future-architect
diff --git a/‎.gitignore‎
Lines changed: 2 additions & 0 deletions b/‎.gitignore‎
Lines changed: 2 additions & 0 deletions
diff --git a/‎GNUmakefile‎
Lines changed: 13 additions & 1 deletion b/‎GNUmakefile‎
Lines changed: 13 additions & 1 deletion
diff --git a/‎go.mod‎
Lines changed: 1 addition & 1 deletion b/‎go.mod‎
Lines changed: 1 addition & 1 deletion
diff --git a/‎scanner/analyze_golden_test.go‎
Lines changed: 317 additions & 0 deletions b/‎scanner/analyze_golden_test.go‎
Lines changed: 317 additions & 0 deletions
@@ -23,3 +23,5 @@ vuls
 /trivy-to-vuls
 snmp2cpe
 !snmp2cpe/
+/scripts/compare-lockfile
+/compare-lockfile
@@ -10,7 +10,8 @@
 	pretest \
 	test \
 	cov \
-	clean
+	clean \
+	compare-lockfile
 
 SRCS = $(shell git ls-files '*.go')
 PKGS = $(shell go list ./...)
@@ -241,6 +242,17 @@ define sed-d
 	find ${ONE_SEC_AFTER_JSON_DIR} -type f -exec sed -i -e '/scannedRevision/d' {} \;
 endef
 
+# Compare AnalyzeLibrary output between current branch and BASE ref.
+# Fetches real-world lockfiles from popular OSS projects and compares results.
+# Usage:
+#   make compare-lockfile              # fetch fixtures and compare against master
+#   make compare-lockfile BASE=commit  # compare against specific ref
+#   make compare-lockfile FETCH=0      # re-run with cached fixtures (skip download)
+BASE ?= master
+FETCH ?= 1
+compare-lockfile:
+	$(GO) run scripts/compare-lockfile.go $(if $(filter 1,$(FETCH)),-fetch) -base $(BASE)
+
 define count-cve
 	for jsonfile in ${NOW_JSON_DIR}/*.json ;  do \
 		echo $$jsonfile; cat $$jsonfile | jq ".scannedCves | length" ; \
 
@@ -59,7 +59,6 @@ require (
 	github.com/vulsio/gost v0.7.2
 	go.etcd.io/bbolt v1.4.3
 	golang.org/x/oauth2 v0.35.0
-	golang.org/x/sync v0.20.0
 	golang.org/x/term v0.40.0
 	golang.org/x/text v0.34.0
 	golang.org/x/xerrors v0.0.0-20240903120638-7835f813f4da
@@ -347,6 +346,7 @@ require (
 	golang.org/x/exp v0.0.0-20251219203646-944ab1f22d93 // indirect
 	golang.org/x/mod v0.33.0 // indirect
 	golang.org/x/net v0.51.0 // indirect
+	golang.org/x/sync v0.20.0 // indirect
 	golang.org/x/sys v0.41.0 // indirect
 	golang.org/x/time v0.14.0 // indirect
 	golang.org/x/tools v0.42.0 // indirect
 
@@ -0,0 +1,317 @@
+package scanner
+
+import (
+	"cmp"
+	"context"
+	"encoding/json"
+	"flag"
+	"os"
+	"path/filepath"
+	"slices"
+	"strings"
+	"testing"
+
+	"github.com/future-architect/vuls/models"
+)
+
+var update = flag.Bool("update", false, "update golden files")
+
+// lockfileEntry defines a test fixture for AnalyzeLibrary golden testing.
+type lockfileEntry struct {
+	// path is the relative path from the fixtures directory.
+	path string
+	// filemode to pass to AnalyzeLibrary (0755 for executables, 0644 otherwise).
+	filemode os.FileMode
+	// binary indicates the fixture is a binary file only available in the
+	// integration submodule (not copied to testdata/fixtures/).
+	binary bool
+	// expectParseError indicates this fixture is known to produce a parse error
+	// (e.g. unsupported lockfile version). The test treats errors as empty result.
