Merge pull request #427 from crytic/update

fix coverage report bugs (source unit ID issues) and method selection process

Author: anishnaik

.github/workflows/ci.yml

@@ -87,7 +87,6 @@ jobs:
           inputs: ./medusa-*.tar.gz
 
       - name: Upload artifact
-        if: github.ref == 'refs/heads/master' || (github.event_name == 'push' && startsWith(github.ref, 'refs/tags/'))
         uses: actions/upload-artifact@v4
         with:
           name: medusa-${{ runner.os }}-${{ runner.arch }}

DEV.md

@@ -0,0 +1,104 @@
+# Debugging and Development
+
+## Debugging
+
+The following scripts are available for Medusa developers for debugging changes to the fuzzer.
+
+### Corpus diff
+
+The corpus diff script is used to compare two corpora and identify the methods that are present in one but not the other. This is useful for identifying methods that are missing from a corpus that should be present.
+
+```shell
+python3 scripts/corpus_diff.py corpus1 corpus2
+```
+
+```shell
+Methods only in ~/corpus1:
+-  clampSplitWeight(uint32,uint32)
+
+Methods only in ~/corpus2:
+  <None>
+```
+
+### Corpus stats
+
+The corpus stats script is used to generate statistics about a corpus. This includes the number of sequences, the average length of sequences, and the frequency of methods called.
+
+```shell
+python3 scripts/corpus_stats.py corpus
+```
+
+```shell
+Number of Sequences in ~/corpus: 130
+
+Average Length of Transactions List: 43
+
+Frequency of Methods Called:
+-  testReceiversReceivedSplit(uint8): 280
+-  setMaxEndHints(uint32,uint32): 174
+-  setStreamBalanceWithdrawAll(uint8): 139
+-  giveClampedAmount(uint8,uint8,uint128): 136
+-  receiveStreamsSplitAndCollectToSelf(uint8): 133
+-  testSqueezeViewVsActual(uint8,uint8): 128
+-  testSqueeze(uint8,uint8): 128
+-  testSetStreamBalance(uint8,int128): 128
+-  addStreamWithClamping(uint8,uint8,uint160,uint32,uint32,int128): 125
+-  removeAllSplits(uint8): 118
+-  testSplittableAfterSplit(uint8): 113
+-  testSqueezableVsReceived(uint8): 111
+-  testBalanceAtInFuture(uint8,uint8,uint160): 108
+-  testRemoveStreamShouldNotRevert(uint8,uint256): 103
+-  invariantWithdrawAllTokensShouldNotRevert(): 103
+-  collect(uint8,uint8): 101
+-  invariantAmtPerSecVsMinAmtPerSec(uint8,uint256): 98
+-  testSqueezableAmountCantBeWithdrawn(uint8,uint8): 97
+-  split(uint8): 97
+-  invariantWithdrawAllTokens(): 95
+-  testReceiveStreams(uint8,uint32): 93
+-  invariantAccountingVsTokenBalance(): 92
+-  testSqueezeWithFuzzedHistoryShouldNotRevert(uint8,uint8,uint256,bytes32): 91
+-  testSqueezableAmountCantBeUndone(uint8,uint8,uint160,uint32,uint32,int128): 87
+-  testCollect(uint8,uint8): 86
+-  testSetStreamBalanceWithdrawAllShouldNotRevert(uint8): 86
+-  testAddStreamShouldNotRevert(uint8,uint8,uint160,uint32,uint32,int128): 85
+-  testReceiveStreamsShouldNotRevert(uint8): 84
+-  addSplitsReceiver(uint8,uint8,uint32): 84
+-  setStreamBalanceWithClamping(uint8,int128): 82
+-  addSplitsReceiverWithClamping(uint8,uint8,uint32): 80
+-  testSetStreamBalanceShouldNotRevert(uint8,int128): 80
+-  testSplitShouldNotRevert(uint8): 80
+-  squeezeAllAndReceiveAndSplitAndCollectToSelf(uint8): 79
+-  addStreamImmediatelySqueezable(uint8,uint8,uint160): 79
+-  testSetSplitsShouldNotRevert(uint8,uint8,uint32): 78
+-  invariantSumAmtDeltaIsZero(uint8): 78
+-  testReceiveStreamsViewConsistency(uint8,uint32): 76
+-  squeezeToSelf(uint8): 74
+-  collectToSelf(uint8): 72
+-  setStreams(uint8,uint8,uint160,uint32,uint32,int128): 70
+-  receiveStreamsAllCycles(uint8): 69
+-  invariantWithdrawShouldAlwaysFail(uint256): 68
+-  addStream(uint8,uint8,uint160,uint32,uint32,int128): 68
+-  squeezeWithFuzzedHistory(uint8,uint8,uint256,bytes32): 67
+-  setStreamsWithClamping(uint8,uint8,uint160,uint32,uint32,int128): 67
+-  splitAndCollectToSelf(uint8): 67
+-  testSqueezeWithFullyHashedHistory(uint8,uint8): 65
+-  give(uint8,uint8,uint128): 65
+-  setSplits(uint8,uint8,uint32): 65
+-  testSqueezeTwice(uint8,uint8,uint256,bytes32): 65
+-  testSetStreamsShouldNotRevert(uint8,uint8,uint160,uint32,uint32,int128): 64
+-  squeezeAllSenders(uint8): 63
+-  removeStream(uint8,uint256): 62
+-  testCollectableAfterSplit(uint8): 58
+-  testCollectShouldNotRevert(uint8,uint8): 56
+-  testReceiveStreamsViewVsActual(uint8,uint32): 55
+-  receiveStreams(uint8,uint32): 55
+-  setSplitsWithClamping(uint8,uint8,uint32): 55
+-  testGiveShouldNotRevert(uint8,uint8,uint128): 47
+-  setStreamBalance(uint8,int128): 47
+-  squeezeWithDefaultHistory(uint8,uint8): 45
+-  testSplitViewVsActual(uint8): 45
+-  testAddSplitsShouldNotRevert(uint8,uint8,uint32): 30
+-  testSqueezeWithDefaultHistoryShouldNotRevert(uint8,uint8): 23
+
+Number of Unique Methods: 65
+```

