libgo/go/encoding/git85/git.go

   1 // Copyright 2009 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 git85 implements the radix 85 data encoding
   6 // used in the Git version control system.
   7 package git85
   8
   9 import (
  10         "bytes"
  11         "io"
  12         "os"
  13         "strconv"
  14 )
  15
  16 type CorruptInputError int64
  17
  18 func (e CorruptInputError) String() string {
  19         return "illegal git85 data at input byte " + strconv.Itoa64(int64(e))
  20 }
  21
  22 const encode = "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz!#$%&()*+-;<=>?@^_`{|}~"
  23
  24 // The decodings are 1+ the actual value, so that the
  25 // default zero value can be used to mean "not valid".
  26 var decode = [256]uint8{
  27         '0': 1, 2, 3, 4, 5, 6, 7, 8, 9, 10,
  28         'A': 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23,
  29         24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36,
  30         'a': 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49,
  31         50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62,
  32         '!': 63,
  33         '#': 64, 65, 66, 67,
  34         '(': 68, 69, 70, 71,
  35         '-': 72,
  36         ';': 73,
  37         '<': 74, 75, 76, 77,
  38         '@': 78,
  39         '^': 79, 80, 81,
  40         '{': 82, 83, 84, 85,
  41 }
  42
  43 // Encode encodes src into EncodedLen(len(src))
  44 // bytes of dst.  As a convenience, it returns the number
  45 // of bytes written to dst, but this value is always EncodedLen(len(src)).
  46 // Encode implements the radix 85 encoding used in the
  47 // Git version control tool.
  48 //
  49 // The encoding splits src into chunks of at most 52 bytes
  50 // and encodes each chunk on its own line.
  51 func Encode(dst, src []byte) int {
  52         ndst := 0
  53         for len(src) > 0 {
  54                 n := len(src)
  55                 if n > 52 {
  56                         n = 52
  57                 }
  58                 if n <= 27 {
  59                         dst[ndst] = byte('A' + n - 1)
  60                 } else {
  61                         dst[ndst] = byte('a' + n - 26 - 1)
  62                 }
  63                 ndst++
  64                 for i := 0; i < n; i += 4 {
  65                         var v uint32
  66                         for j := 0; j < 4 && i+j < n; j++ {
  67                                 v |= uint32(src[i+j]) << uint(24-j*8)
  68                         }
  69                         for j := 4; j >= 0; j-- {
  70                                 dst[ndst+j] = encode[v%85]
  71                                 v /= 85
  72                         }
  73                         ndst += 5
  74                 }
  75                 dst[ndst] = '\n'
  76                 ndst++
  77                 src = src[n:]
  78         }
  79         return ndst
  80 }
  81
  82 // EncodedLen returns the length of an encoding of n source bytes.
  83 func EncodedLen(n int) int {
  84         if n == 0 {
  85                 return 0
  86         }
  87         // 5 bytes per 4 bytes of input, rounded up.
  88         // 2 extra bytes for each line of 52 src bytes, rounded up.
  89         return (n+3)/4*5 + (n+51)/52*2
  90 }
  91
  92 var newline = []byte{'\n'}
  93
  94 // Decode decodes src into at most MaxDecodedLen(len(src))
  95 // bytes, returning the actual number of bytes written to dst.
  96 //
  97 // If Decode encounters invalid input, it returns a CorruptInputError.
  98 //
  99 func Decode(dst, src []byte) (n int, err os.Error) {
 100         ndst := 0
 101         nsrc := 0
 102         for nsrc < len(src) {
 103                 var l int
 104                 switch ch := int(src[nsrc]); {
 105                 case 'A' <= ch && ch <= 'Z':
 106                         l = ch - 'A' + 1
 107                 case 'a' <= ch && ch <= 'z':
 108                         l = ch - 'a' + 26 + 1
 109                 default:
 110                         return ndst, CorruptInputError(nsrc)
 111                 }
 112                 if nsrc+1+l > len(src) {
 113                         return ndst, CorruptInputError(nsrc)
 114                 }
 115                 el := (l + 3) / 4 * 5 // encoded len
 116                 if nsrc+1+el+1 > len(src) || src[nsrc+1+el] != '\n' {
 117                         return ndst, CorruptInputError(nsrc)
 118                 }
 119                 line := src[nsrc+1 : nsrc+1+el]
 120                 for i := 0; i < el; i += 5 {
 121                         var v uint32
 122                         for j := 0; j < 5; j++ {
 123                                 ch := decode[line[i+j]]
 124                                 if ch == 0 {
 125                                         return ndst, CorruptInputError(nsrc + 1 + i + j)
 126                                 }
 127                                 v = v*85 + uint32(ch-1)
 128                         }
 129                         for j := 0; j < 4; j++ {
 130                                 dst[ndst] = byte(v >> 24)
 131                                 v <<= 8
 132                                 ndst++
 133                         }
 134                 }
 135                 // Last fragment may have run too far (but there was room in dst).
