punycode.go

     1  // Copyright 2012 The Go Authors. All rights reserved.
     2  // Use of this source code is governed by a BSD-style
     3  // license that can be found in the LICENSE file.
     4  
     5  package cookiejar
     6  
     7  // This file implements the Punycode algorithm from RFC 3492.
     8  
     9  import (
    10  	"fmt"
    11  	"net/http/internal/ascii"
    12  	"strings"
    13  	"unicode/utf8"
    14  )
    15  
    16  // These parameter values are specified in section 5.
    17  //
    18  // All computation is done with int32s, so that overflow behavior is identical
    19  // regardless of whether int is 32-bit or 64-bit.
    20  const (
    21  	base        int32 = 36
    22  	damp        int32 = 700
    23  	initialBias int32 = 72
    24  	initialN    int32 = 128
    25  	skew        int32 = 38
    26  	tmax        int32 = 26
    27  	tmin        int32 = 1
    28  )
    29  
    30  // encode encodes a string as specified in section 6.3 and prepends prefix to
    31  // the result.
    32  //
    33  // The "while h < length(input)" line in the specification becomes "for
    34  // remaining != 0" in the Go code, because len(s) in Go is in bytes, not runes.
    35  func encode(prefix, s string) (string, error) {
    36  	output := make([]byte, len(prefix), len(prefix)+1+2*len(s))
    37  	copy(output, prefix)
    38  	delta, n, bias := int32(0), initialN, initialBias
    39  	b, remaining := int32(0), int32(0)
    40  	for _, r := range s {
    41  		if r < utf8.RuneSelf {
    42  			b++
    43  			output = append(output, byte(r))
    44  		} else {
    45  			remaining++
    46  		}
    47  	}
    48  	h := b
    49  	if b > 0 {
    50  		output = append(output, '-')
    51  	}
    52  	for remaining != 0 {
    53  		m := int32(0x7fffffff)
    54  		for _, r := range s {
    55  			if m > r && r >= n {
    56  				m = r
    57  			}
    58  		}
    59  		delta += (m - n) * (h + 1)
    60  		if delta < 0 {
    61  			return "", fmt.Errorf("cookiejar: invalid label %q", s)
    62  		}
    63  		n = m
    64  		for _, r := range s {
    65  			if r < n {
    66  				delta++
    67  				if delta < 0 {
    68  					return "", fmt.Errorf("cookiejar: invalid label %q", s)
    69  				}
    70  				continue
    71  			}
    72  			if r > n {
    73  				continue
    74  			}
    75  			q := delta
    76  			for k := base; ; k += base {
    77  				t := k - bias
    78  				if t < tmin {
    79  					t = tmin
    80  				} else if t > tmax {
    81  					t = tmax
    82  				}
    83  				if q < t {
    84  					break
    85  				}
    86  				output = append(output, encodeDigit(t+(q-t)%(base-t)))
    87  				q = (q - t) / (base - t)
    88  			}
    89  			output = append(output, encodeDigit(q))
    90  			bias = adapt(delta, h+1, h == b)
    91  			delta = 0
    92  			h++
    93  			remaining--
    94  		}
    95  		delta++
    96  		n++
    97  	}
    98  	return string(output), nil
    99  }
   100  
   101  func encodeDigit(digit int32) byte {
   102  	switch {
   103  	case 0 <= digit && digit < 26:
   104  		return byte(digit + 'a')
   105  	case 26 <= digit && digit < 36:
   106  		return byte(digit + ('0' - 26))
   107  	}
   108  	panic("cookiejar: internal error in punycode encoding")
   109  }
   110  
   111  // adapt is the bias adaptation function specified in section 6.1.
   112  func adapt(delta, numPoints int32, firstTime bool) int32 {
   113  	if firstTime {
   114  		delta /= damp
   115  	} else {
   116  		delta /= 2
   117  	}
   118  	delta += delta / numPoints
   119  	k := int32(0)
   120  	for delta > ((base-tmin)*tmax)/2 {
   121  		delta /= base - tmin
   122  		k += base
   123  	}
   124  	return k + (base-tmin+1)*delta/(delta+skew)
   125  }
   126  
   127  // Strictly speaking, the remaining code below deals with IDNA (RFC 5890 and
   128  // friends) and not Punycode (RFC 3492) per se.
   129  
   130  // acePrefix is the ASCII Compatible Encoding prefix.
   131  const acePrefix = "xn--"
   132  
   133  // toASCII converts a domain or domain label to its ASCII form. For example,
   134  // toASCII("bücher.example.com") is "xn--bcher-kva.example.com", and
   135  // toASCII("golang") is "golang".
   136  func toASCII(s string) (string, error) {
   137  	if ascii.Is(s) {
   138  		return s, nil
   139  	}
   140  	labels := strings.Split(s, ".")
   141  	for i, label := range labels {
   142  		if !ascii.Is(label) {
   143  			a, err := encode(acePrefix, label)
   144  			if err != nil {
   145  				return "", err
   146  			}
   147  			labels[i] = a
   148  		}
   149  	}
   150  	return strings.Join(labels, "."), nil
   151  }
   152
View as plain text