diff options
author | Daniel Baumann <daniel.baumann@progress-linux.org> | 2024-04-16 19:19:13 +0000 |
---|---|---|
committer | Daniel Baumann <daniel.baumann@progress-linux.org> | 2024-04-16 19:19:13 +0000 |
commit | ccd992355df7192993c666236047820244914598 (patch) | |
tree | f00fea65147227b7743083c6148396f74cd66935 /src/encoding/ascii85 | |
parent | Initial commit. (diff) | |
download | golang-1.21-ccd992355df7192993c666236047820244914598.tar.xz golang-1.21-ccd992355df7192993c666236047820244914598.zip |
Adding upstream version 1.21.8.upstream/1.21.8
Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>
Diffstat (limited to 'src/encoding/ascii85')
-rw-r--r-- | src/encoding/ascii85/ascii85.go | 307 | ||||
-rw-r--r-- | src/encoding/ascii85/ascii85_test.go | 214 |
2 files changed, 521 insertions, 0 deletions
diff --git a/src/encoding/ascii85/ascii85.go b/src/encoding/ascii85/ascii85.go new file mode 100644 index 0000000..1f1fb00 --- /dev/null +++ b/src/encoding/ascii85/ascii85.go @@ -0,0 +1,307 @@ +// 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 ascii85 implements the ascii85 data encoding +// as used in the btoa tool and Adobe's PostScript and PDF document formats. +package ascii85 + +import ( + "io" + "strconv" +) + +/* + * Encoder + */ + +// Encode encodes src into at most MaxEncodedLen(len(src)) +// bytes of dst, returning the actual number of bytes written. +// +// The encoding handles 4-byte chunks, using a special encoding +// for the last fragment, so Encode is not appropriate for use on +// individual blocks of a large data stream. Use NewEncoder() instead. +// +// Often, ascii85-encoded data is wrapped in <~ and ~> symbols. +// Encode does not add these. +func Encode(dst, src []byte) int { + if len(src) == 0 { + return 0 + } + + n := 0 + for len(src) > 0 { + dst[0] = 0 + dst[1] = 0 + dst[2] = 0 + dst[3] = 0 + dst[4] = 0 + + // Unpack 4 bytes into uint32 to repack into base 85 5-byte. + var v uint32 + switch len(src) { + default: + v |= uint32(src[3]) + fallthrough + case 3: + v |= uint32(src[2]) << 8 + fallthrough + case 2: + v |= uint32(src[1]) << 16 + fallthrough + case 1: + v |= uint32(src[0]) << 24 + } + + // Special case: zero (!!!!!) shortens to z. + if v == 0 && len(src) >= 4 { + dst[0] = 'z' + dst = dst[1:] + src = src[4:] + n++ + continue + } + + // Otherwise, 5 base 85 digits starting at !. + for i := 4; i >= 0; i-- { + dst[i] = '!' + byte(v%85) + v /= 85 + } + + // If src was short, discard the low destination bytes. + m := 5 + if len(src) < 4 { + m -= 4 - len(src) + src = nil + } else { + src = src[4:] + } + dst = dst[m:] + n += m + } + return n +} + +// MaxEncodedLen returns the maximum length of an encoding of n source bytes. +func MaxEncodedLen(n int) int { return (n + 3) / 4 * 5 } + +// NewEncoder returns a new ascii85 stream encoder. Data written to +// the returned writer will be encoded and then written to w. +// Ascii85 encodings operate in 32-bit blocks; when finished +// writing, the caller must Close the returned encoder to flush any +// trailing partial block. +func NewEncoder(w io.Writer) io.WriteCloser { return &encoder{w: w} } + +type encoder struct { + err error + w io.Writer + buf [4]byte // buffered data waiting to be encoded + nbuf int // number of bytes in buf + out [1024]byte // output buffer +} + +func (e *encoder) Write(p []byte) (n int, err error) { + if e.err != nil { + return 0, e.err + } + + // Leading fringe. + if e.nbuf > 0 { + var i int + for i = 0; i < len(p) && e.nbuf < 4; i++ { + e.buf[e.nbuf] = p[i] + e.nbuf++ + } + n += i + p = p[i:] + if e.nbuf < 4 { + return + } + nout := Encode(e.out[0:], e.buf[0:]) + if _, e.err = e.w.Write(e.out[0:nout]); e.err != nil { + return n, e.err + } + e.nbuf = 0 + } + + // Large interior chunks. + for len(p) >= 4 { + nn := len(e.out) / 5 * 4 + if nn > len(p) { + nn = len(p) + } + nn -= nn % 4 + if nn > 0 { + nout := Encode(e.out[0:], p[0:nn]) + if _, e.err = e.w.Write(e.out[0:nout]); e.err != nil { + return n, e.err + } + } + n += nn + p = p[nn:] + } + + // Trailing fringe. + copy(e.buf[:], p) + e.nbuf = len(p) + n += len(p) + return +} + +// Close flushes any pending output from the encoder. +// It is an error to call Write after calling Close. +func (e *encoder) Close() error { + // If there's anything left in the buffer, flush it out + if e.err == nil && e.nbuf > 0 { + nout := Encode(e.out[0:], e.buf[0:e.nbuf]) + e.nbuf = 0 + _, e.err = e.w.Write(e.out[0:nout]) + } + return e.err +} + +/* + * Decoder + */ + +type CorruptInputError int64 + +func (e CorruptInputError) Error() string { + return "illegal ascii85 data at input byte " + strconv.FormatInt(int64(e), 10) +} + +// Decode decodes src into dst, returning both the number +// of bytes written to dst and the number consumed from src. +// If src contains invalid ascii85 data, Decode will return the +// number of bytes successfully written and a CorruptInputError. +// Decode ignores space and control characters in src. +// Often, ascii85-encoded data is wrapped in <~ and ~> symbols. +// Decode expects these to have been stripped by the caller. +// +// If flush is true, Decode assumes that src represents the +// end of the input stream and processes it completely rather +// than wait for the completion of another 32-bit block. +// +// NewDecoder wraps an io.Reader interface around Decode. +func Decode(dst, src []byte, flush bool) (ndst, nsrc int, err error) { + var v uint32 + var nb int + for i, b := range src { + if len(dst)-ndst < 4 { + return + } + switch { + case b <= ' ': + continue + case b == 'z' && nb == 0: + nb = 5 + v = 0 + case '!' <= b && b <= 'u': + v = v*85 + uint32(b-'!') + nb++ + default: + return 0, 0, CorruptInputError(i) + } + if nb == 5 { + nsrc = i + 1 + dst[ndst] = byte(v >> 24) + dst[ndst+1] = byte(v >> 16) + dst[ndst+2] = byte(v >> 8) + dst[ndst+3] = byte(v) + ndst += 4 + nb = 0 + v = 0 + } + } + if flush { + nsrc = len(src) + if nb > 0 { + // The number of output bytes in the last fragment + // is the number of leftover input bytes - 1: + // the extra byte provides enough bits to cover + // the inefficiency of the encoding for the block. + if nb == 1 { + return 0, 0, CorruptInputError(len(src)) + } + for i := nb; i < 5; i++ { + // The short encoding truncated the output value. + // We have to assume the worst case values (digit 84) + // in order to ensure that the top bits are correct. + v = v*85 + 84 + } + for i := 0; i < nb-1; i++ { + dst[ndst] = byte(v >> 24) + v <<= 8 + ndst++ + } + } + } + return +} + +// NewDecoder constructs a new ascii85 stream decoder. +func NewDecoder(r io.Reader) io.Reader { return &decoder{r: r} } + +type decoder struct { + err error + readErr error + r io.Reader + buf [1024]byte // leftover input + nbuf int + out []byte // leftover decoded output + outbuf [1024]byte +} + +func (d *decoder) Read(p []byte) (n int, err error) { + if len(p) == 0 { + return 0, nil + } + if d.err != nil { + return 0, d.err + } + + for { + // Copy leftover output from last decode. + if len(d.out) > 0 { + n = copy(p, d.out) + d.out = d.out[n:] + return + } + + // Decode leftover input from last read. + var nn, nsrc, ndst int + if d.nbuf > 0 { + ndst, nsrc, d.err = Decode(d.outbuf[0:], d.buf[0:d.nbuf], d.readErr != nil) + if ndst > 0 { + d.out = d.outbuf[0:ndst] + d.nbuf = copy(d.buf[0:], d.buf[nsrc:d.nbuf]) + continue // copy out and return + } + if ndst == 0 && d.err == nil { + // Special case: input buffer is mostly filled with non-data bytes. + // Filter out such bytes to make room for more input. + off := 0 + for i := 0; i < d.nbuf; i++ { + if d.buf[i] > ' ' { + d.buf[off] = d.buf[i] + off++ + } + } + d.nbuf = off + } + } + + // Out of input, out of decoded output. Check errors. + if d.err != nil { + return 0, d.err + } + if d.readErr != nil { + d.err = d.readErr + return 0, d.err + } + + // Read more data. + nn, d.readErr = d.r.Read(d.buf[d.nbuf:]) + d.nbuf += nn + } +} diff --git a/src/encoding/ascii85/ascii85_test.go b/src/encoding/ascii85/ascii85_test.go new file mode 100644 index 0000000..578829e --- /dev/null +++ b/src/encoding/ascii85/ascii85_test.go @@ -0,0 +1,214 @@ +// 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 ascii85 + +import ( + "bytes" + "io" + "strings" + "testing" +) + +type testpair struct { + decoded, encoded string +} + +var bigtest = testpair{ + "Man is distinguished, not only by his reason, but by this singular passion from " + + "other animals, which is a lust of the mind, that by a perseverance of delight in " + + "the continued and indefatigable generation of knowledge, exceeds the short " + + "vehemence of any carnal pleasure.", + "9jqo^BlbD-BleB1DJ+*+F(f,q/0JhKF<GL>Cj@.4Gp$d7F!,L7@<6@)/0JDEF<G%<+EV:2F!,\n" + + "O<DJ+*.@<*K0@<6L(Df-\\0Ec5e;DffZ(EZee.Bl.9pF\"AGXBPCsi+DGm>@3BB/F*&OCAfu2/AKY\n" + + "i(DIb:@FD,*)+C]U=@3BN#EcYf8ATD3s@q?d$AftVqCh[NqF<G:8+EV:.+Cf>-FD5W8ARlolDIa\n" + + "l(DId<j@<?3r@:F%a+D58'ATD4$Bl@l3De:,-DJs`8ARoFb/0JMK@qB4^F!,R<AKZ&-DfTqBG%G\n" + + ">uD.RTpAKYo'+CT/5+Cei#DII?(E,9)oF*2M7/c\n", +} + +var pairs = []testpair{ + // Encode returns 0 when len(src) is 0 + { + "", + "", + }, + // Wikipedia example + bigtest, + // Special case when shortening !!!!! to z. + { + "\000\000\000\000", + "z", + }, +} + +func testEqual(t *testing.T, msg string, args ...any) bool { + t.Helper() + if args[len(args)-2] != args[len(args)-1] { + t.Errorf(msg, args...) + return false + } + return true +} + +func strip85(s string) string { + t := make([]byte, len(s)) + w := 0 + for r := 0; r < len(s); r++ { + c := s[r] + if c > ' ' { + t[w] = c + w++ + } + } + return string(t[0:w]) +} + +func TestEncode(t *testing.T) { + for _, p := range pairs { + buf := make([]byte, MaxEncodedLen(len(p.decoded))) + n := Encode(buf, []byte(p.decoded)) + buf = buf[0:n] + testEqual(t, "Encode(%q) = %q, want %q", p.decoded, strip85(string(buf)), strip85(p.encoded)) + } +} + +func TestEncoder(t *testing.T) { + for _, p := range pairs { + bb := &strings.Builder{} + encoder := NewEncoder(bb) + encoder.Write([]byte(p.decoded)) + encoder.Close() + testEqual(t, "Encode(%q) = %q, want %q", p.decoded, strip85(bb.String()), strip85(p.encoded)) + } +} + +func TestEncoderBuffering(t *testing.T) { + input := []byte(bigtest.decoded) + for bs := 1; bs <= 12; bs++ { + bb := &strings.Builder{} + encoder := NewEncoder(bb) + for pos := 0; pos < len(input); pos += bs { + end := pos + bs + if end > len(input) { + end = len(input) + } + n, err := encoder.Write(input[pos:end]) + testEqual(t, "Write(%q) gave error %v, want %v", input[pos:end], err, error(nil)) + testEqual(t, "Write(%q) gave