google-protobuf 3.20.0.rc.1-x86-mingw32 → 3.20.1.rc.1-x86-mingw32
Sign up to get free protection for your applications and to get access to all the features.
Potentially problematic release.
This version of google-protobuf might be problematic. Click here for more details.
- checksums.yaml +4 -4
- data/ext/google/protobuf_c/convert.c +4 -4
- data/ext/google/protobuf_c/defs.c +1 -0
- data/ext/google/protobuf_c/extconf.rb +1 -1
- data/ext/google/protobuf_c/message.c +19 -25
- data/ext/google/protobuf_c/protobuf.c +14 -3
- data/ext/google/protobuf_c/ruby-upb.c +399 -316
- data/ext/google/protobuf_c/ruby-upb.h +712 -388
- data/ext/google/protobuf_c/third_party/utf8_range/naive.c +92 -0
- data/ext/google/protobuf_c/third_party/utf8_range/range2-neon.c +157 -0
- data/ext/google/protobuf_c/third_party/utf8_range/range2-sse.c +170 -0
- data/ext/google/protobuf_c/third_party/utf8_range/utf8_range.h +1 -1
- data/lib/google/2.5/protobuf_c.so +0 -0
- data/lib/google/2.6/protobuf_c.so +0 -0
- data/lib/google/2.7/protobuf_c.so +0 -0
- data/lib/google/3.0/protobuf_c.so +0 -0
- data/lib/google/3.1/protobuf_c.so +0 -0
- data/tests/basic.rb +88 -11
- metadata +14 -11
- data/ext/google/protobuf_c/third_party/utf8_range/utf8_range.c +0 -395
@@ -0,0 +1,92 @@
|
|
1
|
+
#include <stdio.h>
|
2
|
+
|
3
|
+
/*
|
4
|
+
* http://www.unicode.org/versions/Unicode6.0.0/ch03.pdf - page 94
|
5
|
+
*
|
6
|
+
* Table 3-7. Well-Formed UTF-8 Byte Sequences
|
7
|
+
*
|
8
|
+
* +--------------------+------------+-------------+------------+-------------+
|
9
|
+
* | Code Points | First Byte | Second Byte | Third Byte | Fourth Byte |
|
10
|
+
* +--------------------+------------+-------------+------------+-------------+
|
11
|
+
* | U+0000..U+007F | 00..7F | | | |
|
12
|
+
* +--------------------+------------+-------------+------------+-------------+
|
13
|
+
* | U+0080..U+07FF | C2..DF | 80..BF | | |
|
14
|
+
* +--------------------+------------+-------------+------------+-------------+
|
15
|
+
* | U+0800..U+0FFF | E0 | A0..BF | 80..BF | |
|
16
|
+
* +--------------------+------------+-------------+------------+-------------+
|
17
|
+
* | U+1000..U+CFFF | E1..EC | 80..BF | 80..BF | |
|
18
|
+
* +--------------------+------------+-------------+------------+-------------+
|
19
|
+
* | U+D000..U+D7FF | ED | 80..9F | 80..BF | |
|
20
|
+
* +--------------------+------------+-------------+------------+-------------+
|
21
|
+
* | U+E000..U+FFFF | EE..EF | 80..BF | 80..BF | |
|
22
|
+
* +--------------------+------------+-------------+------------+-------------+
|
23
|
+
* | U+10000..U+3FFFF | F0 | 90..BF | 80..BF | 80..BF |
|
24
|
+
* +--------------------+------------+-------------+------------+-------------+
|
25
|
+
* | U+40000..U+FFFFF | F1..F3 | 80..BF | 80..BF | 80..BF |
|
26
|
+
* +--------------------+------------+-------------+------------+-------------+
|
27
|
+
* | U+100000..U+10FFFF | F4 | 80..8F | 80..BF | 80..BF |
|
28
|
+
* +--------------------+------------+-------------+------------+-------------+
|
29
|
+
*/
|
30
|
+
|
31
|
+
/* Return 0 - success, >0 - index(1 based) of first error char */
|
32
|
+
int utf8_naive(const unsigned char *data, int len)
|
33
|
+
{
|
34
|
