add package dependency quality notes

Signed-off-by: Alex Goodman <[email protected]>
anchore · Nov 15, 2024 · 5b7ec60 · 5b7ec60
1 parent 70ef3f2
commit 5b7ec60
Show file tree

Hide file tree

Showing 105 changed files with 2,946 additions and 2,273 deletions.
diff --git a/internal/task/package_task_factory.go b/internal/task/package_task_factory.go
@@ -110,6 +110,8 @@ func NewPackageTask(cfg CatalogingFactoryConfig, c pkg.Cataloger, tags ...string
 
 		pkgs, relationships = applyCompliance(cfg.ComplianceConfig, pkgs, relationships)
 
+		finalizeDependencyCompleteness(pkgs)
+
 		sbom.AddPackages(pkgs...)
 		sbom.AddRelationships(relationships...)
 		t.Add(int64(len(pkgs)))
@@ -124,6 +126,13 @@ func NewPackageTask(cfg CatalogingFactoryConfig, c pkg.Cataloger, tags ...string
 	return NewTask(c.Name(), fn, tags...)
 }
 
+func finalizeDependencyCompleteness(pkgs []pkg.Package) {
+	// ensure that all packages have a non-empty dependency list
+	for i := range pkgs {
+		pkgs[i].Dependencies = pkg.ParseDependencyCompleteness(string(pkgs[i].Dependencies))
+	}
+}
+
 func finalizePkgCatalogerResults(cfg CatalogingFactoryConfig, resolver file.PathResolver, catalogerName string, pkgs []pkg.Package, relationships []artifact.Relationship) ([]pkg.Package, []artifact.Relationship) {
 	for i, p := range pkgs {
 		if p.FoundBy == "" {

diff --git a/syft/pkg/cataloger/alpine/cataloger_test.go b/syft/pkg/cataloger/alpine/cataloger_test.go
@@ -23,7 +23,8 @@ func TestApkDBCataloger(t *testing.T) {
 		Licenses: pkg.NewLicenseSet(
 			pkg.NewLicenseFromLocations("GPL-3.0-or-later", dbLocation),
 		),
-		Locations: file.NewLocationSet(dbLocation),
+		Locations:    file.NewLocationSet(dbLocation),
+		Dependencies: pkg.CompleteDependencies,
 		Metadata: pkg.ApkDBEntry{
 			Package:       "bash",
 			OriginPackage: "bash",
@@ -52,7 +53,8 @@ func TestApkDBCataloger(t *testing.T) {
 		Licenses: pkg.NewLicenseSet(
 			pkg.NewLicenseFromLocations("GPL-2.0-only", dbLocation),
 		),
-		Locations: file.NewLocationSet(dbLocation),
+		Locations:    file.NewLocationSet(dbLocation),
+		Dependencies: pkg.CompleteDependencies,
 		Metadata: pkg.ApkDBEntry{
 			Package:       "busybox-binsh",
 			OriginPackage: "busybox",
@@ -81,7 +83,8 @@ func TestApkDBCataloger(t *testing.T) {
 		Licenses: pkg.NewLicenseSet(
 			pkg.NewLicenseFromLocations("MIT", dbLocation),
 		),
-		Locations: file.NewLocationSet(dbLocation),
+		Locations:    file.NewLocationSet(dbLocation),
+		Dependencies: pkg.CompleteDependencies,
 		Metadata: pkg.ApkDBEntry{
 			Package:       "musl",
 			OriginPackage: "musl",
@@ -108,7 +111,8 @@ func TestApkDBCataloger(t *testing.T) {
 		Licenses: pkg.NewLicenseSet(
 			pkg.NewLicenseFromLocations("GPL-2.0-or-later", dbLocation),
 		),
-		Locations: file.NewLocationSet(dbLocation),
+		Locations:    file.NewLocationSet(dbLocation),
+		Dependencies: pkg.CompleteDependencies,
 		Metadata: pkg.ApkDBEntry{
 			Package:       "readline",
 			OriginPackage: "readline",

diff --git a/syft/pkg/cataloger/alpine/package.go b/syft/pkg/cataloger/alpine/package.go
@@ -20,13 +20,14 @@ func newPackage(d parsedData, release *linux.Release, dbLocation file.Location)
 	}
 
