reflect/protoreflect: optimize Name.IsValid and FullName.IsValid
For simplicity, IsValid was implemented in terms of regular expressions,
which are useful for verifying a string according to a strict grammar,
but performs poorly. Implement the check in terms of hand-written code,
which provides a 20x improvement to validate the name "google.protobuf.Any".
name old time/op new time/op delta
FullNameIsValid-8 683ns ± 2% 35ns ± 1% -94.86% (p=0.000 n=10+10)
Change-Id: I980403befca0b72cea22acd274064a46cb02644b
Reviewed-on: https://go-review.googlesource.com/c/protobuf/+/238002
Reviewed-by: Damien Neil <dneil@google.com>
diff --git a/reflect/protoreflect/proto.go b/reflect/protoreflect/proto.go
index b669a4e..dd85915 100644
--- a/reflect/protoreflect/proto.go
+++ b/reflect/protoreflect/proto.go
@@ -128,7 +128,6 @@
import (
"fmt"
- "regexp"
"strings"
"google.golang.org/protobuf/encoding/protowire"
@@ -408,19 +407,14 @@
doNotImplement
}
-var (
- regexName = regexp.MustCompile(`^[_a-zA-Z][_a-zA-Z0-9]*$`)
- regexFullName = regexp.MustCompile(`^[_a-zA-Z][_a-zA-Z0-9]*(\.[_a-zA-Z][_a-zA-Z0-9]*)*$`)
-)
-
// Name is the short name for a proto declaration. This is not the name
// as used in Go source code, which might not be identical to the proto name.
type Name string // e.g., "Kind"
-// IsValid reports whether n is a syntactically valid name.
+// IsValid reports whether s is a syntactically valid name.
// An empty name is invalid.
-func (n Name) IsValid() bool {
- return regexName.MatchString(string(n))
+func (s Name) IsValid() bool {
+ return consumeIdent(string(s)) == len(s)
}
// Names represent a list of names.
@@ -443,10 +437,42 @@
// This should not have any leading or trailing dots.
type FullName string // e.g., "google.protobuf.Field.Kind"
-// IsValid reports whether n is a syntactically valid full name.
+// IsValid reports whether s is a syntactically valid full name.
// An empty full name is invalid.
-func (n FullName) IsValid() bool {
- return regexFullName.MatchString(string(n))
+func (s FullName) IsValid() bool {
+ i := consumeIdent(string(s))
+ if i < 0 {
+ return false
+ }
+ for len(s) > i {
+ if s[i] != '.' {
+ return false
+ }
+ i++
+ n := consumeIdent(string(s[i:]))
+ if n < 0 {
+ return false
+ }
+ i += n
+ }
+ return true
+}
+
+func consumeIdent(s string) (i int) {
+ if len(s) == 0 || !isLetter(s[i]) {
+ return -1
+ }
+ i++
+ for len(s) > i && isLetterDigit(s[i]) {
+ i++
+ }
+ return i
+}
+func isLetter(c byte) bool {
+ return c == '_' || ('a' <= c && c <= 'z') || ('A' <= c && c <= 'Z')
+}
+func isLetterDigit(c byte) bool {
+ return isLetter(c) || ('0' <= c && c <= '9')
}
// Name returns the short name, which is the last identifier segment.
diff --git a/reflect/protoreflect/proto_test.go b/reflect/protoreflect/proto_test.go
index a62a004..97e55cc 100644
--- a/reflect/protoreflect/proto_test.go
+++ b/reflect/protoreflect/proto_test.go
@@ -72,3 +72,11 @@
}
}
}
+
+var sink bool
+
+func BenchmarkFullNameIsValid(b *testing.B) {
+ for i := 0; i < b.N; i++ {
+ sink = FullName("google.protobuf.Any").IsValid()
+ }
+}
