blob: a00f02d8fab698a588872ade0e5693b0d0dc0660 [file] [log] [blame]
// 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 tar
import (
"bytes"
"encoding/hex"
"errors"
"io"
"os"
"path"
"reflect"
"sort"
"strings"
"testing"
"testing/iotest"
"time"
)
func bytediff(a, b []byte) string {
const (
uniqueA = "- "
uniqueB = "+ "
identity = " "
)
var ss []string
sa := strings.Split(strings.TrimSpace(hex.Dump(a)), "\n")
sb := strings.Split(strings.TrimSpace(hex.Dump(b)), "\n")
for len(sa) > 0 && len(sb) > 0 {
if sa[0] == sb[0] {
ss = append(ss, identity+sa[0])
} else {
ss = append(ss, uniqueA+sa[0])
ss = append(ss, uniqueB+sb[0])
}
sa, sb = sa[1:], sb[1:]
}
for len(sa) > 0 {
ss = append(ss, uniqueA+sa[0])
sa = sa[1:]
}
for len(sb) > 0 {
ss = append(ss, uniqueB+sb[0])
sb = sb[1:]
}
return strings.Join(ss, "\n")
}
func TestWriter(t *testing.T) {
type (
testHeader struct { // WriteHeader(hdr) == wantErr
hdr Header
wantErr error
}
testWrite struct { // Write(str) == (wantCnt, wantErr)
str string
wantCnt int
wantErr error
}
testReadFrom struct { // ReadFrom(testFile{ops}) == (wantCnt, wantErr)
ops fileOps
wantCnt int64
wantErr error
}
testClose struct { // Close() == wantErr
wantErr error
}
testFnc interface{} // testHeader | testWrite | testReadFrom | testClose
)
vectors := []struct {
file string // Optional filename of expected output
tests []testFnc
}{{
// The writer test file was produced with this command:
// tar (GNU tar) 1.26
// ln -s small.txt link.txt
// tar -b 1 --format=ustar -c -f writer.tar small.txt small2.txt link.txt
file: "testdata/writer.tar",
tests: []testFnc{
testHeader{Header{
Typeflag: TypeReg,
Name: "small.txt",
Size: 5,
Mode: 0640,
Uid: 73025,
Gid: 5000,
Uname: "dsymonds",
Gname: "eng",
ModTime: time.Unix(1246508266, 0),
}, nil},
testWrite{"Kilts", 5, nil},
testHeader{Header{
Typeflag: TypeReg,
Name: "small2.txt",
Size: 11,
Mode: 0640,
Uid: 73025,
Uname: "dsymonds",
Gname: "eng",
Gid: 5000,
ModTime: time.Unix(1245217492, 0),
}, nil},
testWrite{"Google.com\n", 11, nil},
testHeader{Header{
Typeflag: TypeSymlink,
Name: "link.txt",
Linkname: "small.txt",
Mode: 0777,
Uid: 1000,
Gid: 1000,
Uname: "strings",
Gname: "strings",
ModTime: time.Unix(1314603082, 0),
}, nil},
testWrite{"", 0, nil},
testClose{nil},
},
}, {
// The truncated test file was produced using these commands:
// dd if=/dev/zero bs=1048576 count=16384 > /tmp/16gig.txt
// tar -b 1 -c -f- /tmp/16gig.txt | dd bs=512 count=8 > writer-big.tar
file: "testdata/writer-big.tar",
tests: []testFnc{
testHeader{Header{
Typeflag: TypeReg,
Name: "tmp/16gig.txt",
Size: 16 << 30,
Mode: 0640,
Uid: 73025,
Gid: 5000,
Uname: "dsymonds",
Gname: "eng",
ModTime: time.Unix(1254699560, 0),
Format: FormatGNU,
}, nil},
},
}, {
// This truncated file was produced using this library.
// It was verified to work with GNU tar 1.27.1 and BSD tar 3.1.2.
// dd if=/dev/zero bs=1G count=16 >> writer-big-long.tar
// gnutar -xvf writer-big-long.tar
// bsdtar -xvf writer-big-long.tar
//
// This file is in PAX format.
file: "testdata/writer-big-long.tar",
tests: []testFnc{
testHeader{Header{
Typeflag: TypeReg,
Name: strings.Repeat("longname/", 15) + "16gig.txt",
Size: 16 << 30,
Mode: 0644,
Uid: 1000,
Gid: 1000,
Uname: "guillaume",
Gname: "guillaume",
ModTime: time.Unix(1399583047, 0),
}, nil},
},
}, {
// This file was produced using GNU tar v1.17.
