build source PURL from metadata and gate it to CycloneDX 1.6+

- the source PURL is built from package metadata instead of editing the binary PURL: only `distro` (plus `epoch` and `rpmmod` for rpm) is carried over, subpath is dropped, and the ref is skipped when the PURL type does not match the metadata ecosystem
- one emission rule for all ecosystems: only when the source name or version differs from the binary, so self-sourced rpm and alpm packages no longer get a redundant ref
- source RPM names must end in `.src.rpm` or `.nosrc.rpm`, and `nosrc` maps to `arch=nosrc`
- `source-distribution` refs are dropped when encoding below spec 1.6, since cyclonedx-go does not downgrade them (CycloneDX/cyclonedx-go#291)

Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>

syft/format/cyclonedxjson/encoder_test.go

@@ -128,6 +128,31 @@ func TestCycloneDxSourcePackageEncoder(t *testing.T) {
 	)
 }
 
+func TestCycloneDxSourcePackageEncoder_SpecVersions(t *testing.T) {
+	s := testutil.SourcePackagesInput(t, t.TempDir())
+	tests := []struct {
+		version string
+		want    bool
+	}{
+		{version: "1.4", want: false},
+		{version: "1.5", want: false},
+		{version: "1.6", want: true},
+		{version: cyclonedxutil.DefaultVersion, want: true},
+	}
+	for _, tt := range tests {
+		t.Run(tt.version, func(t *testing.T) {
+			enc, err := NewFormatEncoderWithConfig(EncoderConfig{Version: tt.version})
+			require.NoError(t, err)
+
+			var buf bytes.Buffer
+			require.NoError(t, enc.Encode(&buf, s))
+
+			// the source-distribution reference type only exists from spec 1.6
+			assert.Equal(t, tt.want, strings.Contains(buf.String(), `"source-distribution"`))
+		})
+	}
+}
+
 func redactor(values ...string) testutil.Redactor {
 	return testutil.NewRedactions().
 		WithValuesRedacted(values...).

syft/format/internal/cyclonedxutil/encoder.go

@@ -2,6 +2,7 @@ package cyclonedxutil
 
 import (
 	"io"
+	"slices"
 
 	"github.com/CycloneDX/cyclonedx-go"
 
@@ -29,9 +30,35 @@ func NewEncoder(version string, format cyclonedx.BOMFileFormat, pretty bool) (En
 
 func (e Encoder) Encode(writer io.Writer, s sbom.SBOM) error {
 	bom := cyclonedxhelpers.ToFormatModel(s)
+	if e.version < cyclonedx.SpecVersion1_6 {
+		// cyclonedx-go does not downgrade this 1.6+ reference type for older spec versions
+		// (see https://github.com/CycloneDX/cyclonedx-go/issues/291), so drop it rather than emit an invalid document
+		dropExternalReferenceType(bom.Components, cyclonedx.ERTypeSourceDistribution)
+	}
 	enc := cyclonedx.NewBOMEncoder(writer, e.format)
 	enc.SetPretty(e.pretty)
 	enc.SetEscapeHTML(false)
 
 	return enc.EncodeVersion(bom, e.version)
 }
+
+func dropExternalReferenceType(components *[]cyclonedx.Component, refType cyclonedx.ExternalReferenceType) {
+	if components == nil {
+		return
+	}
+	for i := range *components {
+		c := &(*components)[i]
+		dropExternalReferenceType(c.Components, refType)
+		if c.ExternalReferences == nil {
+			continue
+		}
+		refs := slices.DeleteFunc(*c.ExternalReferences, func(r cyclonedx.ExternalReference) bool {
+			return r.Type == refType
+		})
+		if len(refs) == 0 {
+			c.ExternalReferences = nil
+			continue
+		}
+		c.ExternalReferences = &refs
+	}
+}

syft/format/internal/cyclonedxutil/helpers/component.go

@@ -47,14 +47,6 @@ func EncodeComponent(p pkg.Package, supplier string, locationSorter func(a, b fi
 		componentType = cyclonedx.ComponentTypeMachineLearningModel
 	}
 