length %v, want %v", input[pos:end], n, end-pos) + } + err := encoder.Close() + testEqual(t, "Close gave error %v, want %v", err, error(nil)) + testEqual(t, "Encoding/%d of %q = %q, want %q", bs, bigtest.decoded, strip85(bb.String()), strip85(bigtest.encoded)) + } +} + +func TestDecode(t *testing.T) { + for _, p := range pairs { + dbuf := make([]byte, 4*len(p.encoded)) + ndst, nsrc, err := Decode(dbuf, []byte(p.encoded), true) + testEqual(t, "Decode(%q) = error %v, want %v", p.encoded, err, error(nil)) + testEqual(t, "Decode(%q) = nsrc %v, want %v", p.encoded, nsrc, len(p.encoded)) + testEqual(t, "Decode(%q) = ndst %v, want %v", p.encoded, ndst, len(p.decoded)) + testEqual(t, "Decode(%q) = %q, want %q", p.encoded, string(dbuf[0:ndst]), p.decoded) + } +} + +func TestDecoder(t *testing.T) { + for _, p := range pairs { + decoder := NewDecoder(strings.NewReader(p.encoded)) + dbuf, err := io.ReadAll(decoder) + if err != nil { + t.Fatal("Read failed", err) + } + testEqual(t, "Read from %q = length %v, want %v", p.encoded, len(dbuf), len(p.decoded)) + testEqual(t, "Decoding of %q = %q, want %q", p.encoded, string(dbuf), p.decoded) + if err != nil { + testEqual(t, "Read from %q = %v, want %v", p.encoded, err, io.EOF) + } + } +} + +func TestDecoderBuffering(t *testing.T) { + for bs := 1; bs <= 12; bs++ { + decoder := NewDecoder(strings.NewReader(bigtest.encoded)) + buf := make([]byte, len(bigtest.decoded)+12) + var total int + var n int + var err error + for total = 0; total < len(bigtest.decoded) && err == nil; { + n, err = decoder.Read(buf[total : total+bs]) + total += n + } + if err != nil && err != io.EOF { + t.Errorf("Read from %q at pos %d = %d, unexpected error %v", bigtest.encoded, total, n, err) + } + testEqual(t, "Decoding/%d of %q = %q, want %q", bs, bigtest.encoded, string(buf[0:total]), bigtest.decoded) + } +} + +func TestDecodeCorrupt(t *testing.T) { + type corrupt struct { + e string + p int + } + examples := []corrupt{ + {"v", 0}, + {"!z!!!!!!!!!", 1}, + } + + for _, e := range examples { + dbuf := make([]byte, 4*len(e.e)) + _, _, err := Decode(dbuf, []byte(e.e), true) + switch err := err.(type) { + case CorruptInputError: + testEqual(t, "Corruption in %q at offset %v, want %v", e.e, int(err), e.p) + default: + t.Error("Decoder failed to detect corruption in", e) + } + } +} + +func TestBig(t *testing.T) { + n := 3*1000 + 1 + raw := make([]byte, n) + const alpha = "0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ" + for i := 0; i < n; i++ { + raw[i] = alpha[i%len(alpha)] + } + encoded := new(bytes.Buffer) + w := NewEncoder(encoded) + nn, err := w.Write(raw) + if nn != n || err != nil { + t.Fatalf("Encoder.Write(raw) = %d, %v want %d, nil", nn, err, n) + } + err = w.Close() + if err != nil { + t.Fatalf("Encoder.Close() = %v want nil", err) + } + decoded, err := io.ReadAll(NewDecoder(encoded)) + if err != nil { + t.Fatalf("io.ReadAll(NewDecoder(...)): %v", err) + } + + if !bytes.Equal(raw, decoded) { + var i int + for i = 0; i < len(decoded) && i < len(raw); i++ { + if decoded[i] != raw[i] { + break + } + } + t.Errorf("Decode(Encode(%d-byte string)) failed at offset %d", n, i) + } +} + +func TestDecoderInternalWhitespace(t *testing.T) { + s := strings.Repeat(" ", 2048) + "z" + decoded, err := io.ReadAll(NewDecoder(strings.NewReader(s))) + if err != nil { + t.Errorf("Decode gave error %v", err) + } + if want := []byte("\000\000\000\000"); !bytes.Equal(want, decoded) { + t.Errorf("Decode failed: got %v, want %v", decoded, want) + } +} |