+
int err_pos = 1;
|
35
|
+
|
36
|
+
while (len) {
|
37
|
+
int bytes;
|
38
|
+
const unsigned char byte1 = data[0];
|
39
|
+
|
40
|
+
/* 00..7F */
|
41
|
+
if (byte1 <= 0x7F) {
|
42
|
+
bytes = 1;
|
43
|
+
/* C2..DF, 80..BF */
|
44
|
+
} else if (len >= 2 && byte1 >= 0xC2 && byte1 <= 0xDF &&
|
45
|
+
(signed char)data[1] <= (signed char)0xBF) {
|
46
|
+
bytes = 2;
|
47
|
+
} else if (len >= 3) {
|
48
|
+
const unsigned char byte2 = data[1];
|
49
|
+
|
50
|
+
/* Is byte2, byte3 between 0x80 ~ 0xBF */
|
51
|
+
const int byte2_ok = (signed char)byte2 <= (signed char)0xBF;
|
52
|
+
const int byte3_ok = (signed char)data[2] <= (signed char)0xBF;
|
53
|
+
|
54
|
+
if (byte2_ok && byte3_ok &&
|
55
|
+
/* E0, A0..BF, 80..BF */
|
56
|
+
((byte1 == 0xE0 && byte2 >= 0xA0) ||
|
57
|
+
/* E1..EC, 80..BF, 80..BF */
|
58
|
+
(byte1 >= 0xE1 && byte1 <= 0xEC) ||
|
59
|
+
/* ED, 80..9F, 80..BF */
|
60
|
+
(byte1 == 0xED && byte2 <= 0x9F) ||
|
61
|
+
/* EE..EF, 80..BF, 80..BF */
|
62
|
+
(byte1 >= 0xEE && byte1 <= 0xEF))) {
|
63
|
+
bytes = 3;
|
64
|
+
} else if (len >= 4) {
|
65
|
+
/* Is byte4 between 0x80 ~ 0xBF */
|
66
|
+
const int byte4_ok = (signed char)data[3] <= (signed char)0xBF;
|
67
|
+
|
68
|
+
if (byte2_ok && byte3_ok && byte4_ok &&
|
69
|
+
/* F0, 90..BF, 80..BF, 80..BF */
|
70
|
+
((byte1 == 0xF0 && byte2 >= 0x90) ||
|
71
|
+
/* F1..F3, 80..BF, 80..BF, 80..BF */
|
72
|
+
(byte1 >= 0xF1 && byte1 <= 0xF3) ||
|
73
|
+
/* F4, 80..8F, 80..BF, 80..BF */
|
74
|
+
(byte1 == 0xF4 && byte2 <= 0x8F))) {
|
75
|
+
bytes = 4;
|
76
|
+
} else {
|
77
|
+
return err_pos;
|
78
|
+
}
|
79
|
+
} else {
|
80
|
+
return err_pos;
|
81
|
+
}
|
82
|
+
} else {
|
83
|
+
return err_pos;
|
84
|
+
}
|
85
|
+
|
86
|
+
len -= bytes;
|
87
|
+
err_pos += bytes;
|
88
|
+
data += bytes;
|
89
|
+
}
|
90
|
+
|
91
|
+
return 0;
|
92
|
+
}
|
@@ -0,0 +1,157 @@
|
|
1
|
+
/*
|
2
|
+
* Process 2x16 bytes in each iteration.
|
3
|
+
* Comments removed for brevity. See range-neon.c for details.
|
4
|
+
*/
|
5
|
+
#if defined(__aarch64__) && defined(__ARM_NEON)
|
6
|
+
|
7
|
+
#include <stdio.h>
|
8
|
+
#include <stdint.h>
|
9
|
+
#include <arm_neon.h>
|
10
|
+
|
11
|
+
int utf8_naive(const unsigned char *data, int len);
|
12
|
+
|
13
|
+
static const uint8_t _first_len_tbl[] = {
|
14
|
+
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 2, 3,
|
15
|
+
};
|
16
|
+
|
17
|
+
static const uint8_t _first_range_tbl[] = {
|
18
|
+
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 8, 8, 8, 8,
|
19
|
+
};
|
20
|
+
|
21
|
+
static const uint8_t _range_min_tbl[] = {
|
22
|
+
0x00, 0x80, 0x80, 0x80, 0xA0, 0x80, 0x90, 0x80,
|
23
|
+
0xC2, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF,
|
24
|
+
};
|
25
|
+
static const uint8_t _range_max_tbl[] = {
|
26
|
+
0x7F, 0xBF, 0xBF, 0xBF, 0xBF, 0x9F, 0xBF, 0x8F,
|
27
|
+
0xF4, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
|
28
|
+
};
|
29
|
+
|
30
|
+
static const uint8_t _range_adjust_tbl[] = {
|
31
|
+
2, 3, 0, 0, 0, 0, 0, 0, 0, 4, 0, 0, 0, 0, 0, 0,
|
32
|
+
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 0, 0, 0, 0,
|
33
|
+
};
|
34
|
+
|
35
|
+
/* Return 0 on success, -1 on error */
|
36
|
+
int utf8_range2(const unsigned char *data, int len)