 	p := pkg.Package{
-		Name:      d.Package,
-		Version:   d.Version,
-		Locations: file.NewLocationSet(dbLocation.WithAnnotation(pkg.EvidenceAnnotationKey, pkg.PrimaryEvidenceAnnotation)),
-		Licenses:  pkg.NewLicenseSet(pkg.NewLicensesFromLocation(dbLocation, licenseStrings...)...),
-		PURL:      packageURL(d.ApkDBEntry, release),
-		Type:      pkg.ApkPkg,
-		Metadata:  d.ApkDBEntry,
+		Name:         d.Package,
+		Version:      d.Version,
+		Locations:    file.NewLocationSet(dbLocation.WithAnnotation(pkg.EvidenceAnnotationKey, pkg.PrimaryEvidenceAnnotation)),
+		Licenses:     pkg.NewLicenseSet(pkg.NewLicensesFromLocation(dbLocation, licenseStrings...)...),
+		PURL:         packageURL(d.ApkDBEntry, release),
+		Type:         pkg.ApkPkg,
+		Dependencies: pkg.CompleteDependencies,
+		Metadata:     d.ApkDBEntry,
 	}
 
 	p.SetID()

diff --git a/syft/pkg/cataloger/alpine/parse_apk_db_test.go b/syft/pkg/cataloger/alpine/parse_apk_db_test.go
@@ -90,7 +90,8 @@ func TestSinglePackageDetails(t *testing.T) {
 					pkg.NewLicense("BSD"),
 					pkg.NewLicense("GPL2+"),
 				),
-				Type: pkg.ApkPkg,
+				Type:         pkg.ApkPkg,
+				Dependencies: pkg.CompleteDependencies,
 				Metadata: pkg.ApkDBEntry{
 					Package:       "musl-utils",
 					OriginPackage: "musl",
@@ -177,7 +178,8 @@ func TestSinglePackageDetails(t *testing.T) {
 				Licenses: pkg.NewLicenseSet(
 					pkg.NewLicense("GPL-2.0-only"),
 				),
-				Type: pkg.ApkPkg,
+				Type:         pkg.ApkPkg,
+				Dependencies: pkg.CompleteDependencies,
 				Metadata: pkg.ApkDBEntry{
 					Package:       "alpine-baselayout-data",
 					OriginPackage: "alpine-baselayout",
@@ -221,8 +223,8 @@ func TestSinglePackageDetails(t *testing.T) {
 				Licenses: pkg.NewLicenseSet(
 					pkg.NewLicense("GPL-2.0-only"),
 				),
-				Type: pkg.ApkPkg,
-				PURL: "",
+				Type:         pkg.ApkPkg,
+				Dependencies: pkg.CompleteDependencies,
 				Metadata: pkg.ApkDBEntry{
 					Package:       "alpine-baselayout",
 					OriginPackage: "alpine-baselayout",

diff --git a/syft/pkg/cataloger/arch/cataloger_test.go b/syft/pkg/cataloger/arch/cataloger_test.go
@@ -34,7 +34,8 @@ func TestAlpmCataloger(t *testing.T) {
 		Licenses: pkg.NewLicenseSet(
 			pkg.NewLicenseFromLocations("MIT", treeSitterDbLocation),
 		),
-		Locations: file.NewLocationSet(treeSitterDbLocation),
+		Locations:    file.NewLocationSet(treeSitterDbLocation),
+		Dependencies: pkg.CompleteDependencies,
 		Metadata: pkg.AlpmDBEntry{
 			BasePackage:  "tree-sitter",
 			Package:      "tree-sitter",
@@ -60,7 +61,8 @@ func TestAlpmCataloger(t *testing.T) {
 		Licenses: pkg.NewLicenseSet(
 			pkg.NewLicenseFromLocations("GPL3", emacsDbLocation),
 		),
-		Locations: file.NewLocationSet(emacsDbLocation),
+		Locations:    file.NewLocationSet(emacsDbLocation),
+		Dependencies: pkg.CompleteDependencies,
 		Metadata: pkg.AlpmDBEntry{
 			BasePackage:  "emacs",
 			Package:      "emacs",
@@ -86,6 +88,7 @@ func TestAlpmCataloger(t *testing.T) {
 			fuzzyDbLocation,
 			file.NewLocation("var/lib/pacman/local/fuzzy-1.2-3/files"),
 		),
+		Dependencies: pkg.CompleteDependencies,
 		Metadata: pkg.AlpmDBEntry{
 			Package: "fuzzy",
 			Version: "1.2-3",
@@ -103,11 +106,12 @@ func TestAlpmCataloger(t *testing.T) {
 	}
 
