| 1 |
/* Copyright (c) 1990 Regents of the University of California */
|
| 2 |
|
| 3 |
#ifndef lint
|
| 4 |
static char SCCSid[] = "$SunId$ LBL";
|
| 5 |
#endif
|
| 6 |
|
| 7 |
#include <stdio.h>
|
| 8 |
#include <ctype.h>
|
| 9 |
/*
|
| 10 |
* rexpr.c - regular expression parser (ala grep)
|
| 11 |
*/
|
| 12 |
|
| 13 |
#define CCHR 2
|
| 14 |
#define CDOT 4
|
| 15 |
#define CCL 6
|
| 16 |
#define NCCL 8
|
| 17 |
#define CDOL 10
|
| 18 |
#define CEOF 11
|
| 19 |
#define CBRC 14
|
| 20 |
#define CLET 15
|
| 21 |
#define STAR 01
|
| 22 |
|
| 23 |
#define ESIZE 255
|
| 24 |
|
| 25 |
#define same(a,b) (a==b || (iflag && (a^b)==' ' && isalpha(a)))
|
| 26 |
|
| 27 |
#ifdef BSD
|
| 28 |
#define memcpy(to,from,len) bcopy(from,to,len)
|
| 29 |
#endif
|
| 30 |
|
| 31 |
static int advance(), cclass();
|
| 32 |
|
| 33 |
static char expbuf[ESIZE];
|
| 34 |
static int iflag;
|
| 35 |
static int circf;
|
| 36 |
|
| 37 |
int explen; /* length of the last expression found */
|
| 38 |
|
| 39 |
int
|
| 40 |
ecompile(sp, iflg, wflag) /* compile the expression */
|
| 41 |
register char *sp;
|
| 42 |
int iflg, wflag;
|
| 43 |
{
|
| 44 |
register c;
|
| 45 |
register char *ep;
|
| 46 |
char *lastep;
|
| 47 |
int cclcnt;
|
| 48 |
|
| 49 |
iflag = iflg;
|
| 50 |
ep = expbuf;
|
| 51 |
explen = 0;
|
| 52 |
if (*sp == '^') {
|
| 53 |
circf = 1;
|
| 54 |
sp++;
|
| 55 |
} else
|
| 56 |
circf = 0;
|
| 57 |
if (wflag)
|
| 58 |
*ep++ = CBRC;
|
| 59 |
for (;;) {
|
| 60 |
if (ep >= &expbuf[ESIZE])
|
| 61 |
return(-1);
|
| 62 |
if ((c = *sp++) != '*')
|
| 63 |
lastep = ep;
|
| 64 |
switch (c) {
|
| 65 |
|
| 66 |
case '\0':
|
| 67 |
if (wflag)
|
| 68 |
*ep++ = CLET;
|
| 69 |
*ep++ = CEOF;
|
| 70 |
explen = ep - expbuf;
|
| 71 |
return(0);
|
| 72 |
|
| 73 |
case '.':
|
| 74 |
*ep++ = CDOT;
|
| 75 |
continue;
|
| 76 |
|
| 77 |
case '*':
|
| 78 |
if (lastep==0)
|
| 79 |
goto defchar;
|
| 80 |
*lastep |= STAR;
|
| 81 |
continue;
|
| 82 |
|
| 83 |
case '$':
|
| 84 |
if (*sp != '\0')
|
| 85 |
goto defchar;
|
| 86 |
*ep++ = CDOL;
|
| 87 |
continue;
|
| 88 |
|
| 89 |
case '[':
|
| 90 |
*ep++ = CCL;
|
| 91 |
*ep++ = 0;
|
| 92 |
cclcnt = 1;
|
| 93 |
if ((c = *sp++) == '^') {
|
| 94 |
c = *sp++;
|
| 95 |
ep[-2] = NCCL;
|
| 96 |
}
|
| 97 |
do {
|
| 98 |
*ep++ = c;
|
| 99 |
cclcnt++;
|
| 100 |
if (c=='\0' || ep >= &expbuf[ESIZE])
|
| 101 |
return(-1);
|
| 102 |
} while ((c = *sp++) != ']');
|
| 103 |
lastep[1] = cclcnt;
|
| 104 |
continue;
|
| 105 |
|
| 106 |
case '\\':
|
| 107 |
if ((c = *sp++) == '\0')
|
| 108 |
return(-1);
|
| 109 |
if (c == '<') {
|
| 110 |
if (ep == expbuf || ep[-1] != CBRC)
|
| 111 |
*ep++ = CBRC;
|
| 112 |
continue;
|
| 113 |
}
|
| 114 |
if (c == '>') {
|
| 115 |
*ep++ = CLET;
|
| 116 |
continue;
|
| 117 |
}
|
| 118 |
defchar:
|
| 119 |
default:
|
| 120 |
*ep++ = CCHR;
|
| 121 |
*ep++ = c;
|
| 122 |
}
|
| 123 |
}
|
| 124 |
}
|
| 125 |
|
| 126 |
char *
|
| 127 |
expsave() /* save compiled string */
|
| 128 |
{
|
| 129 |
extern char *malloc();
|
| 130 |
register char *ep;
|
| 131 |
|
| 132 |
if (explen == 0)
|
| 133 |
return(NULL);
|
| 134 |
if ((ep = malloc(explen+3)) == NULL)
|
| 135 |
return(NULL);
|
| 136 |
ep[0] = iflag;
|
| 137 |
ep[1] = circf;
|
| 138 |
ep[2] = explen;
|
| 139 |
(void)memcpy(ep+3, expbuf, explen);
|
| 140 |
return(ep);
|
| 141 |
}
|
| 142 |
|
| 143 |
