lbcd/blockchain/fullblocktests/generate.go

2136 lines
74 KiB
Go
Raw Normal View History

blockchain: Add block validation infrastructure. This adds a full-blown testing infrastructure in order to test consensus validation rules. It is built around the idea of dynamically generating full blocks that target specific rules linked together to form a block chain. In order to properly test the rules, each test instance starts with a valid block that is then modified in the specific way needed to test a specific rule. Blocks which exercise following rules have been added for this initial version. These tests were largely ported from the original Java-based 'official' block acceptance tests as well as some additional tests available in the Core python port. It is expected that further tests can be added over time as consensus rules change. * Enough valid blocks to have a stable base of mature coinbases to spend for futher tests * Basic forking and chain reorganization * Double spends on forks * Too much proof-of-work coinbase (extending main chain, in block that forces a reorg, and in a valid fork) * Max and too many signature operations via various combinations of OP_CHECKSIG, OP_MULTISIG, OP_CHECKSIGVERIFY, and OP_MULTISIGVERIFY * Too many and max signature operations with offending sigop after invalid data push * Max and too many signature operations via pay-to-script-hash redeem scripts * Attempt to spend tx created on a different fork * Attempt to spend immature coinbase (on main chain and fork) * Max size block and block that exceeds the max size * Children of rejected blocks are either orphans or rejected * Coinbase script too small and too large * Max length coinbase script * Attempt to spend tx in blocks that failed to connect * Valid non-coinbase tx in place of coinbase * Block with no transactions * Invalid proof-of-work * Block with a timestamp too far in the future * Invalid merkle root * Invalid proof-of-work limit (bits header field) * Negative proof-of-work limit (bits header field) * Two coinbase transactions * Duplicate transactions * Spend from transaction that does not exist * Timestamp exactly at and one second after the median time * Blocks with same hash via merkle root tricks * Spend from transaction index that is out of range * Transaction that spends more that its inputs provide * Transaction with same hash as an existing tx that has not been fully spent (BIP0030) * Non-final coinbase and non-coinbase txns * Max size block with canonical encoding which exceeds max size with non-canonical encoding * Spend from transaction earlier in same block * Spend from transaction later in same block * Double spend transaction from earlier in same block * Coinbase that pays more than subsidy + fees * Coinbase that includes subsidy + fees * Invalid opcode in dead execution path * Reorganization of txns with OP_RETURN outputs * Spend of an OP_RETURN output * Transaction with multiple OP_RETURN outputs * Large max-sized block reorganization test (disabled by default since it takes a long time and a lot of memory to run) Finally, the README.md files in the main and docs directories have been updated to reflect the use of the new testing framework.
2016-09-14 01:11:12 +02:00
// Copyright (c) 2016 The btcsuite developers
// Use of this source code is governed by an ISC
// license that can be found in the LICENSE file.
// The vast majority of the rules tested in this package were ported from the
// the original Java-based 'official' block acceptance tests at
// https://github.com/TheBlueMatt/test-scripts as well as some additional tests
// available in the Core python port of the same.
package fullblocktests
import (
"bytes"
"encoding/binary"
"errors"
"fmt"
"math"
"runtime"
"time"
"github.com/btcsuite/btcd/blockchain"
"github.com/btcsuite/btcd/btcec"
"github.com/btcsuite/btcd/chaincfg"
"github.com/btcsuite/btcd/chaincfg/chainhash"
"github.com/btcsuite/btcd/txscript"
"github.com/btcsuite/btcd/wire"
"github.com/btcsuite/btcutil"
)
const (
// Intentionally defined here rather than using constants from codebase
// to ensure consensus changes are detected.
maxBlockSigOps = 20000
maxBlockSize = 2000000
blockchain: Add block validation infrastructure. This adds a full-blown testing infrastructure in order to test consensus validation rules. It is built around the idea of dynamically generating full blocks that target specific rules linked together to form a block chain. In order to properly test the rules, each test instance starts with a valid block that is then modified in the specific way needed to test a specific rule. Blocks which exercise following rules have been added for this initial version. These tests were largely ported from the original Java-based 'official' block acceptance tests as well as some additional tests available in the Core python port. It is expected that further tests can be added over time as consensus rules change. * Enough valid blocks to have a stable base of mature coinbases to spend for futher tests * Basic forking and chain reorganization * Double spends on forks * Too much proof-of-work coinbase (extending main chain, in block that forces a reorg, and in a valid fork) * Max and too many signature operations via various combinations of OP_CHECKSIG, OP_MULTISIG, OP_CHECKSIGVERIFY, and OP_MULTISIGVERIFY * Too many and max signature operations with offending sigop after invalid data push * Max and too many signature operations via pay-to-script-hash redeem scripts * Attempt to spend tx created on a different fork * Attempt to spend immature coinbase (on main chain and fork) * Max size block and block that exceeds the max size * Children of rejected blocks are either orphans or rejected * Coinbase script too small and too large * Max length coinbase script * Attempt to spend tx in blocks that failed to connect * Valid non-coinbase tx in place of coinbase * Block with no transactions * Invalid proof-of-work * Block with a timestamp too far in the future * Invalid merkle root * Invalid proof-of-work limit (bits header field) * Negative proof-of-work limit (bits header field) * Two coinbase transactions * Duplicate transactions * Spend from transaction that does not exist * Timestamp exactly at and one second after the median time * Blocks with same hash via merkle root tricks * Spend from transaction index that is out of range * Transaction that spends more that its inputs provide * Transaction with same hash as an existing tx that has not been fully spent (BIP0030) * Non-final coinbase and non-coinbase txns * Max size block with canonical encoding which exceeds max size with non-canonical encoding * Spend from transaction earlier in same block * Spend from transaction later in same block * Double spend transaction from earlier in same block * Coinbase that pays more than subsidy + fees * Coinbase that includes subsidy + fees * Invalid opcode in dead execution path * Reorganization of txns with OP_RETURN outputs * Spend of an OP_RETURN output * Transaction with multiple OP_RETURN outputs * Large max-sized block reorganization test (disabled by default since it takes a long time and a lot of memory to run) Finally, the README.md files in the main and docs directories have been updated to reflect the use of the new testing framework.
2016-09-14 01:11:12 +02:00
minCoinbaseScriptLen = 2
maxCoinbaseScriptLen = 100
medianTimeBlocks = 11
maxScriptElementSize = 520
// numLargeReorgBlocks is the number of blocks to use in the large block
// reorg test (when enabled). This is the equivalent of 1 week's worth
// of blocks.
numLargeReorgBlocks = 1088
)
var (
// opTrueScript is simply a public key script that contains the OP_TRUE
// opcode. It is defined here to reduce garbage creation.
opTrueScript = []byte{txscript.OP_TRUE}
// lowFee is a single satoshi and exists to make the test code more
// readable.
lowFee = btcutil.Amount(1)
)
// TestInstance is an interface that describes a specific test instance returned
// by the tests generated in this package. It should be type asserted to one
// of the concrete test instance types in order to test accordingly.
type TestInstance interface {
FullBlockTestInstance()
}
// AcceptedBlock defines a test instance that expects a block to be accepted to
// the blockchain either by extending the main chain, on a side chain, or as an
// orphan.
type AcceptedBlock struct {
Name string
Block *wire.MsgBlock
Height int32
IsMainChain bool
IsOrphan bool
}
// Ensure AcceptedBlock implements the TestInstance interface.
var _ TestInstance = AcceptedBlock{}
// FullBlockTestInstance only exists to allow AcceptedBlock to be treated as a
// TestInstance.
//
// This implements the TestInstance interface.
func (b AcceptedBlock) FullBlockTestInstance() {}
// RejectedBlock defines a test instance that expects a block to be rejected by
// the blockchain consensus rules.
type RejectedBlock struct {
Name string
Block *wire.MsgBlock
Height int32
RejectCode blockchain.ErrorCode
}
// Ensure RejectedBlock implements the TestInstance interface.
var _ TestInstance = RejectedBlock{}
// FullBlockTestInstance only exists to allow RejectedBlock to be treated as a
// TestInstance.
//
// This implements the TestInstance interface.
func (b RejectedBlock) FullBlockTestInstance() {}
// OrphanOrRejectedBlock defines a test instance that expects a block to either
// be accepted as an orphan or rejected. This is useful since some
// implementations might optimize the immediate rejection of orphan blocks when
// their parent was previously rejected, while others might accept it as an
// orphan that eventually gets flushed (since the parent can never be accepted
// to ultimately link it).
type OrphanOrRejectedBlock struct {
Name string
Block *wire.MsgBlock
Height int32
}
// Ensure ExpectedTip implements the TestInstance interface.
var _ TestInstance = OrphanOrRejectedBlock{}
// FullBlockTestInstance only exists to allow OrphanOrRejectedBlock to be
// treated as a TestInstance.
//
// This implements the TestInstance interface.
func (b OrphanOrRejectedBlock) FullBlockTestInstance() {}
// ExpectedTip defines a test instance that expects a block to be the current
// tip of the main chain.
type ExpectedTip struct {
Name string
Block *wire.MsgBlock
Height int32
}
// Ensure ExpectedTip implements the TestInstance interface.
var _ TestInstance = ExpectedTip{}
// FullBlockTestInstance only exists to allow ExpectedTip to be treated as a
// TestInstance.
//
// This implements the TestInstance interface.
func (b ExpectedTip) FullBlockTestInstance() {}
// RejectedNonCanonicalBlock defines a test instance that expects a serialized
// block that is not canonical and therefore should be rejected.
type RejectedNonCanonicalBlock struct {
Name string
RawBlock []byte
Height int32
}
// FullBlockTestInstance only exists to allow RejectedNonCanonicalBlock to be treated as
// a TestInstance.
//
// This implements the TestInstance interface.
func (b RejectedNonCanonicalBlock) FullBlockTestInstance() {}
// spendableOut represents a transaction output that is spendable along with
// additional metadata such as the block its in and how much it pays.
type spendableOut struct {
prevOut wire.OutPoint
amount btcutil.Amount
}
// makeSpendableOutForTx returns a spendable output for the given transaction
// and transaction output index within the transaction.
func makeSpendableOutForTx(tx *wire.MsgTx, txOutIndex uint32) spendableOut {
return spendableOut{
prevOut: wire.OutPoint{
Hash: tx.TxHash(),
Index: txOutIndex,
},
amount: btcutil.Amount(tx.TxOut[txOutIndex].Value),
}
}
// makeSpendableOut returns a spendable output for the given block, transaction
// index within the block, and transaction output index within the transaction.
func makeSpendableOut(block *wire.MsgBlock, txIndex, txOutIndex uint32) spendableOut {
return makeSpendableOutForTx(block.Transactions[txIndex], txOutIndex)
}
// testGenerator houses state used to easy the process of generating test blocks
// that build from one another along with housing other useful things such as
// available spendable outputs used throughout the tests.
type testGenerator struct {
params *chaincfg.Params
tip *wire.MsgBlock
tipName string
tipHeight int32
blocks map[chainhash.Hash]*wire.MsgBlock
blocksByName map[string]*wire.MsgBlock
blockHeights map[string]int32
// Used for tracking spendable coinbase outputs.
spendableOuts []spendableOut
prevCollectedHash chainhash.Hash
// Common key for any tests which require signed transactions.
privKey *btcec.PrivateKey
}
// makeTestGenerator returns a test generator instance initialized with the
// genesis block as the tip.
func makeTestGenerator(params *chaincfg.Params) (testGenerator, error) {
privKey, _ := btcec.PrivKeyFromBytes(btcec.S256(), []byte{0x01})
genesis := params.GenesisBlock
genesisHash := genesis.BlockHash()
return testGenerator{
params: params,
blocks: map[chainhash.Hash]*wire.MsgBlock{genesisHash: genesis},
blocksByName: map[string]*wire.MsgBlock{"genesis": genesis},
blockHeights: map[string]int32{"genesis": 0},
tip: genesis,
tipName: "genesis",
tipHeight: 0,
privKey: privKey,
}, nil
}
// payToScriptHashScript returns a standard pay-to-script-hash for the provided
// redeem script.
func payToScriptHashScript(redeemScript []byte) []byte {
redeemScriptHash := btcutil.Hash160(redeemScript)
script, err := txscript.NewScriptBuilder().
AddOp(txscript.OP_HASH160).AddData(redeemScriptHash).
AddOp(txscript.OP_EQUAL).Script()
if err != nil {
panic(err)
}
return script
}
// pushDataScript returns a script with the provided items individually pushed
// to the stack.
func pushDataScript(items ...[]byte) []byte {
builder := txscript.NewScriptBuilder()
for _, item := range items {
builder.AddData(item)
}
script, err := builder.Script()
if err != nil {
panic(err)
}
return script
}
// standardCoinbaseScript returns a standard script suitable for use as the
// signature script of the coinbase transaction of a new block. In particular,
// it starts with the block height that is required by version 2 blocks.
func standardCoinbaseScript(blockHeight int32, extraNonce uint64) ([]byte, error) {
return txscript.NewScriptBuilder().AddInt64(int64(blockHeight)).
AddInt64(int64(extraNonce)).Script()
}
// opReturnScript returns a provably-pruneable OP_RETURN script with the
// provided data.
func opReturnScript(data []byte) []byte {
builder := txscript.NewScriptBuilder()
script, err := builder.AddOp(txscript.OP_RETURN).AddData(data).Script()
if err != nil {
panic(err)
}
return script
}
// uniqueOpReturnScript returns a standard provably-pruneable OP_RETURN script
// with a random uint64 encoded as the data.
func uniqueOpReturnScript() []byte {
rand, err := wire.RandomUint64()
if err != nil {
panic(err)
}
data := make([]byte, 8)
blockchain: Add block validation infrastructure. This adds a full-blown testing infrastructure in order to test consensus validation rules. It is built around the idea of dynamically generating full blocks that target specific rules linked together to form a block chain. In order to properly test the rules, each test instance starts with a valid block that is then modified in the specific way needed to test a specific rule. Blocks which exercise following rules have been added for this initial version. These tests were largely ported from the original Java-based 'official' block acceptance tests as well as some additional tests available in the Core python port. It is expected that further tests can be added over time as consensus rules change. * Enough valid blocks to have a stable base of mature coinbases to spend for futher tests * Basic forking and chain reorganization * Double spends on forks * Too much proof-of-work coinbase (extending main chain, in block that forces a reorg, and in a valid fork) * Max and too many signature operations via various combinations of OP_CHECKSIG, OP_MULTISIG, OP_CHECKSIGVERIFY, and OP_MULTISIGVERIFY * Too many and max signature operations with offending sigop after invalid data push * Max and too many signature operations via pay-to-script-hash redeem scripts * Attempt to spend tx created on a different fork * Attempt to spend immature coinbase (on main chain and fork) * Max size block and block that exceeds the max size * Children of rejected blocks are either orphans or rejected * Coinbase script too small and too large * Max length coinbase script * Attempt to spend tx in blocks that failed to connect * Valid non-coinbase tx in place of coinbase * Block with no transactions * Invalid proof-of-work * Block with a timestamp too far in the future * Invalid merkle root * Invalid proof-of-work limit (bits header field) * Negative proof-of-work limit (bits header field) * Two coinbase transactions * Duplicate transactions * Spend from transaction that does not exist * Timestamp exactly at and one second after the median time * Blocks with same hash via merkle root tricks * Spend from transaction index that is out of range * Transaction that spends more that its inputs provide * Transaction with same hash as an existing tx that has not been fully spent (BIP0030) * Non-final coinbase and non-coinbase txns * Max size block with canonical encoding which exceeds max size with non-canonical encoding * Spend from transaction earlier in same block * Spend from transaction later in same block * Double spend transaction from earlier in same block * Coinbase that pays more than subsidy + fees * Coinbase that includes subsidy + fees * Invalid opcode in dead execution path * Reorganization of txns with OP_RETURN outputs * Spend of an OP_RETURN output * Transaction with multiple OP_RETURN outputs * Large max-sized block reorganization test (disabled by default since it takes a long time and a lot of memory to run) Finally, the README.md files in the main and docs directories have been updated to reflect the use of the new testing framework.
