Merge pull request #108011 from cici37/benchmark

Benchmark cel.UnstructuredToVal

staging/src/k8s.io/apiextensions-apiserver/pkg/apiserver/schema/cel/values_test.go

@@ -619,3 +619,47 @@ func TestMapper(t *testing.T) {
 		})
 	}
 }
+
+func BenchmarkUnstructuredToVal(b *testing.B) {
+	b.ReportAllocs()
+	b.ResetTimer()
+
+	for n := 0; n < b.N; n++ {
+		if val := UnstructuredToVal([]interface{}{
+			map[string]interface{}{
+				"key": "a",
+				"val": 1,
+			},
+			map[string]interface{}{
+				"key": "b",
+				"val": 2,
+			},
+			map[string]interface{}{
+				"key": "@b",
+				"val": 2,
+			},
+		}, &mapListSchema); val == nil {
+			b.Fatal(val)
+		}
+	}
+}
+
+func BenchmarkUnstructuredToValWithEscape(b *testing.B) {
+	b.ReportAllocs()
+	b.ResetTimer()
+
+	for n := 0; n < b.N; n++ {
+		if val := UnstructuredToVal([]interface{}{
+			map[string]interface{}{
+				"key": "a.1",
+				"val": "__i.1",
+			},
+			map[string]interface{}{
+				"key": "b.1",
+				"val": 2,
+			},
+		}, &mapListSchema); val == nil {
+			b.Fatal(val)
+		}
+	}
+}

staging/src/k8s.io/apiextensions-apiserver/third_party/forked/celopenapi/model/escaping.go

@@ -36,6 +36,57 @@ var celReservedSymbols = sets.NewString(
 // expandMatcher matches the escape sequence, characters that are escaped, and characters that are unsupported
 var expandMatcher = regexp.MustCompile(`(__|[-./]|[^a-zA-Z0-9-./_])`)
 
+// newCharacterFilter returns a boolean array to indicate the allowed characters
+func newCharacterFilter(characters string) []bool {
+	maxChar := 0
+	for _, c := range characters {
+		if maxChar < int(c) {
+			maxChar = int(c)
+		}
+	}
+	filter := make([]bool, maxChar+1)
+
+	for _, c := range characters {
+		filter[int(c)] = true
+	}
+
+	return filter
+}
+
+type escapeCheck struct {
+	canSkipRegex     bool
+	invalidCharFound bool
+}
+
+// skipRegexCheck checks if escape would be skipped.
+// if invalidCharFound is true, it must have invalid character; if invalidCharFound is false, not sure if it has invalid character or not
+func skipRegexCheck(ident string) escapeCheck {
+	escapeCheck := escapeCheck{canSkipRegex: true, invalidCharFound: false}
+	// skip escape if possible
+	previous_underscore := false
+	for _, c := range ident {
+		if c == '/' || c == '-' || c == '.' {
+			escapeCheck.canSkipRegex = false
+			return escapeCheck
+		}
+		intc := int(c)
+		if intc < 0 || intc >= len(validCharacterFilter) || !validCharacterFilter[intc] {
+			escapeCheck.invalidCharFound = true
+			return escapeCheck
+		}
+		if c == '_' && previous_underscore {
+			escapeCheck.canSkipRegex = false
+			return escapeCheck
+		}
+
+		previous_underscore = c == '_'
+	}
+	return escapeCheck
+}
+
+// validCharacterFilter indicates the allowed characters.
+var validCharacterFilter = newCharacterFilter("abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789_")
+
 // Escape escapes ident and returns a CEL identifier (of the form '[a-zA-Z_][a-zA-Z0-9_]*'), or returns
 // false if the ident does not match the supported input format of `[a-zA-Z_.-/][a-zA-Z0-9_.-/]*`.
 // Escaping Rules:
@@ -53,6 +104,15 @@ func Escape(ident string) (string, bool) {
 	if celReservedSymbols.Has(ident) {
 		return "__" + ident + "__", true
 	}
+
+	escapeCheck := skipRegexCheck(ident)
+	if escapeCheck.invalidCharFound {
+		return "", false
+	}
+	if escapeCheck.canSkipRegex {
+		return ident, true
+	}
+
 	ok := true
 	ident = expandMatcher.ReplaceAllStringFunc(ident, func(s string) string {
 		switch s {

staging/src/k8s.io/apiextensions-apiserver/third_party/forked/celopenapi/model/escaping_test.go

@@ -170,3 +170,37 @@ func TestEscapingFuzz(t *testing.T) {
 }
 
 var validCelIdent = regexp.MustCompile(`^[a-zA-Z_][a-zA-Z0-9_]*$`)
+
+func TestCanSkipRegex(t *testing.T) {
+	cases := []struct {
+		unescaped        string
+		canSkip          bool
+		invalidCharFound bool
+	}{
+		{unescaped: "a.a", canSkip: false},
+		{unescaped: "a-a", canSkip: false},
+		{unescaped: "a__a", canSkip: false},
+		{unescaped: "a.-/__a", canSkip: false},
+		{unescaped: "a_a", canSkip: true},
+		{unescaped: "a_a_a", canSkip: true},
+		{unescaped: "@", invalidCharFound: true},
+		{unescaped: "👑", invalidCharFound: true},
+		// if escape keyword is detected before invalid character, invalidCharFound
+		{unescaped: "/👑", canSkip: false},
+	}
+
+	for _, tc := range cases {
+		t.Run(tc.unescaped, func(t *testing.T) {
+			escapeCheck := skipRegexCheck(tc.unescaped)
+			if escapeCheck.invalidCharFound {
+				if escapeCheck.invalidCharFound != tc.invalidCharFound {
+					t.Errorf("Expected input validation to be %v, but got %t", tc.invalidCharFound, escapeCheck.invalidCharFound)
+				}
+			} else {
+				if escapeCheck.canSkipRegex != tc.canSkip {
+					t.Errorf("Expected %v, but got %t", tc.canSkip, escapeCheck.canSkipRegex)
+				}
+			}
+		})
+	}
+}