[279] | 1 | /**********************************************************************
|
---|
| 2 | regenc.c - Onigmo (Oniguruma-mod) (regular expression library)
|
---|
| 3 | **********************************************************************/
|
---|
| 4 | /*-
|
---|
| 5 | * Copyright (c) 2002-2007 K.Kosako <sndgk393 AT ybb DOT ne DOT jp>
|
---|
| 6 | * Copyright (c) 2011 K.Takata <kentkt AT csc DOT jp>
|
---|
| 7 | * All rights reserved.
|
---|
| 8 | *
|
---|
| 9 | * Redistribution and use in source and binary forms, with or without
|
---|
| 10 | * modification, are permitted provided that the following conditions
|
---|
| 11 | * are met:
|
---|
| 12 | * 1. Redistributions of source code must retain the above copyright
|
---|
| 13 | * notice, this list of conditions and the following disclaimer.
|
---|
| 14 | * 2. Redistributions in binary form must reproduce the above copyright
|
---|
| 15 | * notice, this list of conditions and the following disclaimer in the
|
---|
| 16 | * documentation and/or other materials provided with the distribution.
|
---|
| 17 | *
|
---|
| 18 | * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
|
---|
| 19 | * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
|
---|
| 20 | * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
|
---|
| 21 | * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
|
---|
| 22 | * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
|
---|
| 23 | * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
|
---|
| 24 | * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
|
---|
| 25 | * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
|
---|
| 26 | * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
|
---|
| 27 | * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
|
---|
| 28 | * SUCH DAMAGE.
|
---|
| 29 | */
|
---|
| 30 |
|
---|
| 31 | #include "regint.h"
|
---|
| 32 |
|
---|
| 33 | OnigEncoding OnigEncDefaultCharEncoding = ONIG_ENCODING_INIT_DEFAULT;
|
---|
| 34 |
|
---|
| 35 | extern int
|
---|
| 36 | onigenc_init(void)
|
---|
| 37 | {
|
---|
| 38 | return 0;
|
---|
| 39 | }
|
---|
| 40 |
|
---|
| 41 | extern OnigEncoding
|
---|
| 42 | onigenc_get_default_encoding(void)
|
---|
| 43 | {
|
---|
| 44 | return OnigEncDefaultCharEncoding;
|
---|
| 45 | }
|
---|
| 46 |
|
---|
| 47 | extern int
|
---|
| 48 | onigenc_set_default_encoding(OnigEncoding enc)
|
---|
| 49 | {
|
---|
| 50 | OnigEncDefaultCharEncoding = enc;
|
---|
| 51 | return 0;
|
---|
| 52 | }
|
---|
| 53 |
|
---|
| 54 | extern UChar*
|
---|
| 55 | onigenc_get_right_adjust_char_head(OnigEncoding enc, const UChar* start, const UChar* s)
|
---|
| 56 | {
|
---|
| 57 | UChar* p = ONIGENC_LEFT_ADJUST_CHAR_HEAD(enc, start, s);
|
---|
| 58 | if (p < s) {
|
---|
| 59 | p += enclen(enc, p);
|
---|
| 60 | }
|
---|
| 61 | return p;
|
---|
| 62 | }
|
---|
| 63 |
|
---|
| 64 | extern UChar*
|
---|
| 65 | onigenc_get_right_adjust_char_head_with_prev(OnigEncoding enc,
|
---|
| 66 | const UChar* start, const UChar* s, const UChar** prev)
|
---|
| 67 | {
|
---|
| 68 | UChar* p = ONIGENC_LEFT_ADJUST_CHAR_HEAD(enc, start, s);
|
---|
| 69 |
|
---|
| 70 | if (p < s) {
|
---|
| 71 | if (prev) *prev = (const UChar* )p;
|
---|
| 72 | p += enclen(enc, p);
|
---|
| 73 | }
|
---|
| 74 | else {
|
---|
| 75 | if (prev) *prev = (const UChar* )NULL; /* Sorry */
|
---|
| 76 | }
|
---|
| 77 | return p;
|
---|
| 78 | }
|
---|
| 79 |
|
---|
| 80 | extern UChar*
|
---|
| 81 | onigenc_get_prev_char_head(OnigEncoding enc, const UChar* start, const UChar* s)
|
---|
| 82 | {
|
---|
| 83 | if (s <= start)
|
---|
| 84 | return (UChar* )NULL;
|
---|
| 85 |
|
---|
| 86 | return ONIGENC_LEFT_ADJUST_CHAR_HEAD(enc, start, s - 1);
|
---|
| 87 | }
|
---|
| 88 |
|
---|
| 89 | extern UChar*
|
---|
| 90 | onigenc_step_back(OnigEncoding enc, const UChar* start, const UChar* s, int n)
|
---|
| 91 | {
|
---|
| 92 | while (ONIG_IS_NOT_NULL(s) && n-- > 0) {
|
---|
| 93 | if (s <= start)
|
---|
| 94 | return (UChar* )NULL;
|
---|
| 95 |
|
---|
| 96 | s = ONIGENC_LEFT_ADJUST_CHAR_HEAD(enc, start, s - 1);
|
---|
| 97 | }
|
---|
| 98 | return (UChar* )s;
|
---|
| 99 | }
|
---|
| 100 |
|
---|
| 101 | extern UChar*
|
---|
| 102 | onigenc_step(OnigEncoding enc, const UChar* p, const UChar* end, int n)
|
---|
| 103 | {
|
---|
| 104 | UChar* q = (UChar* )p;
|
---|
| 105 | while (n-- > 0) {
|
---|
| 106 | q += ONIGENC_MBC_ENC_LEN(enc, q);
|
---|
| 107 | }
|
---|
| 108 | return (q <= end ? q : NULL);
|
---|
| 109 | }
|
---|
| 110 |
|
---|
| 111 | extern int
|
---|
| 112 | onigenc_strlen(OnigEncoding enc, const UChar* p, const UChar* end)
|
---|
| 113 | {
|
---|
| 114 | int n = 0;
|
---|
| 115 | UChar* q = (UChar* )p;
|
---|
| 116 |
|
---|
| 117 | while (q < end) {
|
---|
| 118 | q += ONIGENC_MBC_ENC_LEN(enc, q);
|
---|
| 119 | n++;
|
---|
| 120 | }
|
---|
| 121 | return n;
|
---|
| 122 | }
|
---|
| 123 |
|
---|
| 124 | extern int
|
---|
| 125 | onigenc_strlen_null(OnigEncoding enc, const UChar* s)
|
---|
| 126 | {
|
---|
| 127 | int n = 0;
|
---|
| 128 | UChar* p = (UChar* )s;
