1
by brian
clean slate |
1 |
/*
|
2 |
* The matching engine and friends. This file is #included by regexec.c
|
|
3 |
* after suitable #defines of a variety of macros used herein, so that
|
|
4 |
* different state representations can be used without duplicating masses
|
|
5 |
* of code.
|
|
6 |
*/
|
|
7 |
||
8 |
#ifdef SNAMES
|
|
9 |
#define matcher smatcher
|
|
10 |
#define fast sfast
|
|
11 |
#define slow sslow
|
|
12 |
#define dissect sdissect
|
|
13 |
#define backref sbackref
|
|
14 |
#define step sstep
|
|
15 |
#define print sprint
|
|
16 |
#define at sat
|
|
17 |
#define match smat
|
|
18 |
#endif
|
|
19 |
#ifdef LNAMES
|
|
20 |
#define matcher lmatcher
|
|
21 |
#define fast lfast
|
|
22 |
#define slow lslow
|
|
23 |
#define dissect ldissect
|
|
24 |
#define backref lbackref
|
|
25 |
#define step lstep
|
|
26 |
#define print lprint
|
|
27 |
#define at lat
|
|
28 |
#define match lmat
|
|
29 |
#endif
|
|
30 |
||
31 |
/* another structure passed up and down to avoid zillions of parameters */
|
|
32 |
struct match { |
|
33 |
struct re_guts *g; |
|
34 |
int eflags; |
|
35 |
my_regmatch_t *pmatch; /* [nsub+1] (0 element unused) */ |
|
36 |
char *offp; /* offsets work from here */ |
|
37 |
char *beginp; /* start of string -- virtual NUL precedes */ |
|
38 |
char *endp; /* end of string -- virtual NUL here */ |
|
39 |
char *coldp; /* can be no match starting before here */ |
|
40 |
char **lastpos; /* [nplus+1] */ |
|
41 |
STATEVARS; |
|
42 |
states st; /* current states */ |
|
43 |
states fresh; /* states for a fresh start */ |
|
44 |
states tmp; /* temporary */ |
|
45 |
states empty; /* empty set of states */ |
|
46 |
};
|
|
47 |
||
48 |
#include "engine.ih" |
|
49 |
||
50 |
#ifdef REDEBUG
|
|
51 |
#define SP(t, s, c) print(m, t, s, c, stdout)
|
|
52 |
#define AT(t, p1, p2, s1, s2) at(m, t, p1, p2, s1, s2)
|
|
53 |
#define NOTE(str) { if (m->eflags®_TRACE) printf("=%s\n", (str)); }
|
|
54 |
#else
|
|
55 |
#define SP(t, s, c) /* nothing */ |
|
56 |
#define AT(t, p1, p2, s1, s2) /* nothing */ |
|
57 |
#define NOTE(s) /* nothing */ |
|
58 |
#endif
|
|
59 |
||
60 |
/*
|
|
61 |
- matcher - the actual matching engine
|
|
62 |
== static int matcher(register struct re_guts *g, char *string, \
|
|
63 |
== size_t nmatch, regmatch_t pmatch[], int eflags);
|
|
64 |
*/
|
|
65 |
static int /* 0 success, REG_NOMATCH failure */ |
|
66 |
matcher(charset,g, str, nmatch, pmatch, eflags) |
|
67 |
CHARSET_INFO *charset; |
|
68 |
register struct re_guts *g; |
|
69 |
char *str; |
|
70 |
size_t nmatch; |
|
71 |
my_regmatch_t pmatch[]; |
|
72 |
int eflags; |
|
73 |
{
|
|
74 |
register char *endp; |
|
75 |
register uint i; |
|
76 |
struct match mv; |
|
77 |
register struct match *m = &mv; |
|
78 |
register char *dp; |
|
79 |
register const sopno gf = g->firststate+1; /* +1 for OEND */ |
|
80 |
register const sopno gl = g->laststate; |
|
81 |
char *start; |
|
82 |
char *stop; |
|
83 |
||
84 |
/* simplify the situation where possible */
|
|
85 |
if (g->cflags®_NOSUB) |
|
86 |
nmatch = 0; |
|
87 |
if (eflags®_STARTEND) { |
|
88 |
