regex.c revision 1.2
1 /* $NetBSD: regex.c,v 1.2 2018/08/12 13:02:37 christos Exp $ */ 2 3 /* 4 * Copyright (C) Internet Systems Consortium, Inc. ("ISC") 5 * 6 * This Source Code Form is subject to the terms of the Mozilla Public 7 * License, v. 2.0. If a copy of the MPL was not distributed with this 8 * file, You can obtain one at http://mozilla.org/MPL/2.0/. 9 * 10 * See the COPYRIGHT file distributed with this work for additional 11 * information regarding copyright ownership. 12 */ 13 14 #include <config.h> 15 16 #include <isc/file.h> 17 #include <isc/print.h> 18 #include <isc/regex.h> 19 #include <isc/string.h> 20 21 #if VALREGEX_REPORT_REASON 22 #define FAIL(x) do { reason = (x); goto error; } while(/*CONSTCOND*/0) 23 #else 24 #define FAIL(x) goto error 25 #endif 26 27 /* 28 * Validate the regular expression 'C' locale. 29 */ 30 int 31 isc_regex_validate(const char *c) { 32 enum { 33 none, parse_bracket, parse_bound, 34 parse_ce, parse_ec, parse_cc 35 } state = none; 36 /* Well known character classes. */ 37 const char *cc[] = { 38 ":alnum:", ":digit:", ":punct:", ":alpha:", ":graph:", 39 ":space:", ":blank:", ":lower:", ":upper:", ":cntrl:", 40 ":print:", ":xdigit:" 41 }; 42 isc_boolean_t seen_comma = ISC_FALSE; 43 isc_boolean_t seen_high = ISC_FALSE; 44 isc_boolean_t seen_char = ISC_FALSE; 45 isc_boolean_t seen_ec = ISC_FALSE; 46 isc_boolean_t seen_ce = ISC_FALSE; 47 isc_boolean_t have_atom = ISC_FALSE; 48 int group = 0; 49 int range = 0; 50 int sub = 0; 51 isc_boolean_t empty_ok = ISC_FALSE; 52 isc_boolean_t neg = ISC_FALSE; 53 isc_boolean_t was_multiple = ISC_FALSE; 54 unsigned int low = 0; 55 unsigned int high = 0; 56 const char *ccname = NULL; 57 int range_start = 0; 58 #if VALREGEX_REPORT_REASON 59 const char *reason = ""; 60 #endif 61 62 if (c == NULL || *c == 0) 63 FAIL("empty string"); 64 65 while (c != NULL && *c != 0) { 66 switch (state) { 67 case none: 68 switch (*c) { 69 case '\\': /* make literal */ 70 ++c; 71 switch (*c) { 72 case '1': case '2': case '3': 73 case '4': case '5': case '6': 74 case '7': case '8': case '9': 75 if ((*c - '0') > sub) 76 FAIL("bad back reference"); 77 have_atom = ISC_TRUE; 78 was_multiple = ISC_FALSE; 79 break; 80 case 0: 81 FAIL("escaped end-of-string"); 82 default: 83 goto literal; 84 } 85 ++c; 86 break; 87 case '[': /* bracket start */ 88 ++c; 89 neg = ISC_FALSE; 90 was_multiple = ISC_FALSE; 91 seen_char = ISC_FALSE; 92 state = parse_bracket; 93 break; 94 case '{': /* bound start */ 95 switch (c[1]) { 96 case '0': case '1': case '2': case '3': 97 case '4': case '5': case '6': case '7': 98 case '8': case '9': 99 if (!have_atom) 100 FAIL("no atom"); 101 if (was_multiple) 102 FAIL("was multiple"); 103 seen_comma = ISC_FALSE; 104 seen_high = ISC_FALSE; 105 low = high = 0; 106 state = parse_bound; 107 break; 108 default: 109 goto literal; 110 } 111 ++c; 112 have_atom = ISC_TRUE; 113 was_multiple = ISC_TRUE; 114 break; 115 case '}': 116 goto literal; 117 case '(': /* group start */ 118 have_atom = ISC_FALSE; 119 was_multiple = ISC_FALSE; 120 empty_ok = ISC_TRUE; 121 ++group; 122 ++sub; 123 ++c; 124 break; 125 case ')': /* group end */ 126 if (group && !have_atom && !empty_ok) 127 FAIL("empty alternative"); 128 have_atom = ISC_TRUE; 129 was_multiple = ISC_FALSE; 130 if (group != 0) 131 --group; 132 ++c; 133 break; 134 case '|': /* alternative seperator */ 135 if (!have_atom) 136 FAIL("no atom"); 137 have_atom = ISC_FALSE; 138 empty_ok = ISC_FALSE; 139 was_multiple = ISC_FALSE; 140 ++c; 141 break; 142 case '^': 143 case '$': 144 have_atom = ISC_TRUE; 145 was_multiple = ISC_TRUE; 146 ++c; 147 break; 148 case '+': 149 case '*': 150 case '?': 151 if (was_multiple) 152 FAIL("was multiple"); 153 if (!have_atom) 154 FAIL("no atom"); 155 have_atom = ISC_TRUE; 156 was_multiple = ISC_TRUE; 157 ++c; 158 