perf(depinject): optimize isExportedType function

[0x2d3c]

• avoid redundant rune slice creation in unicode check

[aljo242]

@0x2d3c what is the performance impact of this?

Why is this faster and can you show benchmarks verifying that there is improvement?

[0x2d3c]

what is the performance impact of this?

[]rune(name) will initialize a slice of the same length as name, resulting in additional memory allocation. The implementation of utf8.DecodeRuneInString just meets the actual scenario here, avoiding additional memory overhead

[0x2d3c]

Why is this faster and can you show benchmarks verifying that there is improvement?

short string bench output:

BenchmarkUsingRuneSlice-8 25518284 46.3 ns/op 48 B/op 1 allocs/op
BenchmarkUsingDecodeRune-8 86976563 13.6 ns/op 0 B/op 0 allocs/op

long string bench output:

BenchmarkUsingRuneSlice-8 12938479 92.8 ns/op 96 B/op 1 allocs/op
BenchmarkUsingDecodeRune-8 86932154 13.7 ns/op 0 B/op 0 allocs/op

[technicallyty]

what is the performance impact of this?

[]rune(name) will initialize a slice of the same length as name, resulting in additional memory allocation. The implementation of utf8.DecodeRuneInString just meets the actual scenario here, avoiding additional memory overhead

wouldn't it be even faster to just do:

if r := name[0]; unicode.IsLower(rune(r))?

[0x2d3c]

what is the performance impact of this?

[]rune(name) will initialize a slice of the same length as name, resulting in additional memory allocation. The implementation of utf8.DecodeRuneInString just meets the actual scenario here, avoiding additional memory overhead

wouldn't it be even faster to just do:

if r := name[0]; unicode.IsLower(rune(r))?

does not support non-ASCII characters (multi-byte UTF-8 characters will be truncated)

[technicallyty]

Why is this faster and can you show benchmarks verifying that there is improvement?

short string bench output:

BenchmarkUsingRuneSlice-8 25518284 46.3 ns/op 48 B/op 1 allocs/op BenchmarkUsingDecodeRune-8 86976563 13.6 ns/op 0 B/op 0 allocs/op

long string bench output:

BenchmarkUsingRuneSlice-8 12938479 92.8 ns/op 96 B/op 1 allocs/op BenchmarkUsingDecodeRune-8 86932154 13.7 ns/op 0 B/op 0 allocs/op

can you provide your benchmarking code?

[0x2d3c]

Why is this faster and can you show benchmarks verifying that there is improvement?

short string bench output:
BenchmarkUsingRuneSlice-8 25518284 46.3 ns/op 48 B/op 1 allocs/op BenchmarkUsingDecodeRune-8 86976563 13.6 ns/op 0 B/op 0 allocs/op
long string bench output:
BenchmarkUsingRuneSlice-8 12938479 92.8 ns/op 96 B/op 1 allocs/op BenchmarkUsingDecodeRune-8 86932154 13.7 ns/op 0 B/op 0 allocs/op

can you provide your benchmarking code?

this is the bench code for two ways to implement local verification and comparison. I don't know if it has any reference value.

// short（ASCII）
var shortASCII = "hi"

// short（multi-byte UTF-8）
var shortUTF8 = "你好"

// long（ASCII）
var longASCII = "abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ"

// long (multi-byte UTF-8)
var longUTF8 = "你好，世界！🚀🔥"

func BenchmarkRuneShortASCII(b *testing.B) {
	for i := 0; i < b.N; i++ {
		r, _ := utf8.DecodeRuneInString(shortASCII)
		_ = unicode.IsLower(r)
	}
}

func BenchmarkRuneSliceShortASCII(b *testing.B) {
	for i := 0; i < b.N; i++ {
		unicode.IsLower([]rune(shortASCII)[0])
	}
}

func BenchmarkRuneShortUTF8(b *testing.B) {
	for i := 0; i < b.N; i++ {
		r, _ := utf8.DecodeRuneInString(shortUTF8)
		_ = unicode.IsLower(r)
	}
}

func BenchmarkRuneSliceShortUTF8(b *testing.B) {
	for i := 0; i < b.N; i++ {
		unicode.IsLower([]rune(shortUTF8)[0])
	}
}

func BenchmarkRuneLongASCII(b *testing.B) {
	for i := 0; i < b.N; i++ {
		r, _ := utf8.DecodeRuneInString(longASCII)
		_ = unicode.IsLower(r)
	}
}

func BenchmarkRuneSliceLongASCII(b *testing.B) {
	for i := 0; i < b.N; i++ {
		unicode.IsLower([]rune(longASCII)[0])
	}
}

func BenchmarkRuneLongUTF8(b *testing.B) {
	for i := 0; i < b.N; i++ {
		r, _ := utf8.DecodeRuneInString(longUTF8)
		_ = unicode.IsLower(r)
	}
}

func BenchmarkRuneSliceLongUTF8(b *testing.B) {
	for i := 0; i < b.N; i++ {
		unicode.IsLower([]rune(longUTF8)[0])
	}
}