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 char expbuf[ESIZE]; |
32 |
static int iflag; |
33 |
static int circf; |
34 |
|
35 |
int explen; /* length of the last expression found */ |
36 |
|
37 |
int |
38 |
ecompile(sp, iflg, wflag) /* compile the expression */ |
39 |
register char *sp; |
40 |
int iflg, wflag; |
41 |
{ |
42 |
register c; |
43 |
register char *ep; |
44 |
char *lastep; |
45 |
int cclcnt; |
46 |
|
47 |
iflag = iflg; |
48 |
ep = expbuf; |
49 |
explen = 0; |
50 |
if (*sp == '^') { |
51 |
circf = 1; |
52 |
sp++; |
53 |
} else |
54 |
circf = 0; |
55 |
if (wflag) |
56 |
*ep++ = CBRC; |
57 |
for (;;) { |
58 |
if (ep >= &expbuf[ESIZE]) |
59 |
return(-1); |
60 |
if ((c = *sp++) != '*') |
61 |
lastep = ep; |
62 |
switch (c) { |
63 |
|
64 |
case '\0': |
65 |
if (wflag) |
66 |
*ep++ = CLET; |
67 |
*ep++ = CEOF; |
68 |
explen = ep - expbuf; |
69 |
return(0); |
70 |
|
71 |
case '.': |
72 |
*ep++ = CDOT; |
73 |
continue; |
74 |
|
75 |
case '*': |
76 |
if (lastep==0) |
77 |
goto defchar; |
78 |
*lastep |= STAR; |
79 |
continue; |
80 |
|
81 |
case '$': |
82 |
if (*sp != '\0') |
83 |
goto defchar; |
84 |
*ep++ = CDOL; |
85 |
continue; |
86 |
|
87 |
case '[': |
88 |
*ep++ = CCL; |
89 |
*ep++ = 0; |
90 |
cclcnt = 1; |
91 |
if ((c = *sp++) == '^') { |
92 |
c = *sp++; |
93 |
ep[-2] = NCCL; |
94 |
} |
95 |
do { |
96 |
*ep++ = c; |
97 |
cclcnt++; |
98 |
if (c=='\0' || ep >= &expbuf[ESIZE]) |
99 |
return(-1); |
100 |
} while ((c = *sp++) != ']'); |
101 |
lastep[1] = cclcnt; |
102 |
continue; |
103 |
|
104 |
case '\\': |
105 |
if ((c = *sp++) == '\0') |
106 |
return(-1); |
107 |
if (c == '<') { |
108 |
*ep++ = CBRC; |
109 |
continue; |
110 |
} |
111 |
if (c == '>') { |
112 |
*ep++ = CLET; |
113 |
continue; |
114 |
} |
115 |
defchar: |
116 |
default: |
117 |
*ep++ = CCHR; |
118 |
*ep++ = c; |
119 |
} |
120 |
} |
121 |
} |
122 |
|
123 |
char * |
124 |
expsave() /* save compiled string */ |
125 |
{ |
126 |
extern char *malloc(); |
127 |
register char *ep; |
128 |
|
129 |
if (explen == 0) |
130 |
return(NULL); |
131 |
if ((ep = malloc(explen+3)) == NULL) |
132 |
return(NULL); |
133 |
ep[0] = iflag; |
134 |
ep[1] = circf; |
135 |
ep[2] = explen; |
136 |
(void)memcpy(ep+3, expbuf, explen); |
137 |
return(ep); |
138 |
} |
139 |
|
140 |
expset(ep) /* install saved string */ |
141 |
register char *ep; |
142 |
{ |
143 |
iflag = ep[0]; |
144 |
circf = ep[1]; |
145 |
(void)memcpy(expbuf, ep+3, ep[2]&0xff); |
146 |
} |
147 |
|
148 |
char * |
149 |
eindex(sp) /* find the expression in str */ |
150 |
register char *sp; |
151 |
{ |
152 |
if (circf) { |
153 |
if (advance(sp, expbuf)) |
154 |
return(sp); |
155 |
