Add(time.Minute))) {
// Reset the extra nonce and clear all cached template
// variations if the best block changed.
if state.prevHash != nil && !state.prevHash.IsEqual(latestHash) {
state.extraNonce = 0
state.blockInfo = make(map[chainhash.Hash]*workStateBlockInfo)
}
// Reset the previous best hash the block template was generated
// against so any errors below cause the next invocation to try
// again.
state.prevHash = nil
// Choose a payment address at random.
Implement a built-in concurrent CPU miner. This commit implements a built-in concurrent CPU miner that can be enabled with the combination of the --generate and --miningaddr options. The --blockminsize, --blockmaxsize, and --blockprioritysize configuration options wich already existed prior to this commit control the block template generation and hence affect blocks mined via the new CPU miner. The following is a quick overview of the changes and design: - Starting btcd with --generate and no addresses specified via --miningaddr will give an error and exit immediately - Makes use of multiple worker goroutines which independently create block templates, solve them, and submit the solved blocks - The default number of worker threads are based on the number of processor cores in the system and can be dynamically changed at run-time - There is a separate speed monitor goroutine used to collate periodic updates from the workers to calculate overall hashing speed - The current mining state, number of workers, and hashes per second can be queried - Updated sample-btcd.conf file has been updated to include the coin generation (mining) settings - Updated doc.go for the new command line options In addition the old --getworkkey option is now deprecated in favor of the new --miningaddr option. This was changed for a few reasons: - There is no reason to have a separate list of keys for getwork and CPU mining - getwork is deprecated and will be going away in the future so that means the --getworkkey flag will also be going away - Having the work 'key' in the option can be confused with wanting a private key while --miningaddr make it a little more clear it is an address that is required Closes #137. Reviewed by @jrick.
2014-06-12 03:09:38 +02:00
payToAddr := cfg.miningAddrs[rand.Intn(len(cfg.miningAddrs))]
template, err := NewBlockTemplate(s.policy, s.server, payToAddr)
if err != nil {
context := "Failed to create new block template"
return nil, internalRPCError(err.Error(), context)
}
msgBlock = template.Block
// Update work state to ensure another block template isn't
// generated until needed.
state.msgBlock = msgBlock
state.lastGenerated = time.Now()
state.lastTxUpdate = lastTxUpdate
state.prevHash = latestHash
rpcsLog.Debugf("Generated block template (timestamp %v, extra "+
"nonce %d, target %064x, merkle root %s, signature "+
"script %x)", msgBlock.Header.Timestamp,
state.extraNonce,
blockchain.CompactToBig(msgBlock.Header.Bits),
msgBlock.Header.MerkleRoot,
msgBlock.Transactions[0].TxIn[0].SignatureScript)
} else {
// At this point, there is a saved block template and a new
// request for work was made, but either the available
// transactions haven't change or it hasn't been long enough to
// trigger a new block template to be generated. So, update the
// existing block template and track the variations so each
// variation can be regenerated if a caller finds an answer and
// makes a submission against it.
// Update the time of the block template to the current time
// while accounting for the median time of the past several
// blocks per the chain consensus rules.
UpdateBlockTime(msgBlock, s.server.blockManager)
// Increment the extra nonce and update the block template
// with the new value by regenerating the coinbase script and
// setting the merkle root to the new value.
state.extraNonce++
err := UpdateExtraNonce(msgBlock, latestHeight+1, state.extraNonce)
if err != nil {
errStr := fmt.Sprintf("Failed to update extra nonce: "+
"%v", err)
return nil, internalRPCError(errStr, "")
}
rpcsLog.Debugf("Updated block template (timestamp %v, extra "+
"nonce %d, target %064x, merkle root %s, signature "+
"script %x)", msgBlock.Header.Timestamp,
state.extraNonce,
blockchain.CompactToBig(msgBlock.Header.Bits),
msgBlock.Header.MerkleRoot,
msgBlock.Transactions[0].TxIn[0].SignatureScript)
}
// In order to efficiently store the variations of block templates that
// have been provided to callers, save a pointer to the block as well as
// the modified signature script keyed by the merkle root. This
// information, along with the data that is included in a work
// submission, is used to rebuild the block before checking the
// submitted solution.
coinbaseTx := msgBlock.Transactions[0]
state.blockInfo[msgBlock.Header.MerkleRoot] = &workStateBlockInfo{
msgBlock: msgBlock,
signatureScript: coinbaseTx.TxIn[0].SignatureScript,
}
// Serialize the block header into a buffer large enough to hold the
// the block header and the internal sha256 padding that is added and
// retuned as part of the data below.
data := make([]byte, 0, getworkDataLen)
buf := bytes.NewBuffer(data)
err := msgBlock.Header.Serialize(buf)
if err != nil {
errStr := fmt.Sprintf("Failed to serialize data: %v", err)
return nil, internalRPCError(errStr, "")
}
// Calculate the midstate for the block header. The midstate here is
// the internal state of the sha256 algorithm for the first chunk of the
// block header (sha256 operates on 64-byte chunks) which is before the
// nonce. This allows sophisticated callers to avoid hashing the first
// chunk over and over while iterating the nonce range.
data = data[:buf.Len()]
midstate := fastsha256.MidState256(data)
// Expand the data slice to include the full data buffer and apply the
// internal sha256 padding which consists of a single 1 bit followed
// by enough zeros to pad the message out to 56 bytes followed by the
// length of the message in bits encoded as a big-endian uint64
// (8 bytes). Thus, the resulting length is a multiple of the sha256
// block size (64 bytes). This makes the data ready for sophisticated
// caller to make use of only the second chunk along with the midstate
// for the first chunk.
data = data[:getworkDataLen]
data[wire.MaxBlockHeaderPayload] = 0x80
binary.BigEndian.PutUint64(data[len(data)-8:],
wire.MaxBlockHeaderPayload*8)
// Create the hash1 field which is a zero hash along with the internal
// sha256 padding as described above. This field is really quite
// useless, but it is required for compatibility with the reference
// implementation.
var hash1 [hash1Len]byte
hash1[chainhash.HashSize] = 0x80
binary.BigEndian.PutUint64(hash1[len(hash1)-8:], chainhash.HashSize*8)
2016-02-22 11:35:54 +01:00
// The final result reverses each of the fields to little endian.
// In particular, the data, hash1, and midstate fields are treated as
// arrays of uint32s (per the internal sha256 hashing state) which are
// in big endian, and thus each 4 bytes is byte swapped. The target is
// also in big endian, but it is treated as a uint256 and byte swapped
// to little endian accordingly.
//
// The fact the fields are reversed in this way is rather odd and likey
// an artifact of some legacy internal state in the reference
// implementation, but it is required for compatibility.
reverseUint32Array(data)
reverseUint32Array(hash1[:])
reverseUint32Array(midstate[:])
target := bigToLEUint256(blockchain.CompactToBig(msgBlock.Header.Bits))
reply := &btcjson.GetWorkResult{
Data: hex.EncodeToString(data),
Hash1: hex.EncodeToString(hash1[:]),
Midstate: hex.EncodeToString(midstate[:]),
Target: hex.EncodeToString(target[:]),
}
return reply, nil
}
// handleGetWorkSubmission is a helper for handleGetWork which deals with
// the calling submitting work to be verified and processed.
//
// This function MUST be called with the RPC workstate locked.
func handleGetWorkSubmission(s *rpcServer, hexData string) (interface{}, error) {
// Ensure the provided data is sane.
if len(hexData)%2 != 0 {
hexData = "0" + hexData
}
data, err := hex.DecodeString(hexData)
if err != nil {
return false, rpcDecodeHexError(hexData)
}
if len(data) != getworkDataLen {
return false, &btcjson.RPCError{
Code: btcjson.ErrRPCInvalidParameter,
Message: fmt.Sprintf("Argument must be "+
"%d bytes (not %d)", getworkDataLen,
len(data)),
}
}
// Reverse the data as if it were an array of 32-bit unsigned integers.
// The fact the getwork request and submission data is reversed in this
// way is rather odd and likey an artifact of some legacy internal state
// in the reference implementation, but it is required for
// compatibility.
reverseUint32Array(data)
// Deserialize the block header from the data.
var submittedHeader wire.BlockHeader
bhBuf := bytes.NewReader(data[0:wire.MaxBlockHeaderPayload])
err = submittedHeader.Deserialize(bhBuf)
if err != nil {
return false, &btcjson.RPCError{
Code: btcjson.ErrRPCInvalidParameter,
Message: fmt.Sprintf("Argument does not "+
"contain a valid block header: %v", err),
}
}
// Look up the full block for the provided data based on the
// merkle root. Return false to indicate the solve failed if
// it's not available.
state := s.workState
blockInfo, ok := state.blockInfo[submittedHeader.MerkleRoot]
if !ok {
rpcsLog.Debugf("Block submitted via getwork has no matching "+
"template for merkle root %s",
submittedHeader.MerkleRoot)
return false, nil
}
// Reconstruct the block using the submitted header stored block info.
msgBlock := blockInfo.msgBlock
block := btcutil.NewBlock(msgBlock)
msgBlock.Header.Timestamp = submittedHeader.Timestamp
msgBlock.Header.Nonce = submittedHeader.Nonce
msgBlock.Transactions[0].TxIn[0].SignatureScript = blockInfo.signatureScript
merkles := blockchain.BuildMerkleTreeStore(block.Transactions())
msgBlock.Header.MerkleRoot = *merkles[len(merkles)-1]
// Ensure the submitted block hash is less than the target difficulty.
err = blockchain.CheckProofOfWork(block, activeNetParams.PowLimit)
if err != nil {
// Anything other than a rule violation is an unexpected error,
// so return that error as an internal error.