+	expectParseError bool
+}
+
+var lockfiles = []lockfileEntry{
+	// Node.js
+	{"npm-v1/package-lock.json", 0644, false, false},
+	{"npm-v2/package-lock.json", 0644, false, false},
+	{"npm-v3/package-lock.json", 0644, false, false},
+	{"yarn.lock", 0644, false, false},
+	{"pnpm/pnpm-lock.yaml", 0644, false, true}, // pnpm v8: known parse error
+	{"pnpm-v9/pnpm-lock.yaml", 0644, false, false},
+	{"bun.lock", 0644, false, false},
+
+	// Python
+	{"requirements.txt", 0644, false, false},
+	{"Pipfile.lock", 0644, false, false},
+	{"poetry-v1/poetry.lock", 0644, false, false},
+	{"poetry-v2/poetry.lock", 0644, false, false},
+	{"uv.lock", 0644, false, false},
+
+	// Ruby
+	{"Gemfile.lock", 0644, false, false},
+
+	// Rust
+	{"Cargo.lock", 0644, false, false},
+	{"hello-rust", 0755, true, false},
+
+	// PHP
+	{"composer.lock", 0644, false, false},
+	{"installed.json", 0644, false, false},
+
+	// Go
+	{"go.mod", 0644, false, false},
+	{"go.sum", 0644, false, false},
+	{"gobinary", 0755, true, false},
+
+	// Java
+	{"pom.xml", 0644, false, false},
+	{"gradle.lockfile", 0644, false, false},
+	{"log4j-core-2.13.0.jar", 0644, true, false},
+	{"wrong-name-log4j-core.jar", 0644, true, false},
+	{"juddiv3-war-3.3.5.war", 0644, true, false},
+
+	// .NET
+	{"packages.lock.json", 0644, false, false},
+	{"packages.config", 0644, false, false},
+	{"datacollector.deps.json", 0644, false, false},
+	{"Directory.Packages.props", 0644, false, false},
+
+	// C/C++
+	{"conan-v1/conan.lock", 0644, false, false},
+	{"conan-v2/conan.lock", 0644, false, false},
+
+	// Dart
+	{"pubspec.lock", 0644, false, false},
+
+	// Elixir
+	{"mix.lock", 0644, false, false},
+
+	// Swift
+	{"Podfile.lock", 0644, false, false},
+	{"Package.resolved", 0644, false, false},
+}
+
+// goldenFileName converts a lockfile path to a golden file name.
+// e.g. "npm-v3/package-lock.json" -> "npm-v3_package-lock.json"
+// Uses filepath.ToSlash to normalize path separators across platforms.
+func goldenFileName(lockfilePath string) string {
+	return strings.ReplaceAll(filepath.ToSlash(lockfilePath), "/", "_") + ".json"
+}
+
+func TestAnalyzeLibrary_Golden(t *testing.T) {
+	fixturesDir := filepath.Join("testdata", "fixtures")
+	integrationDir := filepath.Join("..", "integration", "data", "lockfile")
+	goldenDir := filepath.Join("testdata", "golden")
+
+	for _, lf := range lockfiles {
+		t.Run(lf.path, func(t *testing.T) {
+			// Test fixtures are in testdata/fixtures/ (committed to repo).
+			// Binary fixtures (JAR, WAR, Go/Rust binaries) are only in the
+			// integration submodule — skip if not available.
+			// NOTE: We intentionally do NOT add submodules: true to CI checkout.
+			// Attack scenario: an attacker forks this repo, edits .gitmodules to
+			// replace the integration submodule URL with their own repo containing
+			// a malicious go.mod or _test.go, then opens a PR. If CI checks out
+			// submodules, `go test` executes attacker-controlled code with access
+			// to the CI environment (secrets, GITHUB_TOKEN, network).
+			// Binary fixture tests therefore run locally only.