// gnutar -b 4 --format=ustar (longname/)*15 + file.txt
file: "testdata/ustar.tar",
tests: []testFnc{
testHeader{Header{
Typeflag: TypeReg,
Name: strings.Repeat("longname/", 15) + "file.txt",
Size: 6,
Mode: 0644,
Uid: 501,
Gid: 20,
Uname: "shane",
Gname: "staff",
ModTime: time.Unix(1360135598, 0),
}, nil},
testWrite{"hello\n", 6, nil},
testClose{nil},
},
}, {
// This file was produced using GNU tar v1.26:
// echo "Slartibartfast" > file.txt
// ln file.txt hard.txt
// tar -b 1 --format=ustar -c -f hardlink.tar file.txt hard.txt
file: "testdata/hardlink.tar",
tests: []testFnc{
testHeader{Header{
Typeflag: TypeReg,
Name: "file.txt",
Size: 15,
Mode: 0644,
Uid: 1000,
Gid: 100,
Uname: "vbatts",
Gname: "users",
ModTime: time.Unix(1425484303, 0),
}, nil},
testWrite{"Slartibartfast\n", 15, nil},
testHeader{Header{
Typeflag: TypeLink,
Name: "hard.txt",
Linkname: "file.txt",
Mode: 0644,
Uid: 1000,
Gid: 100,
Uname: "vbatts",
Gname: "users",
ModTime: time.Unix(1425484303, 0),
}, nil},
testWrite{"", 0, nil},
testClose{nil},
},
}, {
tests: []testFnc{
testHeader{Header{
Typeflag: TypeReg,
Name: "bad-null.txt",
Xattrs: map[string]string{"null\x00null\x00": "fizzbuzz"},
}, headerError{}},
},
}, {
tests: []testFnc{
testHeader{Header{
Typeflag: TypeReg,
Name: "null\x00.txt",
}, headerError{}},
},
}, {
file: "testdata/pax-records.tar",
tests: []testFnc{
testHeader{Header{
Typeflag: TypeReg,
Name: "file",
Uname: strings.Repeat("long", 10),
PAXRecords: map[string]string{
"path": "FILE", // Should be ignored
"GNU.sparse.map": "0,0", // Should be ignored
"comment": "Hello, 世界",
"GOLANG.pkg": "tar",
},
}, nil},
testClose{nil},
},
}, {
// Craft a theoretically valid PAX archive with global headers.
// The GNU and BSD tar tools do not parse these the same way.
//
// BSD tar v3.1.2 parses and ignores all global headers;
// the behavior is verified by researching the source code.
//
// $ bsdtar -tvf pax-global-records.tar
// ---------- 0 0 0 0 Dec 31 1969 file1
// ---------- 0 0 0 0 Dec 31 1969 file2
// ---------- 0 0 0 0 Dec 31 1969 file3
// ---------- 0 0 0 0 May 13 2014 file4
//
// GNU tar v1.27.1 applies global headers to subsequent records,
// but does not do the following properly:
// * It does not treat an empty record as deletion.
// * It does not use subsequent global headers to update previous ones.
//
// $ gnutar -tvf pax-global-records.tar
// ---------- 0/0 0 2017-07-13 19:40 global1
// ---------- 0/0 0 2017-07-13 19:40 file2
// gnutar: Substituting `.' for empty member name
// ---------- 0/0 0 1969-12-31 16:00
// gnutar: Substituting `.' for empty member name
// ---------- 0/0 0 2014-05-13 09:53
//
// According to the PAX specification, this should have been the result:
// ---------- 0/0 0 2017-07-13 19:40 global1
// ---------- 0/0 0 2017-07-13 19:40 file2
// ---------- 0/0 0 2017-07-13 19:40 file3
// ---------- 0/0 0 2014-05-13 09:53 file4
file: "testdata/pax-global-records.tar",
tests: []testFnc{
testHeader{Header{
Typeflag: TypeXGlobalHeader,
PAXRecords: map[string]string{"path": "global1", "mtime": "1500000000.0"},
}, nil},
testHeader{Header{
Typeflag: TypeReg, Name: "file1",
}, nil},
testHeader{Header{
Typeflag: TypeReg,
Name: "file2",
PAXRecords: map[string]string{"path": "file2"},
}, nil},
testHeader{Header{
Typeflag: TypeXGlobalHeader,
PAXRecords: map[string]string{"path": ""}, // Should delete "path", but keep "mtime"
}, nil},
testHeader{Header{
Typeflag: TypeReg, Name: "file3",
}, nil},
testHeader{Header{
Typeflag: TypeReg,
Name: "file4",
ModTime: time.Unix(1400000000, 0),
PAXRecords: map[string]string{"mtime": "1400000000"},
}, nil},
testClose{nil},
},
}, {
file: "testdata/gnu-utf8.tar",
tests: []testFnc{
testHeader{Header{
Typeflag: TypeReg,
Name: "☺☻☹☺☻☹☺☻☹☺☻☹☺☻☹☺☻☹☺☻☹☺☻☹☺☻☹☺☻☹☺☻☹☺☻☹☺☻☹☺☻☹☺☻☹☺☻☹☺☻☹☺☻☹",
Mode: 0644,
Uid: 1000, Gid: 1000,
Uname: "☺",
Gname: "⚹",
ModTime: time.Unix(0, 0),
Format: FormatGNU,
}, nil},
testClose{nil},
},
}, {
file: "testdata/gnu-not-utf8.tar",
tests: []testFnc{
testHeader{Header{
Typeflag: TypeReg,
Name: "hi\x80\x81\x82\x83bye",
Mode: 0644,
Uid: 1000,
Gid: 1000,
Uname: "rawr",
Gname: "dsnet",
ModTime: time.Unix(0, 0),
Format: FormatGNU,
}, nil},
testClose{nil},
},
// TODO(dsnet): Re-enable this test when adding sparse support.
