add source.Location + reorient Resolvers to use it

Signed-off-by: Alex Goodman <alex.goodman@anchore.com>

syft/source/all_layers_resolver.go

@@ -8,6 +8,8 @@ import (
 	"github.com/anchore/stereoscope/pkg/image"
 )
 
+var _ Resolver = &AllLayersResolver{}
+
 // AllLayersResolver implements path and content access for the AllLayers source option for container image data sources.
 type AllLayersResolver struct {
 	img    *image.Image
@@ -61,14 +63,14 @@ func (r *AllLayersResolver) fileByRef(ref file.Reference, uniqueFileIDs file.Ref
 }
 
 // FilesByPath returns all file.References that match the given paths from any layer in the image.
-func (r *AllLayersResolver) FilesByPath(paths ...file.Path) ([]file.Reference, error) {
+func (r *AllLayersResolver) FilesByPath(paths ...string) ([]Location, error) {
 	uniqueFileIDs := file.NewFileReferenceSet()
-	uniqueFiles := make([]file.Reference, 0)
+	uniqueLocations := make([]Location, 0)
 
 	for _, path := range paths {
 		for idx, layerIdx := range r.layers {
 			tree := r.img.Layers[layerIdx].Tree
-			ref := tree.File(path)
+			ref := tree.File(file.Path(path))
 			if ref == nil {
 				// no file found, keep looking through layers
 				continue
@@ -91,17 +93,18 @@ func (r *AllLayersResolver) FilesByPath(paths ...file.Path) ([]file.Reference, e
 			if err != nil {
 				return nil, err
 			}
-			uniqueFiles = append(uniqueFiles, results...)
+			for _, result := range results {
+				uniqueLocations = append(uniqueLocations, newLocationFromImage(result, r.img))
 			}
 		}
-
-	return uniqueFiles, nil
+	}
+	return uniqueLocations, nil
 }
 
 // FilesByGlob returns all file.References that match the given path glob pattern from any layer in the image.
-func (r *AllLayersResolver) FilesByGlob(patterns ...string) ([]file.Reference, error) {
+func (r *AllLayersResolver) FilesByGlob(patterns ...string) ([]Location, error) {
 	uniqueFileIDs := file.NewFileReferenceSet()
-	uniqueFiles := make([]file.Reference, 0)
+	uniqueLocations := make([]Location, 0)
 
 	for _, pattern := range patterns {
 		for idx, layerIdx := range r.layers {
@@ -128,31 +131,63 @@ func (r *AllLayersResolver) FilesByGlob(patterns ...string) ([]file.Reference, e
 				if err != nil {
 					return nil, err
 				}
-				uniqueFiles = append(uniqueFiles, results...)
+				for _, result := range results {
+					uniqueLocations = append(uniqueLocations, newLocationFromImage(result, r.img))
+				}
 			}
 		}
 	}
 
-	return uniqueFiles, nil
+	return uniqueLocations, nil
 }
 
-func (r *AllLayersResolver) RelativeFileByPath(reference file.Reference, path string) (*file.Reference, error) {
-	entry, err := r.img.FileCatalog.Get(reference)
+func (r *AllLayersResolver) RelativeFileByPath(location Location, path string) *Location {
+	entry, err := r.img.FileCatalog.Get(location.ref)
 	if err != nil {
-		return nil, err
+		return nil
 	}
 
-	return entry.Source.SquashedTree.File(file.Path(path)), nil
+	relativeRef := entry.Source.SquashedTree.File(file.Path(path))
+	if relativeRef == nil {
+		return nil
+	}
+
+	relativeLocation := newLocationFromImage(*relativeRef, r.img)
+
+	return &relativeLocation
 }
 
 // MultipleFileContentsByRef returns the file contents for all file.References relative to the image. Note that a
 // file.Reference is a path relative to a particular layer.
-func (r *AllLayersResolver) MultipleFileContentsByRef(f ...file.Reference) (map[file.Reference]string, error) {
-	return r.img.MultipleFileContentsByRef(f...)
+func (r *AllLayersResolver) MultipleFileContentsByRef(locations []Location) (map[Location]string, error) {
+	return mapLocationRefs(r.img.MultipleFileContentsByRef, locations)
 }
 
