1 /* $OpenBSD: main.c,v 1.11 2020/02/14 19:17:34 schwarze Exp $ */ 2 /* $NetBSD: main.c,v 1.2 1995/04/20 22:39:51 cgd Exp $ */ 3 4 #include <stdio.h> 5 #include <stdlib.h> 6 #include <string.h> 7 #include <sys/types.h> 8 #include <regex.h> 9 #include <assert.h> 10 #include <unistd.h> 11 12 #include "main.ih" 13 14 char *progname; 15 int debug = 0; 16 int line = 0; 17 int status = 0; 18 19 int copts = REG_EXTENDED; 20 int eopts = 0; 21 regoff_t startoff = 0; 22 regoff_t endoff = 0; 23 24 25 extern int split(char *, char *[], int, char *); 26 extern void regprint(regex_t *, FILE *); 27 28 /* 29 - main - do the simple case, hand off to regress() for regression 30 */ 31 int 32 main(int argc, char *argv[]) 33 34 { 35 regex_t re; 36 # define NS 10 37 regmatch_t subs[NS]; 38 char erbuf[100]; 39 int err; 40 size_t len; 41 int c; 42 int errflg = 0; 43 register int i; 44 45 progname = argv[0]; 46 47 while ((c = getopt(argc, argv, "c:E:e:S:x")) != -1) 48 switch (c) { 49 case 'c': /* compile options */ 50 copts = options('c', optarg); 51 break; 52 case 'E': /* end offset */ 53 endoff = (regoff_t)atoi(optarg); 54 break; 55 case 'e': /* execute options */ 56 eopts = options('e', optarg); 57 break; 58 case 'S': /* start offset */ 59 startoff = (regoff_t)atoi(optarg); 60 break; 61 case 'x': /* Debugging. */ 62 debug++; 63 break; 64 case '?': 65 default: 66 errflg++; 67 break; 68 } 69 if (errflg) { 70 fprintf(stderr, "usage: %s ", progname); 71 fprintf(stderr, "[-x] [-c copt] [-E endoff] [-e eopt] [-S startoff] [re]\n"); 72 exit(2); 73 } 74 75 if (optind >= argc) { 76 regress(stdin); 77 exit(status); 78 } 79 80 err = regcomp(&re, argv[optind++], copts); 81 if (err) { 82 len = regerror(err, &re, erbuf, sizeof(erbuf)); 83 fprintf(stderr, "error %s, %zu/%zu `%s'\n", 84 eprint(err), len, sizeof(erbuf), erbuf); 85 exit(status); 86 } 87 regprint(&re, stdout); 88 89 if (optind >= argc) { 90 regfree(&re); 91 exit(status); 92 } 93 94 if (eopts®_STARTEND) { 95 subs[0].rm_so = startoff; 96 subs[0].rm_eo = strlen(argv[optind]) - endoff; 97 } 98 err = regexec(&re, argv[optind], (size_t)NS, subs, eopts); 99 if (err) { 100 len = regerror(err, &re, erbuf, sizeof(erbuf)); 101 fprintf(stderr, "error %s, %zu/%zu `%s'\n", 102 eprint(err), len, sizeof(erbuf), erbuf); 103 exit(status); 104 } 105 if (!(copts®_NOSUB)) { 106 len = (size_t)(subs[0].rm_eo - subs[0].rm_so); 107 if (subs[0].rm_so != -1) { 108 if (len != 0) 109 printf("match `%.*s'\n", (int)len, 110 argv[optind] + subs[0].rm_so); 111 else 112 printf("match `'@%.1s\n", 113 argv[optind] + subs[0].rm_so); 114 } 115 for (i = 1; i < NS; i++) 116 if (subs[i].rm_so != -1) 117 printf("(%d) `%.*s'\n", i, 118 (int)(subs[i].rm_eo - subs[i].rm_so), 119 argv[optind] + subs[i].rm_so); 120 } 121 exit(status); 122 } 123 124 /* 125 - regress - main loop of regression test 126 == void regress(FILE *in); 127 */ 128 void 129 regress(in) 130 FILE *in; 131 { 132 char inbuf[1000]; 133 # define MAXF 10 134 char *f[MAXF]; 135 int nf; 136 int i; 137 char erbuf[100]; 138 size_t ne; 139 char *badpat = "invalid regular expression"; 140 # define SHORT 10 141 char *bpname = "REG_BADPAT"; 142 regex_t re; 143 144 while (fgets(inbuf, sizeof(inbuf), in) != NULL) { 145 line++; 146 if (inbuf[0] == '#' || inbuf[0] == '\n') 147 continue; /* NOTE CONTINUE */ 148 inbuf[strcspn(inbuf, "\n")] = '\0'; /* get rid of stupid \n */ 149 if (debug) 150 fprintf(stdout, "%d:\n", line); 151 nf = split(inbuf, f, MAXF, "\t\t"); 152 if (nf < 3) { 153 fprintf(stderr, "bad input, line %d\n", line); 154 exit(1); 155 } 156 for (i = 0; i < nf; i++) 157 if (strcmp(f[i], "\"\"") == 0) 158 f[i] = ""; 159 if (nf <= 3) 160 f[3] = NULL; 161 if (nf <= 4) 162 f[4] = NULL; 163 try(f[0], f[1], f[2], f[3], f[4], options('c', f[1])); 164 if (opt('&', f[1])) /* try with either type of RE */ 165 try(f[0], f[1], f[2], f[3], f[4], 166 options('c', f[1]) &~ REG_EXTENDED); 167 } 168 169 ne = regerror(REG_BADPAT, (regex_t *)NULL, erbuf, sizeof(erbuf)); 170 if (strcmp(erbuf, badpat) != 0 || ne != strlen(badpat)+1) { 171 fprintf(stderr, "end: regerror() test gave `%s' not `%s'\n", 172 erbuf, badpat); 173 status = 1; 174 } 175 ne = regerror(REG_BADPAT, (regex_t *)NULL, erbuf, (size_t)SHORT); 176 if (strncmp(erbuf, badpat, SHORT-1) != 0 || erbuf[SHORT-1] != '\0' || 177 ne != strlen(badpat)+1) { 178 fprintf(stderr, "end: regerror() short test gave `%s' not `%.*s'\n", 179 erbuf, SHORT-1, badpat); 180 status = 1; 181 } 182 ne = regerror(REG_ITOA|REG_BADPAT, (regex_t *)NULL, erbuf, sizeof(erbuf)); 183 if (strcmp(erbuf, bpname) != 0 || ne != strlen(bpname)+1) { 184 fprintf(stderr, "end: regerror() ITOA test gave `%s' not `%s'\n", 185 erbuf, bpname); 186 status = 1; 187 } 188 re.re_endp = bpname; 189 ne = regerror(REG_ATOI, &re, erbuf, sizeof(erbuf)); 190 if (atoi(erbuf) != (int)REG_BADPAT) { 191 fprintf(stderr, "end: regerror() ATOI test gave `%s' not `%ld'\n", 192 erbuf, (long)REG_BADPAT); 193 status = 1; 194 } else if (ne != strlen(erbuf)+1) { 195 fprintf(stderr, "end: regerror() ATOI test len(`%s') = %ld\n", 196 erbuf, (long)REG_BADPAT); 197 status = 1; 198 } 199 } 200 201 /* 202 - try - try it, and report on problems 203 == void try(char *f0, char *f1, char *f2, char *f3, char *f4, int opts); 204 */ 205 void 206 try(f0, f1, f2, f3, f4, opts) 207 char *f0; 208 char *f1; 209 char *f2; 210 char *f3; 211 char *f4; 212 int opts; /* may not match f1 */ 213 { 214 regex_t re; 215 # define NSUBS 10 216 regmatch_t subs[NSUBS]; 217 # define NSHOULD 15 218 char *should[NSHOULD]; 219 int nshould; 220 char erbuf[100]; 221 int err; 222 int len; 223 char *type = (opts & REG_EXTENDED) ? "ERE" : "BRE"; 224 register int i; 225 char *grump; 226 char f0copy[1000]; 227 char f2copy[1000]; 228 229 strlcpy(f0copy, f0, sizeof f0copy); 230 re.re_endp = (opts®_PEND) ? f0copy + strlen(f0copy) : NULL; 231 fixstr(f0copy); 232 err = regcomp(&re, f0copy, opts); 233 if (err != 0 && (!opt('C', f1) || err != efind(f2))) { 234 /* unexpected error or wrong error */ 235 len = regerror(err, &re, erbuf, sizeof(erbuf)); 236 fprintf(stderr, "%d: %s error %s, %d/%zu `%s'\n", 237 line, type, eprint(err), len, 238 sizeof(erbuf), erbuf); 239 status = 1; 240 } else if (err == 0 && opt('C', f1)) { 241 /* unexpected success */ 242 fprintf(stderr, "%d: %s should have given REG_%s\n", 243 line, type, f2); 244 status = 1; 245 err = 1; /* so we won't try regexec */ 246 } 247 248 if (err != 0) { 249 regfree(&re); 250 return; 251 } 252 253 strlcpy(f2copy, f2, sizeof f2copy); 254 fixstr(f2copy); 255 256 if (options('e', f1)®_STARTEND) { 257 if (strchr(f2, '(') == NULL || strchr(f2, ')') == NULL) 258 fprintf(stderr, "%d: bad STARTEND syntax\n", line); 259 subs[0].rm_so = strchr(f2, '(') - f2 + 1; 260 subs[0].rm_eo = strchr(f2, ')') - f2; 261 /* the preceding character is relevant with REG_NOTBOL */ 262 f2copy[subs[0].rm_so - 1] = subs[0].rm_so > 1 ? 263 f2copy[subs[0].rm_so - 2] : 'X'; 264 } 265 err = regexec(&re, f2copy, NSUBS, subs, options('e', f1)); 266 267 if (err != 0 && (f3 != NULL || err != REG_NOMATCH)) { 268 /* unexpected error or wrong error */ 269 len = regerror(err, &re, erbuf, sizeof(erbuf)); 270 fprintf(stderr, "%d: %s exec error %s, %d/%zu `%s'\n", 271 line, type, eprint(err), len, 272 sizeof(erbuf), erbuf); 273 status = 1; 274 } else if (err != 0) { 275 /* nothing more to check */ 276 } else if (f3 == NULL) { 277 /* unexpected success */ 278 fprintf(stderr, "%d: %s exec should have failed\n", 279 line, type); 280 status = 1; 281 err = 1; /* just on principle */ 282 } else if (opts®_NOSUB) { 283 /* nothing more to check */ 284 } else if ((grump = check(f2, subs[0], f3)) != NULL) { 285 fprintf(stderr, "%d: %s %s\n", line, type, grump); 286 status = 1; 287 err = 1; 288 } 289 290 if (err != 0 || f4 == NULL) { 291 regfree(&re); 292 return; 293 } 294 295 for (i = 1; i < NSHOULD; i++) 296 should[i] = NULL; 297 nshould = split(f4, should+1, NSHOULD-1, ","); 298 if (nshould == 0) { 299 nshould = 1; 300 should[1] = ""; 301 } 302 for (i = 1; i < NSUBS; i++) { 303 grump = check(f2, subs[i], should[i]); 304 if (grump != NULL) { 305 fprintf(stderr, "%d: %s $%d %s\n", line, 306 type, i, grump); 307 status = 1; 308 err = 1; 309 } 310 } 311 312 regfree(&re); 313 } 314 315 /* 316 - options - pick options out of a regression-test string 317 == int options(int type, char *s); 318 */ 319 int 320 options(type, s) 321 int type; /* 'c' compile, 'e' exec */ 322 char *s; 323 { 324 register char *p; 325 register int o = (type == 'c') ? copts : eopts; 326 register char *legal = (type == 'c') ? "bisnmp" : "^$#tl"; 327 328 for (p = s; *p != '\0'; p++) 329 if (strchr(legal, *p) != NULL) 330 switch (*p) { 331 case 'b': 332 o &= ~REG_EXTENDED; 333 break; 334 case 'i': 335 o |= REG_ICASE; 336 break; 337 case 's': 338 o |= REG_NOSUB; 339 break; 340 case 'n': 341 o |= REG_NEWLINE; 342 break; 343 case 'm': 344 o &= ~REG_EXTENDED; 345 o |= REG_NOSPEC; 346 break; 347 case 'p': 348 o |= REG_PEND; 349 break; 350 case '^': 351 o |= REG_NOTBOL; 352 break; 353 case '$': 354 o |= REG_NOTEOL; 355 break; 356 case '#': 357 o |= REG_STARTEND; 358 break; 359 case 't': /* trace */ 360 o |= REG_TRACE; 361 break; 362 case 'l': /* force long representation */ 363 o |= REG_LARGE; 364 break; 365 case 'r': /* force backref use */ 366 o |= REG_BACKR; 367 break; 368 } 369 return(o); 370 } 371 372 /* 373 - opt - is a particular option in a regression string? 