chain/test_chain_test.go

@@ -215,15 +215,15 @@ func TestChainDynamicDeployments(t *testing.T) {
 		compilations, _, err := cryticCompile.Compile()
 		assert.NoError(t, err)
 		assert.EqualValues(t, 1, len(compilations))
-		assert.EqualValues(t, 1, len(compilations[0].Sources))
+		assert.EqualValues(t, 1, len(compilations[0].SourcePathToArtifact))
 
 		// Obtain our chain and senders
 		chain, senders := createChain(t)
 
 		// Deploy each contract that has no construct arguments.
 		deployCount := 0
 		for _, compilation := range compilations {
-			for _, source := range compilation.Sources {
+			for _, source := range compilation.SourcePathToArtifact {
 				for _, contract := range source.Contracts {
 					contract := contract
 					if len(contract.Abi.Constructor.Inputs) == 0 {
@@ -329,7 +329,7 @@ func TestChainDeploymentWithArgs(t *testing.T) {
 		compilations, _, err := cryticCompile.Compile()
 		assert.NoError(t, err)
 		assert.EqualValues(t, 1, len(compilations))
-		assert.EqualValues(t, 1, len(compilations[0].Sources))
+		assert.EqualValues(t, 1, len(compilations[0].SourcePathToArtifact))
 
 		// Obtain our chain and senders
 		chain, senders := createChain(t)
@@ -346,7 +346,7 @@ func TestChainDeploymentWithArgs(t *testing.T) {
 		// Deploy each contract
 		deployCount := 0
 		for _, compilation := range compilations {
-			for _, source := range compilation.Sources {
+			for _, source := range compilation.SourcePathToArtifact {
 				for contractName, contract := range source.Contracts {
 					contract := contract
 
