ox 2.14.14 → 2.14.15
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- checksums.yaml +4 -4
- data/CHANGELOG.md +6 -0
- data/README.md +1 -1
- data/ext/ox/attr.h +33 -39
- data/ext/ox/base64.c +48 -42
- data/ext/ox/base64.h +4 -4
- data/ext/ox/buf.h +80 -86
- data/ext/ox/builder.c +378 -423
- data/ext/ox/cache.c +2 -2
- data/ext/ox/cache8.c +37 -40
- data/ext/ox/cache8.h +7 -7
- data/ext/ox/dump.c +838 -867
- data/ext/ox/err.c +16 -13
- data/ext/ox/err.h +11 -12
- data/ext/ox/extconf.rb +5 -5
- data/ext/ox/gen_load.c +135 -137
- data/ext/ox/hash_load.c +130 -148
- data/ext/ox/helper.h +32 -39
- data/ext/ox/intern.c +1 -2
- data/ext/ox/obj_load.c +590 -644
- data/ext/ox/ox.c +2 -2
- data/ext/ox/ox.h +5 -5
- data/ext/ox/parse.c +836 -874
- data/ext/ox/sax.c +38 -23
- data/ext/ox/sax.h +2 -2
- data/ext/ox/sax_as.c +78 -94
- data/ext/ox/sax_buf.c +85 -94
- data/ext/ox/sax_buf.h +101 -120
- data/ext/ox/sax_hint.c +175 -184
- data/ext/ox/sax_hint.h +19 -19
- data/ext/ox/sax_stack.h +59 -45
- data/ext/ox/slotcache.c +2 -2
- data/ext/ox/slotcache.h +4 -4
- data/ext/ox/special.c +320 -327
- data/ext/ox/special.h +2 -2
- data/ext/ox/type.h +19 -19
- data/lib/ox/bag.rb +13 -9
- data/lib/ox/cdata.rb +0 -2
- data/lib/ox/comment.rb +0 -2
- data/lib/ox/doctype.rb +0 -2
- data/lib/ox/document.rb +3 -5
- data/lib/ox/element.rb +41 -26
- data/lib/ox/error.rb +0 -3
- data/lib/ox/hasattrs.rb +7 -8
- data/lib/ox/instruct.rb +4 -6
- data/lib/ox/node.rb +3 -4
- data/lib/ox/raw.rb +0 -2
- data/lib/ox/sax.rb +20 -36
- data/lib/ox/version.rb +1 -2
- data/lib/ox/xmlrpc_adapter.rb +4 -6
- data/lib/ox.rb +15 -16
- metadata +6 -5
data/ext/ox/special.c
CHANGED
|
@@ -3,395 +3,388 @@
|
|
|
3
3
|
* All rights reserved.
|
|
4
4
|
*/
|
|
5
5
|
|
|
6
|
-
#include <string.h>
|
|
7
|
-
#include <stdbool.h>
|
|
8
|
-
|
|
9
6
|
#include "special.h"
|
|
10
7
|
|
|
8
|
+
#include <stdbool.h>
|
|
9
|
+
#include <string.h>
|
|
10
|
+
|
|
11
11
|
/*
|
|
12
12
|
u0000..u007F 00000000000000xxxxxxx 0xxxxxxx
|
|
13
13
|
u0080..u07FF 0000000000yyyyyxxxxxx 110yyyyy 10xxxxxx
|
|
14
14
|
u0800..uD7FF, uE000..uFFFF 00000zzzzyyyyyyxxxxxx 1110zzzz 10yyyyyy 10xxxxxx
|
|
15
15
|
u10000..u10FFFF uuuzzzzzzyyyyyyxxxxxx 11110uuu 10zzzzzz 10yyyyyy 10xxxxxx
|
|
16
16
|
*/
|
|
17
|
-
char*
|
|
18
|
-
|
|
19
|
-
int
|
|
20
|
-
|
|
21
|
-
unsigned char c;
|
|
17
|
+
char *ox_ucs_to_utf8_chars(char *text, uint64_t u) {
|
|
18
|
+
int reading = 0;
|
|
19
|
+
int i;
|
|
20
|
+
unsigned char c;
|
|
22
21
|
|
|
23
22
|
if (u <= 0x000000000000007FULL) {
|
|
24
|
-
|
|
25
|
-
|
|
23
|
+
/* 0xxxxxxx */
|
|
24
|
+
*text++ = (char)u;
