File: decoder.go

package info (click to toggle)
golang-github-dop251-base64dec 0.0~git20231022.c6c9f9a-2
  • links: PTS, VCS
  • area: main
  • in suites: forky, sid
  • size: 80 kB
  • sloc: makefile: 2
file content (223 lines) | stat: -rw-r--r-- 5,350 bytes parent folder | download
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
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
//go:generate go run generate_decode_map.go

// Package base64dec contains a universal base64 decoder that works on both the standard and url-safe variants, padded and raw.
// The code is based on the standard encoding/base64 package.
package base64dec

import (
	"encoding/base64"
	"encoding/binary"
	"strconv"
)

const padChar = '='

type ByteSeq interface {
	[]byte | string
}

// DecodeBase64 decodes src and writes at most base64.RawStdEncoding.DecodedLen(len(src))
// bytes to dst and returns the number of bytes written. If src contains invalid base64 data, it will return the
// number of bytes successfully written and base64.CorruptInputError.
// New line characters (\r and \n) are ignored.
// The input can be in the standard or the alternate (aka url-safe) encoding. It can be padded or un-padded.
// If there is a correct padding, it is consumed and no error is returned. If there is no padding where it's required,
// no error is returned. If there is an incorrect padding (i.e. too many or too few characters) it is treated
// as garbage at the end (i.e. the error will point to the first padding character).
func DecodeBase64[T ByteSeq](dst []byte, src T) (n int, err error) {
	if len(src) == 0 {
		return 0, nil
	}

	si := 0
	for strconv.IntSize >= 64 && len(src)-si >= 8 && len(dst)-n >= 8 {
		src2 := src[si : si+8]
		if dn, ok := assemble64(
			decodeMap[src2[0]],
			decodeMap[src2[1]],
			decodeMap[src2[2]],
			decodeMap[src2[3]],
			decodeMap[src2[4]],
			decodeMap[src2[5]],
			decodeMap[src2[6]],
			decodeMap[src2[7]],
		); ok {
			binary.BigEndian.PutUint64(dst[n:], dn)
			n += 6
			si += 8
		} else {
			var ninc int
			si, ninc, err = decodeQuantum(dst[n:], src, si)
			n += ninc
			if err != nil {
				return n, err
			}
		}
	}

	for len(src)-si >= 4 && len(dst)-n >= 4 {
		src2 := src[si : si+4]
		if dn, ok := assemble32(
			decodeMap[src2[0]],
			decodeMap[src2[1]],
			decodeMap[src2[2]],
			decodeMap[src2[3]],
		); ok {
			binary.BigEndian.PutUint32(dst[n:], dn)
			n += 3
			si += 4
		} else {
			var ninc int
			si, ninc, err = decodeQuantum(dst[n:], src, si)
			n += ninc
			if err != nil {
				return n, err
			}
		}
	}

	for si < len(src) {
		var ninc int
		si, ninc, err = decodeQuantum(dst[n:], src, si)
		n += ninc
		if err != nil {
			return n, err
		}
	}
	return n, err
}

// assemble32 assembles 4 base64 digits into 3 bytes.
// Each digit comes from the decode map, and will be 0xff
// if it came from an invalid character.
func assemble32(n1, n2, n3, n4 byte) (dn uint32, ok bool) {
	// Check that all the digits are valid. If any of them was 0xff, their
	// bitwise OR will be 0xff.
	if n1|n2|n3|n4 == 0xff {
		return 0, false
	}
	return uint32(n1)<<26 |
			uint32(n2)<<20 |
			uint32(n3)<<14 |
			uint32(n4)<<8,
		true
}

// assemble64 assembles 8 base64 digits into 6 bytes.
// Each digit comes from the decode map, and will be 0xff
// if it came from an invalid character.
func assemble64(n1, n2, n3, n4, n5, n6, n7, n8 byte) (dn uint64, ok bool) {
	// Check that all the digits are valid. If any of them was 0xff, their
	// bitwise OR will be 0xff.
	if n1|n2|n3|n4|n5|n6|n7|n8 == 0xff {
		return 0, false
	}
	return uint64(n1)<<58 |
			uint64(n2)<<52 |
			uint64(n3)<<46 |
			uint64(n4)<<40 |
			uint64(n5)<<34 |
			uint64(n6)<<28 |
			uint64(n7)<<22 |
			uint64(n8)<<16,
		true
}

// decodeQuantum decodes up to 4 base64 bytes. The received parameters are
// the destination buffer dst, the source buffer src and an index in the
// source buffer si.
// It returns the number of bytes read from src, the number of bytes written
// to dst, and an error, if any.
func decodeQuantum[T ByteSeq](dst []byte, src T, si int) (nsi, n int, err error) {
	// Decode quantum using the base64 alphabet
	var dbuf [4]byte
	dlen := 4

	for j := 0; j < len(dbuf); j++ {
		if len(src) == si {
			if j == 0 {
				return si, 0, nil
			}
			dlen = j
			break
		}
		in := src[si]
		si++

		out := decodeMap[in]
		if out != 0xff {
			dbuf[j] = out
			continue
		}

		if in == '\n' || in == '\r' {
			j--
			continue
		}

		dlen = j

		if rune(in) != padChar {
			err = base64.CorruptInputError(si - 1)
			break
		}

		// We've reached the end and there's padding
		switch j {
		case 0, 1:
			// incorrect padding
			err = base64.CorruptInputError(si - 1)
		case 2:
			// "==" is expected, the first "=" is already consumed.
			// skip over newlines
			for si < len(src) && (src[si] == '\n' || src[si] == '\r') {
				si++
			}
			if si == len(src) {
				// not enough padding
				err = base64.CorruptInputError(si - 1)
				break
			} else if rune(src[si]) != padChar {
				// incorrect padding
				err = base64.CorruptInputError(si - 1)
				break
			}

			si++
		}

		if err == nil {
			// skip over newlines
			for si < len(src) && (src[si] == '\n' || src[si] == '\r') {
				si++
			}
			if si < len(src) {
				// trailing garbage
				err = base64.CorruptInputError(si)
			}
		}
		break
	}

	if dlen == 0 {
		return si, 0, err
	}

	// Convert 4x 6bit source bytes into 3 bytes
	val := uint(dbuf[0])<<18 | uint(dbuf[1])<<12 | uint(dbuf[2])<<6 | uint(dbuf[3])
	dbuf[2], dbuf[1], dbuf[0] = byte(val>>0), byte(val>>8), byte(val>>16)
	switch dlen {
	case 4:
		dst[2] = dbuf[2]
		dbuf[2] = 0
		fallthrough
	case 3:
		dst[1] = dbuf[1]
		dbuf[1] = 0
		fallthrough
	case 2:
		dst[0] = dbuf[0]
	}

	return si, dlen - 1, err
}