2016-09-14 01:11:12 +02:00
binary.LittleEndian.PutUint64(data[0:8], rand)
return opReturnScript(data)
}
// createCoinbaseTx returns a coinbase transaction paying an appropriate
// subsidy based on the passed block height. The coinbase signature script
// conforms to the requirements of version 2 blocks.
func (g *testGenerator) createCoinbaseTx(blockHeight int32) *wire.MsgTx {
extraNonce := uint64(0)
coinbaseScript, err := standardCoinbaseScript(blockHeight, extraNonce)
if err != nil {
panic(err)
}
tx := wire.NewMsgTx(1)
blockchain: Add block validation infrastructure. This adds a full-blown testing infrastructure in order to test consensus validation rules. It is built around the idea of dynamically generating full blocks that target specific rules linked together to form a block chain. In order to properly test the rules, each test instance starts with a valid block that is then modified in the specific way needed to test a specific rule. Blocks which exercise following rules have been added for this initial version. These tests were largely ported from the original Java-based 'official' block acceptance tests as well as some additional tests available in the Core python port. It is expected that further tests can be added over time as consensus rules change. * Enough valid blocks to have a stable base of mature coinbases to spend for futher tests * Basic forking and chain reorganization * Double spends on forks * Too much proof-of-work coinbase (extending main chain, in block that forces a reorg, and in a valid fork) * Max and too many signature operations via various combinations of OP_CHECKSIG, OP_MULTISIG, OP_CHECKSIGVERIFY, and OP_MULTISIGVERIFY * Too many and max signature operations with offending sigop after invalid data push * Max and too many signature operations via pay-to-script-hash redeem scripts * Attempt to spend tx created on a different fork * Attempt to spend immature coinbase (on main chain and fork) * Max size block and block that exceeds the max size * Children of rejected blocks are either orphans or rejected * Coinbase script too small and too large * Max length coinbase script * Attempt to spend tx in blocks that failed to connect * Valid non-coinbase tx in place of coinbase * Block with no transactions * Invalid proof-of-work * Block with a timestamp too far in the future * Invalid merkle root * Invalid proof-of-work limit (bits header field) * Negative proof-of-work limit (bits header field) * Two coinbase transactions * Duplicate transactions * Spend from transaction that does not exist * Timestamp exactly at and one second after the median time * Blocks with same hash via merkle root tricks * Spend from transaction index that is out of range * Transaction that spends more that its inputs provide * Transaction with same hash as an existing tx that has not been fully spent (BIP0030) * Non-final coinbase and non-coinbase txns * Max size block with canonical encoding which exceeds max size with non-canonical encoding * Spend from transaction earlier in same block * Spend from transaction later in same block * Double spend transaction from earlier in same block * Coinbase that pays more than subsidy + fees * Coinbase that includes subsidy + fees * Invalid opcode in dead execution path * Reorganization of txns with OP_RETURN outputs * Spend of an OP_RETURN output * Transaction with multiple OP_RETURN outputs * Large max-sized block reorganization test (disabled by default since it takes a long time and a lot of memory to run) Finally, the README.md files in the main and docs directories have been updated to reflect the use of the new testing framework.
2016-09-14 01:11:12 +02:00
tx.AddTxIn(&wire.TxIn{
// Coinbase transactions have no inputs, so previous outpoint is
// zero hash and max index.
PreviousOutPoint: *wire.NewOutPoint(&chainhash.Hash{},
wire.MaxPrevOutIndex),
Sequence: wire.MaxTxInSequenceNum,
SignatureScript: coinbaseScript,
})
tx.AddTxOut(&wire.TxOut{
Value: blockchain.CalcBlockSubsidy(blockHeight, g.params),
PkScript: opTrueScript,
})
return tx
}
// calcMerkleRoot creates a merkle tree from the slice of transactions and
// returns the root of the tree.
func calcMerkleRoot(txns []*wire.MsgTx) chainhash.Hash {
if len(txns) == 0 {
return chainhash.Hash{}
}
utilTxns := make([]*btcutil.Tx, 0, len(txns))
for _, tx := range txns {
utilTxns = append(utilTxns, btcutil.NewTx(tx))
}
merkles := blockchain.BuildMerkleTreeStore(utilTxns, false)
blockchain: Add block validation infrastructure. This adds a full-blown testing infrastructure in order to test consensus validation rules. It is built around the idea of dynamically generating full blocks that target specific rules linked together to form a block chain. In order to properly test the rules, each test instance starts with a valid block that is then modified in the specific way needed to test a specific rule. Blocks which exercise following rules have been added for this initial version. These tests were largely ported from the original Java-based 'official' block acceptance tests as well as some additional tests available in the Core python port. It is expected that further tests can be added over time as consensus rules change. * Enough valid blocks to have a stable base of mature coinbases to spend for futher tests * Basic forking and chain reorganization * Double spends on forks * Too much proof-of-work coinbase (extending main chain, in block that forces a reorg, and in a valid fork) * Max and too many signature operations via various combinations of OP_CHECKSIG, OP_MULTISIG, OP_CHECKSIGVERIFY, and OP_MULTISIGVERIFY * Too many and max signature operations with offending sigop after invalid data push * Max and too many signature operations via pay-to-script-hash redeem scripts * Attempt to spend tx created on a different fork * Attempt to spend immature coinbase (on main chain and fork) * Max size block and block that exceeds the max size * Children of rejected blocks are either orphans or rejected * Coinbase script too small and too large * Max length coinbase script * Attempt to spend tx in blocks that failed to connect * Valid non-coinbase tx in place of coinbase * Block with no transactions * Invalid proof-of-work * Block with a timestamp too far in the future * Invalid merkle root * Invalid proof-of-work limit (bits header field) * Negative proof-of-work limit (bits header field) * Two coinbase transactions * Duplicate transactions * Spend from transaction that does not exist * Timestamp exactly at and one second after the median time * Blocks with same hash via merkle root tricks * Spend from transaction index that is out of range * Transaction that spends more that its inputs provide * Transaction with same hash as an existing tx that has not been fully spent (BIP0030) * Non-final coinbase and non-coinbase txns * Max size block with canonical encoding which exceeds max size with non-canonical encoding * Spend from transaction earlier in same block * Spend from transaction later in same block * Double spend transaction from earlier in same block * Coinbase that pays more than subsidy + fees * Coinbase that includes subsidy + fees * Invalid opcode in dead execution path * Reorganization of txns with OP_RETURN outputs * Spend of an OP_RETURN output * Transaction with multiple OP_RETURN outputs * Large max-sized block reorganization test (disabled by default since it takes a long time and a lot of memory to run) Finally, the README.md files in the main and docs directories have been updated to reflect the use of the new testing framework.
2016-09-14 01:11:12 +02:00
return *merkles[len(merkles)-1]
}
// solveBlock attempts to find a nonce which makes the passed block header hash
// to a value less than the target difficulty. When a successful solution is
// found true is returned and the nonce field of the passed header is updated
// with the solution. False is returned if no solution exists.
//
// NOTE: This function will never solve blocks with a nonce of 0. This is done
// so the 'nextBlock' function can properly detect when a nonce was modified by
// a munge function.
func solveBlock(header *wire.BlockHeader) bool {
// sbResult is used by the solver goroutines to send results.
type sbResult struct {
found bool
nonce uint32
}
// solver accepts a block header and a nonce range to test. It is
// intended to be run as a goroutine.
targetDifficulty := blockchain.CompactToBig(header.Bits)
quit := make(chan bool)
results := make(chan sbResult)
solver := func(hdr wire.BlockHeader, startNonce, stopNonce uint32) {
// We need to modify the nonce field of the header, so make sure
// we work with a copy of the original header.
for i := startNonce; i >= startNonce && i <= stopNonce; i++ {
select {
case <-quit:
return
default:
hdr.Nonce = i
hash := hdr.BlockHash()
if blockchain.HashToBig(&hash).Cmp(
targetDifficulty) <= 0 {
results <- sbResult{true, i}
return
}
}
}
results <- sbResult{false, 0}
}
startNonce := uint32(1)
stopNonce := uint32(math.MaxUint32)
numCores := uint32(runtime.NumCPU())
noncesPerCore := (stopNonce - startNonce) / numCores
for i := uint32(0); i < numCores; i++ {
rangeStart := startNonce + (noncesPerCore * i)
rangeStop := startNonce + (noncesPerCore * (i + 1)) - 1
if i == numCores-1 {
rangeStop = stopNonce
}
go solver(*header, rangeStart, rangeStop)
}
for i := uint32(0); i < numCores; i++ {
result := <-results
if result.found {
close(quit)
header.Nonce = result.nonce
return true
}
}
return false
}
// additionalCoinbase returns a function that itself takes a block and
// modifies it by adding the provided amount to coinbase subsidy.
func additionalCoinbase(amount btcutil.Amount) func(*wire.MsgBlock) {
return func(b *wire.MsgBlock) {
// Increase the first proof-of-work coinbase subsidy by the
// provided amount.
b.Transactions[0].TxOut[0].Value += int64(amount)
}
}
// additionalSpendFee returns a function that itself takes a block and modifies
// it by adding the provided fee to the spending transaction.
//
// NOTE: The coinbase value is NOT updated to reflect the additional fee. Use
// 'additionalCoinbase' for that purpose.
func additionalSpendFee(fee btcutil.Amount) func(*wire.MsgBlock) {
return func(b *wire.MsgBlock) {
// Increase the fee of the spending transaction by reducing the
// amount paid.
if int64(fee) > b.Transactions[1].TxOut[0].Value {
panic(fmt.Sprintf("additionalSpendFee: fee of %d "+
"exceeds available spend transaction value",
fee))
}
b.Transactions[1].TxOut[0].Value -= int64(fee)
}
}
// replaceSpendScript returns a function that itself takes a block and modifies
// it by replacing the public key script of the spending transaction.
func replaceSpendScript(pkScript []byte) func(*wire.MsgBlock) {
return func(b *wire.MsgBlock) {
b.Transactions[1].TxOut[0].PkScript = pkScript
}
}
// replaceCoinbaseSigScript returns a function that itself takes a block and
// modifies it by replacing the signature key script of the coinbase.
func replaceCoinbaseSigScript(script []byte) func(*wire.MsgBlock) {
return func(b *wire.MsgBlock) {
b.Transactions[0].TxIn[0].SignatureScript = script
}
}
// additionalTx returns a function that itself takes a block and modifies it by
// adding the the provided transaction.
func additionalTx(tx *wire.MsgTx) func(*wire.MsgBlock) {
return func(b *wire.MsgBlock) {
b.AddTransaction(tx)
}
}
// createSpendTx creates a transaction that spends from the provided spendable
// output and includes an additional unique OP_RETURN output to ensure the
// transaction ends up with a unique hash. The script is a simple OP_TRUE
// script which avoids the need to track addresses and signature scripts in the
// tests.
func createSpendTx(spend *spendableOut, fee btcutil.Amount) *wire.MsgTx {
spendTx := wire.NewMsgTx(1)
blockchain: Add block validation infrastructure. This adds a full-blown testing infrastructure in order to test consensus validation rules. It is built around the idea of dynamically generating full blocks that target specific rules linked together to form a block chain. In order to properly test the rules, each test instance starts with a valid block that is then modified in the specific way needed to test a specific rule. Blocks which exercise following rules have been added for this initial version. These tests were largely ported from the original Java-based 'official' block acceptance tests as well as some additional tests available in the Core python port. It is expected that further tests can be added over time as consensus rules change. * Enough valid blocks to have a stable base of mature coinbases to spend for futher tests * Basic forking and chain reorganization * Double spends on forks * Too much proof-of-work coinbase (extending main chain, in block that forces a reorg, and in a valid fork) * Max and too many signature operations via various combinations of OP_CHECKSIG, OP_MULTISIG, OP_CHECKSIGVERIFY, and OP_MULTISIGVERIFY * Too many and max signature operations with offending sigop after invalid data push * Max and too many signature operations via pay-to-script-hash redeem scripts * Attempt to spend tx created on a different fork * Attempt to spend immature coinbase (on main chain and fork) * Max size block and block that exceeds the max size * Children of rejected blocks are either orphans or rejected * Coinbase script too small and too large * Max length coinbase script * Attempt to spend tx in blocks that failed to connect * Valid non-coinbase tx in place of coinbase * Block with no transactions * Invalid proof-of-work * Block with a timestamp too far in the future * Invalid merkle root * Invalid proof-of-work limit (bits header field) * Negative proof-of-work limit (bits header field) * Two coinbase transactions * Duplicate transactions * Spend from transaction that does not exist * Timestamp exactly at and one second after the median time * Blocks with same hash via merkle root tricks * Spend from transaction index that is out of range * Transaction that spends more that its inputs provide * Transaction with same hash as an existing tx that has not been fully spent (BIP0030) * Non-final coinbase and non-coinbase txns * Max size block with canonical encoding which exceeds max size with non-canonical encoding * Spend from transaction earlier in same block * Spend from transaction later in same block * Double spend transaction from earlier in same block * Coinbase that pays more than subsidy + fees * Coinbase that includes subsidy + fees * Invalid opcode in dead execution path * Reorganization of txns with OP_RETURN outputs * Spend of an OP_RETURN output * Transaction with multiple OP_RETURN outputs * Large max-sized block reorganization test (disabled by default since it takes a long time and a lot of memory to run) Finally, the README.md files in the main and docs directories have been updated to reflect the use of the new testing framework.
2016-09-14 01:11:12 +02:00
spendTx.AddTxIn(&wire.TxIn{
PreviousOutPoint: spend.prevOut,
Sequence: wire.MaxTxInSequenceNum,
SignatureScript: nil,
})
spendTx.AddTxOut(wire.NewTxOut(int64(spend.amount-fee),
opTrueScript))
spendTx.AddTxOut(wire.NewTxOut(0, uniqueOpReturnScript()))
return spendTx
}
// createSpendTxForTx creates a transaction that spends from the first output of
// the provided transaction and includes an additional unique OP_RETURN output
// to ensure the transaction ends up with a unique hash. The public key script
// is a simple OP_TRUE script which avoids the need to track addresses and
// signature scripts in the tests. The signature script is nil.
func createSpendTxForTx(tx *wire.MsgTx, fee btcutil.Amount) *wire.MsgTx {
spend := makeSpendableOutForTx(tx, 0)
return createSpendTx(&spend, fee)
}
// nextBlock builds a new block that extends the current tip associated with the
// generator and updates the generator's tip to the newly generated block.