|
---|
| 129 |
|
---|
| 130 | while (1) {
|
---|
| 131 | if (*p == '\0') {
|
---|
| 132 | UChar* q;
|
---|
| 133 | int len = ONIGENC_MBC_MINLEN(enc);
|
---|
| 134 |
|
---|
| 135 | if (len == 1) return n;
|
---|
| 136 | q = p + 1;
|
---|
| 137 | while (len > 1) {
|
---|
| 138 | if (*q != '\0') break;
|
---|
| 139 | q++;
|
---|
| 140 | len--;
|
---|
| 141 | }
|
---|
| 142 | if (len == 1) return n;
|
---|
| 143 | }
|
---|
| 144 | p += ONIGENC_MBC_ENC_LEN(enc, p);
|
---|
| 145 | n++;
|
---|
| 146 | }
|
---|
| 147 | }
|
---|
| 148 |
|
---|
| 149 | extern int
|
---|
| 150 | onigenc_str_bytelen_null(OnigEncoding enc, const UChar* s)
|
---|
| 151 | {
|
---|
| 152 | UChar* start = (UChar* )s;
|
---|
| 153 | UChar* p = (UChar* )s;
|
---|
| 154 |
|
---|
| 155 | while (1) {
|
---|
| 156 | if (*p == '\0') {
|
---|
| 157 | UChar* q;
|
---|
| 158 | int len = ONIGENC_MBC_MINLEN(enc);
|
---|
| 159 |
|
---|
| 160 | if (len == 1) return (int )(p - start);
|
---|
| 161 | q = p + 1;
|
---|
| 162 | while (len > 1) {
|
---|
| 163 | if (*q != '\0') break;
|
---|
| 164 | q++;
|
---|
| 165 | len--;
|
---|
| 166 | }
|
---|
| 167 | if (len == 1) return (int )(p - start);
|
---|
| 168 | }
|
---|
| 169 | p += ONIGENC_MBC_ENC_LEN(enc, p);
|
---|
| 170 | }
|
---|
| 171 | }
|
---|
| 172 |
|
---|
| 173 | const UChar OnigEncAsciiToLowerCaseTable[] = {
|
---|
| 174 | '\000', '\001', '\002', '\003', '\004', '\005', '\006', '\007',
|
---|
| 175 | '\010', '\011', '\012', '\013', '\014', '\015', '\016', '\017',
|
---|
| 176 | '\020', '\021', '\022', '\023', '\024', '\025', '\026', '\027',
|
---|
| 177 | '\030', '\031', '\032', '\033', '\034', '\035', '\036', '\037',
|
---|
| 178 | '\040', '\041', '\042', '\043', '\044', '\045', '\046', '\047',
|
---|
| 179 | '\050', '\051', '\052', '\053', '\054', '\055', '\056', '\057',
|
---|
| 180 | '\060', '\061', '\062', '\063', '\064', '\065', '\066', '\067',
|
---|
| 181 | '\070', '\071', '\072', '\073', '\074', '\075', '\076', '\077',
|
---|
| 182 | '\100', '\141', '\142', '\143', '\144', '\145', '\146', '\147',
|
---|
| 183 | '\150', '\151', '\152', '\153', '\154', '\155', '\156', '\157',
|
---|
| 184 | '\160', '\161', '\162', '\163', '\164', '\165', '\166', '\167',
|
---|
| 185 | '\170', '\171', '\172', '\133', '\134', '\135', '\136', '\137',
|
---|
| 186 | '\140', '\141', '\142', '\143', '\144', '\145', '\146', '\147',
|
---|
| 187 | '\150', '\151', '\152', '\153', '\154', '\155', '\156', '\157',
|
---|
| 188 | '\160', '\161', '\162', '\163', '\164', '\165', '\166', '\167',
|
---|
| 189 | '\170', '\171', '\172', '\173', '\174', '\175', '\176', '\177',
|
---|
| 190 | '\200', '\201', '\202', '\203', '\204', '\205', '\206', '\207',
|
---|
| 191 | '\210', '\211', '\212', '\213', '\214', '\215', '\216', '\217',
|
---|
| 192 | '\220', '\221', '\222', '\223', '\224', '\225', '\226', '\227',
|
---|
| 193 | '\230', '\231', '\232', '\233', '\234', '\235', '\236', '\237',
|
---|
| 194 | '\240', '\241', '\242', '\243', '\244', '\245', '\246', '\247',
|
---|
| 195 | '\250', '\251', '\252', '\253', '\254', '\255', '\256', '\257',
|
---|
| 196 | '\260', '\261', '\262', '\263', '\264', '\265', '\266', '\267',
|
---|
| 197 | '\270', '\271', '\272', '\273', '\274', '\275', '\276', '\277',
|
---|
| 198 | '\300', '\301', '\302', '\303', '\304', '\305', '\306', '\307',
|
---|
| 199 | '\310', '\311', '\312', '\313', '\314', '\315', '\316', '\317',
|
---|
| 200 | '\320', '\321', '\322', '\323', '\324', '\325', '\326', '\327',
|
---|
| 201 | '\330', '\331', '\332', '\333', '\334', '\335', '\336', '\337',
|
---|
| 202 | '\340', '\341', '\342', '\343', '\344', '\345', '\346', '\347',
|
---|
| 203 | '\350', '\351', '\352', '\353', '\354', '\355', '\356', '\357',
|
---|
| 204 | '\360', '\361', '\362', '\363', '\364', '\365', '\366', '\367',
|
---|
| 205 | '\370', '\371', '\372', '\373', '\374', '\375', '\376', '\377',
|
---|
| 206 | };
|
---|
| 207 |
|
---|
| 208 | #ifdef USE_UPPER_CASE_TABLE
|
---|
| 209 | const UChar OnigEncAsciiToUpperCaseTable[256] = {
|
---|
| 210 | '\000', '\001', '\002', '\003', '\004', '\005', '\006', '\007',
|
---|
| 211 | '\010', '\011', '\012', '\013', '\014', '\015', '\016', '\017',
|
---|
| 212 | '\020', '\021', '\022', '\023', '\024', '\025', '\026', '\027',
|
---|
| 213 | '\030', '\031', '\032', '\033', '\034', '\035', '\036', '\037',
|
---|
| 214 | '\040', '\041', '\042', '\043', '\044', '\045', '\046', '\047',
|
---|
| 215 | '\050', '\051', '\052', '\053', '\054', '\055', '\056', '\057',
|
---|
| 216 | '\060', '\061', '\062', '\063', '\064', '\065', '\066', '\067',
|
---|
| 217 | '\070', '\071', '\072', '\073', '\074', '\075', '\076', '\077',
|
---|
| 218 | '\100', '\101', '\102', '\103', '\104', '\105', '\106', '\107',
|
---|
| 219 | '\110', '\111', '\112', '\113', '\114', '\115', '\116', '\117',
|
---|
| 220 | '\120', '\121', '\122', '\123', '\124', '\125', '\126', '\127',
|
---|
| 221 | '\130', '\131', '\132', '\133', '\134', '\135', '\136', '\137',
|
---|
| 222 | '\140', '\101', '\102', '\103', '\104', '\105', '\106', '\107',