|
|
26
25
|
} else if (u <= 0x00000000000007FFULL) {
|
|
27
|
-
|
|
28
|
-
|
|
29
|
-
|
|
26
|
+
/* 110yyyyy 10xxxxxx */
|
|
27
|
+
*text++ = (char)(0x00000000000000C0ULL | (0x000000000000001FULL & (u >> 6)));
|
|
28
|
+
*text++ = (char)(0x0000000000000080ULL | (0x000000000000003FULL & u));
|
|
30
29
|
} else if (u <= 0x000000000000D7FFULL || (0x000000000000E000ULL <= u && u <= 0x000000000000FFFFULL)) {
|
|
31
|
-
|
|
32
|
-
|
|
33
|
-
|
|
34
|
-
|
|
30
|
+
/* 1110zzzz 10yyyyyy 10xxxxxx */
|
|
31
|
+
*text++ = (char)(0x00000000000000E0ULL | (0x000000000000000FULL & (u >> 12)));
|
|
32
|
+
*text++ = (char)(0x0000000000000080ULL | (0x000000000000003FULL & (u >> 6)));
|
|
33
|
+
*text++ = (char)(0x0000000000000080ULL | (0x000000000000003FULL & u));
|
|
35
34
|
} else if (0x0000000000010000ULL <= u && u <= 0x000000000010FFFFULL) {
|
|
36
|
-
|
|
37
|
-
|
|
38
|
-
|
|
39
|
-
|
|
40
|
-
|
|
35
|
+
/* 11110uuu 10zzzzzz 10yyyyyy 10xxxxxx */
|
|
36
|
+
*text++ = (char)(0x00000000000000F0ULL | (0x0000000000000007ULL & (u >> 18)));
|
|
37
|
+
*text++ = (char)(0x0000000000000080ULL | (0x000000000000003FULL & (u >> 12)));
|
|
38
|
+
*text++ = (char)(0x0000000000000080ULL | (0x000000000000003FULL & (u >> 6)));
|
|
39
|
+
*text++ = (char)(0x0000000000000080ULL | (0x000000000000003FULL & u));
|
|
41
40
|
} else {
|
|
42
|
-
|
|
43
|
-
|
|
44
|
-
|
|
45
|
-
|
|
46
|
-
|
|
47
|
-
|
|
48
|
-
|
|
49
|
-
|
|
50
|
-
|
|
51
|
-
|
|
41
|
+
/* assume it is UTF-8 encoded directly and not UCS */
|
|
42
|
+
for (i = 56; 0 <= i; i -= 8) {
|
|
43
|
+
c = (unsigned char)((u >> i) & 0x00000000000000FFULL);
|
|
44
|
+
if (reading) {
|
|
45
|
+
*text++ = (char)c;
|
|
46
|
+
} else if ('\0' != c) {
|
|
47
|
+
*text++ = (char)c;
|
|
48
|
+
reading = 1;
|
|
49
|
+
}
|
|
50
|
+
}
|
|
52
51
|
}
|
|
53
52
|
return text;
|
|
54
53
|
}
|
|
55
54
|
|
|
56
|
-
#define BUCKET_SIZE
|
|
57
|
-
#define BUCKET_MASK
|
|
55
|
+
#define BUCKET_SIZE 256
|
|
56
|
+
#define BUCKET_MASK 255
|
|
58
57
|
|
|
59
58
|
typedef struct _slot {
|
|
60
|
-
const char
|
|
61
|
-
uint64_t
|
|
62
|
-
struct _slot
|
|
63
|
-
uint64_t
|
|
59
|
+
const char *key;
|
|
60
|
+
uint64_t code;
|
|
61
|
+
struct _slot *next;
|
|
62
|
+
uint64_t hash;
|
|
64
63
|
} *Slot;
|
|
65
64
|
|
|
66
65
|
typedef struct _cache {
|
|
67
|
-
Slot
|
|
66
|
+
Slot buckets[BUCKET_SIZE];
|
|
68
67
|
} *Cache;
|
|
69
68
|
|
|
70
|
-
static struct _cache
|
|
71
|
-
static bool
|
|
69
|
+
static struct _cache entity_cache;
|
|
70
|
+
static bool inited = false;
|
|
72
71
|
|
|
73
72
|
// HTML entities such as &. This is a complete list from the HTML 5 spec.
