|
1 package toml |
|
2 |
|
3 import ( |
|
4 "unicode/utf8" |
|
5 ) |
|
6 |
|
7 type utf8Err struct { |
|
8 Index int |
|
9 Size int |
|
10 } |
|
11 |
|
12 func (u utf8Err) Zero() bool { |
|
13 return u.Size == 0 |
|
14 } |
|
15 |
|
16 // Verified that a given string is only made of valid UTF-8 characters allowed |
|
17 // by the TOML spec: |
|
18 // |
|
19 // Any Unicode character may be used except those that must be escaped: |
|
20 // quotation mark, backslash, and the control characters other than tab (U+0000 |
|
21 // to U+0008, U+000A to U+001F, U+007F). |
|
22 // |
|
23 // It is a copy of the Go 1.17 utf8.Valid implementation, tweaked to exit early |
|
24 // when a character is not allowed. |
|
25 // |
|
26 // The returned utf8Err is Zero() if the string is valid, or contains the byte |
|
27 // index and size of the invalid character. |
|
28 // |
|
29 // quotation mark => already checked |
|
30 // backslash => already checked |
|
31 // 0-0x8 => invalid |
|
32 // 0x9 => tab, ok |
|
33 // 0xA - 0x1F => invalid |
|
34 // 0x7F => invalid |
|
35 func utf8TomlValidAlreadyEscaped(p []byte) (err utf8Err) { |
|
36 // Fast path. Check for and skip 8 bytes of ASCII characters per iteration. |
|
37 offset := 0 |
|
38 for len(p) >= 8 { |
|
39 // Combining two 32 bit loads allows the same code to be used |
|
40 // for 32 and 64 bit platforms. |
|
41 // The compiler can generate a 32bit load for first32 and second32 |
|
42 // on many platforms. See test/codegen/memcombine.go. |
|
43 first32 := uint32(p[0]) | uint32(p[1])<<8 | uint32(p[2])<<16 | uint32(p[3])<<24 |
|
44 second32 := uint32(p[4]) | uint32(p[5])<<8 | uint32(p[6])<<16 | uint32(p[7])<<24 |
|
45 if (first32|second32)&0x80808080 != 0 { |
|
46 // Found a non ASCII byte (>= RuneSelf). |
|
47 break |
|
48 } |
|
49 |
|
50 for i, b := range p[:8] { |
|
51 if invalidAscii(b) { |
|
52 err.Index = offset + i |
|
53 err.Size = 1 |
|
54 return |
|
55 } |
|
56 } |
|
57 |
|
58 p = p[8:] |
|
59 offset += 8 |
|
60 } |
|
61 n := len(p) |
|
62 for i := 0; i < n; { |
|
63 pi := p[i] |
|
64 if pi < utf8.RuneSelf { |
|
65 if invalidAscii(pi) { |
|
66 err.Index = offset + i |
|
67 err.Size = 1 |
|
68 return |
|
69 } |
|
70 i++ |
|
71 continue |
|
72 } |
|
73 x := first[pi] |
|
74 if x == xx { |
|
75 // Illegal starter byte. |
|
76 err.Index = offset + i |
|
77 err.Size = 1 |
|
78 return |
|
79 } |
|
80 size := int(x & 7) |
|
81 if i+size > n { |
|
82 // Short or invalid. |
|
83 err.Index = offset + i |
|
84 err.Size = n - i |
|
85 return |
|
86 } |
|
87 accept := acceptRanges[x>>4] |
|
88 if c := p[i+1]; c < accept.lo || accept.hi < c { |
|
89 err.Index = offset + i |
|
90 err.Size = 2 |
|
91 return |
|
92 } else if size == 2 { |
|
93 } else if c := p[i+2]; c < locb || hicb < c { |
|
94 err.Index = offset + i |
|
95 err.Size = 3 |
|
96 return |
|
97 } else if size == 3 { |
|
98 } else if c := p[i+3]; c < locb || hicb < c { |
|
99 err.Index = offset + i |
|
100 err.Size = 4 |
|
101 return |
|
102 } |
|
103 i += size |
|
104 } |
|
105 return |
|
106 } |
|
107 |
|
108 // Return the size of the next rune if valid, 0 otherwise. |
|
109 func utf8ValidNext(p []byte) int { |
|
110 c := p[0] |
|
111 |
|
112 if c < utf8.RuneSelf { |
|
113 if invalidAscii(c) { |
|
114 return 0 |
|
115 } |
|
116 return 1 |
|
117 } |
|
118 |
|
119 x := first[c] |
|
120 if x == xx { |
|
121 // Illegal starter byte. |
|
122 return 0 |
|
123 } |
|
124 size := int(x & 7) |
|
125 if size > len(p) { |
|
126 // Short or invalid. |
|
127 return 0 |
|
128 } |
|
129 accept := acceptRanges[x>>4] |
|
130 if c := p[1]; c < accept.lo || accept.hi < c { |
|
131 return 0 |
|
132 } else if size == 2 { |
|
133 } else if c := p[2]; c < locb || hicb < c { |
|
134 return 0 |
|
135 } else if size == 3 { |
|