//
// The block will include the following:
// - A coinbase that pays the required subsidy to an OP_TRUE script
// - When a spendable output is provided:
// - A transaction that spends from the provided output the following outputs:
// - One that pays the inputs amount minus 1 atom to an OP_TRUE script
// - One that contains an OP_RETURN output with a random uint64 in order to
// ensure the transaction has a unique hash
//
// Additionally, if one or more munge functions are specified, they will be
// invoked with the block prior to solving it. This provides callers with the
// opportunity to modify the block which is especially useful for testing.
//
// In order to simply the logic in the munge functions, the following rules are
// applied after all munge functions have been invoked:
// - The merkle root will be recalculated unless it was manually changed
// - The block will be solved unless the nonce was changed
func (g *testGenerator) nextBlock(blockName string, spend *spendableOut, mungers ...func(*wire.MsgBlock)) *wire.MsgBlock {
// Create coinbase transaction for the block using any additional
// subsidy if specified.
nextHeight := g.tipHeight + 1
coinbaseTx := g.createCoinbaseTx(nextHeight)
txns := []*wire.MsgTx{coinbaseTx}
if spend != nil {
// Create the transaction with a fee of 1 atom for the
// miner and increase the coinbase subsidy accordingly.
fee := btcutil.Amount(1)
coinbaseTx.TxOut[0].Value += int64(fee)
// Create a transaction that spends from the provided spendable
// output and includes an additional unique OP_RETURN output to
// ensure the transaction ends up with a unique hash, then add
// add it to the list of transactions to include in the block.
// The script is a simple OP_TRUE script in order to avoid the
// need to track addresses and signature scripts in the tests.
txns = append(txns, createSpendTx(spend, fee))
}
// Use a timestamp that is one second after the previous block unless
// this is the first block in which case the current time is used.
var ts time.Time
if nextHeight == 1 {
ts = time.Unix(time.Now().Unix(), 0)
} else {
ts = g.tip.Header.Timestamp.Add(time.Second)
}
block := wire.MsgBlock{
Header: wire.BlockHeader{
Version: 1,
PrevBlock: g.tip.BlockHash(),
MerkleRoot: calcMerkleRoot(txns),
Bits: g.params.PowLimitBits,
Timestamp: ts,
Nonce: 0, // To be solved.
},
Transactions: txns,
}
// Perform any block munging just before solving. Only recalculate the
// merkle root if it wasn't manually changed by a munge function.
curMerkleRoot := block.Header.MerkleRoot
curNonce := block.Header.Nonce
for _, f := range mungers {
f(&block)
}
if block.Header.MerkleRoot == curMerkleRoot {
block.Header.MerkleRoot = calcMerkleRoot(block.Transactions)
}
// Only solve the block if the nonce wasn't manually changed by a munge
// function.
if block.Header.Nonce == curNonce && !solveBlock(&block.Header) {
panic(fmt.Sprintf("Unable to solve block at height %d",
nextHeight))
}
// Update generator state and return the block.
blockHash := block.BlockHash()
g.blocks[blockHash] = &block
g.blocksByName[blockName] = &block
g.blockHeights[blockName] = nextHeight
g.tip = &block
g.tipName = blockName
g.tipHeight = nextHeight
return &block
}
// updateBlockState manually updates the generator state to remove all internal
// map references to a block via its old hash and insert new ones for the new
// block hash. This is useful if the test code has to manually change a block
// after 'nextBlock' has returned.
func (g *testGenerator) updateBlockState(oldBlockName string, oldBlockHash chainhash.Hash, newBlockName string, newBlock *wire.MsgBlock) {
// Look up the height from the existing entries.
blockHeight := g.blockHeights[oldBlockName]
// Remove existing entries.
delete(g.blocks, oldBlockHash)
delete(g.blocksByName, oldBlockName)
delete(g.blockHeights, oldBlockName)
// Add new entries.
newBlockHash := newBlock.BlockHash()
g.blocks[newBlockHash] = newBlock
g.blocksByName[newBlockName] = newBlock
g.blockHeights[newBlockName] = blockHeight
}
// setTip changes the tip of the instance to the block with the provided name.
// This is useful since the tip is used for things such as generating subsequent
// blocks.
func (g *testGenerator) setTip(blockName string) {
g.tip = g.blocksByName[blockName]
g.tipName = blockName
g.tipHeight = g.blockHeights[blockName]
}
// oldestCoinbaseOuts removes the oldest coinbase output that was previously
// saved to the generator and returns the set as a slice.
func (g *testGenerator) oldestCoinbaseOut() spendableOut {
op := g.spendableOuts[0]
g.spendableOuts = g.spendableOuts[1:]
return op
}
// saveTipCoinbaseOut adds the coinbase tx output in the current tip block to
// the list of spendable outputs.
func (g *testGenerator) saveTipCoinbaseOut() {
g.spendableOuts = append(g.spendableOuts, makeSpendableOut(g.tip, 0, 0))
g.prevCollectedHash = g.tip.BlockHash()
}
// saveSpendableCoinbaseOuts adds all coinbase outputs from the last block that
// had its coinbase tx output colleted to the current tip. This is useful to
// batch the collection of coinbase outputs once the tests reach a stable point
// so they don't have to manually add them for the right tests which will
// ultimately end up being the best chain.
func (g *testGenerator) saveSpendableCoinbaseOuts() {
// Ensure tip is reset to the current one when done.
curTipName := g.tipName
defer g.setTip(curTipName)
// Loop through the ancestors of the current tip until the
// reaching the block that has already had the coinbase outputs
// collected.
var collectBlocks []*wire.MsgBlock
for b := g.tip; b != nil; b = g.blocks[b.Header.PrevBlock] {
if b.BlockHash() == g.prevCollectedHash {
break
}
collectBlocks = append(collectBlocks, b)
}
for i := range collectBlocks {
g.tip = collectBlocks[len(collectBlocks)-1-i]
g.saveTipCoinbaseOut()
}
}
// nonCanonicalVarInt return a variable-length encoded integer that is encoded
// with 9 bytes even though it could be encoded with a minimal canonical
// encoding.
func nonCanonicalVarInt(val uint32) []byte {
var rv [9]byte
rv[0] = 0xff
binary.LittleEndian.PutUint64(rv[1:], uint64(val))
return rv[:]
}
// encodeNonCanonicalBlock serializes the block in a non-canonical way by
// encoding the number of transactions using a variable-length encoded integer
// with 9 bytes even though it should be encoded with a minimal canonical
// encoding.
func encodeNonCanonicalBlock(b *wire.MsgBlock) []byte {
var buf bytes.Buffer
b.Header.BtcEncode(&buf, 0, wire.BaseEncoding)
blockchain: Add block validation infrastructure. This adds a full-blown testing infrastructure in order to test consensus validation rules. It is built around the idea of dynamically generating full blocks that target specific rules linked together to form a block chain. In order to properly test the rules, each test instance starts with a valid block that is then modified in the specific way needed to test a specific rule. Blocks which exercise following rules have been added for this initial version. These tests were largely ported from the original Java-based 'official' block acceptance tests as well as some additional tests available in the Core python port. It is expected that further tests can be added over time as consensus rules change. * Enough valid blocks to have a stable base of mature coinbases to spend for futher tests * Basic forking and chain reorganization * Double spends on forks * Too much proof-of-work coinbase (extending main chain, in block that forces a reorg, and in a valid fork) * Max and too many signature operations via various combinations of OP_CHECKSIG, OP_MULTISIG, OP_CHECKSIGVERIFY, and OP_MULTISIGVERIFY * Too many and max signature operations with offending sigop after invalid data push * Max and too many signature operations via pay-to-script-hash redeem scripts * Attempt to spend tx created on a different fork * Attempt to spend immature coinbase (on main chain and fork) * Max size block and block that exceeds the max size * Children of rejected blocks are either orphans or rejected * Coinbase script too small and too large * Max length coinbase script * Attempt to spend tx in blocks that failed to connect * Valid non-coinbase tx in place of coinbase * Block with no transactions * Invalid proof-of-work * Block with a timestamp too far in the future * Invalid merkle root * Invalid proof-of-work limit (bits header field) * Negative proof-of-work limit (bits header field) * Two coinbase transactions * Duplicate transactions * Spend from transaction that does not exist * Timestamp exactly at and one second after the median time * Blocks with same hash via merkle root tricks * Spend from transaction index that is out of range * Transaction that spends more that its inputs provide * Transaction with same hash as an existing tx that has not been fully spent (BIP0030) * Non-final coinbase and non-coinbase txns * Max size block with canonical encoding which exceeds max size with non-canonical encoding * Spend from transaction earlier in same block * Spend from transaction later in same block * Double spend transaction from earlier in same block * Coinbase that pays more than subsidy + fees * Coinbase that includes subsidy + fees * Invalid opcode in dead execution path * Reorganization of txns with OP_RETURN outputs * Spend of an OP_RETURN output * Transaction with multiple OP_RETURN outputs * Large max-sized block reorganization test (disabled by default since it takes a long time and a lot of memory to run) Finally, the README.md files in the main and docs directories have been updated to reflect the use of the new testing framework.
2016-09-14 01:11:12 +02:00
buf.Write(nonCanonicalVarInt(uint32(len(b.Transactions))))
for _, tx := range b.Transactions {
tx.BtcEncode(&buf, 0, wire.BaseEncoding)
blockchain: Add block validation infrastructure. This adds a full-blown testing infrastructure in order to test consensus validation rules. It is built around the idea of dynamically generating full blocks that target specific rules linked together to form a block chain. In order to properly test the rules, each test instance starts with a valid block that is then modified in the specific way needed to test a specific rule. Blocks which exercise following rules have been added for this initial version. These tests were largely ported from the original Java-based 'official' block acceptance tests as well as some additional tests available in the Core python port. It is expected that further tests can be added over time as consensus rules change. * Enough valid blocks to have a stable base of mature coinbases to spend for futher tests * Basic forking and chain reorganization * Double spends on forks * Too much proof-of-work coinbase (extending main chain, in block that forces a reorg, and in a valid fork) * Max and too many signature operations via various combinations of OP_CHECKSIG, OP_MULTISIG, OP_CHECKSIGVERIFY, and OP_MULTISIGVERIFY * Too many and max signature operations with offending sigop after invalid data push * Max and too many signature operations via pay-to-script-hash redeem scripts * Attempt to spend tx created on a different fork * Attempt to spend immature coinbase (on main chain and fork) * Max size block and block that exceeds the max size * Children of rejected blocks are either orphans or rejected * Coinbase script too small and too large * Max length coinbase script * Attempt to spend tx in blocks that failed to connect * Valid non-coinbase tx in place of coinbase * Block with no transactions * Invalid proof-of-work * Block with a timestamp too far in the future * Invalid merkle root * Invalid proof-of-work limit (bits header field) * Negative proof-of-work limit (bits header field) * Two coinbase transactions * Duplicate transactions * Spend from transaction that does not exist * Timestamp exactly at and one second after the median time * Blocks with same hash via merkle root tricks * Spend from transaction index that is out of range * Transaction that spends more that its inputs provide * Transaction with same hash as an existing tx that has not been fully spent (BIP0030) * Non-final coinbase and non-coinbase txns * Max size block with canonical encoding which exceeds max size with non-canonical encoding * Spend from transaction earlier in same block * Spend from transaction later in same block * Double spend transaction from earlier in same block * Coinbase that pays more than subsidy + fees * Coinbase that includes subsidy + fees * Invalid opcode in dead execution path * Reorganization of txns with OP_RETURN outputs * Spend of an OP_RETURN output * Transaction with multiple OP_RETURN outputs * Large max-sized block reorganization test (disabled by default since it takes a long time and a lot of memory to run) Finally, the README.md files in the main and docs directories have been updated to reflect the use of the new testing framework.
2016-09-14 01:11:12 +02:00
}
return buf.Bytes()
}
// cloneBlock returns a deep copy of the provided block.
func cloneBlock(b *wire.MsgBlock) wire.MsgBlock {
var blockCopy wire.MsgBlock
blockCopy.Header = b.Header
for _, tx := range b.Transactions {
blockCopy.AddTransaction(tx.Copy())
}
return blockCopy
}
// repeatOpcode returns a byte slice with the provided opcode repeated the
// specified number of times.
func repeatOpcode(opcode uint8, numRepeats int) []byte {
return bytes.Repeat([]byte{opcode}, numRepeats)
}
// assertScriptSigOpsCount panics if the provided script does not have the
// specified number of signature operations.
func assertScriptSigOpsCount(script []byte, expected int) {
numSigOps := txscript.GetSigOpCount(script)
if numSigOps != expected {
_, file, line, _ := runtime.Caller(1)
panic(fmt.Sprintf("assertion failed at %s:%d: generated number "+
"of sigops for script is %d instead of expected %d",
file, line, numSigOps, expected))
}
}
// countBlockSigOps returns the number of legacy signature operations in the
// scripts in the passed block.
func countBlockSigOps(block *wire.MsgBlock) int {
totalSigOps := 0
for _, tx := range block.Transactions {
for _, txIn := range tx.TxIn {
numSigOps := txscript.GetSigOpCount(txIn.SignatureScript)
totalSigOps += numSigOps
}
for _, txOut := range tx.TxOut {
numSigOps := txscript.GetSigOpCount(txOut.PkScript)
totalSigOps += numSigOps
}
}
return totalSigOps
}
// assertTipBlockSigOpsCount panics if the current tip block associated with the
// generator does not have the specified number of signature operations.
func (g *testGenerator) assertTipBlockSigOpsCount(expected int) {
numSigOps := countBlockSigOps(g.tip)
if numSigOps != expected {
panic(fmt.Sprintf("generated number of sigops for block %q "+
"(height %d) is %d instead of expected %d", g.tipName,
g.tipHeight, numSigOps, expected))
}
}
// assertTipBlockSize panics if the if the current tip block associated with the
// generator does not have the specified size when serialized.
func (g *testGenerator) assertTipBlockSize(expected int) {
serializeSize := g.tip.SerializeSize()
if serializeSize != expected {
panic(fmt.Sprintf("block size of block %q (height %d) is %d "+
"instead of expected %d", g.tipName, g.tipHeight,
serializeSize, expected))
}
}
// assertTipNonCanonicalBlockSize panics if the if the current tip block
// associated with the generator does not have the specified non-canonical size
// when serialized.
func (g *testGenerator) assertTipNonCanonicalBlockSize(expected int) {
serializeSize := len(encodeNonCanonicalBlock(g.tip))
if serializeSize != expected {
panic(fmt.Sprintf("block size of block %q (height %d) is %d "+
"instead of expected %d", g.tipName, g.tipHeight,
serializeSize, expected))
}
}
// assertTipBlockNumTxns panics if the number of transactions in the current tip
// block associated with the generator does not match the specified value.
func (g *testGenerator) assertTipBlockNumTxns(expected int) {
numTxns := len(g.tip.Transactions)
if numTxns != expected {
panic(fmt.Sprintf("number of txns in block %q (height %d) is "+
"%d instead of expected %d", g.tipName, g.tipHeight,
numTxns, expected))
}
}
// assertTipBlockHash panics if the current tip block associated with the
// generator does not match the specified hash.
func (g *testGenerator) assertTipBlockHash(expected chainhash.Hash) {
hash := g.tip.BlockHash()
if hash != expected {
panic(fmt.Sprintf("block hash of block %q (height %d) is %v "+
"instead of expected %v", g.tipName, g.tipHeight, hash,
expected))
}
}
// assertTipBlockMerkleRoot panics if the merkle root in header of the current
// tip block associated with the generator does not match the specified hash.