|
37
|
+
{
|
38
|
+
if (len >= 32) {
|
39
|
+
uint8x16_t prev_input = vdupq_n_u8(0);
|
40
|
+
uint8x16_t prev_first_len = vdupq_n_u8(0);
|
41
|
+
|
42
|
+
const uint8x16_t first_len_tbl = vld1q_u8(_first_len_tbl);
|
43
|
+
const uint8x16_t first_range_tbl = vld1q_u8(_first_range_tbl);
|
44
|
+
const uint8x16_t range_min_tbl = vld1q_u8(_range_min_tbl);
|
45
|
+
const uint8x16_t range_max_tbl = vld1q_u8(_range_max_tbl);
|
46
|
+
const uint8x16x2_t range_adjust_tbl = vld2q_u8(_range_adjust_tbl);
|
47
|
+
|
48
|
+
const uint8x16_t const_1 = vdupq_n_u8(1);
|
49
|
+
const uint8x16_t const_2 = vdupq_n_u8(2);
|
50
|
+
const uint8x16_t const_e0 = vdupq_n_u8(0xE0);
|
51
|
+
|
52
|
+
uint8x16_t error1 = vdupq_n_u8(0);
|
53
|
+
uint8x16_t error2 = vdupq_n_u8(0);
|
54
|
+
uint8x16_t error3 = vdupq_n_u8(0);
|
55
|
+
uint8x16_t error4 = vdupq_n_u8(0);
|
56
|
+
|
57
|
+
while (len >= 32) {
|
58
|
+
/******************* two blocks interleaved **********************/
|
59
|
+
|
60
|
+
#if defined(__GNUC__) && !defined(__clang__) && (__GNUC__ < 8)
|
61
|
+
/* gcc doesn't support vldq1_u8_x2 until version 8 */
|
62
|
+
const uint8x16_t input_a = vld1q_u8(data);
|
63
|
+
const uint8x16_t input_b = vld1q_u8(data + 16);
|
64
|
+
#else
|
65
|
+
/* Forces a double load on Clang */
|
66
|
+
const uint8x16x2_t input_pair = vld1q_u8_x2(data);
|
67
|
+
const uint8x16_t input_a = input_pair.val[0];
|
68
|
+
const uint8x16_t input_b = input_pair.val[1];
|
69
|
+
#endif
|
70
|
+
|
71
|
+
const uint8x16_t high_nibbles_a = vshrq_n_u8(input_a, 4);
|
72
|
+
const uint8x16_t high_nibbles_b = vshrq_n_u8(input_b, 4);
|
73
|
+
|
74
|
+
const uint8x16_t first_len_a =
|
75
|
+
vqtbl1q_u8(first_len_tbl, high_nibbles_a);
|
76
|
+
const uint8x16_t first_len_b =
|
77
|
+
vqtbl1q_u8(first_len_tbl, high_nibbles_b);
|
78
|
+
|
79
|
+
uint8x16_t range_a = vqtbl1q_u8(first_range_tbl, high_nibbles_a);
|
80
|
+
uint8x16_t range_b = vqtbl1q_u8(first_range_tbl, high_nibbles_b);
|
81
|
+
|
82
|
+
range_a =
|
83
|
+
vorrq_u8(range_a, vextq_u8(prev_first_len, first_len_a, 15));
|
84
|
+
range_b =
|
85
|
+
vorrq_u8(range_b, vextq_u8(first_len_a, first_len_b, 15));
|
86
|
+
|
87
|
+
uint8x16_t tmp1_a, tmp2_a, tmp1_b, tmp2_b;
|
88
|
+
tmp1_a = vextq_u8(prev_first_len, first_len_a, 14);
|
89
|
+
tmp1_a = vqsubq_u8(tmp1_a, const_1);
|
90
|
+
range_a = vorrq_u8(range_a, tmp1_a);
|
91
|
+
|
92
|
+
tmp1_b = vextq_u8(first_len_a, first_len_b, 14);
|
93
|
+
tmp1_b = vqsubq_u8(tmp1_b, const_1);
|
94
|
+
range_b = vorrq_u8(range_b, tmp1_b);
|
95
|
+
|
96
|
+
tmp2_a = vextq_u8(prev_first_len, first_len_a, 13);
|
97
|
+
tmp2_a = vqsubq_u8(tmp2_a, const_2);
|
98
|
+
range_a = vorrq_u8(range_a, tmp2_a);
|
99
|
+
|
100
|
+
tmp2_b = vextq_u8(first_len_a, first_len_b, 13);
|
101
|
+
tmp2_b = vqsubq_u8(tmp2_b, const_2);
|
102
|
+
range_b = vorrq_u8(range_b, tmp2_b);
|
103
|
+
|
104
|
+
uint8x16_t shift1_a = vextq_u8(prev_input, input_a, 15);
|
105
|
+
uint8x16_t pos_a = vsubq_u8(shift1_a, const_e0);
|
106
|
+
range_a = vaddq_u8(range_a, vqtbl2q_u8(range_adjust_tbl, pos_a));
|
107
|
+
|
108
|
+
uint8x16_t shift1_b = vextq_u8(input_a, input_b, 15);
|
109
|
+
uint8x16_t pos_b = vsubq_u8(shift1_b, const_e0);