|
|
74
|
-
static struct _slot
|
|
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
|
-
{
|
|
224
|
-
{
|
|
225
|
-
{
|
|
226
|
-
{
|
|
227
|
-
{
|
|
228
|
-
{
|
|
229
|
-
{
|
|
230
|
-
{
|
|
231
|
-
{
|
|
232
|
-
{
|
|
233
|
-
{
|
|
234
|
-
{
|
|
235
|
-
{
|
|
236
|
-
{
|
|
237
|
-
{
|
|
238
|
-
{
|
|
239
|
-
{
|
|
240
|
-
{
|
|
241
|
-
{
|
|
242
|
-
{
|
|
243
|
-
{
|
|
244
|
-
{
|
|
245
|
-
{
|
|
246
|
-
{
|
|
247
|
-
{
|
|
248
|
-
{
|
|
249
|
-
{
|
|
250
|
-
{
|
|
251
|
-
{
|
|
252
|
-
{
|
|
253
|
-
{
|
|
254
|
-
{
|
|
255
|
-
{
|
|
256
|
-
{
|
|
257
|
-
{
|
|
258
|
-
{
|
|
259
|
-
{
|
|
260
|
-
{
|
|
261
|
-
{
|
|
262
|
-
{
|
|
263
|
-
{
|
|
264
|
-
{
|
|
265
|
-
{
|
|
266
|
-
{
|
|
267
|
-
{
|
|
268
|
-
{
|
|
269
|
-
{
|
|
270
|
-
{
|
|
271
|
-
{
|
|
272
|
-
{
|
|
273
|
-
{
|
|
274
|
-
{
|
|
275
|
-
{
|
|
276
|
-
{
|
|
277
|
-
{
|
|
278
|
-
{
|
|
279
|
-
{
|
|
280
|
-
{
|
|
281
|
-
{
|
|
282
|
-
{
|
|
283
|
-
{
|
|
284
|
-
{
|
|
285
|
-
{
|
|
286
|
-
{
|
|
287
|
-
{
|
|
288
|
-
{
|
|
289
|
-
{
|
|
290
|
-
{
|
|
291
|
-
{
|
|
292
|
-
{
|
|
293
|
-
{
|
|
294
|
-
{
|
|
295
|
-
{
|
|
296
|
-
{
|
|
297
|
-
{
|
|
298
|
-
{
|
|
299
|
-
{
|
|
300
|
-
{
|
|
301
|
-
{
|
|
302
|
-
{
|
|
303
|
-
{
|
|
304
|
-
{
|
|
305
|
-
{
|
|
306
|
-
{
|
|
307
|
-
{
|
|
308
|
-
{
|
|
309
|
-
{
|
|
310
|
-
{
|
|
311
|
-
{
|
|
312
|
-
{
|
|
313
|
-
{
|
|
314
|
-
{
|
|
315
|
-
{
|
|
316
|
-
{
|
|
317
|
-
{
|
|
318
|
-
{
|
|
319
|
-
{
|
|
320
|
-
{
|
|
321
|
-
{
|
|
322
|
-
{
|
|
323
|
-
{
|
|
324
|
-
{
|
|
325
|
-
{
|
|
326
|
-
{
|
|
327
|
-
{
|
|
328
|
-
{
|
|
73
|
+
static struct _slot entities[] = {
|
|
74
|
+
{"AElig", 198}, // latin capital letter AE
|
|
75
|
+
{"Aacute", 193}, // latin capital letter A with acute
|
|
76
|
+
{"Acirc", 194}, // latin capital letter A with circumflex
|
|
77
|
+
{"Agrave", 192}, // latin capital letter A with grave
|
|
78
|
+
{"Alpha", 913}, // greek capital letter alpha, U+0391
|
|
79
|
+
{"Aring", 197}, // latin capital letter A with ring above
|
|
80
|
+
{"Atilde", 195}, // latin capital letter A with tilde
|
|
81
|
+
{"Auml", 196}, // latin capital letter A with diaeresis
|
|
82
|
+
{"Beta", 914}, // greek capital letter beta, U+0392
|
|
83
|
+
{"Ccedil", 199}, // latin capital letter C with cedilla
|
|
84
|
+
{"Chi", 935}, // greek capital letter chi, U+03A7
|
|
85
|
+
{"Dagger", 8225}, // double dagger, U+2021 ISOpub
|
|
86
|
+
{"Delta", 916}, // greek capital letter delta
|
|
87
|
+
{"ETH", 208}, // latin capital letter ETH, U+00D0 ISOlat1
|
|
88
|
+
{"Eacute", 201}, // latin capital letter E with acute
|
|
89
|
+
{"Ecirc", 202}, // latin capital letter E with circumflex
|
|
90
|
+
{"Egrave", 200}, // latin capital letter E with grave
|
|
91
|
+
{"Epsilon", 917}, // greek capital letter epsilon, U+0395
|
|
92
|
+
{"Eta", 919}, // greek capital letter eta, U+0397
|
|
93
|
+
{"Euml", 203}, // latin capital letter E with diaeresis
|
|
94
|
+
{"Gamma", 915}, // greek capital letter gamma
|
|
95
|
+
{"Iacute", 205}, // latin capital letter I with acute
|
|
96
|
+
{"Icirc", 206}, // latin capital letter I with circumflex
|
|
97
|
+
{"Igrave", 204}, // latin capital letter I with grave
|
|
98
|
+
{"Iota", 921}, // greek capital letter iota, U+0399
|
|
99
|
+
{"Iuml", 207}, // latin capital letter I with diaeresis
|
|
100
|
+
{"Kappa", 922}, // greek capital letter kappa, U+039A
|
|
101