func (g *testGenerator) assertTipBlockMerkleRoot(expected chainhash.Hash) {
hash := g.tip.Header.MerkleRoot
if hash != expected {
panic(fmt.Sprintf("merkle root of block %q (height %d) is %v "+
"instead of expected %v", g.tipName, g.tipHeight, hash,
expected))
}
}
// assertTipBlockTxOutOpReturn panics if the current tip block associated with
// the generator does not have an OP_RETURN script for the transaction output at
// the provided tx index and output index.
func (g *testGenerator) assertTipBlockTxOutOpReturn(txIndex, txOutIndex uint32) {
if txIndex >= uint32(len(g.tip.Transactions)) {
panic(fmt.Sprintf("Transaction index %d in block %q "+
"(height %d) does not exist", txIndex, g.tipName,
g.tipHeight))
}
tx := g.tip.Transactions[txIndex]
if txOutIndex >= uint32(len(tx.TxOut)) {
panic(fmt.Sprintf("transaction index %d output %d in block %q "+
"(height %d) does not exist", txIndex, txOutIndex,
g.tipName, g.tipHeight))
}
txOut := tx.TxOut[txOutIndex]
if txOut.PkScript[0] != txscript.OP_RETURN {
panic(fmt.Sprintf("transaction index %d output %d in block %q "+
"(height %d) is not an OP_RETURN", txIndex, txOutIndex,
g.tipName, g.tipHeight))
}
}
// Generate returns a slice of tests that can be used to exercise the consensus
// validation rules. The tests are intended to be flexible enough to allow both
// unit-style tests directly against the blockchain code as well as integration
// style tests over the peer-to-peer network. To achieve that goal, each test
// contains additional information about the expected result, however that
// information can be ignored when doing comparison tests between two
// independent versions over the peer-to-peer network.
func Generate(includeLargeReorg bool) (tests [][]TestInstance, err error) {
// In order to simplify the generation code which really should never
// fail unless the test code itself is broken, panics are used
// internally. This deferred func ensures any panics don't escape the
// generator by replacing the named error return with the underlying
// panic error.
defer func() {
if r := recover(); r != nil {
tests = nil
switch rt := r.(type) {
case string:
err = errors.New(rt)
case error:
err = rt
default:
err = errors.New("Unknown panic")
}
}
}()
// Create a test generator instance initialized with the genesis block
// as the tip.
g, err := makeTestGenerator(regressionNetParams)
if err != nil {
return nil, err
}
// Define some convenience helper functions to return an individual test
// instance that has the described characteristics.
//
// acceptBlock creates a test instance that expects the provided block
// to be accepted by the consensus rules.
//
// rejectBlock creates a test instance that expects the provided block
// to be rejected by the consensus rules.
//
// rejectNonCanonicalBlock creates a test instance that encodes the
// provided block using a non-canonical encoded as described by the
// encodeNonCanonicalBlock function and expected it to be rejected.
//
// orphanOrRejectBlock creates a test instance that expected the
// provided block to either by accepted as an orphan or rejected by the
// consensus rules.
//
// expectTipBlock creates a test instance that expects the provided
// block to be the current tip of the block chain.
acceptBlock := func(blockName string, block *wire.MsgBlock, isMainChain, isOrphan bool) TestInstance {
blockHeight := g.blockHeights[blockName]
return AcceptedBlock{blockName, block, blockHeight, isMainChain,
isOrphan}
}
rejectBlock := func(blockName string, block *wire.MsgBlock, code blockchain.ErrorCode) TestInstance {
blockHeight := g.blockHeights[blockName]
return RejectedBlock{blockName, block, blockHeight, code}
}
rejectNonCanonicalBlock := func(blockName string, block *wire.MsgBlock) TestInstance {
blockHeight := g.blockHeights[blockName]
encoded := encodeNonCanonicalBlock(block)
return RejectedNonCanonicalBlock{blockName, encoded, blockHeight}
}
orphanOrRejectBlock := func(blockName string, block *wire.MsgBlock) TestInstance {
blockHeight := g.blockHeights[blockName]
return OrphanOrRejectedBlock{blockName, block, blockHeight}
}
expectTipBlock := func(blockName string, block *wire.MsgBlock) TestInstance {
blockHeight := g.blockHeights[blockName]
return ExpectedTip{blockName, block, blockHeight}
}
// Define some convenience helper functions to populate the tests slice
// with test instances that have the described characteristics.
//
// accepted creates and appends a single acceptBlock test instance for
// the current tip which expects the block to be accepted to the main
// chain.
//
// acceptedToSideChainWithExpectedTip creates an appends a two-instance
// test. The first instance is an acceptBlock test instance for the
// current tip which expects the block to be accepted to a side chain.
// The second instance is an expectBlockTip test instance for provided
// values.
//
// rejected creates and appends a single rejectBlock test instance for
// the current tip.
//
// rejectedNonCanonical creates and appends a single
// rejectNonCanonicalBlock test instance for the current tip.
//
// orphanedOrRejected creates and appends a single orphanOrRejectBlock
// test instance for the current tip.
accepted := func() {
tests = append(tests, []TestInstance{
acceptBlock(g.tipName, g.tip, true, false),
})
}
acceptedToSideChainWithExpectedTip := func(tipName string) {
tests = append(tests, []TestInstance{
acceptBlock(g.tipName, g.tip, false, false),
expectTipBlock(tipName, g.blocksByName[tipName]),
})
}
rejected := func(code blockchain.ErrorCode) {
tests = append(tests, []TestInstance{
rejectBlock(g.tipName, g.tip, code),
})
}
rejectedNonCanonical := func() {
tests = append(tests, []TestInstance{
rejectNonCanonicalBlock(g.tipName, g.tip),
})
}
orphanedOrRejected := func() {
tests = append(tests, []TestInstance{
orphanOrRejectBlock(g.tipName, g.tip),
})
}
// ---------------------------------------------------------------------
// Generate enough blocks to have mature coinbase outputs to work with.
//
// genesis -> bm0 -> bm1 -> ... -> bm99
// ---------------------------------------------------------------------
coinbaseMaturity := g.params.CoinbaseMaturity
var testInstances []TestInstance
for i := uint16(0); i < coinbaseMaturity; i++ {
blockName := fmt.Sprintf("bm%d", i)
g.nextBlock(blockName, nil)
g.saveTipCoinbaseOut()
testInstances = append(testInstances, acceptBlock(g.tipName,
g.tip, true, false))
}
tests = append(tests, testInstances)
// Collect spendable outputs. This simplifies the code below.
var outs []*spendableOut
for i := uint16(0); i < coinbaseMaturity; i++ {
op := g.oldestCoinbaseOut()
outs = append(outs, &op)
}
// ---------------------------------------------------------------------
// Basic forking and reorg tests.
// ---------------------------------------------------------------------
// ---------------------------------------------------------------------
// The comments below identify the structure of the chain being built.
//
// The values in parenthesis repesent which outputs are being spent.
//
// For example, b1(0) indicates the first collected spendable output
// which, due to the code above to create the correct number of blocks,
// is the first output that can be spent at the current block height due
// to the coinbase maturity requirement.
// ---------------------------------------------------------------------
// Start by building a couple of blocks at current tip (value in parens
// is which output is spent):
//
// ... -> b1(0) -> b2(1)
g.nextBlock("b1", outs[0])
accepted()
g.nextBlock("b2", outs[1])
accepted()
// Create a fork from b1. There should not be a reorg since b2 was seen
// first.
//
// ... -> b1(0) -> b2(1)
// \-> b3(1)
g.setTip("b1")
g.nextBlock("b3", outs[1])
b3Tx1Out := makeSpendableOut(g.tip, 1, 0)
acceptedToSideChainWithExpectedTip("b2")
// Extend b3 fork to make the alternative chain longer and force reorg.
//
// ... -> b1(0) -> b2(1)
// \-> b3(1) -> b4(2)
g.nextBlock("b4", outs[2])
accepted()
// Extend b2 fork twice to make first chain longer and force reorg.
//
// ... -> b1(0) -> b2(1) -> b5(2) -> b6(3)
// \-> b3(1) -> b4(2)
g.setTip("b2")
g.nextBlock("b5", outs[2])
acceptedToSideChainWithExpectedTip("b4")
g.nextBlock("b6", outs[3])
accepted()
// ---------------------------------------------------------------------
// Double spend tests.
// ---------------------------------------------------------------------
// Create a fork that double spends.
//
// ... -> b1(0) -> b2(1) -> b5(2) -> b6(3)
// \-> b7(2) -> b8(4)
// \-> b3(1) -> b4(2)
g.setTip("b5")
g.nextBlock("b7", outs[2])
acceptedToSideChainWithExpectedTip("b6")
g.nextBlock("b8", outs[4])
rejected(blockchain.ErrMissingTxOut)
blockchain: Add block validation infrastructure. This adds a full-blown testing infrastructure in order to test consensus validation rules. It is built around the idea of dynamically generating full blocks that target specific rules linked together to form a block chain. In order to properly test the rules, each test instance starts with a valid block that is then modified in the specific way needed to test a specific rule. Blocks which exercise following rules have been added for this initial version. These tests were largely ported from the original Java-based 'official' block acceptance tests as well as some additional tests available in the Core python port. It is expected that further tests can be added over time as consensus rules change. * Enough valid blocks to have a stable base of mature coinbases to spend for futher tests * Basic forking and chain reorganization * Double spends on forks * Too much proof-of-work coinbase (extending main chain, in block that forces a reorg, and in a valid fork) * Max and too many signature operations via various combinations of OP_CHECKSIG, OP_MULTISIG, OP_CHECKSIGVERIFY, and OP_MULTISIGVERIFY * Too many and max signature operations with offending sigop after invalid data push * Max and too many signature operations via pay-to-script-hash redeem scripts * Attempt to spend tx created on a different fork * Attempt to spend immature coinbase (on main chain and fork) * Max size block and block that exceeds the max size * Children of rejected blocks are either orphans or rejected * Coinbase script too small and too large * Max length coinbase script * Attempt to spend tx in blocks that failed to connect * Valid non-coinbase tx in place of coinbase * Block with no transactions * Invalid proof-of-work * Block with a timestamp too far in the future * Invalid merkle root * Invalid proof-of-work limit (bits header field) * Negative proof-of-work limit (bits header field) * Two coinbase transactions * Duplicate transactions * Spend from transaction that does not exist * Timestamp exactly at and one second after the median time * Blocks with same hash via merkle root tricks * Spend from transaction index that is out of range * Transaction that spends more that its inputs provide * Transaction with same hash as an existing tx that has not been fully spent (BIP0030) * Non-final coinbase and non-coinbase txns * Max size block with canonical encoding which exceeds max size with non-canonical encoding * Spend from transaction earlier in same block * Spend from transaction later in same block * Double spend transaction from earlier in same block * Coinbase that pays more than subsidy + fees * Coinbase that includes subsidy + fees * Invalid opcode in dead execution path * Reorganization of txns with OP_RETURN outputs * Spend of an OP_RETURN output * Transaction with multiple OP_RETURN outputs * Large max-sized block reorganization test (disabled by default since it takes a long time and a lot of memory to run) Finally, the README.md files in the main and docs directories have been updated to reflect the use of the new testing framework.
2016-09-14 01:11:12 +02:00
// ---------------------------------------------------------------------
// Too much proof-of-work coinbase tests.
// ---------------------------------------------------------------------
// Create a block that generates too coinbase.
//
// ... -> b1(0) -> b2(1) -> b5(2) -> b6(3)
// \-> b9(4)
// \-> b3(1) -> b4(2)
g.setTip("b6")
g.nextBlock("b9", outs[4], additionalCoinbase(1))
rejected(blockchain.ErrBadCoinbaseValue)
// Create a fork that ends with block that generates too much coinbase.
//
// ... -> b1(0) -> b2(1) -> b5(2) -> b6(3)
// \-> b10(3) -> b11(4)
// \-> b3(1) -> b4(2)
g.setTip("b5")
g.nextBlock("b10", outs[3])
acceptedToSideChainWithExpectedTip("b6")
g.nextBlock("b11", outs[4], additionalCoinbase(1))
rejected(blockchain.ErrBadCoinbaseValue)
// Create a fork that ends with block that generates too much coinbase
// as before, but with a valid fork first.
//
// ... -> b1(0) -> b2(1) -> b5(2) -> b6(3)
// | \-> b12(3) -> b13(4) -> b14(5)
// | (b12 added last)
// \-> b3(1) -> b4(2)
g.setTip("b5")
b12 := g.nextBlock("b12", outs[3])
b13 := g.nextBlock("b13", outs[4])
b14 := g.nextBlock("b14", outs[5], additionalCoinbase(1))
tests = append(tests, []TestInstance{
acceptBlock("b13", b13, false, true),
acceptBlock("b14", b14, false, true),
rejectBlock("b12", b12, blockchain.ErrBadCoinbaseValue),
expectTipBlock("b13", b13),
})
// ---------------------------------------------------------------------
// Checksig signature operation count tests.
// ---------------------------------------------------------------------
// Add a block with max allowed signature operations using OP_CHECKSIG.
//
// ... -> b5(2) -> b12(3) -> b13(4) -> b15(5)
// \-> b3(1) -> b4(2)
g.setTip("b13")
manySigOps := repeatOpcode(txscript.OP_CHECKSIG, maxBlockSigOps)
g.nextBlock("b15", outs[5], replaceSpendScript(manySigOps))
g.assertTipBlockSigOpsCount(maxBlockSigOps)
accepted()
// Attempt to add block with more than max allowed signature operations
// using OP_CHECKSIG.
//
// ... -> b5(2) -> b12(3) -> b13(4) -> b15(5)
// \ \-> b16(7)
// \-> b3(1) -> b4(2)
tooManySigOps := repeatOpcode(txscript.OP_CHECKSIG, maxBlockSigOps+1)
g.nextBlock("b16", outs[6], replaceSpendScript(tooManySigOps))
g.assertTipBlockSigOpsCount(maxBlockSigOps + 1)
rejected(blockchain.ErrTooManySigOps)
// ---------------------------------------------------------------------
// Cross-fork spend tests.
// ---------------------------------------------------------------------
// Create block that spends a tx created on a different fork.