|
110
|
+
range_b = vaddq_u8(range_b, vqtbl2q_u8(range_adjust_tbl, pos_b));
|
111
|
+
|
112
|
+
uint8x16_t minv_a = vqtbl1q_u8(range_min_tbl, range_a);
|
113
|
+
uint8x16_t maxv_a = vqtbl1q_u8(range_max_tbl, range_a);
|
114
|
+
|
115
|
+
uint8x16_t minv_b = vqtbl1q_u8(range_min_tbl, range_b);
|
116
|
+
uint8x16_t maxv_b = vqtbl1q_u8(range_max_tbl, range_b);
|
117
|
+
|
118
|
+
error1 = vorrq_u8(error1, vcltq_u8(input_a, minv_a));
|
119
|
+
error2 = vorrq_u8(error2, vcgtq_u8(input_a, maxv_a));
|
120
|
+
|
121
|
+
error3 = vorrq_u8(error3, vcltq_u8(input_b, minv_b));
|
122
|
+
error4 = vorrq_u8(error4, vcgtq_u8(input_b, maxv_b));
|
123
|
+
|
124
|
+
/************************ next iteration *************************/
|
125
|
+
prev_input = input_b;
|
126
|
+
prev_first_len = first_len_b;
|
127
|
+
|
128
|
+
data += 32;
|
129
|
+
len -= 32;
|
130
|
+
}
|
131
|
+
error1 = vorrq_u8(error1, error2);
|
132
|
+
error1 = vorrq_u8(error1, error3);
|
133
|
+
error1 = vorrq_u8(error1, error4);
|
134
|
+
|
135
|
+
if (vmaxvq_u8(error1))
|
136
|
+
return -1;
|
137
|
+
|
138
|
+
uint32_t token4;
|
139
|
+
vst1q_lane_u32(&token4, vreinterpretq_u32_u8(prev_input), 3);
|
140
|
+
|
141
|
+
const int8_t *token = (const int8_t *)&token4;
|
142
|
+
int lookahead = 0;
|
143
|
+
if (token[3] > (int8_t)0xBF)
|
144
|
+
lookahead = 1;
|
145
|
+
else if (token[2] > (int8_t)0xBF)
|
146
|
+
lookahead = 2;
|
147
|
+
else if (token[1] > (int8_t)0xBF)
|
148
|
+
lookahead = 3;
|
149
|
+
|
150
|
+
data -= lookahead;
|
151
|
+
len += lookahead;
|
152
|
+
}
|
153
|
+
|
154
|
+
return utf8_naive(data, len);
|
155
|
+
}
|
156
|
+
|
157
|
+
#endif
|
@@ -0,0 +1,170 @@
|
|
1
|
+
/*
|
2
|
+
* Process 2x16 bytes in each iteration.
|
3
|
+
* Comments removed for brevity. See range-sse.c for details.
|
4
|
+
*/
|
5
|
+
#ifdef __SSE4_1__
|
6
|
+
|
7
|
+
#include <stdio.h>
|
8
|
+
#include <stdint.h>
|
9
|
+
#include <x86intrin.h>
|
10
|
+
|
11
|
+
int utf8_naive(const unsigned char *data, int len);
|
12
|
+
|
13
|
+
static const int8_t _first_len_tbl[] = {
|
14
|
+
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 2, 3,
|
15
|
+
};
|
16
|
+
|
17
|
+
static const int8_t _first_range_tbl[] = {
|
18
|
+
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 8, 8, 8, 8,
|
19
|
+
};
|
20
|
+
|
21
|
+
static const int8_t _range_min_tbl[] = {
|
22
|
+
0x00, 0x80, 0x80, 0x80, 0xA0, 0x80, 0x90, 0x80,
|
23
|
+
0xC2, 0x7F, 0x7F, 0x7F, 0x7F, 0x7F, 0x7F, 0x7F,
|
24
|
+
};
|
25
|
+
static const int8_t _range_max_tbl[] = {
|
26
|
+
0x7F, 0xBF, 0xBF, 0xBF, 0xBF, 0x9F, 0xBF, 0x8F,
|
27
|
+
0xF4, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80,
|
28
|
+
};
|
29
|
+
|
30
|
+
static const int8_t _df_ee_tbl[] = {
|
31
|
+
0, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0,
|
32
|
+
};
|
33
|
+
static const int8_t _ef_fe_tbl[] = {
|
34
|
+
0, 3, 0, 0, 0, 4, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
|
35
|
+
};
|
36
|
+
|
37
|
+
/* Return 0 on success, -1 on error */
|
38
|
+
int utf8_range2(const unsigned char *data, int len)
|
39
|
+
{
|
40
|
+
if (len >= 32) {
|
41
|
+
__m128i prev_input = _mm_set1_epi8(0);
|
42
|
+
__m128i prev_first_len = _mm_set1_epi8(0);
|
43
|
+
|
44
|
+