|
---|
| 223 | '\110', '\111', '\112', '\113', '\114', '\115', '\116', '\117',
|
---|
| 224 | '\120', '\121', '\122', '\123', '\124', '\125', '\126', '\127',
|
---|
| 225 | '\130', '\131', '\132', '\173', '\174', '\175', '\176', '\177',
|
---|
| 226 | '\200', '\201', '\202', '\203', '\204', '\205', '\206', '\207',
|
---|
| 227 | '\210', '\211', '\212', '\213', '\214', '\215', '\216', '\217',
|
---|
| 228 | '\220', '\221', '\222', '\223', '\224', '\225', '\226', '\227',
|
---|
| 229 | '\230', '\231', '\232', '\233', '\234', '\235', '\236', '\237',
|
---|
| 230 | '\240', '\241', '\242', '\243', '\244', '\245', '\246', '\247',
|
---|
| 231 | '\250', '\251', '\252', '\253', '\254', '\255', '\256', '\257',
|
---|
| 232 | '\260', '\261', '\262', '\263', '\264', '\265', '\266', '\267',
|
---|
| 233 | '\270', '\271', '\272', '\273', '\274', '\275', '\276', '\277',
|
---|
| 234 | '\300', '\301', '\302', '\303', '\304', '\305', '\306', '\307',
|
---|
| 235 | '\310', '\311', '\312', '\313', '\314', '\315', '\316', '\317',
|
---|
| 236 | '\320', '\321', '\322', '\323', '\324', '\325', '\326', '\327',
|
---|
| 237 | '\330', '\331', '\332', '\333', '\334', '\335', '\336', '\337',
|
---|
| 238 | '\340', '\341', '\342', '\343', '\344', '\345', '\346', '\347',
|
---|
| 239 | '\350', '\351', '\352', '\353', '\354', '\355', '\356', '\357',
|
---|
| 240 | '\360', '\361', '\362', '\363', '\364', '\365', '\366', '\367',
|
---|
| 241 | '\370', '\371', '\372', '\373', '\374', '\375', '\376', '\377',
|
---|
| 242 | };
|
---|
| 243 | #endif
|
---|
| 244 |
|
---|
| 245 | const unsigned short OnigEncAsciiCtypeTable[256] = {
|
---|
| 246 | 0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008,
|
---|
| 247 | 0x4008, 0x420c, 0x4209, 0x4208, 0x4208, 0x4208, 0x4008, 0x4008,
|
---|
| 248 | 0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008,
|
---|
| 249 | 0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008,
|
---|
| 250 | 0x4284, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0,
|
---|
| 251 | 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0,
|
---|
| 252 | 0x78b0, 0x78b0, 0x78b0, 0x78b0, 0x78b0, 0x78b0, 0x78b0, 0x78b0,
|
---|
| 253 | 0x78b0, 0x78b0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0,
|
---|
| 254 | 0x41a0, 0x7ca2, 0x7ca2, 0x7ca2, 0x7ca2, 0x7ca2, 0x7ca2, 0x74a2,
|
---|
| 255 | 0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2,
|
---|
| 256 | 0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2,
|
---|
| 257 | 0x74a2, 0x74a2, 0x74a2, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x51a0,
|
---|
| 258 | 0x41a0, 0x78e2, 0x78e2, 0x78e2, 0x78e2, 0x78e2, 0x78e2, 0x70e2,
|
---|
| 259 | 0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2,
|
---|
| 260 | 0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2,
|
---|
| 261 | 0x70e2, 0x70e2, 0x70e2, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x4008,
|
---|
| 262 | 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000,
|
---|
| 263 | 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000,
|
---|
| 264 | 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000,
|
---|
| 265 | 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000,
|
---|
| 266 | 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000,
|
---|
| 267 | 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000,
|
---|
| 268 | 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000,
|
---|
| 269 | 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000,
|
---|
| 270 | 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000,
|
---|
| 271 | 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000,
|
---|
| 272 | 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000,
|
---|
| 273 | 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000,
|
---|
| 274 | 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000,
|
---|
| 275 | 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000,
|
---|
| 276 | 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000,
|
---|
| 277 | 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000
|
---|
| 278 | };
|
---|
| 279 |
|
---|
| 280 | const UChar OnigEncISO_8859_1_ToLowerCaseTable[256] = {
|
---|
| 281 | '\000', '\001', '\002', '\003', '\004', '\005', '\006', '\007',
|
---|
| 282 | '\010', '\011', '\012', '\013', '\014', '\015', '\016', '\017',
|
---|
| 283 | '\020', '\021', '\022', '\023', '\024', '\025', '\026', '\027',
|
---|
| 284 | '\030', '\031', '\032', '\033', '\034', '\035', '\036', '\037',
|
---|
| 285 | '\040', '\041', '\042', '\043', '\044', '\045', '\046', '\047',
|
---|
| 286 | '\050', '\051', '\052', '\053', '\054', '\055', '\056', '\057',
|
---|
| 287 | '\060', '\061', '\062', '\063', '\064', '\065', '\066', '\067',
|
---|
| 288 | '\070', '\071', '\072', '\073', '\074', '\075', '\076', '\077',
|
---|
| 289 | '\100', '\141', '\142', '\143', '\144', '\145', '\146', '\147',
|
---|
| 290 | '\150', '\151', '\152', '\153', '\154', '\155', '\156', '\157',
|
---|
| 291 | '\160', '\161', '\162', '\163', '\164', '\165', '\166', '\167',