-	externalRefs := encodeExternalReferences(p)
-	if srcRef := encodeSourcePackageExternalReference(p); srcRef != nil {
-		if externalRefs == nil {
-			externalRefs = &[]cyclonedx.ExternalReference{}
-		}
-		*externalRefs = append(*externalRefs, *srcRef)
-	}
-
 	return cyclonedx.Component{
 		Type:               componentType,
 		Name:               p.Name,
@@ -67,7 +59,7 @@ func EncodeComponent(p pkg.Package, supplier string, locationSorter func(a, b fi
 		Author:             encodeAuthor(p),
 		Publisher:          encodePublisher(p),
 		Description:        encodeDescription(p),
-		ExternalReferences: externalRefs,
+		ExternalReferences: encodeExternalReferences(p),
 		Properties:         properties,
 		BOMRef:             DeriveBomRef(p),
 	}

syft/format/internal/cyclonedxutil/helpers/external_references.go

@@ -80,42 +80,51 @@ func encodeExternalReferences(p pkg.Package) *[]cyclonedx.ExternalReference {
 			}
 		}
 	}
-
+	if srcRef := encodeSourcePackageExternalReference(p); srcRef != nil {
+		refs = append(refs, *srcRef)
+	}
 	if len(refs) > 0 {
 		return &refs
 	}
 	return nil
 }
 
+// encodeSourcePackageExternalReference returns a "source-distribution" reference holding the PURL of the source
+// package that an OS package was built from. It is only emitted when the source differs from the binary package
+// (by name or version) and when the package PURL is of the type implied by the metadata.
 func encodeSourcePackageExternalReference(p pkg.Package) *cyclonedx.ExternalReference {
-	name, version, arch := sourceCoordinates(p)
-	if name == "" || version == "" {
-		return nil
-	}
 	if p.PURL == "" {
 		return nil
 	}
-
-	pkgPurl, err := packageurl.FromString(p.PURL)
-	if err != nil {
+	src := sourcePackageOf(p)
+	if src == nil {
 		return nil
 	}
 
-	qualifiers := pkgPurl.Qualifiers.Map()
-	delete(qualifiers, pkg.PURLQualifierUpstream)
-	delete(qualifiers, pkg.PURLQualifierArch)
+	// the binary PURL is only consulted for the namespace and a small set of qualifiers (e.g. distro), everything
+	// else describes the binary package and must not be restated as a fact about the source package.
+	binPurl, err := packageurl.FromString(p.PURL)
+	if err != nil || binPurl.Type != src.purlType {
+		return nil
+	}
 
-	if arch != "" {
-		qualifiers[pkg.PURLQualifierArch] = arch
+	qualifiers := map[string]string{}
+	for _, q := range binPurl.Qualifiers {
+		if src.keepQualifier(q.Key) {
+			qualifiers[q.Key] = q.Value
+		}
+	}
+	if src.arch != "" {
+		qualifiers[pkg.PURLQualifierArch] = src.arch
 	}
 
 	srcPurl := packageurl.NewPackageURL(
-		pkgPurl.Type,
-		pkgPurl.Namespace,
-		name,
-		version,
+		src.purlType,
+		binPurl.Namespace,
+		src.name,
+		src.version,
 		packageurl.QualifiersFromMap(qualifiers),
-		pkgPurl.Subpath,
+		"",
 	)
 
 	return &cyclonedx.ExternalReference{
@@ -124,51 +133,91 @@ func encodeSourcePackageExternalReference(p pkg.Package) *cyclonedx.ExternalRefe
 	}
 }
 