start = str + pmatch[0].rm_so; |
|
89 |
stop = str + pmatch[0].rm_eo; |
|
90 |
} else { |
|
91 |
start = str; |
|
92 |
stop = start + strlen(start); |
|
93 |
}
|
|
94 |
if (stop < start) |
|
95 |
return(REG_INVARG); |
|
96 |
||
97 |
/* prescreening; this does wonders for this rather slow code */
|
|
98 |
if (g->must != NULL) { |
|
99 |
for (dp = start; dp < stop; dp++) |
|
100 |
if (*dp == g->must[0] && stop - dp >= g->mlen && |
|
101 |
memcmp(dp, g->must, (size_t)g->mlen) == 0) |
|
102 |
break; |
|
103 |
if (dp == stop) /* we didn't find g->must */ |
|
104 |
return(REG_NOMATCH); |
|
105 |
}
|
|
106 |
||
107 |
/* match struct setup */
|
|
108 |
m->g = g; |
|
109 |
m->eflags = eflags; |
|
110 |
m->pmatch = NULL; |
|
111 |
m->lastpos = NULL; |
|
112 |
m->offp = str; |
|
113 |
m->beginp = start; |
|
114 |
m->endp = stop; |
|
115 |
STATESETUP(m, 4); |
|
116 |
SETUP(m->st); |
|
117 |
SETUP(m->fresh); |
|
118 |
SETUP(m->tmp); |
|
119 |
SETUP(m->empty); |
|
120 |
CLEAR(m->empty); |
|
121 |
||
122 |
/* this loop does only one repetition except for backrefs */
|
|
123 |
for (;;) { |
|
124 |
endp = fast(charset, m, start, stop, gf, gl); |
|
125 |
if (endp == NULL) { /* a miss */ |
|
126 |
if (m->pmatch != NULL) |
|
127 |
free((char *)m->pmatch); |
|
128 |
if (m->lastpos != NULL) |
|
129 |
free((char *)m->lastpos); |
|
130 |
STATETEARDOWN(m); |
|
131 |
return(REG_NOMATCH); |
|
132 |
}
|
|
133 |
if (nmatch == 0 && !g->backrefs) |
|
134 |
break; /* no further info needed */ |
|
135 |
||
136 |
/* where? */
|
|
137 |
assert(m->coldp != NULL); |
|
138 |
for (;;) { |
|
139 |
NOTE("finding start"); |
|
140 |
endp = slow(charset, m, m->coldp, stop, gf, gl); |
|
141 |
if (endp != NULL) |
|
142 |
break; |
|
143 |
assert(m->coldp < m->endp); |
|
144 |
m->coldp++; |
|
145 |
}
|
|
146 |
if (nmatch == 1 && !g->backrefs) |
|
147 |
break; /* no further info needed */ |
|
148 |
||
149 |
/* oh my, he wants the subexpressions... */
|
|
150 |
if (m->pmatch == NULL) |
|
151 |
m->pmatch = (my_regmatch_t *)malloc((m->g->nsub + 1) * |
|
152 |
sizeof(my_regmatch_t)); |
|
153 |
if (m->pmatch == NULL) { |
|
154 |
if (m->lastpos != NULL) |
|
155 |
free((char *)m->lastpos); |
|
156 |
STATETEARDOWN(m); |
|
157 |
return(REG_ESPACE); |
|
158 |
}
|
|
159 |
for (i = 1; i <= m->g->nsub; i++) |
|
160 |
m->pmatch[i].rm_so = m->pmatch[i].rm_eo = -1; |
|
161 |
if (!g->backrefs && !(m->eflags®_BACKR)) { |
|
162 |
NOTE("dissecting"); |
|
163 |
dp = dissect(charset, m, m->coldp, endp, gf, gl); |
|
164 |
} else { |
|
165 |
if (g->nplus > 0 && m->lastpos == NULL) |
|
166 |
m->lastpos = (char **)malloc((g->nplus+1) * |
|
167 |
sizeof(char *)); |
|
168 |
if (g->nplus > 0 && m->lastpos == NULL) { |
|
169 |
free(m->pmatch); |
|
170 |
STATETEARDOWN(m); |
|
171 |
return(REG_ESPACE); |
|
172 |
}
|
|
173 |
NOTE("backref dissect"); |
|
174 |
dp = backref(charset, m, m->coldp, endp, gf, gl, (sopno)0); |
|
175 |
}
|
|
176 |
if (dp != NULL) |
|
177 |
break; |
|
178 |
||
179 |
/* uh-oh... we couldn't find a subexpression-level match */
|
|
180 |
