reader.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  /*
     6  Package zlib implements reading and writing of zlib format compressed data,
     7  as specified in RFC 1950.
     8  
     9  The implementation provides filters that uncompress during reading
    10  and compress during writing.  For example, to write compressed data
    11  to a buffer:
    12  
    13  	var b bytes.Buffer
    14  	w := zlib.NewWriter(&b)
    15  	w.Write([]byte("hello, world\n"))
    16  	w.Close()
    17  
    18  and to read that data back:
    19  
    20  	r, err := zlib.NewReader(&b)
    21  	io.Copy(os.Stdout, r)
    22  	r.Close()
    23  */
    24  package zlib
    25  
    26  import (
    27  	"bufio"
    28  	"compress/flate"
    29  	"errors"
    30  	"hash"
    31  	"hash/adler32"
    32  	"io"
    33  )
    34  
    35  const (
    36  	zlibDeflate   = 8
    37  	zlibMaxWindow = 7
    38  )
    39  
    40  var (
    41  	// ErrChecksum is returned when reading ZLIB data that has an invalid checksum.
    42  	ErrChecksum = errors.New("zlib: invalid checksum")
    43  	// ErrDictionary is returned when reading ZLIB data that has an invalid dictionary.
    44  	ErrDictionary = errors.New("zlib: invalid dictionary")
    45  	// ErrHeader is returned when reading ZLIB data that has an invalid header.
    46  	ErrHeader = errors.New("zlib: invalid header")
    47  )
    48  
    49  type reader struct {
    50  	r            flate.Reader
    51  	decompressor io.ReadCloser
    52  	digest       hash.Hash32
    53  	err          error
    54  	scratch      [4]byte
    55  }
    56  
    57  // Resetter resets a ReadCloser returned by NewReader or NewReaderDict
    58  // to switch to a new underlying Reader. This permits reusing a ReadCloser
    59  // instead of allocating a new one.
    60  type Resetter interface {
    61  	// Reset discards any buffered data and resets the Resetter as if it was
    62  	// newly initialized with the given reader.
    63  	Reset(r io.Reader, dict []byte) error
    64  }
    65  
    66  // NewReader creates a new ReadCloser.
    67  // Reads from the returned ReadCloser read and decompress data from r.
    68  // If r does not implement io.ByteReader, the decompressor may read more
    69  // data than necessary from r.
    70  // It is the caller's responsibility to call Close on the ReadCloser when done.
    71  //
    72  // The ReadCloser returned by NewReader also implements Resetter.
    73  func NewReader(r io.Reader) (io.ReadCloser, error) {
    74  	return NewReaderDict(r, nil)
    75  }
    76  
    77  // NewReaderDict is like NewReader but uses a preset dictionary.
    78  // NewReaderDict ignores the dictionary if the compressed data does not refer to it.
    79  // If the compressed data refers to a different dictionary, NewReaderDict returns ErrDictionary.
    80  //
    81  // The ReadCloser returned by NewReaderDict also implements Resetter.
    82  func NewReaderDict(r io.Reader, dict []byte) (io.ReadCloser, error) {
    83  	z := new(reader)
    84  	err := z.Reset(r, dict)
    85  	if err != nil {
    86  		return nil, err
    87  	}
    88  	return z, nil
    89  }
    90  
    91  func (z *reader) Read(p []byte) (int, error) {
    92  	if z.err != nil {
    93  		return 0, z.err
    94  	}
    95  
    96  	var n int
    97  	n, z.err = z.decompressor.Read(p)
    98  	z.digest.Write(p[0:n])
    99  	if z.err != io.EOF {
   100  		// In the normal case we return here.
   101  		return n, z.err
   102  	}
   103  
   104  	// Finished file; check checksum.
   105  	if _, err := io.ReadFull(z.r, z.scratch[0:4]); err != nil {
   106  		if err == io.EOF {
   107  			err = io.ErrUnexpectedEOF
   108  		}
   109  		z.err = err
   110  		return n, z.err
   111  	}
   112  	// ZLIB (RFC 1950) is big-endian, unlike GZIP (RFC 1952).
   113  	checksum := uint32(z.scratch[0])<<24 | uint32(z.scratch[1])<<16 | uint32(z.scratch[2])<<8 | uint32(z.scratch[3])
   114  	if checksum != z.digest.Sum32() {
   115  		z.err = ErrChecksum
   116  		return n, z.err
   117  	}
   118  	return n, io.EOF
   119  }
   120  
   121  // Calling Close does not close the wrapped io.Reader originally passed to NewReader.
   122  // In order for the ZLIB checksum to be verified, the reader must be
   123  // fully consumed until the io.EOF.
   124  func (z *reader) Close() error {
   125  	if z.err != nil && z.err != io.EOF {
   126  		return z.err
   127  	}
   128  	z.err = z.decompressor.Close()
   129  	return z.err
   130  }
   131  
   132  func (z *reader) Reset(r io.Reader, dict []byte) error {
   133  	*z = reader{decompressor: z.decompressor}
   134  	if fr, ok := r.(flate.Reader); ok {
   135  		z.r = fr
   136  	} else {
   137  		z.r = bufio.NewReader(r)
   138  	}
   139  
   140  	// Read the header (RFC 1950 section 2.2.).
   141  	_, z.err = io.ReadFull(z.r, z.scratch[0:2])
   142  	if z.err != nil {
   143  		if z.err == io.EOF {
   144  			z.err = io.ErrUnexpectedEOF
   145  		}
   146  		return z.err
   147  	}
   148  	h := uint(z.scratch[0])<<8 | uint(z.scratch[1])
   149  	if (z.scratch[0]&0x0f != zlibDeflate) || (z.scratch[0]>>4 > zlibMaxWindow) || (h%31 != 0) {
   150  		z.err = ErrHeader
   151  		return z.err
   152  	}
   153  	haveDict := z.scratch[1]&0x20 != 0
   154  	if haveDict {
   155  		_, z.err = io.ReadFull(z.r, z.scratch[0:4])
   156  		if z.err != nil {
   157  			if z.err == io.EOF {
   158  				z.err = io.ErrUnexpectedEOF
   159  			}
   160  			return z.err
   161  		}
   162  		checksum := uint32(z.scratch[0])<<24 | uint32(z.scratch[1])<<16 | uint32(z.scratch[2])<<8 | uint32(z.scratch[3])
   163  		if checksum != adler32.Checksum(dict) {
   164  			z.err = ErrDictionary
   165  			return z.err
   166  		}
   167  	}
   168  
   169  	if z.decompressor == nil {
   170  		if haveDict {
   171  			z.decompressor = flate.NewReaderDict(z.r, dict)
   172  		} else {
   173  			z.decompressor = flate.NewReader(z.r)
   174  		}
   175  	} else {
   176  		z.decompressor.(flate.Resetter).Reset(z.r, dict)
   177  	}
   178  	z.digest = adler32.New()
   179  	return nil
   180  }
   181
View as plain text