...
Run Format

Source file src/encoding/base64/base64.go

Documentation: encoding/base64

  // Copyright 2009 The Go Authors. All rights reserved.
  // Use of this source code is governed by a BSD-style
  // license that can be found in the LICENSE file.
  
  // Package base64 implements base64 encoding as specified by RFC 4648.
  package base64
  
  import (
  	"io"
  	"strconv"
  )
  
  /*
   * Encodings
   */
  
  // An Encoding is a radix 64 encoding/decoding scheme, defined by a
  // 64-character alphabet. The most common encoding is the "base64"
  // encoding defined in RFC 4648 and used in MIME (RFC 2045) and PEM
  // (RFC 1421).  RFC 4648 also defines an alternate encoding, which is
  // the standard encoding with - and _ substituted for + and /.
  type Encoding struct {
  	encode    [64]byte
  	decodeMap [256]byte
  	padChar   rune
  	strict    bool
  }
  
  const (
  	StdPadding rune = '=' // Standard padding character
  	NoPadding  rune = -1  // No padding
  )
  
  const encodeStd = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/"
  const encodeURL = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789-_"
  
  // NewEncoding returns a new padded Encoding defined by the given alphabet,
  // which must be a 64-byte string.
  // The resulting Encoding uses the default padding character ('='),
  // which may be changed or disabled via WithPadding.
  func NewEncoding(encoder string) *Encoding {
  	if len(encoder) != 64 {
  		panic("encoding alphabet is not 64-bytes long")
  	}
  
  	e := new(Encoding)
  	e.padChar = StdPadding
  	copy(e.encode[:], encoder)
  
  	for i := 0; i < len(e.decodeMap); i++ {
  		e.decodeMap[i] = 0xFF
  	}
  	for i := 0; i < len(encoder); i++ {
  		e.decodeMap[encoder[i]] = byte(i)
  	}
  	return e
  }
  
  // WithPadding creates a new encoding identical to enc except
  // with a specified padding character, or NoPadding to disable padding.
  func (enc Encoding) WithPadding(padding rune) *Encoding {
  	enc.padChar = padding
  	return &enc
  }
  
  // Strict creates a new encoding identical to enc except with
  // strict decoding enabled. In this mode, the decoder requires that
  // trailing padding bits are zero, as described in RFC 4648 section 3.5.
  func (enc Encoding) Strict() *Encoding {
  	enc.strict = true
  	return &enc
  }
  
  // StdEncoding is the standard base64 encoding, as defined in
  // RFC 4648.
  var StdEncoding = NewEncoding(encodeStd)
  
  // URLEncoding is the alternate base64 encoding defined in RFC 4648.
  // It is typically used in URLs and file names.
  var URLEncoding = NewEncoding(encodeURL)
  
  // RawStdEncoding is the standard raw, unpadded base64 encoding,
  // as defined in RFC 4648 section 3.2.
  // This is the same as StdEncoding but omits padding characters.
  var RawStdEncoding = StdEncoding.WithPadding(NoPadding)
  
  // RawURLEncoding is the unpadded alternate base64 encoding defined in RFC 4648.
  // It is typically used in URLs and file names.
  // This is the same as URLEncoding but omits padding characters.
  var RawURLEncoding = URLEncoding.WithPadding(NoPadding)
  
  /*
   * Encoder
   */
  
  // Encode encodes src using the encoding enc, writing
  // EncodedLen(len(src)) bytes to dst.
  //
  // The encoding pads the output to a multiple of 4 bytes,
  // so Encode is not appropriate for use on individual blocks
  // of a large data stream. Use NewEncoder() instead.
  func (enc *Encoding) Encode(dst, src []byte) {
  	if len(src) == 0 {
  		return
  	}
  
  	di, si := 0, 0
  	n := (len(src) / 3) * 3
  	for si < n {
  		// Convert 3x 8bit source bytes into 4 bytes
  		val := uint(src[si+0])<<16 | uint(src[si+1])<<8 | uint(src[si+2])
  