assert(g->backrefs); /* must be back references doing it */ |
|
181 |
assert(g->nplus == 0 || m->lastpos != NULL); |
|
182 |
for (;;) { |
|
183 |
if (dp != NULL || endp <= m->coldp) |
|
184 |
break; /* defeat */ |
|
185 |
NOTE("backoff"); |
|
186 |
endp = slow(charset, m, m->coldp, endp-1, gf, gl); |
|
187 |
if (endp == NULL) |
|
188 |
break; /* defeat */ |
|
189 |
/* try it on a shorter possibility */
|
|
190 |
#ifndef NDEBUG
|
|
191 |
for (i = 1; i <= m->g->nsub; i++) { |
|
192 |
assert(m->pmatch[i].rm_so == -1); |
|
193 |
assert(m->pmatch[i].rm_eo == -1); |
|
194 |
}
|
|
195 |
#endif
|
|
196 |
NOTE("backoff dissect"); |
|
197 |
dp = backref(charset, m, m->coldp, endp, gf, gl, (sopno)0); |
|
198 |
}
|
|
199 |
assert(dp == NULL || dp == endp); |
|
200 |
if (dp != NULL) /* found a shorter one */ |
|
201 |
break; |
|
202 |
||
203 |
/* despite initial appearances, there is no match here */
|
|
204 |
NOTE("false alarm"); |
|
205 |
start = m->coldp + 1; /* recycle starting later */ |
|
206 |
assert(start <= stop); |
|
207 |
}
|
|
208 |
||
209 |
/* fill in the details if requested */
|
|
210 |
if (nmatch > 0) { |
|
211 |
pmatch[0].rm_so = m->coldp - m->offp; |
|
212 |
pmatch[0].rm_eo = endp - m->offp; |
|
213 |
}
|
|
214 |
if (nmatch > 1) { |
|
215 |
assert(m->pmatch != NULL); |
|
216 |
for (i = 1; i < nmatch; i++) |
|
217 |
if (i <= m->g->nsub) |
|
218 |
pmatch[i] = m->pmatch[i]; |
|
219 |
else { |
|
220 |
pmatch[i].rm_so = -1; |
|
221 |
pmatch[i].rm_eo = -1; |
|
222 |
}
|
|
223 |
}
|
|
224 |
||
225 |
if (m->pmatch != NULL) |
|
226 |
free((char *)m->pmatch); |
|
227 |
if (m->lastpos != NULL) |
|
228 |
free((char *)m->lastpos); |
|
229 |
STATETEARDOWN(m); |
|
230 |
return(0); |
|
231 |
}
|
|
232 |
||
233 |
/*
|
|
234 |
- dissect - figure out what matched what, no back references
|
|
235 |
== static char *dissect(register struct match *m, char *start, \
|
|
236 |
== char *stop, sopno startst, sopno stopst);
|
|
237 |
*/
|
|
238 |
static char * /* == stop (success) always */ |
|
239 |
dissect(charset, m, start, stop, startst, stopst) |
|
240 |
CHARSET_INFO *charset; |
|
241 |
register struct match *m; |
|
242 |
char *start; |
|
243 |
char *stop; |
|
244 |
sopno startst; |
|
245 |
sopno stopst; |
|
246 |
{
|
|
247 |
register uint i; |
|
248 |
register sopno ss; /* start sop of current subRE */ |
|
249 |
register sopno es; /* end sop of current subRE */ |
|
250 |
register char *sp; /* start of string matched by it */ |
|
251 |
register char *stp; /* string matched by it cannot pass here */ |
|
252 |
register char *rest; /* start of rest of string */ |
|
253 |
register char *tail; /* string unmatched by rest of RE */ |
|
254 |
register sopno ssub; /* start sop of subsubRE */ |
|
255 |
register sopno esub; /* end sop of subsubRE */ |
|
256 |
register char *ssp; /* start of string matched by subsubRE */ |
|
257 |
register char *sep; /* end of string matched by subsubRE */ |
|
258 |
register char *oldssp; /* previous ssp */ |
|
259 |
register char *dp; /* used in debug mode to check asserts */ |
|
260 |
||
261 |
AT("diss", start, stop, startst, stopst); |