|
+
{"Lambda", 923}, // greek capital letter lambda
|
|
102
|
+
{"Mu", 924}, // greek capital letter mu, U+039C
|
|
103
|
+
{"Ntilde", 209}, // latin capital letter N with tilde
|
|
104
|
+
{"Nu", 925}, // greek capital letter nu, U+039D
|
|
105
|
+
{"OElig", 338}, // - latin capital ligature OE
|
|
106
|
+
{"Oacute", 211}, // latin capital letter O with acute
|
|
107
|
+
{"Ocirc", 212}, // latin capital letter O with circumflex
|
|
108
|
+
{"Ograve", 210}, // latin capital letter O with grave
|
|
109
|
+
{"Omega", 937}, // greek capital letter omega
|
|
110
|
+
{"Omicron", 927}, // greek capital letter omicron, U+039F
|
|
111
|
+
{"Oslash", 216}, // latin capital letter O with stroke
|
|
112
|
+
{"Otilde", 213}, // latin capital letter O with tilde
|
|
113
|
+
{"Ouml", 214}, // latin capital letter O with diaeresis
|
|
114
|
+
{"Phi", 934}, // greek capital letter phi
|
|
115
|
+
{"Pi", 928}, // greek capital letter pi, U+03A0 ISOgrk3
|
|
116
|
+
{"Prime", 8243}, // double prime = seconds = inches
|
|
117
|
+
{"Psi", 936}, // greek capital letter psi
|
|
118
|
+
{"Rho", 929}, // greek capital letter rho, U+03A1
|
|
119
|
+
{"Scaron", 352}, // - latin capital letter S with caron
|
|
120
|
+
{"Sigma", 931}, // greek capital letter sigma
|
|
121
|
+
{"THORN", 222}, // latin capital letter THORN
|
|
122
|
+
{"Tau", 932}, // greek capital letter tau, U+03A4
|
|
123
|
+
{"Theta", 920}, // greek capital letter theta
|
|
124
|
+
{"Uacute", 218}, // latin capital letter U with acute
|
|
125
|
+
{"Ucirc", 219}, // latin capital letter U with circumflex
|
|
126
|
+
{"Ugrave", 217}, // latin capital letter U with grave
|
|
127
|
+
{"Upsilon", 933}, // greek capital letter upsilon
|
|
128
|
+
{"Uuml", 220}, // latin capital letter U with diaeresis
|
|
129
|
+
{"Xi", 926}, // greek capital letter xi, U+039E ISOgrk3
|
|
130
|
+
{"Yacute", 221}, // latin capital letter Y with acute
|
|
131
|
+
{"Yuml", 376}, // - latin capital letter Y with diaeresis
|
|
132
|
+
{"Zeta", 918}, // greek capital letter zeta, U+0396
|
|
133
|
+
{"aacute", 225}, // latin small letter a with acute
|
|
134
|
+
{"acirc", 226}, // latin small letter a with circumflex
|
|
135
|
+
{"acute", 180}, // acute accent = spacing acute
|
|
136
|
+
{"aelig", 230}, // latin small letter ae
|
|
137
|
+
{"agrave", 224}, // latin small letter a with grave
|
|
138
|
+
{"alefsym", 8501}, // alef symbol = first transfinite cardinal
|
|
139
|
+
{"alpha", 945}, // greek small letter alpha
|
|
140
|
+
{"amp", 38}, // -- ampersand, U+0026 ISOnum
|
|
141
|
+
{"and", 8743}, // logical and = wedge, U+2227 ISOtech
|
|
142
|
+
{"ang", 8736}, // angle, U+2220 ISOamso
|
|
143
|
+
{"apos", 39}, // -- single quote
|
|
144
|
+
{"aring", 229}, // latin small letter a with ring above
|
|
145
|
+
{"asymp", 8776}, // almost equal to = asymptotic to
|
|
146
|
+
{"atilde", 227}, // latin small letter a with tilde
|
|
147
|
+
{"auml", 228}, // latin small letter a with diaeresis
|
|
148
|
+
{"bdquo", 8222}, // double low-9 quotation mark, U+201E NEW
|
|
149
|
+
{"beta", 946}, // greek small letter beta, U+03B2 ISOgrk3
|
|
150
|
+
{"brvbar", 166}, // broken bar = broken vertical bar
|
|
151
|
+
{"bull", 8226}, // bullet = black small circle
|
|
152
|
+
{"cap", 8745}, // intersection = cap, U+2229 ISOtech
|
|
153
|
+
{"ccedil", 231}, // latin small letter c with cedilla
|
|
154
|
+