|
---|
| 292 | '\170', '\171', '\172', '\133', '\134', '\135', '\136', '\137',
|
---|
| 293 | '\140', '\141', '\142', '\143', '\144', '\145', '\146', '\147',
|
---|
| 294 | '\150', '\151', '\152', '\153', '\154', '\155', '\156', '\157',
|
---|
| 295 | '\160', '\161', '\162', '\163', '\164', '\165', '\166', '\167',
|
---|
| 296 | '\170', '\171', '\172', '\173', '\174', '\175', '\176', '\177',
|
---|
| 297 | '\200', '\201', '\202', '\203', '\204', '\205', '\206', '\207',
|
---|
| 298 | '\210', '\211', '\212', '\213', '\214', '\215', '\216', '\217',
|
---|
| 299 | '\220', '\221', '\222', '\223', '\224', '\225', '\226', '\227',
|
---|
| 300 | '\230', '\231', '\232', '\233', '\234', '\235', '\236', '\237',
|
---|
| 301 | '\240', '\241', '\242', '\243', '\244', '\245', '\246', '\247',
|
---|
| 302 | '\250', '\251', '\252', '\253', '\254', '\255', '\256', '\257',
|
---|
| 303 | '\260', '\261', '\262', '\263', '\264', '\265', '\266', '\267',
|
---|
| 304 | '\270', '\271', '\272', '\273', '\274', '\275', '\276', '\277',
|
---|
| 305 | '\340', '\341', '\342', '\343', '\344', '\345', '\346', '\347',
|
---|
| 306 | '\350', '\351', '\352', '\353', '\354', '\355', '\356', '\357',
|
---|
| 307 | '\360', '\361', '\362', '\363', '\364', '\365', '\366', '\327',
|
---|
| 308 | '\370', '\371', '\372', '\373', '\374', '\375', '\376', '\337',
|
---|
| 309 | '\340', '\341', '\342', '\343', '\344', '\345', '\346', '\347',
|
---|
| 310 | '\350', '\351', '\352', '\353', '\354', '\355', '\356', '\357',
|
---|
| 311 | '\360', '\361', '\362', '\363', '\364', '\365', '\366', '\367',
|
---|
| 312 | '\370', '\371', '\372', '\373', '\374', '\375', '\376', '\377'
|
---|
| 313 | };
|
---|
| 314 |
|
---|
| 315 | #ifdef USE_UPPER_CASE_TABLE
|
---|
| 316 | const UChar OnigEncISO_8859_1_ToUpperCaseTable[256] = {
|
---|
| 317 | '\000', '\001', '\002', '\003', '\004', '\005', '\006', '\007',
|
---|
| 318 | '\010', '\011', '\012', '\013', '\014', '\015', '\016', '\017',
|
---|
| 319 | '\020', '\021', '\022', '\023', '\024', '\025', '\026', '\027',
|
---|
| 320 | '\030', '\031', '\032', '\033', '\034', '\035', '\036', '\037',
|
---|
| 321 | '\040', '\041', '\042', '\043', '\044', '\045', '\046', '\047',
|
---|
| 322 | '\050', '\051', '\052', '\053', '\054', '\055', '\056', '\057',
|
---|
| 323 | '\060', '\061', '\062', '\063', '\064', '\065', '\066', '\067',
|
---|
| 324 | '\070', '\071', '\072', '\073', '\074', '\075', '\076', '\077',
|
---|
| 325 | '\100', '\101', '\102', '\103', '\104', '\105', '\106', '\107',
|
---|
| 326 | '\110', '\111', '\112', '\113', '\114', '\115', '\116', '\117',
|
---|
| 327 | '\120', '\121', '\122', '\123', '\124', '\125', '\126', '\127',
|
---|
| 328 | '\130', '\131', '\132', '\133', '\134', '\135', '\136', '\137',
|
---|
| 329 | '\140', '\101', '\102', '\103', '\104', '\105', '\106', '\107',
|
---|
| 330 | '\110', '\111', '\112', '\113', '\114', '\115', '\116', '\117',
|
---|
| 331 | '\120', '\121', '\122', '\123', '\124', '\125', '\126', '\127',
|
---|
| 332 | '\130', '\131', '\132', '\173', '\174', '\175', '\176', '\177',
|
---|
| 333 | '\200', '\201', '\202', '\203', '\204', '\205', '\206', '\207',
|
---|
| 334 | '\210', '\211', '\212', '\213', '\214', '\215', '\216', '\217',
|
---|
| 335 | '\220', '\221', '\222', '\223', '\224', '\225', '\226', '\227',
|
---|
| 336 | '\230', '\231', '\232', '\233', '\234', '\235', '\236', '\237',
|
---|
| 337 | '\240', '\241', '\242', '\243', '\244', '\245', '\246', '\247',
|
---|
| 338 | '\250', '\251', '\252', '\253', '\254', '\255', '\256', '\257',
|
---|
| 339 | '\260', '\261', '\262', '\263', '\264', '\265', '\266', '\267',
|
---|
| 340 | '\270', '\271', '\272', '\273', '\274', '\275', '\276', '\277',
|
---|
| 341 | '\300', '\301', '\302', '\303', '\304', '\305', '\306', '\307',
|
---|
| 342 | '\310', '\311', '\312', '\313', '\314', '\315', '\316', '\317',
|
---|
| 343 | '\320', '\321', '\322', '\323', '\324', '\325', '\326', '\327',
|
---|
| 344 | '\330', '\331', '\332', '\333', '\334', '\335', '\336', '\337',
|
---|
| 345 | '\300', '\301', '\302', '\303', '\304', '\305', '\306', '\307',
|
---|
| 346 | '\310', '\311', '\312', '\313', '\314', '\315', '\316', '\317',
|
---|
| 347 | '\320', '\321', '\322', '\323', '\324', '\325', '\326', '\367',
|
---|
| 348 | '\330', '\331', '\332', '\333', '\334', '\335', '\336', '\377',
|
---|
| 349 | };
|
---|
| 350 | #endif
|
---|
| 351 |
|
---|
| 352 | extern void
|
---|
| 353 | onigenc_set_default_caseconv_table(const UChar* table ARG_UNUSED)
|
---|
| 354 | {
|
---|
| 355 | /* nothing */
|
---|
| 356 | /* obsoleted. */
|
---|
| 357 | }
|
---|
| 358 |
|
---|
| 359 | extern UChar*
|
---|
| 360 | onigenc_get_left_adjust_char_head(OnigEncoding enc, const UChar* start, const UChar* s)
|
---|
| 361 | {
|
---|
| 362 | return ONIGENC_LEFT_ADJUST_CHAR_HEAD(enc, start, s);
|
---|
| 363 | }
|
---|
| 364 |
|
---|
| 365 | const OnigPairCaseFoldCodes OnigAsciiLowerMap[] = {
|
---|
| 366 | { 0x41, 0x61 },
|
---|
| 367 | { 0x42, 0x62 },
|
---|
| 368 | { 0x43, 0x63 },