  		dst[di+0] = enc.encode[val>>18&0x3F]
  		dst[di+1] = enc.encode[val>>12&0x3F]
  		dst[di+2] = enc.encode[val>>6&0x3F]
  		dst[di+3] = enc.encode[val&0x3F]
  
  		si += 3
  		di += 4
  	}
  
  	remain := len(src) - si
  	if remain == 0 {
  		return
  	}
  	// Add the remaining small block
  	val := uint(src[si+0]) << 16
  	if remain == 2 {
  		val |= uint(src[si+1]) << 8
  	}
  
  	dst[di+0] = enc.encode[val>>18&0x3F]
  	dst[di+1] = enc.encode[val>>12&0x3F]
  
  	switch remain {
  	case 2:
  		dst[di+2] = enc.encode[val>>6&0x3F]
  		if enc.padChar != NoPadding {
  			dst[di+3] = byte(enc.padChar)
  		}
  	case 1:
  		if enc.padChar != NoPadding {
  			dst[di+2] = byte(enc.padChar)
  			dst[di+3] = byte(enc.padChar)
  		}
  	}
  }
  
  // EncodeToString returns the base64 encoding of src.
  func (enc *Encoding) EncodeToString(src []byte) string {
  	buf := make([]byte, enc.EncodedLen(len(src)))
  	enc.Encode(buf, src)
  	return string(buf)
  }
  
  type encoder struct {
  	err  error
  	enc  *Encoding
  	w    io.Writer
  	buf  [3]byte    // buffered data waiting to be encoded
  	nbuf int        // number of bytes in buf
  	out  [1024]byte // output buffer
  }
  
  func (e *encoder) Write(p []byte) (n int, err error) {
  	if e.err != nil {
  		return 0, e.err
  	}
  
  	// Leading fringe.
  	if e.nbuf > 0 {
  		var i int
  		for i = 0; i < len(p) && e.nbuf < 3; i++ {
  			e.buf[e.nbuf] = p[i]
  			e.nbuf++
  		}
  		n += i
  		p = p[i:]
  		if e.nbuf < 3 {
  			return
  		}
  		e.enc.Encode(e.out[:], e.buf[:])
  		if _, e.err = e.w.Write(e.out[:4]); e.err != nil {
  			return n, e.err
  		}
  		e.nbuf = 0
  	}
  
  	// Large interior chunks.
  	for len(p) >= 3 {
  		nn := len(e.out) / 4 * 3
  		if nn > len(p) {
  			nn = len(p)
  			nn -= nn % 3
  		}
  		e.enc.Encode(e.out[:], p[:nn])
  		if _, e.err = e.w.Write(e.out[0 : nn/3*4]); e.err != nil {
  			return n, e.err
  		}
  		n += nn
  		p = p[nn:]
  	}
  
  	// Trailing fringe.
  	for i := 0; i < len(p); i++ {
  		e.buf[i] = p[i]
  	}
  	e.nbuf = len(p)
  	n += len(p)
  	return
  }
  
  // Close flushes any pending output from the encoder.
  // It is an error to call Write after calling Close.
  func (e *encoder) Close() error {
  	// If there's anything left in the buffer, flush it out
  	if e.err == nil && e.nbuf > 0 {
  		e.enc.Encode(e.out[:], e.buf[:e.nbuf])
  		_, e.err = e.w.Write(e.out[:e.enc.EncodedLen(e.nbuf)])
  		e.nbuf = 0
  	}
  	return e.err
  }
  
  // NewEncoder returns a new base64 stream encoder. Data written to
  // the returned writer will be encoded using enc and then written to w.
  // Base64 encodings operate in 4-byte blocks; when finished
  // writing, the caller must Close the returned encoder to flush any
  // partially written blocks.
  func NewEncoder(enc *Encoding, w io.Writer) io.WriteCloser {
  	return &encoder{enc: enc, w: w}
  }
  
  // EncodedLen returns the length in bytes of the base64 encoding
  // of an input buffer of length n.
  func (enc *Encoding) EncodedLen(n int) int {
  	if enc.padChar == NoPadding {
  		return (n*8 + 5) / 6 // minimum # chars at 6 bits per char
  	}
  	return (n + 2) / 3 * 4 // minimum # 4-char quanta, 3 bytes each
  }
  