{"cedil", 184}, // cedilla = spacing cedilla, U+00B8 ISOdia
|
|
155
|
+
{"cent", 162}, // cent sign, U+00A2 ISOnum
|
|
156
|
+
{"chi", 967}, // greek small letter chi, U+03C7 ISOgrk3
|
|
157
|
+
{"circ", 710}, // - modifier letter circumflex accent
|
|
158
|
+
{"clubs", 9827}, // black club suit = shamrock
|
|
159
|
+
{"cong", 8773}, // approximately equal to, U+2245 ISOtech
|
|
160
|
+
{"copy", 169}, // copyright sign, U+00A9 ISOnum
|
|
161
|
+
{"crarr", 8629}, // downwards arrow with corner leftwards
|
|
162
|
+
{"cup", 8746}, // union = cup, U+222A ISOtech
|
|
163
|
+
{"curren", 164}, // currency sign, U+00A4 ISOnum
|
|
164
|
+
{"dArr", 8659}, // downwards double arrow, U+21D3 ISOamsa
|
|
165
|
+
{"dagger", 8224}, // dagger, U+2020 ISOpub
|
|
166
|
+
{"darr", 8595}, // downwards arrow, U+2193 ISOnum
|
|
167
|
+
{"deg", 176}, // degree sign, U+00B0 ISOnum
|
|
168
|
+
{"delta", 948}, // greek small letter delta
|
|
169
|
+
{"diams", 9830}, // black diamond suit, U+2666 ISOpub
|
|
170
|
+
{"divide", 247}, // division sign, U+00F7 ISOnum
|
|
171
|
+
{"eacute", 233}, // latin small letter e with acute
|
|
172
|
+
{"ecirc", 234}, // latin small letter e with circumflex
|
|
173
|
+
{"egrave", 232}, // latin small letter e with grave
|
|
174
|
+
{"empty", 8709}, // empty set = null set = diameter
|
|
175
|
+
{"emsp", 8195}, // em space, U+2003 ISOpub
|
|
176
|
+
{"ensp", 8194}, // en space, U+2002 ISOpub
|
|
177
|
+
{"epsilon", 949}, // greek small letter epsilon
|
|
178
|
+
{"equiv", 8801}, // identical to, U+2261 ISOtech
|
|
179
|
+
{"eta", 951}, // greek small letter eta, U+03B7 ISOgrk3
|
|
180
|
+
{"eth", 240}, // latin small letter eth, U+00F0 ISOlat1
|
|
181
|
+
{"euml", 235}, // latin small letter e with diaeresis
|
|
182
|
+
{"euro", 8364}, // - euro sign, U+20AC NEW
|
|
183
|
+
{"exist", 8707}, // there exists, U+2203 ISOtech
|
|
184
|
+
{"fnof", 402}, // latin small f with hook = function
|
|
185
|
+
{"forall", 8704}, // for all, U+2200 ISOtech
|
|
186
|
+
{"frac12", 189}, // vulgar fraction one half
|
|
187
|
+
{"frac14", 188}, // vulgar fraction one quarter
|
|
188
|
+
{"frac34", 190}, // vulgar fraction three quarters
|
|
189
|
+
{"frasl", 8260}, // fraction slash, U+2044 NEW
|
|
190
|
+
{"gamma", 947}, // greek small letter gamma
|
|
191
|
+
{"ge", 8805}, // greater-than or equal to
|
|
192
|
+
{"gt", 62}, // -- greater-than sign, U+003E ISOnum
|
|
193
|
+
{"hArr", 8660}, // left right double arrow
|
|
194
|
+
{"harr", 8596}, // left right arrow, U+2194 ISOamsa
|
|
195
|
+
{"hearts", 9829}, // black heart suit = valentine
|
|
196
|
+
{"hellip", 8230}, // horizontal ellipsis = three dot leader
|
|
197
|
+
{"iacute", 237}, // latin small letter i with acute
|
|
198
|
+
{"icirc", 238}, // latin small letter i with circumflex
|
|
199
|
+
{"iexcl", 161}, // inverted exclamation mark, U+00A1 ISOnum
|
|
200
|
+
{"igrave", 236}, // latin small letter i with grave
|
|
201
|
+
{"image", 8465}, // blackletter capital I = imaginary part
|
|
202
|
+
{"infin", 8734}, // infinity, U+221E ISOtech
|
|
203
|
+
{"int", 8747}, // integral, U+222B ISOtech
|
|
204
|
+
{"iota", 953}, // greek small letter iota, U+03B9 ISOgrk3
|
|
205
|
+
{"iquest", 191}, // inverted question mark
|
|
206
|
+
{"isin", 8712}, // element of, U+2208 ISOtech
|
|
207
|
+
{"iuml", 239}, // latin small letter i with diaeresis
|
|
208
|
+
