Merge pull request #1888 from stevvooe/more-efficient-path-comparison

registry/storage: more efficient path compare in catalog

registry/storage/catalog.go

@@ -38,7 +38,7 @@ func (reg *registry) Repositories(ctx context.Context, repos []string, last stri
 		_, file := path.Split(repoPath)
 		if file == "_layers" {
 			repoPath = strings.TrimSuffix(repoPath, "/_layers")
-			if pathGreaterThan(repoPath, last) {
+			if lessPath(last, repoPath) {
 				foundRepos = append(foundRepos, repoPath)
 			}
 			return ErrSkipDir
@@ -100,22 +100,53 @@ func (reg *registry) Enumerate(ctx context.Context, ingester func(string) error)
 
 }
 
-func pathGreaterThan(pathX, pathY string) (b bool) {
+// lessPath returns true if one path a is less than path b.
-	splitPathX := strings.SplitN(pathX, "/", 2)
+//
-	splitPathY := strings.SplitN(pathY, "/", 2)
+// A component-wise comparison is done, rather than the lexical comparison of
+// strings.
+func lessPath(a, b string) bool {
+	// we provide this behavior by making separator always sort first.
+	return compareReplaceInline(a, b, '/', '\x00') < 0
+}
 
-	if splitPathX[0] == splitPathY[0] {
+// compareReplaceInline modifies runtime.cmpstring to replace old with new
-		if len(splitPathX) == 1 && len(splitPathY) == 1 {
+// during a byte-wise comparison.
-			return false
+func compareReplaceInline(s1, s2 string, old, new byte) int {
-		} else if len(splitPathX) == 1 && len(splitPathY) != 1 {
+	// TODO(stevvooe): We are missing an optimization when the s1 and s2 have
-			return false
+	// the exact same slice header. It will make the code unsafe but can
-		} else if len(splitPathX) != 1 && len(splitPathY) == 1 {
+	// provide some extra performance.
-			return true
-		}
-
-		return pathGreaterThan(splitPathX[1], splitPathY[1])
 
+	l := len(s1)
+	if len(s2) < l {
+		l = len(s2)
 	}
 
-	return splitPathX[0] > splitPathY[0]
+	for i := 0; i < l; i++ {
+		c1, c2 := s1[i], s2[i]
+		if c1 == old {
+			c1 = new
+		}
+
+		if c2 == old {
+			c2 = new
+		}
+
+		if c1 < c2 {
+			return -1
+		}
+
+		if c1 > c2 {
+			return +1
+		}
+	}
+
+	if len(s1) < len(s2) {
+		return -1
+	}
+
+	if len(s1) > len(s2) {
+		return +1
+	}
+
+	return 0
 }

registry/storage/catalog_test.go

@@ -3,6 +3,7 @@ package storage
 import (
 	"fmt"
 	"io"
+	"math/rand"
 	"testing"
 
 	"github.com/docker/distribution"
@@ -220,3 +221,83 @@ func TestCatalogWalkError(t *testing.T) {
 		t.Errorf("Expected catalog driver list error")
 	}
 }
+
+func BenchmarkPathCompareEqual(B *testing.B) {
+	B.StopTimer()
+	pp := randomPath(100)
+	// make a real copy
+	ppb := append([]byte{}, []byte(pp)...)
+	a, b := pp, string(ppb)
+
+	B.StartTimer()
+	for i := 0; i < B.N; i++ {
+		lessPath(a, b)
+	}
+}
+
+func BenchmarkPathCompareNotEqual(B *testing.B) {
+	B.StopTimer()
+	a, b := randomPath(100), randomPath(100)
+	B.StartTimer()
+
+	for i := 0; i < B.N; i++ {
+		lessPath(a, b)
+	}
+}
+
+func BenchmarkPathCompareNative(B *testing.B) {
+	B.StopTimer()
+	a, b := randomPath(100), randomPath(100)
+	B.StartTimer()
+
+	for i := 0; i < B.N; i++ {
+		c := a < b
+		c = c && false
+	}
+}
+
+func BenchmarkPathCompareNativeEqual(B *testing.B) {
+	B.StopTimer()
+	pp := randomPath(100)
+	a, b := pp, pp
+	B.StartTimer()
+
+	for i := 0; i < B.N; i++ {
+		c := a < b
+		c = c && false
+	}
+}
+
+var filenameChars = []byte("abcdefghijklmnopqrstuvwxyz0123456789")
+var separatorChars = []byte("._-")
+
+func randomPath(length int64) string {
+	path := "/"
+	for int64(len(path)) < length {
+		chunkLength := rand.Int63n(length-int64(len(path))) + 1
+		chunk := randomFilename(chunkLength)
+		path += chunk
+		remaining := length - int64(len(path))
+		if remaining == 1 {
+			path += randomFilename(1)
+		} else if remaining > 1 {
+			path += "/"
+		}
+	}
+	return path
+}
+
+func randomFilename(length int64) string {
+	b := make([]byte, length)
+	wasSeparator := true
+	for i := range b {
+		if !wasSeparator && i < len(b)-1 && rand.Intn(4) == 0 {
+			b[i] = separatorChars[rand.Intn(len(separatorChars))]
+			wasSeparator = true
+		} else {
+			b[i] = filenameChars[rand.Intn(len(filenameChars))]
+			wasSeparator = false
+		}
+	}
+	return string(b)
+}