add VirtualPath to source.Location

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

syft/source/all_layers_resolver.go

@@ -114,8 +114,7 @@ func (r *AllLayersResolver) FilesByPath(paths ...string) ([]Location, error) {
 				return nil, err
 			}
 			for _, result := range results {
-				// we always prefer the REAL path (not the user given path which may have symlinks)
-				uniqueLocations = append(uniqueLocations, NewLocationFromImage(string(result.RealPath), result, r.img))
+				uniqueLocations = append(uniqueLocations, NewLocationFromImage(path, result, r.img))
 			}
 		}
 	}
@@ -154,8 +153,7 @@ func (r *AllLayersResolver) FilesByGlob(patterns ...string) ([]Location, error)
 					return nil, err
 				}
 				for _, refResult := range refResults {
-					// we always prefer the REAL path (not the user given path which may have symlinks)
-					uniqueLocations = append(uniqueLocations, NewLocationFromImage(string(refResult.RealPath), refResult, r.img))
+					uniqueLocations = append(uniqueLocations, NewLocationFromImage(string(result.MatchPath), refResult, r.img))
 				}
 			}
 		}

syft/source/all_layers_resolver_test.go

@@ -117,7 +117,7 @@ func TestAllLayersResolver_FilesByPath(t *testing.T) {
 					t.Errorf("bad resolve path: '%s'!='%s'", string(actual.ref.RealPath), expected.path)
 				}
 
-				if expected.path != "" && string(actual.ref.RealPath) != actual.Path {
+				if expected.path != "" && string(actual.ref.RealPath) != actual.RealPath {
 					t.Errorf("we should always prefer real paths over ones with links")
 				}
 
@@ -225,7 +225,7 @@ func TestAllLayersResolver_FilesByGlob(t *testing.T) {
 					t.Errorf("bad resolve path: '%s'!='%s'", string(actual.ref.RealPath), expected.path)
 				}
 
-				if expected.path != "" && string(actual.ref.RealPath) != actual.Path {
+				if expected.path != "" && string(actual.ref.RealPath) != actual.RealPath {
 					t.Errorf("we should always prefer real paths over ones with links")
 				}
 

syft/source/content_requester_test.go

@@ -54,20 +54,20 @@ func TestContentRequester(t *testing.T) {
 			}
 
 			for _, entry := range data {
-				if expected, ok := test.expectedContents[entry.Location.Path]; ok {
+				if expected, ok := test.expectedContents[entry.Location.RealPath]; ok {
 					actualBytes, err := ioutil.ReadAll(entry.Contents)
 					if err != nil {
-						t.Fatalf("could not read %q: %+v", entry.Location.Path, err)
+						t.Fatalf("could not read %q: %+v", entry.Location.RealPath, err)
 					}
 					for expected != string(actualBytes) {
-						t.Errorf("mismatched contents for %q", entry.Location.Path)
+						t.Errorf("mismatched contents for %q", entry.Location.RealPath)
 						dmp := diffmatchpatch.New()
 						diffs := dmp.DiffMain(expected, string(actualBytes), true)
 						t.Errorf("diff: %s", dmp.DiffPrettyText(diffs))
 					}
 					continue
 				}
-				t.Errorf("could not find %q", entry.Location.Path)
+				t.Errorf("could not find %q", entry.Location.RealPath)
 			}
 		})
 	}

syft/source/directory_resolver.go

@@ -110,7 +110,7 @@ func (r *DirectoryResolver) RelativeFileByPath(_ Location, path string) *Locatio
 func (r DirectoryResolver) MultipleFileContentsByLocation(locations []Location) (map[Location]io.ReadCloser, error) {
 	refContents := make(map[Location]io.ReadCloser)
 	for _, location := range locations {
-		refContents[location] = file.NewDeferredReadCloser(location.Path)
+		refContents[location] = file.NewDeferredReadCloser(location.RealPath)
 	}
 	return refContents, nil
 }
@@ -118,5 +118,5 @@ func (r DirectoryResolver) MultipleFileContentsByLocation(locations []Location)
 // FileContentsByLocation fetches file contents for a single file reference relative to a directory.
 // If the path does not exist an error is returned.
 func (r DirectoryResolver) FileContentsByLocation(location Location) (io.ReadCloser, error) {
-	return file.NewDeferredReadCloser(location.Path), nil
+	return file.NewDeferredReadCloser(location.RealPath), nil
 }

syft/source/directory_resolver_test.go

@@ -80,8 +80,8 @@ func TestDirectoryResolver_FilesByPath(t *testing.T) {
 			}
 
 			for _, actual := range refs {
-				if actual.Path != c.expected {
-					t.Errorf("bad resolve path: '%s'!='%s'", actual.Path, c.expected)
+				if actual.RealPath != c.expected {
+					t.Errorf("bad resolve path: '%s'!='%s'", actual.RealPath, c.expected)
 				}
 			}
 		})

syft/source/image_squash_resolver.go

@@ -66,8 +66,7 @@ func (r *ImageSquashResolver) FilesByPath(paths ...string) ([]Location, error) {
 
 		if resolvedRef != nil && !uniqueFileIDs.Contains(*resolvedRef) {
 			uniqueFileIDs.Add(*resolvedRef)
-			// we always prefer the REAL path (not the user given path which may have symlinks)
-			uniqueLocations = append(uniqueLocations, NewLocationFromImage(string(resolvedRef.RealPath), *resolvedRef, r.img))
+			uniqueLocations = append(uniqueLocations, NewLocationFromImage(path, *resolvedRef, r.img))
 		}
 	}
 

syft/source/image_squash_resolver_test.go

@@ -106,7 +106,7 @@ func TestImageSquashResolver_FilesByPath(t *testing.T) {
 				t.Errorf("bad resolve path: '%s'!='%s'", string(actual.ref.RealPath), c.resolvePath)
 			}
 