if _, ok := err.(blockchain.RuleError); !ok {
return false, internalRPCError("Unexpected error "+
"while checking proof of work: "+err.Error(),
"")
}
rpcsLog.Debugf("Block submitted via getwork does not meet "+
"the required proof of work: %v", err)
return false, nil
}
latestHash, _ := s.server.blockManager.chainState.Best()
if !msgBlock.Header.PrevBlock.IsEqual(latestHash) {
rpcsLog.Debugf("Block submitted via getwork with previous "+
"block %s is stale", msgBlock.Header.PrevBlock)
return false, nil
}
// Process this block using the same rules as blocks coming from other
// nodes. This will in turn relay it to the network like normal.
isOrphan, err := s.server.blockManager.ProcessBlock(block, blockchain.BFNone)
if err != nil || isOrphan {
// Anything other than a rule violation is an unexpected error,
// so return that error as an internal error.
if _, ok := err.(blockchain.RuleError); !ok {
return false, internalRPCError("Unexpected error "+
"while processing block: "+err.Error(), "")
}
rpcsLog.Infof("Block submitted via getwork rejected: %v", err)
return false, nil
}
// The block was accepted.
rpcsLog.Infof("Block submitted via getwork accepted: %s", block.Hash())
return true, nil
}
// handleGetWork implements the getwork command.
func handleGetWork(s *rpcServer, cmd interface{}, closeChan <-chan struct{}) (interface{}, error) {
c := cmd.(*btcjson.GetWorkCmd)
Implement a built-in concurrent CPU miner. This commit implements a built-in concurrent CPU miner that can be enabled with the combination of the --generate and --miningaddr options. The --blockminsize, --blockmaxsize, and --blockprioritysize configuration options wich already existed prior to this commit control the block template generation and hence affect blocks mined via the new CPU miner. The following is a quick overview of the changes and design: - Starting btcd with --generate and no addresses specified via --miningaddr will give an error and exit immediately - Makes use of multiple worker goroutines which independently create block templates, solve them, and submit the solved blocks - The default number of worker threads are based on the number of processor cores in the system and can be dynamically changed at run-time - There is a separate speed monitor goroutine used to collate periodic updates from the workers to calculate overall hashing speed - The current mining state, number of workers, and hashes per second can be queried - Updated sample-btcd.conf file has been updated to include the coin generation (mining) settings - Updated doc.go for the new command line options In addition the old --getworkkey option is now deprecated in favor of the new --miningaddr option. This was changed for a few reasons: - There is no reason to have a separate list of keys for getwork and CPU mining - getwork is deprecated and will be going away in the future so that means the --getworkkey flag will also be going away - Having the work 'key' in the option can be confused with wanting a private key while --miningaddr make it a little more clear it is an address that is required Closes #137. Reviewed by @jrick.
2014-06-12 03:09:38 +02:00
// Respond with an error if there are no addresses to pay the created
// blocks to.
Implement a built-in concurrent CPU miner. This commit implements a built-in concurrent CPU miner that can be enabled with the combination of the --generate and --miningaddr options. The --blockminsize, --blockmaxsize, and --blockprioritysize configuration options wich already existed prior to this commit control the block template generation and hence affect blocks mined via the new CPU miner. The following is a quick overview of the changes and design: - Starting btcd with --generate and no addresses specified via --miningaddr will give an error and exit immediately - Makes use of multiple worker goroutines which independently create block templates, solve them, and submit the solved blocks - The default number of worker threads are based on the number of processor cores in the system and can be dynamically changed at run-time - There is a separate speed monitor goroutine used to collate periodic updates from the workers to calculate overall hashing speed - The current mining state, number of workers, and hashes per second can be queried - Updated sample-btcd.conf file has been updated to include the coin generation (mining) settings - Updated doc.go for the new command line options In addition the old --getworkkey option is now deprecated in favor of the new --miningaddr option. This was changed for a few reasons: - There is no reason to have a separate list of keys for getwork and CPU mining - getwork is deprecated and will be going away in the future so that means the --getworkkey flag will also be going away - Having the work 'key' in the option can be confused with wanting a private key while --miningaddr make it a little more clear it is an address that is required Closes #137. Reviewed by @jrick.
2014-06-12 03:09:38 +02:00
if len(cfg.miningAddrs) == 0 {
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCInternal.Code,
Implement a built-in concurrent CPU miner. This commit implements a built-in concurrent CPU miner that can be enabled with the combination of the --generate and --miningaddr options. The --blockminsize, --blockmaxsize, and --blockprioritysize configuration options wich already existed prior to this commit control the block template generation and hence affect blocks mined via the new CPU miner. The following is a quick overview of the changes and design: - Starting btcd with --generate and no addresses specified via --miningaddr will give an error and exit immediately - Makes use of multiple worker goroutines which independently create block templates, solve them, and submit the solved blocks - The default number of worker threads are based on the number of processor cores in the system and can be dynamically changed at run-time - There is a separate speed monitor goroutine used to collate periodic updates from the workers to calculate overall hashing speed - The current mining state, number of workers, and hashes per second can be queried - Updated sample-btcd.conf file has been updated to include the coin generation (mining) settings - Updated doc.go for the new command line options In addition the old --getworkkey option is now deprecated in favor of the new --miningaddr option. This was changed for a few reasons: - There is no reason to have a separate list of keys for getwork and CPU mining - getwork is deprecated and will be going away in the future so that means the --getworkkey flag will also be going away - Having the work 'key' in the option can be confused with wanting a private key while --miningaddr make it a little more clear it is an address that is required Closes #137. Reviewed by @jrick.
2014-06-12 03:09:38 +02:00
Message: "No payment addresses specified via --miningaddr",
}
}
// Return an error if there are no peers connected since there is no
// way to relay a found block or receive transactions to work on.
// However, allow this state when running in the regression test or
// simulation test mode.
if !(cfg.RegressionTest || cfg.SimNet) && s.server.ConnectedCount() == 0 {
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCClientNotConnected,
Message: "Bitcoin is not connected",
}
}
// No point in generating or accepting work before the chain is synced.
_, currentHeight := s.server.blockManager.chainState.Best()
if currentHeight != 0 && !s.server.blockManager.IsCurrent() {
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCClientInInitialDownload,
Message: "Bitcoin is downloading blocks...",
}
}
// Protect concurrent access from multiple RPC invocations for work
// requests and submission.
s.workState.Lock()
defer s.workState.Unlock()
// When the caller provides data, it is a submission of a supposedly
// solved block that needs to be checked and submitted to the network
// if valid.
if c.Data != nil && *c.Data != "" {
return handleGetWorkSubmission(s, *c.Data)
}
// No data was provided, so the caller is requesting work.
return handleGetWorkRequest(s)
}
// handleHelp implements the help command.
func handleHelp(s *rpcServer, cmd interface{}, closeChan <-chan struct{}) (interface{}, error) {
c := cmd.(*btcjson.HelpCmd)
// Provide a usage overview of all commands when no specific command
// was specified.
var command string
if c.Command != nil {
command = *c.Command
}
if command == "" {
usage, err := s.helpCacher.rpcUsage(false)
if err != nil {
context := "Failed to generate RPC usage"
return nil, internalRPCError(err.Error(), context)
}
return usage, nil
}
// Check that the command asked for is supported and implemented. Only
// search the main list of handlers since help should not be provided
// for commands that are unimplemented or related to wallet
// functionality.
if _, ok := rpcHandlers[command]; !ok {
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCInvalidParameter,
Message: "Unknown command: " + command,
}
}
// Get the help for the command.
help, err := s.helpCacher.rpcMethodHelp(command)
if err != nil {
context := "Failed to generate help"
return nil, internalRPCError(err.Error(), context)
}
return help, nil
}
// handlePing implements the ping command.
func handlePing(s *rpcServer, cmd interface{}, closeChan <-chan struct{}) (interface{}, error) {
// Ask server to ping \o_
nonce, err := wire.RandomUint64()
if err != nil {
return nil, internalRPCError("Not sending ping - failed to "+
"generate nonce: "+err.Error(), "")
}
s.server.BroadcastMessage(wire.NewMsgPing(nonce))
return nil, nil
}
indexers: Implement optional tx/address indexes. This introduces a new indexing infrastructure for supporting optional indexes using the new database and blockchain infrastructure along with two concrete indexer implementations which provide both a transaction-by-hash and a transaction-by-address index. The new infrastructure is mostly separated into a package named indexers which is housed under the blockchain package. In order to support this, a new interface named IndexManager has been introduced in the blockchain package which provides methods to be notified when the chain has been initialized and when blocks are connected and disconnected from the main chain. A concrete implementation of an index manager is provided by the new indexers package. The new indexers package also provides a new interface named Indexer which allows the index manager to manage concrete index implementations which conform to the interface. The following is high level overview of the main index infrastructure changes: - Define a new IndexManager interface in the blockchain package and modify the package to make use of the interface when specified - Create a new indexers package - Provides an Index interface which allows concrete indexes to plugin to an index manager - Provides a concrete IndexManager implementation - Handles the lifecycle of all indexes it manages - Tracks the index tips - Handles catching up disabled indexes that have been reenabled - Handles reorgs while the index was disabled - Invokes the appropriate methods for all managed indexes to allow them to index and deindex the blocks and transactions - Implement a transaction-by-hash index - Makes use of internal block IDs to save a significant amount of space and indexing costs over the old transaction index format - Implement a transaction-by-address index - Makes use of a leveling scheme in order to provide a good tradeoff between space required and indexing costs - Supports enabling and disabling indexes at will - Support the ability to drop indexes if they are no longer desired The following is an overview of the btcd changes: - Add a new index logging subsystem - Add new options --txindex and --addrindex in order to enable the optional indexes - NOTE: The transaction index will automatically be enabled when the address index is enabled because it depends on it - Add new options --droptxindex and --dropaddrindex to allow the indexes to be removed - NOTE: The address index will also be removed when the transaction index is dropped because it depends on it - Update getrawtransactions RPC to make use of the transaction index - Reimplement the searchrawtransaction RPC that makes use of the address index - Update sample-btcd.conf to include sample usage for the new optional index flags
2016-02-19 05:51:18 +01:00
// retrievedTx represents a transaction that was either loaded from the
// transaction memory pool or from the database. When a transaction is loaded
// from the database, it is loaded with the raw serialized bytes while the
// mempool has the fully deserialized structure. This structure therefore will
// have one of the two fields set depending on where is was retrieved from.