 136                 // Back up.
 137                 if l%4 != 0 {
 138                         ndst -= 4 - l%4
 139                 }
 140                 nsrc += 1 + el + 1
 141         }
 142         return ndst, nil
 143 }
 144
 145 func MaxDecodedLen(n int) int { return n / 5 * 4 }
 146
 147 // NewEncoder returns a new Git base85 stream encoder.  Data written to
 148 // the returned writer will be encoded and then written to w.
 149 // The Git encoding operates on 52-byte blocks; when finished
 150 // writing, the caller must Close the returned encoder to flush any
 151 // partially written blocks.
 152 func NewEncoder(w io.Writer) io.WriteCloser { return &encoder{w: w} }
 153
 154 type encoder struct {
 155         w    io.Writer
 156         err  os.Error
 157         buf  [52]byte
 158         nbuf int
 159         out  [1024]byte
 160         nout int
 161 }
 162
 163 func (e *encoder) Write(p []byte) (n int, err os.Error) {
 164         if e.err != nil {
 165                 return 0, e.err
 166         }
 167
 168         // Leading fringe.
 169         if e.nbuf > 0 {
 170                 var i int
 171                 for i = 0; i < len(p) && e.nbuf < 52; i++ {
 172                         e.buf[e.nbuf] = p[i]
 173                         e.nbuf++
 174                 }
 175                 n += i
 176                 p = p[i:]
 177                 if e.nbuf < 52 {
 178                         return
 179                 }
 180                 nout := Encode(e.out[0:], e.buf[0:])
 181                 if _, e.err = e.w.Write(e.out[0:nout]); e.err != nil {
 182                         return n, e.err
 183                 }
 184                 e.nbuf = 0
 185         }
 186
 187         // Large interior chunks.
 188         for len(p) >= 52 {
 189                 nn := len(e.out) / (1 + 52/4*5 + 1) * 52
 190                 if nn > len(p) {
 191                         nn = len(p) / 52 * 52
 192                 }
 193                 if nn > 0 {
 194                         nout := Encode(e.out[0:], p[0:nn])
 195                         if _, e.err = e.w.Write(e.out[0:nout]); e.err != nil {
 196                                 return n, e.err
 197                         }
 198                 }
 199                 n += nn
 200                 p = p[nn:]
 201         }
 202
 203         // Trailing fringe.
 204         for i := 0; i < len(p); i++ {
 205                 e.buf[i] = p[i]
 206         }
 207         e.nbuf = len(p)
 208         n += len(p)
 209         return
 210 }
 211
 212 func (e *encoder) Close() os.Error {
 213         // If there's anything left in the buffer, flush it out
 214         if e.err == nil && e.nbuf > 0 {
 215                 nout := Encode(e.out[0:], e.buf[0:e.nbuf])
 216                 e.nbuf = 0
 217                 _, e.err = e.w.Write(e.out[0:nout])
 218         }
 219         return e.err
 220 }
 221
 222 // NewDecoder returns a new Git base85 stream decoder.
 223 func NewDecoder(r io.Reader) io.Reader { return &decoder{r: r} }
 224
 225 type decoder struct {
 226         r       io.Reader
 227         err     os.Error
 228         readErr os.Error
 229         buf     [1024]byte
 230         nbuf    int
 231         out     []byte
 232         outbuf  [1024]byte
 233         off     int64
 234 }
 235
 236 func (d *decoder) Read(p []byte) (n int, err os.Error) {
 237         if len(p) == 0 {
 238                 return 0, nil
 239         }
 240
 241         for {
 242                 // Copy leftover output from last decode.
 243                 if len(d.out) > 0 {
 244                         n = copy(p, d.out)
 245                         d.out = d.out[n:]
 246                         return
 247                 }
 248
 249                 // Out of decoded output.  Check errors.
 250                 if d.err != nil {
 251                         return 0, d.err
 252                 }
 253                 if d.readErr != nil {
 254                         d.err = d.readErr
 255                         return 0, d.err
 256                 }
 257
 258                 // Read and decode more input.
 259                 var nn int
 260                 nn, d.readErr = d.r.Read(d.buf[d.nbuf:])
 261                 d.nbuf += nn
 262
 263                 // Send complete lines to Decode.
 264                 nl := bytes.LastIndex(d.buf[0:d.nbuf], newline)
 265                 if nl < 0 {
 266                         continue
 267                 }
 268                 nn, d.err = Decode(d.outbuf[0:], d.buf[0:nl+1])
 269                 if e, ok := d.err.(CorruptInputError); ok {
 270                         d.err = CorruptInputError(int64(e) + d.off)
 271                 }
 272                 d.out = d.outbuf[0:nn]
 273                 d.nbuf = copy(d.buf[0:], d.buf[nl+1:d.nbuf])
 274                 d.off += int64(nl + 1)
 275         }
 276         panic("unreacahable")
 277 }