// TestNonRepertoire tests that codes outside of an Encoding's repertoire are
// converted:
// - to the Unicode replacement character '\ufffd' when decoding to UTF-8,
// - to the ASCII substitute character '\x1a' when encoding from UTF-8.
func TestNonRepertoire(t *testing.T) {
testCases := []struct {
e encoding.Encoding
dSrc, eSrc string
}{
{charmap.Windows1252, "\x81", "갂"},
{japanese.EUCJP, "\xfe\xfc", "갂"},
{japanese.ISO2022JP, "\x1b$B\x7e\x7e", "갂"},
{japanese.ShiftJIS, "\xef\xfc", "갂"},
{korean.EUCKR, "\xfe\xfe", "א"},
{simplifiedchinese.GBK, "\xfe\xfe", "갂"},
{simplifiedchinese.HZGB2312, "~{z~", "갂"},
{traditionalchinese.Big5, "\x81\x40", "갂"},
}
for _, tc := range testCases {
for _, direction := range []string{"Decode", "Encode"} {
enc, want, src := (transform.Transformer)(nil), "", ""
if direction == "Decode" {
enc, want, src = tc.e.NewDecoder(), "\ufffd", tc.dSrc
} else {
enc, want, src = tc.e.NewEncoder(), "\x1a", tc.eSrc
}
dst, err := ioutil.ReadAll(transform.NewReader(strings.NewReader(src), enc))
if err != nil {
t.Errorf("%s %v: %v", direction, tc.e, err)
continue
}
if got := string(dst); got != want {
t.Errorf("%s %v:\ngot %q\nwant %q", direction, tc.e, got, want)
continue
}
}
}
}
func TestBasics(t *testing.T) {
for _, tc := range basicTestCases {
for _, direction := range []string{"Decode", "Encode"} {
newTransformer, want, src := (func() transform.Transformer)(nil), "", ""
wPrefix, sPrefix := "", ""
if direction == "Decode" {
newTransformer, want, src = tc.e.NewDecoder, tc.utf8, tc.encoded
wPrefix, sPrefix = "", tc.encPrefix
} else {
newTransformer, want, src = tc.e.NewEncoder, tc.encoded, tc.utf8
wPrefix, sPrefix = tc.encPrefix, ""
}
dst := make([]byte, len(wPrefix)+len(want))
nDst, nSrc, err := newTransformer().Transform(dst, []byte(sPrefix+src), true)
if err != nil {
t.Errorf("%v: %s: %v", tc.e, direction, err)
continue
}
if nDst != len(wPrefix)+len(want) {
t.Errorf("%v: %s: nDst got %d, want %d",
tc.e, direction, nDst, len(wPrefix)+len(want))
continue
}
if nSrc != len(sPrefix)+len(src) {
t.Errorf("%v: %s: nSrc got %d, want %d",
tc.e, direction, nSrc, len(sPrefix)+len(src))
continue
}
if got := string(dst); got != wPrefix+want {
t.Errorf("%v: %s:\ngot %q\nwant %q",
tc.e, direction, got, wPrefix+want)
continue
}
for _, n := range []int{0, 1, 2, 10, 123, 4567} {
input := sPrefix + strings.Repeat(src, n)
sr := strings.NewReader(input)
g, err := ioutil.ReadAll(transform.NewReader(sr, newTransformer()))
if err != nil {
t.Errorf("%v: %s: ReadAll: n=%d: %v", tc.e, direction, n, err)
continue
}
if len(g) == 0 && len(input) == 0 {
// If the input is empty then the output can be empty,
// regardless of whatever wPrefix is.
continue
}
got1, want1 := string(g), wPrefix+strings.Repeat(want, n)
if got1 != want1 {
t.Errorf("%v: %s: ReadAll: n=%d\ngot %q\nwant %q",
tc.e, direction, n, trim(got1), trim(want1))
continue
}
}
}
}
}
func benchmark(b *testing.B, direction string, enc encoding.Encoding) {
_, src, newTransformer, err := load(direction, enc)
if err != nil {
b.Fatal(err)
}
b.SetBytes(int64(len(src)))
b.ResetTimer()
for i := 0; i < b.N; i++ {
r := transform.NewReader(bytes.NewReader(src), newTransformer())
io.Copy(ioutil.Discard, r)
}
}
// TestBig5CircumflexAndMacron tests the special cases listed in
// http://encoding.spec.whatwg.org/#big5
// Note that these special cases aren't preserved by round-tripping through
// decoding and encoding (since
// http://encoding.spec.whatwg.org/index-big5.txt does not have an entry for
// U+0304 or U+030C), so we can't test this in TestBasics.