136 } else if c := p[3]; c < locb || hicb < c { |
|
137 return 0 |
|
138 } |
|
139 |
|
140 return size |
|
141 } |
|
142 |
|
143 var invalidAsciiTable = [256]bool{ |
|
144 0x00: true, |
|
145 0x01: true, |
|
146 0x02: true, |
|
147 0x03: true, |
|
148 0x04: true, |
|
149 0x05: true, |
|
150 0x06: true, |
|
151 0x07: true, |
|
152 0x08: true, |
|
153 // 0x09 TAB |
|
154 // 0x0A LF |
|
155 0x0B: true, |
|
156 0x0C: true, |
|
157 // 0x0D CR |
|
158 0x0E: true, |
|
159 0x0F: true, |
|
160 0x10: true, |
|
161 0x11: true, |
|
162 0x12: true, |
|
163 0x13: true, |
|
164 0x14: true, |
|
165 0x15: true, |
|
166 0x16: true, |
|
167 0x17: true, |
|
168 0x18: true, |
|
169 0x19: true, |
|
170 0x1A: true, |
|
171 0x1B: true, |
|
172 0x1C: true, |
|
173 0x1D: true, |
|
174 0x1E: true, |
|
175 0x1F: true, |
|
176 // 0x20 - 0x7E Printable ASCII characters |
|
177 0x7F: true, |
|
178 } |
|
179 |
|
180 func invalidAscii(b byte) bool { |
|
181 return invalidAsciiTable[b] |
|
182 } |
|
183 |
|
184 // acceptRange gives the range of valid values for the second byte in a UTF-8 |
|
185 // sequence. |
|
186 type acceptRange struct { |
|
187 lo uint8 // lowest value for second byte. |
|
188 hi uint8 // highest value for second byte. |
|
189 } |
|
190 |
|
191 // acceptRanges has size 16 to avoid bounds checks in the code that uses it. |
|
192 var acceptRanges = [16]acceptRange{ |
|
193 0: {locb, hicb}, |
|
194 1: {0xA0, hicb}, |
|
195 2: {locb, 0x9F}, |
|
196 3: {0x90, hicb}, |
|
197 4: {locb, 0x8F}, |
|
198 } |
|
199 |
|
200 // first is information about the first byte in a UTF-8 sequence. |
|
201 var first = [256]uint8{ |
|
202 // 1 2 3 4 5 6 7 8 9 A B C D E F |
|
203 as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, // 0x00-0x0F |
|
204 as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, // 0x10-0x1F |
|
205 as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, // 0x20-0x2F |
|
206 as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, // 0x30-0x3F |
|
207 as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, // 0x40-0x4F |
|
208 as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, // 0x50-0x5F |
|
209 as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, // 0x60-0x6F |
|
210 as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, // 0x70-0x7F |
|
211 // 1 2 3 4 5 6 7 8 9 A B C D E F |
|
212 xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, // 0x80-0x8F |
|
213 xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, // 0x90-0x9F |
|
214 xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, // 0xA0-0xAF |
|
215 xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, // 0xB0-0xBF |
|
216 xx, xx, s1, s1, s1, s1, s1, s1, s1, s1, s1, s1, s1, s1, s1, s1, // 0xC0-0xCF |
|
217 s1, s1, s1, s1, s1, s1, s1, s1, s1, s1, s1, s1, s1, s1, s1, s1, // 0xD0-0xDF |
|
218 s2, s3, s3, s3, s3, s3, s3, s3, s3, s3, s3, s3, s3, s4, s3, s3, // 0xE0-0xEF |
|
219 s5, s6, s6, s6, s7, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, // 0xF0-0xFF |
|
220 } |
|
221 |
|
222 const ( |
|
223 // The default lowest and highest continuation byte. |
|
224 locb = 0b10000000 |
|
225 hicb = 0b10111111 |
|
226 |
|
227 // These names of these constants are chosen to give nice alignment in the |
|
228 // table below. The first nibble is an index into acceptRanges or F for |
|
229 // special one-byte cases. The second nibble is the Rune length or the |
|
230 // Status for the special one-byte case. |
|
231 xx = 0xF1 // invalid: size 1 |
|
232 as = 0xF0 // ASCII: size 1 |
|
233 s1 = 0x02 // accept 0, size 2 |
|
234 s2 = 0x13 // accept 1, size 3 |
|
235 s3 = 0x03 // accept 0, size 3 |
|
236 s4 = 0x23 // accept 2, size 3 |
|
237 s5 = 0x34 // accept 3, size 4 |
|
238 s6 = 0x04 // accept 0, size 4 |
|
239 s7 = 0x44 // accept 4, size 4 |
|
240 ) |