fix: handle individual cataloger panics (anchore#1636)

cmd/syft/cli/eventloop/tasks.go

@@ -46,14 +46,11 @@ func generateCatalogPackagesTask(app *config.Application) (Task, error) {
 
 	task := func(results *sbom.Artifacts, src *source.Source) ([]artifact.Relationship, error) {
 		packageCatalog, relationships, theDistro, err := syft.CatalogPackages(src, app.ToCatalogerConfig())
-		if err != nil {
-			return nil, err
-		}
 
 		results.PackageCatalog = packageCatalog
 		results.LinuxDistribution = theDistro
 
-		return relationships, nil
+		return relationships, err
 	}
 
 	return task, nil

syft/lib.go

@@ -70,13 +70,10 @@ func CatalogPackages(src *source.Source, cfg cataloger.Config) (*pkg.Catalog, []
 	}
 
 	catalog, relationships, err := cataloger.Catalog(resolver, release, cfg.Parallelism, catalogers...)
-	if err != nil {
-		return nil, nil, nil, err
-	}
 
 	relationships = append(relationships, newSourceRelationshipsFromCatalog(src, catalog)...)
 
-	return catalog, relationships, release, nil
+	return catalog, relationships, release, err
 }
 
 func newSourceRelationshipsFromCatalog(src *source.Source, c *pkg.Catalog) []artifact.Relationship {

syft/pkg/cataloger/catalog.go

@@ -3,6 +3,7 @@ package cataloger
 import (
 	"fmt"
 	"math"
+	"runtime/debug"
 	"sync"
 
 	"github.com/hashicorp/go-multierror"
@@ -49,8 +50,15 @@ func newMonitor() (*progress.Manual, *progress.Manual) {
 	return &filesProcessed, &packagesDiscovered
 }
 
-func runCataloger(cataloger pkg.Cataloger, resolver source.FileResolver) (*catalogResult, error) {
-	catalogerResult := new(catalogResult)
+func runCataloger(cataloger pkg.Cataloger, resolver source.FileResolver) (catalogerResult *catalogResult, err error) {
+	// handle individual cataloger panics
+	defer func() {
+		if e := recover(); e != nil {
+			err = fmt.Errorf("%v at:\n%s", e, string(debug.Stack()))
+		}
+	}()
+
+	catalogerResult = new(catalogResult)
 
 	// find packages from the underlying raw data
 	log.WithFields("cataloger", cataloger.Name()).Trace("cataloging started")
@@ -88,7 +96,7 @@ func runCataloger(cataloger pkg.Cataloger, resolver source.FileResolver) (*catal
 	}
 	catalogerResult.Relationships = append(catalogerResult.Relationships, relationships...)
 	log.WithFields("cataloger", cataloger.Name()).Trace("cataloging complete")
-	return catalogerResult, nil
+	return catalogerResult, err
 }
 
 // Catalog a given source (container image or filesystem) with the given catalogers, returning all discovered packages.
@@ -100,7 +108,11 @@ func runCataloger(cataloger pkg.Cataloger, resolver source.FileResolver) (*catal
 func Catalog(resolver source.FileResolver, release *linux.Release, parallelism int, catalogers ...pkg.Cataloger) (*pkg.Catalog, []artifact.Relationship, error) {
 	catalog := pkg.NewCatalog()
 	var allRelationships []artifact.Relationship
+
 	filesProcessed, packagesDiscovered := newMonitor()
+	defer filesProcessed.SetCompleted()
+	defer packagesDiscovered.SetCompleted()
+
 	// perform analysis, accumulating errors for each failed analysis
 	var errs error
 
@@ -158,7 +170,6 @@ func Catalog(resolver source.FileResolver, release *linux.Release, parallelism i
 	for result := range results {
 		if result.Error != nil {
 			errs = multierror.Append(errs, result.Error)
-			continue
 		}
 		for _, p := range result.Packages {
 			catalog.Add(p)
@@ -168,14 +179,7 @@ func Catalog(resolver source.FileResolver, release *linux.Release, parallelism i
 
 	allRelationships = append(allRelationships, pkg.NewRelationships(catalog)...)
 
-	if errs != nil {
-		return nil, nil, errs
-	}
-
-	filesProcessed.SetCompleted()
-	packagesDiscovered.SetCompleted()
-
-	return catalog, allRelationships, nil
+	return catalog, allRelationships, errs
 }
 
 func packageFileOwnershipRelationships(p pkg.Package, resolver source.FilePathResolver) ([]artifact.Relationship, error) {

syft/pkg/cataloger/catalog_test.go

@@ -0,0 +1,67 @@
+package cataloger
+
+import (
+	"testing"
+
+	"github.com/stretchr/testify/require"
+
+	"github.com/anchore/syft/syft/artifact"
+	"github.com/anchore/syft/syft/linux"
+	"github.com/anchore/syft/syft/pkg"
+	"github.com/anchore/syft/syft/source"
+)
+
+func Test_CatalogPanicHandling(t *testing.T) {
+	catalog, relationships, err := Catalog(
+		source.NewMockResolverForPaths(),
+		&linux.Release{},
+		1,
+		panickingCataloger{},
+		returningCataloger{},
+	)
+
+	require.Error(t, err)
+	require.Contains(t, err.Error(), "catalog_test.go")
+	require.Len(t, catalog.Sorted(), 2)
+	require.Len(t, relationships, 1)
+}
+
+type panickingCataloger struct{}
+
+func (p panickingCataloger) Name() string {
+	return "panicking-cataloger"
+}
+
+func (p panickingCataloger) Catalog(_ source.FileResolver) ([]pkg.Package, []artifact.Relationship, error) {
+	panic("something bad happened")
+}
+
+var _ pkg.Cataloger = (*panickingCataloger)(nil)
+
+type returningCataloger struct{}
+
+func (p returningCataloger) Name() string {
+	return "returning-cataloger"
+}
+
+func (p returningCataloger) Catalog(_ source.FileResolver) ([]pkg.Package, []artifact.Relationship, error) {
+	pkg1 := pkg.Package{
+		Name:    "package-1",
+		Version: "1.0",
+	}
+	pkg1.SetID()
+	pkg2 := pkg.Package{
+		Name:    "package-2",
+		Version: "2.0",
+	}
+	pkg2.SetID()
+	return []pkg.Package{pkg1, pkg2}, []artifact.Relationship{
+		{
+			From: pkg1,
+			To:   pkg2,
+			Type: artifact.DependencyOfRelationship,
+		},
+	}, nil
+}
+
+var _ pkg.Cataloger = (*returningCataloger)(nil)