 	madeupPkg := pkg.Package{
-		Name:      "madeup",
-		Version:   "20.30-4",
-		Type:      pkg.AlpmPkg,
-		FoundBy:   "alpm-db-cataloger",
-		Locations: file.NewLocationSet(madeupDbLocation),
+		Name:         "madeup",
+		Version:      "20.30-4",
+		Type:         pkg.AlpmPkg,
+		FoundBy:      "alpm-db-cataloger",
+		Locations:    file.NewLocationSet(madeupDbLocation),
+		Dependencies: pkg.CompleteDependencies,
 		Metadata: pkg.AlpmDBEntry{
 			Package: "madeup",
 			Version: "20.30-4",
@@ -131,6 +135,7 @@ func TestAlpmCataloger(t *testing.T) {
 			file.NewLocation("var/lib/pacman/local/gmp-6.2.1-2/files"),
 			file.NewLocation("var/lib/pacman/local/gmp-6.2.1-2/mtree"),
 		),
+		Dependencies: pkg.CompleteDependencies,
 		Metadata: pkg.AlpmDBEntry{
 			BasePackage:  "gmp",
 			Package:      "gmp",

diff --git a/syft/pkg/cataloger/arch/package.go b/syft/pkg/cataloger/arch/package.go
@@ -16,13 +16,14 @@ func newPackage(m *parsedData, release *linux.Release, dbLocation file.Location,
 	locs.Add(otherLocations...)
 
 	p := pkg.Package{
-		Name:      m.Package,
-		Version:   m.Version,
-		Locations: locs,
-		Licenses:  pkg.NewLicenseSet(pkg.NewLicensesFromLocation(dbLocation.WithoutAnnotations(), licenseCandidates...)...),
-		Type:      pkg.AlpmPkg,
-		PURL:      packageURL(m, release),
-		Metadata:  m.AlpmDBEntry,
+		Name:         m.Package,
+		Version:      m.Version,
+		Locations:    locs,
+		Licenses:     pkg.NewLicenseSet(pkg.NewLicensesFromLocation(dbLocation.WithoutAnnotations(), licenseCandidates...)...),
+		Type:         pkg.AlpmPkg,
+		PURL:         packageURL(m, release),
+		Dependencies: pkg.CompleteDependencies,
+		Metadata:     m.AlpmDBEntry,
 	}
 	p.SetID()
 

diff --git a/syft/pkg/cataloger/binary/classifier_package.go b/syft/pkg/cataloger/binary/classifier_package.go
@@ -35,6 +35,9 @@ func newClassifierPackage(classifier Classifier, location file.Location, matchMe
 		Type:    pkg.BinaryPkg,
 		CPEs:    cpes,
 		FoundBy: catalogerName,
+		// classifiers are limited to identifying package identities, but have no information about dependencies.
+		// Furthermore, we have no understanding about the mechanisms for dependency resolution for these packages.
+		Dependencies: pkg.UnknownDependencyCompleteness,
 		Metadata: pkg.BinarySignature{
 			Matches: []pkg.ClassifierMatch{
 				{

diff --git a/syft/pkg/cataloger/binary/elf_package.go b/syft/pkg/cataloger/binary/elf_package.go
@@ -16,7 +16,10 @@ func newELFPackage(metadata elfBinaryPackageNotes, locations file.LocationSet) p
 		PURL:      packageURL(metadata),
 		Type:      pkgType(metadata.Type),
 		Locations: locations,
-		Metadata:  metadata.ELFBinaryPackageNoteJSONPayload,
+		// though we can look for shared libs, we cannot see static dependencies nor dynamic dependencies using dlopen.
+		// this means that, even in cases where the dep info is actually complete, we can't programmatically determine that.
+		Dependencies: pkg.IncompleteDependencies,
+		Metadata:     metadata.ELFBinaryPackageNoteJSONPayload,
 	}
 