break; 159 case '.': 160 default: 161 literal: 162 have_atom = ISC_TRUE; 163 was_multiple = ISC_FALSE; 164 ++c; 165 break; 166 } 167 break; 168 case parse_bound: 169 switch (*c) { 170 case '0': case '1': case '2': case '3': case '4': 171 case '5': case '6': case '7': case '8': case '9': 172 if (!seen_comma) { 173 low = low * 10 + *c - '0'; 174 if (low > 255) 175 FAIL("lower bound too big"); 176 } else { 177 seen_high = ISC_TRUE; 178 high = high * 10 + *c - '0'; 179 if (high > 255) 180 FAIL("upper bound too big"); 181 } 182 ++c; 183 break; 184 case ',': 185 if (seen_comma) 186 FAIL("multiple commas"); 187 seen_comma = ISC_TRUE; 188 ++c; 189 break; 190 default: 191 case '{': 192 FAIL("non digit/comma"); 193 case '}': 194 if (seen_high && low > high) 195 FAIL("bad parse bound"); 196 seen_comma = ISC_FALSE; 197 state = none; 198 ++c; 199 break; 200 } 201 break; 202 case parse_bracket: 203 switch (*c) { 204 case '^': 205 if (seen_char || neg) goto inside; 206 neg = ISC_TRUE; 207 ++c; 208 break; 209 case '-': 210 if (range == 2) goto inside; 211 if (!seen_char) goto inside; 212 if (range == 1) 213 FAIL("bad range"); 214 range = 2; 215 ++c; 216 break; 217 case '[': 218 ++c; 219 switch (*c) { 220 case '.': /* collating element */ 221 if (range != 0) --range; 222 ++c; 223 state = parse_ce; 224 seen_ce = ISC_FALSE; 225 break; 226 case '=': /* equivalence class */ 227 if (range == 2) 228 FAIL("equivalence class in range"); 229 ++c; 230 state = parse_ec; 231 seen_ec = ISC_FALSE; 232 break; 233 case ':': /* character class */ 234 if (range == 2) 235 FAIL("character class in range"); 236 ccname = c; 237 ++c; 238 state = parse_cc; 239 break; 240 } 241 seen_char = ISC_TRUE; 242 break; 243 case ']': 244 if (!c[1] && !seen_char) 245 FAIL("unfinished brace"); 246 if (!seen_char) 247 goto inside; 248 ++c; 249 range = 0; 250 have_atom = ISC_TRUE; 251 state = none; 252 break; 253 default: 254 inside: 255 seen_char = ISC_TRUE; 256 if (range == 2 && (*c & 0xff) < range_start) 257 FAIL("out of order range"); 258 if (range != 0) 259 --range; 260 range_start = *c & 0xff; 261 ++c; 262 break; 263 }; 264 break; 265 case parse_ce: 266 switch (*c) { 267 case '.': 268 ++c; 269 switch (*c) { 270 case ']': 271 if (!seen_ce) 272 FAIL("empty ce"); 273 ++c; 274 state = parse_bracket; 275 break; 276 default: 277 if (seen_ce) 278 range_start = 256; 279 else 280 range_start = '.'; 281 seen_ce = ISC_TRUE; 282 break; 283 } 284 break; 285 default: 286 if (seen_ce) 287 range_start = 256; 288 else 289 range_start = *c; 290 seen_ce = ISC_TRUE; 291 ++c; 292 break; 293 } 294 break; 295 case parse_ec: 296 switch (*c) { 297 case '=': 298 ++c; 299 switch (*c) { 300 case ']': 301 if (!seen_ec) 302 FAIL("no ec"); 303 ++c; 304 state = parse_bracket; 305 break; 306 default: 307 seen_ec = ISC_TRUE; 308 break; 309 } 310 break; 311 default: 312 seen_ec = ISC_TRUE; 313 ++c; 314 break; 315 } 316 break; 317 case parse_cc: 318 switch (*c) { 319 case ':': 320 ++c; 321 switch (*c) { 322 case ']': { 323 unsigned int i; 324 isc_boolean_t found = ISC_FALSE; 325 for (i = 0; 326 i < sizeof(cc)/sizeof(*cc); 327 i++) 328 { 329 unsigned int len; 330 len = strlen(cc[i]); 331 if (len != 332 (unsigned int)(c - ccname)) 333 continue; 334 if (strncmp(cc[i], ccname, len)) 335 continue; 336 found = ISC_TRUE; 337 } 338 if (!found) 339 FAIL("unknown cc"); 340 ++c; 341 state = parse_bracket; 342 break; 343 } 344 default: 345 break; 346 } 347 break; 348 default: 349 ++c; 350 break; 351 } 352 break; 353 } 354 } 355 if (group != 0) 356 FAIL("group open"); 357 if (state != none) 358 FAIL("incomplete"); 359 if (!have_atom) 360 FAIL("no atom"); 361 return (sub); 362 363 error: 364 #if VALREGEX_REPORT_REASON 365 fprintf(stderr, "%s\n", reason); 366 #endif 367 return (-1); 368 } 369
Indexes created Sun Mar 01 05:31:48 UTC 2026