 // FileContentsByRef fetches file contents for a single file reference, irregardless of the source layer.
 // If the path does not exist an error is returned.
-func (r *AllLayersResolver) FileContentsByRef(ref file.Reference) (string, error) {
-	return r.img.FileContentsByRef(ref)
+func (r *AllLayersResolver) FileContentsByRef(location Location) (string, error) {
+	return r.img.FileContentsByRef(location.ref)
+}
+
+type multiContentFetcher func(refs ...file.Reference) (map[file.Reference]string, error)
+
+func mapLocationRefs(callback multiContentFetcher, locations []Location) (map[Location]string, error) {
+	var fileRefs = make([]file.Reference, len(locations))
+	var locationByRefs = make(map[file.Reference]Location)
+	var results = make(map[Location]string)
+
+	for i, location := range locations {
+		locationByRefs[location.ref] = location
+		fileRefs[i] = location.ref
+	}
+
+	contentsByRef, err := callback(fileRefs...)
+	if err != nil {
+		return nil, err
+	}
+
+	for ref, content := range contentsByRef {
+		results[locationByRefs[ref]] = content
+	}
+	return results, nil
 }

syft/source/directory_resolver.go

@@ -7,11 +7,14 @@ import (
 	"path"
 	"path/filepath"
 
-	"github.com/anchore/stereoscope/pkg/file"
+	"github.com/docker/distribution/reference"
+
 	"github.com/anchore/syft/internal/log"
 	"github.com/bmatcuk/doublestar"
 )
 
+var _ Resolver = &DirectoryResolver{}
+
 // DirectoryResolver implements path and content access for the directory data source.
 type DirectoryResolver struct {
 	Path string
@@ -23,11 +26,11 @@ func (s DirectoryResolver) String() string {
 }
 
 // FilesByPath returns all file.References that match the given paths from the directory.
-func (s DirectoryResolver) FilesByPath(userPaths ...file.Path) ([]file.Reference, error) {
-	var references = make([]file.Reference, 0)
+func (s DirectoryResolver) FilesByPath(userPaths ...string) ([]Location, error) {
+	var references = make([]Location, 0)
 
 	for _, userPath := range userPaths {
-		userStrPath := string(userPath)
+		userStrPath := userPath
 
 		if filepath.IsAbs(userStrPath) {
 			// a path relative to root should be prefixed with the resolvers directory path, otherwise it should be left as is
@@ -45,33 +48,24 @@ func (s DirectoryResolver) FilesByPath(userPaths ...file.Path) ([]file.Reference
 			continue
 		}
 
-		references = append(references, file.NewFileReference(file.Path(userStrPath)))
+		references = append(references, newLocation(userStrPath))
 	}
 
 	return references, nil
 }
 
-func fileContents(path file.Path) ([]byte, error) {
-	contents, err := ioutil.ReadFile(string(path))
-
-	if err != nil {
-		return nil, err
-	}
-	return contents, nil
-}
-
 // FilesByGlob returns all file.References that match the given path glob pattern from any layer in the image.
-func (s DirectoryResolver) FilesByGlob(patterns ...string) ([]file.Reference, error) {
-	result := make([]file.Reference, 0)
+func (s DirectoryResolver) FilesByGlob(patterns ...string) ([]Location, error) {
+	result := make([]Location, 0)
 
 	for _, pattern := range patterns {
 		pathPattern := path.Join(s.Path, pattern)
-		matches, err := doublestar.Glob(pathPattern)
+		pathMatches, err := doublestar.Glob(pathPattern)
 		if err != nil {
-			return result, err
+			return nil, err
 		}
-		for _, match := range matches {
-			fileMeta, err := os.Stat(match)
+		for _, matchedPath := range pathMatches {
+			fileMeta, err := os.Stat(matchedPath)
 			if err != nil {
 				continue
 			}
@@ -81,47 +75,55 @@ func (s DirectoryResolver) FilesByGlob(patterns ...string) ([]file.Reference, er
 				continue
 			}
 
-			matchedPath := file.Path(match)
-			result = append(result, file.NewFileReference(matchedPath))
+			result = append(result, newLocation(matchedPath))
 		}
 	}
 
 	return result, nil
 }
 
-func (s *DirectoryResolver) RelativeFileByPath(_ file.Reference, path string) (*file.Reference, error) {
-	paths, err := s.FilesByPath(file.Path(path))
+func (s *DirectoryResolver) RelativeFileByPath(_ Location, path string) *Location {
+	paths, err := s.FilesByPath(path)
 	if err != nil {
-		return nil, err
+		return nil
 	}
 	if len(paths) == 0 {
-		return nil, nil
+		return nil
 	}
 
-	return &paths[0], nil
+	return &paths[0]
 }
 
 // MultipleFileContentsByRef returns the file contents for all file.References relative a directory.
-func (s DirectoryResolver) MultipleFileContentsByRef(f ...file.Reference) (map[file.Reference]string, error) {
-	refContents := make(map[file.Reference]string)
-	for _, fileRef := range f {
-		contents, err := fileContents(fileRef.Path)
+func (s DirectoryResolver) MultipleFileContentsByRef(locations []Location) (map[Location]string, error) {
+	refContents := make(map[Location]string)
+	for _, location := range locations {
+		contents, err := fileContents(location.Path)
 
 		if err != nil {
-			return nil, fmt.Errorf("could not read contents of file: %s", fileRef.Path)
+			return nil, fmt.Errorf("could not read contents of file: %s", location.Path)
 		}
-		refContents[fileRef] = string(contents)
+		refContents[location] = string(contents)
 	}
 	return refContents, nil
 }
 