|
---|
| 369 | { 0x44, 0x64 },
|
---|
| 370 | { 0x45, 0x65 },
|
---|
| 371 | { 0x46, 0x66 },
|
---|
| 372 | { 0x47, 0x67 },
|
---|
| 373 | { 0x48, 0x68 },
|
---|
| 374 | { 0x49, 0x69 },
|
---|
| 375 | { 0x4a, 0x6a },
|
---|
| 376 | { 0x4b, 0x6b },
|
---|
| 377 | { 0x4c, 0x6c },
|
---|
| 378 | { 0x4d, 0x6d },
|
---|
| 379 | { 0x4e, 0x6e },
|
---|
| 380 | { 0x4f, 0x6f },
|
---|
| 381 | { 0x50, 0x70 },
|
---|
| 382 | { 0x51, 0x71 },
|
---|
| 383 | { 0x52, 0x72 },
|
---|
| 384 | { 0x53, 0x73 },
|
---|
| 385 | { 0x54, 0x74 },
|
---|
| 386 | { 0x55, 0x75 },
|
---|
| 387 | { 0x56, 0x76 },
|
---|
| 388 | { 0x57, 0x77 },
|
---|
| 389 | { 0x58, 0x78 },
|
---|
| 390 | { 0x59, 0x79 },
|
---|
| 391 | { 0x5a, 0x7a }
|
---|
| 392 | };
|
---|
| 393 |
|
---|
| 394 | extern int
|
---|
| 395 | onigenc_ascii_apply_all_case_fold(OnigCaseFoldType flag ARG_UNUSED,
|
---|
| 396 | OnigApplyAllCaseFoldFunc f, void* arg)
|
---|
| 397 | {
|
---|
| 398 | OnigCodePoint code;
|
---|
| 399 | int i, r;
|
---|
| 400 |
|
---|
| 401 | for (i = 0; i < numberof(OnigAsciiLowerMap); i++) {
|
---|
| 402 | code = OnigAsciiLowerMap[i].to;
|
---|
| 403 | r = (*f)(OnigAsciiLowerMap[i].from, &code, 1, arg);
|
---|
| 404 | if (r != 0) return r;
|
---|
| 405 |
|
---|
| 406 | code = OnigAsciiLowerMap[i].from;
|
---|
| 407 | r = (*f)(OnigAsciiLowerMap[i].to, &code, 1, arg);
|
---|
| 408 | if (r != 0) return r;
|
---|
| 409 | }
|
---|
| 410 |
|
---|
| 411 | return 0;
|
---|
| 412 | }
|
---|
| 413 |
|
---|
| 414 | extern int
|
---|
| 415 | onigenc_ascii_get_case_fold_codes_by_str(OnigCaseFoldType flag ARG_UNUSED,
|
---|
| 416 | const OnigUChar* p, const OnigUChar* end ARG_UNUSED,
|
---|
| 417 | OnigCaseFoldCodeItem items[])
|
---|
| 418 | {
|
---|
| 419 | if (0x41 <= *p && *p <= 0x5a) {
|
---|
| 420 | items[0].byte_len = 1;
|
---|
| 421 | items[0].code_len = 1;
|
---|
| 422 | items[0].code[0] = (OnigCodePoint )(*p + 0x20);
|
---|
| 423 | return 1;
|
---|
| 424 | }
|
---|
| 425 | else if (0x61 <= *p && *p <= 0x7a) {
|
---|
| 426 | items[0].byte_len = 1;
|
---|
| 427 | items[0].code_len = 1;
|
---|
| 428 | items[0].code[0] = (OnigCodePoint )(*p - 0x20);
|
---|
| 429 | return 1;
|
---|
| 430 | }
|
---|
| 431 | else
|
---|
| 432 | return 0;
|
---|
| 433 | }
|
---|
| 434 |
|
---|
| 435 | static int
|
---|
| 436 | ss_apply_all_case_fold(OnigCaseFoldType flag ARG_UNUSED,
|
---|
| 437 | OnigApplyAllCaseFoldFunc f, void* arg)
|
---|
| 438 | {
|
---|
| 439 | OnigCodePoint ss[] = { 0x73, 0x73 };
|
---|
| 440 |
|
---|
| 441 | return (*f)((OnigCodePoint )0xdf, ss, 2, arg);
|
---|
| 442 | }
|
---|
| 443 |
|
---|
| 444 | extern int
|
---|
| 445 | onigenc_apply_all_case_fold_with_map(int map_size,
|
---|
| 446 | const OnigPairCaseFoldCodes map[],
|
---|
| 447 | int ess_tsett_flag, OnigCaseFoldType flag,
|
---|
| 448 | OnigApplyAllCaseFoldFunc f, void* arg)
|
---|
| 449 | {
|
---|
| 450 | OnigCodePoint code;
|
---|
| 451 | int i, r;
|
---|
| 452 |
|
---|
| 453 | r = onigenc_ascii_apply_all_case_fold(flag, f, arg);
|
---|
| 454 | if (r != 0) return r;
|
---|
| 455 |
|
---|
| 456 | for (i = 0; i < map_size; i++) {
|
---|
| 457 | code = map[i].to;
|
---|
| 458 | r = (*f)(map[i].from, &code, 1, arg);
|
---|
| 459 | if (r != 0) return r;
|
---|
| 460 |
|
---|
| 461 | code = map[i].from;
|
---|
| 462 | r = (*f)(map[i].to, &code, 1, arg);
|
---|
| 463 | if (r != 0) return r;
|
---|
| 464 | }
|
---|
| 465 |
|
---|
| 466 | if (ess_tsett_flag != 0)
|
---|
| 467 | return ss_apply_all_case_fold(flag, f, arg);
|
---|
| 468 |
|
---|
| 469 | return 0;
|
---|
| 470 | }
|
---|
| 471 |
|
---|
| 472 | extern int
|
---|
| 473 | onigenc_get_case_fold_codes_by_str_with_map(int map_size,
|
---|
| 474 | const OnigPairCaseFoldCodes map[],
|
---|
| 475 | int ess_tsett_flag, OnigCaseFoldType flag ARG_UNUSED,
|
---|
| 476 | const OnigUChar* p, const OnigUChar* end, OnigCaseFoldCodeItem items[])
|
---|
| 477 | {
|
---|
| 478 | if (0x41 <= *p && *p <= 0x5a) {
|
---|
| 479 | items[0].byte_len = 1;
|
---|
| 480 | items[0].code_len = 1;
|
---|
| 481 | items[0].code[0] = (OnigCodePoint )(*p + 0x20);
|
---|
| 482 | if (*p == 0x53 && ess_tsett_flag != 0 && end > p + 1
|
---|
| 483 | && (*(p+1) == 0x53 || *(p+1) == 0x73)) {
|
---|
| 484 | /* SS */
|
---|
| 485 | items[1].byte_len = 2;
|
---|
| 486 | items[1].code_len = 1;
|
---|
| 487 | items[1].code[0] = (OnigCodePoint )0xdf;
|
---|
| 488 | return 2;
|
---|
| 489 | }
|
---|
| 490 | else
|
---|
| 491 | return 1;
|
---|
| 492 | }
|
---|
| 493 | else if (0x61 <= *p && *p <= 0x7a) {
|
---|
| 494 | items[0].byte_len = 1;
|
---|
| 495 | items[0].code_len = 1;
|
---|
| 496 | items[0].code[0] = (OnigCodePoint )(*p - 0x20);
|
---|
| 497 | if (*p == 0x73 && ess_tsett_flag != 0 && end > p + 1
|
---|
| 498 | && (*(p+1) == 0x73 || *(p+1) == 0x53)) {
|
---|
| 499 | /* ss */
|
---|
| 500 | items[1].byte_len = 2;
|
---|
| 501 | items[1].code_len = 1;
|
---|
| 502 | items[1].code[0] = (OnigCodePoint )0xdf;
|
---|
| 503 | return 2;
|
---|
| 504 | }
|
---|
| 505 | else
|
---|
| 506 | return 1;
|
---|