|
262 |
sp = start; |
|
263 |
for (ss = startst; ss < stopst; ss = es) { |
|
264 |
/* identify end of subRE */
|
|
265 |
es = ss; |
|
266 |
switch (OP(m->g->strip[es])) { |
|
267 |
case OPLUS_: |
|
268 |
case OQUEST_: |
|
269 |
es += OPND(m->g->strip[es]); |
|
270 |
break; |
|
271 |
case OCH_: |
|
272 |
while (OP(m->g->strip[es]) != O_CH) |
|
273 |
es += OPND(m->g->strip[es]); |
|
274 |
break; |
|
275 |
}
|
|
276 |
es++; |
|
277 |
||
278 |
/* figure out what it matched */
|
|
279 |
switch (OP(m->g->strip[ss])) { |
|
280 |
case OEND: |
|
281 |
assert(nope); |
|
282 |
break; |
|
283 |
case OCHAR: |
|
284 |
sp++; |
|
285 |
break; |
|
286 |
case OBOL: |
|
287 |
case OEOL: |
|
288 |
case OBOW: |
|
289 |
case OEOW: |
|
290 |
break; |
|
291 |
case OANY: |
|
292 |
case OANYOF: |
|
293 |
sp++; |
|
294 |
break; |
|
295 |
case OBACK_: |
|
296 |
case O_BACK: |
|
297 |
assert(nope); |
|
298 |
break; |
|
299 |
/* cases where length of match is hard to find */
|
|
300 |
case OQUEST_: |
|
301 |
stp = stop; |
|
302 |
for (;;) { |
|
303 |
/* how long could this one be? */
|
|
304 |
rest = slow(charset, m, sp, stp, ss, es); |
|
305 |
assert(rest != NULL); /* it did match */ |
|
306 |
/* could the rest match the rest? */
|
|
307 |
tail = slow(charset, m, rest, stop, es, stopst); |
|
308 |
if (tail == stop) |
|
309 |
break; /* yes! */ |
|
310 |
/* no -- try a shorter match for this one */
|
|
311 |
stp = rest - 1; |
|
312 |
assert(stp >= sp); /* it did work */ |
|
313 |
}
|
|
314 |
ssub = ss + 1; |
|
315 |
esub = es - 1; |
|
316 |
/* did innards match? */
|
|
317 |
if (slow(charset, m, sp, rest, ssub, esub) != NULL) { |
|
318 |
dp = dissect(charset, m, sp, rest, ssub, esub); |
|
319 |
assert(dp == rest); |
|
320 |
} else /* no */ |
|
321 |
assert(sp == rest); |
|
322 |
sp = rest; |
|
323 |
break; |
|
324 |
case OPLUS_: |
|
325 |
stp = stop; |
|
326 |
for (;;) { |
|
327 |
/* how long could this one be? */
|
|
328 |
rest = slow(charset, m, sp, stp, ss, es); |
|
329 |
assert(rest != NULL); /* it did match */ |
|
330 |
/* could the rest match the rest? */
|
|
331 |
tail = slow(charset, m, rest, stop, es, stopst); |
|
332 |
if (tail == stop) |
|
333 |
break; /* yes! */ |
|
334 |
/* no -- try a shorter match for this one */
|
|
335 |
stp = rest - 1; |
|
336 |
assert(stp >= sp); /* it did work */ |
|
337 |
}
|
|
338 |
ssub = ss + 1; |
|
339 |
esub = es - 1; |
|
340 |
ssp = sp; |
|
341 |
oldssp = ssp; |
|
342 |
for (;;) { /* find last match of innards */ |
|
343 |
sep = slow(charset, m, ssp, rest, ssub, esub); |
|
344 |
if (sep == NULL || sep == ssp) |
|
345 |
break; /* failed or matched null */ |
|
346 |
oldssp = ssp; /* on to next try */ |
|
347 |
ssp = sep; |
|
348 |
}
|
|
349 |
if (sep == NULL) { |
|
350 |
/* last successful match */
|
|
351 |
sep = ssp; |
|
352 |
ssp = oldssp; |
|
353 |
}
|
|
354 |
assert(sep == rest); /* must exhaust substring */ |
|
355 |
assert(slow(charset, m, ssp, sep, ssub, esub) == rest); |
|
356 |
dp = dissect(charset, m, ssp, sep, ssub, esub); |
|
357 |
assert(dp == sep); |
|
358 |