 	p.SetID()

diff --git a/syft/pkg/cataloger/binary/elf_package_cataloger_test.go b/syft/pkg/cataloger/binary/elf_package_cataloger_test.go
@@ -35,7 +35,8 @@ func Test_ELF_Package_Cataloger(t *testing.T) {
 						pkg.License{Value: "MIT", SPDXExpression: "MIT", Type: "declared"},
 					),
 
-					Type: pkg.BinaryPkg,
+					Type:         pkg.BinaryPkg,
+					Dependencies: pkg.IncompleteDependencies,
 					Metadata: pkg.ELFBinaryPackageNoteJSONPayload{
 						Type:       "testfixture",
 						Vendor:     "syft",
@@ -56,7 +57,8 @@ func Test_ELF_Package_Cataloger(t *testing.T) {
 					Licenses: pkg.NewLicenseSet(
 						pkg.License{Value: "MIT", SPDXExpression: "MIT", Type: "declared"},
 					),
-					Type: pkg.BinaryPkg,
+					Type:         pkg.BinaryPkg,
+					Dependencies: pkg.IncompleteDependencies,
 					Metadata: pkg.ELFBinaryPackageNoteJSONPayload{
 						Type:       "testfixture",
 						Vendor:     "syft",
@@ -80,8 +82,9 @@ func Test_ELF_Package_Cataloger(t *testing.T) {
 						file.NewLocation("/sha256sum").WithAnnotation(pkg.EvidenceAnnotationKey, pkg.PrimaryEvidenceAnnotation),
 						file.NewLocation("/sha1sum").WithAnnotation(pkg.EvidenceAnnotationKey, pkg.PrimaryEvidenceAnnotation),
 					),
-					Licenses: pkg.NewLicenseSet(),
-					Type:     pkg.RpmPkg,
+					Licenses:     pkg.NewLicenseSet(),
+					Type:         pkg.RpmPkg,
+					Dependencies: pkg.IncompleteDependencies,
 					Metadata: pkg.ELFBinaryPackageNoteJSONPayload{
 						Type:         "rpm",
 						Architecture: "x86_64",
@@ -102,8 +105,9 @@ func Test_ELF_Package_Cataloger(t *testing.T) {
 						file.NewLocation("/sha256sum").WithAnnotation(pkg.EvidenceAnnotationKey, pkg.PrimaryEvidenceAnnotation),
 						file.NewLocation("/sha1sum").WithAnnotation(pkg.EvidenceAnnotationKey, pkg.PrimaryEvidenceAnnotation),
 					),
-					Licenses: pkg.NewLicenseSet(),
-					Type:     pkg.RpmPkg,
+					Licenses:     pkg.NewLicenseSet(),
+					Type:         pkg.RpmPkg,
+					Dependencies: pkg.IncompleteDependencies,
 					Metadata: pkg.ELFBinaryPackageNoteJSONPayload{
 						Type:         "rpm",
 						Architecture: "arm",

diff --git a/syft/pkg/cataloger/binary/elf_package_test.go b/syft/pkg/cataloger/binary/elf_package_test.go
@@ -154,10 +154,11 @@ func Test_newELFPackage(t *testing.T) {
 			},
 
 			expected: pkg.Package{
-				Name:    "syfttestfixture",
-				Version: "0.01",
-				Type:    "binary",
-				PURL:    "pkg:generic/syftsys/[email protected]",
+				Name:         "syfttestfixture",
+				Version:      "0.01",
+				Type:         "binary",
+				PURL:         "pkg:generic/syftsys/[email protected]",
+				Dependencies: pkg.IncompleteDependencies,
 				Metadata: pkg.ELFBinaryPackageNoteJSONPayload{
 					Type:   "binary",
 					System: "syftsys",

diff --git a/syft/pkg/cataloger/cpp/package.go b/syft/pkg/cataloger/cpp/package.go
@@ -67,35 +67,43 @@ func splitConanRef(ref string) *conanRef {
 }
 
 func newConanfilePackage(m pkg.ConanfileEntry, locations ...file.Location) *pkg.Package {
-	return newConanPackage(m.Ref, m, locations...)
+	// though a conanfile is a listing of direct dependencies, we are not capturing these today, so are forced to answer incomplete
+	return newConanPackage(m.Ref, m, pkg.IncompleteDependencies, locations...)
 }
 