const __m128i first_len_tbl =
|
45
|
+
_mm_loadu_si128((const __m128i *)_first_len_tbl);
|
46
|
+
const __m128i first_range_tbl =
|
47
|
+
_mm_loadu_si128((const __m128i *)_first_range_tbl);
|
48
|
+
const __m128i range_min_tbl =
|
49
|
+
_mm_loadu_si128((const __m128i *)_range_min_tbl);
|
50
|
+
const __m128i range_max_tbl =
|
51
|
+
_mm_loadu_si128((const __m128i *)_range_max_tbl);
|
52
|
+
const __m128i df_ee_tbl =
|
53
|
+
_mm_loadu_si128((const __m128i *)_df_ee_tbl);
|
54
|
+
const __m128i ef_fe_tbl =
|
55
|
+
_mm_loadu_si128((const __m128i *)_ef_fe_tbl);
|
56
|
+
|
57
|
+
__m128i error = _mm_set1_epi8(0);
|
58
|
+
|
59
|
+
while (len >= 32) {
|
60
|
+
/***************************** block 1 ****************************/
|
61
|
+
const __m128i input_a = _mm_loadu_si128((const __m128i *)data);
|
62
|
+
|
63
|
+
__m128i high_nibbles =
|
64
|
+
_mm_and_si128(_mm_srli_epi16(input_a, 4), _mm_set1_epi8(0x0F));
|
65
|
+
|
66
|
+
__m128i first_len_a = _mm_shuffle_epi8(first_len_tbl, high_nibbles);
|
67
|
+
|
68
|
+
__m128i range_a = _mm_shuffle_epi8(first_range_tbl, high_nibbles);
|
69
|
+
|
70
|
+
range_a = _mm_or_si128(
|
71
|
+
range_a, _mm_alignr_epi8(first_len_a, prev_first_len, 15));
|
72
|
+
|
73
|
+
__m128i tmp;
|
74
|
+
tmp = _mm_alignr_epi8(first_len_a, prev_first_len, 14);
|
75
|
+
tmp = _mm_subs_epu8(tmp, _mm_set1_epi8(1));
|
76
|
+
range_a = _mm_or_si128(range_a, tmp);
|
77
|
+
|
78
|
+
tmp = _mm_alignr_epi8(first_len_a, prev_first_len, 13);
|
79
|
+
tmp = _mm_subs_epu8(tmp, _mm_set1_epi8(2));
|
80
|
+
range_a = _mm_or_si128(range_a, tmp);
|
81
|
+
|
82
|
+
__m128i shift1, pos, range2;
|
83
|
+
shift1 = _mm_alignr_epi8(input_a, prev_input, 15);
|
84
|
+
pos = _mm_sub_epi8(shift1, _mm_set1_epi8(0xEF));
|
85
|
+
tmp = _mm_subs_epu8(pos, _mm_set1_epi8(0xF0));
|
86
|
+
range2 = _mm_shuffle_epi8(df_ee_tbl, tmp);
|
87
|
+
tmp = _mm_adds_epu8(pos, _mm_set1_epi8(0x70));
|
88
|
+
range2 = _mm_add_epi8(range2, _mm_shuffle_epi8(ef_fe_tbl, tmp));
|
89
|
+
|
90
|
+
range_a = _mm_add_epi8(range_a, range2);
|
91
|
+
|
92
|
+
__m128i minv = _mm_shuffle_epi8(range_min_tbl, range_a);
|
93
|
+
__m128i maxv = _mm_shuffle_epi8(range_max_tbl, range_a);
|
94
|
+
|
95
|
+
tmp = _mm_or_si128(
|
96
|
+
_mm_cmplt_epi8(input_a, minv),
|
97
|
+
_mm_cmpgt_epi8(input_a, maxv)
|
98
|
+
);
|
99
|
+
error = _mm_or_si128(error, tmp);
|
100
|
+
|
101
|
+
/***************************** block 2 ****************************/
|
102
|
+
const __m128i input_b = _mm_loadu_si128((const __m128i *)(data+16));
|
103
|
+
|
104
|
+
high_nibbles =
|
105
|
+
_mm_and_si128(_mm_srli_epi16(input_b, 4), _mm_set1_epi8(0x0F));
|
106
|
+
|
107
|
+
__m128i first_len_b = _mm_shuffle_epi8(first_len_tbl, high_nibbles);
|
108
|
+
|
109
|
+
__m128i range_b = _mm_shuffle_epi8(first_range_tbl, high_nibbles);
|
110
|
+
|
111
|
+
range_b = _mm_or_si128(
|
112
|
+
range_b, _mm_alignr_epi8(first_len_b, first_len_a, 15));
|
113
|
+
|
114
|
+
|
115
|
+
tmp = _mm_alignr_epi8(first_len_b, first_len_a, 14);
|
116
|
+
tmp = _mm_subs_epu8(tmp, _mm_set1_epi8(1));
|
117
|
+
range_b = _mm_or_si128(range_b, tmp);
|
118
|
+
|
119
|
+
tmp = _mm_alignr_epi8(first_len_b, first_len_a, 13);
|
120
|
+