// This is mainly done for efficiency to avoid extra serialization steps when
// possible.
type retrievedTx struct {
txBytes []byte
blkHash *chainhash.Hash // Only set when transaction is in a block.
indexers: Implement optional tx/address indexes. This introduces a new indexing infrastructure for supporting optional indexes using the new database and blockchain infrastructure along with two concrete indexer implementations which provide both a transaction-by-hash and a transaction-by-address index. The new infrastructure is mostly separated into a package named indexers which is housed under the blockchain package. In order to support this, a new interface named IndexManager has been introduced in the blockchain package which provides methods to be notified when the chain has been initialized and when blocks are connected and disconnected from the main chain. A concrete implementation of an index manager is provided by the new indexers package. The new indexers package also provides a new interface named Indexer which allows the index manager to manage concrete index implementations which conform to the interface. The following is high level overview of the main index infrastructure changes: - Define a new IndexManager interface in the blockchain package and modify the package to make use of the interface when specified - Create a new indexers package - Provides an Index interface which allows concrete indexes to plugin to an index manager - Provides a concrete IndexManager implementation - Handles the lifecycle of all indexes it manages - Tracks the index tips - Handles catching up disabled indexes that have been reenabled - Handles reorgs while the index was disabled - Invokes the appropriate methods for all managed indexes to allow them to index and deindex the blocks and transactions - Implement a transaction-by-hash index - Makes use of internal block IDs to save a significant amount of space and indexing costs over the old transaction index format - Implement a transaction-by-address index - Makes use of a leveling scheme in order to provide a good tradeoff between space required and indexing costs - Supports enabling and disabling indexes at will - Support the ability to drop indexes if they are no longer desired The following is an overview of the btcd changes: - Add a new index logging subsystem - Add new options --txindex and --addrindex in order to enable the optional indexes - NOTE: The transaction index will automatically be enabled when the address index is enabled because it depends on it - Add new options --droptxindex and --dropaddrindex to allow the indexes to be removed - NOTE: The address index will also be removed when the transaction index is dropped because it depends on it - Update getrawtransactions RPC to make use of the transaction index - Reimplement the searchrawtransaction RPC that makes use of the address index - Update sample-btcd.conf to include sample usage for the new optional index flags
2016-02-19 05:51:18 +01:00
tx *btcutil.Tx
}
// fetchInputTxos fetches the outpoints from all transactions referenced by the
// inputs to the passed transaction by checking the transaction mempool first
// then the transaction index for those already mined into blocks.
func fetchInputTxos(s *rpcServer, tx *wire.MsgTx) (map[wire.OutPoint]wire.TxOut, error) {
mp := s.server.txMemPool
originOutputs := make(map[wire.OutPoint]wire.TxOut)
for txInIndex, txIn := range tx.TxIn {
// Attempt to fetch and use the referenced transaction from the
// memory pool.
origin := &txIn.PreviousOutPoint
originTx, err := mp.FetchTransaction(&origin.Hash)
if err == nil {
txOuts := originTx.MsgTx().TxOut
if origin.Index >= uint32(len(txOuts)) {
errStr := fmt.Sprintf("unable to find output "+
"%v referenced from transaction %s:%d",
origin, tx.TxHash(), txInIndex)
indexers: Implement optional tx/address indexes. This introduces a new indexing infrastructure for supporting optional indexes using the new database and blockchain infrastructure along with two concrete indexer implementations which provide both a transaction-by-hash and a transaction-by-address index. The new infrastructure is mostly separated into a package named indexers which is housed under the blockchain package. In order to support this, a new interface named IndexManager has been introduced in the blockchain package which provides methods to be notified when the chain has been initialized and when blocks are connected and disconnected from the main chain. A concrete implementation of an index manager is provided by the new indexers package. The new indexers package also provides a new interface named Indexer which allows the index manager to manage concrete index implementations which conform to the interface. The following is high level overview of the main index infrastructure changes: - Define a new IndexManager interface in the blockchain package and modify the package to make use of the interface when specified - Create a new indexers package - Provides an Index interface which allows concrete indexes to plugin to an index manager - Provides a concrete IndexManager implementation - Handles the lifecycle of all indexes it manages - Tracks the index tips - Handles catching up disabled indexes that have been reenabled - Handles reorgs while the index was disabled - Invokes the appropriate methods for all managed indexes to allow them to index and deindex the blocks and transactions - Implement a transaction-by-hash index - Makes use of internal block IDs to save a significant amount of space and indexing costs over the old transaction index format - Implement a transaction-by-address index - Makes use of a leveling scheme in order to provide a good tradeoff between space required and indexing costs - Supports enabling and disabling indexes at will - Support the ability to drop indexes if they are no longer desired The following is an overview of the btcd changes: - Add a new index logging subsystem - Add new options --txindex and --addrindex in order to enable the optional indexes - NOTE: The transaction index will automatically be enabled when the address index is enabled because it depends on it - Add new options --droptxindex and --dropaddrindex to allow the indexes to be removed - NOTE: The address index will also be removed when the transaction index is dropped because it depends on it - Update getrawtransactions RPC to make use of the transaction index - Reimplement the searchrawtransaction RPC that makes use of the address index - Update sample-btcd.conf to include sample usage for the new optional index flags
2016-02-19 05:51:18 +01:00
return nil, internalRPCError(errStr, "")
}
originOutputs[*origin] = *txOuts[origin.Index]
continue
}
// Look up the location of the transaction.
blockRegion, err := s.server.txIndex.TxBlockRegion(&origin.Hash)
if err != nil {
context := "Failed to retrieve transaction location"
return nil, internalRPCError(err.Error(), context)
}
if blockRegion == nil {
return nil, rpcNoTxInfoError(&origin.Hash)
}
// Load the raw transaction bytes from the database.
var txBytes []byte
err = s.server.db.View(func(dbTx database.Tx) error {
var err error
txBytes, err = dbTx.FetchBlockRegion(blockRegion)
return err
})
if err != nil {
return nil, rpcNoTxInfoError(&origin.Hash)
}
// Deserialize the transaction
var msgTx wire.MsgTx
err = msgTx.Deserialize(bytes.NewReader(txBytes))
if err != nil {
context := "Failed to deserialize transaction"
return nil, internalRPCError(err.Error(), context)
}
// Add the referenced output to the map.
if origin.Index >= uint32(len(msgTx.TxOut)) {
errStr := fmt.Sprintf("unable to find output %v "+
"referenced from transaction %s:%d", origin,
tx.TxHash(), txInIndex)
indexers: Implement optional tx/address indexes. This introduces a new indexing infrastructure for supporting optional indexes using the new database and blockchain infrastructure along with two concrete indexer implementations which provide both a transaction-by-hash and a transaction-by-address index. The new infrastructure is mostly separated into a package named indexers which is housed under the blockchain package. In order to support this, a new interface named IndexManager has been introduced in the blockchain package which provides methods to be notified when the chain has been initialized and when blocks are connected and disconnected from the main chain. A concrete implementation of an index manager is provided by the new indexers package. The new indexers package also provides a new interface named Indexer which allows the index manager to manage concrete index implementations which conform to the interface. The following is high level overview of the main index infrastructure changes: - Define a new IndexManager interface in the blockchain package and modify the package to make use of the interface when specified - Create a new indexers package - Provides an Index interface which allows concrete indexes to plugin to an index manager - Provides a concrete IndexManager implementation - Handles the lifecycle of all indexes it manages - Tracks the index tips - Handles catching up disabled indexes that have been reenabled - Handles reorgs while the index was disabled - Invokes the appropriate methods for all managed indexes to allow them to index and deindex the blocks and transactions - Implement a transaction-by-hash index - Makes use of internal block IDs to save a significant amount of space and indexing costs over the old transaction index format - Implement a transaction-by-address index - Makes use of a leveling scheme in order to provide a good tradeoff between space required and indexing costs - Supports enabling and disabling indexes at will - Support the ability to drop indexes if they are no longer desired The following is an overview of the btcd changes: - Add a new index logging subsystem - Add new options --txindex and --addrindex in order to enable the optional indexes - NOTE: The transaction index will automatically be enabled when the address index is enabled because it depends on it - Add new options --droptxindex and --dropaddrindex to allow the indexes to be removed - NOTE: The address index will also be removed when the transaction index is dropped because it depends on it - Update getrawtransactions RPC to make use of the transaction index - Reimplement the searchrawtransaction RPC that makes use of the address index - Update sample-btcd.conf to include sample usage for the new optional index flags
2016-02-19 05:51:18 +01:00
return nil, internalRPCError(errStr, "")
}
originOutputs[*origin] = *msgTx.TxOut[origin.Index]
}
return originOutputs, nil
}
// createVinListPrevOut returns a slice of JSON objects for the inputs of the
// passed transaction.
func createVinListPrevOut(s *rpcServer, mtx *wire.MsgTx, chainParams *chaincfg.Params, vinExtra bool, filterAddrMap map[string]struct{}) ([]btcjson.VinPrevOut, error) {
// Coinbase transactions only have a single txin by definition.