{"kappa", 954}, // greek small letter kappa
|
|
209
|
+
{"lArr", 8656}, // leftwards double arrow, U+21D0 ISOtech
|
|
210
|
+
{"lambda", 955}, // greek small letter lambda
|
|
211
|
+
{"lang", 9001}, // left-pointing angle bracket = bra
|
|
212
|
+
{"laquo", 171}, // left-pointing double angle quotation mark
|
|
213
|
+
{"larr", 8592}, // leftwards arrow, U+2190 ISOnum
|
|
214
|
+
{"lceil", 8968}, // left ceiling = apl upstile
|
|
215
|
+
{"ldquo", 8220}, // left double quotation mark
|
|
216
|
+
{"le", 8804}, // less-than or equal to, U+2264 ISOtech
|
|
217
|
+
{"lfloor", 8970}, // left floor = apl downstile
|
|
218
|
+
{"lowast", 8727}, // asterisk operator, U+2217 ISOtech
|
|
219
|
+
{"loz", 9674}, // lozenge, U+25CA ISOpub
|
|
220
|
+
{"lrm", 8206}, // left-to-right mark, U+200E NEW RFC 2070
|
|
221
|
+
{"lsaquo", 8249}, // single left-pointing angle quotation mark
|
|
222
|
+
{"lsquo", 8216}, // left single quotation mark
|
|
223
|
+
{"lt", 60}, // -- less-than sign, U+003C ISOnum
|
|
224
|
+
{"macr", 175}, // macron = spacing macron = overline
|
|
225
|
+
{"mdash", 8212}, // em dash, U+2014 ISOpub
|
|
226
|
+
{"micro", 181}, // micro sign, U+00B5 ISOnum
|
|
227
|
+
{"middot", 183}, // middle dot = Georgian comma
|
|
228
|
+
{"minus", 8722}, // minus sign, U+2212 ISOtech
|
|
229
|
+
{"mu", 956}, // greek small letter mu, U+03BC ISOgrk3
|
|
230
|
+
{"nabla", 8711}, // nabla = backward difference
|
|
231
|
+
{"nbsp", 160}, // no-break space = non-breaking space
|
|
232
|
+
{"ndash", 8211}, // en dash, U+2013 ISOpub
|
|
233
|
+
{"ne", 8800}, // not equal to, U+2260 ISOtech
|
|
234
|
+
{"ni", 8715}, // contains as member, U+220B ISOtech
|
|
235
|
+
{"not", 172}, // not sign, U+00AC ISOnum
|
|
236
|
+
{"notin", 8713}, // not an element of, U+2209 ISOtech
|
|
237
|
+
{"nsub", 8836}, // not a subset of, U+2284 ISOamsn
|
|
238
|
+
{"ntilde", 241}, // latin small letter n with tilde
|
|
239
|
+
{"nu", 957}, // greek small letter nu, U+03BD ISOgrk3
|
|
240
|
+
{"oacute", 243}, // latin small letter o with acute
|
|
241
|
+
{"ocirc", 244}, // latin small letter o with circumflex
|
|
242
|
+
{"oelig", 339}, // - latin small ligature oe, U+0153 ISOlat2
|
|
243
|
+
{"ograve", 242}, // latin small letter o with grave
|
|
244
|
+
{"oline", 8254}, // overline = spacing overscore
|
|
245
|
+
{"omega", 969}, // greek small letter omega
|
|
246
|
+
{"omicron", 959}, // greek small letter omicron, U+03BF NEW
|
|
247
|
+
{"oplus", 8853}, // circled plus = direct sum
|
|
248
|
+
{"or", 8744}, // logical or = vee, U+2228 ISOtech
|
|
249
|
+
{"ordf", 170}, // feminine ordinal indicator, U+00AA ISOnum
|
|
250
|
+
{"ordm", 186}, // masculine ordinal indicator
|
|
251
|
+
{"oslash", 248}, // latin small letter o with stroke
|
|
252
|
+
{"otilde", 245}, // latin small letter o with tilde
|
|
253
|
+
{"otimes", 8855}, // circled times = vector product
|
|
254
|
+
{"ouml", 246}, // latin small letter o with diaeresis
|
|
255
|
+
{"para", 182}, // pilcrow sign = paragraph sign
|
|
256
|
+
{"part", 8706}, // partial differential, U+2202 ISOtech
|
|
257
|
+
{"permil", 8240}, // per mille sign, U+2030 ISOtech
|
|
258
|
+
{"perp", 8869}, // up tack = orthogonal to = perpendicular
|
|
259
|
+
{"phi", 966}, // greek small letter phi, U+03C6 ISOgrk3
|
|
260
|
+
{"pi", 960}, // greek small letter pi, U+03C0 ISOgrk3
|
|
261
|
+