// See https://golang.org/issue/22735
/*
}, {
file: "testdata/gnu-nil-sparse-data.tar",
tests: []testFnc{
testHeader{Header{
Typeflag: TypeGNUSparse,
Name: "sparse.db",
Size: 1000,
SparseHoles: []sparseEntry{{Offset: 1000, Length: 0}},
}, nil},
testWrite{strings.Repeat("0123456789", 100), 1000, nil},
testClose{},
},
}, {
file: "testdata/gnu-nil-sparse-hole.tar",
tests: []testFnc{
testHeader{Header{
Typeflag: TypeGNUSparse,
Name: "sparse.db",
Size: 1000,
SparseHoles: []sparseEntry{{Offset: 0, Length: 1000}},
}, nil},
testWrite{strings.Repeat("\x00", 1000), 1000, nil},
testClose{},
},
}, {
file: "testdata/pax-nil-sparse-data.tar",
tests: []testFnc{
testHeader{Header{
Typeflag: TypeReg,
Name: "sparse.db",
Size: 1000,
SparseHoles: []sparseEntry{{Offset: 1000, Length: 0}},
}, nil},
testWrite{strings.Repeat("0123456789", 100), 1000, nil},
testClose{},
},
}, {
file: "testdata/pax-nil-sparse-hole.tar",
tests: []testFnc{
testHeader{Header{
Typeflag: TypeReg,
Name: "sparse.db",
Size: 1000,
SparseHoles: []sparseEntry{{Offset: 0, Length: 1000}},
}, nil},
testWrite{strings.Repeat("\x00", 1000), 1000, nil},
testClose{},
},
}, {
file: "testdata/gnu-sparse-big.tar",
tests: []testFnc{
testHeader{Header{
Typeflag: TypeGNUSparse,
Name: "gnu-sparse",
Size: 6e10,
SparseHoles: []sparseEntry{
{Offset: 0e10, Length: 1e10 - 100},
{Offset: 1e10, Length: 1e10 - 100},
{Offset: 2e10, Length: 1e10 - 100},
{Offset: 3e10, Length: 1e10 - 100},
{Offset: 4e10, Length: 1e10 - 100},
{Offset: 5e10, Length: 1e10 - 100},
},
}, nil},
testReadFrom{fileOps{
int64(1e10 - blockSize),
strings.Repeat("\x00", blockSize-100) + strings.Repeat("0123456789", 10),
int64(1e10 - blockSize),
strings.Repeat("\x00", blockSize-100) + strings.Repeat("0123456789", 10),
int64(1e10 - blockSize),
strings.Repeat("\x00", blockSize-100) + strings.Repeat("0123456789", 10),
int64(1e10 - blockSize),
strings.Repeat("\x00", blockSize-100) + strings.Repeat("0123456789", 10),
int64(1e10 - blockSize),
strings.Repeat("\x00", blockSize-100) + strings.Repeat("0123456789", 10),
int64(1e10 - blockSize),
strings.Repeat("\x00", blockSize-100) + strings.Repeat("0123456789", 10),
}, 6e10, nil},
testClose{nil},
},
}, {
file: "testdata/pax-sparse-big.tar",
tests: []testFnc{
testHeader{Header{
Typeflag: TypeReg,
Name: "pax-sparse",
Size: 6e10,
SparseHoles: []sparseEntry{
{Offset: 0e10, Length: 1e10 - 100},
{Offset: 1e10, Length: 1e10 - 100},
{Offset: 2e10, Length: 1e10 - 100},
{Offset: 3e10, Length: 1e10 - 100},
{Offset: 4e10, Length: 1e10 - 100},
{Offset: 5e10, Length: 1e10 - 100},
},
}, nil},
testReadFrom{fileOps{
int64(1e10 - blockSize),
strings.Repeat("\x00", blockSize-100) + strings.Repeat("0123456789", 10),
int64(1e10 - blockSize),
strings.Repeat("\x00", blockSize-100) + strings.Repeat("0123456789", 10),
int64(1e10 - blockSize),
strings.Repeat("\x00", blockSize-100) + strings.Repeat("0123456789", 10),
int64(1e10 - blockSize),
strings.Repeat("\x00", blockSize-100) + strings.Repeat("0123456789", 10),
int64(1e10 - blockSize),
strings.Repeat("\x00", blockSize-100) + strings.Repeat("0123456789", 10),
int64(1e10 - blockSize),
strings.Repeat("\x00", blockSize-100) + strings.Repeat("0123456789", 10),
}, 6e10, nil},
testClose{nil},
},
*/
}, {
file: "testdata/trailing-slash.tar",
tests: []testFnc{
testHeader{Header{Name: strings.Repeat("123456789/", 30)}, nil},
testClose{nil},
},
}, {
// Automatically promote zero value of Typeflag depending on the name.
file: "testdata/file-and-dir.tar",
tests: []testFnc{
testHeader{Header{Name: "small.txt", Size: 5}, nil},
testWrite{"Kilts", 5, nil},
testHeader{Header{Name: "dir/"}, nil},
testClose{nil},
},
}}
equalError := func(x, y error) bool {
_, ok1 := x.(headerError)
_, ok2 := y.(headerError)
if ok1 || ok2 {
return ok1 && ok2
}
return x == y
}
for _, v := range vectors {
t.Run(path.Base(v.file), func(t *testing.T) {
const maxSize = 10 << 10 // 10KiB
buf := new(bytes.Buffer)
tw := NewWriter(iotest.TruncateWriter(buf, maxSize))
for i, tf := range v.tests {
switch tf := tf.(type) {
case testHeader:
err := tw.WriteHeader(&tf.hdr)
if !equalError(err, tf.wantErr) {
t.Fatalf("test %d, WriteHeader() = %v, want %v", i, err, tf.wantErr)
}
case testWrite:
got, err := tw.Write([]byte(tf.str))
if got != tf.wantCnt || !equalError(err, tf.wantErr) {
t.Fatalf("test %d, Write() = (%d, %v), want (%d, %v)", i, got, err, tf.wantCnt, tf.wantErr)
}
case testReadFrom:
f := &testFile{ops: tf.ops}
got, err := tw.readFrom(f)
if _, ok := err.(testError); ok {
t.Errorf("test %d, ReadFrom(): %v", i, err)
} else if got != tf.wantCnt || !equalError(err, tf.wantErr) {
t.Errorf("test %d, ReadFrom() = (%d, %v), want (%d, %v)", i, got, err, tf.wantCnt, tf.wantErr)
}
if len(f.ops) > 0 {
t.Errorf("test %d, expected %d more operations", i, len(f.ops))
}
case testClose:
err := tw.Close()
if !equalError(err, tf.wantErr) {
t.Fatalf("test %d, Close() = %v, want %v", i, err, tf.wantErr)
}
default:
t.Fatalf("test %d, unknown test operation: %T", i, tf)
}
}
if v.file != "" {
want, err := os.ReadFile(v.file)
if err != nil {
t.Fatalf("ReadFile() = %v, want nil", err)
}
got := buf.Bytes()
if !bytes.Equal(want, got) {
t.Fatalf("incorrect result: (-got +want)\n%v", bytediff(got, want))
}
}
})
}
}
func TestPax(t *testing.T) {
// Create an archive with a large name
fileinfo, err := os.Stat("testdata/small.txt")
if err != nil {
t.Fatal(err)
}
hdr, err := FileInfoHeader(fileinfo, "")
if err != nil {
t.Fatalf("os.Stat: %v", err)
}
// Force a PAX long name to be written
longName := strings.Repeat("ab", 100)
contents := strings.Repeat(" ", int(hdr.Size))
hdr.Name = longName
var buf bytes.Buffer
writer := NewWriter(&buf)
if err := writer.WriteHeader(hdr); err != nil {
t.Fatal(err)
}
if _, err = writer.Write([]byte(contents)); err != nil {
t.Fatal(err)
}
if err := writer.Close(); err != nil {
t.Fatal(err)
}
// Simple test to make sure PAX extensions are in effect
if !bytes.Contains(buf.Bytes(), []byte("PaxHeaders.0")) {
t.Fatal("Expected at least one PAX header to be written.")
