feat: update licenses to including license content when SPDX expressions are unable to be determined (#3366)

---------
Signed-off-by: HeyeOpenSource <opensource@heye-international.com>
Signed-off-by: Christopher Phillips <32073428+spiffcs@users.noreply.github.com>
Co-authored-by: Christopher Phillips <32073428+spiffcs@users.noreply.github.com>

syft/format/common/spdxhelpers/to_format_model.go

@@ -15,6 +15,7 @@ import (
 	"github.com/spdx/tools-golang/spdx"
 
 	"github.com/anchore/packageurl-go"
+	internallicenses "github.com/anchore/syft/internal/licenses"
 	"github.com/anchore/syft/internal/log"
 	"github.com/anchore/syft/internal/mimetype"
 	"github.com/anchore/syft/internal/relationship"
@@ -762,10 +763,16 @@ func toOtherLicenses(catalog *pkg.Collection) []*spdx.OtherLicense {
 		if license.Value == "" {
 			value, _ = strings.CutPrefix(license.ID, "LicenseRef-")
 		}
-		result = append(result, &spdx.OtherLicense{
+		other := &spdx.OtherLicense{
 			LicenseIdentifier: license.ID,
 			ExtractedText:     value,
-		})
+		}
+		customPrefix := spdxlicense.LicenseRefPrefix + helpers.SanitizeElementID(internallicenses.UnknownLicensePrefix)
+		if strings.HasPrefix(license.ID, customPrefix) {
+			other.LicenseName = strings.TrimPrefix(license.ID, customPrefix)
+			other.LicenseComment = strings.Trim(internallicenses.UnknownLicensePrefix, "-_")
+		}
+		result = append(result, other)
 	}
 	return result
 }

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

@@ -1,10 +1,12 @@
 package helpers
 
 import (
+	"encoding/base64"
 	"strings"
 
 	"github.com/CycloneDX/cyclonedx-go"
 
+	"github.com/anchore/syft/internal/licenses"
 	"github.com/anchore/syft/internal/spdxlicense"
 	"github.com/anchore/syft/syft/pkg"
 )
@@ -110,7 +112,7 @@ func separateLicenses(p pkg.Package) (spdx, other cyclonedx.Licenses, expression
 			continue
 		}
 
-		if l.SPDXExpression != "" {
+		if l.SPDXExpression != "" && !strings.HasPrefix(l.SPDXExpression, licenses.UnknownLicensePrefix) {
 			// COMPLEX EXPRESSION CASE
 			ex = append(ex, l.SPDXExpression)
 			continue
@@ -118,17 +120,43 @@ func separateLicenses(p pkg.Package) (spdx, other cyclonedx.Licenses, expression
 
 		// license string that are not valid spdx expressions or ids
 		// we only use license Name here since we cannot guarantee that the license is a valid SPDX expression
-		if len(l.URLs) > 0 {
+		if len(l.URLs) > 0 && !strings.HasPrefix(l.SPDXExpression, licenses.UnknownLicensePrefix) {
 			processLicenseURLs(l, "", &otherc)
 			continue
 		}
-		otherc = append(otherc, cyclonedx.LicenseChoice{
+
+		otherc = append(otherc, processCustomLicense(l)...)
+	}
+	return spdxc, otherc, ex
+}
+
+func processCustomLicense(l pkg.License) cyclonedx.Licenses {
+	result := cyclonedx.Licenses{}
+	if strings.HasPrefix(l.SPDXExpression, licenses.UnknownLicensePrefix) {
+		cyclonedxLicense := &cyclonedx.License{
+			Name: l.SPDXExpression,
+		}
+		if len(l.URLs) > 0 {
+			cyclonedxLicense.URL = l.URLs[0]
+		}
+		if len(l.Contents) > 0 {
+			cyclonedxLicense.Text = &cyclonedx.AttachedText{
+				Content: base64.StdEncoding.EncodeToString([]byte(l.Contents)),
+			}
+			cyclonedxLicense.Text.ContentType = "text/plain"
+			cyclonedxLicense.Text.Encoding = "base64"
+		}
+		result = append(result, cyclonedx.LicenseChoice{
+			License: cyclonedxLicense,
+		})
+	} else {
+		result = append(result, cyclonedx.LicenseChoice{
 			License: &cyclonedx.License{
 				Name: l.Value,
 			},
 		})
 	}
-	return spdxc, otherc, ex
+	return result
 }
 
 func processLicenseURLs(l pkg.License, spdxID string, populate *cyclonedx.Licenses) {

syft/format/internal/spdxutil/helpers/license.go

@@ -5,6 +5,7 @@ import (
 	"fmt"
 	"strings"
 