-func sourceCoordinates(p pkg.Package) (name, version, arch string) {
-	switch metadata := p.Metadata.(type) {
+// sourcePackage is the identity of the source package a binary OS package was built from.
+type sourcePackage struct {
+	purlType string
+	name     string
+	version  string
+	// arch is the ecosystem's source architecture marker, empty when the ecosystem has no such convention
+	arch string
+}
+
+func (s sourcePackage) keepQualifier(key string) bool {
+	switch key {
+	case pkg.PURLQualifierDistro:
+		return true
+	case pkg.PURLQualifierEpoch, pkg.PURLQualifierRpmModularity:
+		return s.purlType == pkg.RpmPkg.PackageURLType()
+	}
+	return false
+}
+
+// sourcePackageOf returns the source package for p, or nil when there is none or it is the same as the binary.
+func sourcePackageOf(p pkg.Package) *sourcePackage {
+	var src sourcePackage
+	var binName, binVersion string
+	switch m := p.Metadata.(type) {
 	case pkg.ApkDBEntry:
-		if metadata.OriginPackage != "" && metadata.OriginPackage != metadata.Package {
-			return metadata.OriginPackage, metadata.Version, ""
-		}
+		// apk has no source arch convention, so the source PURL carries no arch
+		src = sourcePackage{purlType: pkg.ApkPkg.PackageURLType(), name: m.OriginPackage, version: m.Version}
+		binName, binVersion = m.Package, m.Version
 	case pkg.DpkgDBEntry:
-		return dpkgSourceCoordinates(metadata)
+		src = dpkgSourcePackage(m)
+		binName, binVersion = m.Package, m.Version
 	case pkg.DpkgArchiveEntry:
-		return dpkgSourceCoordinates(pkg.DpkgDBEntry(metadata))
+		src = dpkgSourcePackage(pkg.DpkgDBEntry(m))
+		binName, binVersion = m.Package, m.Version
 	case pkg.RpmDBEntry:
-		return rpmSourceCoordinates(metadata.SourceRpm)
+		src = rpmSourcePackage(m.SourceRpm)
+		binName, binVersion = m.Name, m.Version+"-"+m.Release
 	case pkg.RpmArchive:
-		return rpmSourceCoordinates(metadata.SourceRpm)
+		src = rpmSourcePackage(m.SourceRpm)
+		binName, binVersion = m.Name, m.Version+"-"+m.Release
 	case pkg.AlpmDBEntry:
-		if metadata.BasePackage != "" {
-			return metadata.BasePackage, metadata.Version, ""
-		}
+		// alpm has no source arch convention, so the source PURL carries no arch
+		src = sourcePackage{purlType: pkg.AlpmPkg.PackageURLType(), name: m.BasePackage, version: m.Version}
+		binName, binVersion = m.Package, m.Version
+	default:
+		return nil
 	}
-	return "", "", ""
+	if src.name == "" || src.version == "" {
+		return nil
+	}
+	if src.name == binName && src.version == binVersion {
+		return nil
+	}
+	return &src
 }
 
-func dpkgSourceCoordinates(entry pkg.DpkgDBEntry) (name, version, arch string) {
-	if entry.Source == "" {
-		return "", "", ""
+func dpkgSourcePackage(entry pkg.DpkgDBEntry) sourcePackage {
+	version := entry.SourceVersion
+	if version == "" {
+		version = entry.Version
 	}
-	if entry.SourceVersion != "" {
-		return entry.Source, entry.SourceVersion, "source"
-	}
-	return entry.Source, entry.Version, "source"
+	return sourcePackage{purlType: pkg.DebPkg.PackageURLType(), name: entry.Source, version: version, arch: "source"}
 }
 
-func rpmSourceCoordinates(sourceRpm string) (name, version, arch string) {
-	// <name>-<version>-<release>.src.rpm
-	sourceRpm = strings.TrimSuffix(sourceRpm, ".rpm")
-	sourceRpm = strings.TrimSuffix(strings.TrimSuffix(sourceRpm, ".src"), ".nosrc")
-	release := strings.LastIndex(sourceRpm, "-")
+// rpmSourcePackage parses a source RPM filename of the form <name>-<version>-<release>.(no)src.rpm.
+func rpmSourcePackage(sourceRpm string) sourcePackage {
+	var arch string
+	switch {
+	case strings.HasSuffix(sourceRpm, ".src.rpm"):
+		arch = "src"
+	case strings.HasSuffix(sourceRpm, ".nosrc.rpm"):
+		arch = "nosrc"
+	default:
+		return sourcePackage{}
+	}
+	nvr := strings.TrimSuffix(sourceRpm, "."+arch+".rpm")
+	release := strings.LastIndex(nvr, "-")
 	if release < 0 {
-		return "", "", ""
+		return sourcePackage{}
 	}
-	ver := strings.LastIndex(sourceRpm[:release], "-")
+	ver := strings.LastIndex(nvr[:release], "-")
 	if ver < 0 {
-		return "", "", ""
+		return sourcePackage{}
 	}
-	return sourceRpm[:ver], sourceRpm[ver+1:], "src"
+	return sourcePackage{purlType: pkg.RpmPkg.PackageURLType(), name: nvr[:ver], version: nvr[ver+1:], arch: arch}
 }
 