sp = rest; |
|
359 |
break; |
|
360 |
case OCH_: |
|
361 |
stp = stop; |
|
362 |
for (;;) { |
|
363 |
/* how long could this one be? */
|
|
364 |
rest = slow(charset, m, sp, stp, ss, es); |
|
365 |
assert(rest != NULL); /* it did match */ |
|
366 |
/* could the rest match the rest? */
|
|
367 |
tail = slow(charset, m, rest, stop, es, stopst); |
|
368 |
if (tail == stop) |
|
369 |
break; /* yes! */ |
|
370 |
/* no -- try a shorter match for this one */
|
|
371 |
stp = rest - 1; |
|
372 |
assert(stp >= sp); /* it did work */ |
|
373 |
}
|
|
374 |
ssub = ss + 1; |
|
375 |
esub = ss + OPND(m->g->strip[ss]) - 1; |
|
376 |
assert(OP(m->g->strip[esub]) == OOR1); |
|
377 |
for (;;) { /* find first matching branch */ |
|
378 |
if (slow(charset, m, sp, rest, ssub, esub) == rest) |
|
379 |
break; /* it matched all of it */ |
|
380 |
/* that one missed, try next one */
|
|
381 |
assert(OP(m->g->strip[esub]) == OOR1); |
|
382 |
esub++; |
|
383 |
assert(OP(m->g->strip[esub]) == OOR2); |
|
384 |
ssub = esub + 1; |
|
385 |
esub += OPND(m->g->strip[esub]); |
|
386 |
if (OP(m->g->strip[esub]) == OOR2) |
|
387 |
esub--; |
|
388 |
else
|
|
389 |
assert(OP(m->g->strip[esub]) == O_CH); |
|
390 |
}
|
|
391 |
dp = dissect(charset, m, sp, rest, ssub, esub); |
|
392 |
assert(dp == rest); |
|
393 |
sp = rest; |
|
394 |
break; |
|
395 |
case O_PLUS: |
|
396 |
case O_QUEST: |
|
397 |
case OOR1: |
|
398 |
case OOR2: |
|
399 |
case O_CH: |
|
400 |
assert(nope); |
|
401 |
break; |
|
402 |
case OLPAREN: |
|
403 |
i = OPND(m->g->strip[ss]); |
|
404 |
assert(0 < i && i <= m->g->nsub); |
|
405 |
m->pmatch[i].rm_so = sp - m->offp; |
|
406 |
break; |
|
407 |
case ORPAREN: |
|
408 |
i = OPND(m->g->strip[ss]); |
|
409 |
assert(0 < i && i <= m->g->nsub); |
|
410 |
m->pmatch[i].rm_eo = sp - m->offp; |
|
411 |
break; |
|
412 |
default: /* uh oh */ |
|
413 |
assert(nope); |
|
414 |
break; |
|
415 |
}
|
|
416 |
}
|
|
417 |
||
418 |
assert(sp == stop); |
|
419 |
return(sp); |
|
420 |
}
|
|
421 |
||
422 |
/*
|
|
423 |
- backref - figure out what matched what, figuring in back references
|
|
424 |
== static char *backref(register struct match *m, char *start, \
|
|
425 |
== char *stop, sopno startst, sopno stopst, sopno lev);
|
|
426 |
*/
|
|
427 |
static char * /* == stop (success) or NULL (failure) */ |
|
428 |
backref(charset,m, start, stop, startst, stopst, lev) |
|
429 |
CHARSET_INFO *charset; |
|
430 |
register struct match *m; |
|
431 |
char *start; |
|
432 |
char *stop; |
|
433 |
sopno startst; |
|
434 |
sopno stopst; |
|
435 |
sopno lev; /* PLUS nesting level */ |
|
436 |
{
|
|
437 |
register uint i; |
|
438 |
register sopno ss; /* start sop of current subRE */ |
|
439 |
register char *sp; /* start of string matched by it */ |
|
440 |
register sopno ssub; /* start sop of subsubRE */ |
|
441 |
register sopno esub; /* end sop of subsubRE */ |
|
442 |
register char *ssp; /* start of string matched by subsubRE */ |
|
443 |
register char *dp; |
|
444 |
register size_t len; |
|
445 |
register int hard; |
|
446 |
register sop s; |
|
447 |
