1 files changed, 97 insertions, 119 deletions

diff --git a/libgo/go/compress/gzip/gunzip.go b/libgo/go/compress/gzip/gunzip.go
index 3d331454a69..7e640692f3f 100644
--- a/libgo/go/compress/gzip/gunzip.go
+++ b/libgo/go/compress/gzip/gunzip.go
@@ -9,8 +9,8 @@ package gzip
 import (
 	"bufio"
 	"compress/flate"
+	"encoding/binary"
 	"errors"
-	"hash"
 	"hash/crc32"
 	"io"
 	"time"
@@ -27,13 +27,6 @@ const (
 	flagComment = 1 << 4
 )
 
-func makeReader(r io.Reader) flate.Reader {
-	if rr, ok := r.(flate.Reader); ok {
-		return rr
-	}
-	return bufio.NewReader(r)
-}
-
 var (
 	// ErrChecksum is returned when reading GZIP data that has an invalid checksum.
 	ErrChecksum = errors.New("gzip: invalid checksum")
@@ -41,6 +34,16 @@ var (
 	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.
 //
@@ -58,23 +61,22 @@ type Header struct {
 // 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
+// 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
+// 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       hash.Hash32
-	size         uint32
-	flg          byte
+	digest       uint32 // CRC-32, IEEE polynomial (section 8)
+	size         uint32 // Uncompressed size (section 2.3.1)
 	buf          [512]byte
 	err          error
 	multistream  bool
@@ -89,10 +91,7 @@ type Reader struct {
 // The Reader.Header fields will be valid in the Reader returned.
 func NewReader(r io.Reader) (*Reader, error) {
 	z := new(Reader)
-	z.r = makeReader(r)
-	z.multistream = true
-	z.digest = crc32.NewIEEE()
-	if err := z.readHeader(true); err != nil {
+	if err := z.Reset(r); err != nil {
 		return nil, err
 	}
 	return z, nil
@@ -102,16 +101,17 @@ func NewReader(r io.Reader) (*Reader, error) {
 // 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.r = makeReader(r)
-	if z.digest == nil {
-		z.digest = crc32.NewIEEE()
+	*z = Reader{
+		decompressor: z.decompressor,
+		multistream:  true,
+	}
+	if rr, ok := r.(flate.Reader); ok {
+		z.r = rr
 	} else {
-		z.digest.Reset()
+		z.r = bufio.NewReader(r)
 	}
-	z.size = 0
-	z.err = nil
-	z.multistream = true
-	return z.readHeader(true)
+	z.Header, z.err = z.readHeader()
+	return z.err
 }
 
 // Multistream controls whether the reader supports multistream files.
@@ -134,14 +134,13 @@ func (z *Reader) Multistream(ok bool) {
 	z.multistream = ok
 }
 
-// GZIP (RFC 1952) is little-endian, unlike ZLIB (RFC 1950).
-func get4(p []byte) uint32 {
-	return uint32(p[0]) | uint32(p[1])<<8 | uint32(p[2])<<16 | uint32(p[3])<<24
-}
-
+// 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
+	needConv := false
 	for i := 0; ; i++ {
 		if i >= len(z.buf) {
 			return "", ErrHeader
@@ -151,159 +150,138 @@ func (z *Reader) readString() (string, error) {
 			return "", err
 		}
 		if z.buf[i] > 0x7f {
-			needconv = true
+			needConv = true
 		}
 		if z.buf[i] == 0 {
-			// GZIP (RFC 1952) specifies that strings are NUL-terminated ISO 8859-1 (Latin-1).
-			if needconv {
+			// 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[0:i] {
+				for _, v := range z.buf[:i] {
 					s = append(s, rune(v))
 				}
 				return string(s), nil
 			}
-			return string(z.buf[0:i]), nil
-		}
-	}
-}
-
-func (z *Reader) read2() (uint32, error) {
-	_, err := io.ReadFull(z.r, z.buf[0:2])
-	if err != nil {
-		if err == io.EOF {
-			err = io.ErrUnexpectedEOF
+			return string(z.buf[:i]), nil
 		}
-		return 0, err
 	}
-	return uint32(z.buf[0]) | uint32(z.buf[1])<<8, nil
 }
 