//
// ... -> b5(2) -> b12(3) -> b13(4) -> b15(5)
// \ \-> b17(b3.tx[1])
// \-> b3(1) -> b4(2)
g.setTip("b15")
g.nextBlock("b17", &b3Tx1Out)
rejected(blockchain.ErrMissingTxOut)
blockchain: Add block validation infrastructure. This adds a full-blown testing infrastructure in order to test consensus validation rules. It is built around the idea of dynamically generating full blocks that target specific rules linked together to form a block chain. In order to properly test the rules, each test instance starts with a valid block that is then modified in the specific way needed to test a specific rule. Blocks which exercise following rules have been added for this initial version. These tests were largely ported from the original Java-based 'official' block acceptance tests as well as some additional tests available in the Core python port. It is expected that further tests can be added over time as consensus rules change. * Enough valid blocks to have a stable base of mature coinbases to spend for futher tests * Basic forking and chain reorganization * Double spends on forks * Too much proof-of-work coinbase (extending main chain, in block that forces a reorg, and in a valid fork) * Max and too many signature operations via various combinations of OP_CHECKSIG, OP_MULTISIG, OP_CHECKSIGVERIFY, and OP_MULTISIGVERIFY * Too many and max signature operations with offending sigop after invalid data push * Max and too many signature operations via pay-to-script-hash redeem scripts * Attempt to spend tx created on a different fork * Attempt to spend immature coinbase (on main chain and fork) * Max size block and block that exceeds the max size * Children of rejected blocks are either orphans or rejected * Coinbase script too small and too large * Max length coinbase script * Attempt to spend tx in blocks that failed to connect * Valid non-coinbase tx in place of coinbase * Block with no transactions * Invalid proof-of-work * Block with a timestamp too far in the future * Invalid merkle root * Invalid proof-of-work limit (bits header field) * Negative proof-of-work limit (bits header field) * Two coinbase transactions * Duplicate transactions * Spend from transaction that does not exist * Timestamp exactly at and one second after the median time * Blocks with same hash via merkle root tricks * Spend from transaction index that is out of range * Transaction that spends more that its inputs provide * Transaction with same hash as an existing tx that has not been fully spent (BIP0030) * Non-final coinbase and non-coinbase txns * Max size block with canonical encoding which exceeds max size with non-canonical encoding * Spend from transaction earlier in same block * Spend from transaction later in same block * Double spend transaction from earlier in same block * Coinbase that pays more than subsidy + fees * Coinbase that includes subsidy + fees * Invalid opcode in dead execution path * Reorganization of txns with OP_RETURN outputs * Spend of an OP_RETURN output * Transaction with multiple OP_RETURN outputs * Large max-sized block reorganization test (disabled by default since it takes a long time and a lot of memory to run) Finally, the README.md files in the main and docs directories have been updated to reflect the use of the new testing framework.
2016-09-14 01:11:12 +02:00
// Create block that forks and spends a tx created on a third fork.
//
// ... -> b5(2) -> b12(3) -> b13(4) -> b15(5)
// | \-> b18(b3.tx[1]) -> b19(6)
// \-> b3(1) -> b4(2)
g.setTip("b13")
g.nextBlock("b18", &b3Tx1Out)
acceptedToSideChainWithExpectedTip("b15")
g.nextBlock("b19", outs[6])
rejected(blockchain.ErrMissingTxOut)
blockchain: Add block validation infrastructure. This adds a full-blown testing infrastructure in order to test consensus validation rules. It is built around the idea of dynamically generating full blocks that target specific rules linked together to form a block chain. In order to properly test the rules, each test instance starts with a valid block that is then modified in the specific way needed to test a specific rule. Blocks which exercise following rules have been added for this initial version. These tests were largely ported from the original Java-based 'official' block acceptance tests as well as some additional tests available in the Core python port. It is expected that further tests can be added over time as consensus rules change. * Enough valid blocks to have a stable base of mature coinbases to spend for futher tests * Basic forking and chain reorganization * Double spends on forks * Too much proof-of-work coinbase (extending main chain, in block that forces a reorg, and in a valid fork) * Max and too many signature operations via various combinations of OP_CHECKSIG, OP_MULTISIG, OP_CHECKSIGVERIFY, and OP_MULTISIGVERIFY * Too many and max signature operations with offending sigop after invalid data push * Max and too many signature operations via pay-to-script-hash redeem scripts * Attempt to spend tx created on a different fork * Attempt to spend immature coinbase (on main chain and fork) * Max size block and block that exceeds the max size * Children of rejected blocks are either orphans or rejected * Coinbase script too small and too large * Max length coinbase script * Attempt to spend tx in blocks that failed to connect * Valid non-coinbase tx in place of coinbase * Block with no transactions * Invalid proof-of-work * Block with a timestamp too far in the future * Invalid merkle root * Invalid proof-of-work limit (bits header field) * Negative proof-of-work limit (bits header field) * Two coinbase transactions * Duplicate transactions * Spend from transaction that does not exist * Timestamp exactly at and one second after the median time * Blocks with same hash via merkle root tricks * Spend from transaction index that is out of range * Transaction that spends more that its inputs provide * Transaction with same hash as an existing tx that has not been fully spent (BIP0030) * Non-final coinbase and non-coinbase txns * Max size block with canonical encoding which exceeds max size with non-canonical encoding * Spend from transaction earlier in same block * Spend from transaction later in same block * Double spend transaction from earlier in same block * Coinbase that pays more than subsidy + fees * Coinbase that includes subsidy + fees * Invalid opcode in dead execution path * Reorganization of txns with OP_RETURN outputs * Spend of an OP_RETURN output * Transaction with multiple OP_RETURN outputs * Large max-sized block reorganization test (disabled by default since it takes a long time and a lot of memory to run) Finally, the README.md files in the main and docs directories have been updated to reflect the use of the new testing framework.
2016-09-14 01:11:12 +02:00
// ---------------------------------------------------------------------
// Immature coinbase tests.
// ---------------------------------------------------------------------
// Create block that spends immature coinbase.
//
// ... -> b13(4) -> b15(5)
// \-> b20(7)
g.setTip("b15")
g.nextBlock("b20", outs[7])
rejected(blockchain.ErrImmatureSpend)
// Create block that spends immature coinbase on a fork.
//
// ... -> b13(4) -> b15(5)
// \-> b21(5) -> b22(7)
g.setTip("b13")
g.nextBlock("b21", outs[5])
acceptedToSideChainWithExpectedTip("b15")
g.nextBlock("b22", outs[7])
rejected(blockchain.ErrImmatureSpend)
// ---------------------------------------------------------------------
// Max block size tests.
// ---------------------------------------------------------------------
// Create block that is the max allowed size.
//
// ... -> b15(5) -> b23(6)
g.setTip("b15")
g.nextBlock("b23", outs[6], func(b *wire.MsgBlock) {
bytesToMaxSize := maxBlockSize - b.SerializeSize() - 3
sizePadScript := repeatOpcode(0x00, bytesToMaxSize)
replaceSpendScript(sizePadScript)(b)
})
g.assertTipBlockSize(maxBlockSize)
accepted()
// Create block that is the one byte larger than max allowed size. This
// is done on a fork and should be rejected regardless.
//
// ... -> b15(5) -> b23(6)
// \-> b24(6) -> b25(7)
g.setTip("b15")
g.nextBlock("b24", outs[6], func(b *wire.MsgBlock) {
bytesToMaxSize := maxBlockSize - b.SerializeSize() - 3
sizePadScript := repeatOpcode(0x00, bytesToMaxSize+1)
replaceSpendScript(sizePadScript)(b)
})
g.assertTipBlockSize(maxBlockSize + 1)
rejected(blockchain.ErrBlockTooBig)
// Parent was rejected, so this block must either be an orphan or
// outright rejected due to an invalid parent.
g.nextBlock("b25", outs[7])
orphanedOrRejected()
// ---------------------------------------------------------------------
// Coinbase script length limits tests.
// ---------------------------------------------------------------------
// Create block that has a coinbase script that is smaller than the
// required length. This is done on a fork and should be rejected
// regardless. Also, create a block that builds on the rejected block.
//
// ... -> b15(5) -> b23(6)
// \-> b26(6) -> b27(7)
g.setTip("b15")
tooSmallCbScript := repeatOpcode(0x00, minCoinbaseScriptLen-1)
g.nextBlock("b26", outs[6], replaceCoinbaseSigScript(tooSmallCbScript))
rejected(blockchain.ErrBadCoinbaseScriptLen)
// Parent was rejected, so this block must either be an orphan or
// outright rejected due to an invalid parent.
g.nextBlock("b27", outs[7])
orphanedOrRejected()
// Create block that has a coinbase script that is larger than the
// allowed length. This is done on a fork and should be rejected
// regardless. Also, create a block that builds on the rejected block.
//
// ... -> b15(5) -> b23(6)
// \-> b28(6) -> b29(7)
g.setTip("b15")
tooLargeCbScript := repeatOpcode(0x00, maxCoinbaseScriptLen+1)
g.nextBlock("b28", outs[6], replaceCoinbaseSigScript(tooLargeCbScript))
rejected(blockchain.ErrBadCoinbaseScriptLen)
// Parent was rejected, so this block must either be an orphan or
// outright rejected due to an invalid parent.
g.nextBlock("b29", outs[7])
orphanedOrRejected()
// Create block that has a max length coinbase script.
//
// ... -> b23(6) -> b30(7)
g.setTip("b23")
maxSizeCbScript := repeatOpcode(0x00, maxCoinbaseScriptLen)
g.nextBlock("b30", outs[7], replaceCoinbaseSigScript(maxSizeCbScript))
accepted()
// ---------------------------------------------------------------------
// Multisig[Verify]/ChecksigVerifiy signature operation count tests.
// ---------------------------------------------------------------------
// Create block with max signature operations as OP_CHECKMULTISIG.
//
// ... -> b30(7) -> b31(8)
//
// OP_CHECKMULTISIG counts for 20 sigops.
manySigOps = repeatOpcode(txscript.OP_CHECKMULTISIG, maxBlockSigOps/20)
g.nextBlock("b31", outs[8], replaceSpendScript(manySigOps))
g.assertTipBlockSigOpsCount(maxBlockSigOps)
accepted()
// Create block with more than max allowed signature operations using
// OP_CHECKMULTISIG.
//
// ... -> b31(8)
// \-> b32(9)
//
// OP_CHECKMULTISIG counts for 20 sigops.
tooManySigOps = repeatOpcode(txscript.OP_CHECKMULTISIG, maxBlockSigOps/20)
tooManySigOps = append(manySigOps, txscript.OP_CHECKSIG)
g.nextBlock("b32", outs[9], replaceSpendScript(tooManySigOps))
g.assertTipBlockSigOpsCount(maxBlockSigOps + 1)
rejected(blockchain.ErrTooManySigOps)
// Create block with max signature operations as OP_CHECKMULTISIGVERIFY.
//
// ... -> b31(8) -> b33(9)
g.setTip("b31")
manySigOps = repeatOpcode(txscript.OP_CHECKMULTISIGVERIFY, maxBlockSigOps/20)
g.nextBlock("b33", outs[9], replaceSpendScript(manySigOps))
g.assertTipBlockSigOpsCount(maxBlockSigOps)
accepted()
// Create block with more than max allowed signature operations using
// OP_CHECKMULTISIGVERIFY.
//
// ... -> b33(9)
// \-> b34(10)
//
tooManySigOps = repeatOpcode(txscript.OP_CHECKMULTISIGVERIFY, maxBlockSigOps/20)
tooManySigOps = append(manySigOps, txscript.OP_CHECKSIG)
g.nextBlock("b34", outs[10], replaceSpendScript(tooManySigOps))
g.assertTipBlockSigOpsCount(maxBlockSigOps + 1)
rejected(blockchain.ErrTooManySigOps)
// Create block with max signature operations as OP_CHECKSIGVERIFY.
//
// ... -> b33(9) -> b35(10)
//
g.setTip("b33")
manySigOps = repeatOpcode(txscript.OP_CHECKSIGVERIFY, maxBlockSigOps)
g.nextBlock("b35", outs[10], replaceSpendScript(manySigOps))
g.assertTipBlockSigOpsCount(maxBlockSigOps)
accepted()
// Create block with more than max allowed signature operations using
// OP_CHECKSIGVERIFY.
//
// ... -> b35(10)
// \-> b36(11)
//
tooManySigOps = repeatOpcode(txscript.OP_CHECKSIGVERIFY, maxBlockSigOps+1)
g.nextBlock("b36", outs[11], replaceSpendScript(tooManySigOps))
g.assertTipBlockSigOpsCount(maxBlockSigOps + 1)
rejected(blockchain.ErrTooManySigOps)
// ---------------------------------------------------------------------
// Spending of tx outputs in block that failed to connect tests.
// ---------------------------------------------------------------------
// Create block that spends a transaction from a block that failed to
// connect (due to containing a double spend).
//
// ... -> b35(10)
// \-> b37(11)
// \-> b38(b37.tx[1])
//
g.setTip("b35")
doubleSpendTx := createSpendTx(outs[11], lowFee)
g.nextBlock("b37", outs[11], additionalTx(doubleSpendTx))
b37Tx1Out := makeSpendableOut(g.tip, 1, 0)
rejected(blockchain.ErrMissingTxOut)
blockchain: Add block validation infrastructure. This adds a full-blown testing infrastructure in order to test consensus validation rules. It is built around the idea of dynamically generating full blocks that target specific rules linked together to form a block chain. In order to properly test the rules, each test instance starts with a valid block that is then modified in the specific way needed to test a specific rule. Blocks which exercise following rules have been added for this initial version. These tests were largely ported from the original Java-based 'official' block acceptance tests as well as some additional tests available in the Core python port. It is expected that further tests can be added over time as consensus rules change. * Enough valid blocks to have a stable base of mature coinbases to spend for futher tests * Basic forking and chain reorganization * Double spends on forks * Too much proof-of-work coinbase (extending main chain, in block that forces a reorg, and in a valid fork) * Max and too many signature operations via various combinations of OP_CHECKSIG, OP_MULTISIG, OP_CHECKSIGVERIFY, and OP_MULTISIGVERIFY * Too many and max signature operations with offending sigop after invalid data push * Max and too many signature operations via pay-to-script-hash redeem scripts * Attempt to spend tx created on a different fork * Attempt to spend immature coinbase (on main chain and fork) * Max size block and block that exceeds the max size * Children of rejected blocks are either orphans or rejected * Coinbase script too small and too large * Max length coinbase script * Attempt to spend tx in blocks that failed to connect * Valid non-coinbase tx in place of coinbase * Block with no transactions * Invalid proof-of-work * Block with a timestamp too far in the future * Invalid merkle root * Invalid proof-of-work limit (bits header field) * Negative proof-of-work limit (bits header field) * Two coinbase transactions * Duplicate transactions * Spend from transaction that does not exist * Timestamp exactly at and one second after the median time * Blocks with same hash via merkle root tricks * Spend from transaction index that is out of range * Transaction that spends more that its inputs provide * Transaction with same hash as an existing tx that has not been fully spent (BIP0030) * Non-final coinbase and non-coinbase txns * Max size block with canonical encoding which exceeds max size with non-canonical encoding * Spend from transaction earlier in same block * Spend from transaction later in same block * Double spend transaction from earlier in same block * Coinbase that pays more than subsidy + fees * Coinbase that includes subsidy + fees * Invalid opcode in dead execution path * Reorganization of txns with OP_RETURN outputs * Spend of an OP_RETURN output * Transaction with multiple OP_RETURN outputs * Large max-sized block reorganization test (disabled by default since it takes a long time and a lot of memory to run) Finally, the README.md files in the main and docs directories have been updated to reflect the use of the new testing framework.