if blockchain.IsCoinBaseTx(mtx) {
// Only include the transaction if the filter map is empty
// because a coinbase input has no addresses and so would never
// match a non-empty filter.
if len(filterAddrMap) != 0 {
return nil, nil
}
txIn := mtx.TxIn[0]
vinList := make([]btcjson.VinPrevOut, 1)
vinList[0].Coinbase = hex.EncodeToString(txIn.SignatureScript)
vinList[0].Sequence = txIn.Sequence
return vinList, nil
}
// Use a dynamically sized list to accomodate the address filter.
vinList := make([]btcjson.VinPrevOut, 0, len(mtx.TxIn))
// Lookup all of the referenced transaction outputs needed to populate
// the previous output information if requested.
var originOutputs map[wire.OutPoint]wire.TxOut
if vinExtra || len(filterAddrMap) > 0 {
var err error
originOutputs, err = fetchInputTxos(s, mtx)
if err != nil {
return nil, err
}
}
for _, txIn := range mtx.TxIn {
// The disassembled string will contain [error] inline
// if the script doesn't fully parse, so ignore the
// error here.
disbuf, _ := txscript.DisasmString(txIn.SignatureScript)
// Create the basic input entry without the additional optional
// previous output details which will be added later if
// requested and available.
prevOut := &txIn.PreviousOutPoint
vinEntry := btcjson.VinPrevOut{
Txid: prevOut.Hash.String(),
Vout: prevOut.Index,
Sequence: txIn.Sequence,
ScriptSig: &btcjson.ScriptSig{
Asm: disbuf,
Hex: hex.EncodeToString(txIn.SignatureScript),
},
}
// Add the entry to the list now if it already passed the filter
// since the previous output might not be available.
passesFilter := len(filterAddrMap) == 0
if passesFilter {
vinList = append(vinList, vinEntry)
}
// Only populate previous output information if requested and
// available.
if len(originOutputs) == 0 {
continue
}
originTxOut, ok := originOutputs[*prevOut]
if !ok {
continue
}
// Ignore the error here since an error means the script
// couldn't parse and there is no additional information about
// it anyways.
_, addrs, _, _ := txscript.ExtractPkScriptAddrs(
originTxOut.PkScript, chainParams)
// Encode the addresses while checking if the address passes the
// filter when needed.
encodedAddrs := make([]string, len(addrs))
for j, addr := range addrs {
encodedAddr := addr.EncodeAddress()
encodedAddrs[j] = encodedAddr
// No need to check the map again if the filter already
// passes.
if passesFilter {
continue
}
if _, exists := filterAddrMap[encodedAddr]; exists {
passesFilter = true
}
}
// Ignore the entry if it doesn't pass the filter.
if !passesFilter {
continue
}
// Add entry to the list if it wasn't already done above.
if len(filterAddrMap) != 0 {
vinList = append(vinList, vinEntry)
}
// Update the entry with previous output information if
// requested.
if vinExtra {
vinListEntry := &vinList[len(vinList)-1]
vinListEntry.PrevOut = &btcjson.PrevOut{
Addresses: encodedAddrs,
Value: btcutil.Amount(originTxOut.Value).ToBTC(),
}
}
}
return vinList, nil
}
// fetchMempoolTxnsForAddress queries the address index for all unconfirmed
// transactions that involve the provided address. The results will be limited
// by the number to skip and the number requested.
func fetchMempoolTxnsForAddress(s *rpcServer, addr btcutil.Address, numToSkip, numRequested uint32) ([]*btcutil.Tx, uint32) {
// There are no entries to return when there are less available than the
// number being skipped.
mpTxns := s.server.addrIndex.UnconfirmedTxnsForAddress(addr)
numAvailable := uint32(len(mpTxns))
if numToSkip > numAvailable {
return nil, numAvailable
}
// Filter the available entries based on the number to skip and number
// requested.
rangeEnd := numToSkip + numRequested
if rangeEnd > numAvailable {
rangeEnd = numAvailable
}
return mpTxns[numToSkip:rangeEnd], numToSkip
}
// handleSearchRawTransactions implements the searchrawtransactions command.
func handleSearchRawTransactions(s *rpcServer, cmd interface{}, closeChan <-chan struct{}) (interface{}, error) {
// Respond with an error if the address index is not enabled.
addrIndex := s.server.addrIndex
if addrIndex == nil {
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCMisc,
Message: "Address index must be enabled (--addrindex)",
}
}
// Override the flag for including extra previous output information in
// each input if needed.
c := cmd.(*btcjson.SearchRawTransactionsCmd)
vinExtra := false
if c.VinExtra != nil {
vinExtra = *c.VinExtra != 0
}
// Including the extra previous output information requires the
// transaction index. Currently the address index relies on the
// transaction index, so this check is redundant, but it's better to be
// safe in case the address index is ever changed to not rely on it.
if vinExtra && s.server.txIndex == nil {
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCMisc,
Message: "Transaction index must be enabled (--txindex)",
}
}
// Attempt to decode the supplied address.
addr, err := btcutil.DecodeAddress(c.Address, s.server.chainParams)
if err != nil {
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCInvalidAddressOrKey,
Message: "Invalid address or key: " + err.Error(),
}
}
// Override the default number of requested entries if needed. Also,
// just return now if the number of requested entries is zero to avoid
// extra work.
numRequested := 100
if c.Count != nil {
numRequested = *c.Count
if numRequested < 0 {
numRequested = 1
}
}
if numRequested == 0 {
return nil, nil
}
// Override the default number of entries to skip if needed.
var numToSkip int
if c.Skip != nil {
numToSkip = *c.Skip
if numToSkip < 0 {
numToSkip = 0
}
}
// Override the reverse flag if needed.
var reverse bool
if c.Reverse != nil {
reverse = *c.Reverse
}
// Add transactions from mempool first if client asked for reverse
// order. Otherwise, they will be added last (as needed depending on
// the requested counts).
//
// NOTE: This code doesn't sort by dependency. This might be something
// to do in the future for the client's convenience, or leave it to the
// client.
numSkipped := uint32(0)
addressTxns := make([]retrievedTx, 0, numRequested)
if reverse {
// Transactions in the mempool are not in a block header yet,
// so the block header field in the retieved transaction struct
// is left nil.
mpTxns, mpSkipped := fetchMempoolTxnsForAddress(s, addr,
uint32(numToSkip), uint32(numRequested))
numSkipped += mpSkipped
for _, tx := range mpTxns {
addressTxns = append(addressTxns, retrievedTx{tx: tx})
}
}
// Fetch transactions from the database in the desired order if more are
// needed.
if len(addressTxns) < numRequested {
err = s.server.db.View(func(dbTx database.Tx) error {
regions, dbSkipped, err := addrIndex.TxRegionsForAddress(
dbTx, addr, uint32(numToSkip)-numSkipped,
uint32(numRequested-len(addressTxns)), reverse)
if err != nil {
return err
}
// Load the raw transaction bytes from the database.
serializedTxns, err := dbTx.FetchBlockRegions(regions)
if err != nil {
return err
}
// Add the transaction and the hash of the block it is
// contained in to the list. Note that the transaction
// is left serialized here since the caller might have
// requested non-verbose output and hence there would be
// no point in deserializing it just to reserialize it
// later.
for i, serializedTx := range serializedTxns {
addressTxns = append(addressTxns, retrievedTx{
txBytes: serializedTx,
blkHash: regions[i].Hash,
})
}
numSkipped += dbSkipped
return nil
})
if err != nil {
context := "Failed to load address index entries"
return nil, internalRPCError(err.Error(), context)
}
}
// Add transactions from mempool last if client did not request reverse
// order and the number of results is still under the number requested.
if !reverse && len(addressTxns) < numRequested {
// Transactions in the mempool are not in a block header yet,
// so the block header field in the retieved transaction struct
// is left nil.
mpTxns, mpSkipped := fetchMempoolTxnsForAddress(s, addr,
uint32(numToSkip)-numSkipped, uint32(numRequested-
len(addressTxns)))
numSkipped += mpSkipped
for _, tx := range mpTxns {
addressTxns = append(addressTxns, retrievedTx{tx: tx})
}
}
// Address has never been used if neither source yielded any results.
if len(addressTxns) == 0 {
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCNoTxInfo,
Message: "No information available about address",
}
}
// Serialize all of the transactions to hex.
hexTxns := make([]string, len(addressTxns))
for i := range addressTxns {
// Simply encode the raw bytes to hex when the retrieved
// transaction is already in serialized form.
rtx := &addressTxns[i]
if rtx.txBytes != nil {
hexTxns[i] = hex.EncodeToString(rtx.txBytes)
continue
}
// Serialize the transaction first and convert to hex when the
// retrieved transaction is the deserialized structure.
hexTxns[i], err = messageToHex(rtx.tx.MsgTx())
if err != nil {
return nil, err
}
}
// When not in verbose mode, simply return a list of serialized txns.
if c.Verbose != nil && *c.Verbose == 0 {
return hexTxns, nil
}
// Normalize the provided filter addresses (if any) to ensure there are
// no duplicates.
filterAddrMap := make(map[string]struct{})
if c.FilterAddrs != nil && len(*c.FilterAddrs) > 0 {
for _, addr := range *c.FilterAddrs {
filterAddrMap[addr] = struct{}{}
}
}
// The verbose flag is set, so generate the JSON object and return it.
best := s.chain.BestSnapshot()
chainParams := s.server.chainParams
srtList := make([]btcjson.SearchRawTransactionsResult, len(addressTxns))
for i := range addressTxns {
// The deserialized transaction is needed, so deserialize the
// retrieved transaction if it's in serialized form (which will
// be the case when it was lookup up from the database).