{"piv", 982}, // greek pi symbol, U+03D6 ISOgrk3
|
|
262
|
+
{"plusmn", 177}, // plus-minus sign = plus-or-minus sign
|
|
263
|
+
{"pound", 163}, // pound sign, U+00A3 ISOnum
|
|
264
|
+
{"prime", 8242}, // prime = minutes = feet, U+2032 ISOtech
|
|
265
|
+
{"prod", 8719}, // n-ary product = product sign
|
|
266
|
+
{"prop", 8733}, // proportional to, U+221D ISOtech
|
|
267
|
+
{"psi", 968}, // greek small letter psi, U+03C8 ISOgrk3
|
|
268
|
+
{"quot", 34}, // -- quotation mark = APL quote
|
|
269
|
+
{"rArr", 8658}, // rightwards double arrow
|
|
270
|
+
{"radic", 8730}, // square root = radical sign
|
|
271
|
+
{"rang", 9002}, // right-pointing angle bracket = ket
|
|
272
|
+
{"raquo", 187}, // right-pointing double angle quotation mark
|
|
273
|
+
{"rarr", 8594}, // rightwards arrow, U+2192 ISOnum
|
|
274
|
+
{"rceil", 8969}, // right ceiling, U+2309 ISOamsc
|
|
275
|
+
{"rdquo", 8221}, // right double quotation mark
|
|
276
|
+
{"real", 8476}, // blackletter capital R = real part symbol
|
|
277
|
+
{"reg", 174}, // registered sign = registered trade mark sign
|
|
278
|
+
{"rfloor", 8971}, // right floor, U+230B ISOamsc
|
|
279
|
+
{"rho", 961}, // greek small letter rho, U+03C1 ISOgrk3
|
|
280
|
+
{"rlm", 8207}, // right-to-left mark, U+200F NEW RFC 2070
|
|
281
|
+
{"rsaquo", 8250}, // single right-pointing angle quotation mark
|
|
282
|
+
{"rsquo", 8217}, // right single quotation mark
|
|
283
|
+
{"sbquo", 8218}, // single low-9 quotation mark, U+201A NEW
|
|
284
|
+
{"scaron", 353}, // - latin small letter s with caron
|
|
285
|
+
{"sdot", 8901}, // dot operator, U+22C5 ISOamsb
|
|
286
|
+
{"sect", 167}, // section sign, U+00A7 ISOnum
|
|
287
|
+
{"shy", 173}, // soft hyphen = discretionary hyphen
|
|
288
|
+
{"sigma", 963}, // greek small letter sigma
|
|
289
|
+
{"sigmaf", 962}, // greek small letter final sigma
|
|
290
|
+
{"sim", 8764}, // tilde operator = varies with = similar to
|
|
291
|
+
{"spades", 9824}, // black spade suit, U+2660 ISOpub
|
|
292
|
+
{"sub", 8834}, // subset of, U+2282 ISOtech
|
|
293
|
+
{"sube", 8838}, // subset of or equal to, U+2286 ISOtech
|
|
294
|
+
{"sum", 8721}, // n-ary sumation, U+2211 ISOamsb
|
|
295
|
+
{"sup", 8835}, // superset of, U+2283 ISOtech
|
|
296
|
+
{"sup1", 185}, // superscript one = superscript digit one
|
|
297
|
+
{"sup2", 178}, // superscript two = superscript digit two
|
|
298
|
+
{"sup3", 179}, // superscript three = superscript digit three
|
|
299
|
+
{"supe", 8839}, // superset of or equal to
|
|
300
|
+
{"szlig", 223}, // latin small letter sharp s = ess-zed
|
|
301
|
+
{"tau", 964}, // greek small letter tau, U+03C4 ISOgrk3
|
|
302
|
+
{"there4", 8756}, // therefore, U+2234 ISOtech
|
|
303
|
+
{"theta", 952}, // greek small letter theta
|
|
304
|
+
{"thetasym", 977}, // greek small letter theta symbol
|
|
305
|
+
{"thinsp", 8201}, // thin space, U+2009 ISOpub
|
|
306
|
+
{"thorn", 254}, // latin small letter thorn
|
|
307
|
+
{"tilde", 732}, // - small tilde, U+02DC ISOdia
|
|
308
|
+
{"times", 215}, // multiplication sign, U+00D7 ISOnum
|
|
309
|
+
{"trade", 8482}, // trade mark sign, U+2122 ISOnum
|
|
310
|
+
{"uArr", 8657}, // upwards double arrow, U+21D1 ISOamsa
|
|
311
|
+
{"uacute", 250}, // latin small letter u with acute
|
|
312
|
+
{"uarr", 8593}, // upwards arrow, U+2191 ISOnum-->
|
|
313
|
+
{"ucirc", 251}, // latin small letter u with circumflex
