[SumDB] Improved MapFn performance 10x

vindex/cmd/sumdbindex/main.go

@@ -29,7 +29,7 @@ import (
 	"os"
 	"os/signal"
 	"path"
-	"regexp"
+
 	"strings"
 	"syscall"
 	"time"
@@ -54,15 +54,6 @@ var (
 	listen                   = flag.String("listen", ":8088", "Address to set up HTTP server listening on")
 )
 
-var (
-	// Example leaf:
-	// golang.org/x/text v0.3.0 h1:g61tztE5qeGQ89tm6NTjjM9VPIm088od1l6aSorWRWg=
-	// golang.org/x/text v0.3.0/go.mod h1:NqM8EUOU14njkJ3fqMW+pc6Ldnwhi/IjpwHt7yyuwOQ=
-	//
-	line0RE = regexp.MustCompile(`(.*) (.*) h1:(.*)`)
-	line1RE = regexp.MustCompile(`(.*) (.*)/go.mod h1:(.*)`)
-)
-
 func main() {
 	klog.InitFlags(nil)
 	flag.Parse()
@@ -233,24 +224,16 @@ func mapFn(data []byte) [][32]byte {
 		panic(fmt.Errorf("expected 2 lines but got %d", len(lines)))
 	}
 
-	line0Parts := line0RE.FindStringSubmatch(lines[0])
-	line0Module, line0Version := line0Parts[1], line0Parts[2]
-
-	line1Parts := line1RE.FindStringSubmatch(lines[1])
-	line1Module, line1Version := line1Parts[1], line1Parts[2]
-
-	if line0Module != line1Module {
-		klog.Errorf("mismatched module names: (%s, %s)", line0Module, line1Module)
-	}
-	if line0Version != line1Version {
-		klog.Errorf("mismatched version names: (%s, %s)", line0Version, line0Version)
+	line0Parts := strings.Fields(lines[0])
+	if len(line0Parts) < 2 {
+		panic(fmt.Errorf("expected at least 2 parts in line 0 but got %d", len(line0Parts)))
 	}
+	line0Module, line0Version := line0Parts[0], line0Parts[1]
+
 	if module.IsPseudoVersion(line0Version) {
 		// Drop any emphemeral builds
 		return nil
 	}
 
-	klog.V(2).Infof("MapFn found: Module: %s:\t%s", line0Module, line0Version)
-
 	return [][32]byte{sha256.Sum256([]byte(line0Module))}
 }

vindex/cmd/sumdbindex/main_test.go

@@ -0,0 +1,52 @@
+package main
+
+import (
+	"bytes"
+	"crypto/sha256"
+	"testing"
+)
+
+var exampleLeaf = []byte(`golang.org/x/text v0.3.0 h1:g61tztE5qeGQ89tm6NTjjM9VPIm088od1l6aSorWRWg=
+golang.org/x/text v0.3.0/go.mod h1:NqM8EUOU14njkJ3fqMW+pc6Ldnwhi/IjpwHt7yyuwOQ=
+`)
+
+var pseudoVersionLeaf = []byte(`github.com/transparency-dev/tessera v0.0.0-20240222160914-411202e8d356 h1:4jV/qA6RzP7Z6s+/vQ0W2RjM3FjC6B2M3r8=
+github.com/transparency-dev/tessera v0.0.0-20240222160914-411202e8d356/go.mod h1:T/Ym+5H1e28Qv6iMzT3w=
+`)
+
+func TestMapFn(t *testing.T) {
+	for _, tc := range []struct {
+		name string
+		leaf []byte
+		want [][32]byte
+	}{
+		{
+			name: "valid",
+			leaf: exampleLeaf,
+			want: [][32]byte{sha256.Sum256([]byte("golang.org/x/text"))},
+		},
+		{
+			name: "pseudo_version",
+			leaf: pseudoVersionLeaf,
+			want: nil,
+		},
+	} {
+		t.Run(tc.name, func(t *testing.T) {
+			got := mapFn(tc.leaf)
+			if len(got) != len(tc.want) {
+				t.Fatalf("mapFn() returned %d keys, want %d", len(got), len(tc.want))
+			}
+			for i := range got {
+				if !bytes.Equal(got[i][:], tc.want[i][:]) {
+					t.Errorf("mapFn()[%d] = %x, want %x", i, got[i], tc.want[i])
+				}
+			}
+		})
+	}
+}
+
+func BenchmarkMapFn(b *testing.B) {
+	for b.Loop() {
+		mapFn(exampleLeaf)
+	}
+}