// Otherwise, use the existing deserialized transaction.
rtx := &addressTxns[i]
var mtx *wire.MsgTx
if rtx.tx == nil {
// Deserialize the transaction.
mtx = new(wire.MsgTx)
err := mtx.Deserialize(bytes.NewReader(rtx.txBytes))
if err != nil {
context := "Failed to deserialize transaction"
return nil, internalRPCError(err.Error(),
context)
}
} else {
mtx = rtx.tx.MsgTx()
}
result := &srtList[i]
result.Hex = hexTxns[i]
result.Txid = mtx.TxHash().String()
indexers: Implement optional tx/address indexes. This introduces a new indexing infrastructure for supporting optional indexes using the new database and blockchain infrastructure along with two concrete indexer implementations which provide both a transaction-by-hash and a transaction-by-address index. The new infrastructure is mostly separated into a package named indexers which is housed under the blockchain package. In order to support this, a new interface named IndexManager has been introduced in the blockchain package which provides methods to be notified when the chain has been initialized and when blocks are connected and disconnected from the main chain. A concrete implementation of an index manager is provided by the new indexers package. The new indexers package also provides a new interface named Indexer which allows the index manager to manage concrete index implementations which conform to the interface. The following is high level overview of the main index infrastructure changes: - Define a new IndexManager interface in the blockchain package and modify the package to make use of the interface when specified - Create a new indexers package - Provides an Index interface which allows concrete indexes to plugin to an index manager - Provides a concrete IndexManager implementation - Handles the lifecycle of all indexes it manages - Tracks the index tips - Handles catching up disabled indexes that have been reenabled - Handles reorgs while the index was disabled - Invokes the appropriate methods for all managed indexes to allow them to index and deindex the blocks and transactions - Implement a transaction-by-hash index - Makes use of internal block IDs to save a significant amount of space and indexing costs over the old transaction index format - Implement a transaction-by-address index - Makes use of a leveling scheme in order to provide a good tradeoff between space required and indexing costs - Supports enabling and disabling indexes at will - Support the ability to drop indexes if they are no longer desired The following is an overview of the btcd changes: - Add a new index logging subsystem - Add new options --txindex and --addrindex in order to enable the optional indexes - NOTE: The transaction index will automatically be enabled when the address index is enabled because it depends on it - Add new options --droptxindex and --dropaddrindex to allow the indexes to be removed - NOTE: The address index will also be removed when the transaction index is dropped because it depends on it - Update getrawtransactions RPC to make use of the transaction index - Reimplement the searchrawtransaction RPC that makes use of the address index - Update sample-btcd.conf to include sample usage for the new optional index flags
2016-02-19 05:51:18 +01:00
result.Vin, err = createVinListPrevOut(s, mtx, chainParams,
vinExtra, filterAddrMap)
if err != nil {
return nil, err
}
result.Vout = createVoutList(mtx, chainParams, filterAddrMap)
result.Version = mtx.Version
result.LockTime = mtx.LockTime
// Transactions grabbed from the mempool aren't yet in a block,
// so conditionally fetch block details here. This will be
// reflected in the final JSON output (mempool won't have
// confirmations or block information).
var blkHeader *wire.BlockHeader
var blkHashStr string
var blkHeight int32
if blkHash := rtx.blkHash; blkHash != nil {
// Load the raw header bytes from the database.
var headerBytes []byte
err := s.server.db.View(func(dbTx database.Tx) error {
var err error
headerBytes, err = dbTx.FetchBlockHeader(blkHash)
return err
})
if err != nil {
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCBlockNotFound,
Message: "Block not found",
}
}
// Deserialize the block header.
var header wire.BlockHeader
err = header.Deserialize(bytes.NewReader(headerBytes))
if err != nil {
context := "Failed to deserialize block header"
return nil, internalRPCError(err.Error(), context)
}
// Get the block height from chain.
height, err := s.chain.BlockHeightByHash(blkHash)
if err != nil {
context := "Failed to obtain block height"
return nil, internalRPCError(err.Error(), context)
}
blkHeader = &header
blkHashStr = blkHash.String()
blkHeight = height
}
// Add the block information to the result if there is any.
if blkHeader != nil {
// This is not a typo, they are identical in Bitcoin
// Core as well.
result.Time = blkHeader.Timestamp.Unix()
result.Blocktime = blkHeader.Timestamp.Unix()
result.BlockHash = blkHashStr
result.Confirmations = uint64(1 + best.Height - blkHeight)
}
}
return srtList, nil
}
// handleSendRawTransaction implements the sendrawtransaction command.
func handleSendRawTransaction(s *rpcServer, cmd interface{}, closeChan <-chan struct{}) (interface{}, error) {
c := cmd.(*btcjson.SendRawTransactionCmd)
// Deserialize and send off to tx relay
hexStr := c.HexTx
if len(hexStr)%2 != 0 {
hexStr = "0" + hexStr
}
serializedTx, err := hex.DecodeString(hexStr)
if err != nil {
return nil, rpcDecodeHexError(hexStr)
}
msgtx := wire.NewMsgTx()
err = msgtx.Deserialize(bytes.NewReader(serializedTx))
if err != nil {
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCDeserialization,
Message: "TX decode failed: " + err.Error(),
}
}
tx := btcutil.NewTx(msgtx)
acceptedTxs, err := s.server.txMemPool.ProcessTransaction(tx, false, false)
if err != nil {
// When the error is a rule error, it means the transaction was
// simply rejected as opposed to something actually going wrong,
// so log it as such. Otherwise, something really did go wrong,
// so log it as an actual error. In both cases, a JSON-RPC
// error is returned to the client with the deserialization
// error code (to match bitcoind behavior).
if _, ok := err.(RuleError); ok {
rpcsLog.Debugf("Rejected transaction %v: %v", tx.Hash(),
err)
} else {
rpcsLog.Errorf("Failed to process transaction %v: %v",
tx.Hash(), err)
}
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCDeserialization,
Message: "TX rejected: " + err.Error(),
}
}
s.server.AnnounceNewTransactions(acceptedTxs)
// Keep track of all the sendrawtransaction request txns so that they
// can be rebroadcast if they don't make their way into a block.
iv := wire.NewInvVect(wire.InvTypeTx, tx.Hash())
s.server.AddRebroadcastInventory(iv, tx)
return tx.Hash().String(), nil
}
// handleSetGenerate implements the setgenerate command.
func handleSetGenerate(s *rpcServer, cmd interface{}, closeChan <-chan struct{}) (interface{}, error) {
c := cmd.(*btcjson.SetGenerateCmd)
// Disable generation regardless of the provided generate flag if the
// maximum number of threads (goroutines for our purposes) is 0.
// Otherwise enable or disable it depending on the provided flag.
generate := c.Generate
genProcLimit := -1
if c.GenProcLimit != nil {
genProcLimit = *c.GenProcLimit
}
if genProcLimit == 0 {
generate = false
}
if !generate {
s.server.cpuMiner.Stop()
} else {
// Respond with an error if there are no addresses to pay the
// created blocks to.
if len(cfg.miningAddrs) == 0 {
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCInternal.Code,
Message: "No payment addresses specified " +
"via --miningaddr",
}
}
// It's safe to call start even if it's already started.
s.server.cpuMiner.SetNumWorkers(int32(genProcLimit))
s.server.cpuMiner.Start()
}
return nil, nil
}
// handleStop implements the stop command.
func handleStop(s *rpcServer, cmd interface{}, closeChan <-chan struct{}) (interface{}, error) {
select {
case s.requestProcessShutdown <- struct{}{}:
default:
}
return "btcd stopping.", nil
}
// handleSubmitBlock implements the submitblock command.
func handleSubmitBlock(s *rpcServer, cmd interface{}, closeChan <-chan struct{}) (interface{}, error) {
c := cmd.(*btcjson.SubmitBlockCmd)
// Deserialize the submitted block.
hexStr := c.HexBlock
if len(hexStr)%2 != 0 {
hexStr = "0" + c.HexBlock
}
serializedBlock, err := hex.DecodeString(hexStr)
if err != nil {
return nil, rpcDecodeHexError(hexStr)
}
block, err := btcutil.NewBlockFromBytes(serializedBlock)
if err != nil {
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCDeserialization,
Message: "Block decode failed: " + err.Error(),
}
}
_, err = s.server.blockManager.ProcessBlock(block, blockchain.BFNone)
if err != nil {
return fmt.Sprintf("rejected: %s", err.Error()), nil
}
rpcsLog.Infof("Accepted block %s via submitblock", block.Hash())
return nil, nil
}
// handleValidateAddress implements the validateaddress command.
func handleValidateAddress(s *rpcServer, cmd interface{}, closeChan <-chan struct{}) (interface{}, error) {
c := cmd.(*btcjson.ValidateAddressCmd)
result := btcjson.ValidateAddressChainResult{}
addr, err := btcutil.DecodeAddress(c.Address, activeNetParams.Params)
if err != nil {
// Return the default value (false) for IsValid.
return result, nil
}
result.Address = addr.EncodeAddress()
result.IsValid = true
return result, nil
}
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
func verifyChain(s *rpcServer, level, depth int32) error {
best := s.chain.BestSnapshot()
finishHeight := best.Height - depth
if finishHeight < 0 {
finishHeight = 0
}
rpcsLog.Infof("Verifying chain for %d blocks at level %d",
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
best.Height-finishHeight, level)
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
for height := best.Height; height > finishHeight; height-- {
// Level 0 just looks up the block.