|
|
314
|
+
{"ugrave", 249}, // latin small letter u with grave
|
|
315
|
+
{"uml", 168}, // diaeresis = spacing diaeresis
|
|
316
|
+
{"upsih", 978}, // greek upsilon with hook symbol
|
|
317
|
+
{"upsilon", 965}, // greek small letter upsilon
|
|
318
|
+
{"uuml", 252}, // latin small letter u with diaeresis
|
|
319
|
+
{"weierp", 8472}, // script capital P = power set
|
|
320
|
+
{"xi", 958}, // greek small letter xi, U+03BE ISOgrk3
|
|
321
|
+
{"yacute", 253}, // latin small letter y with acute
|
|
322
|
+
{"yen", 165}, // yen sign = yuan sign, U+00A5 ISOnum
|
|
323
|
+
{"yuml", 255}, // latin small letter y with diaeresis
|
|
324
|
+
{"zeta", 950}, // greek small letter zeta, U+03B6 ISOgrk3
|
|
325
|
+
{"zwj", 8205}, // zero width joiner, U+200D NEW RFC 2070
|
|
326
|
+
{"zwnj", 8204}, // zero width non-joiner
|
|
327
|
+
{NULL, 0},
|
|
329
328
|
};
|
|
330
329
|
|
|
331
|
-
static uint64_t
|
|
332
|
-
|
|
333
|
-
uint64_t h = 0;
|
|
330
|
+
static uint64_t calc_hash(const char *key) {
|
|
331
|
+
uint64_t h = 0;
|
|
334
332
|
|
|
335
333
|
if (NULL != key) {
|
|
336
|
-
|
|
334
|
+
const uint8_t *k = (const uint8_t *)key;
|
|
337
335
|
|
|
338
|
-
|
|
339
|
-
|
|
340
|
-
|
|
341
|
-
|
|
336
|
+
for (; 0 != *k; k++) {
|
|
337
|
+
// narrow to most used range of 0x4D (77) in size
|
|
338
|
+
h = 77 * h + ((*k | 0x20) - 0x2D);
|
|
339
|
+
}
|
|
342
340
|
}
|
|
343
341
|
return h;
|
|
344
342
|
}
|
|
345
343
|
|
|
346
|
-
static Slot*
|
|
347
|
-
get_bucketp(uint64_t h) {
|
|
344
|
+
static Slot *get_bucketp(uint64_t h) {
|
|
348
345
|
return entity_cache.buckets + (BUCKET_MASK & (h ^ (h << 5) ^ (h >> 7)));
|
|
349
346
|
}
|
|
350
347
|
|
|
351
|
-
static void
|
|
352
|
-
|
|
353
|
-
|
|
354
|
-
Slot *bucket = get_bucketp(h);
|
|
348
|
+
static void cache_set(Slot s) {
|
|
349
|
+
int64_t h = calc_hash(s->key);
|
|
350
|
+
Slot *bucket = get_bucketp(h);
|
|
355
351
|
|
|
356
352
|
s->hash = h;
|
|
357
353
|
s->next = *bucket;
|
|
358
354
|
*bucket = s;
|
|
359
355
|
}
|
|
360
356
|
|
|
361
|
-
static Slot
|
|
362
|
-
|
|
363
|
-
|
|
364
|
-
Slot
|
|
365
|
-
Slot s;
|
|
357
|
+
static Slot cache_get(const char *key) {
|
|
358
|
+
int64_t h = calc_hash(key);
|
|
359
|
+
Slot *bucket = get_bucketp(h);
|
|
360
|
+
Slot s;
|
|
366
361
|
|
|
367
362
|
for (s = *bucket; NULL != s; s = s->next) {
|
|
368
|
-
|
|
369
|
-
|
|
370
|
-
|
|
363
|
+
if (h == (int64_t)s->hash && 0 == strcasecmp(s->key, key)) {
|
|
364
|
+
return s;
|
|
365
|
+
}
|
|
371
366
|
}
|
|
372
367
|
return NULL;
|
|
373
368
|
}
|
|
374
369
|
|
|
375
|
-
static void
|
|
376
|
-
|
|
377
|
-
Slot e = entities;
|
|
370
|
+
static void cache_init() {
|
|
371
|
+
Slot e = entities;
|
|
378
372
|
|
|
379
373
|
memset(&entity_cache, 0, sizeof(struct _cache));
|
|
380
374
|
for (; NULL != e->key; e++) {
|
|
381
|
-
|
|
375
|
+
cache_set(e);
|
|
382
376
|
}
|
|
383
377
|
inited = true;
|
|
384
378
|
}
|
|
385
379
|
|
|
386
|
-
char*
|
|
387
|
-
|
|
388
|
-
Slot s = entities;
|
|
380
|
+
char *ox_entity_lookup(char *text, const char *key) {
|
|
381
|
+
Slot s = entities;
|
|
389
382
|
|
|
390
383
|
if (!inited) {
|
|
391
|
-
|
|
384
|
+
cache_init();
|
|
392
385
|
}
|
|
393
386
|
if (NULL == (s = cache_get(key))) {
|
|
394
|
-
|
|
387
|
+
return NULL;
|
|
395
388
|
}
|
|
396
389
|
return ox_ucs_to_utf8_chars(text, s->code);
|
|
397
390
|
}
|