register regoff_t offsave; |
|
448 |
register cset *cs; |
|
449 |
||
450 |
AT("back", start, stop, startst, stopst); |
|
451 |
sp = start; |
|
452 |
||
453 |
/* get as far as we can with easy stuff */
|
|
454 |
hard = 0; |
|
455 |
for (ss = startst; !hard && ss < stopst; ss++) |
|
456 |
switch (OP(s = m->g->strip[ss])) { |
|
457 |
case OCHAR: |
|
458 |
if (sp == stop || *sp++ != (char)OPND(s)) |
|
459 |
return(NULL); |
|
460 |
break; |
|
461 |
case OANY: |
|
462 |
if (sp == stop) |
|
463 |
return(NULL); |
|
464 |
sp++; |
|
465 |
break; |
|
466 |
case OANYOF: |
|
467 |
cs = &m->g->sets[OPND(s)]; |
|
468 |
if (sp == stop || !CHIN(cs, *sp++)) |
|
469 |
return(NULL); |
|
470 |
break; |
|
471 |
case OBOL: |
|
472 |
if ( (sp == m->beginp && !(m->eflags®_NOTBOL)) || |
|
473 |
(sp < m->endp && *(sp-1) == '\n' && |
|
474 |
(m->g->cflags®_NEWLINE)) ) |
|
475 |
{ /* yes */ } |
|
476 |
else
|
|
477 |
return(NULL); |
|
478 |
break; |
|
479 |
case OEOL: |
|
480 |
if ( (sp == m->endp && !(m->eflags®_NOTEOL)) || |
|
481 |
(sp < m->endp && *sp == '\n' && |
|
482 |
(m->g->cflags®_NEWLINE)) ) |
|
483 |
{ /* yes */ } |
|
484 |
else
|
|
485 |
return(NULL); |
|
486 |
break; |
|
487 |
case OBOW: |
|
488 |
if (( (sp == m->beginp && !(m->eflags®_NOTBOL)) || |
|
489 |
(sp < m->endp && *(sp-1) == '\n' && |
|
490 |
(m->g->cflags®_NEWLINE)) || |
|
491 |
(sp > m->beginp && |
|
492 |
!ISWORD(charset,*(sp-1))) ) && |
|
493 |
(sp < m->endp && ISWORD(charset,*sp)) ) |
|
494 |
{ /* yes */ } |
|
495 |
else
|
|
496 |
return(NULL); |
|
497 |
break; |
|
498 |
case OEOW: |
|
499 |
if (( (sp == m->endp && !(m->eflags®_NOTEOL)) || |
|
500 |
(sp < m->endp && *sp == '\n' && |
|
501 |
(m->g->cflags®_NEWLINE)) || |
|
502 |
(sp < m->endp && !ISWORD(charset,*sp)) ) && |
|
503 |
(sp > m->beginp && ISWORD(charset,*(sp-1))) ) |
|
504 |
{ /* yes */ } |
|
505 |
else
|
|
506 |
return(NULL); |
|
507 |
break; |
|
508 |
case O_QUEST: |
|
509 |
break; |
|
510 |
case OOR1: /* matches null but needs to skip */ |
|
511 |
ss++; |
|
512 |
s = m->g->strip[ss]; |
|
513 |
do { |
|
514 |
assert(OP(s) == OOR2); |
|
515 |
ss += OPND(s); |
|
516 |
} while (OP(s = m->g->strip[ss]) != O_CH); |
|
517 |
/* note that the ss++ gets us past the O_CH */
|
|
518 |
break; |
|
519 |
default: /* have to make a choice */ |
|
520 |
hard = 1; |
|
521 |
break; |
|
522 |
}
|
|
523 |
if (!hard) { /* that was it! */ |
|
524 |
if (sp != stop) |
|
525 |
return(NULL); |
|
526 |
return(sp); |
|
527 |
}
|
|
528 |
ss--; /* adjust for the for's final increment */ |
|
529 |
||
530 |
/* the hard stuff */
|
|
531 |
AT("hard", sp, stop, ss, stopst); |
|
532 |
s = m->g->strip[ss]; |
|
533 |
switch (OP(s)) { |
|
534 |
case OBACK_: /* the vilest depths */ |
|
535 |
i = OPND(s); |
|
536 |
assert(0 < i && i <= m->g->nsub); |
|
537 |
if (m->pmatch[i].rm_eo == -1) |
|
538 |
return(NULL); |
|
539 |
assert(m->pmatch[i].rm_so != -1); |
|
540 |
len = m->pmatch[i].rm_eo - m->pmatch[i].rm_so; |
|
541 |
assert((size_t) (stop - m->beginp) >= len); |
|
542 |
if (sp > stop - len) |
|
543 |
return(NULL); /* not enough left to match */ |