return(NULL); |
156 |
} |
157 |
/* fast check for first character */ |
158 |
if (expbuf[0]==CCHR) { |
159 |
register c = expbuf[1]; |
160 |
do { |
161 |
if (same(*sp, c) && advance(sp, expbuf)) |
162 |
return(sp); |
163 |
} while (*sp++); |
164 |
return(NULL); |
165 |
} |
166 |
/* regular algorithm */ |
167 |
do { |
168 |
if (advance(sp, expbuf)) |
169 |
return(sp); |
170 |
} while (*sp++); |
171 |
return(NULL); |
172 |
} |
173 |
|
174 |
static int |
175 |
advance(alp, ep) |
176 |
char *alp; |
177 |
register char *ep; |
178 |
{ |
179 |
register char *lp; |
180 |
char *curlp; |
181 |
|
182 |
lp = alp; |
183 |
for (;;) switch (*ep++) { |
184 |
|
185 |
case CCHR: |
186 |
if (!same(*ep, *lp)) |
187 |
return (0); |
188 |
ep++, lp++; |
189 |
continue; |
190 |
|
191 |
case CDOT: |
192 |
if (*lp++) |
193 |
continue; |
194 |
return(0); |
195 |
|
196 |
case CDOL: |
197 |
if (*lp==0) |
198 |
continue; |
199 |
return(0); |
200 |
|
201 |
case CEOF: |
202 |
explen = lp - alp; |
203 |
return(1); |
204 |
|
205 |
case CCL: |
206 |
if (cclass(ep, *lp++, 1)) { |
207 |
ep += *ep; |
208 |
continue; |
209 |
} |
210 |
return(0); |
211 |
|
212 |
case NCCL: |
213 |
if (cclass(ep, *lp++, 0)) { |
214 |
ep += *ep; |
215 |
continue; |
216 |
} |
217 |
return(0); |
218 |
|
219 |
case CDOT|STAR: |
220 |
curlp = lp; |
221 |
while (*lp++); |
222 |
goto star; |
223 |
|
224 |
case CCHR|STAR: |
225 |
curlp = lp; |
226 |
while (same(*lp, *ep)) |
227 |
lp++; |
228 |
lp++; |
229 |
ep++; |
230 |
goto star; |
231 |
|
232 |
case CCL|STAR: |
233 |
case NCCL|STAR: |
234 |
curlp = lp; |
235 |
while (cclass(ep, *lp++, ep[-1]==(CCL|STAR))); |
236 |
ep += *ep; |
237 |
star: |
238 |
do { |
239 |
lp--; |
240 |
if (advance(lp, ep)) { |
241 |
explen += lp - alp; |
242 |
return(1); |
243 |
} |
244 |
} while (lp > curlp); |
245 |
return(0); |
246 |
|
247 |
case CBRC: |
248 |
if (lp == expbuf) |
249 |
continue; |
250 |
if ((isalnum(*lp) || *lp == '_') && !(isalnum(lp[-1]) || lp[-1] == '_')) |
251 |
continue; |
252 |
return (0); |
253 |
|
254 |
case CLET: |
255 |
if (!isalnum(*lp) && *lp != '_') |
256 |
continue; |
257 |
return (0); |
258 |
|
259 |
default: |
260 |
fprintf(stderr, "RE botch\n"); |
261 |
} |
262 |
} |
263 |
|
264 |
static int |
265 |
cclass(set, c, af) |
266 |
register char *set; |
267 |
register c; |
268 |
{ |
269 |
register n; |
270 |
|
271 |
if (c == 0) |
272 |
return(0); |
273 |
n = *set++; |
274 |
while (--n) |
275 |
if (n > 2 && set[1] == '-') { |
276 |
if (c >= set[0] && c <= set[2]) |
277 |
return (af); |
278 |
set += 3; |
279 |
n -= 2; |
280 |
} else |
281 |
if (*set++ == c) |
282 |
return(af); |
283 |
return(!af); |
284 |
} |