Add nested archive cataloging (#5260)

Signed-off-by: Kudryavcev Nikolay <kydry.nikolau@gmail.com> Signed-off-by: Thomas Bechtold <thomasbechtold@jpberlin.de> Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com> Signed-off-by: Keith Zantow <kzantow@gmail.com> Co-authored-by: Kudryavcev Nikolay <kydry.nikolau@gmail.com> Co-authored-by: Thomas Bechtold <thomasbechtold@jpberlin.de> Co-authored-by: Alex Goodman <wagoodman@users.noreply.github.com>
2026-11-01 15:06:36 +01:00 · 2026-10-29 17:33:07 -04:00
parent 14a0beb93c
commit 4b8a3db5f3
224 changed files with 15918 additions and 1281 deletions
@@ -0,0 +1,120 @@
+package task
+
+import (
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+
+	"github.com/anchore/syft/syft/artifact"
+	"github.com/anchore/syft/syft/cataloging"
+	"github.com/anchore/syft/syft/file"
+	"github.com/anchore/syft/syft/pkg"
+	"github.com/anchore/syft/syft/sbom"
+)
+
+func Test_unknownsLabeler_archiveReasonsSurvivePackages(t *testing.T) {
+	jar := file.NewCoordinates("/app.jar", "")
+	p := pkg.Package{Name: "app", Locations: file.NewLocationSet(file.NewLocationFromCoordinates(jar))}
+	p.SetID()
+	s := &sbom.SBOM{Artifacts: sbom.Artifacts{
+		Packages: pkg.NewCollection(p),
+		Unknowns: map[file.Coordinates][]string{jar: {
+			formatUnknown("archive cataloged from part of its contents: too big", ArchiveCatalogerTaskName),
+			formatUnknown("unable to read manifest", "java-archive-cataloger"),
+		}},
+	}}
+
+	unknownsLabelerTask{cataloging.UnknownsConfig{RemoveWhenPackagesDefined: true}}.finalize(file.NewMockResolverForPaths(), s)
+
+	assert.Equal(t, []string{formatUnknown("archive cataloged from part of its contents: too big", ArchiveCatalogerTaskName)}, s.Artifacts.Unknowns[jar],
+		"a package describing the archive explains a parse failure, not contents left out")
+}
+
+func Test_unknownsLabeler_archiveContentsAreReportedByFullPath(t *testing.T) {
+	// a package describing an archive does not account for the files within it
+	war := file.NewCoordinates("/app.war", "layer")
+	in := func(fullPath string) file.Coordinates {
+		return file.Coordinates{RealPath: "/app.war", FileSystemID: "layer", FullPath: fullPath}
+	}
+	nativeLib := in("/app.war!/WEB-INF/lib/jna.jar!/linux-x86-64/libjnidispatch.so")
+	unidentifiedJar := in("/app.war!/WEB-INF/lib/javax.inject-1.jar")
+	p := pkg.Package{Name: "app", Locations: file.NewLocationSet(file.NewLocationFromCoordinates(war))}
+	p.SetID()
+	noPackage := formatUnknown("no package identified in archive", "java-archive-cataloger")
+
+	s := &sbom.SBOM{
+		Artifacts: sbom.Artifacts{
+			Packages:    pkg.NewCollection(p),
+			Executables: map[file.Coordinates]file.Executable{nativeLib: {}},
+			Unknowns:    map[file.Coordinates][]string{unidentifiedJar: {noPackage}},
+		},
+		Relationships: []artifact.Relationship{{From: war, To: unidentifiedJar, Type: artifact.ContainsRelationship}},
+	}
+
+	unknownsLabelerTask{cataloging.UnknownsConfig{RemoveWhenPackagesDefined: true, IncludeExecutablesWithoutPackages: true}}.finalize(file.NewMockResolverForPaths(), s)
+
+	assert.Equal(t, map[file.Coordinates][]string{
+		nativeLib:       {formatUnknown("no package identified in executable file", unknownsLabelerTaskName)},
+		unidentifiedJar: {noPackage},
+	}, s.Artifacts.Unknowns)
+}
+
+func Test_unknownsLabeler_relativeOwnedPathsNameNothingInAnArchive(t *testing.T) {
+	// a python RECORD lists paths relative to site-packages, which resolve nowhere at the scan root either
+	whl := func(fullPath string) file.Coordinates {
+		return file.Coordinates{RealPath: "/pip.whl", FullPath: fullPath}
+	}
+	tool := whl("/pip.whl!/pip/tool")
+	p := pkg.Package{
+		Name:      "pip",
+		Locations: file.NewLocationSet(file.NewLocationFromCoordinates(whl("/pip.whl!/pip.dist-info/METADATA"))),
+		Metadata:  pkg.PythonPackage{Files: []pkg.PythonFileRecord{{Path: "pip/tool"}}},
+	}
+	p.SetID()
+	s := &sbom.SBOM{Artifacts: sbom.Artifacts{
+		Packages:    pkg.NewCollection(p),
+		Executables: map[file.Coordinates]file.Executable{tool: {}},
+	}}
+
+	unknownsLabelerTask{cataloging.UnknownsConfig{IncludeExecutablesWithoutPackages: true}}.finalize(file.NewMockResolverForPaths(), s)
+
+	assert.NotEmpty(t, s.Artifacts.Unknowns[tool])
+}
+
+func Test_unknownsLabeler_ownedFilesAreWithinTheOwnersContainer(t *testing.T) {
+	atRoot := file.NewCoordinates("/usr/bin/foo", "")
+	inArchive := file.Coordinates{RealPath: "/rootfs.tar", FullPath: "/rootfs.tar!/usr/bin/foo"}
+	owner := func(at file.Coordinates) pkg.Package {
+		p := pkg.Package{
+			Name:      "owner",
+			Type:      pkg.RpmPkg,
+			Locations: file.NewLocationSet(file.NewLocationFromCoordinates(at)),
+			Metadata:  pkg.RpmDBEntry{Files: []pkg.RpmFileRecord{{Path: "/usr/bin/foo"}}},
+		}
+		p.SetID()
+		return p
+	}
+
+	tests := []struct {
+		name       string
+		owner      file.Coordinates
+		referenced file.Coordinates
+	}{
+		{name: "an owner at the scan root", owner: file.NewCoordinates("/var/lib/rpm/Packages", ""), referenced: atRoot},
+		{name: "an owner inside an archive", owner: file.Coordinates{RealPath: "/rootfs.tar", FullPath: "/rootfs.tar!/var/lib/rpm/Packages"}, referenced: inArchive},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			s := &sbom.SBOM{Artifacts: sbom.Artifacts{
+				Packages:    pkg.NewCollection(owner(tt.owner)),
+				Executables: map[file.Coordinates]file.Executable{atRoot: {}, inArchive: {}},
+			}}
+
+			unknownsLabelerTask{cataloging.UnknownsConfig{IncludeExecutablesWithoutPackages: true}}.finalize(file.NewMockResolverForPaths("/usr/bin/foo"), s)
+
+			for _, exe := range []file.Coordinates{atRoot, inArchive} {
+				assert.Equal(t, exe != tt.referenced, len(s.Artifacts.Unknowns[exe]) > 0, exe.FullNestedPath())
+			}
+		})
+	}
+}