2016-09-14 01:11:12 +02:00
g.setTip("b35")
g.nextBlock("b38", &b37Tx1Out)
rejected(blockchain.ErrMissingTxOut)
blockchain: Add block validation infrastructure. This adds a full-blown testing infrastructure in order to test consensus validation rules. It is built around the idea of dynamically generating full blocks that target specific rules linked together to form a block chain. In order to properly test the rules, each test instance starts with a valid block that is then modified in the specific way needed to test a specific rule. Blocks which exercise following rules have been added for this initial version. These tests were largely ported from the original Java-based 'official' block acceptance tests as well as some additional tests available in the Core python port. It is expected that further tests can be added over time as consensus rules change. * Enough valid blocks to have a stable base of mature coinbases to spend for futher tests * Basic forking and chain reorganization * Double spends on forks * Too much proof-of-work coinbase (extending main chain, in block that forces a reorg, and in a valid fork) * Max and too many signature operations via various combinations of OP_CHECKSIG, OP_MULTISIG, OP_CHECKSIGVERIFY, and OP_MULTISIGVERIFY * Too many and max signature operations with offending sigop after invalid data push * Max and too many signature operations via pay-to-script-hash redeem scripts * Attempt to spend tx created on a different fork * Attempt to spend immature coinbase (on main chain and fork) * Max size block and block that exceeds the max size * Children of rejected blocks are either orphans or rejected * Coinbase script too small and too large * Max length coinbase script * Attempt to spend tx in blocks that failed to connect * Valid non-coinbase tx in place of coinbase * Block with no transactions * Invalid proof-of-work * Block with a timestamp too far in the future * Invalid merkle root * Invalid proof-of-work limit (bits header field) * Negative proof-of-work limit (bits header field) * Two coinbase transactions * Duplicate transactions * Spend from transaction that does not exist * Timestamp exactly at and one second after the median time * Blocks with same hash via merkle root tricks * Spend from transaction index that is out of range * Transaction that spends more that its inputs provide * Transaction with same hash as an existing tx that has not been fully spent (BIP0030) * Non-final coinbase and non-coinbase txns * Max size block with canonical encoding which exceeds max size with non-canonical encoding * Spend from transaction earlier in same block * Spend from transaction later in same block * Double spend transaction from earlier in same block * Coinbase that pays more than subsidy + fees * Coinbase that includes subsidy + fees * Invalid opcode in dead execution path * Reorganization of txns with OP_RETURN outputs * Spend of an OP_RETURN output * Transaction with multiple OP_RETURN outputs * Large max-sized block reorganization test (disabled by default since it takes a long time and a lot of memory to run) Finally, the README.md files in the main and docs directories have been updated to reflect the use of the new testing framework.
2016-09-14 01:11:12 +02:00
// ---------------------------------------------------------------------
// Pay-to-script-hash signature operation count tests.
// ---------------------------------------------------------------------
// Create a pay-to-script-hash redeem script that consists of 9
// signature operations to be used in the next three blocks.
const redeemScriptSigOps = 9
redeemScript := pushDataScript(g.privKey.PubKey().SerializeCompressed())
redeemScript = append(redeemScript, bytes.Repeat([]byte{txscript.OP_2DUP,
txscript.OP_CHECKSIGVERIFY}, redeemScriptSigOps-1)...)
redeemScript = append(redeemScript, txscript.OP_CHECKSIG)
assertScriptSigOpsCount(redeemScript, redeemScriptSigOps)
// Create a block that has enough pay-to-script-hash outputs such that
// another block can be created that consumes them all and exceeds the
// max allowed signature operations per block.
//
// ... -> b35(10) -> b39(11)
g.setTip("b35")
b39 := g.nextBlock("b39", outs[11], func(b *wire.MsgBlock) {
// Create a chain of transactions each spending from the
// previous one such that each contains an output that pays to
// the redeem script and the total number of signature
// operations in those redeem scripts will be more than the
// max allowed per block.
p2shScript := payToScriptHashScript(redeemScript)
txnsNeeded := (maxBlockSigOps / redeemScriptSigOps) + 1
prevTx := b.Transactions[1]
for i := 0; i < txnsNeeded; i++ {
prevTx = createSpendTxForTx(prevTx, lowFee)
prevTx.TxOut[0].Value -= 2
prevTx.AddTxOut(wire.NewTxOut(2, p2shScript))
b.AddTransaction(prevTx)
}
})
g.assertTipBlockNumTxns((maxBlockSigOps / redeemScriptSigOps) + 3)
accepted()
// Create a block with more than max allowed signature operations where
// the majority of them are in pay-to-script-hash scripts.
//
// ... -> b35(10) -> b39(11)
// \-> b40(12)
g.setTip("b39")
g.nextBlock("b40", outs[12], func(b *wire.MsgBlock) {
txnsNeeded := (maxBlockSigOps / redeemScriptSigOps)
for i := 0; i < txnsNeeded; i++ {
// Create a signed transaction that spends from the
// associated p2sh output in b39.
spend := makeSpendableOutForTx(b39.Transactions[i+2], 2)
tx := createSpendTx(&spend, lowFee)
sig, err := txscript.RawTxInSignature(tx, 0,
redeemScript, txscript.SigHashAll, g.privKey)
if err != nil {
panic(err)
}
tx.TxIn[0].SignatureScript = pushDataScript(sig,
redeemScript)
b.AddTransaction(tx)
}
// Create a final tx that includes a non-pay-to-script-hash
// output with the number of signature operations needed to push
// the block one over the max allowed.
fill := maxBlockSigOps - (txnsNeeded * redeemScriptSigOps) + 1
finalTx := b.Transactions[len(b.Transactions)-1]
tx := createSpendTxForTx(finalTx, lowFee)
tx.TxOut[0].PkScript = repeatOpcode(txscript.OP_CHECKSIG, fill)
b.AddTransaction(tx)
})
rejected(blockchain.ErrTooManySigOps)
// Create a block with the max allowed signature operations where the
// majority of them are in pay-to-script-hash scripts.
//
// ... -> b35(10) -> b39(11) -> b41(12)
g.setTip("b39")
g.nextBlock("b41", outs[12], func(b *wire.MsgBlock) {
txnsNeeded := (maxBlockSigOps / redeemScriptSigOps)
for i := 0; i < txnsNeeded; i++ {
spend := makeSpendableOutForTx(b39.Transactions[i+2], 2)
tx := createSpendTx(&spend, lowFee)
sig, err := txscript.RawTxInSignature(tx, 0,
redeemScript, txscript.SigHashAll, g.privKey)
if err != nil {
panic(err)
}
tx.TxIn[0].SignatureScript = pushDataScript(sig,
redeemScript)
b.AddTransaction(tx)
}
// Create a final tx that includes a non-pay-to-script-hash
// output with the number of signature operations needed to push
// the block to exactly the max allowed.
fill := maxBlockSigOps - (txnsNeeded * redeemScriptSigOps)
if fill == 0 {
return
}
finalTx := b.Transactions[len(b.Transactions)-1]
tx := createSpendTxForTx(finalTx, lowFee)
tx.TxOut[0].PkScript = repeatOpcode(txscript.OP_CHECKSIG, fill)
b.AddTransaction(tx)
})
accepted()
// ---------------------------------------------------------------------
// Reset the chain to a stable base.
//
// ... -> b35(10) -> b39(11) -> b42(12) -> b43(13)
// \-> b41(12)
// ---------------------------------------------------------------------
g.setTip("b39")
g.nextBlock("b42", outs[12])
acceptedToSideChainWithExpectedTip("b41")
g.nextBlock("b43", outs[13])
accepted()
// ---------------------------------------------------------------------
// Various malformed block tests.
// ---------------------------------------------------------------------
// Create block with an otherwise valid transaction in place of where
// the coinbase must be.
//
// ... -> b43(13)
// \-> b44(14)
g.nextBlock("b44", nil, func(b *wire.MsgBlock) {
nonCoinbaseTx := createSpendTx(outs[14], lowFee)
b.Transactions[0] = nonCoinbaseTx
})
rejected(blockchain.ErrFirstTxNotCoinbase)
// Create block with no transactions.
//
// ... -> b43(13)
// \-> b45(_)
g.setTip("b43")
g.nextBlock("b45", nil, func(b *wire.MsgBlock) {
b.Transactions = nil
})
rejected(blockchain.ErrNoTransactions)
// Create block with invalid proof of work.
//
// ... -> b43(13)
// \-> b46(14)
g.setTip("b43")
b46 := g.nextBlock("b46", outs[14])
// This can't be done inside a munge function passed to nextBlock
// because the block is solved after the function returns and this test
// requires an unsolved block.
{
origHash := b46.BlockHash()
for {
// Keep incrementing the nonce until the hash treated as
// a uint256 is higher than the limit.
b46.Header.Nonce++
blockHash := b46.BlockHash()
hashNum := blockchain.HashToBig(&blockHash)
if hashNum.Cmp(g.params.PowLimit) >= 0 {
break
}
}
g.updateBlockState("b46", origHash, "b46", b46)
}
rejected(blockchain.ErrHighHash)
// Create block with a timestamp too far in the future.
//
// ... -> b43(13)
// \-> b47(14)
g.setTip("b43")
g.nextBlock("b47", outs[14], func(b *wire.MsgBlock) {
// 3 hours in the future clamped to 1 second precision.
nowPlus3Hours := time.Now().Add(time.Hour * 3)
b.Header.Timestamp = time.Unix(nowPlus3Hours.Unix(), 0)
})
rejected(blockchain.ErrTimeTooNew)
// Create block with an invalid merkle root.
//
// ... -> b43(13)
// \-> b48(14)
g.setTip("b43")
g.nextBlock("b48", outs[14], func(b *wire.MsgBlock) {
b.Header.MerkleRoot = chainhash.Hash{}
})
rejected(blockchain.ErrBadMerkleRoot)
// Create block with an invalid proof-of-work limit.
//
// ... -> b43(13)
// \-> b49(14)
g.setTip("b43")
g.nextBlock("b49", outs[14], func(b *wire.MsgBlock) {
b.Header.Bits--
})
rejected(blockchain.ErrUnexpectedDifficulty)
// Create block with an invalid negative proof-of-work limit.
//
// ... -> b43(13)
// \-> b49a(14)
g.setTip("b43")
b49a := g.nextBlock("b49a", outs[14])
// This can't be done inside a munge function passed to nextBlock
// because the block is solved after the function returns and this test
// involves an unsolvable block.
{
origHash := b49a.BlockHash()
b49a.Header.Bits = 0x01810000 // -1 in compact form.
g.updateBlockState("b49a", origHash, "b49a", b49a)
}
rejected(blockchain.ErrUnexpectedDifficulty)
// Create block with two coinbase transactions.
//
// ... -> b43(13)
// \-> b50(14)
g.setTip("b43")
coinbaseTx := g.createCoinbaseTx(g.tipHeight + 1)
g.nextBlock("b50", outs[14], additionalTx(coinbaseTx))
rejected(blockchain.ErrMultipleCoinbases)
// Create block with duplicate transactions.
//
// This test relies on the shape of the shape of the merkle tree to test
// the intended condition and thus is asserted below.
//
// ... -> b43(13)
// \-> b51(14)
g.setTip("b43")
g.nextBlock("b51", outs[14], func(b *wire.MsgBlock) {
b.AddTransaction(b.Transactions[1])
})
g.assertTipBlockNumTxns(3)
rejected(blockchain.ErrDuplicateTx)
// Create a block that spends a transaction that does not exist.
//
// ... -> b43(13)
// \-> b52(14)
g.setTip("b43")
g.nextBlock("b52", outs[14], func(b *wire.MsgBlock) {
hash := newHashFromStr("00000000000000000000000000000000" +
"00000000000000000123456789abcdef")
b.Transactions[1].TxIn[0].PreviousOutPoint.Hash = *hash
b.Transactions[1].TxIn[0].PreviousOutPoint.Index = 0
})
rejected(blockchain.ErrMissingTxOut)
blockchain: Add block validation infrastructure. This adds a full-blown testing infrastructure in order to test consensus validation rules. It is built around the idea of dynamically generating full blocks that target specific rules linked together to form a block chain. In order to properly test the rules, each test instance starts with a valid block that is then modified in the specific way needed to test a specific rule. Blocks which exercise following rules have been added for this initial version. These tests were largely ported from the original Java-based 'official' block acceptance tests as well as some additional tests available in the Core python port. It is expected that further tests can be added over time as consensus rules change. * Enough valid blocks to have a stable base of mature coinbases to spend for futher tests * Basic forking and chain reorganization * Double spends on forks * Too much proof-of-work coinbase (extending main chain, in block that forces a reorg, and in a valid fork) * Max and too many signature operations via various combinations of OP_CHECKSIG, OP_MULTISIG, OP_CHECKSIGVERIFY, and OP_MULTISIGVERIFY * Too many and max signature operations with offending sigop after invalid data push * Max and too many signature operations via pay-to-script-hash redeem scripts * Attempt to spend tx created on a different fork * Attempt to spend immature coinbase (on main chain and fork) * Max size block and block that exceeds the max size * Children of rejected blocks are either orphans or rejected * Coinbase script too small and too large * Max length coinbase script * Attempt to spend tx in blocks that failed to connect * Valid non-coinbase tx in place of coinbase * Block with no transactions * Invalid proof-of-work * Block with a timestamp too far in the future * Invalid merkle root * Invalid proof-of-work limit (bits header field) * Negative proof-of-work limit (bits header field) * Two coinbase transactions * Duplicate transactions * Spend from transaction that does not exist * Timestamp exactly at and one second after the median time * Blocks with same hash via merkle root tricks * Spend from transaction index that is out of range * Transaction that spends more that its inputs provide * Transaction with same hash as an existing tx that has not been fully spent (BIP0030) * Non-final coinbase and non-coinbase txns * Max size block with canonical encoding which exceeds max size with non-canonical encoding * Spend from transaction earlier in same block * Spend from transaction later in same block * Double spend transaction from earlier in same block * Coinbase that pays more than subsidy + fees * Coinbase that includes subsidy + fees * Invalid opcode in dead execution path * Reorganization of txns with OP_RETURN outputs * Spend of an OP_RETURN output * Transaction with multiple OP_RETURN outputs * Large max-sized block reorganization test (disabled by default since it takes a long time and a lot of memory to run) Finally, the README.md files in the main and docs directories have been updated to reflect the use of the new testing framework.
2016-09-14 01:11:12 +02:00
// ---------------------------------------------------------------------
// Block header median time tests.
// ---------------------------------------------------------------------
// Reset the chain to a stable base.
//
// ... -> b33(9) -> b35(10) -> b39(11) -> b42(12) -> b43(13) -> b53(14)
g.setTip("b43")
g.nextBlock("b53", outs[14])
accepted()
// Create a block with a timestamp that is exactly the median time. The
// block must be rejected.
//
// ... -> b33(9) -> b35(10) -> b39(11) -> b42(12) -> b43(13) -> b53(14)
// \-> b54(15)
g.nextBlock("b54", outs[15], func(b *wire.MsgBlock) {
medianBlock := g.blocks[b.Header.PrevBlock]
for i := 0; i < medianTimeBlocks/2; i++ {
medianBlock = g.blocks[medianBlock.Header.PrevBlock]
}
b.Header.Timestamp = medianBlock.Header.Timestamp
})
rejected(blockchain.ErrTimeTooOld)
// Create a block with a timestamp that is one second after the median
// time. The block must be accepted.
//
// ... -> b33(9) -> b35(10) -> b39(11) -> b42(12) -> b43(13) -> b53(14) -> b55(15)
g.setTip("b53")
g.nextBlock("b55", outs[15], func(b *wire.MsgBlock) {
medianBlock := g.blocks[b.Header.PrevBlock]
for i := 0; i < medianTimeBlocks/2; i++ {
medianBlock = g.blocks[medianBlock.Header.PrevBlock]
}
medianBlockTime := medianBlock.Header.Timestamp
b.Header.Timestamp = medianBlockTime.Add(time.Second)
})
accepted()
// ---------------------------------------------------------------------
// CVE-2012-2459 (block hash collision due to merkle tree algo) tests.