  /*
   * Decoder
   */
  
  type CorruptInputError int64
  
  func (e CorruptInputError) Error() string {
  	return "illegal base64 data at input byte " + strconv.FormatInt(int64(e), 10)
  }
  
  // decode is like Decode but returns an additional 'end' value, which
  // indicates if end-of-message padding or a partial quantum was encountered
  // and thus any additional data is an error.
  func (enc *Encoding) decode(dst, src []byte) (n int, end bool, err error) {
  	si := 0
  
  	// skip over newlines
  	for si < len(src) && (src[si] == '\n' || src[si] == '\r') {
  		si++
  	}
  
  	for si < len(src) && !end {
  		// Decode quantum using the base64 alphabet
  		var dbuf [4]byte
  		dinc, dlen := 3, 4
  
  		for j := range dbuf {
  			if len(src) == si {
  				if enc.padChar != NoPadding || j < 2 {
  					return n, false, CorruptInputError(si - j)
  				}
  				dinc, dlen, end = j-1, j, true
  				break
  			}
  			in := src[si]
  
  			si++
  			// skip over newlines
  			for si < len(src) && (src[si] == '\n' || src[si] == '\r') {
  				si++
  			}
  
  			if rune(in) == enc.padChar {
  				// We've reached the end and there's padding
  				switch j {
  				case 0, 1:
  					// incorrect padding
  					return n, false, CorruptInputError(si - 1)
  				case 2:
  					// "==" is expected, the first "=" is already consumed.
  					if si == len(src) {
  						// not enough padding
  						return n, false, CorruptInputError(len(src))
  					}
  					if rune(src[si]) != enc.padChar {
  						// incorrect padding
  						return n, false, CorruptInputError(si - 1)
  					}
  
  					si++
  					// skip over newlines
  					for si < len(src) && (src[si] == '\n' || src[si] == '\r') {
  						si++
  					}
  				}
  				if si < len(src) {
  					// trailing garbage
  					err = CorruptInputError(si)
  				}
  				dinc, dlen, end = 3, j, true
  				break
  			}
  			dbuf[j] = enc.decodeMap[in]
  			if dbuf[j] == 0xFF {
  				return n, false, CorruptInputError(si - 1)
  			}
  		}
  
  		// Convert 4x 6bit source bytes into 3 bytes
  		val := uint(dbuf[0])<<18 | uint(dbuf[1])<<12 | uint(dbuf[2])<<6 | uint(dbuf[3])
  		dbuf[2], dbuf[1], dbuf[0] = byte(val>>0), byte(val>>8), byte(val>>16)
  		switch dlen {
  		case 4:
  			dst[2] = dbuf[2]
  			dbuf[2] = 0
  			fallthrough
  		case 3:
  			dst[1] = dbuf[1]
  			if enc.strict && dbuf[2] != 0 {
  				return n, end, CorruptInputError(si - 1)
  			}
  			dbuf[1] = 0
  			fallthrough
  		case 2:
  			dst[0] = dbuf[0]
  			if enc.strict && (dbuf[1] != 0 || dbuf[2] != 0) {
  				return n, end, CorruptInputError(si - 2)
  			}
  		}
  		dst = dst[dinc:]
  		n += dlen - 1
  	}
  
  	return n, end, err
  }
  
  // Decode decodes src using the encoding enc. It writes at most
  // DecodedLen(len(src)) bytes to dst and returns the number of bytes
  // written. If src contains invalid base64 data, it will return the
  // number of bytes successfully written and CorruptInputError.
  // New line characters (\r and \n) are ignored.
  func (enc *Encoding) Decode(dst, src []byte) (n int, err error) {
  	n, _, err = enc.decode(dst, src)
  	return
  }
  
