works

2 files changed, 98 insertions, 30 deletions

diff --git a/main.go b/main.go
index 88ff7c1..a0d217c 100644
--- a/main.go
+++ b/main.go
@@ -2,6 +2,7 @@ package main
 
 import (
 	"bytes"
+	"compress/flate"
 	"encoding/json"
 	"errors"
 	"fmt"
@@ -9,6 +10,7 @@ import (
 	"io"
 	"log"
 	"os"
+	"path"
 	"time"
 
 	flag "github.com/spf13/pflag"
@@ -16,7 +18,9 @@ import (
 
 func main() {
 	var zipFilename string
+	var output string
 	flag.StringVarP(&zipFilename, "filename", "f", "", "filename")
+	flag.StringVarP(&output, "output", "o", "./out/", "output directory")
 	flag.Parse()
 
 	f, err := os.Open(zipFilename)
@@ -27,7 +31,10 @@ func main() {
 
 	for {
 		var sig [4]byte
-		if _, err = io.ReadFull(f, sig[:]); err != nil {
+		_, err = io.ReadFull(f, sig[:])
+		if err == io.EOF {
+			break
+		} else if err != nil {
 			log.Fatal("failed to read signature: ", err)
 		}
 
@@ -41,22 +48,47 @@ func main() {
 				pos, _ := f.Seek(0, io.SeekCurrent)
 				log.Fatalf("failed to read file header at 0x%x: %s", pos, err)
 			}
-			hdr.HeaderOffset, _ = f.Seek(0, io.SeekCurrent)
+			hdr.Offset, _ = f.Seek(0, io.SeekCurrent)
 			j, _ := json.MarshalIndent(hdr, "", "  ")
-			log.Printf("File at 0x%x: %s", hdr.HeaderOffset, string(j))
+			log.Printf("File at 0x%x: %s", hdr.Offset, string(j))
+			hasDD := false
 			if hdr.Flags&0x08 > 0 {
-				log.Printf("Searching for file's end...")
-				if err = findAndUseDataDescriptor(&hdr, f); err != nil {
-					log.Fatalf("failed to udpate using data descriptor at 0x%x: %s", hdr.HeaderOffset, err)
+				if hdr.CompressedSize == 0 {
+					log.Printf("Searching for file's end...")
+					if err = findAndUseDataDescriptor(&hdr, f); err != nil {
+						log.Fatalf("failed to udpate using data descriptor at 0x%x: %s", hdr.Offset, err)
+					}
 				}
+				hasDD = true
+			}
+			err := extract(f, &hdr, output)
+			if err != nil {
+				log.Fatalf("failed to extract file: %s", err)
 			}
-			if pos, err := f.Seek(hdr.HeaderOffset+int64(hdr.CompressedSize), io.SeekStart); err != nil {
+			size := int64(hdr.CompressedSize)
+			if size == 0xFFFFFFFF {
+				size = int64(hdr.CompressedSize64)
+			}
+			pos, err := f.Seek(hdr.Offset+size, io.SeekStart)
+			if err != nil {
 				log.Fatalf("failed to seek to next header, stopped at 0x%x: %s", pos, err)
 			}
-
-		case dataDescriptorSignature:
-			io.CopyN(io.Discard, f, 12)
-
+			if hasDD {
+				if _, err = io.ReadFull(f, sig[:]); err != nil {
+					log.Fatal("failed to read DD signature: ", err)
+				}
+				sigb := readBuf(sig[:])
+				signature = sigb.uint32()
+				if signature != dataDescriptorSignature {
+					log.Fatalf("unexpected signature at 0x%x: got %08x", pos, signature)
+				}
+				log.Printf("Skipping data descriptor at 0x%x, zip64=%t...", pos, hdr.zip64)
+				if hdr.zip64 {
+					io.CopyN(io.Discard, f, 20)
+				} else {
+					io.CopyN(io.Discard, f, 12)
+				}
+			}
 		case directoryHeaderSignature:
 			var hdr File
 			if err = readDirectoryHeader(&hdr, f); err != nil {
@@ -109,7 +141,7 @@ func parseFileHeader(f *File, r io.Reader) error {
 
 	needUSize := f.UncompressedSize == ^uint32(0)
 	needCSize := f.CompressedSize == ^uint32(0)
-	needHeaderOffset := f.HeaderOffset == int64(^uint32(0))
+	needHeaderOffset := f.Offset == int64(^uint32(0))
 
