summaryrefslogtreecommitdiff
path: root/src/pkg/net/http/httputil/chunked.go
blob: 9632bfd19d5286d181fba0187495c37af68f0469 (plain)
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
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
// 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.

// The wire protocol for HTTP's "chunked" Transfer-Encoding.

// This code is duplicated in net/http and net/http/httputil.
// Please make any changes in both files.

package httputil

import (
	"bufio"
	"bytes"
	"errors"
	"fmt"
	"io"
)

const maxLineLength = 4096 // assumed <= bufio.defaultBufSize

var ErrLineTooLong = errors.New("header line too long")

// newChunkedReader returns a new chunkedReader that translates the data read from r
// out of HTTP "chunked" format before returning it.
// The chunkedReader returns io.EOF when the final 0-length chunk is read.
//
// newChunkedReader is not needed by normal applications. The http package
// automatically decodes chunking when reading response bodies.
func newChunkedReader(r io.Reader) io.Reader {
	br, ok := r.(*bufio.Reader)
	if !ok {
		br = bufio.NewReader(r)
	}
	return &chunkedReader{r: br}
}

type chunkedReader struct {
	r   *bufio.Reader
	n   uint64 // unread bytes in chunk
	err error
	buf [2]byte
}

func (cr *chunkedReader) beginChunk() {
	// chunk-size CRLF
	var line []byte
	line, cr.err = readLine(cr.r)
	if cr.err != nil {
		return
	}
	cr.n, cr.err = parseHexUint(line)
	if cr.err != nil {
		return
	}
	if cr.n == 0 {
		cr.err = io.EOF
	}
}

func (cr *chunkedReader) chunkHeaderAvailable() bool {
	n := cr.r.Buffered()
	if n > 0 {
		peek, _ := cr.r.Peek(n)
		return bytes.IndexByte(peek, '\n') >= 0
	}
	return false
}

func (cr *chunkedReader) Read(b []uint8) (n int, err error) {
	for cr.err == nil {
		if cr.n == 0 {
			if n > 0 && !cr.chunkHeaderAvailable() {
				// We've read enough. Don't potentially block
				// reading a new chunk header.
				break
			}
			cr.beginChunk()
			continue
		}
		if len(b) == 0 {
			break
		}
		rbuf := b
		if uint64(len(rbuf)) > cr.n {
			rbuf = rbuf[:cr.n]
		}
		var n0 int
		n0, cr.err = cr.r.Read(rbuf)
		n += n0
		b = b[n0:]
		cr.n -= uint64(n0)
		// If we're at the end of a chunk, read the next two
		// bytes to verify they are "\r\n".
		if cr.n == 0 && cr.err == nil {
			if _, cr.err = io.ReadFull(cr.r, cr.buf[:2]); cr.err == nil {
				if cr.buf[0] != '\r' || cr.buf[1] != '\n' {
					cr.err = errors.New("malformed chunked encoding")
				}
			}
		}
	}
	return n, cr.err
}

// Read a line of bytes (up to \n) from b.
// Give up if the line exceeds maxLineLength.
// The returned bytes are a pointer into storage in
// the bufio, so they are only valid until the next bufio read.
func readLine(b *bufio.Reader) (p []byte, err error) {
	if p, err = b.ReadSlice('\n'); err != nil {
		// We always know when EOF is coming.
		// If the caller asked for a line, there should be a line.
		if err == io.EOF {
			err = io.ErrUnexpectedEOF
		} else if err == bufio.ErrBufferFull {
			err = ErrLineTooLong
		}
		return nil, err
	}
	if len(p) >= maxLineLength {
		return nil, ErrLineTooLong
	}
	return trimTrailingWhitespace(p), nil
}

func trimTrailingWhitespace(b []byte) []byte {
	for len(b) > 0 && isASCIISpace(b[len(b)-1]) {
		b = b[:len(b)-1]
	}
	return b
}

func isASCIISpace(b byte) bool {
	return b == ' ' || b == '\t' || b == '\n' || b == '\r'
}

// newChunkedWriter returns a new chunkedWriter that translates writes into HTTP
// "chunked" format before writing them to w. Closing the returned chunkedWriter
// sends the final 0-length chunk that marks the end of the stream.
//
// newChunkedWriter is not needed by normal applications. The http
// package adds chunking automatically if handlers don't set a
// Content-Length header. Using newChunkedWriter inside a handler
// would result in double chunking or chunking with a Content-Length
// length, both of which are wrong.
func newChunkedWriter(w io.Writer) io.WriteCloser {
	return &chunkedWriter{w}
}

// Writing to chunkedWriter translates to writing in HTTP chunked Transfer
// Encoding wire format to the underlying Wire chunkedWriter.
type chunkedWriter struct {
	Wire io.Writer
}

// Write the contents of data as one chunk to Wire.
// NOTE: Note that the corresponding chunk-writing procedure in Conn.Write has
// a bug since it does not check for success of io.WriteString
func (cw *chunkedWriter) Write(data []byte) (n int, err error) {

	// Don't send 0-length data. It looks like EOF for chunked encoding.
	if len(data) == 0 {
		return 0, nil
	}

	if _, err = fmt.Fprintf(cw.Wire, "%x\r\n", len(data)); err != nil {
		return 0, err
	}
	if n, err = cw.Wire.Write(data); err != nil {
		return
	}
	if n != len(data) {
		err = io.ErrShortWrite
		return
	}
	_, err = io.WriteString(cw.Wire, "\r\n")

	return
}

func (cw *chunkedWriter) Close() error {
	_, err := io.WriteString(cw.Wire, "0\r\n")
	return err
}

func parseHexUint(v []byte) (n uint64, err error) {
	for _, b := range v {
		n <<= 4
		switch {
		case '0' <= b && b <= '9':
			b = b - '0'
		case 'a' <= b && b <= 'f':
			b = b - 'a' + 10
		case 'A' <= b && b <= 'F':
			b = b - 'A' + 10
		default:
			return 0, errors.New("invalid byte in chunk length")
		}
		n |= uint64(b)
	}
	return
}