  // DecodeString returns the bytes represented by the base64 string s.
  func (enc *Encoding) DecodeString(s string) ([]byte, error) {
  	dbuf := make([]byte, enc.DecodedLen(len(s)))
  	n, _, err := enc.decode(dbuf, []byte(s))
  	return dbuf[:n], err
  }
  
  type decoder struct {
  	err     error
  	readErr error // error from r.Read
  	enc     *Encoding
  	r       io.Reader
  	end     bool       // saw end of message
  	buf     [1024]byte // leftover input
  	nbuf    int
  	out     []byte // leftover decoded output
  	outbuf  [1024 / 4 * 3]byte
  }
  
  func (d *decoder) Read(p []byte) (n int, err error) {
  	// Use leftover decoded output from last read.
  	if len(d.out) > 0 {
  		n = copy(p, d.out)
  		d.out = d.out[n:]
  		return n, nil
  	}
  
  	if d.err != nil {
  		return 0, d.err
  	}
  
  	// This code assumes that d.r strips supported whitespace ('\r' and '\n').
  
  	// Refill buffer.
  	for d.nbuf < 4 && d.readErr == nil {
  		nn := len(p) / 3 * 4
  		if nn < 4 {
  			nn = 4
  		}
  		if nn > len(d.buf) {
  			nn = len(d.buf)
  		}
  		nn, d.readErr = d.r.Read(d.buf[d.nbuf:nn])
  		d.nbuf += nn
  	}
  
  	if d.nbuf < 4 {
  		if d.enc.padChar == NoPadding && d.nbuf > 0 {
  			// Decode final fragment, without padding.
  			var nw int
  			nw, _, d.err = d.enc.decode(d.outbuf[:], d.buf[:d.nbuf])
  			d.nbuf = 0
  			d.end = true
  			d.out = d.outbuf[:nw]
  			n = copy(p, d.out)
  			d.out = d.out[n:]
  			if n > 0 || len(p) == 0 && len(d.out) > 0 {
  				return n, nil
  			}
  			if d.err != nil {
  				return 0, d.err
  			}
  		}
  		d.err = d.readErr
  		if d.err == io.EOF && d.nbuf > 0 {
  			d.err = io.ErrUnexpectedEOF
  		}
  		return 0, d.err
  	}
  
  	// Decode chunk into p, or d.out and then p if p is too small.
  	nr := d.nbuf / 4 * 4
  	nw := d.nbuf / 4 * 3
  	if nw > len(p) {
  		nw, d.end, d.err = d.enc.decode(d.outbuf[:], d.buf[:nr])
  		d.out = d.outbuf[:nw]
  		n = copy(p, d.out)
  		d.out = d.out[n:]
  	} else {
  		n, d.end, d.err = d.enc.decode(p, d.buf[:nr])
  	}
  	d.nbuf -= nr
  	copy(d.buf[:d.nbuf], d.buf[nr:])
  	return n, d.err
  }
  
  type newlineFilteringReader struct {
  	wrapped io.Reader
  }
  
  func (r *newlineFilteringReader) Read(p []byte) (int, error) {
  	n, err := r.wrapped.Read(p)
  	for n > 0 {
  		offset := 0
  		for i, b := range p[:n] {
  			if b != '\r' && b != '\n' {
  				if i != offset {
  					p[offset] = b
  				}
  				offset++
  			}
  		}
  		if offset > 0 {
  			return offset, err
  		}
  		// Previous buffer entirely whitespace, read again
  		n, err = r.wrapped.Read(p)
  	}
  	return n, err
  }
  
  // NewDecoder constructs a new base64 stream decoder.
  func NewDecoder(enc *Encoding, r io.Reader) io.Reader {
  	return &decoder{enc: enc, r: &newlineFilteringReader{r}}
  }
  
  // DecodedLen returns the maximum length in bytes of the decoded data
  // corresponding to n bytes of base64-encoded data.
  func (enc *Encoding) DecodedLen(n int) int {
  	if enc.padChar == NoPadding {
  		// Unpadded data may end with partial block of 2-3 characters.
  		return n * 6 / 8
  	}
  	// Padded base64 should always be a multiple of 4 characters in length.
  	return n / 4 * 3
  }
  

View as plain text