-func newConanlockPackage(m pkg.ConanV1LockEntry, locations ...file.Location) *pkg.Package {
-	return newConanPackage(m.Ref, m, locations...)
+func newConanlockV1Package(m pkg.ConanV1LockEntry, locations ...file.Location) *pkg.Package {
+	// conan.lock is primarily used to lock the dependency graph at specific versions to ensure consistent builds,
+	// and the dependency tree can be inferred from the contents (distinguishing between direct and transitive dependencies)
+	return newConanPackage(m.Ref, m, pkg.CompleteDependencies, locations...)
 }
 
-func newConanReferencePackage(m pkg.ConanV2LockEntry, locations ...file.Location) *pkg.Package {
-	return newConanPackage(m.Ref, m, locations...)
+func newConanLockv2Package(m pkg.ConanV2LockEntry, locations ...file.Location) *pkg.Package {
+	// conan.lock is primarily used to lock the dependency graph at specific versions to ensure consistent builds,
+	// and the dependency tree can be inferred from the contents (distinguishing between direct and transitive dependencies)
+	return newConanPackage(m.Ref, m, pkg.CompleteDependencies, locations...)
 }
 
 func newConaninfoPackage(m pkg.ConaninfoEntry, locations ...file.Location) *pkg.Package {
-	return newConanPackage(m.Ref, m, locations...)
+	// conaninfo.txt is generated during the build and contains detailed information about the entire dependency tree,
+	// and today the codebase does infer direct dependencies from this source
+	return newConanPackage(m.Ref, m, pkg.CompleteDependencies, locations...)
 }
 
-func newConanPackage(refStr string, metadata any, locations ...file.Location) *pkg.Package {
+func newConanPackage(refStr string, metadata any, dep pkg.DependencyCompleteness, locations ...file.Location) *pkg.Package {
 	ref := splitConanRef(refStr)
 	if ref == nil {
 		return nil
 	}
 
 	p := pkg.Package{
-		Name:      ref.Name,
-		Version:   ref.Version,
-		Locations: file.NewLocationSet(locations...),
-		PURL:      packageURL(ref),
-		Language:  pkg.CPP,
-		Type:      pkg.ConanPkg,
-		Metadata:  metadata,
+		Name:         ref.Name,
+		Version:      ref.Version,
+		Locations:    file.NewLocationSet(locations...),
+		PURL:         packageURL(ref),
+		Language:     pkg.CPP,
+		Type:         pkg.ConanPkg,
+		Dependencies: dep,
+		Metadata:     metadata,
 	}
 
 	p.SetID()

diff --git a/syft/pkg/cataloger/cpp/parse_conanfile.go b/syft/pkg/cataloger/cpp/parse_conanfile.go
@@ -22,11 +22,12 @@ func parseConanfile(_ context.Context, _ file.Resolver, _ *generic.Environment,
 	r := bufio.NewReader(reader)
 	inRequirements := false
 	var pkgs []pkg.Package
+loop:
 	for {
 		line, err := r.ReadString('\n')
 		switch {
 		case errors.Is(err, io.EOF):
-			return pkgs, nil, unknown.IfEmptyf(pkgs, "unable to determine packages")
+			break loop
 		case err != nil:
 			return nil, nil, fmt.Errorf("failed to parse conanfile.txt file: %w", err)
 		}
@@ -56,4 +57,5 @@ func parseConanfile(_ context.Context, _ file.Resolver, _ *generic.Environment,
 
 		pkgs = append(pkgs, *p)
 	}
+	return pkgs, nil, unknown.IfEmptyf(pkgs, "unable to determine packages")
 }