+			srcPath := filepath.Join(fixturesDir, lf.path)
+			if lf.binary {
+				srcPath = filepath.Join(integrationDir, lf.path)
+			}
+			contents, err := os.ReadFile(srcPath)
+			if err != nil {
+				if lf.binary {
+					t.Skipf("Binary fixture not found: %s (requires: git submodule update --init)", srcPath)
+				}
+				t.Fatalf("Failed to read %s: %v", srcPath, err)
+			}
+
+			got, err := AnalyzeLibrary(context.Background(), lf.path, contents, lf.filemode, true)
+			if err != nil {
+				if lf.expectParseError {
+					// Verify the error is actually a parse error (contains "parse error" or the parser type)
+					errMsg := err.Error()
+					if !strings.Contains(errMsg, "parse error") && !strings.Contains(errMsg, "Failed to parse") {
+						t.Fatalf("AnalyzeLibrary(%s) expected parse error but got: %v", lf.path, err)
+					}
+					t.Logf("AnalyzeLibrary(%s) returned expected parse error: %v", lf.path, err)
+					got = nil
+				} else {
+					t.Fatalf("AnalyzeLibrary(%s) unexpected error: %v", lf.path, err)
+				}
+			}
+
+			gotJSON, err := json.MarshalIndent(normalizeResult(got), "", "  ")
+			if err != nil {
+				t.Fatalf("Failed to marshal result: %v", err)
+			}
+
+			goldenPath := filepath.Join(goldenDir, goldenFileName(lf.path))
+
+			if *update {
+				if err := os.MkdirAll(goldenDir, 0755); err != nil {
+					t.Fatalf("Failed to create golden dir: %v", err)
+				}
+				if err := os.WriteFile(goldenPath, gotJSON, 0644); err != nil {
+					t.Fatalf("Failed to write golden file: %v", err)
+				}
+				t.Logf("Updated golden file: %s", goldenPath)
+				return
+			}
+
+			wantJSON, err := os.ReadFile(goldenPath)
+			if err != nil {
+				t.Fatalf("Golden file not found: %s (run with -update to generate)", goldenPath)
+			}
+
+			if string(gotJSON) != string(wantJSON) {
+				t.Errorf("AnalyzeLibrary(%s) output differs from golden file.\nGot:\n%s\nWant:\n%s",
+					lf.path, string(gotJSON), string(wantJSON))
+			}
+		})
+	}
+}
+
+// TestAnalyzeLibrary_PomOnline verifies that pom.xml parsing in online mode
+// (resolving transitive dependencies from Maven Central) works correctly.
+// Skipped with -short since it requires network access.
+func TestAnalyzeLibrary_PomOnline(t *testing.T) {
+	if testing.Short() {
+		t.Skip("skipping online pom.xml test (requires network access)")
+	}
+
+	fixturesDir := filepath.Join("testdata", "fixtures")
+	goldenDir := filepath.Join("testdata", "golden")
+
+	contents, err := os.ReadFile(filepath.Join(fixturesDir, "pom.xml"))
+	if err != nil {
+		t.Fatalf("Failed to read pom.xml: %v", err)
+	}
+
+	got, err := AnalyzeLibrary(context.Background(), "pom.xml", contents, 0644, false)
+	if err != nil {
+		t.Fatalf("AnalyzeLibrary(pom.xml, online) unexpected error: %v", err)
+	}
+
+	gotJSON, err := json.MarshalIndent(normalizeResult(got), "", "  ")
+	if err != nil {
+		t.Fatalf("Failed to marshal result: %v", err)
+	}
+
+	goldenPath := filepath.Join(goldenDir, "pom.xml.online.json")
+
+	if *update {
+		if err := os.MkdirAll(goldenDir, 0755); err != nil {
+			t.Fatalf("Failed to create golden dir: %v", err)
+		}
+		if err := os.WriteFile(goldenPath, gotJSON, 0644); err != nil {
+			t.Fatalf("Failed to write golden file: %v", err)
+		}
+		t.Logf("Updated golden file: %s", goldenPath)
+		return
+	}
+
+	wantJSON, err := os.ReadFile(goldenPath)
+	if err != nil {
+		t.Fatalf("Golden file not found: %s (run with -update to generate)", goldenPath)
+	}
+
+	if string(gotJSON) != string(wantJSON) {
+		t.Errorf("AnalyzeLibrary(pom.xml, online) output differs from golden file.\nGot:\n%s\nWant:\n%s",
+			string(gotJSON), string(wantJSON))
+	}
+
+	// Online mode should resolve transitive dependencies, producing more results than offline.