 // supported algorithm in cycloneDX as of 1.4

syft/format/internal/cyclonedxutil/helpers/external_references_test.go

@@ -164,7 +164,7 @@ func Test_encodeSourcePackage(t *testing.T) {
 				Type:    pkg.ApkPkg,
 				Metadata: pkg.ApkDBEntry{
 					Package:       "some-package",
-					OriginPackage: "some-package",
+					OriginPackage: "some-origin",
 					Version:       "1.0.0-r1",
 					Architecture:  "x86_64",
 				},
@@ -190,7 +190,6 @@ func Test_encodeSourcePackage(t *testing.T) {
 			},
 		},
 		{
-			// Matching existing implementation where upstream qualifier is omitted when OriginPackage = Package for Alpine
 			name: "from apk with source = bin",
 			input: pkg.Package{
 				Name:    "some-package",
@@ -312,9 +311,7 @@ func Test_encodeSourcePackage(t *testing.T) {
 					SourceRpm: "some-package-1.0.0-1.el8.src.rpm",
 				},
 			},
-			expected: &cyclonedx.ExternalReference{
-				Type: cyclonedx.ERTypeSourceDistribution, URL: "pkg:rpm/centos/some-package@1.0.0-1.el8?arch=src&distro=centos-8&epoch=32",
-			},
+			expected: nil,
 		},
 		{
 			name: "from rpm archive",
@@ -355,8 +352,6 @@ func Test_encodeSourcePackage(t *testing.T) {
 			},
 		},
 		{
-			// Matching existing implementation where upstream qualifier is *not* omitted when BasePackage == Package for Alpm
-			// This differs from the Alpine case where it is omitted.
 			name: "from alpm with source = bin",
 			input: pkg.Package{
 				Name:    "some-package",
@@ -370,9 +365,147 @@ func Test_encodeSourcePackage(t *testing.T) {
 					Architecture: "x86_64",
 				},
 			},
-			expected: &cyclonedx.ExternalReference{
-				Type: cyclonedx.ERTypeSourceDistribution, URL: "pkg:alpm/arch/some-package@1.0.0?distro=arch-rolling",
+			expected: nil,
+		},
+		{
+			name: "from rpm with dashes in the source name",
+			input: pkg.Package{
+				Name: "perl-Foo-Bar-libs",
+				Type: pkg.RpmPkg,
+				PURL: "pkg:rpm/redhat/perl-Foo-Bar-libs@1.0-1.el8?arch=x86_64&distro=rhel-8",
+				Metadata: pkg.RpmDBEntry{
+					Name:      "perl-Foo-Bar-libs",
+					Version:   "1.0",
+					Release:   "1.el8",
+					SourceRpm: "perl-Foo-Bar-1.0-1.el8.src.rpm",
+				},
 			},
+			expected: &cyclonedx.ExternalReference{
+				Type: cyclonedx.ERTypeSourceDistribution, URL: "pkg:rpm/redhat/perl-Foo-Bar@1.0-1.el8?arch=src&distro=rhel-8",
+			},
+		},
+		{
+			name: "from rpm nosrc",
+			input: pkg.Package{
+				Name: "some-firmware",
+				Type: pkg.RpmPkg,
+				PURL: "pkg:rpm/redhat/some-firmware@1.0-1.el8?arch=noarch&distro=rhel-8",
+				Metadata: pkg.RpmDBEntry{
+					Name:      "some-firmware",
+					Version:   "1.0",
+					Release:   "1.el8",
+					SourceRpm: "firmware-1.0-1.el8.nosrc.rpm",
+				},
+			},
+			expected: &cyclonedx.ExternalReference{
+				Type: cyclonedx.ERTypeSourceDistribution, URL: "pkg:rpm/redhat/firmware@1.0-1.el8?arch=nosrc&distro=rhel-8",
