utf8.go

package main

import (
	"fmt"
	"io/ioutil"
	"math/rand"
	"os"
	"time"
	"unicode/utf8"
)

func main() {
	// create a 2mb file of valid UTF-8 (worst case)
	rand.Seed(time.Now().UnixNano())
	path := "gibberish.txt"
	if err := writeFile(path, 2*1024*1024); err != nil {
		panic(err)
	}

	data, err := ioutil.ReadFile(path)

	if err != nil {
		panic(err)
	}

	start := time.Now()
	valid := utf8.Valid(data)
	duration := time.Since(start)

	defer os.Remove(path)

	fmt.Printf("UTF-8 valid: %t\n", valid)
	fmt.Printf("Time taken: %s\n", duration)
}

func writeFile(path string, size int) error {
	file, err := os.Create(path)
	if err != nil {
		return err
	}
	defer file.Close()

	// only use bytes we know are utf-8 valid
	for i := 0; i < size; i++ {
		char := byte(rand.Intn(95) + 32) // 32-126 inclusive
		if _, err := file.Write([]byte{char}); err != nil {
			return err
		}
	}

	// add one invalid sequence at the end
	if _, err := file.Write([]byte{0xF0, 0x41}); err != nil {
		return err
	}

	return nil
}

~
$ go run foo.go
UTF-8 valid: true
Time taken: 223.542µs

~
$ go run foo.go
UTF-8 valid: true
Time taken: 196.167µs

~
$ go run foo.go
UTF-8 valid: true
Time taken: 209.667µs

~/w/dgp (yarn-parser)
$ go run foo.go
UTF-8 valid: false
Time taken: 214.916µs

~/w/dgp (yarn-parser)
$ go run foo.go
UTF-8 valid: false
Time taken: 248.625µs

~/w/dgp (yarn-parser)
$ go run foo.go
UTF-8 valid: false
Time taken: 195.458µs