 	// Best effort to find what we need.
 	// Other zip authors might not even follow the basic format,
@@ -151,7 +183,7 @@ parseExtras:
 				if len(fieldBuf) < 8 {
 					return ErrFormat
 				}
-				f.HeaderOffset = int64(fieldBuf.uint64())
+				f.Offset = int64(fieldBuf.uint64())
 			}
 		case ntfsExtraID:
 			if len(fieldBuf) < 4 {
@@ -299,11 +331,11 @@ func findAndUseDataDescriptor(f *File, r io.ReadSeeker) error {
 		if err := parseDataDescriptor(f, r); err != nil {
 			return fmt.Errorf("failed to read descriptor : %s", err)
 		}
-		if descriptorPos-int64(f.DescriptorCompressedSize) == f.HeaderOffset+4 {
+		if descriptorPos-int64(f.DescriptorCompressedSize) == f.Offset+4 {
 			break
 		} else {
 			log.Printf("invalid size: expected offset %x (hex) != %x (hex), compsize=%d, size=%d",
-				f.HeaderOffset+4, descriptorPos-int64(f.DescriptorCompressedSize), f.DescriptorCompressedSize, f.DescriptorUncompressedSize)
+				f.Offset+4, descriptorPos-int64(f.DescriptorCompressedSize), f.DescriptorCompressedSize, f.DescriptorUncompressedSize)
 		}
 		if pos, err = r.Seek(descriptorPos+4, io.SeekStart); err != nil {
 			return fmt.Errorf("failed to seek to descriptor at 0x%x: %s", pos, err)
@@ -313,8 +345,8 @@ func findAndUseDataDescriptor(f *File, r io.ReadSeeker) error {
 	}
 
 	// bring back at the beginning of the header (due to chunking)
-	if _, err := r.Seek(f.HeaderOffset, io.SeekStart); err != nil {
-		return fmt.Errorf("failed to seek to beginning at %d: %s", f.HeaderOffset, err)
+	if _, err := r.Seek(f.Offset, io.SeekStart); err != nil {
+		return fmt.Errorf("failed to seek to beginning at %d: %s", f.Offset, err)
 	}
 	crc := crc32.NewIEEE()
 	if _, err := io.CopyN(crc, r, int64(f.DescriptorCompressedSize)); err != nil {
@@ -331,3 +363,39 @@ func findAndUseDataDescriptor(f *File, r io.ReadSeeker) error {
 
 	return nil
 }
+
+func extract(r io.ReadSeeker, f *File, outputDirectory string) error {
+	dir, base := path.Split(path.Clean(f.Name))
+	directory := path.Join(outputDirectory, dir)
+	err := os.MkdirAll(directory, 0o755)
+	if err != nil {
+		return err
+	}
+
+	w, err := os.OpenFile(path.Join(directory, base), os.O_WRONLY|os.O_CREATE, 0o644)
+	if err != nil {
+		return err
+	}
+	defer w.Close()
+	if _, err := r.Seek(f.Offset, io.SeekStart); err != nil {
+		return err
+	}
+	size := int64(f.UncompressedSize)
+	if size == 0xFFFFFFFF {
+		size = int64(f.UncompressedSize64)
+	}
+	if f.Method == Deflate {
+		fr := flate.NewReader(r)
+		if _, err := io.CopyN(w, fr, size); err != nil {
+			return err
+		}
+	} else if f.Method == Store {
+		if _, err := io.CopyN(w, r, size); err != nil {
+			return err
+		}
+	} else {
+		return errors.New("invalid method")
+	}
+
+	return nil
+}
diff --git a/reader.go b/reader.go
index 6ef8885..1042d92 100644
--- a/reader.go
+++ b/reader.go
@@ -54,10 +54,10 @@ type ReadCloser struct {
 // The file content can be accessed by calling Open.
 type File struct {
 	FileHeader
-	zip          *Reader
-	zipr         io.ReaderAt
-	HeaderOffset int64 // includes overall ZIP archive baseOffset
-	zip64        bool  // zip64 extended information extra field presence
+	zip    *Reader
+	zipr   io.ReaderAt
+	Offset int64 // includes overall ZIP archive baseOffset
+	zip64  bool  // zip64 extended information extra field presence
 }
 