@@ -467,7 +467,7 @@ func TestChainCloning(t *testing.T) {
 
 		// Deploy each contract that has no construct arguments 10 times.
 		for _, compilation := range compilations {
-			for _, source := range compilation.Sources {
+			for _, source := range compilation.SourcePathToArtifact {
 				for _, contract := range source.Contracts {
 					contract := contract
 					if len(contract.Abi.Constructor.Inputs) == 0 {
@@ -563,7 +563,7 @@ func TestChainCallSequenceReplayMatchSimple(t *testing.T) {
 
 		// Deploy each contract that has no construct arguments 10 times.
 		for _, compilation := range compilations {
-			for _, source := range compilation.Sources {
+			for _, source := range compilation.SourcePathToArtifact {
 				for _, contract := range source.Contracts {
 					contract := contract
 					if len(contract.Abi.Constructor.Inputs) == 0 {

compilation/platforms/crytic_compile.go

@@ -143,7 +143,7 @@ func (c *CryticCompilationConfig) Compile() ([]types.Compilation, string, error)
 	var compilationList []types.Compilation
 
 	// Define the structure of our crytic-compile export data.
-	type solcExportSource struct {
+	type solcSourceUnit struct {
 		AST any `json:"AST"`
 	}
 	type solcExportContract struct {
@@ -154,9 +154,8 @@ func (c *CryticCompilationConfig) Compile() ([]types.Compilation, string, error)
 		BinRuntime    string `json:"bin-runtime"`
 	}
 	type solcExportData struct {
-		Sources    map[string]solcExportSource   `json:"sources"`
-		Contracts  map[string]solcExportContract `json:"contracts"`
-		SourceList []string                      `json:"sourceList"`
+		Sources   map[string]solcSourceUnit     `json:"sources"`
+		Contracts map[string]solcExportContract `json:"contracts"`
 	}
 
 	// Loop through each .json file for compilation units.
@@ -176,14 +175,41 @@ func (c *CryticCompilationConfig) Compile() ([]types.Compilation, string, error)
 
 		// Create a compilation object that will store the contracts and source information.
 		compilation := types.NewCompilation()
-		compilation.SourceList = solcExport.SourceList
+
+		// Create a map of contract names to their kinds
+		contractKinds := make(map[string]types.ContractKind)
 
 		// Loop through all sources and parse them into our types.
 		for sourcePath, source := range solcExport.Sources {
-			compilation.Sources[sourcePath] = types.CompiledSource{
-				Ast:       source.AST,
-				Contracts: make(map[string]types.CompiledContract),
+			// Convert the AST into our version of the AST (types.AST)
+			var ast types.AST
+			b, err = json.Marshal(source.AST)
+			if err != nil {
+				return nil, "", fmt.Errorf("could not encode AST from sources: %v", err)
+			}
+			err = json.Unmarshal(b, &ast)
+			if err != nil {
+				return nil, "", fmt.Errorf("could not parse AST from sources: %v", err)
+			}
+
+			// From the AST, extract the contract kinds where the contract definition could be for a contract, library,
+			// or interface
+			for _, node := range ast.Nodes {
+				if node.GetNodeType() == "ContractDefinition" {
+					contractDefinition := node.(types.ContractDefinition)
+					contractKinds[contractDefinition.CanonicalName] = contractDefinition.Kind
+				}
+			}
+
+			// Retrieve the source unit ID
+			sourceUnitId := ast.GetSourceUnitID()
+			compilation.SourcePathToArtifact[sourcePath] = types.SourceArtifact{
+				// TODO: Our types.AST is not the same as the original AST but we could parse it and avoid using "any"
+				Ast:          source.AST,
+				Contracts:    make(map[string]types.CompiledContract),
+				SourceUnitId: sourceUnitId,
 			}
+			compilation.SourceIdToPath[sourceUnitId] = sourcePath
 		}
 
 		// Loop through all contracts and parse them into our types.
@@ -198,12 +224,12 @@ func (c *CryticCompilationConfig) Compile() ([]types.Compilation, string, error)
 