// ---------------------------------------------------------------------
// Create two blocks that have the same hash via merkle tree tricks to
// ensure that the valid block is accepted even though it has the same
// hash as the invalid block that was rejected first.
//
// This is accomplished by building the blocks as follows:
//
// b57 (valid block):
//
// root = h1234 = h(h12 || h34)
// // \\
// h12 = h(h(cb) || h(tx2)) h34 = h(h(tx3) || h(tx3))
// // \\ // \\
// coinbase tx2 tx3 nil
//
// transactions: coinbase, tx2, tx3
// merkle tree level 1: h12 = h(h(cb) || h(tx2))
// h34 = h(h(tx3) || h(tx3)) // Algo reuses tx3
// merkle tree root: h(h12 || h34)
//
// b56 (invalid block with the same hash):
//
// root = h1234 = h(h12 || h34)
// // \\
// h12 = h(h(cb) || h(tx2)) h34 = h(h(tx3) || h(tx3))
// // \\ // \\
// coinbase tx2 tx3 tx3
//
// transactions: coinbase, tx2, tx3, tx3
// merkle tree level 1: h12 = h(h(cb) || h(tx2))
// h34 = h(h(tx3) || h(tx3)) // real tx3 dup
// merkle tree root: h(h12 || h34)
//
// ... -> b55(15) -> b57(16)
// \-> b56(16)
g.setTip("b55")
b57 := g.nextBlock("b57", outs[16], func(b *wire.MsgBlock) {
tx2 := b.Transactions[1]
tx3 := createSpendTxForTx(tx2, lowFee)
b.AddTransaction(tx3)
})
g.assertTipBlockNumTxns(3)
g.setTip("b55")
b56 := g.nextBlock("b56", nil, func(b *wire.MsgBlock) {
*b = cloneBlock(b57)
b.AddTransaction(b.Transactions[2])
})
g.assertTipBlockNumTxns(4)
g.assertTipBlockHash(b57.BlockHash())
g.assertTipBlockMerkleRoot(b57.Header.MerkleRoot)
rejected(blockchain.ErrDuplicateTx)
// Since the two blocks have the same hash and the generator state now
// has b56 associated with the hash, manually remove b56, replace it
// with b57, and then reset the tip to it.
g.updateBlockState("b56", b56.BlockHash(), "b57", b57)
g.setTip("b57")
accepted()
// Create a block that contains two duplicate txns that are not in a
// consecutive position within the merkle tree.
//
// This is accomplished by building the block as follows:
//
// transactions: coinbase, tx2, tx3, tx4, tx5, tx6, tx3, tx4
// merkle tree level 2: h12 = h(h(cb) || h(tx2))
// h34 = h(h(tx3) || h(tx4))
// h56 = h(h(tx5) || h(tx6))
// h78 = h(h(tx3) || h(tx4)) // Same as h34
// merkle tree level 1: h1234 = h(h12 || h34)
// h5678 = h(h56 || h78)
// merkle tree root: h(h1234 || h5678)
//
//
// ... -> b55(15) -> b57(16)
// \-> b56p2(16)
g.setTip("b55")
g.nextBlock("b56p2", outs[16], func(b *wire.MsgBlock) {
// Create 4 transactions that each spend from the previous tx
// in the block.
spendTx := b.Transactions[1]
for i := 0; i < 4; i++ {
spendTx = createSpendTxForTx(spendTx, lowFee)
b.AddTransaction(spendTx)
}
// Add the duplicate transactions (3rd and 4th).
b.AddTransaction(b.Transactions[2])
b.AddTransaction(b.Transactions[3])
})
g.assertTipBlockNumTxns(8)
rejected(blockchain.ErrDuplicateTx)
// ---------------------------------------------------------------------
// Invalid transaction type tests.
// ---------------------------------------------------------------------
// Create block with a transaction that tries to spend from an index
// that is out of range from an otherwise valid and existing tx.
//
// ... -> b57(16)
// \-> b58(17)
g.setTip("b57")
g.nextBlock("b58", outs[17], func(b *wire.MsgBlock) {
b.Transactions[1].TxIn[0].PreviousOutPoint.Index = 42
})
rejected(blockchain.ErrMissingTxOut)
blockchain: Add block validation infrastructure. This adds a full-blown testing infrastructure in order to test consensus validation rules. It is built around the idea of dynamically generating full blocks that target specific rules linked together to form a block chain. In order to properly test the rules, each test instance starts with a valid block that is then modified in the specific way needed to test a specific rule. Blocks which exercise following rules have been added for this initial version. These tests were largely ported from the original Java-based 'official' block acceptance tests as well as some additional tests available in the Core python port. It is expected that further tests can be added over time as consensus rules change. * Enough valid blocks to have a stable base of mature coinbases to spend for futher tests * Basic forking and chain reorganization * Double spends on forks * Too much proof-of-work coinbase (extending main chain, in block that forces a reorg, and in a valid fork) * Max and too many signature operations via various combinations of OP_CHECKSIG, OP_MULTISIG, OP_CHECKSIGVERIFY, and OP_MULTISIGVERIFY * Too many and max signature operations with offending sigop after invalid data push * Max and too many signature operations via pay-to-script-hash redeem scripts * Attempt to spend tx created on a different fork * Attempt to spend immature coinbase (on main chain and fork) * Max size block and block that exceeds the max size * Children of rejected blocks are either orphans or rejected * Coinbase script too small and too large * Max length coinbase script * Attempt to spend tx in blocks that failed to connect * Valid non-coinbase tx in place of coinbase * Block with no transactions * Invalid proof-of-work * Block with a timestamp too far in the future * Invalid merkle root * Invalid proof-of-work limit (bits header field) * Negative proof-of-work limit (bits header field) * Two coinbase transactions * Duplicate transactions * Spend from transaction that does not exist * Timestamp exactly at and one second after the median time * Blocks with same hash via merkle root tricks * Spend from transaction index that is out of range * Transaction that spends more that its inputs provide * Transaction with same hash as an existing tx that has not been fully spent (BIP0030) * Non-final coinbase and non-coinbase txns * Max size block with canonical encoding which exceeds max size with non-canonical encoding * Spend from transaction earlier in same block * Spend from transaction later in same block * Double spend transaction from earlier in same block * Coinbase that pays more than subsidy + fees * Coinbase that includes subsidy + fees * Invalid opcode in dead execution path * Reorganization of txns with OP_RETURN outputs * Spend of an OP_RETURN output * Transaction with multiple OP_RETURN outputs * Large max-sized block reorganization test (disabled by default since it takes a long time and a lot of memory to run) Finally, the README.md files in the main and docs directories have been updated to reflect the use of the new testing framework.
2016-09-14 01:11:12 +02:00
// Create block with transaction that pays more than its inputs.
//
// ... -> b57(16)
// \-> b59(17)
g.setTip("b57")
g.nextBlock("b59", outs[17], func(b *wire.MsgBlock) {
b.Transactions[1].TxOut[0].Value = int64(outs[17].amount) + 1
})
rejected(blockchain.ErrSpendTooHigh)
// ---------------------------------------------------------------------
// BIP0030 tests.
// ---------------------------------------------------------------------
// Create a good block to reset the chain to a stable base.
//
// ... -> b57(16) -> b60(17)
g.setTip("b57")
g.nextBlock("b60", outs[17])
accepted()
// Create block that has a tx with the same hash as an existing tx that
// has not been fully spent.
//
// ... -> b60(17)
// \-> b61(18)
g.nextBlock("b61", outs[18], func(b *wire.MsgBlock) {
// Duplicate the coinbase of the parent block to force the
// condition.
parent := g.blocks[b.Header.PrevBlock]
b.Transactions[0] = parent.Transactions[0]
})
rejected(blockchain.ErrOverwriteTx)
// ---------------------------------------------------------------------
// Blocks with non-final transaction tests.
// ---------------------------------------------------------------------
// Create block that contains a non-final non-coinbase transaction.
//
// ... -> b60(17)
// \-> b62(18)
g.setTip("b60")
g.nextBlock("b62", outs[18], func(b *wire.MsgBlock) {
// A non-final transaction must have at least one input with a
// non-final sequence number in addition to a non-final lock
// time.
b.Transactions[1].LockTime = 0xffffffff
b.Transactions[1].TxIn[0].Sequence = 0
})
rejected(blockchain.ErrUnfinalizedTx)
// Create block that contains a non-final coinbase transaction.
//
// ... -> b60(17)
// \-> b63(18)
g.setTip("b60")
g.nextBlock("b63", outs[18], func(b *wire.MsgBlock) {
// A non-final transaction must have at least one input with a
// non-final sequence number in addition to a non-final lock
// time.
b.Transactions[0].LockTime = 0xffffffff
b.Transactions[0].TxIn[0].Sequence = 0
})
rejected(blockchain.ErrUnfinalizedTx)
// ---------------------------------------------------------------------
// Non-canonical variable-length integer tests.
// ---------------------------------------------------------------------
// Create a max size block with the variable-length integer for the
// number of transactions replaced with a larger non-canonical version
// that causes the block size to exceed the max allowed size. Then,
// create another block that is identical except with the canonical
// encoding and ensure it is accepted. The intent is to verify the
// implementation does not reject the second block, which will have the
// same hash, due to the first one already being rejected.
//
// ... -> b60(17) -> b64(18)
// \-> b64a(18)
g.setTip("b60")
b64a := g.nextBlock("b64a", outs[18], func(b *wire.MsgBlock) {
bytesToMaxSize := maxBlockSize - b.SerializeSize() - 3
sizePadScript := repeatOpcode(0x00, bytesToMaxSize)
replaceSpendScript(sizePadScript)(b)
})
g.assertTipNonCanonicalBlockSize(maxBlockSize + 8)
rejectedNonCanonical()
g.setTip("b60")
b64 := g.nextBlock("b64", outs[18], func(b *wire.MsgBlock) {
*b = cloneBlock(b64a)
})
// Since the two blocks have the same hash and the generator state now
// has b64a associated with the hash, manually remove b64a, replace it
// with b64, and then reset the tip to it.
g.updateBlockState("b64a", b64a.BlockHash(), "b64", b64)
g.setTip("b64")
g.assertTipBlockHash(b64a.BlockHash())
g.assertTipBlockSize(maxBlockSize)
accepted()
// ---------------------------------------------------------------------
// Same block transaction spend tests.
// ---------------------------------------------------------------------
// Create block that spends an output created earlier in the same block.
//
// ... b64(18) -> b65(19)
g.setTip("b64")
g.nextBlock("b65", outs[19], func(b *wire.MsgBlock) {
tx3 := createSpendTxForTx(b.Transactions[1], lowFee)
b.AddTransaction(tx3)
})
accepted()
// Create block that spends an output created later in the same block.
//
// ... -> b65(19)
// \-> b66(20)
g.nextBlock("b66", nil, func(b *wire.MsgBlock) {
tx2 := createSpendTx(outs[20], lowFee)
tx3 := createSpendTxForTx(tx2, lowFee)
b.AddTransaction(tx3)
b.AddTransaction(tx2)
})
rejected(blockchain.ErrMissingTxOut)
blockchain: Add block validation infrastructure. This adds a full-blown testing infrastructure in order to test consensus validation rules. It is built around the idea of dynamically generating full blocks that target specific rules linked together to form a block chain. In order to properly test the rules, each test instance starts with a valid block that is then modified in the specific way needed to test a specific rule. Blocks which exercise following rules have been added for this initial version. These tests were largely ported from the original Java-based 'official' block acceptance tests as well as some additional tests available in the Core python port. It is expected that further tests can be added over time as consensus rules change. * Enough valid blocks to have a stable base of mature coinbases to spend for futher tests * Basic forking and chain reorganization * Double spends on forks * Too much proof-of-work coinbase (extending main chain, in block that forces a reorg, and in a valid fork) * Max and too many signature operations via various combinations of OP_CHECKSIG, OP_MULTISIG, OP_CHECKSIGVERIFY, and OP_MULTISIGVERIFY * Too many and max signature operations with offending sigop after invalid data push * Max and too many signature operations via pay-to-script-hash redeem scripts * Attempt to spend tx created on a different fork * Attempt to spend immature coinbase (on main chain and fork) * Max size block and block that exceeds the max size * Children of rejected blocks are either orphans or rejected * Coinbase script too small and too large * Max length coinbase script * Attempt to spend tx in blocks that failed to connect * Valid non-coinbase tx in place of coinbase * Block with no transactions * Invalid proof-of-work * Block with a timestamp too far in the future * Invalid merkle root * Invalid proof-of-work limit (bits header field) * Negative proof-of-work limit (bits header field) * Two coinbase transactions * Duplicate transactions * Spend from transaction that does not exist * Timestamp exactly at and one second after the median time * Blocks with same hash via merkle root tricks * Spend from transaction index that is out of range * Transaction that spends more that its inputs provide * Transaction with same hash as an existing tx that has not been fully spent (BIP0030) * Non-final coinbase and non-coinbase txns * Max size block with canonical encoding which exceeds max size with non-canonical encoding * Spend from transaction earlier in same block * Spend from transaction later in same block * Double spend transaction from earlier in same block * Coinbase that pays more than subsidy + fees * Coinbase that includes subsidy + fees * Invalid opcode in dead execution path * Reorganization of txns with OP_RETURN outputs * Spend of an OP_RETURN output * Transaction with multiple OP_RETURN outputs * Large max-sized block reorganization test (disabled by default since it takes a long time and a lot of memory to run) Finally, the README.md files in the main and docs directories have been updated to reflect the use of the new testing framework.
2016-09-14 01:11:12 +02:00
// Create block that double spends a transaction created in the same
// block.
//
// ... -> b65(19)
// \-> b67(20)
g.setTip("b65")
g.nextBlock("b67", outs[20], func(b *wire.MsgBlock) {
tx2 := b.Transactions[1]
tx3 := createSpendTxForTx(tx2, lowFee)
tx4 := createSpendTxForTx(tx2, lowFee)
b.AddTransaction(tx3)
b.AddTransaction(tx4)
})
rejected(blockchain.ErrMissingTxOut)
blockchain: Add block validation infrastructure. This adds a full-blown testing infrastructure in order to test consensus validation rules. It is built around the idea of dynamically generating full blocks that target specific rules linked together to form a block chain. In order to properly test the rules, each test instance starts with a valid block that is then modified in the specific way needed to test a specific rule. Blocks which exercise following rules have been added for this initial version. These tests were largely ported from the original Java-based 'official' block acceptance tests as well as some additional tests available in the Core python port. It is expected that further tests can be added over time as consensus rules change. * Enough valid blocks to have a stable base of mature coinbases to spend for futher tests * Basic forking and chain reorganization * Double spends on forks * Too much proof-of-work coinbase (extending main chain, in block that forces a reorg, and in a valid fork) * Max and too many signature operations via various combinations of OP_CHECKSIG, OP_MULTISIG, OP_CHECKSIGVERIFY, and OP_MULTISIGVERIFY * Too many and max signature operations with offending sigop after invalid data push * Max and too many signature operations via pay-to-script-hash redeem scripts * Attempt to spend tx created on a different fork * Attempt to spend immature coinbase (on main chain and fork) * Max size block and block that exceeds the max size * Children of rejected blocks are either orphans or rejected * Coinbase script too small and too large * Max length coinbase script * Attempt to spend tx in blocks that failed to connect * Valid non-coinbase tx in place of coinbase * Block with no transactions * Invalid proof-of-work * Block with a timestamp too far in the future * Invalid merkle root * Invalid proof-of-work limit (bits header field) * Negative proof-of-work limit (bits header field) * Two coinbase transactions * Duplicate transactions * Spend from transaction that does not exist * Timestamp exactly at and one second after the median time * Blocks with same hash via merkle root tricks * Spend from transaction index that is out of range * Transaction that spends more that its inputs provide * Transaction with same hash as an existing tx that has not been fully spent (BIP0030) * Non-final coinbase and non-coinbase txns * Max size block with canonical encoding which exceeds max size with non-canonical encoding * Spend from transaction earlier in same block * Spend from transaction later in same block * Double spend transaction from earlier in same block * Coinbase that pays more than subsidy + fees * Coinbase that includes subsidy + fees * Invalid opcode in dead execution path * Reorganization of txns with OP_RETURN outputs * Spend of an OP_RETURN output * Transaction with multiple OP_RETURN outputs * Large max-sized block reorganization test (disabled by default since it takes a long time and a lot of memory to run) Finally, the README.md files in the main and docs directories have been updated to reflect the use of the new testing framework.