 // OpenReader will open the Zip file specified by name and return a ReadCloser.
@@ -143,7 +143,7 @@ func (z *Reader) init(r io.ReaderAt, size int64) error {
 		if err != nil {
 			return err
 		}
-		f.HeaderOffset += z.baseOffset
+		f.Offset += z.baseOffset
 		z.File = append(z.File, f)
 	}
 	if uint16(len(z.File)) != uint16(end.DirectoryRecords) { // only compare 16 bits here
@@ -187,7 +187,7 @@ func (f *File) DataOffset() (offset int64, err error) {
 	if err != nil {
 		return
 	}
-	return f.HeaderOffset + bodyOffset, nil
+	return f.Offset + bodyOffset, nil
 }
 
 // Open returns a ReadCloser that provides access to the File's contents.
@@ -198,7 +198,7 @@ func (f *File) Open() (io.ReadCloser, error) {
 		return nil, err
 	}
 	size := int64(f.CompressedSize64)
-	r := io.NewSectionReader(f.zipr, f.HeaderOffset+bodyOffset, size)
+	r := io.NewSectionReader(f.zipr, f.Offset+bodyOffset, size)
 	dcomp := f.zip.decompressor(f.Method)
 	if dcomp == nil {
 		return nil, ErrAlgorithm
@@ -206,7 +206,7 @@ func (f *File) Open() (io.ReadCloser, error) {
 	var rc io.ReadCloser = dcomp(r)
 	var desr io.Reader
 	if f.hasDataDescriptor() {
-		desr = io.NewSectionReader(f.zipr, f.HeaderOffset+bodyOffset+size, dataDescriptorLen)
+		desr = io.NewSectionReader(f.zipr, f.Offset+bodyOffset+size, dataDescriptorLen)
 	}
 	rc = &checksumReader{
 		rc:   rc,
@@ -224,7 +224,7 @@ func (f *File) OpenRaw() (io.Reader, error) {
 	if err != nil {
 		return nil, err
 	}
-	r := io.NewSectionReader(f.zipr, f.HeaderOffset+bodyOffset, int64(f.CompressedSize64))
+	r := io.NewSectionReader(f.zipr, f.Offset+bodyOffset, int64(f.CompressedSize64))
 	return r, nil
 }
 
@@ -287,7 +287,7 @@ func (r *checksumReader) Close() error { return r.rc.Close() }
 // and returns the file body offset.
 func (f *File) findBodyOffset() (int64, error) {
 	var buf [fileHeaderLen]byte
-	if _, err := f.zipr.ReadAt(buf[:], f.HeaderOffset); err != nil {
+	if _, err := f.zipr.ReadAt(buf[:], f.Offset); err != nil {
 		return 0, err
 	}
 	b := readBuf(buf[:])
@@ -325,7 +325,7 @@ func readDirectoryHeader(f *File, r io.Reader) error {
 	commentLen := int(b.uint16())
 	b = b[4:] // skipped start disk number and internal attributes (2x uint16)
 	f.ExternalAttrs = b.uint32()
-	f.HeaderOffset = int64(b.uint32())
+	f.Offset = int64(b.uint32())
 	d := make([]byte, filenameLen+extraLen+commentLen)
 	if _, err := io.ReadFull(r, d); err != nil {
 		return err
@@ -354,7 +354,7 @@ func readDirectoryHeader(f *File, r io.Reader) error {
 
 	needUSize := f.UncompressedSize == ^uint32(0)
 	needCSize := f.CompressedSize == ^uint32(0)
-	needHeaderOffset := f.HeaderOffset == int64(^uint32(0))
+	needHeaderOffset := f.Offset == int64(^uint32(0))
 
 	// Best effort to find what we need.
 	// Other zip authors might not even follow the basic format,
@@ -396,7 +396,7 @@ parseExtras:
 				if len(fieldBuf) < 8 {
 					return ErrFormat
 				}
-				f.HeaderOffset = int64(fieldBuf.uint64())
+				f.Offset = int64(fieldBuf.uint64())
 			}
 		case ntfsExtraID:
 			if len(fieldBuf) < 4 {