fix(walker) clean for all temporary data after all analyzes

analyzer/analyzer.go

@@ -185,28 +185,23 @@ func (a Analyzer) ImageConfigAnalyzerVersions() map[string]int {
 }
 
 func (a Analyzer) AnalyzeFile(ctx context.Context, wg *sync.WaitGroup, limit *semaphore.Weighted, result *AnalysisResult,
-	dir, filePath string, info os.FileInfo, opener Opener) (err error) {
+	dir, filePath string, info os.FileInfo, opener Opener) (cleaner func() error, err error) {
 	if info.IsDir() {
-		return nil
+		return nil, nil
 	}
 	for _, d := range a.drivers {
 		// filepath extracted from tar file doesn't have the prefix "/"
 		if !d.Required(strings.TrimLeft(filePath, "/"), info) {
 			continue
 		}
-		rc, cleaner, err := opener()
+		rc, c, err := opener()
 		if err != nil {
-			return xerrors.Errorf("unable to open a file (%s): %w", filePath, err)
+			return nil, xerrors.Errorf("unable to open a file (%s): %w", filePath, err)
 		}
-		defer func() {
-			err := cleaner()
-			if err != nil {
-				log.Logger.Warn("Clean temp directory error: %s", err)
-			}
-		}()
+		cleaner = c
 
 		if err = limit.Acquire(ctx, 1); err != nil {
-			return xerrors.Errorf("semaphore acquire: %w", err)
+			return nil, xerrors.Errorf("semaphore acquire: %w", err)
 		}
 		wg.Add(1)
 
@@ -223,9 +218,8 @@ func (a Analyzer) AnalyzeFile(ctx context.Context, wg *sync.WaitGroup, limit *se
 			result.Merge(ret)
 		}(d, AnalysisTarget{Dir: dir, FilePath: filePath, ContentReader: rc})
 	}
-	wg.Wait()
 
-	return nil
+	return cleaner, nil
 }
 
 func (a Analyzer) AnalyzeImageConfig(targetOS types.OS, configBlob []byte) []types.Package {

artifact/image/image.go

@@ -195,23 +195,37 @@ func (a Artifact) inspectLayer(ctx context.Context, diffID string) (types.BlobIn
 		return types.BlobInfo{}, xerrors.Errorf("unable to get uncompressed layer %s: %w", diffID, err)
 	}
 
+	var cleaners []func() error
 	var wg sync.WaitGroup
 	result := new(analyzer.AnalysisResult)
 	limit := semaphore.NewWeighted(parallel)
 
 	opqDirs, whFiles, err := a.walker.Walk(r, func(filePath string, info os.FileInfo, opener analyzer.Opener) error {
-		if err = a.analyzer.AnalyzeFile(ctx, &wg, limit, result, "", filePath, info, opener); err != nil {
+		cleaner, err := a.analyzer.AnalyzeFile(ctx, &wg, limit, result, "", filePath, info, opener)
+		if err != nil {
 			return xerrors.Errorf("failed to analyze %s: %w", filePath, err)
 		}
+		if cleaner != nil {
+			cleaners = append(cleaners, cleaner)
+		}
+
 		return nil
 	})
 	if err != nil {
 		return types.BlobInfo{}, xerrors.Errorf("walk error: %w", err)
+
 	}
 
 	// Wait for all the goroutine to finish.
 	wg.Wait()
 
+	// Clean for all temporary data after all analyzers.
+	for _, c := range cleaners {
+		if err := c(); err != nil {
+			log.Logger.Warn("Clean temp directory error: %s", err)
+		}
+	}
+
 	// Sort the analysis result for consistent results
 	result.Sort()
 

artifact/local/fs.go

@@ -20,6 +20,7 @@ import (
 	"github.com/aquasecurity/fanal/cache"
 	"github.com/aquasecurity/fanal/config/scanner"
 	"github.com/aquasecurity/fanal/hook"
+	"github.com/aquasecurity/fanal/log"
 	"github.com/aquasecurity/fanal/types"
 	"github.com/aquasecurity/fanal/walker"
 )
@@ -77,6 +78,7 @@ func buildAbsPaths(base string, paths []string) []string {
 }
 
 func (a Artifact) Inspect(ctx context.Context) (types.ArtifactReference, error) {
+	var cleaners []func() error
 	var wg sync.WaitGroup
 	result := new(analyzer.AnalysisResult)
 	limit := semaphore.NewWeighted(parallel)
@@ -87,9 +89,14 @@ func (a Artifact) Inspect(ctx context.Context) (types.ArtifactReference, error)
 		if err != nil {
 			return xerrors.Errorf("filepath rel (%s): %w", filePath, err)
 		}
-		if err = a.analyzer.AnalyzeFile(ctx, &wg, limit, result, a.dir, filePath, info, opener); err != nil {
+		cleaner, err := a.analyzer.AnalyzeFile(ctx, &wg, limit, result, a.dir, filePath, info, opener)
+		if err != nil {
 			return xerrors.Errorf("analyze file (%s): %w", filePath, err)
 		}
+		if cleaner != nil {
+			cleaners = append(cleaners, cleaner)
+		}
+
 		return nil
 	})
 	if err != nil {
@@ -99,6 +106,13 @@ func (a Artifact) Inspect(ctx context.Context) (types.ArtifactReference, error)
 	// Wait for all the goroutine to finish.
 	wg.Wait()
 
+	// Clean for all temporary data after all analyzers.
+	for _, c := range cleaners {
+		if err := c(); err != nil {
+			log.Logger.Warn("Clean temp directory error: %s", err)
+		}
+	}
+
 	// Sort the analysis result for consistent results
 	result.Sort()
 

walker/walk.go

@@ -14,7 +14,7 @@ var (
 	systemDirs = []string{"proc", "sys", "dev"}
 )
 
-const ThresholdSize = int64(200) << 20
+const ThresholdSize = int64(200) << 20 // 200MB
 
 type WalkFunc func(filePath string, info os.FileInfo, opener analyzer.Opener) error