+	offlineGoldenPath := filepath.Join(goldenDir, "pom.xml.json")
+	offlineJSON, err := os.ReadFile(offlineGoldenPath)
+	if err != nil {
+		t.Logf("Offline golden file not found, skipping comparison: %s", offlineGoldenPath)
+		return
+	}
+
+	var onlineRes []goldenLibraryScanner
+	if err := json.Unmarshal(gotJSON, &onlineRes); err != nil {
+		t.Fatalf("Failed to unmarshal online JSON result: %v", err)
+	}
+	var offlineRes []goldenLibraryScanner
+	if err := json.Unmarshal(offlineJSON, &offlineRes); err != nil {
+		t.Fatalf("Failed to unmarshal offline golden JSON: %v", err)
+	}
+	var onlineLibs, offlineLibs int
+	for _, s := range onlineRes {
+		onlineLibs += len(s.Libs)
+	}
+	for _, s := range offlineRes {
+		offlineLibs += len(s.Libs)
+	}
+	if onlineLibs <= offlineLibs {
+		t.Errorf("Online mode should resolve more dependencies than offline mode.\nOnline libs: %d\nOffline libs: %d",
+			onlineLibs, offlineLibs)
+	}
+}
+
+// normalizeResult produces a stable, comparable representation of the scan result.
+// It sorts libraries by name+version to avoid ordering-dependent diffs.
+type goldenLibraryScanner struct {
+	Type         string          `json:"type"`
+	LockfilePath string          `json:"lockfilePath"`
+	Libs         []goldenLibrary `json:"libs"`
+}
+
+type goldenLibrary struct {
+	Name     string `json:"name"`
+	Version  string `json:"version"`
+	PURL     string `json:"purl,omitempty"`
+	FilePath string `json:"filePath,omitempty"`
+	Digest   string `json:"digest,omitempty"`
+	Dev      bool   `json:"dev,omitempty"`
+}
+
+func normalizeResult(scanners []models.LibraryScanner) []goldenLibraryScanner {
+	result := make([]goldenLibraryScanner, 0, len(scanners))
+	for _, s := range scanners {
+		gs := goldenLibraryScanner{
+			Type:         string(s.Type),
+			LockfilePath: s.LockfilePath,
+			Libs:         make([]goldenLibrary, 0, len(s.Libs)),
+		}
+		for _, lib := range s.Libs {
+			gs.Libs = append(gs.Libs, goldenLibrary{
+				Name:     lib.Name,
+				Version:  lib.Version,
+				PURL:     lib.PURL,
+				FilePath: lib.FilePath,
+				Digest:   lib.Digest,
+				Dev:      lib.Dev,
+			})
+		}
+		slices.SortFunc(gs.Libs, func(a, b goldenLibrary) int {
+			return cmp.Or(
+				cmp.Compare(a.Name, b.Name),
+				cmp.Compare(a.Version, b.Version),
+				cmp.Compare(a.PURL, b.PURL),
+				cmp.Compare(a.FilePath, b.FilePath),
+				cmp.Compare(a.Digest, b.Digest),
+				func() int {
+					switch {
+					case !a.Dev && b.Dev:
+						return -1
+					case a.Dev && !b.Dev:
+						return +1
+					default:
+						return 0
+					}
+				}(),
+			)
+		})
+		result = append(result, gs)
+	}
+	slices.SortFunc(result, func(a, b goldenLibraryScanner) int {
+		return cmp.Or(
+			cmp.Compare(a.Type, b.Type),
+			cmp.Compare(a.LockfilePath, b.LockfilePath),
+		)
+	})
+	return result
+}