tmp = _mm_subs_epu8(tmp, _mm_set1_epi8(2));
|
121
|
+
range_b = _mm_or_si128(range_b, tmp);
|
122
|
+
|
123
|
+
shift1 = _mm_alignr_epi8(input_b, input_a, 15);
|
124
|
+
pos = _mm_sub_epi8(shift1, _mm_set1_epi8(0xEF));
|
125
|
+
tmp = _mm_subs_epu8(pos, _mm_set1_epi8(0xF0));
|
126
|
+
range2 = _mm_shuffle_epi8(df_ee_tbl, tmp);
|
127
|
+
tmp = _mm_adds_epu8(pos, _mm_set1_epi8(0x70));
|
128
|
+
range2 = _mm_add_epi8(range2, _mm_shuffle_epi8(ef_fe_tbl, tmp));
|
129
|
+
|
130
|
+
range_b = _mm_add_epi8(range_b, range2);
|
131
|
+
|
132
|
+
minv = _mm_shuffle_epi8(range_min_tbl, range_b);
|
133
|
+
maxv = _mm_shuffle_epi8(range_max_tbl, range_b);
|
134
|
+
|
135
|
+
|
136
|
+
tmp = _mm_or_si128(
|
137
|
+
_mm_cmplt_epi8(input_b, minv),
|
138
|
+
_mm_cmpgt_epi8(input_b, maxv)
|
139
|
+
);
|
140
|
+
error = _mm_or_si128(error, tmp);
|
141
|
+
|
142
|
+
/************************ next iteration **************************/
|
143
|
+
prev_input = input_b;
|
144
|
+
prev_first_len = first_len_b;
|
145
|
+
|
146
|
+
data += 32;
|
147
|
+
len -= 32;
|
148
|
+
}
|
149
|
+
|
150
|
+
if (!_mm_testz_si128(error, error))
|
151
|
+
return -1;
|
152
|
+
|
153
|
+
int32_t token4 = _mm_extract_epi32(prev_input, 3);
|
154
|
+
const int8_t *token = (const int8_t *)&token4;
|
155
|
+
int lookahead = 0;
|
156
|
+
if (token[3] > (int8_t)0xBF)
|
157
|
+
lookahead = 1;
|
158
|
+
else if (token[2] > (int8_t)0xBF)
|
159
|
+
lookahead = 2;
|
160
|
+
else if (token[1] > (int8_t)0xBF)
|
161
|
+
lookahead = 3;
|
162
|
+
|
163
|
+
data -= lookahead;
|
164
|
+
len += lookahead;
|
165
|
+
}
|
166
|
+
|
167
|
+
return utf8_naive(data, len);
|
168
|
+
}
|
169
|
+
|
170
|
+
#endif
|
Binary file
|
Binary file
|
Binary file
|
Binary file
|
Binary file
|
data/tests/basic.rb
CHANGED
@@ -66,8 +66,11 @@ module BasicTest
|
|
66
66
|
def test_issue_8559_crash
|
67
67
|
msg = TestMessage.new
|
68
68
|
msg.repeated_int32 = ::Google::Protobuf::RepeatedField.new(:int32, [1, 2, 3])
|
69
|
-
|
70
|
-
|
69
|
+
|
70
|
+
# https://github.com/jruby/jruby/issues/6818 was fixed in JRuby 9.3.0.0
|
71
|
+
if cruby_or_jruby_9_3_or_higher?
|
72
|
+
GC.start(full_mark: true, immediate_sweep: true)
|
73
|
+
end
|
71
74
|
TestMessage.encode(msg)
|
72
75
|
end
|
73
76
|
|
@@ -79,6 +82,34 @@ module BasicTest
|
|
79
82
|
assert_equal 8, msg.id
|
80
83
|
end
|
81
84
|
|
85
|
+
def test_issue_9507
|
86
|
+
pool = Google::Protobuf::DescriptorPool.new
|
87
|
+
pool.build do
|
88
|
+
add_message "NpeMessage" do
|
89
|
+
optional :type, :enum, 1, "TestEnum"
|
90
|
+
optional :other, :string, 2
|
91
|
+
end
|
92
|
+
add_enum "TestEnum" do
|
93
|
+
value :Something, 0
|
94
|
+
end
|
95
|
+
end
|
96
|
+
|
97
|
+
msgclass = pool.lookup("NpeMessage").msgclass
|
98
|
+
|
99
|
+
m = msgclass.new(
|
100
|
+
other: "foo" # must be set, but can be blank
|
101
|
+
)
|
102
|
+
|
103
|
+
begin
|
104
|
+
encoded = msgclass.encode(m)
|
105
|
+
rescue java.lang.NullPointerException
|
106
|
+
flunk "NPE rescued"
|
107
|
+
end
|
108
|
+
decoded = msgclass.decode(encoded)
|
109
|
+
decoded.inspect
|
110
|
+
decoded.to_proto
|
111
|
+
end
|
112
|
+
|
82
113
|
def test_has_field
|
83
114
|
m = TestSingularFields.new
|
84
115
|
assert !m.has_singular_msg?