}
// Test that we can get a long name back out of the archive.
reader := NewReader(&buf)
hdr, err = reader.Next()
if err != nil {
t.Fatal(err)
}
if hdr.Name != longName {
t.Fatal("Couldn't recover long file name")
}
}
func TestPaxSymlink(t *testing.T) {
// Create an archive with a large linkname
fileinfo, err := os.Stat("testdata/small.txt")
if err != nil {
t.Fatal(err)
}
hdr, err := FileInfoHeader(fileinfo, "")
hdr.Typeflag = TypeSymlink
if err != nil {
t.Fatalf("os.Stat:1 %v", err)
}
// Force a PAX long linkname to be written
longLinkname := strings.Repeat("1234567890/1234567890", 10)
hdr.Linkname = longLinkname
hdr.Size = 0
var buf bytes.Buffer
writer := NewWriter(&buf)
if err := writer.WriteHeader(hdr); err != nil {
t.Fatal(err)
}
if err := writer.Close(); err != nil {
t.Fatal(err)
}
// Simple test to make sure PAX extensions are in effect
if !bytes.Contains(buf.Bytes(), []byte("PaxHeaders.0")) {
t.Fatal("Expected at least one PAX header to be written.")
}
// Test that we can get a long name back out of the archive.
reader := NewReader(&buf)
hdr, err = reader.Next()
if err != nil {
t.Fatal(err)
}
if hdr.Linkname != longLinkname {
t.Fatal("Couldn't recover long link name")
}
}
func TestPaxNonAscii(t *testing.T) {
// Create an archive with non ascii. These should trigger a pax header
// because pax headers have a defined utf-8 encoding.
fileinfo, err := os.Stat("testdata/small.txt")
if err != nil {
t.Fatal(err)
}
hdr, err := FileInfoHeader(fileinfo, "")
if err != nil {
t.Fatalf("os.Stat:1 %v", err)
}
// some sample data
chineseFilename := "文件名"
chineseGroupname := "組"
chineseUsername := "用戶名"
hdr.Name = chineseFilename
hdr.Gname = chineseGroupname
hdr.Uname = chineseUsername
contents := strings.Repeat(" ", int(hdr.Size))
var buf bytes.Buffer
writer := NewWriter(&buf)
if err := writer.WriteHeader(hdr); err != nil {
t.Fatal(err)
}
if _, err = writer.Write([]byte(contents)); err != nil {
t.Fatal(err)
}
if err := writer.Close(); err != nil {
t.Fatal(err)
}
// Simple test to make sure PAX extensions are in effect
if !bytes.Contains(buf.Bytes(), []byte("PaxHeaders.0")) {
t.Fatal("Expected at least one PAX header to be written.")
}
// Test that we can get a long name back out of the archive.
reader := NewReader(&buf)
hdr, err = reader.Next()
if err != nil {
t.Fatal(err)
}
if hdr.Name != chineseFilename {
t.Fatal("Couldn't recover unicode name")
}
if hdr.Gname != chineseGroupname {
t.Fatal("Couldn't recover unicode group")
}
if hdr.Uname != chineseUsername {
t.Fatal("Couldn't recover unicode user")
}
}
func TestPaxXattrs(t *testing.T) {
xattrs := map[string]string{
"user.key": "value",
}
// Create an archive with an xattr
fileinfo, err := os.Stat("testdata/small.txt")
if err != nil {
t.Fatal(err)
}
hdr, err := FileInfoHeader(fileinfo, "")
if err != nil {
t.Fatalf("os.Stat: %v", err)
}
contents := "Kilts"
hdr.Xattrs = xattrs
var buf bytes.Buffer
writer := NewWriter(&buf)
if err := writer.WriteHeader(hdr); err != nil {
t.Fatal(err)
}
if _, err = writer.Write([]byte(contents)); err != nil {
t.Fatal(err)
}
if err := writer.Close(); err != nil {
t.Fatal(err)
}
// Test that we can get the xattrs back out of the archive.
