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
|
#include <stdio.h>
#include <string.h>
#include <assert.h>
#include "test_framework.h"
#include "../global.h"
#include "../utf8.h"
// Returns the number of bytes in the utf8 unit, or -1 on invalid input.
// If the parse is successful, puts the bits that are part of the unit being
// parsed in *unit.
static int parse_utf8_prefix_byte(uint8_t b, int64_t *unit) {
if ((b & 0b10000000) == 0b00000000) {*unit = b & 0b01111111; return 1;}
if ((b & 0b11100000) == 0b11000000) {*unit = b & 0b00011111; return 2;}
if ((b & 0b11110000) == 0b11100000) {*unit = b & 0b00001111; return 3;}
if ((b & 0b11111000) == 0b11110000) {*unit = b & 0b00000111; return 4;}
return -1;
}
// Returns length of the parsed utf8 unit, and puts the parsed value in *unitp.
// No range checking is done.
static int parse_utf8_unit(const uint8_t *buf, size_t length, int64_t *unitp, bool debug) {
if (length == 0) {
if (debug) fprintf(stderr, "[utf8ref] unit at EOS\n");
return -1;
}
int64_t unit;
const int num_bytes = parse_utf8_prefix_byte(buf[0], &unit);
assert(num_bytes == -1 || (1 <= num_bytes && num_bytes <= 4));
if (num_bytes == -1) {
if (debug) fprintf(stderr, "[utf8ref] invalid prefix byte %x\n", (unsigned)buf[0]);
return -1;
}
assert(unit >= 0);
if (length < (size_t)num_bytes) {
if (debug) fprintf(stderr, "[utf8ref] prefix byte %x specifies length %d, but EOS\n", (unsigned)buf[0], num_bytes);
return -1;
}
for (int i = 1; i < num_bytes; i++) {
if ((buf[i] & 0b11000000) != 0b10000000) {
if (debug) fprintf(stderr, "[utf8ref] invalid continuation byte %x\n", (unsigned)buf[i]);
return -1;
}
unit = (unit << 6) | (buf[i] & 0b00111111);
}
// check for overlong encodings
if ((num_bytes >= 2 && unit <= 0x7F) ||
(num_bytes >= 3 && unit <= 0x7FF) ||
(num_bytes >= 4 && unit <= 0xFFFF)) {
if (debug) fprintf(stderr, "[utf8ref] overlong encoding with prefix byte %x\n", (unsigned)buf[0]);
return -1;
}
*unitp = unit;
return num_bytes;
}
static bool validate_utf8_reference(const char *buf_, size_t length, bool debug) {
const uint8_t *buf = (const uint8_t*)buf_;
size_t cursor = 0;
while (cursor < length) {
int64_t unit;
int len = parse_utf8_unit(buf + cursor, length - cursor, &unit, debug);
assert(len == -1 || (1 <= len && len <= 4));
if (len == -1) return false;
assert(unit >= 0);
// fprintf(stderr, "unit = 0x%lx\n", unit);
// Surrogate code point
if (0xD800 <= unit && unit <= 0xDFFF) {
if (debug) fprintf(stderr, "[utf8ref] surrogate code point %lx (prefix byte %x)\n", unit, (unsigned)buf[cursor]);
return false;
}
// Maximal unicode value
if (unit > 0x10FFFF) {
if (debug) fprintf(stderr, "[utf8ref] out of range code point %lx (prefix byte %x)\n", unit, (unsigned)buf[cursor]);
return false;
}
cursor += len;
}
return true;
}
// Requires that the buffer has space for at least 4 bytes.