| 507 | }
|
---|
| 508 | else if (*p == 0xdf && ess_tsett_flag != 0) {
|
---|
| 509 | items[0].byte_len = 1;
|
---|
| 510 | items[0].code_len = 2;
|
---|
| 511 | items[0].code[0] = (OnigCodePoint )'s';
|
---|
| 512 | items[0].code[1] = (OnigCodePoint )'s';
|
---|
| 513 |
|
---|
| 514 | items[1].byte_len = 1;
|
---|
| 515 | items[1].code_len = 2;
|
---|
| 516 | items[1].code[0] = (OnigCodePoint )'S';
|
---|
| 517 | items[1].code[1] = (OnigCodePoint )'S';
|
---|
| 518 |
|
---|
| 519 | items[2].byte_len = 1;
|
---|
| 520 | items[2].code_len = 2;
|
---|
| 521 | items[2].code[0] = (OnigCodePoint )'s';
|
---|
| 522 | items[2].code[1] = (OnigCodePoint )'S';
|
---|
| 523 |
|
---|
| 524 | items[3].byte_len = 1;
|
---|
| 525 | items[3].code_len = 2;
|
---|
| 526 | items[3].code[0] = (OnigCodePoint )'S';
|
---|
| 527 | items[3].code[1] = (OnigCodePoint )'s';
|
---|
| 528 |
|
---|
| 529 | return 4;
|
---|
| 530 | }
|
---|
| 531 | else {
|
---|
| 532 | int i;
|
---|
| 533 |
|
---|
| 534 | for (i = 0; i < map_size; i++) {
|
---|
| 535 | if (*p == map[i].from) {
|
---|
| 536 | items[0].byte_len = 1;
|
---|
| 537 | items[0].code_len = 1;
|
---|
| 538 | items[0].code[0] = map[i].to;
|
---|
| 539 | return 1;
|
---|
| 540 | }
|
---|
| 541 | else if (*p == map[i].to) {
|
---|
| 542 | items[0].byte_len = 1;
|
---|
| 543 | items[0].code_len = 1;
|
---|
| 544 | items[0].code[0] = map[i].from;
|
---|
| 545 | return 1;
|
---|
| 546 | }
|
---|
| 547 | }
|
---|
| 548 | }
|
---|
| 549 |
|
---|
| 550 | return 0;
|
---|
| 551 | }
|
---|
| 552 |
|
---|
| 553 |
|
---|
| 554 | extern int
|
---|
| 555 | onigenc_not_support_get_ctype_code_range(OnigCtype ctype ARG_UNUSED,
|
---|
| 556 | OnigCodePoint* sb_out ARG_UNUSED,
|
---|
| 557 | const OnigCodePoint* ranges[] ARG_UNUSED)
|
---|
| 558 | {
|
---|
| 559 | return ONIG_NO_SUPPORT_CONFIG;
|
---|
| 560 | }
|
---|
| 561 |
|
---|
| 562 | extern int
|
---|
| 563 | onigenc_is_mbc_newline_0x0a(const UChar* p, const UChar* end)
|
---|
| 564 | {
|
---|
| 565 | if (p < end) {
|
---|
| 566 | if (*p == 0x0a) return 1;
|
---|
| 567 | }
|
---|
| 568 | return 0;
|
---|
| 569 | }
|
---|
| 570 |
|
---|
| 571 | /* for single byte encodings */
|
---|
| 572 | extern int
|
---|
| 573 | onigenc_ascii_mbc_case_fold(OnigCaseFoldType flag ARG_UNUSED, const UChar** p,
|
---|
| 574 | const UChar*end ARG_UNUSED, UChar* lower)
|
---|
| 575 | {
|
---|
| 576 | *lower = ONIGENC_ASCII_CODE_TO_LOWER_CASE(**p);
|
---|
| 577 |
|
---|
| 578 | (*p)++;
|
---|
| 579 | return 1; /* return byte length of converted char to lower */
|
---|
| 580 | }
|
---|
| 581 |
|
---|
| 582 | #if 0
|
---|
| 583 | extern int
|
---|
| 584 | onigenc_ascii_is_mbc_ambiguous(OnigCaseFoldType flag ARG_UNUSED,
|
---|
| 585 | const UChar** pp, const UChar* end ARG_UNUSED)
|
---|
| 586 | {
|
---|
| 587 | const UChar* p = *pp;
|
---|
| 588 |
|
---|
| 589 | (*pp)++;
|
---|
| 590 | return ONIGENC_IS_ASCII_CODE_CASE_AMBIG(*p);
|
---|
| 591 | }
|
---|
| 592 | #endif
|
---|
| 593 |
|
---|
| 594 | extern int
|
---|
| 595 | onigenc_single_byte_mbc_enc_len(const UChar* p ARG_UNUSED)
|
---|
| 596 | {
|
---|
| 597 | return 1;
|
---|
| 598 | }
|
---|
| 599 |
|
---|
| 600 | extern OnigCodePoint
|
---|
| 601 | onigenc_single_byte_mbc_to_code(const UChar* p, const UChar* end ARG_UNUSED)
|
---|
| 602 | {
|
---|
| 603 | return (OnigCodePoint )(*p);
|
---|
| 604 | }
|
---|
| 605 |
|
---|
| 606 | extern int
|
---|
| 607 | onigenc_single_byte_code_to_mbclen(OnigCodePoint code ARG_UNUSED)
|
---|
| 608 | {
|
---|
| 609 | return (code < 0x100 ? 1 : ONIGERR_INVALID_CODE_POINT_VALUE);
|
---|
| 610 | }
|
---|
| 611 |
|
---|
| 612 | extern int
|
---|
| 613 | onigenc_single_byte_code_to_mbc(OnigCodePoint code, UChar *buf)
|
---|
| 614 | {
|
---|
| 615 | *buf = (UChar )(code & 0xff);
|
---|
| 616 | return 1;
|
---|
| 617 | }
|
---|
| 618 |
|
---|
| 619 | extern UChar*
|
---|
| 620 | onigenc_single_byte_left_adjust_char_head(const UChar* start ARG_UNUSED,
|
---|
| 621 | const UChar* s)
|
---|
| 622 | {
|
---|
| 623 | return (UChar* )s;
|
---|
| 624 | }
|
---|
| 625 |
|
---|
| 626 | extern int
|
---|
| 627 | onigenc_always_true_is_allowed_reverse_match(const UChar* s ARG_UNUSED,
|
---|
| 628 | const UChar* end ARG_UNUSED)
|
---|
| 629 | {
|
---|
| 630 | return TRUE;
|
---|
| 631 | }
|
---|
| 632 |
|
---|
| 633 | extern int
|
---|
| 634 | onigenc_always_false_is_allowed_reverse_match(const UChar* s ARG_UNUSED,
|
---|
| 635 | const UChar* end ARG_UNUSED)
|
---|
| 636 | {
|
---|
| 637 | return FALSE;
|
---|
| 638 | }
|
---|
| 639 |
|
---|
| 640 | extern int
|
---|
| 641 | onigenc_ascii_is_code_ctype(OnigCodePoint code, unsigned int ctype)
|
---|
| 642 | {
|
---|
| 643 | if (code < 128)
|
---|
| 644 | return ONIGENC_IS_ASCII_CODE_CTYPE(code, ctype);
|
---|
| 645 | else
|
---|
| 646 | return FALSE;
|
---|
| 647 | }
|
---|
| 648 |
|
---|
| 649 | extern OnigCodePoint
|
---|
| 650 | onigenc_mbn_mbc_to_code(OnigEncoding enc, const UChar* p, const UChar* end)
|
---|
| 651 | {
|
---|
| 652 | int c, i, len;
|
---|
| 653 | OnigCodePoint n;