func TestBig5CircumflexAndMacron(t *testing.T) {
src := "\x88\x5f\x88\x60\x88\x61\x88\x62\x88\x63\x88\x64\x88\x65\x88\x66 " +
"\x88\xa2\x88\xa3\x88\xa4\x88\xa5\x88\xa6"
want := "ÓǑÒ\u00ca\u0304Ế\u00ca\u030cỀÊ " +
"ü\u00ea\u0304ế\u00ea\u030cề"
dst, err := ioutil.ReadAll(transform.NewReader(
strings.NewReader(src), traditionalchinese.Big5.NewDecoder()))
if err != nil {
t.Fatal(err)
}
if got := string(dst); got != want {
t.Fatalf("\ngot %q\nwant %q", got, want)
}
}
// Parse the html while handling the charset
func ParseHTML(r io.Reader, cs string) (*html.Node, error) {
var err error
if cs == "" {
// attempt to guess the charset of the HTML document
r, err = charset.NewReader(r, "")
if err != nil {
return nil, err
}
} else {
// let the user specify the charset
e, name := charset.Lookup(cs)
if name == "" {
return nil, fmt.Errorf("'%s' is not a valid charset", cs)
}
r = transform.NewReader(r, e.NewDecoder())
}
return html.Parse(r)
}
// NewReader returns an io.Reader that converts the content of r to UTF-8.
// It calls DetermineEncoding to find out what r's encoding is.
func NewReader(r io.Reader, contentType string) (io.Reader, error) {
preview := make([]byte, 1024)
n, err := io.ReadFull(r, preview)
switch {
case err == io.ErrUnexpectedEOF:
preview = preview[:n]
r = bytes.NewReader(preview)
case err != nil:
return nil, err
default:
r = io.MultiReader(bytes.NewReader(preview), r)
}
if e, _, _ := DetermineEncoding(preview, contentType); e != encoding.Nop {
r = transform.NewReader(r, e.NewDecoder())
}
return r, nil
}
func TestFiles(t *testing.T) {
for _, dir := range []string{"Decode", "Encode"} {
for _, tf := range testdataFiles {
dst, src, newTransformer, err := load(dir, tf.enc)
if err != nil {
t.Errorf("%s, %s: load: %v", dir, tf.enc, err)
continue
}
buf := bytes.NewBuffer(nil)
r := transform.NewReader(bytes.NewReader(src), newTransformer())
if _, err := io.Copy(buf, r); err != nil {
t.Errorf("%s, %s: copy: %v", dir, tf.enc, err)
continue
}
if !bytes.Equal(buf.Bytes(), dst) {
t.Errorf("%s, %s: transformed bytes did not match golden file", dir, tf.enc)
continue
}
}
}
}
func TestReplacement(t *testing.T) {
for _, direction := range []string{"Decode", "Encode"} {
enc, want := (transform.Transformer)(nil), ""
if direction == "Decode" {
enc = encoding.Replacement.NewDecoder()
want = "\ufffd"
} else {
enc = encoding.Replacement.NewEncoder()
want = "AB\x00CD\ufffdYZ"
}
sr := strings.NewReader("AB\x00CD\x80YZ")
g, err := ioutil.ReadAll(transform.NewReader(sr, enc))
if err != nil {
t.Errorf("%s: ReadAll: %v", direction, err)
continue
}
if got := string(g); got != want {
t.Errorf("%s:\ngot %q\nwant %q", direction, got, want)
continue
}
}
}
func TestReader(t *testing.T) {
switch runtime.GOOS {
case "nacl": // platforms that don't permit direct file system access
t.Skipf("not supported on %q", runtime.GOOS)
}
for _, tc := range sniffTestCases {
content, err := ioutil.ReadFile("testdata/" + tc.filename)
if err != nil {
t.Errorf("%s: error reading file: %v", tc.filename, err)
continue
}
r, err := NewReader(bytes.NewReader(content), tc.declared)
if err != nil {
t.Errorf("%s: error creating reader: %v", tc.filename, err)
continue
}
got, err := ioutil.ReadAll(r)
if err != nil {
t.Errorf("%s: error reading from charset.NewReader: %v", tc.filename, err)
continue
}
e, _ := Lookup(tc.want)
want, err := ioutil.ReadAll(transform.NewReader(bytes.NewReader(content), e.NewDecoder()))
if err != nil {
t.Errorf("%s: error decoding with hard-coded charset name: %v", tc.filename, err)
continue
}
if !bytes.Equal(got, want) {
t.Errorf("%s: got %q, want %q", tc.filename, got, want)
continue
}
}
}
func ExampleDecodeWindows1252() {
sr := strings.NewReader("Gar\xe7on !")
tr := transform.NewReader(sr, charmap.Windows1252.NewDecoder())
io.Copy(os.Stdout, tr)
// Output: Garçon !
}
func transformString(t transform.Transformer, s string) (string, error) {
r := transform.NewReader(strings.NewReader(s), t)
b, err := ioutil.ReadAll(r)
return string(b), err
}