str.c revision 1590
11901Swollman/*- 274462Salfred * Copyright (c) 1991, 1993 350476Speter * The Regents of the University of California. All rights reserved. 448794Snik * 574462Salfred * Redistribution and use in source and binary forms, with or without 620779Smpp * modification, are permitted provided that the following conditions 720779Smpp * are met: 820779Smpp * 1. Redistributions of source code must retain the above copyright 956629Sshin * notice, this list of conditions and the following disclaimer. 1056660Sbde * 2. Redistributions in binary form must reproduce the above copyright 1168947Sru * notice, this list of conditions and the following disclaimer in the 1274462Salfred * documentation and/or other materials provided with the distribution. 1374462Salfred * 3. All advertising materials mentioning features or use of this software 1420779Smpp * must display the following acknowledgement: 1584306Sru * This product includes software developed by the University of 1684306Sru * California, Berkeley and its contributors. 1720779Smpp * 4. Neither the name of the University nor the names of its contributors 1856629Sshin * may be used to endorse or promote products derived from this software 1956629Sshin * without specific prior written permission. 2056629Sshin * 2120779Smpp * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 22108037Sru * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 2356660Sbde * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 2456629Sshin * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 2556660Sbde * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 26108037Sru * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 2756629Sshin * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 2821064Speter * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 291901Swollman * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 301901Swollman * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 3120779Smpp * SUCH DAMAGE. 3274462Salfred */ 3374462Salfred 3474462Salfred#ifndef lint 3574462Salfredstatic char sccsid[] = "@(#)str.c 8.1 (Berkeley) 6/6/93"; 3674462Salfred#endif /* not lint */ 3774462Salfred 3874462Salfred#include <sys/cdefs.h> 3956629Sshin#include <sys/types.h> 4074462Salfred 4174462Salfred#include <errno.h> 4274462Salfred#include <stddef.h> 4374462Salfred#include <stdio.h> 4474462Salfred#include <stdlib.h> 4574462Salfred#include <string.h> 4674462Salfred 4774462Salfred#include "extern.h" 4874462Salfred 4974462Salfredstatic int backslash __P((STR *)); 5074462Salfredstatic int bracket __P((STR *)); 5156629Sshinstatic int c_class __P((const void *, const void *)); 5274462Salfredstatic void genclass __P((STR *)); 5374462Salfredstatic void genequiv __P((STR *)); 5474462Salfredstatic int genrange __P((STR *)); 5574462Salfredstatic void genseq __P((STR *)); 5674462Salfred 5774462Salfredint 5874462Salfrednext(s) 5956629Sshin register STR *s; 6074462Salfred{ 6156629Sshin register int ch; 6274462Salfred 6374462Salfred switch (s->state) { 6474462Salfred case EOS: 6556629Sshin return (0); 6656629Sshin case INFINITE: 6756629Sshin return (1); 6856629Sshin case NORMAL: 6956629Sshin switch (ch = *s->str) { 70108037Sru case '\0': 7156629Sshin s->state = EOS; 72108037Sru return (0); 7356629Sshin case '\\': 7456629Sshin s->lastch = backslash(s); 7556629Sshin break; 7656629Sshin case '[': 7756629Sshin if (bracket(s)) 7856629Sshin return (next(s)); 7956629Sshin /* FALLTHROUGH */ 8056629Sshin default: 8174462Salfred ++s->str; 8256629Sshin s->lastch = ch; 8356629Sshin break; 8456629Sshin } 8574462Salfred 8674462Salfred /* We can start a range at any time. */ 8756629Sshin if (s->str[0] == '-' && genrange(s)) 8874462Salfred return (next(s)); 89108037Sru return (1); 9074462Salfred case RANGE: 91108037Sru if (s->cnt-- == 0) { 9274462Salfred s->state = NORMAL; 9374462Salfred return (next(s)); 9474462Salfred } 9556629Sshin ++s->lastch; 9674462Salfred return (1); 9779754Sdd case SEQUENCE: 9874462Salfred if (s->cnt-- == 0) { 9974462Salfred s->state = NORMAL; 10074462Salfred return (next(s)); 10174462Salfred } 10274462Salfred return (1); 10374462Salfred case SET: 104 if ((s->lastch = s->set[s->cnt++]) == OOBCH) { 105 s->state = NORMAL; 106 return (next(s)); 107 } 108 return (1); 109 } 110 /* NOTREACHED */ 111} 112 113static int 114bracket(s) 115 register STR *s; 116{ 117 register char *p; 118 119 switch (s->str[1]) { 120 case ':': /* "[:class:]" */ 121 if ((p = strstr(s->str + 2, ":]")) == NULL) 122 return (0); 123 *p = '\0'; 124 s->str += 2; 125 genclass(s); 126 s->str = p + 2; 127 return (1); 128 case '=': /* "[=equiv=]" */ 129 if ((p = strstr(s->str + 2, "=]")) == NULL) 130 return (0); 131 s->str += 2; 132 genequiv(s); 133 return (1); 134 default: /* "[\###*n]" or "[#*n]" */ 135 if ((p = strpbrk(s->str + 2, "*]")) == NULL) 136 return (0); 137 if (p[0] != '*' || index(p, ']') == NULL) 138 return (0); 139 s->str += 1; 140 genseq(s); 141 return (1); 142 } 143 /* NOTREACHED */ 144} 145 146int isalnum __P((int)), 147 isalpha __P((int)), 148 isblank __P((int)), 149 isspace __P((int)), 150 iscntrl __P((int)), 151 isdigit __P((int)), 152 isgraph __P((int)), 153 islower __P((int)), 154 isprint __P((int)), 155 ispunct __P((int)), 156 isupper __P((int)), 157 isxdigit __P((int)); 158 159typedef struct { 160 char *name; 161 int (*func) __P((int)); 162 int *set; 163} CLASS; 164 165static CLASS classes[] = { 166 { "alnum", isalnum, }, 167 { "alpha", isalpha, }, 168 { "blank", isblank, }, 169 { "cntrl", iscntrl, }, 170 { "digit", isdigit, }, 171 { "graph", isgraph, }, 172 { "lower", islower, }, 173 { "print", isupper, }, 174 { "punct", ispunct, }, 175 { "space", isspace, }, 176 { "upper", isupper, }, 177 { "xdigit", isxdigit, }, 178}; 179 180static void 181genclass(s) 182 STR *s; 183{ 184 register int cnt, (*func) __P((int)); 185 CLASS *cp, tmp; 186 int *p; 187 188 tmp.name = s->str; 189 if ((cp = (CLASS *)bsearch(&tmp, classes, sizeof(classes) / 190 sizeof(CLASS), sizeof(CLASS), c_class)) == NULL) 191 err("unknown class %s", s->str); 192 193 if ((cp->set = p = malloc((NCHARS + 1) * sizeof(int))) == NULL) 194 err("%s", strerror(errno)); 195 bzero(p, (NCHARS + 1) * sizeof(int)); 196 for (cnt = 0, func = cp->func; cnt < NCHARS; ++cnt) 197 if ((func)(cnt)) 198 *p++ = cnt; 199 *p = OOBCH; 200 201 s->cnt = 0; 202 s->state = SET; 203 s->set = cp->set; 204} 205 206static int 207c_class(a, b) 208 const void *a, *b; 209{ 210 return (strcmp(((CLASS *)a)->name, ((CLASS *)b)->name)); 211} 212 213/* 214 * English doesn't have any equivalence classes, so for now 215 * we just syntax check and grab the character. 216 */ 217static void 218genequiv(s) 219 STR *s; 220{ 221 if (*s->str == '\\') { 222 s->equiv[0] = backslash(s); 223 if (*s->str != '=') 224 err("misplaced equivalence equals sign"); 225 } else { 226 s->equiv[0] = s->str[0]; 227 if (s->str[1] != '=') 228 err("misplaced equivalence equals sign"); 229 } 230 s->str += 2; 231 s->cnt = 0; 232 s->state = SET; 233 s->set = s->equiv; 234} 235 236static int 237genrange(s) 238 STR *s; 239{ 240 int stopval; 241 char *savestart; 242 243 savestart = s->str; 244 stopval = *++s->str == '\\' ? backslash(s) : *s->str; 245 if (stopval < s->lastch) { 246 s->str = savestart; 247 return (0); 248 } 249 s->cnt = stopval - s->lastch + 1; 250 s->state = RANGE; 251 --s->lastch; 252 return (1); 253} 254 255static void 256genseq(s) 257 STR *s; 258{ 259 char *ep; 260 261 if (s->which == STRING1) 262 err("sequences only valid in string2"); 263 264 if (*s->str == '\\') 265 s->lastch = backslash(s); 266 else 267 s->lastch = *s->str++; 268 if (*s->str != '*') 269 err("misplaced sequence asterisk"); 270 271 switch (*++s->str) { 272 case '\\': 273 s->cnt = backslash(s); 274 break; 275 case ']': 276 s->cnt = 0; 277 ++s->str; 278 break; 279 default: 280 if (isdigit(*s->str)) { 281 s->cnt = strtol(s->str, &ep, 0); 282 if (*ep == ']') { 283 s->str = ep + 1; 284 break; 285 } 286 } 287 err("illegal sequence count"); 288 /* NOTREACHED */ 289 } 290 291 s->state = s->cnt ? SEQUENCE : INFINITE; 292} 293 294/* Use the #defines isXXX() here, DON'T use them above. */ 295#include <ctype.h> 296 297/* 298 * Translate \??? into a character. Up to 3 octal digits, if no digits either 299 * an escape code or a literal character. 300 */ 301static int 302backslash(s) 303 register STR *s; 304{ 305 register int ch, cnt, val; 306 307 for (cnt = val = 0;;) { 308 ch = *++s->str; 309 if (!isascii(ch) || !isdigit(ch)) 310 break; 311 val = val * 8 + ch - '0'; 312 if (++cnt == 3) { 313 ++s->str; 314 break; 315 } 316 } 317 if (cnt) 318 return (val); 319 if (ch != '\0') 320 ++s->str; 321 switch (ch) { 322 case 'a': /* escape characters */ 323 return ('\7'); 324 case 'b': 325 return ('\b'); 326 case 'f': 327 return ('\f'); 328 case 'n': 329 return ('\n'); 330 case 'r': 331 return ('\r'); 332 case 't': 333 return ('\t'); 334 case 'v': 335 return ('\13'); 336 case '\0': /* \" -> \ */ 337 s->state = EOS; 338 return ('\\'); 339 default: /* \x" -> x */ 340 return (ch); 341 } 342} 343