|
@@ -145,7 +176,7 @@ module BasicTest
|
|
145
176
|
m = TestSingularFields.new
|
146
177
|
|
147
178
|
m.singular_int32 = -42
|
148
|
-
assert_equal -42, m.singular_int32
|
179
|
+
assert_equal( -42, m.singular_int32 )
|
149
180
|
m.clear_singular_int32
|
150
181
|
assert_equal 0, m.singular_int32
|
151
182
|
|
@@ -540,8 +571,6 @@ module BasicTest
|
|
540
571
|
|
541
572
|
|
542
573
|
def test_json_maps
|
543
|
-
# TODO: Fix JSON in JRuby version.
|
544
|
-
return if RUBY_PLATFORM == "java"
|
545
574
|
m = MapMessage.new(:map_string_int32 => {"a" => 1})
|
546
575
|
expected = {mapStringInt32: {a: 1}, mapStringMsg: {}, mapStringEnum: {}}
|
547
576
|
expected_preserve = {map_string_int32: {a: 1}, map_string_msg: {}, map_string_enum: {}}
|
@@ -555,8 +584,6 @@ module BasicTest
|
|
555
584
|
end
|
556
585
|
|
557
586
|
def test_json_maps_emit_defaults_submsg
|
558
|
-
# TODO: Fix JSON in JRuby version.
|
559
|
-
return if RUBY_PLATFORM == "java"
|
560
587
|
m = MapMessage.new(:map_string_msg => {"a" => TestMessage2.new(foo: 0)})
|
561
588
|
expected = {mapStringInt32: {}, mapStringMsg: {a: {foo: 0}}, mapStringEnum: {}}
|
562
589
|
|
@@ -566,8 +593,6 @@ module BasicTest
|
|
566
593
|
end
|
567
594
|
|
568
595
|
def test_json_emit_defaults_submsg
|
569
|
-
# TODO: Fix JSON in JRuby version.
|
570
|
-
return if RUBY_PLATFORM == "java"
|
571
596
|
m = TestSingularFields.new(singular_msg: proto_module::TestMessage2.new)
|
572
597
|
|
573
598
|
expected = {
|
@@ -590,8 +615,6 @@ module BasicTest
|
|
590
615
|
end
|
591
616
|
|
592
617
|
def test_respond_to
|
593
|
-
# This test fails with JRuby 1.7.23, likely because of an old JRuby bug.
|
594
|
-
return if RUBY_PLATFORM == "java"
|
595
618
|
msg = MapMessage.new
|
596
619
|
assert msg.respond_to?(:map_string_int32)
|
597
620
|
assert !msg.respond_to?(:bacon)
|
@@ -658,5 +681,59 @@ module BasicTest
|
|
658
681
|
assert_equal str, m.optional_string
|
659
682
|
assert_equal str, m.optional_bytes
|
660
683
|
end
|
684
|
+
|
685
|
+
def test_utf8
|
686
|
+
m = proto_module::TestMessage.new(
|
687
|
+
optional_string: "µpb",
|
688
|
+
)
|
689
|
+
m2 = proto_module::TestMessage.decode(proto_module::TestMessage.encode(m))
|
690
|
+
assert_equal m2, m
|
691
|
+
end
|
692
|
+
|
693
|
+
def test_map_fields_respond_to? # regression test for issue 9202
|
694
|
+
msg = proto_module::MapMessage.new
|
695
|
+
assert msg.respond_to?(:map_string_int32=)
|
696
|
+
msg.map_string_int32 = Google::Protobuf::Map.new(:string, :int32)
|
697
|
+
assert msg.respond_to?(:map_string_int32)
|
698
|
+
assert_equal( Google::Protobuf::Map.new(:string, :int32), msg.map_string_int32 )
|
699
|
+
assert msg.respond_to?(:clear_map_string_int32)
|
700
|
+
msg.clear_map_string_int32
|
701
|
+
|
702
|
+
assert !msg.respond_to?(:has_map_string_int32?)
|
703
|
+
assert_raise NoMethodError do
|
704
|
+
msg.has_map_string_int32?
|
705
|
+
end
|
706
|
+
assert !msg.respond_to?(:map_string_int32_as_value)
|
707
|
+
assert_raise NoMethodError do
|
708
|
+
msg.map_string_int32_as_value
|
709
|
+
end
|
710
|
+
assert !msg.respond_to?(:map_string_int32_as_value=)
|
711
|
+
assert_raise NoMethodError do
|
712
|
+
msg.map_string_int32_as_value = :boom
|
713
|
+
end
|
714
|
+
end
|
715
|
+
end
|
716
|
+
|
717
|
+
def test_oneof_fields_respond_to? # regression test for issue 9202
|
718
|
+
msg = proto_module::OneofMessage.new
|
719
|
+
# `has_` prefix + "?" suffix actions should only work for oneofs fields.
|
720
|
+
assert msg.has_my_oneof?
|
721
|
+
assert msg.respond_to? :has_my_oneof?
|
722
|
+
assert !msg.respond_to?( :has_a? )
|
723
|
+
assert_raise NoMethodError do
|
724
|
+
msg.has_a?
|
725
|
+
end
|
726
|
+
assert !msg.respond_to?( :has_b? )
|
727
|
+
assert_raise NoMethodError do
|
728
|
+
msg.has_b?