|
544 |
ssp = m->offp + m->pmatch[i].rm_so; |
|
545 |
if (memcmp(sp, ssp, len) != 0) |
|
546 |
return(NULL); |
|
547 |
while (m->g->strip[ss] != SOP(O_BACK, i)) |
|
548 |
ss++; |
|
549 |
return(backref(charset, m, sp+len, stop, ss+1, stopst, lev)); |
|
550 |
break; |
|
551 |
case OQUEST_: /* to null or not */ |
|
552 |
dp = backref(charset, m, sp, stop, ss+1, stopst, lev); |
|
553 |
if (dp != NULL) |
|
554 |
return(dp); /* not */ |
|
555 |
return(backref(charset, m, sp, stop, ss+OPND(s)+1, stopst, lev)); |
|
556 |
break; |
|
557 |
case OPLUS_: |
|
558 |
assert(m->lastpos != NULL); |
|
559 |
assert(lev+1 <= m->g->nplus); |
|
560 |
m->lastpos[lev+1] = sp; |
|
561 |
return(backref(charset, m, sp, stop, ss+1, stopst, lev+1)); |
|
562 |
break; |
|
563 |
case O_PLUS: |
|
564 |
if (sp == m->lastpos[lev]) /* last pass matched null */ |
|
565 |
return(backref(charset, m, sp, stop, ss+1, stopst, lev-1)); |
|
566 |
/* try another pass */
|
|
567 |
m->lastpos[lev] = sp; |
|
568 |
dp = backref(charset, m, sp, stop, ss-OPND(s)+1, stopst, lev); |
|
569 |
if (dp == NULL) |
|
570 |
return(backref(charset, m, sp, stop, ss+1, stopst, lev-1)); |
|
571 |
else
|
|
572 |
return(dp); |
|
573 |
break; |
|
574 |
case OCH_: /* find the right one, if any */ |
|
575 |
ssub = ss + 1; |
|
576 |
esub = ss + OPND(s) - 1; |
|
577 |
assert(OP(m->g->strip[esub]) == OOR1); |
|
578 |
for (;;) { /* find first matching branch */ |
|
579 |
dp = backref(charset, m, sp, stop, ssub, esub, lev); |
|
580 |
if (dp != NULL) |
|
581 |
return(dp); |
|
582 |
/* that one missed, try next one */
|
|
583 |
if (OP(m->g->strip[esub]) == O_CH) |
|
584 |
return(NULL); /* there is none */ |
|
585 |
esub++; |
|
586 |
assert(OP(m->g->strip[esub]) == OOR2); |
|
587 |
ssub = esub + 1; |
|
588 |
esub += OPND(m->g->strip[esub]); |
|
589 |
if (OP(m->g->strip[esub]) == OOR2) |
|
590 |
esub--; |
|
591 |
else
|
|
592 |
assert(OP(m->g->strip[esub]) == O_CH); |
|
593 |
}
|
|
594 |
break; |
|
595 |
case OLPAREN: /* must undo assignment if rest fails */ |
|
596 |
i = OPND(s); |
|
597 |
assert(0 < i && i <= m->g->nsub); |
|
598 |
offsave = m->pmatch[i].rm_so; |
|
599 |
m->pmatch[i].rm_so = sp - m->offp; |
|
600 |
dp = backref(charset, m, sp, stop, ss+1, stopst, lev); |
|
601 |
if (dp != NULL) |
|
602 |
return(dp); |
|
603 |
m->pmatch[i].rm_so = offsave; |
|
604 |
return(NULL); |
|
605 |
break; |
|
606 |
case ORPAREN: /* must undo assignment if rest fails */ |
|
607 |
i = OPND(s); |
|
608 |
assert(0 < i && i <= m->g->nsub); |
|
609 |
offsave = m->pmatch[i].rm_eo; |
|
610 |
m->pmatch[i].rm_eo = sp - m->offp; |
|
611 |
dp = backref(charset, m, sp, stop, ss+1, stopst, lev); |
|
612 |
if (dp != NULL) |
|
613 |
return(dp); |
|
614 |
m->pmatch[i].rm_eo = offsave; |
|
615 |
return(NULL); |
|
616 |
break; |
|
617 |
default: /* uh oh */ |
|
618 |
assert(nope); |
|
619 |
break; |
|
620 |
}
|
|
621 |
||
622 |
/* "can't happen" */
|
|
623 |
assert(nope); |
|
624 |
/* NOTREACHED */
|
|
625 |
return 0; /* Keep gcc happy */ |
|
626 |
}
|
|
627 |
||
628 |
/*
|
|
629 |