 // FileContentsByRef fetches file contents for a single file reference relative to a directory.
 // If the path does not exist an error is returned.
-func (s DirectoryResolver) FileContentsByRef(reference file.Reference) (string, error) {
-	contents, err := fileContents(reference.Path)
+func (s DirectoryResolver) FileContentsByRef(location Location) (string, error) {
+	contents, err := fileContents(location.Path)
 	if err != nil {
 		return "", fmt.Errorf("could not read contents of file: %s", reference.Path)
 	}
 
 	return string(contents), nil
 }
+
+func fileContents(path string) ([]byte, error) {
+	contents, err := ioutil.ReadFile(path)
+
+	if err != nil {
+		return nil, err
+	}
+	return contents, nil
+}

syft/source/image_squash_resolver.go

@@ -7,6 +7,8 @@ import (
 	"github.com/anchore/stereoscope/pkg/image"
 )
 
+var _ Resolver = &ImageSquashResolver{}
+
 // ImageSquashResolver implements path and content access for the Squashed source option for container image data sources.
 type ImageSquashResolver struct {
 	img *image.Image
@@ -21,13 +23,13 @@ func NewImageSquashResolver(img *image.Image) (*ImageSquashResolver, error) {
 }
 
 // FilesByPath returns all file.References that match the given paths within the squashed representation of the image.
-func (r *ImageSquashResolver) FilesByPath(paths ...file.Path) ([]file.Reference, error) {
+func (r *ImageSquashResolver) FilesByPath(paths ...string) ([]Location, error) {
 	uniqueFileIDs := file.NewFileReferenceSet()
-	uniqueFiles := make([]file.Reference, 0)
+	uniqueLocations := make([]Location, 0)
 
 	for _, path := range paths {
 		tree := r.img.SquashedTree()
-		ref := tree.File(path)
+		ref := tree.File(file.Path(path))
 		if ref == nil {
 			// no file found, keep looking through layers
 			continue
@@ -54,17 +56,17 @@ func (r *ImageSquashResolver) FilesByPath(paths ...file.Path) ([]file.Reference,
 
 		if resolvedRef != nil && !uniqueFileIDs.Contains(*resolvedRef) {
 			uniqueFileIDs.Add(*resolvedRef)
-			uniqueFiles = append(uniqueFiles, *resolvedRef)
+			uniqueLocations = append(uniqueLocations, newLocationFromImage(*resolvedRef, r.img))
 		}
 	}
 
-	return uniqueFiles, nil
+	return uniqueLocations, nil
 }
 
 // FilesByGlob returns all file.References that match the given path glob pattern within the squashed representation of the image.
-func (r *ImageSquashResolver) FilesByGlob(patterns ...string) ([]file.Reference, error) {
+func (r *ImageSquashResolver) FilesByGlob(patterns ...string) ([]Location, error) {
 	uniqueFileIDs := file.NewFileReferenceSet()
-	uniqueFiles := make([]file.Reference, 0)
+	uniqueLocations := make([]Location, 0)
 
 	for _, pattern := range patterns {
 		refs, err := r.img.SquashedTree().FilesByGlob(pattern)
@@ -86,42 +88,42 @@ func (r *ImageSquashResolver) FilesByGlob(patterns ...string) ([]file.Reference,
 				}
 			}
 
-			resolvedRefs, err := r.FilesByPath(ref.Path)
+			resolvedLocations, err := r.FilesByPath(string(ref.Path))
 			if err != nil {
 				return nil, fmt.Errorf("failed to find files by path (ref=%+v): %w", ref, err)
 			}
-			for _, resolvedRef := range resolvedRefs {
-				if !uniqueFileIDs.Contains(resolvedRef) {
-					uniqueFileIDs.Add(resolvedRef)
-					uniqueFiles = append(uniqueFiles, resolvedRef)
+			for _, resolvedLocation := range resolvedLocations {
+				if !uniqueFileIDs.Contains(resolvedLocation.ref) {
+					uniqueFileIDs.Add(resolvedLocation.ref)
+					uniqueLocations = append(uniqueLocations, resolvedLocation)
 				}
 			}
 		}
 	}
 
-	return uniqueFiles, nil
+	return uniqueLocations, nil
 }
 
-func (r *ImageSquashResolver) RelativeFileByPath(reference file.Reference, path string) (*file.Reference, error) {
-	paths, err := r.FilesByPath(file.Path(path))
+func (r *ImageSquashResolver) RelativeFileByPath(_ Location, path string) *Location {
+	paths, err := r.FilesByPath(path)
 	if err != nil {
-		return nil, err
+		return nil
 	}
 	if len(paths) == 0 {
-		return nil, nil
+		return nil
 	}
 