|
---|
| 654 |
|
---|
| 655 | len = enclen(enc, p);
|
---|
| 656 | n = (OnigCodePoint )(*p++);
|
---|
| 657 | if (len == 1) return n;
|
---|
| 658 |
|
---|
| 659 | for (i = 1; i < len; i++) {
|
---|
| 660 | if (p >= end) break;
|
---|
| 661 | c = *p++;
|
---|
| 662 | n <<= 8; n += c;
|
---|
| 663 | }
|
---|
| 664 | return n;
|
---|
| 665 | }
|
---|
| 666 |
|
---|
| 667 | extern int
|
---|
| 668 | onigenc_mbn_mbc_case_fold(OnigEncoding enc, OnigCaseFoldType flag ARG_UNUSED,
|
---|
| 669 | const UChar** pp, const UChar* end ARG_UNUSED,
|
---|
| 670 | UChar* lower)
|
---|
| 671 | {
|
---|
| 672 | int len;
|
---|
| 673 | const UChar *p = *pp;
|
---|
| 674 |
|
---|
| 675 | if (ONIGENC_IS_MBC_ASCII(p)) {
|
---|
| 676 | *lower = ONIGENC_ASCII_CODE_TO_LOWER_CASE(*p);
|
---|
| 677 | (*pp)++;
|
---|
| 678 | return 1;
|
---|
| 679 | }
|
---|
| 680 | else {
|
---|
| 681 | int i;
|
---|
| 682 |
|
---|
| 683 | len = enclen(enc, p);
|
---|
| 684 | for (i = 0; i < len; i++) {
|
---|
| 685 | *lower++ = *p++;
|
---|
| 686 | }
|
---|
| 687 | (*pp) += len;
|
---|
| 688 | return len; /* return byte length of converted to lower char */
|
---|
| 689 | }
|
---|
| 690 | }
|
---|
| 691 |
|
---|
| 692 | #if 0
|
---|
| 693 | extern int
|
---|
| 694 | onigenc_mbn_is_mbc_ambiguous(OnigEncoding enc, OnigCaseFoldType flag,
|
---|
| 695 | const UChar** pp, const UChar* end ARG_UNUSED)
|
---|
| 696 | {
|
---|
| 697 | const UChar* p = *pp;
|
---|
| 698 |
|
---|
| 699 | if (ONIGENC_IS_MBC_ASCII(p)) {
|
---|
| 700 | (*pp)++;
|
---|
| 701 | return ONIGENC_IS_ASCII_CODE_CASE_AMBIG(*p);
|
---|
| 702 | }
|
---|
| 703 |
|
---|
| 704 | (*pp) += enclen(enc, p);
|
---|
| 705 | return FALSE;
|
---|
| 706 | }
|
---|
| 707 | #endif
|
---|
| 708 |
|
---|
| 709 | extern int
|
---|
| 710 | onigenc_mb2_code_to_mbclen(OnigCodePoint code)
|
---|
| 711 | {
|
---|
| 712 | if ((code & 0xff00) != 0) return 2;
|
---|
| 713 | else return 1;
|
---|
| 714 | }
|
---|
| 715 |
|
---|
| 716 | extern int
|
---|
| 717 | onigenc_mb4_code_to_mbclen(OnigCodePoint code)
|
---|
| 718 | {
|
---|
| 719 | if ((code & 0xff000000) != 0) return 4;
|
---|
| 720 | else if ((code & 0xff0000) != 0) return 3;
|
---|
| 721 | else if ((code & 0xff00) != 0) return 2;
|
---|
| 722 | else return 1;
|
---|
| 723 | }
|
---|
| 724 |
|
---|
| 725 | extern int
|
---|
| 726 | onigenc_mb2_code_to_mbc(OnigEncoding enc, OnigCodePoint code, UChar *buf)
|
---|
| 727 | {
|
---|
| 728 | UChar *p = buf;
|
---|
| 729 |
|
---|
| 730 | if ((code & 0xff00) != 0) {
|
---|
| 731 | *p++ = (UChar )((code >> 8) & 0xff);
|
---|
| 732 | }
|
---|
| 733 | *p++ = (UChar )(code & 0xff);
|
---|
| 734 |
|
---|
| 735 | #if 1
|
---|
| 736 | if (enclen(enc, buf) != (p - buf))
|
---|
| 737 | return ONIGERR_INVALID_CODE_POINT_VALUE;
|
---|
| 738 | #endif
|
---|
| 739 | return (int )(p - buf);
|
---|
| 740 | }
|
---|
| 741 |
|
---|
| 742 | extern int
|
---|
| 743 | onigenc_mb4_code_to_mbc(OnigEncoding enc, OnigCodePoint code, UChar *buf)
|
---|
| 744 | {
|
---|
| 745 | UChar *p = buf;
|
---|
| 746 |
|
---|
| 747 | if ((code & 0xff000000) != 0) {
|
---|
| 748 | *p++ = (UChar )((code >> 24) & 0xff);
|
---|
| 749 | }
|
---|
| 750 | if ((code & 0xff0000) != 0 || p != buf) {
|
---|
| 751 | *p++ = (UChar )((code >> 16) & 0xff);
|
---|
| 752 | }
|
---|
| 753 | if ((code & 0xff00) != 0 || p != buf) {
|
---|
| 754 | *p++ = (UChar )((code >> 8) & 0xff);
|
---|
| 755 | }
|
---|
| 756 | *p++ = (UChar )(code & 0xff);
|
---|
| 757 |
|
---|
| 758 | #if 1
|
---|
| 759 | if (enclen(enc, buf) != (p - buf))
|
---|
| 760 | return ONIGERR_INVALID_CODE_POINT_VALUE;
|
---|
| 761 | #endif
|
---|
| 762 | return (int )(p - buf);
|
---|
| 763 | }
|
---|
| 764 |
|
---|
| 765 | extern int
|
---|
| 766 | onigenc_minimum_property_name_to_ctype(OnigEncoding enc, UChar* p, UChar* end)
|
---|
| 767 | {
|
---|
| 768 | static const PosixBracketEntryType PBS[] = {
|
---|
| 769 | POSIX_BRACKET_ENTRY_INIT("Alnum", ONIGENC_CTYPE_ALNUM),
|
---|
| 770 | POSIX_BRACKET_ENTRY_INIT("Alpha", ONIGENC_CTYPE_ALPHA),
|
---|
| 771 | POSIX_BRACKET_ENTRY_INIT("Blank", ONIGENC_CTYPE_BLANK),
|
---|
| 772 | POSIX_BRACKET_ENTRY_INIT("Cntrl", ONIGENC_CTYPE_CNTRL),
|
---|
| 773 | POSIX_BRACKET_ENTRY_INIT("Digit", ONIGENC_CTYPE_DIGIT),
|
---|
| 774 | POSIX_BRACKET_ENTRY_INIT("Graph", ONIGENC_CTYPE_GRAPH),
|
---|
| 775 | POSIX_BRACKET_ENTRY_INIT("Lower", ONIGENC_CTYPE_LOWER),
|
---|
| 776 | POSIX_BRACKET_ENTRY_INIT("Print", ONIGENC_CTYPE_PRINT),
|
---|
| 777 | POSIX_BRACKET_ENTRY_INIT("Punct", ONIGENC_CTYPE_PUNCT),
|
---|
| 778 | POSIX_BRACKET_ENTRY_INIT("Space", ONIGENC_CTYPE_SPACE),
|
---|
| 779 | POSIX_BRACKET_ENTRY_INIT("Upper", ONIGENC_CTYPE_UPPER),
|
---|
| 780 | POSIX_BRACKET_ENTRY_INIT("XDigit", ONIGENC_CTYPE_XDIGIT),
|
---|
| 781 | POSIX_BRACKET_ENTRY_INIT("ASCII", ONIGENC_CTYPE_ASCII),
|
---|
| 782 | POSIX_BRACKET_ENTRY_INIT("Word", ONIGENC_CTYPE_WORD),
|
---|
| 783 | };
|
---|
| 784 |
|
---|
| 785 | const PosixBracketEntryType *pb;
|
---|
| 786 | int len;
|
---|
| 787 |
|
---|
| 788 | len = onigenc_strlen(enc, p, end);