 			// Ensure a source exists for this, or create one if our path somehow differed from any
 			// path not existing in the "sources" key at the root of the export.
-			if _, ok := compilation.Sources[sourcePath]; !ok {
-				parentSource := types.CompiledSource{
+			if _, ok := compilation.SourcePathToArtifact[sourcePath]; !ok {
+				parentSource := types.SourceArtifact{
 					Ast:       nil,
 					Contracts: make(map[string]types.CompiledContract),
 				}
-				compilation.Sources[sourcePath] = parentSource
+				compilation.SourcePathToArtifact[sourcePath] = parentSource
 			}
 
 			// Parse the ABI
@@ -223,12 +249,13 @@ func (c *CryticCompilationConfig) Compile() ([]types.Compilation, string, error)
 			}
 
 			// Add contract details
-			compilation.Sources[sourcePath].Contracts[contractName] = types.CompiledContract{
+			compilation.SourcePathToArtifact[sourcePath].Contracts[contractName] = types.CompiledContract{
 				Abi:             *contractAbi,
 				InitBytecode:    initBytecode,
 				RuntimeBytecode: runtimeBytecode,
 				SrcMapsInit:     contract.SrcMap,
 				SrcMapsRuntime:  contract.SrcMapRuntime,
+				Kind:            contractKinds[contractName],
 			}
 		}
 

compilation/platforms/crytic_compile_test.go

@@ -1,22 +1,23 @@
 package platforms
 
 import (
-	"github.com/crytic/medusa/compilation/types"
-	"github.com/crytic/medusa/utils"
-	"github.com/crytic/medusa/utils/testutils"
-	"github.com/stretchr/testify/assert"
 	"os"
 	"os/exec"
 	"path/filepath"
 	"strings"
 	"testing"
+
+	"github.com/crytic/medusa/compilation/types"
+	"github.com/crytic/medusa/utils"
+	"github.com/crytic/medusa/utils/testutils"
+	"github.com/stretchr/testify/assert"
 )
 
 // testCryticGetCompiledSourceByBaseName checks if a given source file exists in a given compilation's map of sources.
 // The source file is the file name of a specific file. This function simply checks one of the paths ends with
 // this name. Avoid including any directories in case the path separators differ per system.
 // Returns the types.CompiledSource (mapping value) associated to the path if it is found. Returns nil otherwise.
-func testCryticGetCompiledSourceByBaseName(sources map[string]types.CompiledSource, name string) *types.CompiledSource {
+func testCryticGetCompiledSourceByBaseName(sources map[string]types.SourceArtifact, name string) *types.SourceArtifact {
 	// Obtain a lower case version of our name to search for
 	lowerName := strings.ToLower(name)
 