|
729
|
+
end
|
730
|
+
assert !msg.respond_to?( :has_c? )
|
731
|
+
assert_raise NoMethodError do
|
732
|
+
msg.has_c?
|
733
|
+
end
|
734
|
+
assert !msg.respond_to?( :has_d? )
|
735
|
+
assert_raise NoMethodError do
|
736
|
+
msg.has_d?
|
737
|
+
end
|
661
738
|
end
|
662
739
|
end
|
metadata
CHANGED
@@ -1,14 +1,14 @@
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
2
2
|
name: google-protobuf
|
3
3
|
version: !ruby/object:Gem::Version
|
4
|
-
version: 3.20.
|
4
|
+
version: 3.20.1.rc.1
|
5
5
|
platform: x86-mingw32
|
6
6
|
authors:
|
7
7
|
- Protobuf Authors
|
8
|
-
autorequire:
|
8
|
+
autorequire:
|
9
9
|
bindir: bin
|
10
10
|
cert_chain: []
|
11
|
-
date: 2022-
|
11
|
+
date: 2022-04-06 00:00:00.000000000 Z
|
12
12
|
dependencies:
|
13
13
|
- !ruby/object:Gem::Dependency
|
14
14
|
name: rake-compiler-dock
|
@@ -16,14 +16,14 @@ dependencies:
|
|
16
16
|
requirements:
|
17
17
|
- - '='
|
18
18
|
- !ruby/object:Gem::Version
|
19
|
-
version: 1.1
|
19
|
+
version: 1.2.1
|
20
20
|
type: :development
|
21
21
|
prerelease: false
|
22
22
|
version_requirements: !ruby/object:Gem::Requirement
|
23
23
|
requirements:
|
24
24
|
- - '='
|
25
25
|
- !ruby/object:Gem::Version
|
26
|
-
version: 1.1
|
26
|
+
version: 1.2.1
|
27
27
|
- !ruby/object:Gem::Dependency
|
28
28
|
name: rake-compiler
|
29
29
|
requirement: !ruby/object:Gem::Requirement
|
@@ -80,13 +80,16 @@ files:
|
|
80
80
|
- ext/google/protobuf_c/ruby-upb.c
|
81
81
|
- ext/google/protobuf_c/ruby-upb.h
|
82
82
|
- ext/google/protobuf_c/third_party/utf8_range/LICENSE
|
83
|
-
- ext/google/protobuf_c/third_party/utf8_range/
|
83
|
+
- ext/google/protobuf_c/third_party/utf8_range/naive.c
|
84
|
+
- ext/google/protobuf_c/third_party/utf8_range/range2-neon.c
|
85
|
+
- ext/google/protobuf_c/third_party/utf8_range/range2-sse.c
|
84
86
|
- ext/google/protobuf_c/third_party/utf8_range/utf8_range.h
|
85
87
|
- ext/google/protobuf_c/wrap_memcpy.c
|
86
88
|
- lib/google/2.5/protobuf_c.so
|
87
89
|
- lib/google/2.6/protobuf_c.so
|
88
90
|
- lib/google/2.7/protobuf_c.so
|
89
91
|
- lib/google/3.0/protobuf_c.so
|
92
|
+
- lib/google/3.1/protobuf_c.so
|
90
93
|
- lib/google/protobuf.rb
|
91
94
|
- lib/google/protobuf/any_pb.rb
|
92
95
|
- lib/google/protobuf/api_pb.rb
|
@@ -110,8 +113,8 @@ homepage: https://developers.google.com/protocol-buffers
|
|
110
113
|
licenses:
|
111
114
|
- BSD-3-Clause
|
112
115
|
metadata:
|
113
|
-
source_code_uri: https://github.com/protocolbuffers/protobuf/tree/v3.20.
|
114
|
-
post_install_message:
|
116
|
+
source_code_uri: https://github.com/protocolbuffers/protobuf/tree/v3.20.1-rc1/ruby
|
117
|
+
post_install_message:
|
115
118
|
rdoc_options: []
|
116
119
|
require_paths:
|
117
120
|
- lib
|
@@ -122,15 +125,15 @@ required_ruby_version: !ruby/object:Gem::Requirement
|
|
122
125
|
version: '2.5'
|
123
126
|
- - "<"
|
124
127
|
- !ruby/object:Gem::Version
|
125
|
-
version: 3.
|
128
|
+
version: 3.2.dev
|
126
129
|
required_rubygems_version: !ruby/object:Gem::Requirement
|
127
130
|
requirements:
|
128
131
|
- - ">"
|
129
132
|
- !ruby/object:Gem::Version
|
130
133
|
version: 1.3.1
|
131
134
|
requirements: []
|
132
|
-
rubygems_version: 3.
|
133
|
-
signing_key:
|
135
|
+
rubygems_version: 3.3.4
|
136
|
+
signing_key:
|
134
137
|
specification_version: 4
|
135
138
|
summary: Protocol Buffers
|
136
139
|
test_files:
|