|
---|
| 789 | for (pb = PBS; pb < PBS + numberof(PBS); pb++) {
|
---|
| 790 | if (len == pb->len &&
|
---|
| 791 | onigenc_with_ascii_strnicmp(enc, p, end, pb->name, pb->len) == 0)
|
---|
| 792 | return pb->ctype;
|
---|
| 793 | }
|
---|
| 794 |
|
---|
| 795 | return ONIGERR_INVALID_CHAR_PROPERTY_NAME;
|
---|
| 796 | }
|
---|
| 797 |
|
---|
| 798 | extern int
|
---|
| 799 | onigenc_mb2_is_code_ctype(OnigEncoding enc, OnigCodePoint code,
|
---|
| 800 | unsigned int ctype)
|
---|
| 801 | {
|
---|
| 802 | if (code < 128)
|
---|
| 803 | return ONIGENC_IS_ASCII_CODE_CTYPE(code, ctype);
|
---|
| 804 | else {
|
---|
| 805 | if (CTYPE_IS_WORD_GRAPH_PRINT(ctype)) {
|
---|
| 806 | return (ONIGENC_CODE_TO_MBCLEN(enc, code) > 1 ? TRUE : FALSE);
|
---|
| 807 | }
|
---|
| 808 | }
|
---|
| 809 |
|
---|
| 810 | return FALSE;
|
---|
| 811 | }
|
---|
| 812 |
|
---|
| 813 | extern int
|
---|
| 814 | onigenc_mb4_is_code_ctype(OnigEncoding enc, OnigCodePoint code,
|
---|
| 815 | unsigned int ctype)
|
---|
| 816 | {
|
---|
| 817 | if (code < 128)
|
---|
| 818 | return ONIGENC_IS_ASCII_CODE_CTYPE(code, ctype);
|
---|
| 819 | else {
|
---|
| 820 | if (CTYPE_IS_WORD_GRAPH_PRINT(ctype)) {
|
---|
| 821 | return (ONIGENC_CODE_TO_MBCLEN(enc, code) > 1 ? TRUE : FALSE);
|
---|
| 822 | }
|
---|
| 823 | }
|
---|
| 824 |
|
---|
| 825 | return FALSE;
|
---|
| 826 | }
|
---|
| 827 |
|
---|
| 828 | extern int
|
---|
| 829 | onigenc_with_ascii_strncmp(OnigEncoding enc, const UChar* p, const UChar* end,
|
---|
| 830 | const UChar* sascii /* ascii */, int n)
|
---|
| 831 | {
|
---|
| 832 | int x, c;
|
---|
| 833 |
|
---|
| 834 | while (n-- > 0) {
|
---|
| 835 | if (p >= end) return (int )(*sascii);
|
---|
| 836 |
|
---|
| 837 | c = (int )ONIGENC_MBC_TO_CODE(enc, p, end);
|
---|
| 838 | x = *sascii - c;
|
---|
| 839 | if (x) return x;
|
---|
| 840 |
|
---|
| 841 | sascii++;
|
---|
| 842 | p += enclen(enc, p);
|
---|
| 843 | }
|
---|
| 844 | return 0;
|
---|
| 845 | }
|
---|
| 846 |
|
---|
| 847 | extern int
|
---|
| 848 | onigenc_with_ascii_strnicmp(OnigEncoding enc, const UChar* p, const UChar* end,
|
---|
| 849 | const UChar* sascii /* ascii */, int n)
|
---|
| 850 | {
|
---|
| 851 | int x, c;
|
---|
| 852 |
|
---|
| 853 | while (n-- > 0) {
|
---|
| 854 | if (p >= end) return (int )(*sascii);
|
---|
| 855 |
|
---|
| 856 | c = (int )ONIGENC_MBC_TO_CODE(enc, p, end);
|
---|
| 857 | if (ONIGENC_IS_ASCII_CODE(c))
|
---|
| 858 | c = ONIGENC_ASCII_CODE_TO_LOWER_CASE(c);
|
---|
| 859 | x = ONIGENC_ASCII_CODE_TO_LOWER_CASE(*sascii) - c;
|
---|
| 860 | if (x) return x;
|
---|
| 861 |
|
---|
| 862 | sascii++;
|
---|
| 863 | p += enclen(enc, p);
|
---|
| 864 | }
|
---|
| 865 | return 0;
|
---|
| 866 | }
|
---|
| 867 |
|
---|
| 868 | /* Property management */
|
---|
| 869 | static int
|
---|
| 870 | resize_property_list(int new_size, const OnigCodePoint*** plist, int* psize)
|
---|
| 871 | {
|
---|
| 872 | size_t size;
|
---|
| 873 | const OnigCodePoint **list = *plist;
|
---|
| 874 |
|
---|
| 875 | size = sizeof(OnigCodePoint*) * new_size;
|
---|
| 876 | if (IS_NULL(list)) {
|
---|
| 877 | list = (const OnigCodePoint** )xmalloc(size);
|
---|
| 878 | if (IS_NULL(list)) return ONIGERR_MEMORY;
|
---|
| 879 | }
|
---|
| 880 | else {
|
---|
| 881 | const OnigCodePoint **tmp;
|
---|
| 882 | tmp = (const OnigCodePoint** )xrealloc((void* )list, size);
|
---|
| 883 | if (IS_NULL(tmp)) return ONIGERR_MEMORY;
|
---|
| 884 | list = tmp;
|
---|
| 885 | }
|
---|
| 886 |
|
---|
| 887 | *plist = list;
|
---|
| 888 | *psize = new_size;
|
---|
| 889 |
|
---|
| 890 | return 0;
|
---|
| 891 | }
|
---|
| 892 |
|
---|
| 893 | extern int
|
---|
| 894 | onigenc_property_list_add_property(UChar* name, const OnigCodePoint* prop,
|
---|
| 895 | hash_table_type **table, const OnigCodePoint*** plist, int *pnum,
|
---|
| 896 | int *psize)
|
---|
| 897 | {
|
---|
| 898 | #define PROP_INIT_SIZE 16
|
---|
| 899 |
|
---|
| 900 | int r;
|
---|
| 901 |
|
---|
| 902 | if (*psize <= *pnum) {
|
---|
| 903 | int new_size = (*psize == 0 ? PROP_INIT_SIZE : *psize * 2);
|
---|
| 904 | r = resize_property_list(new_size, plist, psize);
|
---|
| 905 | if (r != 0) return r;
|
---|
| 906 | }
|
---|
| 907 |
|
---|
| 908 | (*plist)[*pnum] = prop;
|
---|
| 909 |
|
---|
| 910 | if (ONIG_IS_NULL(*table)) {
|
---|
| 911 | *table = onig_st_init_strend_table_with_size(PROP_INIT_SIZE);
|
---|
| 912 | if (ONIG_IS_NULL(*table)) return ONIGERR_MEMORY;
|
---|
| 913 | }
|
---|
| 914 |
|
---|
| 915 | *pnum = *pnum + 1;
|
---|
| 916 | onig_st_insert_strend(*table, name, name + strlen((char* )name),
|
---|
| 917 | (hash_data_type )(*pnum + ONIGENC_MAX_STD_CTYPE));
|
---|
| 918 | return 0;
|
---|
| 919 | }
|
---|
| 920 |
|
---|
| 921 | extern int
|
---|
| 922 | onigenc_property_list_init(int (*f)(void))
|
---|
| 923 | {
|
---|
| 924 | int r;
|
---|
| 925 |
|
---|
| 926 | THREAD_ATOMIC_START;
|
---|
| 927 |
|
---|
| 928 | r = f();
|
---|
| 929 |
|
---|
| 930 | THREAD_ATOMIC_END;
|
---|
| 931 | return r;
|
---|
| 932 | }
|
---|