...
Run Format

Source file src/compress/gzip/gunzip.go

  // 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 gzip implements reading and writing of gzip format compressed files,
  // as specified in RFC 1952.
  package gzip
  
  import (
  	"bufio"
  	"compress/flate"
  	"encoding/binary"
  	"errors"
  	"hash/crc32"
  	"io"
  	"time"
  )
  
  const (
  	gzipID1     = 0x1f
  	gzipID2     = 0x8b
  	gzipDeflate = 8
  	flagText    = 1 << 0
  	flagHdrCrc  = 1 << 1
  	flagExtra   = 1 << 2
  	flagName    = 1 << 3
  	flagComment = 1 << 4
  )
  
  var (
  	// ErrChecksum is returned when reading GZIP data that has an invalid checksum.
  	ErrChecksum = errors.New("gzip: invalid checksum")
  	// ErrHeader is returned when reading GZIP data that has an invalid header.
  	ErrHeader = errors.New("gzip: invalid header")
  )
  
  var le = binary.LittleEndian
  
  // noEOF converts io.EOF to io.ErrUnexpectedEOF.
  func noEOF(err error) error {
  	if err == io.EOF {
  		return io.ErrUnexpectedEOF
  	}
  	return err
  }
  
  // The gzip file stores a header giving metadata about the compressed file.
  // That header is exposed as the fields of the Writer and Reader structs.
  //
  // Strings must be UTF-8 encoded and may only contain Unicode code points
  // U+0001 through U+00FF, due to limitations of the GZIP file format.
  type Header struct {
  	Comment string    // comment
  	Extra   []byte    // "extra data"
  	ModTime time.Time // modification time
  	Name    string    // file name
  	OS      byte      // operating system type
  }
  
  // A Reader is an io.Reader that can be read to retrieve
  // uncompressed data from a gzip-format compressed file.
  //
  // In general, a gzip file can be a concatenation of gzip files,
  // each with its own header. Reads from the Reader
  // return the concatenation of the uncompressed data of each.
  // Only the first header is recorded in the Reader fields.
  //
  // Gzip files store a length and checksum of the uncompressed data.
  // The Reader will return a ErrChecksum when Read
  // reaches the end of the uncompressed data if it does not
  // have the expected length or checksum. Clients should treat data
  // returned by Read as tentative until they receive the io.EOF
  // marking the end of the data.
  type Reader struct {
  	Header       // valid after NewReader or Reader.Reset
  	r            flate.Reader
  	decompressor io.ReadCloser
  	digest       uint32 // CRC-32, IEEE polynomial (section 8)
  	size         uint32 // Uncompressed size (section 2.3.1)
  	buf          [512]byte
  	err          error
  	multistream  bool
  }
  
  // NewReader creates a new Reader reading the given reader.
  // If r does not also implement io.ByteReader,
  // the decompressor may read more data than necessary from r.
  //
  // It is the caller's responsibility to call Close on the Reader when done.
  //
  // The Reader.Header fields will be valid in the Reader returned.
  func NewReader(r io.Reader) (*Reader, error) {
  	z := new(Reader)
  	if err := z.Reset(r); err != nil {
  		return nil, err
  	}
  	return z, nil
  }
  
  // Reset discards the Reader z's state and makes it equivalent to the
  // result of its original state from NewReader, but reading from r instead.
  // This permits reusing a Reader rather than allocating a new one.
  func (z *Reader) Reset(r io.Reader) error {
  	*z = Reader{
  		decompressor: z.decompressor,
  		multistream:  true,
  	}
  	if rr, ok := r.(flate.Reader); ok {
  		z.r = rr
  	} else {
  		z.r = bufio.NewReader(r)
  	}
  	z.Header, z.err = z.readHeader()
  	return z.err
  }
  
  // Multistream controls whether the reader supports multistream files.
  //
  // If enabled (the default), the Reader expects the input to be a sequence
  // of individually gzipped data streams, each with its own header and
  // trailer, ending at EOF. The effect is that the concatenation of a sequence
  // of gzipped files is treated as equivalent to the gzip of the concatenation
  // of the sequence. This is standard behavior for gzip readers.
  //
  // Calling Multistream(false) disables this behavior; disabling the behavior
  // can be useful when reading file formats that distinguish individual gzip
  // data streams or mix gzip data streams with other data streams.
  // In this mode, when the Reader reaches the end of the data stream,
  // Read returns io.EOF. If the underlying reader implements io.ByteReader,
  // it will be left positioned just after the gzip stream.
  // To start the next stream, call z.Reset(r) followed by z.Multistream(false).
  // If there is no next stream, z.Reset(r) will return io.EOF.
  func (z *Reader) Multistream(ok bool) {
  	z.multistream = ok
  }
  