reader := NewReader(&buf)
hdr, err = reader.Next()
if err != nil {
t.Fatal(err)
}
if !reflect.DeepEqual(hdr.Xattrs, xattrs) {
t.Fatalf("xattrs did not survive round trip: got %+v, want %+v",
hdr.Xattrs, xattrs)
}
}
func TestPaxHeadersSorted(t *testing.T) {
fileinfo, err := os.Stat("testdata/small.txt")
if err != nil {
t.Fatal(err)
}
hdr, err := FileInfoHeader(fileinfo, "")
if err != nil {
t.Fatalf("os.Stat: %v", err)
}
contents := strings.Repeat(" ", int(hdr.Size))
hdr.Xattrs = map[string]string{
"foo": "foo",
"bar": "bar",
"baz": "baz",
"qux": "qux",
}
var buf bytes.Buffer
writer := NewWriter(&buf)
if err := writer.WriteHeader(hdr); err != nil {
t.Fatal(err)
}
if _, err = writer.Write([]byte(contents)); err != nil {
t.Fatal(err)
}
if err := writer.Close(); err != nil {
t.Fatal(err)
}
// Simple test to make sure PAX extensions are in effect
if !bytes.Contains(buf.Bytes(), []byte("PaxHeaders.0")) {
t.Fatal("Expected at least one PAX header to be written.")
}
// xattr bar should always appear before others
indices := []int{
bytes.Index(buf.Bytes(), []byte("bar=bar")),
bytes.Index(buf.Bytes(), []byte("baz=baz")),
bytes.Index(buf.Bytes(), []byte("foo=foo")),
bytes.Index(buf.Bytes(), []byte("qux=qux")),
}
if !sort.IntsAreSorted(indices) {
t.Fatal("PAX headers are not sorted")
}
}
func TestUSTARLongName(t *testing.T) {
// Create an archive with a path that failed to split with USTAR extension in previous versions.
fileinfo, err := os.Stat("testdata/small.txt")
if err != nil {
t.Fatal(err)
}
hdr, err := FileInfoHeader(fileinfo, "")
hdr.Typeflag = TypeDir
if err != nil {
t.Fatalf("os.Stat:1 %v", err)
}
// Force a PAX long name to be written. The name was taken from a practical example
// that fails and replaced ever char through numbers to anonymize the sample.
longName := "/0000_0000000/00000-000000000/0000_0000000/00000-0000000000000/0000_0000000/00000-0000000-00000000/0000_0000000/00000000/0000_0000000/000/0000_0000000/00000000v00/0000_0000000/000000/0000_0000000/0000000/0000_0000000/00000y-00/0000/0000/00000000/0x000000/"
hdr.Name = longName
hdr.Size = 0
var buf bytes.Buffer
writer := NewWriter(&buf)
if err := writer.WriteHeader(hdr); err != nil {
t.Fatal(err)
}
if err := writer.Close(); err != nil {
t.Fatal(err)
}
// Test that we can get a long name back out of the archive.
reader := NewReader(&buf)
hdr, err = reader.Next()
if err != nil {
t.Fatal(err)
}
if hdr.Name != longName {
t.Fatal("Couldn't recover long name")
}
}
func TestValidTypeflagWithPAXHeader(t *testing.T) {
var buffer bytes.Buffer
tw := NewWriter(&buffer)
fileName := strings.Repeat("ab", 100)
hdr := &Header{
Name: fileName,
Size: 4,
Typeflag: 0,
}
if err := tw.WriteHeader(hdr); err != nil {
t.Fatalf("Failed to write header: %s", err)
}
if _, err := tw.Write([]byte("fooo")); err != nil {
t.Fatalf("Failed to write the file's data: %s", err)
}
tw.Close()
tr := NewReader(&buffer)
for {
header, err := tr.Next()
if err == io.EOF {
break
}
if err != nil {
t.Fatalf("Failed to read header: %s", err)
}
if header.Typeflag != TypeReg {
t.Fatalf("Typeflag should've been %d, found %d", TypeReg, header.Typeflag)
}
}
}
// failOnceWriter fails exactly once and then always reports success.
type failOnceWriter bool
func (w *failOnceWriter) Write(b []byte) (int, error) {
if !*w {
return 0, io.ErrShortWrite
}
*w = true
return len(b), nil
}
func TestWriterErrors(t *testing.T) {
t.Run("HeaderOnly", func(t *testing.T) {
tw := NewWriter(new(bytes.Buffer))
hdr := &Header{Name: "dir/", Typeflag: TypeDir}
if err := tw.WriteHeader(hdr); err != nil {
t.Fatalf("WriteHeader() = %v, want nil", err)
}
if _, err := tw.Write([]byte{0x00}); err != ErrWriteTooLong {
t.Fatalf("Write() = %v, want %v", err, ErrWriteTooLong)
}
})
t.Run("NegativeSize", func(t *testing.T) {
tw := NewWriter(new(bytes.Buffer))
hdr := &Header{Name: "small.txt", Size: -1}
if err := tw.WriteHeader(hdr); err == nil {
t.Fatalf("WriteHeader() = nil, want non-nil error")
}
})
t.Run("BeforeHeader", func(t *testing.T) {
tw := NewWriter(new(bytes.Buffer))
if _, err := tw.Write([]byte("Kilts")); err != ErrWriteTooLong {
t.Fatalf("Write() = %v, want %v", err, ErrWriteTooLong)
}
})
t.Run("AfterClose", func(t *testing.T) {
tw := NewWriter(new(bytes.Buffer))
hdr := &Header{Name: "small.txt"}
if err := tw.WriteHeader(hdr); err != nil {
t.Fatalf("WriteHeader() = %v, want nil", err)
}
if err := tw.Close(); err != nil {
t.Fatalf("Close() = %v, want nil", err)
}
if _, err := tw.Write([]byte("Kilts")); err != ErrWriteAfterClose {