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
block, err := s.chain.BlockByHeight(height)
if err != nil {
rpcsLog.Errorf("Verify is unable to fetch block at "+
"height %d: %v", height, err)
return err
}
// Level 1 does basic chain sanity checks.
if level > 0 {
err := blockchain.CheckBlockSanity(block,
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
activeNetParams.PowLimit, s.server.timeSource)
if err != nil {
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
rpcsLog.Errorf("Verify is unable to validate "+
"block at hash %v height %d: %v",
block.Hash(), height, err)
return err
}
}
}
rpcsLog.Infof("Chain verify completed successfully")
return nil
}
// handleVerifyChain implements the verifychain command.
func handleVerifyChain(s *rpcServer, cmd interface{}, closeChan <-chan struct{}) (interface{}, error) {
c := cmd.(*btcjson.VerifyChainCmd)
var checkLevel, checkDepth int32
if c.CheckLevel != nil {
checkLevel = *c.CheckLevel
}
if c.CheckDepth != nil {
checkDepth = *c.CheckDepth
}
blockchain: Rework to use new db interface. This commit is the first stage of several that are planned to convert the blockchain package into a concurrent safe package that will ultimately allow support for multi-peer download and concurrent chain processing. The goal is to update btcd proper after each step so it can take advantage of the enhancements as they are developed. In addition to the aforementioned benefit, this staged approach has been chosen since it is absolutely critical to maintain consensus. Separating the changes into several stages makes it easier for reviewers to logically follow what is happening and therefore helps prevent consensus bugs. Naturally there are significant automated tests to help prevent consensus issues as well. The main focus of this stage is to convert the blockchain package to use the new database interface and implement the chain-related functionality which it no longer handles. It also aims to improve efficiency in various areas by making use of the new database and chain capabilities. The following is an overview of the chain changes: - Update to use the new database interface - Add chain-related functionality that the old database used to handle - Main chain structure and state - Transaction spend tracking - Implement a new pruned unspent transaction output (utxo) set - Provides efficient direct access to the unspent transaction outputs - Uses a domain specific compression algorithm that understands the standard transaction scripts in order to significantly compress them - Removes reliance on the transaction index and paves the way toward eventually enabling block pruning - Modify the New function to accept a Config struct instead of inidividual parameters - Replace the old TxStore type with a new UtxoViewpoint type that makes use of the new pruned utxo set - Convert code to treat the new UtxoViewpoint as a rolling view that is used between connects and disconnects to improve efficiency - Make best chain state always set when the chain instance is created - Remove now unnecessary logic for dealing with unset best state - Make all exported functions concurrent safe - Currently using a single chain state lock as it provides a straight forward and easy to review path forward however this can be improved with more fine grained locking - Optimize various cases where full blocks were being loaded when only the header is needed to help reduce the I/O load - Add the ability for callers to get a snapshot of the current best chain stats in a concurrent safe fashion - Does not block callers while new blocks are being processed - Make error messages that reference transaction outputs consistently use <transaction hash>:<output index> - Introduce a new AssertError type an convert internal consistency checks to use it - Update tests and examples to reflect the changes - Add a full suite of tests to ensure correct functionality of the new code The following is an overview of the btcd changes: - Update to use the new database and chain interfaces - Temporarily remove all code related to the transaction index - Temporarily remove all code related to the address index - Convert all code that uses transaction stores to use the new utxo view - Rework several calls that required the block manager for safe concurrency to use the chain package directly now that it is concurrent safe - Change all calls to obtain the best hash to use the new best state snapshot capability from the chain package - Remove workaround for limits on fetching height ranges since the new database interface no longer imposes them - Correct the gettxout RPC handler to return the best chain hash as opposed the hash the txout was found in - Optimize various RPC handlers: - Change several of the RPC handlers to use the new chain snapshot capability to avoid needlessly loading data - Update several handlers to use new functionality to avoid accessing the block manager so they are able to return the data without blocking when the server is busy processing blocks - Update non-verbose getblock to avoid deserialization and serialization overhead - Update getblockheader to request the block height directly from chain and only load the header - Update getdifficulty to use the new cached data from chain - Update getmininginfo to use the new cached data from chain - Update non-verbose getrawtransaction to avoid deserialization and serialization overhead - Update gettxout to use the new utxo store versus loading full transactions using the transaction index The following is an overview of the utility changes: - Update addblock to use the new database and chain interfaces - Update findcheckpoint to use the new database and chain interfaces - Remove the dropafter utility which is no longer supported NOTE: The transaction index and address index will be reimplemented in another commit.
2015-08-26 06:03:18 +02:00
err := verifyChain(s, checkLevel, checkDepth)
return err == nil, nil
}
// handleVerifyMessage implements the verifymessage command.
func handleVerifyMessage(s *rpcServer, cmd interface{}, closeChan <-chan struct{}) (interface{}, error) {
c := cmd.(*btcjson.VerifyMessageCmd)
// Decode the provided address.
addr, err := btcutil.DecodeAddress(c.Address, activeNetParams.Params)
if err != nil {
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCInvalidAddressOrKey,
Message: "Invalid address or key: " + err.Error(),
}
}
// Only P2PKH addresses are valid for signing.
if _, ok := addr.(*btcutil.AddressPubKeyHash); !ok {
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCType,
Message: "Address is not a pay-to-pubkey-hash address",
}
}
// Decode base64 signature.
sig, err := base64.StdEncoding.DecodeString(c.Signature)
if err != nil {
return nil, &btcjson.RPCError{
Code: btcjson.ErrRPCParse.Code,
Message: "Malformed base64 encoding: " + err.Error(),
}
}
// Validate the signature - this just shows that it was valid at all.
// we will compare it with the key next.
var buf bytes.Buffer
wire.WriteVarString(&buf, 0, "Bitcoin Signed Message:\n")
wire.WriteVarString(&buf, 0, c.Message)
expectedMessageHash := chainhash.DoubleHashB(buf.Bytes())
pk, wasCompressed, err := btcec.RecoverCompact(btcec.S256(), sig,
expectedMessageHash)
if err != nil {
// Mirror Bitcoin Core behavior, which treats error in
// RecoverCompact as invalid signature.
return false, nil
}
// Reconstruct the pubkey hash.
btcPK := (*btcec.PublicKey)(pk)
var serializedPK []byte
if wasCompressed {
serializedPK = btcPK.SerializeCompressed()
} else {
serializedPK = btcPK.SerializeUncompressed()
}
address, err := btcutil.NewAddressPubKey(serializedPK,
activeNetParams.Params)
if err != nil {
// Again mirror Bitcoin Core behavior, which treats error in public key
// reconstruction as invalid signature.
return false, nil
}
// Return boolean if addresses match.
return address.EncodeAddress() == c.Address, nil
}
// rpcServer holds the items the rpc server may need to access (config,
// shutdown, main server, etc.)
type rpcServer struct {
started int32
shutdown int32
policy *mining.Policy
server *server
chain *blockchain.BlockChain
authsha [fastsha256.Size]byte
limitauthsha [fastsha256.Size]byte
ntfnMgr *wsNotificationManager
numClients int32
statusLines map[int]string
statusLock sync.RWMutex
wg sync.WaitGroup
listeners []net.Listener
workState *workState
gbtWorkState *gbtWorkState
helpCacher *helpCacher
requestProcessShutdown chan struct{}
quit chan int
}
// httpStatusLine returns a response Status-Line (RFC 2616 Section 6.1)
// for the given request and response status code. This function was lifted and
// adapted from the standard library HTTP server code since it's not exported.
func (s *rpcServer) httpStatusLine(req *http.Request, code int) string {
// Fast path:
key := code
proto11 := req.ProtoAtLeast(1, 1)
if !proto11 {
key = -key
}
s.statusLock.RLock()
line, ok := s.statusLines[key]
s.statusLock.RUnlock()
if ok {
return line
}
// Slow path:
proto := "HTTP/1.0"
if proto11 {
proto = "HTTP/1.1"
}
codeStr := strconv.Itoa(code)
text := http.StatusText(code)
if text != "" {
line = proto + " " + codeStr + " " + text + "\r\n"
s.statusLock.Lock()
s.statusLines[key] = line
s.statusLock.Unlock()
} else {
text = "status code " + codeStr
line = proto + " " + codeStr + " " + text + "\r\n"
}
return line
}
// writeHTTPResponseHeaders writes the necessary response headers prior to
// writing an HTTP body given a request to use for protocol negotiation, headers
// to write, a status code, and a writer.
func (s *rpcServer) writeHTTPResponseHeaders(req *http.Request, headers http.Header, code int, w io.Writer) error {
_, err := io.WriteString(w, s.httpStatusLine(req, code))
if err != nil {
return err
}
err = headers.Write(w)
if err != nil {
return err
}
_, err = io.WriteString(w, "\r\n")
if err != nil {
return err
}
return nil
}
// Stop is used by server.go to stop the rpc listener.
func (s *rpcServer) Stop() error {
if atomic.AddInt32(&s.shutdown, 1) != 1 {
rpcsLog.Infof("RPC server is already in the process of shutting down")
return nil
}
rpcsLog.Warnf("RPC server shutting down")
for _, listener := range s.listeners {
err := listener.Close()
if err != nil {
rpcsLog.Errorf("Problem shutting down rpc: %v", err)
return err
}
}
s.ntfnMgr.Shutdown()
s.ntfnMgr.WaitForShutdown()
close(s.quit)
s.wg.Wait()
rpcsLog.Infof("RPC server shutdown complete")
return nil
}
// RequestedProcessShutdown returns a channel that is sent to when an authorized
// RPC client requests the process to shutdown. If the request can not be read
// immediately, it is dropped.
func (s *rpcServer) RequestedProcessShutdown() <-chan struct{} {
return s.requestProcessShutdown
}
// limitConnections responds with a 503 service unavailable and returns true if
// adding another client would exceed the maximum allow RPC clients.