-	return &paths[0], nil
+	return &paths[0]
 }
 
 // MultipleFileContentsByRef returns the file contents for all file.References relative to the image. Note that a
 // file.Reference is a path relative to a particular layer, in this case only from the squashed representation.
-func (r *ImageSquashResolver) MultipleFileContentsByRef(f ...file.Reference) (map[file.Reference]string, error) {
-	return r.img.MultipleFileContentsByRef(f...)
+func (r *ImageSquashResolver) MultipleFileContentsByRef(locations []Location) (map[Location]string, error) {
+	return mapLocationRefs(r.img.MultipleFileContentsByRef, locations)
 }
 
 // FileContentsByRef fetches file contents for a single file reference, irregardless of the source layer.
 // If the path does not exist an error is returned.
-func (r *ImageSquashResolver) FileContentsByRef(ref file.Reference) (string, error) {
-	return r.img.FileContentsByRef(ref)
+func (r *ImageSquashResolver) FileContentsByRef(location Location) (string, error) {
+	return r.img.FileContentsByRef(location.ref)
 }

syft/source/location.go

@@ -0,0 +1,46 @@
+package source
+
+import (
+	"github.com/anchore/syft/internal/log"
+
+	"github.com/anchore/stereoscope/pkg/file"
+	"github.com/anchore/stereoscope/pkg/image"
+)
+
+type Location struct {
+	Path       string `json:"path"`
+	LayerIndex uint   `json:"layerIndex"`
+	LayerID    string `json:"layerID"`
+	ref        file.Reference
+}
+
+func newLocation(path string) Location {
+	return Location{
+		Path: path,
+	}
+}
+
+func newLocationFromRef(ref file.Reference) Location {
+	return Location{
+		Path: string(ref.Path),
+		ref:  ref,
+	}
+}
+
+func newLocationFromImage(ref file.Reference, img *image.Image) Location {
+	entry, err := img.FileCatalog.Get(ref)
+	if err != nil {
+		log.Warnf("unable to find file catalog entry for ref=%+v", ref)
+		return Location{
+			Path: string(ref.Path),
+			ref:  ref,
+		}
+	}
+
+	return Location{
+		Path:       string(ref.Path),
+		LayerIndex: entry.Source.Metadata.Index,
+		LayerID:    entry.Source.Metadata.Digest,
+		ref:        ref,
+	}
+}

syft/source/resolver.go

@@ -3,7 +3,6 @@ package source
 import (
 	"fmt"
 
-	"github.com/anchore/stereoscope/pkg/file"
 	"github.com/anchore/stereoscope/pkg/image"
 )
 
@@ -15,30 +14,30 @@ type Resolver interface {
 
 // ContentResolver knows how to get file content for given file.References
 type ContentResolver interface {
-	FileContentsByRef(ref file.Reference) (string, error)
-	MultipleFileContentsByRef(f ...file.Reference) (map[file.Reference]string, error)
+	FileContentsByRef(Location) (string, error)
+	MultipleFileContentsByRef([]Location) (map[Location]string, error)
 	// TODO: we should consider refactoring to return a set of io.Readers or file.Openers instead of the full contents themselves (allow for optional buffering).
 }
 
 //  FileResolver knows how to get file.References for given string paths and globs
 type FileResolver interface {
 	// FilesByPath fetches a set of file references which have the given path (for an image, there may be multiple matches)
-	FilesByPath(paths ...file.Path) ([]file.Reference, error)
+	FilesByPath(paths ...string) ([]Location, error)
 	// FilesByGlob fetches a set of file references which the given glob matches
-	FilesByGlob(patterns ...string) ([]file.Reference, error)
+	FilesByGlob(patterns ...string) ([]Location, error)
 	// RelativeFileByPath fetches a single file at the given path relative to the layer squash of the given reference.
 	// This is helpful when attempting to find a file that is in the same layer or lower as another file.
-	RelativeFileByPath(reference file.Reference, path string) (*file.Reference, error)
+	RelativeFileByPath(_ Location, path string) *Location
 }
 
 // getImageResolver returns the appropriate resolve for a container image given the source option
-func getImageResolver(img *image.Image, option Scope) (Resolver, error) {
-	switch option {
+func getImageResolver(img *image.Image, scope Scope) (Resolver, error) {
+	switch scope {
 	case SquashedScope:
 		return NewImageSquashResolver(img)
 	case AllLayersScope:
 		return NewAllLayersResolver(img)
 	default:
-		return nil, fmt.Errorf("bad option provided: %+v", option)
+		return nil, fmt.Errorf("bad scope provided: %+v", scope)
 	}
 }