archive/zip: permit zip files to have prefixes

A Java jar file is a zip file, but it can have a prefix that is a bash
script that unpacks the zip file. Most zip programs ignore such prefixes.
This CL changes the archive/zip package to do the same.

Fixes #10464
Fixes #51337

Change-Id: I976e9c64684644317bd21077bc5b4a2baf626ee6
Reviewed-on: https://go-review.googlesource.com/c/go/+/387976
Run-TryBot: Ian Lance Taylor <iant@google.com>
Reviewed-by: Ian Lance Taylor <iant@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Run-TryBot: Ian Lance Taylor <iant@golang.org>
Reviewed-by: David Chase <drchase@google.com>
Reviewed-by: Joseph Tsai <joetsai@digital-static.net>
Auto-Submit: Ian Lance Taylor <iant@google.com>

4 files changed, 41 insertions, 13 deletions

diff --git a/src/archive/zip/reader.go b/src/archive/zip/reader.go
index 19a9c3b2db..d875c7be25 100644
--- a/src/archive/zip/reader.go
+++ b/src/archive/zip/reader.go
@@ -33,6 +33,10 @@ type Reader struct {
 	Comment       string
 	decompressors map[uint16]Decompressor
 
+	// Some JAR files are zip files with a prefix that is a bash script.
+	// The baseOffset field is the start of the zip file proper.
+	baseOffset int64
+
 	// fileList is a list of files sorted by ename,
 	// for use by the Open method.
 	fileListOnce sync.Once
@@ -52,7 +56,7 @@ type File struct {
 	FileHeader
 	zip          *Reader
 	zipr         io.ReaderAt
-	headerOffset int64
+	headerOffset int64 // includes overall ZIP archive baseOffset
 	zip64        bool // zip64 extended information extra field presence
 }
 
@@ -90,11 +94,12 @@ func NewReader(r io.ReaderAt, size int64) (*Reader, error) {
 }
 
 func (z *Reader) init(r io.ReaderAt, size int64) error {
-	end, err := readDirectoryEnd(r, size)
+	end, baseOffset, err := readDirectoryEnd(r, size)
 	if err != nil {
 		return err
 	}
 	z.r = r
+	z.baseOffset = baseOffset
 	// Since the number of directory records is not validated, it is not
 	// safe to preallocate z.File without first checking that the specified
 	// number of files is reasonable, since a malformed archive may
@@ -106,7 +111,7 @@ func (z *Reader) init(r io.ReaderAt, size int64) error {
 	}
 	z.Comment = end.comment
 	rs := io.NewSectionReader(r, 0, size)
-	if _, err = rs.Seek(int64(end.directoryOffset), io.SeekStart); err != nil {
+	if _, err = rs.Seek(z.baseOffset+int64(end.directoryOffset), io.SeekStart); err != nil {
 		return err
 	}
 	buf := bufio.NewReader(rs)
@@ -124,6 +129,7 @@ func (z *Reader) init(r io.ReaderAt, size int64) error {
 		if err != nil {
 			return err
 		}
+		f.headerOffset += z.baseOffset
 		z.File = append(z.File, f)
 	}
 	if uint16(len(z.File)) != uint16(end.directoryRecords) { // only compare 16 bits here
@@ -494,7 +500,7 @@ func readDataDescriptor(r io.Reader, f *File) error {
 	return nil
 }
 
-func readDirectoryEnd(r io.ReaderAt, size int64) (dir *directoryEnd, err error) {
+func readDirectoryEnd(r io.ReaderAt, size int64) (dir *directoryEnd, baseOffset int64, err error) {
 	// look for directoryEndSignature in the last 1k, then in the last 65k
 	var buf []byte
 	var directoryEndOffset int64
@@ -504,7 +510,7 @@ func readDirectoryEnd(r io.ReaderAt, size int64) (dir *directoryEnd, err error)
 		}
 		buf = make([]byte, int(bLen))
 		if _, err := r.ReadAt(buf, size-bLen); err != nil && err != io.EOF {
-			return nil, err
+			return nil, 0, err
 		}
 		if p := findSignatureInBlock(buf); p >= 0 {
 			buf = buf[p:]
@@ -512,7 +518,7 @@ func readDirectoryEnd(r io.ReaderAt, size int64) (dir *directoryEnd, err error)
 			break
 		}
 		if i == 1 || bLen == size {
-			return nil, ErrFormat
+			return nil, 0, ErrFormat
 		}
 	}
 