+			},
+		},
+		{
+			name: "from rpm keeps rpmmod",
+			input: pkg.Package{
+				Name: "nodejs-libs",
+				Type: pkg.RpmPkg,
+				PURL: "pkg:rpm/redhat/nodejs-libs@18.0-1.el8?arch=x86_64&distro=rhel-8&rpmmod=nodejs%3A18",
+				Metadata: pkg.RpmDBEntry{
+					Name:      "nodejs-libs",
+					Version:   "18.0",
+					Release:   "1.el8",
+					SourceRpm: "nodejs-18.0-1.el8.src.rpm",
+				},
+			},
+			expected: &cyclonedx.ExternalReference{
+				Type: cyclonedx.ERTypeSourceDistribution, URL: "pkg:rpm/redhat/nodejs@18.0-1.el8?arch=src&distro=rhel-8&rpmmod=nodejs%3A18",
+			},
+		},
+		{
+			name: "from rpm with malformed source rpm",
+			input: pkg.Package{
+				Name: "gpg-pubkey",
+				Type: pkg.RpmPkg,
+				PURL: "pkg:rpm/redhat/gpg-pubkey@1-1?distro=rhel-8",
+				Metadata: pkg.RpmDBEntry{
+					Name:      "gpg-pubkey",
+					Version:   "1",
+					Release:   "1",
+					SourceRpm: "(none)",
+				},
+			},
+			expected: nil,
+		},
+		{
+			name: "from rpm with non-source rpm filename",
+			input: pkg.Package{
+				Name: "bash",
+				Type: pkg.RpmPkg,
+				PURL: "pkg:rpm/redhat/bash@5.1-2?arch=x86_64&distro=rhel-8",
+				Metadata: pkg.RpmDBEntry{
+					Name:      "bash",
+					Version:   "5.1",
+					Release:   "2",
+					SourceRpm: "bash-5.1-2.x86_64.rpm",
+				},
+			},
+			expected: nil,
+		},
+		{
+			name: "from rpm with too few dashes",
+			input: pkg.Package{
+				Name: "bash",
+				Type: pkg.RpmPkg,
+				PURL: "pkg:rpm/redhat/bash@5.1-2?arch=x86_64&distro=rhel-8",
+				Metadata: pkg.RpmDBEntry{
+					Name:      "bash",
+					Version:   "5.1",
+					Release:   "2",
+					SourceRpm: "bash-5.1.src.rpm",
+				},
+			},
+			expected: nil,
+		},
+		{
+			name: "binary qualifiers and subpath are not carried over",
+			input: pkg.Package{
+				Name: "libssl3",
+				Type: pkg.DebPkg,
+				PURL: "pkg:deb/debian/libssl3@3.0.0-1?arch=amd64&distro=debian-12&checksum=sha256:dead&download_url=https://example.com/x.deb&cpes=cpe:2.3:a:x:x:1:*:*:*:*:*:*:*#sub/dir",
+				Metadata: pkg.DpkgDBEntry{
+					Package: "libssl3",
+					Version: "3.0.0-1",
+					Source:  "openssl",
+				},
+			},
+			expected: &cyclonedx.ExternalReference{
+				Type: cyclonedx.ERTypeSourceDistribution, URL: "pkg:deb/debian/openssl@3.0.0-1?arch=source&distro=debian-12",
+			},
+		},
+		{
+			name: "PURL type does not match metadata",
+			input: pkg.Package{
+				Name: "x",
+				Type: pkg.DebPkg,
+				PURL: "pkg:npm/x@1",
+				Metadata: pkg.DpkgDBEntry{
+					Package: "x",
+					Version: "1",
+					Source:  "lodash",
+				},
+			},
+			expected: nil,
+		},
+		{
+			name: "invalid PURL",
+			input: pkg.Package{
+				Name: "x",
+				Type: pkg.DebPkg,
+				PURL: "not-a-purl",
+				Metadata: pkg.DpkgDBEntry{
+					Package: "x",
+					Version: "1",
+					Source:  "y",
+				},
+			},
+			expected: nil,
 		},
 	}
 	for _, test := range tests {