//
// This function is safe for concurrent access.
func (s *rpcServer) limitConnections(w http.ResponseWriter, remoteAddr string) bool {
if int(atomic.LoadInt32(&s.numClients)+1) > cfg.RPCMaxClients {
rpcsLog.Infof("Max RPC clients exceeded [%d] - "+
"disconnecting client %s", cfg.RPCMaxClients,
remoteAddr)
http.Error(w, "503 Too busy. Try again later.",
http.StatusServiceUnavailable)
return true
}
return false
}
// incrementClients adds one to the number of connected RPC clients. Note
// this only applies to standard clients. Websocket clients have their own
// limits and are tracked separately.
//
// This function is safe for concurrent access.
func (s *rpcServer) incrementClients() {
atomic.AddInt32(&s.numClients, 1)
}
// decrementClients subtracts one from the number of connected RPC clients.
// Note this only applies to standard clients. Websocket clients have their own
// limits and are tracked separately.
//
// This function is safe for concurrent access.
func (s *rpcServer) decrementClients() {
atomic.AddInt32(&s.numClients, -1)
}
// checkAuth checks the HTTP Basic authentication supplied by a wallet
// or RPC client in the HTTP request r. If the supplied authentication
// does not match the username and password expected, a non-nil error is
// returned.
//
// This check is time-constant.
//
// The first bool return value signifies auth success (true if successful) and
// the second bool return value specifies whether the user can change the state
// of the server (true) or whether the user is limited (false). The second is
// always false if the first is.
indexers: Implement optional tx/address indexes. This introduces a new indexing infrastructure for supporting optional indexes using the new database and blockchain infrastructure along with two concrete indexer implementations which provide both a transaction-by-hash and a transaction-by-address index. The new infrastructure is mostly separated into a package named indexers which is housed under the blockchain package. In order to support this, a new interface named IndexManager has been introduced in the blockchain package which provides methods to be notified when the chain has been initialized and when blocks are connected and disconnected from the main chain. A concrete implementation of an index manager is provided by the new indexers package. The new indexers package also provides a new interface named Indexer which allows the index manager to manage concrete index implementations which conform to the interface. The following is high level overview of the main index infrastructure changes: - Define a new IndexManager interface in the blockchain package and modify the package to make use of the interface when specified - Create a new indexers package - Provides an Index interface which allows concrete indexes to plugin to an index manager - Provides a concrete IndexManager implementation - Handles the lifecycle of all indexes it manages - Tracks the index tips - Handles catching up disabled indexes that have been reenabled - Handles reorgs while the index was disabled - Invokes the appropriate methods for all managed indexes to allow them to index and deindex the blocks and transactions - Implement a transaction-by-hash index - Makes use of internal block IDs to save a significant amount of space and indexing costs over the old transaction index format - Implement a transaction-by-address index - Makes use of a leveling scheme in order to provide a good tradeoff between space required and indexing costs - Supports enabling and disabling indexes at will - Support the ability to drop indexes if they are no longer desired The following is an overview of the btcd changes: - Add a new index logging subsystem - Add new options --txindex and --addrindex in order to enable the optional indexes - NOTE: The transaction index will automatically be enabled when the address index is enabled because it depends on it - Add new options --droptxindex and --dropaddrindex to allow the indexes to be removed - NOTE: The address index will also be removed when the transaction index is dropped because it depends on it - Update getrawtransactions RPC to make use of the transaction index - Reimplement the searchrawtransaction RPC that makes use of the address index - Update sample-btcd.conf to include sample usage for the new optional index flags
2016-02-19 05:51:18 +01:00
func (s *rpcServer) checkAuth(r *http.Request, require bool) (bool, bool, error) {
authhdr := r.Header["Authorization"]
if len(authhdr) <= 0 {
if require {
rpcsLog.Warnf("RPC authentication failure from %s",
r.RemoteAddr)
return false, false, errors.New("auth failure")
}
return false, false, nil
}
authsha := fastsha256.Sum256([]byte(authhdr[0]))
// Check for limited auth first as in environments with limited users, those
// are probably expected to have a higher volume of calls
limitcmp := subtle.ConstantTimeCompare(authsha[:], s.limitauthsha[:])
if limitcmp == 1 {
return true, false, nil
}
// Check for admin-level auth
cmp := subtle.ConstantTimeCompare(authsha[:], s.authsha[:])
if cmp == 1 {
return true, true, nil
}
// Request's auth doesn't match either user
rpcsLog.Warnf("RPC authentication failure from %s", r.RemoteAddr)
return false, false, errors.New("auth failure")
}
// parsedRPCCmd represents a JSON-RPC request object that has been parsed into
// a known concrete command along with any error that might have happened while
// parsing it.
type parsedRPCCmd struct {
id interface{}
method string
cmd interface{}
err *btcjson.RPCError
}
// standardCmdResult checks that a parsed command is a standard Bitcoin JSON-RPC
// command and runs the appropriate handler to reply to the command. Any
// commands which are not recognized or not implemented will return an error
// suitable for use in replies.
func (s *rpcServer) standardCmdResult(cmd *parsedRPCCmd, closeChan <-chan struct{}) (interface{}, error) {
handler, ok := rpcHandlers[cmd.method]
if ok {
goto handled
}
_, ok = rpcAskWallet[cmd.method]
if ok {
handler = handleAskWallet
goto handled
}
_, ok = rpcUnimplemented[cmd.method]
if ok {
handler = handleUnimplemented
goto handled
}
return nil, btcjson.ErrRPCMethodNotFound
handled:
return handler(s, cmd.cmd, closeChan)
}
// parseCmd parses a JSON-RPC request object into known concrete command. The
// err field of the returned parsedRPCCmd struct will contain an RPC error that
// is suitable for use in replies if the command is invalid in some way such as
// an unregistered command or invalid parameters.
func parseCmd(request *btcjson.Request) *parsedRPCCmd {
var parsedCmd parsedRPCCmd
parsedCmd.id = request.ID
parsedCmd.method = request.Method
cmd, err := btcjson.UnmarshalCmd(request)
if err != nil {
// When the error is because the method is not registered,
// produce a method not found RPC error.
if jerr, ok := err.(btcjson.Error); ok &&
jerr.ErrorCode == btcjson.ErrUnregisteredMethod {
parsedCmd.err = btcjson.ErrRPCMethodNotFound
return &parsedCmd
}
// Otherwise, some type of invalid parameters is the
// cause, so produce the equivalent RPC error.
parsedCmd.err = btcjson.NewRPCError(
btcjson.ErrRPCInvalidParams.Code, err.Error())
return &parsedCmd
}
parsedCmd.cmd = cmd
return &parsedCmd
}
// createMarshalledReply returns a new marshalled JSON-RPC response given the
// passed parameters. It will automatically convert errors that are not of
// the type *btcjson.RPCError to the appropriate type as needed.
func createMarshalledReply(id, result interface{}, replyErr error) ([]byte, error) {
var jsonErr *btcjson.RPCError
if replyErr != nil {
if jErr, ok := replyErr.(*btcjson.RPCError); ok {
jsonErr = jErr
} else {
jsonErr = internalRPCError(replyErr.Error(), "")
}
}
return btcjson.MarshalResponse(id, result, jsonErr)
}
// jsonRPCRead handles reading and responding to RPC messages.
indexers: Implement optional tx/address indexes. This introduces a new indexing infrastructure for supporting optional indexes using the new database and blockchain infrastructure along with two concrete indexer implementations which provide both a transaction-by-hash and a transaction-by-address index. The new infrastructure is mostly separated into a package named indexers which is housed under the blockchain package. In order to support this, a new interface named IndexManager has been introduced in the blockchain package which provides methods to be notified when the chain has been initialized and when blocks are connected and disconnected from the main chain. A concrete implementation of an index manager is provided by the new indexers package. The new indexers package also provides a new interface named Indexer which allows the index manager to manage concrete index implementations which conform to the interface. The following is high level overview of the main index infrastructure changes: - Define a new IndexManager interface in the blockchain package and modify the package to make use of the interface when specified - Create a new indexers package - Provides an Index interface which allows concrete indexes to plugin to an index manager - Provides a concrete IndexManager implementation - Handles the lifecycle of all indexes it manages - Tracks the index tips - Handles catching up disabled indexes that have been reenabled - Handles reorgs while the index was disabled - Invokes the appropriate methods for all managed indexes to allow them to index and deindex the blocks and transactions - Implement a transaction-by-hash index - Makes use of internal block IDs to save a significant amount of space and indexing costs over the old transaction index format - Implement a transaction-by-address index - Makes use of a leveling scheme in order to provide a good tradeoff between space required and indexing costs - Supports enabling and disabling indexes at will - Support the ability to drop indexes if they are no longer desired The following is an overview of the btcd changes: - Add a new index logging subsystem - Add new options --txindex and --addrindex in order to enable the optional indexes - NOTE: The transaction index will automatically be enabled when the address index is enabled because it depends on it - Add new options --droptxindex and --dropaddrindex to allow the indexes to be removed - NOTE: The address index will also be removed when the transaction index is dropped because it depends on it - Update getrawtransactions RPC to make use of the transaction index - Reimplement the searchrawtransaction RPC that makes use of the address index - Update sample-btcd.conf to include sample usage for the new optional index flags
2016-02-19 05:51:18 +01:00
func (s *rpcServer) jsonRPCRead(w http.ResponseWriter, r *http.Request, isAdmin bool) {
if atomic.LoadInt32(&s.shutdown) != 0 {
return
}
// Read and close the JSON-RPC request body from the caller.
body, err := ioutil.ReadAll(r.Body)
r.Body.Close()
2014-01-15 16:01:12 +01:00
if err != nil {
errMsg := fmt.Sprintf("error reading JSON message: %v", err)
errCode := http.StatusBadRequest
http.Error(w, strconv.FormatInt(int64(errCode), 10)+" "+errMsg,
errCode)
return
2014-01-15 16:01:12 +01:00
}
// Unfortunately, the http server doesn't provide the ability to
// change the read deadline for the new connection and having one breaks
// long polling. However, not having a read deadline on the initial
// connection would mean clients can connect and idle forever. Thus,
// hijack the connecton from the HTTP server, clear the read deadline,
// and handle writing the response manually.
