From e7b65c2c58dd28a0ae102c04c77b42f2461458f5 Mon Sep 17 00:00:00 2001
From: Christopher Angelo Phillips <32073428+spiffcs@users.noreply.github.com>
Date: Tue, 19 Nov 2024 10:00:59 -0500
Subject: [PATCH] 3030 license declared spdx correction (#3461)

* feat: update hasExtractedLicense field to include license-ref candidates
---------
Signed-off-by: Christopher Phillips <32073428+spiffcs@users.noreply.github.com>
---
 .../common/spdxhelpers/to_format_model.go     | 26 ++++++++++++++++---
 .../spdxhelpers/to_format_model_test.go       | 23 ++++++++++++++++
 2 files changed, 46 insertions(+), 3 deletions(-)

diff --git a/syft/format/common/spdxhelpers/to_format_model.go b/syft/format/common/spdxhelpers/to_format_model.go
index 3f93883cf..643c5aa9f 100644
--- a/syft/format/common/spdxhelpers/to_format_model.go
+++ b/syft/format/common/spdxhelpers/to_format_model.go
@@ -5,6 +5,7 @@ import (
 	"crypto/sha1"
 	"fmt"
 	"path"
+	"regexp"
 	"slices"
 	"sort"
 	"strings"
@@ -713,8 +714,8 @@ func toFileTypes(metadata *file.Metadata) (ty []string) {
 	return ty
 }
 
-// other licenses are for licenses from the pkg.Package that do not have an SPDXExpression
-// field. The spdxexpression field is only filled given a validated Value field.
+// other licenses are for licenses from the pkg.Package that do not have a valid SPDX Expression
+// OR are an expression that is a single `License-Ref-*`
 func toOtherLicenses(catalog *pkg.Collection) []*spdx.OtherLicense {
 	licenses := map[string]helpers.SPDXLicense{}
 
@@ -724,11 +725,17 @@ func toOtherLicenses(catalog *pkg.Collection) []*spdx.OtherLicense {
 			if l.Value != "" {
 				licenses[l.ID] = l
 			}
+			if l.ID != "" && isLicenseRef(l.ID) {
+				licenses[l.ID] = l
+			}
 		}
 		for _, l := range concludedLicenses {
 			if l.Value != "" {
 				licenses[l.ID] = l
 			}
+			if l.ID != "" && isLicenseRef(l.ID) {
+				licenses[l.ID] = l
+			}
 		}
 	}
 
@@ -742,14 +749,27 @@ func toOtherLicenses(catalog *pkg.Collection) []*spdx.OtherLicense {
 	slices.Sort(ids)
 	for _, id := range ids {
 		license := licenses[id]
+		value := license.Value
+		// handle cases where LicenseRef needs to be included in hasExtractedLicensingInfos
+		if license.Value == "" {
+			value, _ = strings.CutPrefix(license.ID, "LicenseRef-")
+		}
 		result = append(result, &spdx.OtherLicense{
 			LicenseIdentifier: license.ID,
-			ExtractedText:     license.Value,
+			ExtractedText:     value,
 		})
 	}
 	return result
 }
 
+var licenseRefRegEx = regexp.MustCompile(`^LicenseRef-[A-Za-z0-9_-]+$`)
+
+// isSingularLicenseRef checks if the string is a singular LicenseRef-* identifier
+func isLicenseRef(s string) bool {
+	// Match the input string against the regex
+	return licenseRefRegEx.MatchString(s)
+}
+
 // TODO: handle SPDX excludes file case
 // f file is an "excludes" file, skip it /* exclude SPDX analysis file(s) */
 // see: https://spdx.github.io/spdx-spec/v2.3/package-information/#79-package-verification-code-field
diff --git a/syft/format/common/spdxhelpers/to_format_model_test.go b/syft/format/common/spdxhelpers/to_format_model_test.go
index 934fc95dd..36201bc85 100644
--- a/syft/format/common/spdxhelpers/to_format_model_test.go
+++ b/syft/format/common/spdxhelpers/to_format_model_test.go
@@ -753,6 +753,29 @@ func Test_OtherLicenses(t *testing.T) {
 				},
 			},
 		},
+		{
+			name: "LicenseRef as a valid spdx expression",
+			pkg: pkg.Package{
+				Licenses: pkg.NewLicenseSet(
+					pkg.NewLicense("LicenseRef-Fedora-Public-Domain"),
+				),
+			},
+			expected: []*spdx.OtherLicense{
+				{
+					LicenseIdentifier: "LicenseRef-Fedora-Public-Domain",
+					ExtractedText:     "Fedora-Public-Domain",
+				},
+			},
+		},
+		{
+			name: "LicenseRef as a valid spdx expression does not otherize compound spdx expressions",
+			pkg: pkg.Package{
+				Licenses: pkg.NewLicenseSet(
+					pkg.NewLicense("(MIT AND LicenseRef-Fedora-Public-Domain)"),
+				),
+			},
+			expected: nil,
+		},
 	}
 
 	for _, test := range tests {