t.Fatalf("Write() = %v, want %v", err, ErrWriteAfterClose)
}
if err := tw.Flush(); err != ErrWriteAfterClose {
t.Fatalf("Flush() = %v, want %v", err, ErrWriteAfterClose)
}
if err := tw.Close(); err != nil {
t.Fatalf("Close() = %v, want nil", err)
}
})
t.Run("PrematureFlush", func(t *testing.T) {
tw := NewWriter(new(bytes.Buffer))
hdr := &Header{Name: "small.txt", Size: 5}
if err := tw.WriteHeader(hdr); err != nil {
t.Fatalf("WriteHeader() = %v, want nil", err)
}
if err := tw.Flush(); err == nil {
t.Fatalf("Flush() = %v, want non-nil error", err)
}
})
t.Run("PrematureClose", func(t *testing.T) {
tw := NewWriter(new(bytes.Buffer))
hdr := &Header{Name: "small.txt", Size: 5}
if err := tw.WriteHeader(hdr); err != nil {
t.Fatalf("WriteHeader() = %v, want nil", err)
}
if err := tw.Close(); err == nil {
t.Fatalf("Close() = %v, want non-nil error", err)
}
})
t.Run("Persistence", func(t *testing.T) {
tw := NewWriter(new(failOnceWriter))
if err := tw.WriteHeader(&Header{}); err != io.ErrShortWrite {
t.Fatalf("WriteHeader() = %v, want %v", err, io.ErrShortWrite)
}
if err := tw.WriteHeader(&Header{Name: "small.txt"}); err == nil {
t.Errorf("WriteHeader() = got %v, want non-nil error", err)
}
if _, err := tw.Write(nil); err == nil {
t.Errorf("Write() = %v, want non-nil error", err)
}
if err := tw.Flush(); err == nil {
t.Errorf("Flush() = %v, want non-nil error", err)
}
if err := tw.Close(); err == nil {
t.Errorf("Close() = %v, want non-nil error", err)
}
})
}
func TestSplitUSTARPath(t *testing.T) {
sr := strings.Repeat
vectors := []struct {
input string // Input path
prefix string // Expected output prefix
suffix string // Expected output suffix
ok bool // Split success?
}{
{"", "", "", false},
{"abc", "", "", false},
{"用戶名", "", "", false},
{sr("a", nameSize), "", "", false},
{sr("a", nameSize) + "/", "", "", false},
{sr("a", nameSize) + "/a", sr("a", nameSize), "a", true},
{sr("a", prefixSize) + "/", "", "", false},
{sr("a", prefixSize) + "/a", sr("a", prefixSize), "a", true},
{sr("a", nameSize+1), "", "", false},
{sr("/", nameSize+1), sr("/", nameSize-1), "/", true},
{sr("a", prefixSize) + "/" + sr("b", nameSize),
sr("a", prefixSize), sr("b", nameSize), true},
{sr("a", prefixSize) + "//" + sr("b", nameSize), "", "", false},
{sr("a/", nameSize), sr("a/", 77) + "a", sr("a/", 22), true},
}
for _, v := range vectors {
prefix, suffix, ok := splitUSTARPath(v.input)
if prefix != v.prefix || suffix != v.suffix || ok != v.ok {
t.Errorf("splitUSTARPath(%q):\ngot (%q, %q, %v)\nwant (%q, %q, %v)",
v.input, prefix, suffix, ok, v.prefix, v.suffix, v.ok)
}
}
}
// TestIssue12594 tests that the Writer does not attempt to populate the prefix
// field when encoding a header in the GNU format. The prefix field is valid
// in USTAR and PAX, but not GNU.
func TestIssue12594(t *testing.T) {
names := []string{
"0/1/2/3/4/5/6/7/8/9/10/11/12/13/14/15/16/17/18/19/20/21/22/23/24/25/26/27/28/29/30/file.txt",
"0/1/2/3/4/5/6/7/8/9/10/11/12/13/14/15/16/17/18/19/20/21/22/23/24/25/26/27/28/29/30/31/32/33/file.txt",
"0/1/2/3/4/5/6/7/8/9/10/11/12/13/14/15/16/17/18/19/20/21/22/23/24/25/26/27/28/29/30/31/32/333/file.txt",
"0/1/2/3/4/5/6/7/8/9/10/11/12/13/14/15/16/17/18/19/20/21/22/23/24/25/26/27/28/29/30/31/32/33/34/35/36/37/38/39/40/file.txt",
"0000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000/file.txt",
"/home/support/.openoffice.org/3/user/uno_packages/cache/registry/com.sun.star.comp.deployment.executable.PackageRegistryBackend",
}
for i, name := range names {
var b bytes.Buffer
tw := NewWriter(&b)
if err := tw.WriteHeader(&Header{
Name: name,
Uid: 1 << 25, // Prevent USTAR format
}); err != nil {
t.Errorf("test %d, unexpected WriteHeader error: %v", i, err)
}
if err := tw.Close(); err != nil {
t.Errorf("test %d, unexpected Close error: %v", i, err)
}
// The prefix field should never appear in the GNU format.
var blk block
copy(blk[:], b.Bytes())
prefix := string(blk.USTAR().Prefix())
if i := strings.IndexByte(prefix, 0); i >= 0 {
prefix = prefix[:i] // Truncate at the NUL terminator
}
if blk.GetFormat() == FormatGNU && len(prefix) > 0 && strings.HasPrefix(name, prefix) {
t.Errorf("test %d, found prefix in GNU format: %s", i, prefix)
}
tr := NewReader(&b)
hdr, err := tr.Next()
if err != nil {
t.Errorf("test %d, unexpected Next error: %v", i, err)
}
if hdr.Name != name {
t.Errorf("test %d, hdr.Name = %s, want %s", i, hdr.Name, name)
}
}
}
// testNonEmptyWriter wraps an io.Writer and ensures that
// Write is never called with an empty buffer.