374 == int opt(int c, char *s); 375 */ 376 int /* predicate */ 377 opt(c, s) 378 int c; 379 char *s; 380 { 381 return(strchr(s, c) != NULL); 382 } 383 384 /* 385 - fixstr - transform magic characters in strings 386 == void fixstr(register char *p); 387 */ 388 void 389 fixstr(p) 390 register char *p; 391 { 392 if (p == NULL) 393 return; 394 395 for (; *p != '\0'; p++) 396 if (*p == 'N') 397 *p = '\n'; 398 else if (*p == 'T') 399 *p = '\t'; 400 else if (*p == 'S') 401 *p = ' '; 402 else if (*p == 'Z') 403 *p = '\0'; 404 } 405 406 /* 407 - check - check a substring match 408 == char *check(char *str, regmatch_t sub, char *should); 409 */ 410 char * /* NULL or complaint */ 411 check(str, sub, should) 412 char *str; 413 regmatch_t sub; 414 char *should; 415 { 416 register int len; 417 register int shlen; 418 register char *p; 419 static char grump[500]; 420 register char *at = NULL; 421 422 if (should != NULL && strcmp(should, "-") == 0) 423 should = NULL; 424 if (should != NULL && should[0] == '@') { 425 at = should + 1; 426 should = ""; 427 } 428 429 /* check rm_so and rm_eo for consistency */ 430 if (sub.rm_so > sub.rm_eo || (sub.rm_so == -1 && sub.rm_eo != -1) || 431 (sub.rm_so != -1 && sub.rm_eo == -1) || 432 (sub.rm_so != -1 && sub.rm_so < 0) || 433 (sub.rm_eo != -1 && sub.rm_eo < 0) ) { 434 snprintf(grump, sizeof grump, 435 "start %ld end %ld", (long)sub.rm_so, 436 (long)sub.rm_eo); 437 return(grump); 438 } 439 440 /* check for no match */ 441 if (sub.rm_so == -1 && should == NULL) 442 return(NULL); 443 if (sub.rm_so == -1) 444 return("did not match"); 445 446 /* check for in range */ 447 if (sub.rm_eo > strlen(str)) { 448 snprintf(grump, sizeof grump, 449 "start %ld end %ld, past end of string", 450 (long)sub.rm_so, (long)sub.rm_eo); 451 return(grump); 452 } 453 454 len = (int)(sub.rm_eo - sub.rm_so); 455 p = str + sub.rm_so; 456 457 /* check for not supposed to match */ 458 if (should == NULL) { 459 snprintf(grump, sizeof grump, "matched `%.*s'", len, p); 460 return(grump); 461 } 462 463 /* check for wrong match */ 464 shlen = (int)strlen(should); 465 if (len != shlen || strncmp(p, should, (size_t)shlen) != 0) { 466 snprintf(grump, sizeof grump, "matched `%.*s' instead", len, p); 467 return(grump); 468 } 469 if (shlen > 0) 470 return(NULL); 471 472 /* check null match in right place */ 473 if (at == NULL) 474 return(NULL); 475 shlen = strlen(at); 476 if (shlen == 0) 477 shlen = 1; /* force check for end-of-string */ 478 if (strncmp(p, at, shlen) != 0) { 479 snprintf(grump, sizeof grump, "matched null at `%.20s'", p); 480 return(grump); 481 } 482 return(NULL); 483 } 484 485 /* 486 - eprint - convert error number to name 487 == static char *eprint(int err); 488 */ 489 static char * 490 eprint(err) 491 int err; 492 { 493 static char epbuf[100]; 494 size_t len; 495 496 len = regerror(REG_ITOA|err, (regex_t *)NULL, epbuf, sizeof(epbuf)); 497 assert(len <= sizeof(epbuf)); 498 return(epbuf); 499 } 500 501 /* 502 - efind - convert error name to number 503 == static int efind(char *name); 504 */ 505 static int 506 efind(name) 507 char *name; 508 { 509 static char efbuf[100]; 510 regex_t re; 511 512 snprintf(efbuf, sizeof efbuf, "REG_%s", name); 513 assert(strlen(efbuf) < sizeof(efbuf)); 514 re.re_endp = efbuf; 515 (void) regerror(REG_ATOI, &re, efbuf, sizeof(efbuf)); 516 return(atoi(efbuf)); 517 } 518