  // readString reads a NUL-terminated string from z.r.
  // It treats the bytes read as being encoded as ISO 8859-1 (Latin-1) and
  // will output a string encoded using UTF-8.
  // This method always updates z.digest with the data read.
  func (z *Reader) readString() (string, error) {
  	var err error
  	needConv := false
  	for i := 0; ; i++ {
  		if i >= len(z.buf) {
  			return "", ErrHeader
  		}
  		z.buf[i], err = z.r.ReadByte()
  		if err != nil {
  			return "", err
  		}
  		if z.buf[i] > 0x7f {
  			needConv = true
  		}
  		if z.buf[i] == 0 {
  			// Digest covers the NUL terminator.
  			z.digest = crc32.Update(z.digest, crc32.IEEETable, z.buf[:i+1])
  
  			// Strings are ISO 8859-1, Latin-1 (RFC 1952, section 2.3.1).
  			if needConv {
  				s := make([]rune, 0, i)
  				for _, v := range z.buf[:i] {
  					s = append(s, rune(v))
  				}
  				return string(s), nil
  			}
  			return string(z.buf[:i]), nil
  		}
  	}
  }
  
  // readHeader reads the GZIP header according to section 2.3.1.
  // This method does not set z.err.
  func (z *Reader) readHeader() (hdr Header, err error) {
  	if _, err = io.ReadFull(z.r, z.buf[:10]); err != nil {
  		// RFC 1952, section 2.2, says the following:
  		//	A gzip file consists of a series of "members" (compressed data sets).
  		//
  		// Other than this, the specification does not clarify whether a
  		// "series" is defined as "one or more" or "zero or more". To err on the
  		// side of caution, Go interprets this to mean "zero or more".
  		// Thus, it is okay to return io.EOF here.
  		return hdr, err
  	}
  	if z.buf[0] != gzipID1 || z.buf[1] != gzipID2 || z.buf[2] != gzipDeflate {
  		return hdr, ErrHeader
  	}
  	flg := z.buf[3]
  	if t := int64(le.Uint32(z.buf[4:8])); t > 0 {
  		// Section 2.3.1, the zero value for MTIME means that the
  		// modified time is not set.
  		hdr.ModTime = time.Unix(t, 0)
  	}
  	// z.buf[8] is XFL and is currently ignored.
  	hdr.OS = z.buf[9]
  	z.digest = crc32.ChecksumIEEE(z.buf[:10])
  
  	if flg&flagExtra != 0 {
  		if _, err = io.ReadFull(z.r, z.buf[:2]); err != nil {
  			return hdr, noEOF(err)
  		}
  		z.digest = crc32.Update(z.digest, crc32.IEEETable, z.buf[:2])
  		data := make([]byte, le.Uint16(z.buf[:2]))
  		if _, err = io.ReadFull(z.r, data); err != nil {
  			return hdr, noEOF(err)
  		}
  		z.digest = crc32.Update(z.digest, crc32.IEEETable, data)
  		hdr.Extra = data
  	}
  
  	var s string
  	if flg&flagName != 0 {
  		if s, err = z.readString(); err != nil {
  			return hdr, err
  		}
  		hdr.Name = s
  	}
  
  	if flg&flagComment != 0 {
  		if s, err = z.readString(); err != nil {
  			return hdr, err
  		}
  		hdr.Comment = s
  	}
  
  	if flg&flagHdrCrc != 0 {
  		if _, err = io.ReadFull(z.r, z.buf[:2]); err != nil {
  			return hdr, noEOF(err)
  		}
  		digest := le.Uint16(z.buf[:2])
  		if digest != uint16(z.digest) {
  			return hdr, ErrHeader
  		}
  	}
  
  	z.digest = 0
  	if z.decompressor == nil {
  		z.decompressor = flate.NewReader(z.r)
  	} else {
  		z.decompressor.(flate.Resetter).Reset(z.r, nil)
  	}
  	return hdr, nil
  }
  
  // Read implements io.Reader, reading uncompressed bytes from its underlying Reader.
  func (z *Reader) Read(p []byte) (n int, err error) {
  	if z.err != nil {
  		return 0, z.err
  	}
  
  	n, z.err = z.decompressor.Read(p)
  	z.digest = crc32.Update(z.digest, crc32.IEEETable, p[:n])
  	z.size += uint32(n)
  	if z.err != io.EOF {
  		// In the normal case we return here.
  		return n, z.err
  	}
  
  	// Finished file; check checksum and size.
  	if _, err := io.ReadFull(z.r, z.buf[:8]); err != nil {
  		z.err = noEOF(err)
  		return n, z.err
  	}
  	digest := le.Uint32(z.buf[:4])
  	size := le.Uint32(z.buf[4:8])
  	if digest != z.digest || size != z.size {
  		z.err = ErrChecksum
  		return n, z.err
  	}
  	z.digest, z.size = 0, 0
  
  	// File is ok; check if there is another.
  	if !z.multistream {
  		return n, io.EOF
  	}
  	z.err = nil // Remove io.EOF
  
  	if _, z.err = z.readHeader(); z.err != nil {
  		return n, z.err
  	}
  
  	// Read from next file, if necessary.
  	if n > 0 {
  		return n, nil
  	}
  	return z.Read(p)
  }
  
  // Close closes the Reader. It does not close the underlying io.Reader.
  // In order for the GZIP checksum to be verified, the reader must be
  // fully consumed until the io.EOF.
  func (z *Reader) Close() error { return z.decompressor.Close() }
  

View as plain text