type testNonEmptyWriter struct{ io.Writer }
func (w testNonEmptyWriter) Write(b []byte) (int, error) {
if len(b) == 0 {
return 0, errors.New("unexpected empty Write call")
}
return w.Writer.Write(b)
}
func TestFileWriter(t *testing.T) {
type (
testWrite struct { // Write(str) == (wantCnt, wantErr)
str string
wantCnt int
wantErr error
}
testReadFrom struct { // ReadFrom(testFile{ops}) == (wantCnt, wantErr)
ops fileOps
wantCnt int64
wantErr error
}
testRemaining struct { // LogicalRemaining() == wantLCnt, PhysicalRemaining() == wantPCnt
wantLCnt int64
wantPCnt int64
}
testFnc interface{} // testWrite | testReadFrom | testRemaining
)
type (
makeReg struct {
size int64
wantStr string
}
makeSparse struct {
makeReg makeReg
sph sparseHoles
size int64
}
fileMaker interface{} // makeReg | makeSparse
)
vectors := []struct {
maker fileMaker
tests []testFnc
}{{
maker: makeReg{0, ""},
tests: []testFnc{
testRemaining{0, 0},
testWrite{"", 0, nil},
testWrite{"a", 0, ErrWriteTooLong},
testReadFrom{fileOps{""}, 0, nil},
testReadFrom{fileOps{"a"}, 0, ErrWriteTooLong},
testRemaining{0, 0},
},
}, {
maker: makeReg{1, "a"},
tests: []testFnc{
testRemaining{1, 1},
testWrite{"", 0, nil},
testWrite{"a", 1, nil},
testWrite{"bcde", 0, ErrWriteTooLong},
testWrite{"", 0, nil},
testReadFrom{fileOps{""}, 0, nil},
testReadFrom{fileOps{"a"}, 0, ErrWriteTooLong},
testRemaining{0, 0},
},
}, {
maker: makeReg{5, "hello"},
tests: []testFnc{
testRemaining{5, 5},
testWrite{"hello", 5, nil},
testRemaining{0, 0},
},
}, {
maker: makeReg{5, "\x00\x00\x00\x00\x00"},
tests: []testFnc{
testRemaining{5, 5},
testReadFrom{fileOps{"\x00\x00\x00\x00\x00"}, 5, nil},
testRemaining{0, 0},
},
}, {
maker: makeReg{5, "\x00\x00\x00\x00\x00"},
tests: []testFnc{
testRemaining{5, 5},
testReadFrom{fileOps{"\x00\x00\x00\x00\x00extra"}, 5, ErrWriteTooLong},
testRemaining{0, 0},
},
}, {
maker: makeReg{5, "abc\x00\x00"},
tests: []testFnc{
testRemaining{5, 5},
testWrite{"abc", 3, nil},
testRemaining{2, 2},
testReadFrom{fileOps{"\x00\x00"}, 2, nil},
testRemaining{0, 0},
},
}, {
maker: makeReg{5, "\x00\x00abc"},
tests: []testFnc{
testRemaining{5, 5},
testWrite{"\x00\x00", 2, nil},
testRemaining{3, 3},
testWrite{"abc", 3, nil},
testReadFrom{fileOps{"z"}, 0, ErrWriteTooLong},
testWrite{"z", 0, ErrWriteTooLong},
testRemaining{0, 0},
},
}, {
maker: makeSparse{makeReg{5, "abcde"}, sparseHoles{{2, 3}}, 8},
tests: []testFnc{
testRemaining{8, 5},
testWrite{"ab\x00\x00\x00cde", 8, nil},
testWrite{"a", 0, ErrWriteTooLong},
testRemaining{0, 0},
},
}, {
maker: makeSparse{makeReg{5, "abcde"}, sparseHoles{{2, 3}}, 8},
tests: []testFnc{
testWrite{"ab\x00\x00\x00cdez", 8, ErrWriteTooLong},
testRemaining{0, 0},
},
}, {
maker: makeSparse{makeReg{5, "abcde"}, sparseHoles{{2, 3}}, 8},
tests: []testFnc{
testWrite{"ab\x00", 3, nil},
testRemaining{5, 3},
testWrite{"\x00\x00cde", 5, nil},
testWrite{"a", 0, ErrWriteTooLong},
testRemaining{0, 0},
},
}, {
maker: makeSparse{makeReg{5, "abcde"}, sparseHoles{{2, 3}}, 8},
tests: []testFnc{
testWrite{"ab", 2, nil},
testRemaining{6, 3},
testReadFrom{fileOps{int64(3), "cde"}, 6, nil},
testRemaining{0, 0},
},
}, {
maker: makeSparse{makeReg{5, "abcde"}, sparseHoles{{2, 3}}, 8},
tests: []testFnc{
testReadFrom{fileOps{"ab", int64(3), "cde"}, 8, nil},
testRemaining{0, 0},
},
}, {
maker: makeSparse{makeReg{5, "abcde"}, sparseHoles{{2, 3}}, 8},
tests: []testFnc{
testReadFrom{fileOps{"ab", int64(3), "cdeX"}, 8, ErrWriteTooLong},
testRemaining{0, 0},
},
}, {
maker: makeSparse{makeReg{4, "abcd"}, sparseHoles{{2, 3}}, 8},
tests: []testFnc{
testReadFrom{fileOps{"ab", int64(3), "cd"}, 7, io.ErrUnexpectedEOF},
testRemaining{1, 0},
},
}, {
maker: makeSparse{makeReg{4, "abcd"}, sparseHoles{{2, 3}}, 8},
tests: []testFnc{
testReadFrom{fileOps{"ab", int64(3), "cde"}, 7, errMissData},
testRemaining{1, 0},
},
}, {