hj, ok := w.(http.Hijacker)
if !ok {
errMsg := "webserver doesn't support hijacking"
rpcsLog.Warnf(errMsg)
errCode := http.StatusInternalServerError
http.Error(w, strconv.FormatInt(int64(errCode), 10)+" "+errMsg,
errCode)
return
}
conn, buf, err := hj.Hijack()
2014-01-15 16:01:12 +01:00
if err != nil {
rpcsLog.Warnf("Failed to hijack HTTP connection: %v", err)
errCode := http.StatusInternalServerError
http.Error(w, strconv.FormatInt(int64(errCode), 10)+" "+
err.Error(), errCode)
return
2014-01-15 16:01:12 +01:00
}
defer conn.Close()
defer buf.Flush()
conn.SetReadDeadline(timeZeroVal)
2014-01-15 16:01:12 +01:00
// Attempt to parse the raw body into a JSON-RPC request.
var responseID interface{}
var jsonErr error
var result interface{}
var request btcjson.Request
if err := json.Unmarshal(body, &request); err != nil {
jsonErr = &btcjson.RPCError{
Code: btcjson.ErrRPCParse.Code,
Message: "Failed to parse request: " + err.Error(),
}
}
if jsonErr == nil {
// Requests with no ID (notifications) must not have a response
// per the JSON-RPC spec.
if request.ID == nil {
return
}
// The parse was at least successful enough to have an ID so
// set it for the response.
responseID = request.ID
// Setup a close notifier. Since the connection is hijacked,
// the CloseNotifer on the ResponseWriter is not available.
closeChan := make(chan struct{}, 1)
go func() {
_, err := conn.Read(make([]byte, 1))
if err != nil {
close(closeChan)
}
}()
// Check if the user is limited and set error if method unauthorized
if !isAdmin {
if _, ok := rpcLimited[request.Method]; !ok {
jsonErr = &btcjson.RPCError{
Code: btcjson.ErrRPCInvalidParams.Code,
Message: "limited user not authorized for this method",
}
}
}
if jsonErr == nil {
// Attempt to parse the JSON-RPC request into a known concrete
// command.
parsedCmd := parseCmd(&request)
if parsedCmd.err != nil {
jsonErr = parsedCmd.err
} else {
result, jsonErr = s.standardCmdResult(parsedCmd, closeChan)
}
}
}
// Marshal the response.
msg, err := createMarshalledReply(responseID, result, jsonErr)
2014-01-15 16:01:12 +01:00
if err != nil {
rpcsLog.Errorf("Failed to marshal reply: %v", err)
return
2014-01-15 16:01:12 +01:00
}
// Write the response.
err = s.writeHTTPResponseHeaders(r, w.Header(), http.StatusOK, buf)
if err != nil {
rpcsLog.Error(err)
return
}
if _, err := buf.Write(msg); err != nil {
rpcsLog.Errorf("Failed to write marshalled reply: %v", err)
}
2014-01-15 16:01:12 +01:00
}
// jsonAuthFail sends a message back to the client if the http auth is rejected.
func jsonAuthFail(w http.ResponseWriter) {
w.Header().Add("WWW-Authenticate", `Basic realm="btcd RPC"`)
http.Error(w, "401 Unauthorized.", http.StatusUnauthorized)
}
// Start is used by server.go to start the rpc listener.
func (s *rpcServer) Start() {
if atomic.AddInt32(&s.started, 1) != 1 {
return
}
rpcsLog.Trace("Starting RPC server")
rpcServeMux := http.NewServeMux()
httpServer := &http.Server{
Handler: rpcServeMux,
// Timeout connections which don't complete the initial
// handshake within the allowed timeframe.
ReadTimeout: time.Second * rpcAuthTimeoutSeconds,
}
rpcServeMux.HandleFunc("/", func(w http.ResponseWriter, r *http.Request) {
w.Header().Set("Connection", "close")
w.Header().Set("Content-Type", "application/json")
r.Close = true
// Limit the number of connections to max allowed.
if s.limitConnections(w, r.RemoteAddr) {
return
}
// Keep track of the number of connected clients.
s.incrementClients()
defer s.decrementClients()
_, isAdmin, err := s.checkAuth(r, true)
if err != nil {
jsonAuthFail(w)
return
}
// Read and respond to the request.
s.jsonRPCRead(w, r, isAdmin)
})
// Websocket endpoint.
rpcServeMux.HandleFunc("/ws", func(w http.ResponseWriter, r *http.Request) {
authenticated, isAdmin, err := s.checkAuth(r, false)
if err != nil {
jsonAuthFail(w)
return
}
// Attempt to upgrade the connection to a websocket connection
// using the default size for read/write buffers.
ws, err := websocket.Upgrade(w, r, nil, 0, 0)
if err != nil {
if _, ok := err.(websocket.HandshakeError); !ok {
rpcsLog.Errorf("Unexpected websocket error: %v",
err)
}
http.Error(w, "400 Bad Request.", http.StatusBadRequest)
return
}
s.WebsocketHandler(ws, r.RemoteAddr, authenticated, isAdmin)
})
for _, listener := range s.listeners {
s.wg.Add(1)
go func(listener net.Listener) {
rpcsLog.Infof("RPC server listening on %s", listener.Addr())
httpServer.Serve(listener)
rpcsLog.Tracef("RPC listener done for %s", listener.Addr())
s.wg.Done()
}(listener)
}
s.ntfnMgr.Start()
}
// genCertPair generates a key/cert pair to the paths provided.
func genCertPair(certFile, keyFile string) error {
rpcsLog.Infof("Generating TLS certificates...")
org := "btcd autogenerated cert"
validUntil := time.Now().Add(10 * 365 * 24 * time.Hour)
cert, key, err := btcutil.NewTLSCertPair(org, validUntil, nil)
if err != nil {
return err
}
// Write cert and key files.
if err = ioutil.WriteFile(certFile, cert, 0666); err != nil {
return err
}
if err = ioutil.WriteFile(keyFile, key, 0600); err != nil {
os.Remove(certFile)
return err
}
rpcsLog.Infof("Done generating TLS certificates")
return nil
}
// newRPCServer returns a new instance of the rpcServer struct.
func newRPCServer(listenAddrs []string, policy *mining.Policy, s *server) (*rpcServer, error) {
rpc := rpcServer{
policy: policy,
server: s,
chain: s.blockManager.chain,
statusLines: make(map[int]string),
workState: newWorkState(),
gbtWorkState: newGbtWorkState(s.timeSource),
helpCacher: newHelpCacher(),
requestProcessShutdown: make(chan struct{}),
quit: make(chan int),
2014-08-22 23:22:46 +02:00
}
if cfg.RPCUser != "" && cfg.RPCPass != "" {
login := cfg.RPCUser + ":" + cfg.RPCPass
auth := "Basic " + base64.StdEncoding.EncodeToString([]byte(login))
rpc.authsha = fastsha256.Sum256([]byte(auth))
}
if cfg.RPCLimitUser != "" && cfg.RPCLimitPass != "" {
login := cfg.RPCLimitUser + ":" + cfg.RPCLimitPass
auth := "Basic " + base64.StdEncoding.EncodeToString([]byte(login))
rpc.limitauthsha = fastsha256.Sum256([]byte(auth))
}
rpc.ntfnMgr = newWsNotificationManager(&rpc)
2014-08-22 23:22:46 +02:00
// Setup TLS if not disabled.
listenFunc := net.Listen
if !cfg.DisableTLS {
// Generate the TLS cert and key file if both don't already
// exist.
if !fileExists(cfg.RPCKey) && !fileExists(cfg.RPCCert) {
err := genCertPair(cfg.RPCCert, cfg.RPCKey)
if err != nil {
return nil, err
}
}
keypair, err := tls.LoadX509KeyPair(cfg.RPCCert, cfg.RPCKey)
if err != nil {
return nil, err
2014-08-22 23:22:46 +02:00
}
tlsConfig := tls.Config{
Certificates: []tls.Certificate{keypair},
MinVersion: tls.VersionTLS12,
2014-08-22 23:22:46 +02:00
}
// Change the standard net.Listen function to the tls one.
listenFunc = func(net string, laddr string) (net.Listener, error) {
return tls.Listen(net, laddr, &tlsConfig)
}
2014-08-22 23:22:46 +02:00
}
// TODO(oga) this code is similar to that in server, should be
// factored into something shared.
ipv4ListenAddrs, ipv6ListenAddrs, _, err := parseListeners(listenAddrs)
2014-08-22 23:22:46 +02:00
if err != nil {
return nil, err
2014-08-22 23:22:46 +02:00
}
listeners := make([]net.Listener, 0,
len(ipv6ListenAddrs)+len(ipv4ListenAddrs))
for _, addr := range ipv4ListenAddrs {
listener, err := listenFunc("tcp4", addr)
if err != nil {
rpcsLog.Warnf("Can't listen on %s: %v", addr, err)
continue
2013-10-29 01:43:09 +01:00
}
listeners = append(listeners, listener)
}
for _, addr := range ipv6ListenAddrs {
listener, err := listenFunc("tcp6", addr)
if err != nil {
rpcsLog.Warnf("Can't listen on %s: %v", addr, err)
continue
}
listeners = append(listeners, listener)
}
if len(listeners) == 0 {
return nil, errors.New("RPCS: No valid listen address")
}
rpc.listeners = listeners
2013-08-06 23:55:22 +02:00
return &rpc, nil
2013-08-06 23:55:22 +02:00
}
func init() {
rpcHandlers = rpcHandlersBeforeInit
rand.Seed(time.Now().UnixNano())
}