-func (z *Reader) readHeader(save bool) error {
-	_, err := io.ReadFull(z.r, z.buf[0:10])
-	if err != nil {
-		// RFC1952 section 2.2 says the following:
+// 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 err
+		return hdr, err
 	}
 	if z.buf[0] != gzipID1 || z.buf[1] != gzipID2 || z.buf[2] != gzipDeflate {
-		return ErrHeader
-	}
-	z.flg = z.buf[3]
-	if save {
-		z.ModTime = time.Unix(int64(get4(z.buf[4:8])), 0)
-		// z.buf[8] is xfl, ignored
-		z.OS = z.buf[9]
+		return hdr, ErrHeader
 	}
-	z.digest.Reset()
-	z.digest.Write(z.buf[0:10])
+	flg := z.buf[3]
+	hdr.ModTime = time.Unix(int64(le.Uint32(z.buf[4:8])), 0)
+	// z.buf[8] is XFL and is currently ignored.
+	hdr.OS = z.buf[9]
+	z.digest = crc32.ChecksumIEEE(z.buf[:10])
 
-	if z.flg&flagExtra != 0 {
-		n, err := z.read2()
-		if err != nil {
-			return err
+	if flg&flagExtra != 0 {
+		if _, err = io.ReadFull(z.r, z.buf[:2]); err != nil {
+			return hdr, noEOF(err)
 		}
-		data := make([]byte, n)
+		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 {
-			if err == io.EOF {
-				err = io.ErrUnexpectedEOF
-			}
-			return err
-		}
-		if save {
-			z.Extra = data
+			return hdr, noEOF(err)
 		}
+		z.digest = crc32.Update(z.digest, crc32.IEEETable, data)
+		hdr.Extra = data
 	}
 
 	var s string
-	if z.flg&flagName != 0 {
+	if flg&flagName != 0 {
 		if s, err = z.readString(); err != nil {
-			return err
-		}
-		if save {
-			z.Name = s
+			return hdr, err
 		}
+		hdr.Name = s
 	}
 
-	if z.flg&flagComment != 0 {
+	if flg&flagComment != 0 {
 		if s, err = z.readString(); err != nil {
-			return err
-		}
-		if save {
-			z.Comment = s
+			return hdr, err
 		}
+		hdr.Comment = s
 	}
 
-	if z.flg&flagHdrCrc != 0 {
-		n, err := z.read2()
-		if err != nil {
-			return err
+	if flg&flagHdrCrc != 0 {
+		if _, err = io.ReadFull(z.r, z.buf[:2]); err != nil {
+			return hdr, noEOF(err)
 		}
-		sum := z.digest.Sum32() & 0xFFFF
-		if n != sum {
-			return ErrHeader
+		digest := le.Uint16(z.buf[:2])
+		if digest != uint16(z.digest) {
+			return hdr, ErrHeader
 		}
 	}
 
-	z.digest.Reset()
+	z.digest = 0
 	if z.decompressor == nil {
 		z.decompressor = flate.NewReader(z.r)
 	} else {
 		z.decompressor.(flate.Resetter).Reset(z.r, nil)
 	}
-	return nil
+	return hdr, nil
 }
 
 func (z *Reader) Read(p []byte) (n int, err error) {
 	if z.err != nil {
 		return 0, z.err
 	}
-	if len(p) == 0 {
-		return 0, nil
-	}
 
-	n, err = z.decompressor.Read(p)
-	z.digest.Write(p[0:n])
+	n, z.err = z.decompressor.Read(p)
+	z.digest = crc32.Update(z.digest, crc32.IEEETable, p[:n])
 	z.size += uint32(n)
-	if n != 0 || err != io.EOF {
-		z.err = err
-		return
+	if z.err != io.EOF {
+		// In the normal case we return here.
+		return n, z.err
 	}
 
-	// Finished file; check checksum + size.
-	if _, err := io.ReadFull(z.r, z.buf[0:8]); err != nil {
-		if err == io.EOF {
-			err = io.ErrUnexpectedEOF
-		}
-		z.err = err
-		return 0, 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
 	}
-	crc32, isize := get4(z.buf[0:4]), get4(z.buf[4:8])
-	sum := z.digest.Sum32()
-	if sum != crc32 || isize != z.size {
+	digest := le.Uint32(z.buf[:4])
+	size := le.Uint32(z.buf[4:8])
+	if digest != z.digest || size != z.size {
 		z.err = ErrChecksum
-		return 0, z.err
+		return n, z.err
 	}
+	z.digest, z.size = 0, 0
 
-	// File is ok; is there another?
+	// File is ok; check if there is another.
 	if !z.multistream {
-		return 0, io.EOF
+		return n, io.EOF
 	}
+	z.err = nil // Remove io.EOF
 
-	if err = z.readHeader(false); err != nil {
-		z.err = err
-		return
+	if _, z.err = z.readHeader(); z.err != nil {
+		return n, z.err
 	}
 
-	// Yes.  Reset and read from it.
-	z.digest.Reset()
-	z.size = 0
+	// 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() }