maker: makeSparse{makeReg{6, "abcde"}, sparseHoles{{2, 3}}, 8},
tests: []testFnc{
testReadFrom{fileOps{"ab", int64(3), "cde"}, 8, errUnrefData},
testRemaining{0, 1},
},
}, {
maker: makeSparse{makeReg{4, "abcd"}, sparseHoles{{2, 3}}, 8},
tests: []testFnc{
testWrite{"ab", 2, nil},
testRemaining{6, 2},
testWrite{"\x00\x00\x00", 3, nil},
testRemaining{3, 2},
testWrite{"cde", 2, errMissData},
testRemaining{1, 0},
},
}, {
maker: makeSparse{makeReg{6, "abcde"}, sparseHoles{{2, 3}}, 8},
tests: []testFnc{
testWrite{"ab", 2, nil},
testRemaining{6, 4},
testWrite{"\x00\x00\x00", 3, nil},
testRemaining{3, 4},
testWrite{"cde", 3, errUnrefData},
testRemaining{0, 1},
},
}, {
maker: makeSparse{makeReg{3, "abc"}, sparseHoles{{0, 2}, {5, 2}}, 7},
tests: []testFnc{
testRemaining{7, 3},
testWrite{"\x00\x00abc\x00\x00", 7, nil},
testRemaining{0, 0},
},
}, {
maker: makeSparse{makeReg{3, "abc"}, sparseHoles{{0, 2}, {5, 2}}, 7},
tests: []testFnc{
testRemaining{7, 3},
testReadFrom{fileOps{int64(2), "abc", int64(1), "\x00"}, 7, nil},
testRemaining{0, 0},
},
}, {
maker: makeSparse{makeReg{3, ""}, sparseHoles{{0, 2}, {5, 2}}, 7},
tests: []testFnc{
testWrite{"abcdefg", 0, errWriteHole},
},
}, {
maker: makeSparse{makeReg{3, "abc"}, sparseHoles{{0, 2}, {5, 2}}, 7},
tests: []testFnc{
testWrite{"\x00\x00abcde", 5, errWriteHole},
},
}, {
maker: makeSparse{makeReg{3, "abc"}, sparseHoles{{0, 2}, {5, 2}}, 7},
tests: []testFnc{
testWrite{"\x00\x00abc\x00\x00z", 7, ErrWriteTooLong},
testRemaining{0, 0},
},
}, {
maker: makeSparse{makeReg{3, "abc"}, sparseHoles{{0, 2}, {5, 2}}, 7},
tests: []testFnc{
testWrite{"\x00\x00", 2, nil},
testRemaining{5, 3},
testWrite{"abc", 3, nil},
testRemaining{2, 0},
testWrite{"\x00\x00", 2, nil},
testRemaining{0, 0},
},
}, {
maker: makeSparse{makeReg{2, "ab"}, sparseHoles{{0, 2}, {5, 2}}, 7},
tests: []testFnc{
testWrite{"\x00\x00", 2, nil},
testWrite{"abc", 2, errMissData},
testWrite{"\x00\x00", 0, errMissData},
},
}, {
maker: makeSparse{makeReg{4, "abc"}, sparseHoles{{0, 2}, {5, 2}}, 7},
tests: []testFnc{
testWrite{"\x00\x00", 2, nil},
testWrite{"abc", 3, nil},
testWrite{"\x00\x00", 2, errUnrefData},
},
}}
for i, v := range vectors {
var wantStr string
bb := new(bytes.Buffer)
w := testNonEmptyWriter{bb}
var fw fileWriter
switch maker := v.maker.(type) {
case makeReg:
fw = &regFileWriter{w, maker.size}
wantStr = maker.wantStr
case makeSparse:
if !validateSparseEntries(maker.sph, maker.size) {
t.Fatalf("invalid sparse map: %v", maker.sph)
}
spd := invertSparseEntries(maker.sph, maker.size)
fw = &regFileWriter{w, maker.makeReg.size}
fw = &sparseFileWriter{fw, spd, 0}
wantStr = maker.makeReg.wantStr
default:
t.Fatalf("test %d, unknown make operation: %T", i, maker)
}
for j, tf := range v.tests {
switch tf := tf.(type) {
case testWrite:
got, err := fw.Write([]byte(tf.str))
if got != tf.wantCnt || err != tf.wantErr {
t.Errorf("test %d.%d, Write(%s):\ngot (%d, %v)\nwant (%d, %v)", i, j, tf.str, got, err, tf.wantCnt, tf.wantErr)
}
case testReadFrom:
f := &testFile{ops: tf.ops}
got, err := fw.ReadFrom(f)
if _, ok := err.(testError); ok {
t.Errorf("test %d.%d, ReadFrom(): %v", i, j, err)
} else if got != tf.wantCnt || err != tf.wantErr {
t.Errorf("test %d.%d, ReadFrom() = (%d, %v), want (%d, %v)", i, j, got, err, tf.wantCnt, tf.wantErr)
}
if len(f.ops) > 0 {
t.Errorf("test %d.%d, expected %d more operations", i, j, len(f.ops))
}
case testRemaining:
if got := fw.LogicalRemaining(); got != tf.wantLCnt {
t.Errorf("test %d.%d, LogicalRemaining() = %d, want %d", i, j, got, tf.wantLCnt)
}
if got := fw.PhysicalRemaining(); got != tf.wantPCnt {
t.Errorf("test %d.%d, PhysicalRemaining() = %d, want %d", i, j, got, tf.wantPCnt)
}
default:
t.Fatalf("test %d.%d, unknown test operation: %T", i, j, tf)
}
}
if got := bb.String(); got != wantStr {
t.Fatalf("test %d, String() = %q, want %q", i, got, wantStr)
}
}
}