"strings"
	"text/scanner"

	"cmd/internal/src"
)

// A ScanToken represents an input item. It is a simple wrapping of rune, as
// returned by text/scanner.Scanner, plus a couple of extra values.
type ScanToken rune

const (
	// Asm defines some two-character lexemes. We make up
	// a rune/ScanToken value for them - ugly but simple.
	LSH          ScanToken = -1000 - iota // << Left shift.

The [RuneLen] function returns the number of 16-bit words in
the UTF-16 encoding of the rune. It returns -1 if the rune

		}
		// Handle multi-byte Unicode.
		r, n := utf8.DecodeRune(in[i:])
		out = utf8.AppendRune(out, foldRune(r))
		i += n
	}
	return out
}

// foldRune is returns the smallest rune for all runes in the same fold set.
func foldRune(r rune) rune {
	for {
		r2 := unicode.SimpleFold(r)
		if r2 <= r {
			return r2
		}
		r = r2
	}

}

// hexDecode decodes a short hex digit sequence: "10" -> 16.
func hexDecode(s []byte) rune {
	n := '\x00'
	for _, c := range s {
		n <<= 4
		switch {
		case '0' <= c && c <= '9':
			n |= rune(c - '0')
		case 'a' <= c && c <= 'f':
			n |= rune(c-'a') + 10
		case 'A' <= c && c <= 'F':
			n |= rune(c-'A') + 10
		default:
			panic(fmt.Sprintf("Bad hex digit in %q", s))
		}
	}
	return n

		sep := ""
		for i := 0; i < len(re.Rune); i += 2 {
			b.WriteString(sep)
			sep = " "
			lo, hi := re.Rune[i], re.Rune[i+1]
			if lo == hi {
				fmt.Fprintf(b, "%#x", lo)
			} else {
				fmt.Fprintf(b, "%#x-%#x", lo, hi)
			}
		}
	}
	b.WriteByte('}')
}

func mkCharClass(f func(rune) bool) string {
	re := &Regexp{Op: OpCharClass}
	lo := rune(-1)
	for i := rune(0); i <= unicode.MaxRune; i++ {

package syntax

import "testing"

var compileTests = []struct {
	Regexp string
	Prog   string
}{
	{"a", `  0	fail
  1*	rune1 "a" -> 2
  2	match
`},
	{"[A-M][n-z]", `  0	fail
  1*	rune "AM" -> 2
  2	rune "nz" -> 3
  3	match
`},
	{"", `  0	fail
  1*	nop -> 2
  2	match
`},
	{"a?", `  0	fail
  1	rune1 "a" -> 3
  2*	alt -> 1, 3
  3	match
`},
	{"a??", `  0	fail
  1	rune1 "a" -> 3

var notletterTest = []rune{
	0x20,
	0x35,
	0x375,
	0x619,
	0x700,
	0x1885,
	0xfffe,
	0x1ffff,
	0x10ffff,
}

// Contains all the special cased Latin-1 chars.
var spaceTest = []rune{
	0x09,
	0x0a,
	0x0b,
	0x0c,
	0x0d,
	0x20,
	0x85,
	0xA0,
	0x2000,
	0x3000,
}

type caseT struct {
	cas     int
	in, out rune
}

var caseTest = []caseT{
	// errors

			panic("imperfect hash")
		}
		keywordMap[h] = tok
	}
}

func lower(ch rune) rune     { return ('a' - 'A') | ch } // returns lower-case ch iff ch is ASCII letter
func isLetter(ch rune) bool  { return 'a' <= lower(ch) && lower(ch) <= 'z' || ch == '_' }
func isDecimal(ch rune) bool { return '0' <= ch && ch <= '9' }
func isHex(ch rune) bool     { return '0' <= ch && ch <= '9' || 'a' <= lower(ch) && lower(ch) <= 'f' }

}

func digitVal(ch rune) int {
	switch {
	case '0' <= ch && ch <= '9':
		return int(ch - '0')
	case 'a' <= lower(ch) && lower(ch) <= 'f':
		return int(lower(ch) - 'a' + 10)
	}
	return 16 // larger than any legal digit val
}

func lower(ch rune) rune     { return ('a' - 'A') | ch } // returns lower-case ch iff ch is ASCII letter
func isDecimal(ch rune) bool { return '0' <= ch && ch <= '9' }

	rBuffer := make([]rune, nRunes)
	buf := bytes.NewBuffer([]byte(string(rBuffer)))

	seed := time.Now().UnixNano()
	t.Logf("rand seed: %v", seed)
	rng := rand.New(rand.NewSource(seed))

	for i := 0; i < 50; i++ {
		// Generate a random string of runes.
		buf.Reset()
		for buf.Len() < sys.ExecArgLengthLimit+1 {
			var r rune
			for {
				r = rune(rng.Intn(utf8.MaxRune + 1))
				if utf8.ValidRune(r) {