@@ -53,10 +54,10 @@ func TestCryticSingleFileAbsolutePath(t *testing.T) {
 		// One compilation object
 		assert.EqualValues(t, 1, len(compilations))
 		// One source because we specified one file
-		assert.EqualValues(t, 1, len(compilations[0].Sources))
+		assert.EqualValues(t, 1, len(compilations[0].SourcePathToArtifact))
 		// Two contracts in SimpleContract.sol
 		contractCount := 0
-		for _, source := range compilations[0].Sources {
+		for _, source := range compilations[0].SourcePathToArtifact {
 			contractCount += len(source.Contracts)
 		}
 		assert.EqualValues(t, 2, contractCount)
@@ -82,10 +83,10 @@ func TestCryticSingleFileRelativePathSameDirectory(t *testing.T) {
 		// One compilation object
 		assert.EqualValues(t, 1, len(compilations))
 		// One source because we specified one file
-		assert.EqualValues(t, 1, len(compilations[0].Sources))
+		assert.EqualValues(t, 1, len(compilations[0].SourcePathToArtifact))
 		// Two contracts in SimpleContract.sol
 		contractCount := 0
-		for _, source := range compilations[0].Sources {
+		for _, source := range compilations[0].SourcePathToArtifact {
 			contractCount += len(source.Contracts)
 		}
 		assert.EqualValues(t, 2, contractCount)
@@ -118,10 +119,10 @@ func TestCryticSingleFileRelativePathChildDirectory(t *testing.T) {
 		// One compilation object
 		assert.EqualValues(t, 1, len(compilations))
 		// One source because we specified one file
-		assert.EqualValues(t, 1, len(compilations[0].Sources))
+		assert.EqualValues(t, 1, len(compilations[0].SourcePathToArtifact))
 		// Two contracts in SimpleContract.sol
 		contractCount := 0
-		for _, source := range compilations[0].Sources {
+		for _, source := range compilations[0].SourcePathToArtifact {
 			contractCount += len(source.Contracts)
 		}
 		assert.EqualValues(t, 2, contractCount)
@@ -160,9 +161,9 @@ func TestCryticSingleFileBuildDirectoryArgRelativePath(t *testing.T) {
 		// One compilation object
 		assert.EqualValues(t, 1, len(compilations))
 		// One source because we specified one file
-		assert.EqualValues(t, 1, len(compilations[0].Sources))
+		assert.EqualValues(t, 1, len(compilations[0].SourcePathToArtifact))
 		// Two contracts in SimpleContract.sol.
-		compiledSource := testCryticGetCompiledSourceByBaseName(compilations[0].Sources, contractName)
+		compiledSource := testCryticGetCompiledSourceByBaseName(compilations[0].SourcePathToArtifact, contractName)
 		assert.NotNil(t, compiledSource, "source file could not be resolved in compilation sources")
 		assert.EqualValues(t, 2, len(compiledSource.Contracts))
 	})
@@ -215,11 +216,11 @@ func TestCryticMultipleFiles(t *testing.T) {
 		// Verify there is one compilation object
 		assert.EqualValues(t, 1, len(compilations))
 		// Verify there are two sources
-		assert.EqualValues(t, 2, len(compilations[0].Sources))
+		assert.EqualValues(t, 2, len(compilations[0].SourcePathToArtifact))
 
 		// Verify there are three contracts
 		contractCount := 0
-		for _, source := range compilations[0].Sources {
+		for _, source := range compilations[0].SourcePathToArtifact {
 			contractCount += len(source.Contracts)
 		}
 		assert.EqualValues(t, 3, contractCount)
@@ -247,16 +248,16 @@ func TestCryticDirectoryNoArgs(t *testing.T) {
 		// Two compilation objects
 		assert.EqualValues(t, 2, len(compilations))
 		// One source per compilation unit
-		assert.EqualValues(t, 1, len(compilations[0].Sources))
-		assert.EqualValues(t, 1, len(compilations[1].Sources))
+		assert.EqualValues(t, 1, len(compilations[0].SourcePathToArtifact))
+		assert.EqualValues(t, 1, len(compilations[1].SourcePathToArtifact))
 
 		// Obtain the compiled source from both compilation units
 		firstContractName := "FirstContract.sol"
 		secondContractName := "SecondContract.sol"
-		firstUnitFirstContractSource := testCryticGetCompiledSourceByBaseName(compilations[0].Sources, firstContractName)
-		firstUnitSecondContractSource := testCryticGetCompiledSourceByBaseName(compilations[0].Sources, secondContractName)
-		secondUnitFirstContractSource := testCryticGetCompiledSourceByBaseName(compilations[1].Sources, firstContractName)
-		secondUnitSecondContractSource := testCryticGetCompiledSourceByBaseName(compilations[1].Sources, secondContractName)
+		firstUnitFirstContractSource := testCryticGetCompiledSourceByBaseName(compilations[0].SourcePathToArtifact, firstContractName)
+		firstUnitSecondContractSource := testCryticGetCompiledSourceByBaseName(compilations[0].SourcePathToArtifact, secondContractName)
+		secondUnitFirstContractSource := testCryticGetCompiledSourceByBaseName(compilations[1].SourcePathToArtifact, firstContractName)
+		secondUnitSecondContractSource := testCryticGetCompiledSourceByBaseName(compilations[1].SourcePathToArtifact, secondContractName)
 
 		// Assert that each compilation unit should have two contracts in it.
 		// Compilation unit ordering is non-deterministic in JSON output