expset(ep) /* install saved string */
|
| 144 |
register char *ep;
|
| 145 |
{
|
| 146 |
iflag = ep[0];
|
| 147 |
circf = ep[1];
|
| 148 |
(void)memcpy(expbuf, ep+3, ep[2]&0xff);
|
| 149 |
}
|
| 150 |
|
| 151 |
char *
|
| 152 |
eindex(sp) /* find the expression in string sp */
|
| 153 |
register char *sp;
|
| 154 |
{
|
| 155 |
/* check for match at beginning of line, watch CBRC */
|
| 156 |
if (advance(sp, expbuf[0]==CBRC ? expbuf+1 : expbuf))
|
| 157 |
return(sp);
|
| 158 |
if (circf)
|
| 159 |
return(NULL);
|
| 160 |
/* fast check for first character */
|
| 161 |
if (expbuf[0]==CCHR) {
|
| 162 |
register c = expbuf[1];
|
| 163 |
while (*++sp)
|
| 164 |
if (same(*sp, c) && advance(sp, expbuf))
|
| 165 |
return(sp);
|
| 166 |
return(NULL);
|
| 167 |
}
|
| 168 |
/* regular algorithm */
|
| 169 |
while (*++sp)
|
| 170 |
if (advance(sp, expbuf))
|
| 171 |
return(sp);
|
| 172 |
return(NULL);
|
| 173 |
}
|
| 174 |
|
| 175 |
static int
|
| 176 |
advance(alp, ep)
|
| 177 |
char *alp;
|
| 178 |
register char *ep;
|
| 179 |
{
|
| 180 |
register char *lp;
|
| 181 |
char *curlp;
|
| 182 |
|
| 183 |
lp = alp;
|
| 184 |
for (;;) switch (*ep++) {
|
| 185 |
|
| 186 |
case CCHR:
|
| 187 |
if (!same(*ep, *lp))
|
| 188 |
return (0);
|
| 189 |
ep++, lp++;
|
| 190 |
continue;
|
| 191 |
|
| 192 |
case CDOT:
|
| 193 |
if (*lp++)
|
| 194 |
continue;
|
| 195 |
return(0);
|
| 196 |
|
| 197 |
case CDOL:
|
| 198 |
if (*lp==0)
|
| 199 |
continue;
|
| 200 |
return(0);
|
| 201 |
|
| 202 |
case CEOF:
|
| 203 |
explen = lp - alp;
|
| 204 |
return(1);
|
| 205 |
|
| 206 |
case CCL:
|
| 207 |
if (cclass(ep, *lp++, 1)) {
|
| 208 |
ep += *ep;
|
| 209 |
continue;
|
| 210 |
}
|
| 211 |
return(0);
|
| 212 |
|
| 213 |
case NCCL:
|
| 214 |
if (cclass(ep, *lp++, 0)) {
|
| 215 |
ep += *ep;
|
| 216 |
continue;
|
| 217 |
}
|
| 218 |
return(0);
|
| 219 |
|
| 220 |
case CDOT|STAR:
|
| 221 |
curlp = lp;
|
| 222 |
while (*lp++);
|
| 223 |
goto star;
|
| 224 |
|
| 225 |
case CCHR|STAR:
|
| 226 |
curlp = lp;
|
| 227 |
while (same(*lp, *ep))
|
| 228 |
lp++;
|
| 229 |
lp++;
|
| 230 |
ep++;
|
| 231 |
goto star;
|
| 232 |
|
| 233 |
case CCL|STAR:
|
| 234 |
case NCCL|STAR:
|
| 235 |
curlp = lp;
|
| 236 |
while (cclass(ep, *lp++, ep[-1]==(CCL|STAR)));
|
| 237 |
ep += *ep;
|
| 238 |
star:
|
| 239 |
do {
|
| 240 |
lp--;
|
| 241 |
if (advance(lp, ep)) {
|
| 242 |
explen += lp - alp;
|
| 243 |
return(1);
|
| 244 |
}
|
| 245 |
} while (lp > curlp);
|
| 246 |
return(0);
|
| 247 |
|
| 248 |
case CBRC:
|
| 249 |
if ((isalnum(*lp) || *lp == '_') && !(isalnum(lp[-1]) || lp[-1] == '_'))
|
| 250 |
continue;
|
| 251 |
return (0);
|
| 252 |
|
| 253 |
case CLET:
|
| 254 |
if (!isalnum(*lp) && *lp != '_')
|
| 255 |
continue;
|
| 256 |
return (0);
|
| 257 |
|
| 258 |
default:
|
| 259 |
fprintf(stderr, "RE botch\n");
|
| 260 |
}
|
| 261 |
}
|
| 262 |
|
| 263 |
static int
|
| 264 |
cclass(set, c, af)
|
| 265 |
register char *set;
|
| 266 |
register c;
|
| 267 |
{
|
| 268 |
register n;
|
| 269 |
|
| 270 |
if (c == 0)
|
| 271 |
return(0);
|
| 272 |
n = *set++;
|
| 273 |
while (--n)
|
| 274 |
if (n > 2 && set[1] == '-') {
|
| 275 |
if (c >= set[0] && c <= set[2])
|
| 276 |
return (af);
|
| 277 |
set += 3;
|
| 278 |
n -= 2;
|
| 279 |
} else
|
| 280 |
if (*set++ == c)
|
| 281 |
return(af);
|
| 282 |
return(!af);
|
| 283 |
}
|