// Returns the number of bytes written.
static int utf8_serialise(char *buf_, int64_t unit) {
uint8_t *buf = (uint8_t*)buf_;
#define PLACE_CONTINUATION_BYTE(idx_) \
{buf[(idx_)] = 0x80 | (unit & 0x3F); unit >>= 6;}
if (unit <= 0x7F) {
buf[0] = unit;
return 1;
}
if (unit <= 0x7FF) {
PLACE_CONTINUATION_BYTE(1);
buf[0] = 0xC0 | (unit & 0x1F);
return 2;
}
if (unit <= 0xFFFF) {
PLACE_CONTINUATION_BYTE(2);
PLACE_CONTINUATION_BYTE(1);
buf[0] = 0xE0 | (unit & 0x0F);
return 3;
}
if (unit <= 0x10FFFF) {
PLACE_CONTINUATION_BYTE(3);
PLACE_CONTINUATION_BYTE(2);
PLACE_CONTINUATION_BYTE(1);
buf[0] = 0xF0 | (unit & 0x07);
return 4;
}
assert(false && "Invalid unit in utf8_serialise");
#undef PLACE_CONTINUATION_BYTE
}
static void fill_random_buffer(char *buf, size_t length) {
size_t i = 0;
while (i + sizeof(long) < length) {
*(long*)&buf[i] = random();
i += sizeof(long);
}
while (i < length) buf[i++] = random();
}
DEFINE_TEST(utf8_unit1) {
EXPECT(validate_utf8("hello", 5));
EXPECT(validate_utf8_reference("hello", 5, true));
const char *str = "hello 🧀🇳🇱";
EXPECT(validate_utf8(str, strlen(str)));
EXPECT(validate_utf8_reference(str, strlen(str), true));
EXPECT(validate_utf8("\xe0\xad\xbc`j", 5));
EXPECT(validate_utf8_reference("\xe0\xad\xbc`j", 5, true));
EXPECT(validate_utf8("\xd3\xb0\\i\x00\x00\x00\x001\xc7\xaa_", 12));
EXPECT(validate_utf8("\xc7\xaa_", 3));
EXPECT(validate_utf8_reference("\xd3\xb0\\i\x00\x00\x00\x001\xc7\xaa_", 12, true));
EXPECT(!validate_utf8("\xf2\x98\xbcx", 4));
EXPECT(!validate_utf8_reference("\xf2\x98\xbcx", 4, false));
return 0;
}
DEFINE_TEST(utf8_random) {
const int max_length = 100;
const int num_tests = 10000000;
char *buffer = malloc(max_length + 1, char);
for (int test = 0; test < num_tests; test++) {
// fprintf(stderr, "== test = %d\n", test);
const int length = random() % max_length;
fill_random_buffer(buffer, length);
const bool ret_ref = validate_utf8_reference(buffer, length, false);
const bool ret_impl = validate_utf8(buffer, length);
if (ret_ref != ret_impl) {
fprintf(stderr, "buffer: ");
print_buffer(stderr, buffer, length);
fprintf(stderr, "\n");
fprintf(stderr, "reference -> %d, implementation -> %d\n", ret_ref, ret_impl);
EXPECTRET(1, false && "validate_utf8_reference == validate_utf8");
}
}
free(buffer);
return 0;
}
DEFINE_TEST(utf8_random_valid) {
const int max_length = 100;
const int num_tests = 3000000;
char *buffer = malloc(max_length + 1, char);
for (int test = 0; test < num_tests; test++) {
int length = random() % max_length;
int cursor = 0;
while (cursor + 4 <= length) {
const int64_t unit = random() % 0x110000;
if (0xD800 <= unit && unit <= 0xDFFF) continue; // surrogate
cursor += utf8_serialise(buffer + cursor, unit);
}
length = cursor;
const bool ret_ref = validate_utf8_reference(buffer, length, true);
if (!ret_ref) {
fprintf(stderr, "buffer: ");
print_buffer(stderr, buffer, length);
fprintf(stderr, "\n");
EXPECTRET(1, false && "validate_utf8_reference on valid string");
}
const bool ret_impl = validate_utf8(buffer, length);
if (!ret_impl) {
fprintf(stderr, "buffer: ");
print_buffer(stderr, buffer, length);
fprintf(stderr, "\n");
EXPECTRET(1, false && "validate_utf8 on valid string");
}
}
free(buffer);
return 0;
}
DEFINE_TEST(utf8_exhaustive_1) {
for (int64_t number = 0; number < 0x100000000LL; number++) {
EXPECT(
validate_utf8_reference((const char*)&number, 4, false)
== validate_utf8((const char*)&number, 4)
);
}
return 0;
}
|