-			if c.resolvePath != "" && string(actual.ref.RealPath) != actual.Path {
+			if c.resolvePath != "" && string(actual.ref.RealPath) != actual.RealPath {
 				t.Errorf("we should always prefer real paths over ones with links")
 			}
 
@@ -212,7 +212,7 @@ func TestImageSquashResolver_FilesByGlob(t *testing.T) {
 				t.Errorf("bad resolve path: '%s'!='%s'", string(actual.ref.RealPath), c.resolvePath)
 			}
 
-			if c.resolvePath != "" && string(actual.ref.RealPath) != actual.Path {
+			if c.resolvePath != "" && string(actual.ref.RealPath) != actual.RealPath {
 				t.Errorf("we should always prefer real paths over ones with links")
 			}
 

syft/source/location.go

@@ -7,9 +7,11 @@ import (
 	"github.com/anchore/stereoscope/pkg/image"
 )
 
-// Location represents a path relative to a particular filesystem.
+// Location represents a path relative to a particular filesystem resolved to a specific file.Reference. This struct is used as a key
+// in content fetching to uniquely identify a file relative to a request (the VirtualPath).
 type Location struct {
-	Path         string         `json:"path"`              // The string path of the location (e.g. /etc/hosts)
+	RealPath     string         `json:"path"`              // The path where all path ancestors have no hardlinks / symlinks
+	VirtualPath  string         `json:"-"`                 // The path to the file which may or may not have hardlinks / symlinks
 	FileSystemID string         `json:"layerID,omitempty"` // An ID representing the filesystem. For container images this is a layer digest, directories or root filesystem this is blank.
 	ref          file.Reference // The file reference relative to the stereoscope.FileCatalog that has more information about this location.
 }
@@ -17,23 +19,25 @@ type Location struct {
 // NewLocation creates a new Location representing a path without denoting a filesystem or FileCatalog reference.
 func NewLocation(path string) Location {
 	return Location{
-		Path: path,
+		RealPath: path,
 	}
 }
 
 // NewLocationFromImage creates a new Location representing the given path (extracted from the ref) relative to the given image.
-func NewLocationFromImage(path string, ref file.Reference, img *image.Image) Location {
+func NewLocationFromImage(virtualPath string, 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: path,
-			ref:  ref,
+			VirtualPath: virtualPath,
+			RealPath:    string(ref.RealPath),
+			ref:         ref,
 		}
 	}
 
 	return Location{
-		Path:         path,
+		VirtualPath:  virtualPath,
+		RealPath:     string(ref.RealPath),
 		FileSystemID: entry.Layer.Metadata.Digest,
 		ref:          ref,
 	}

syft/source/mock_resolver.go

@@ -31,7 +31,7 @@ func NewMockResolverForPaths(paths ...string) *MockResolver {
 // HasPath indicates if the given path exists in the underlying source.
 func (r MockResolver) HasPath(path string) bool {
 	for _, l := range r.Locations {
-		if l.Path == path {
+		if l.RealPath == path {
 			return true
 		}
 	}
@@ -40,7 +40,7 @@ func (r MockResolver) HasPath(path string) bool {
 
 // String returns the string representation of the MockResolver.
 func (r MockResolver) String() string {
-	return fmt.Sprintf("mock:(%s,...)", r.Locations[0].Path)
+	return fmt.Sprintf("mock:(%s,...)", r.Locations[0].RealPath)
 }
 
 // FileContentsByLocation fetches file contents for a single location. If the
@@ -48,7 +48,7 @@ func (r MockResolver) String() string {
 func (r MockResolver) FileContentsByLocation(location Location) (io.ReadCloser, error) {
 	for _, l := range r.Locations {
 		if l == location {
-			return os.Open(location.Path)
+			return os.Open(location.RealPath)
 		}
 	}
 
@@ -74,7 +74,7 @@ func (r MockResolver) FilesByPath(paths ...string) ([]Location, error) {
 	var results []Location
 	for _, p := range paths {
 		for _, location := range r.Locations {
-			if p == location.Path {
+			if p == location.RealPath {
 				results = append(results, NewLocation(p))
 			}
 		}
@@ -88,7 +88,7 @@ func (r MockResolver) FilesByGlob(patterns ...string) ([]Location, error) {
 	var results []Location
 	for _, pattern := range patterns {
 		for _, location := range r.Locations {
-			if file.GlobMatch(pattern, location.Path) {
+			if file.GlobMatch(pattern, location.RealPath) {
 				results = append(results, location)
 			}
 		}

syft/source/resolver.go

@@ -16,8 +16,8 @@ type Resolver interface {
 // ContentResolver knows how to get file content for given file.References
 type ContentResolver interface {
 	FileContentsByLocation(Location) (io.ReadCloser, error)
+	// TODO: it is possible to be given duplicate locations that will be overridden in the map (key), a subtle problem that coule easily be misued.
 	MultipleFileContentsByLocation([]Location) (map[Location]io.ReadCloser, 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 a Location for given string paths and globs