2016-09-14 01:11:12 +02:00
// ---------------------------------------------------------------------
// Extra subsidy tests.
// ---------------------------------------------------------------------
// Create block that pays 10 extra to the coinbase and a tx that only
// pays 9 fee.
//
// ... -> b65(19)
// \-> b68(20)
g.setTip("b65")
g.nextBlock("b68", outs[20], additionalCoinbase(10), additionalSpendFee(9))
rejected(blockchain.ErrBadCoinbaseValue)
// Create block that pays 10 extra to the coinbase and a tx that pays
// the extra 10 fee.
//
// ... -> b65(19) -> b69(20)
g.setTip("b65")
g.nextBlock("b69", outs[20], additionalCoinbase(10), additionalSpendFee(10))
accepted()
// ---------------------------------------------------------------------
// More signature operations counting tests.
//
// The next several tests ensure signature operations are counted before
// script elements that cause parse failure while those after are
// ignored and that signature operations after script elements that
// successfully parse even if that element will fail at run-time are
// counted.
// ---------------------------------------------------------------------
// Create block with more than max allowed signature operations such
// that the signature operation that pushes it over the limit is after
// a push data with a script element size that is larger than the max
// allowed size when executed. The block must be rejected because the
// signature operation after the script element must be counted since
// the script parses validly.
//
// The script generated consists of the following form:
//
// Comment assumptions:
// maxBlockSigOps = 20000
// maxScriptElementSize = 520
//
// [0-19999] : OP_CHECKSIG
// [20000] : OP_PUSHDATA4
// [20001-20004]: 521 (little-endian encoded maxScriptElementSize+1)
// [20005-20525]: too large script element
// [20526] : OP_CHECKSIG (goes over the limit)
//
// ... -> b69(20)
// \-> b70(21)
scriptSize := maxBlockSigOps + 5 + (maxScriptElementSize + 1) + 1
tooManySigOps = repeatOpcode(txscript.OP_CHECKSIG, scriptSize)
tooManySigOps[maxBlockSigOps] = txscript.OP_PUSHDATA4
binary.LittleEndian.PutUint32(tooManySigOps[maxBlockSigOps+1:],
maxScriptElementSize+1)
g.nextBlock("b70", outs[21], replaceSpendScript(tooManySigOps))
g.assertTipBlockSigOpsCount(maxBlockSigOps + 1)
rejected(blockchain.ErrTooManySigOps)
// Create block with more than max allowed signature operations such
// that the signature operation that pushes it over the limit is before
// an invalid push data that claims a large amount of data even though
// that much data is not provided.
//
// ... -> b69(20)
// \-> b71(21)
g.setTip("b69")
scriptSize = maxBlockSigOps + 5 + maxScriptElementSize + 1
tooManySigOps = repeatOpcode(txscript.OP_CHECKSIG, scriptSize)
tooManySigOps[maxBlockSigOps+1] = txscript.OP_PUSHDATA4
binary.LittleEndian.PutUint32(tooManySigOps[maxBlockSigOps+2:], 0xffffffff)
g.nextBlock("b71", outs[21], replaceSpendScript(tooManySigOps))
g.assertTipBlockSigOpsCount(maxBlockSigOps + 1)
rejected(blockchain.ErrTooManySigOps)
// Create block with the max allowed signature operations such that all
// counted signature operations are before an invalid push data that
// claims a large amount of data even though that much data is not
// provided. The pushed data itself consists of OP_CHECKSIG so the
// block would be rejected if any of them were counted.
//
// ... -> b69(20) -> b72(21)
g.setTip("b69")
scriptSize = maxBlockSigOps + 5 + maxScriptElementSize
manySigOps = repeatOpcode(txscript.OP_CHECKSIG, scriptSize)
manySigOps[maxBlockSigOps] = txscript.OP_PUSHDATA4
binary.LittleEndian.PutUint32(manySigOps[maxBlockSigOps+1:], 0xffffffff)
g.nextBlock("b72", outs[21], replaceSpendScript(manySigOps))
g.assertTipBlockSigOpsCount(maxBlockSigOps)
accepted()
// Create block with the max allowed signature operations such that all
// counted signature operations are before an invalid push data that
// contains OP_CHECKSIG in the number of bytes to push. The block would
// be rejected if any of them were counted.
//
// ... -> b72(21) -> b73(22)
scriptSize = maxBlockSigOps + 5 + (maxScriptElementSize + 1)
manySigOps = repeatOpcode(txscript.OP_CHECKSIG, scriptSize)
manySigOps[maxBlockSigOps] = txscript.OP_PUSHDATA4
g.nextBlock("b73", outs[22], replaceSpendScript(manySigOps))
g.assertTipBlockSigOpsCount(maxBlockSigOps)
accepted()
// ---------------------------------------------------------------------
// Dead execution path tests.
// ---------------------------------------------------------------------
// Create block with an invalid opcode in a dead execution path.
//
// ... -> b73(22) -> b74(23)
script := []byte{txscript.OP_IF, txscript.OP_INVALIDOPCODE,
txscript.OP_ELSE, txscript.OP_TRUE, txscript.OP_ENDIF}
g.nextBlock("b74", outs[23], replaceSpendScript(script), func(b *wire.MsgBlock) {
tx2 := b.Transactions[1]
tx3 := createSpendTxForTx(tx2, lowFee)
tx3.TxIn[0].SignatureScript = []byte{txscript.OP_FALSE}
b.AddTransaction(tx3)
})
accepted()
// ---------------------------------------------------------------------
// Various OP_RETURN tests.
// ---------------------------------------------------------------------
// Create a block that has multiple transactions each with a single
// OP_RETURN output.
//
// ... -> b74(23) -> b75(24)
g.nextBlock("b75", outs[24], func(b *wire.MsgBlock) {
// Add 4 outputs to the spending transaction that are spent
// below.
const numAdditionalOutputs = 4
const zeroCoin = int64(0)
spendTx := b.Transactions[1]
for i := 0; i < numAdditionalOutputs; i++ {
spendTx.AddTxOut(wire.NewTxOut(zeroCoin, opTrueScript))
}
// Add transactions spending from the outputs added above that
// each contain an OP_RETURN output.
//
// NOTE: The createSpendTx func adds the OP_RETURN output.
zeroFee := btcutil.Amount(0)
for i := uint32(0); i < numAdditionalOutputs; i++ {
spend := makeSpendableOut(b, 1, i+2)
tx := createSpendTx(&spend, zeroFee)
b.AddTransaction(tx)
}
})
g.assertTipBlockNumTxns(6)
g.assertTipBlockTxOutOpReturn(5, 1)
b75OpReturnOut := makeSpendableOut(g.tip, 5, 1)
accepted()
// Reorg to a side chain that does not contain the OP_RETURNs.
//
// ... -> b74(23) -> b75(24)
// \-> b76(24) -> b77(25)
g.setTip("b74")
g.nextBlock("b76", outs[24])
acceptedToSideChainWithExpectedTip("b75")
g.nextBlock("b77", outs[25])
accepted()
// Reorg back to the original chain that contains the OP_RETURNs.
//
// ... -> b74(23) -> b75(24) -> b78(25) -> b79(26)
// \-> b76(24) -> b77(25)
g.setTip("b75")
g.nextBlock("b78", outs[25])
acceptedToSideChainWithExpectedTip("b77")
g.nextBlock("b79", outs[26])
accepted()
// Create a block that spends an OP_RETURN.
//
// ... -> b74(23) -> b75(24) -> b78(25) -> b79(26)
// \-> b76(24) -> b77(25) \-> b80(b75.tx[5].out[1])
//
// An OP_RETURN output doesn't have any value and the default behavior
// of nextBlock is to assign a fee of one, so increment the amount here
// to effective negate that behavior.
b75OpReturnOut.amount++
g.nextBlock("b80", &b75OpReturnOut)
rejected(blockchain.ErrMissingTxOut)
blockchain: Add block validation infrastructure. This adds a full-blown testing infrastructure in order to test consensus validation rules. It is built around the idea of dynamically generating full blocks that target specific rules linked together to form a block chain. In order to properly test the rules, each test instance starts with a valid block that is then modified in the specific way needed to test a specific rule. Blocks which exercise following rules have been added for this initial version. These tests were largely ported from the original Java-based 'official' block acceptance tests as well as some additional tests available in the Core python port. It is expected that further tests can be added over time as consensus rules change. * Enough valid blocks to have a stable base of mature coinbases to spend for futher tests * Basic forking and chain reorganization * Double spends on forks * Too much proof-of-work coinbase (extending main chain, in block that forces a reorg, and in a valid fork) * Max and too many signature operations via various combinations of OP_CHECKSIG, OP_MULTISIG, OP_CHECKSIGVERIFY, and OP_MULTISIGVERIFY * Too many and max signature operations with offending sigop after invalid data push * Max and too many signature operations via pay-to-script-hash redeem scripts * Attempt to spend tx created on a different fork * Attempt to spend immature coinbase (on main chain and fork) * Max size block and block that exceeds the max size * Children of rejected blocks are either orphans or rejected * Coinbase script too small and too large * Max length coinbase script * Attempt to spend tx in blocks that failed to connect * Valid non-coinbase tx in place of coinbase * Block with no transactions * Invalid proof-of-work * Block with a timestamp too far in the future * Invalid merkle root * Invalid proof-of-work limit (bits header field) * Negative proof-of-work limit (bits header field) * Two coinbase transactions * Duplicate transactions * Spend from transaction that does not exist * Timestamp exactly at and one second after the median time * Blocks with same hash via merkle root tricks * Spend from transaction index that is out of range * Transaction that spends more that its inputs provide * Transaction with same hash as an existing tx that has not been fully spent (BIP0030) * Non-final coinbase and non-coinbase txns * Max size block with canonical encoding which exceeds max size with non-canonical encoding * Spend from transaction earlier in same block * Spend from transaction later in same block * Double spend transaction from earlier in same block * Coinbase that pays more than subsidy + fees * Coinbase that includes subsidy + fees * Invalid opcode in dead execution path * Reorganization of txns with OP_RETURN outputs * Spend of an OP_RETURN output * Transaction with multiple OP_RETURN outputs * Large max-sized block reorganization test (disabled by default since it takes a long time and a lot of memory to run) Finally, the README.md files in the main and docs directories have been updated to reflect the use of the new testing framework.
2016-09-14 01:11:12 +02:00
// Create a block that has a transaction with multiple OP_RETURNs. Even
// though it's not considered a standard transaction, it is still valid
// by the consensus rules.
//
// ... -> b79(26) -> b81(27)
//
g.setTip("b79")
g.nextBlock("b81", outs[27], func(b *wire.MsgBlock) {
const numAdditionalOutputs = 4
const zeroCoin = int64(0)
spendTx := b.Transactions[1]
for i := 0; i < numAdditionalOutputs; i++ {
opRetScript := uniqueOpReturnScript()
spendTx.AddTxOut(wire.NewTxOut(zeroCoin, opRetScript))
}
})
for i := uint32(2); i < 6; i++ {
g.assertTipBlockTxOutOpReturn(1, i)
}
accepted()
// ---------------------------------------------------------------------
// Large block re-org test.
// ---------------------------------------------------------------------
if !includeLargeReorg {
return tests, nil
}
// Ensure the tip the re-org test builds on is the best chain tip.
//
// ... -> b81(27) -> ...
g.setTip("b81")
// Collect all of the spendable coinbase outputs from the previous
// collection point up to the current tip.
g.saveSpendableCoinbaseOuts()
spendableOutOffset := g.tipHeight - int32(coinbaseMaturity)
// Extend the main chain by a large number of max size blocks.
//
// ... -> br0 -> br1 -> ... -> br#
testInstances = nil
reorgSpend := *outs[spendableOutOffset]
reorgStartBlockName := g.tipName
chain1TipName := g.tipName
for i := int32(0); i < numLargeReorgBlocks; i++ {
chain1TipName = fmt.Sprintf("br%d", i)
g.nextBlock(chain1TipName, &reorgSpend, func(b *wire.MsgBlock) {
bytesToMaxSize := maxBlockSize - b.SerializeSize() - 3
sizePadScript := repeatOpcode(0x00, bytesToMaxSize)
replaceSpendScript(sizePadScript)(b)
})
g.assertTipBlockSize(maxBlockSize)
g.saveTipCoinbaseOut()
testInstances = append(testInstances, acceptBlock(g.tipName,
g.tip, true, false))
// Use the next available spendable output. First use up any
// remaining spendable outputs that were already popped into the
// outs slice, then just pop them from the stack.
if spendableOutOffset+1+i < int32(len(outs)) {
reorgSpend = *outs[spendableOutOffset+1+i]
} else {
reorgSpend = g.oldestCoinbaseOut()
}
}
tests = append(tests, testInstances)
// Create a side chain that has the same length.
//
// ... -> br0 -> ... -> br#
// \-> bralt0 -> ... -> bralt#
g.setTip(reorgStartBlockName)
testInstances = nil
chain2TipName := g.tipName
for i := uint16(0); i < numLargeReorgBlocks; i++ {
chain2TipName = fmt.Sprintf("bralt%d", i)
g.nextBlock(chain2TipName, nil)
testInstances = append(testInstances, acceptBlock(g.tipName,
g.tip, false, false))
}
testInstances = append(testInstances, expectTipBlock(chain1TipName,
g.blocksByName[chain1TipName]))
tests = append(tests, testInstances)
// Extend the side chain by one to force the large reorg.
//
// ... -> bralt0 -> ... -> bralt# -> bralt#+1
// \-> br0 -> ... -> br#
g.nextBlock(fmt.Sprintf("bralt%d", g.tipHeight+1), nil)
chain2TipName = g.tipName
accepted()
// Extend the first chain by two to force a large reorg back to it.
//
// ... -> br0 -> ... -> br# -> br#+1 -> br#+2
// \-> bralt0 -> ... -> bralt# -> bralt#+1
g.setTip(chain1TipName)
g.nextBlock(fmt.Sprintf("br%d", g.tipHeight+1), nil)
chain1TipName = g.tipName
acceptedToSideChainWithExpectedTip(chain2TipName)
g.nextBlock(fmt.Sprintf("br%d", g.tipHeight+2), nil)
chain1TipName = g.tipName
accepted()
return tests, nil
}