+	"github.com/anchore/syft/internal/licenses"
 	"github.com/anchore/syft/internal/spdxlicense"
 	"github.com/anchore/syft/syft/license"
 	"github.com/anchore/syft/syft/pkg"
@@ -69,19 +70,26 @@ func ParseLicenses(raw []pkg.License) (concluded, declared []SPDXLicense) {
 		}
 
 		candidate := SPDXLicense{}
-		if l.SPDXExpression != "" {
+		if l.SPDXExpression != "" && !strings.HasPrefix(l.SPDXExpression, licenses.UnknownLicensePrefix) {
 			candidate.ID = l.SPDXExpression
 		} else {
-			// we did not find a valid SPDX license ID so treat as separate license
-			if len(l.Value) <= 64 {
-				// if the license text is less than the size of the hash,
-				// just use it directly so the id is more readable
-				candidate.ID = spdxlicense.LicenseRefPrefix + SanitizeElementID(l.Value)
-			} else {
-				hash := sha256.Sum256([]byte(l.Value))
-				candidate.ID = fmt.Sprintf("%s%x", spdxlicense.LicenseRefPrefix, hash)
-			}
 			candidate.Value = l.Value
+			// we did not find a valid SPDX license ID so treat as separate license
+			if strings.HasPrefix(l.SPDXExpression, licenses.UnknownLicensePrefix) {
+				candidate.ID = spdxlicense.LicenseRefPrefix + SanitizeElementID(l.SPDXExpression)
+				if len(l.Contents) > 0 {
+					candidate.Value = l.Contents
+				}
+			} else {
+				if len(l.Value) <= 64 {
+					// if the license text is less than the size of the hash,
+					// just use it directly so the id is more readable
+					candidate.ID = spdxlicense.LicenseRefPrefix + SanitizeElementID(l.Value)
+				} else {
+					hash := sha256.Sum256([]byte(l.Value))
+					candidate.ID = fmt.Sprintf("%s%x", spdxlicense.LicenseRefPrefix, hash)
+				}
+			}
 		}
 
 		switch l.Type {

syft/format/syftjson/model/package.go

@@ -51,6 +51,7 @@ type License struct {
 	Type           license.Type    `json:"type"`
 	URLs           []string        `json:"urls"`
 	Locations      []file.Location `json:"locations"`
+	Contents       string          `json:"contents,omitempty"`
 }
 
 func newModelLicensesFromValues(licenses []string) (ml []License) {

syft/format/syftjson/to_format_model.go

@@ -234,6 +234,7 @@ func toLicenseModel(pkgLicenses []pkg.License) (modelLicenses []model.License) {
 			Type:           l.Type,
 			URLs:           urls,
 			Locations:      locations,
+			Contents:       l.Contents,
 		})
 	}
 	return

syft/format/syftjson/to_syft_model.go

@@ -164,6 +164,7 @@ func toSyftLicenses(m []model.License) (p []pkg.License) {
 			Type:           l.Type,
 			URLs:           l.URLs,
 			Locations:      file.NewLocationSet(l.Locations...),
+			Contents:       l.Contents,
 		})
 	}
 	return

syft/pkg/cataloger/golang/licenses.go

@@ -36,6 +36,7 @@ type goLicense struct {
 	Type           license.Type `json:"type,omitempty"`
 	URLs           []string     `json:"urls,omitempty"`
 	Locations      []string     `json:"locations,omitempty"`
+	Contents       string       `json:"contents,omitempty"`
 }
 
 type goLicenseResolver struct {
@@ -449,6 +450,7 @@ func toPkgLicenses(goLicenses []goLicense) []pkg.License {
 			Type:           l.Type,
 			URLs:           l.URLs,
 			Locations:      toPkgLocations(l.Locations),
+			Contents:       l.Contents,
 		})
 	}
 	return requireCollection(out)
@@ -471,6 +473,7 @@ func toGoLicenses(pkgLicenses []pkg.License) []goLicense {
 			Type:           l.Type,
 			URLs:           l.URLs,
 			Locations:      toGoLocations(l.Locations),
+			Contents:       l.Contents,
 		})
 	}
 	return out

syft/pkg/cataloger/golang/licenses_test.go

@@ -14,6 +14,7 @@ import (
 	"strings"
 	"testing"
 
+	"github.com/google/licensecheck"
 	"github.com/stretchr/testify/require"
 
 	"github.com/anchore/syft/internal/licenses"
@@ -70,7 +71,7 @@ func Test_LicenseSearch(t *testing.T) {
 
 	localVendorDir := filepath.Join(wd, "test-fixtures", "licenses-vendor")
 