@@ -529,7 +535,7 @@ func readDirectoryEnd(r io.ReaderAt, size int64) (dir *directoryEnd, err error)
 	}
 	l := int(d.commentLen)
 	if l > len(b) {
-		return nil, errors.New("zip: invalid comment length")
+		return nil, 0, errors.New("zip: invalid comment length")
 	}
 	d.comment = string(b[:l])
 
@@ -537,17 +543,21 @@ func readDirectoryEnd(r io.ReaderAt, size int64) (dir *directoryEnd, err error)
 	if d.directoryRecords == 0xffff || d.directorySize == 0xffff || d.directoryOffset == 0xffffffff {
 		p, err := findDirectory64End(r, directoryEndOffset)
 		if err == nil && p >= 0 {
+			directoryEndOffset = p
 			err = readDirectory64End(r, p, d)
 		}
 		if err != nil {
-			return nil, err
+			return nil, 0, err
 		}
 	}
+
+	baseOffset = directoryEndOffset - int64(d.directorySize) - int64(d.directoryOffset)
+
 	// Make sure directoryOffset points to somewhere in our file.
-	if o := int64(d.directoryOffset); o < 0 || o >= size {
-		return nil, ErrFormat
+	if o := baseOffset + int64(d.directoryOffset); o < 0 || o >= size {
+		return nil, 0, ErrFormat
 	}
-	return d, nil
+	return d, baseOffset, nil
 }
 
 // findDirectory64End tries to read the zip64 locator just before the
diff --git a/src/archive/zip/reader_test.go b/src/archive/zip/reader_test.go
index fd0a171304..4c1e82b9d4 100644
--- a/src/archive/zip/reader_test.go
+++ b/src/archive/zip/reader_test.go
@@ -91,6 +91,24 @@ var tests = []ZipTest{
 		},
 	},
 	{
+		Name:    "test-prefix.zip",
+		Comment: "This is a zipfile comment.",
+		File: []ZipTestFile{
+			{
+				Name:     "test.txt",
+				Content:  []byte("This is a test text file.\n"),
+				Modified: time.Date(2010, 9, 5, 12, 12, 1, 0, timeZone(+10*time.Hour)),
+				Mode:     0644,
+			},
+			{
+				Name:     "gophercolor16x16.png",
+				File:     "gophercolor16x16.png",
+				Modified: time.Date(2010, 9, 5, 15, 52, 58, 0, timeZone(+10*time.Hour)),
+				Mode:     0644,
+			},
+		},
+	},
+	{
 		Name:   "r.zip",
 		Source: returnRecursiveZip,
 		File: []ZipTestFile{
@@ -1011,7 +1029,7 @@ func TestIssue10957(t *testing.T) {
 		"\x00\x00\x00\x00\x0000000000\x00\x00\x00\x00000" +
 		"00000000PK\x01\x0200000000" +
 		"0000000000000000\v\x00\x00\x00" +
-		"\x00\x0000PK\x05\x06000000\x05\x000000" +
+		"\x00\x0000PK\x05\x06000000\x05\x00\xfd\x00\x00\x00" +
 		"\v\x00\x00\x00\x00\x00")
 	z, err := NewReader(bytes.NewReader(data), int64(len(data)))
 	if err != nil {
@@ -1056,7 +1074,7 @@ func TestIssue11146(t *testing.T) {
 		"0000000000000000PK\x01\x02" +
 		"0000\b0\b\x00000000000000" +
 		"\x00\x00\x00\x00\x00\x00\x00\x00\x00\x000000PK\x05\x06\x00\x00" +
-		"\x00\x0000\x01\x0000008\x00\x00\x00\x00\x00")
+		"\x00\x0000\x01\x00\x26\x00\x00\x008\x00\x00\x00\x00\x00")
 	z, err := NewReader(bytes.NewReader(data), int64(len(data)))
 	if err != nil {
 		t.Fatal(err)
diff --git a/src/archive/zip/testdata/readme.notzip b/src/archive/zip/testdata/readme.notzip
index 81737275c6..79b1cb6de3 100644
--- a/src/archive/zip/testdata/readme.notzip
+++ b/src/archive/zip/testdata/readme.notzip
diff --git a/src/archive/zip/testdata/test-prefix.zip b/src/archive/zip/testdata/test-prefix.zip
new file mode 100644
index 0000000000..1eabb4861e
--- /dev/null
+++ b/src/archive/zip/testdata/test-prefix.zip