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 |
*ep++ = CBRC; |
111 |
continue; |
112 |
} |
113 |
if (c == '>') { |
114 |
*ep++ = CLET; |
115 |
continue; |
116 |
} |
117 |
defchar: |
118 |
default: |
119 |
*ep++ = CCHR; |
120 |
*ep++ = c; |
121 |
} |
122 |
} |
123 |
} |
124 |
|
125 |
char * |
126 |
expsave() /* save compiled string */ |
127 |
{ |
128 |
extern char *malloc(); |
129 |
register char *ep; |
130 |
|
131 |
if (explen == 0) |
132 |
return(NULL); |
133 |
if ((ep = malloc(explen+3)) == NULL) |
134 |
return(NULL); |
135 |
ep[0] = iflag; |
136 |
ep[1] = circf; |
137 |
ep[2] = explen; |
138 |
(void)memcpy(ep+3, expbuf, explen); |
139 |
return(ep); |
140 |
} |
141 |
|
142 |
expset(ep) /* install saved string */ |
143 |
register char *ep; |
144 |
{ |
145 |
iflag = ep[0]; |
146 |
circf = ep[1]; |
147 |
(void)memcpy(expbuf, ep+3, ep[2]&0xff); |
148 |
} |
149 |
|
150 |
char * |
151 |
eindex(sp) /* find the expression in str */ |
152 |
register char *sp; |
153 |
{ |
154 |
if (circf) { |
155 |
if (advance(sp, expbuf)) |
156 |
return(sp); |
157 |
return(NULL); |
158 |
} |
159 |
/* fast check for first character */ |
160 |
if (expbuf[0]==CCHR) { |
161 |
register c = expbuf[1]; |
162 |
do { |
163 |
if (same(*sp, c) && advance(sp, expbuf)) |
164 |
return(sp); |
165 |
} while (*sp++); |
166 |
return(NULL); |
167 |
} |
168 |
/* regular algorithm */ |
169 |
do { |
170 |
if (advance(sp, expbuf)) |
171 |
return(sp); |
172 |
} while (*sp++); |
173 |
return(NULL); |
174 |
} |
175 |
|
176 |
static int |
177 |
advance(alp, ep) |
178 |
char *alp; |
179 |
register char *ep; |
180 |
{ |
181 |
register char *lp; |
182 |
char *curlp; |
183 |
|
184 |
lp = alp; |
185 |
for (;;) switch (*ep++) { |
186 |
|
187 |
case CCHR: |
188 |
if (!same(*ep, *lp)) |
189 |
return (0); |
190 |
ep++, lp++; |
191 |
continue; |
192 |
|
193 |
case CDOT: |
194 |
if (*lp++) |
195 |
continue; |
196 |
return(0); |
197 |
|
198 |
case CDOL: |
199 |
if (*lp==0) |
200 |
continue; |
201 |
return(0); |
202 |
|
203 |
case CEOF: |
204 |
explen = lp - alp; |
205 |
return(1); |
206 |
|
207 |
case CCL: |
208 |
if (cclass(ep, *lp++, 1)) { |
209 |
ep += *ep; |
210 |
continue; |
211 |
} |
212 |
return(0); |
213 |
|
214 |
case NCCL: |
215 |
if (cclass(ep, *lp++, 0)) { |
216 |
ep += *ep; |
217 |
continue; |
218 |
} |
219 |
return(0); |
220 |
|
221 |
case CDOT|STAR: |
222 |
curlp = lp; |
223 |
while (*lp++); |
224 |
goto star; |
225 |
|
226 |
case CCHR|STAR: |
227 |
curlp = lp; |
228 |
while (same(*lp, *ep)) |
229 |
lp++; |
230 |
lp++; |
231 |
ep++; |
232 |
goto star; |
233 |
|
234 |
case CCL|STAR: |
235 |
case NCCL|STAR: |
236 |
curlp = lp; |
237 |
while (cclass(ep, *lp++, ep[-1]==(CCL|STAR))); |
238 |
ep += *ep; |
239 |
star: |
240 |
do { |
241 |
lp--; |
242 |
if (advance(lp, ep)) { |
243 |
explen += lp - alp; |
244 |
return(1); |
245 |
} |
246 |
} while (lp > curlp); |
247 |
return(0); |
248 |
|
249 |
case CBRC: |
250 |
if (lp == alp) |
251 |
continue; |
252 |
if ((isalnum(*lp) || *lp == '_') && !(isalnum(lp[-1]) || lp[-1] == '_')) |
253 |
continue; |
254 |
return (0); |
255 |
|
256 |
case CLET: |
257 |
if (!isalnum(*lp) && *lp != '_') |
258 |
continue; |
259 |
return (0); |
260 |
|
261 |
default: |
262 |
fprintf(stderr, "RE botch\n"); |
263 |
} |
264 |
} |
265 |
|
266 |
static int |
267 |
cclass(set, c, af) |
268 |
register char *set; |
269 |
register c; |
270 |
{ |
271 |
register n; |
272 |
|
273 |
if (c == 0) |
274 |
return(0); |
275 |
n = *set++; |
276 |
while (--n) |
277 |
if (n > 2 && set[1] == '-') { |
278 |
if (c >= set[0] && c <= set[2]) |
279 |
return (af); |
280 |
set += 3; |
281 |
n -= 2; |
282 |
} else |
283 |
if (*set++ == c) |
284 |
return(af); |
285 |
return(!af); |
286 |
} |