-	licenseScanner := licenses.TestingOnlyScanner()
+	licenseScanner := licenses.NewScanner(licensecheck.Scan, float64(75))
 
 	tests := []struct {
 		name     string
@@ -295,7 +296,7 @@ func Test_findVersionPath(t *testing.T) {
 
 func Test_walkDirErrors(t *testing.T) {
 	resolver := newGoLicenseResolver("", CatalogerConfig{})
-	_, err := resolver.findLicensesInFS(context.Background(), licenses.TestingOnlyScanner(), "somewhere", badFS{})
+	_, err := resolver.findLicensesInFS(context.Background(), licenses.NewScanner(licensecheck.Scan, float64(75)), "somewhere", badFS{})
 	require.Error(t, err)
 }
 
@@ -313,8 +314,7 @@ func Test_noLocalGoModDir(t *testing.T) {
 	validTmp := t.TempDir()
 	require.NoError(t, os.MkdirAll(filepath.Join(validTmp, "mod@ver"), 0700|os.ModeDir))
 
-	licenseScanner := licenses.TestingOnlyScanner()
-
+	licenseScanner := licenses.NewScanner(licensecheck.Scan, float64(75))
 	tests := []struct {
 		name    string
 		dir     string
@@ -353,3 +353,30 @@ func Test_noLocalGoModDir(t *testing.T) {
 		})
 	}
 }
+
+func TestLicenseConversion(t *testing.T) {
+	inputLicenses := []pkg.License{
+		{
+			Value:          "Apache-2.0",
+			SPDXExpression: "Apache-2.0",
+			Type:           "concluded",
+			URLs:           nil,
+			Locations:      file.NewLocationSet(file.NewLocation("LICENSE")),
+			Contents:       "",
+		},
+		{
+			Value:          "UNKNOWN",
+			SPDXExpression: "UNKNOWN_4d1cffe420916f2b706300ab63fcafaf35226a0ad3725cb9f95b26036cefae32",
+			Type:           "declared",
+			URLs:           nil,
+			Locations:      file.NewLocationSet(file.NewLocation("LICENSE2")),
+			Contents:       "NVIDIA Software License Agreement and CUDA Supplement to Software License Agreement",
+		},
+	}
+
+	goLicenses := toGoLicenses(inputLicenses)
+
+	result := toPkgLicenses(goLicenses)
+
+	require.Equal(t, inputLicenses, result)
+}

syft/pkg/cataloger/golang/parse_go_binary_test.go

@@ -15,6 +15,7 @@ import (
 	"syscall"
 	"testing"
 
+	"github.com/google/licensecheck"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 
@@ -169,7 +170,7 @@ func TestBuildGoPkgInfo(t *testing.T) {
 		},
 	}
 
-	licenseScanner := licenses.TestingOnlyScanner()
+	licenseScanner := licenses.NewScanner(licensecheck.Scan, float64(75))
 
 	tests := []struct {
 		name          string

syft/pkg/cataloger/java/archive_parser_test.go

@@ -14,6 +14,7 @@ import (
 
 	"github.com/google/go-cmp/cmp"
 	"github.com/google/go-cmp/cmp/cmpopts"
+	"github.com/google/licensecheck"
 	"github.com/gookit/color"
 	"github.com/scylladb/go-set/strset"
 	"github.com/stretchr/testify/assert"
@@ -32,7 +33,7 @@ import (
 func TestSearchMavenForLicenses(t *testing.T) {
 	url := maventest.MockRepo(t, "internal/maven/test-fixtures/maven-repo")
 
-	ctx := licenses.SetContextLicenseScanner(context.Background(), licenses.TestingOnlyScanner())
+	ctx := licenses.SetContextLicenseScanner(context.Background(), licenses.NewScanner(licensecheck.Scan, float64(75)))
 
 	tests := []struct {
 		name             string
@@ -91,7 +92,7 @@ func TestSearchMavenForLicenses(t *testing.T) {
 }
 
 func TestParseJar(t *testing.T) {
-	ctx := licenses.SetContextLicenseScanner(context.Background(), licenses.TestingOnlyScanner())
+	ctx := licenses.SetContextLicenseScanner(context.Background(), licenses.NewScanner(licensecheck.Scan, float64(75)))
 
 	tests := []struct {
 		name         string
@@ -1374,7 +1375,7 @@ func Test_parseJavaArchive_regressions(t *testing.T) {
 }
 
 func Test_deterministicMatchingPomProperties(t *testing.T) {
-	ctx := licenses.SetContextLicenseScanner(context.Background(), licenses.TestingOnlyScanner())
+	ctx := licenses.SetContextLicenseScanner(context.Background(), licenses.NewScanner(licensecheck.Scan, float64(75)))
 
 	tests := []struct {
 		fixture  string

syft/pkg/license.go

@@ -31,6 +31,7 @@ type License struct {
 	Type           license.Type
 	URLs           []string         `hash:"ignore"`
 	Locations      file.LocationSet `hash:"ignore"`
+	Contents       string           `hash:"ignore"` // The optional binary contents of the license file
 }
 
 type Licenses []License