- fast - step through the string at top speed
|
|
630 |
== static char *fast(register struct match *m, char *start, \
|
|
631 |
== char *stop, sopno startst, sopno stopst);
|
|
632 |
*/
|
|
633 |
static char * /* where tentative match ended, or NULL */ |
|
634 |
fast(charset, m, start, stop, startst, stopst) |
|
635 |
CHARSET_INFO *charset; |
|
636 |
register struct match *m; |
|
637 |
char *start; |
|
638 |
char *stop; |
|
639 |
sopno startst; |
|
640 |
sopno stopst; |
|
641 |
{
|
|
642 |
register states st = m->st; |
|
643 |
register states fresh = m->fresh; |
|
644 |
register states tmp = m->tmp; |
|
645 |
register char *p = start; |
|
646 |
register int c = (start == m->beginp) ? OUT : *(start-1); |
|
647 |
register int lastc; /* previous c */ |
|
648 |
register int flagch; |
|
649 |
register int i; |
|
650 |
register char *coldp; /* last p after which no match was underway */ |
|
651 |
||
652 |
CLEAR(st); |
|
653 |
SET1(st, startst); |
|
654 |
st = step(m->g, startst, stopst, st, NOTHING, st); |
|
655 |
ASSIGN(fresh, st); |
|
656 |
SP("start", st, *p); |
|
657 |
coldp = NULL; |
|
658 |
for (;;) { |
|
659 |
/* next character */
|
|
660 |
lastc = c; |
|
661 |
c = (p == m->endp) ? OUT : *p; |
|
662 |
if (EQ(st, fresh)) |
|
663 |
coldp = p; |
|
664 |
||
665 |
/* is there an EOL and/or BOL between lastc and c? */
|
|
666 |
flagch = '\0'; |
|
667 |
i = 0; |
|
668 |
if ( (lastc == '\n' && m->g->cflags®_NEWLINE) || |
|
669 |
(lastc == OUT && !(m->eflags®_NOTBOL)) ) { |
|
670 |
flagch = BOL; |
|
671 |
i = m->g->nbol; |
|
672 |
}
|
|
673 |
if ( (c == '\n' && m->g->cflags®_NEWLINE) || |
|
674 |
(c == OUT && !(m->eflags®_NOTEOL)) ) { |
|
675 |
flagch = (flagch == BOL) ? BOLEOL : EOL; |
|
676 |
i += m->g->neol; |
|
677 |
}
|
|
678 |
if (i != 0) { |
|
679 |
for (; i > 0; i--) |
|
680 |
st = step(m->g, startst, stopst, st, flagch, st); |
|
681 |
SP("boleol", st, c); |
|
682 |
}
|
|
683 |
||
684 |
/* how about a word boundary? */
|
|
685 |
if ( (flagch == BOL || (lastc != OUT && !ISWORD(charset,lastc))) && |
|
686 |
(c != OUT && ISWORD(charset,c)) ) { |
|
687 |
flagch = BOW; |
|
688 |
}
|
|
689 |
if ( (lastc != OUT && ISWORD(charset,lastc)) && |
|
690 |
(flagch == EOL || (c != OUT && !ISWORD(charset,c))) ) { |
|
691 |
flagch = EOW; |
|
692 |
}
|
|
693 |
if (flagch == BOW || flagch == EOW) { |
|
694 |
st = step(m->g, startst, stopst, st, flagch, st); |
|
695 |
SP("boweow", st, c); |
|
696 |
}
|
|
697 |
||
698 |
/* are we done? */
|
|
699 |
if (ISSET(st, stopst) || p == stop) |
|
700 |
break; /* NOTE BREAK OUT */ |
|
701 |
||
702 |
/* no, we must deal with this character */
|
|
703 |
ASSIGN(tmp, st); |
|
704 |
ASSIGN(st, fresh); |
|
705 |
assert(c != OUT); |
|
706 |
st = step(m->g, startst, stopst, tmp, c, st); |
|
707 |
SP("aft", st, c); |
|
708 |
assert(EQ(step(m->g, startst, stopst, st, NOTHING, st), st)); |
|
709 |
p++; |
|
710 |
}
|
|
711 |
||
712 |
assert(coldp != NULL); |
|
713 |
m->coldp = coldp; |
|
714 |
if (ISSET(st, stopst)) |
|
715 |
return(p+1); |
|
716 |
else
|
|
717 |
return(NULL); |
|
718 |
}
|
|
719 |
||
720 |
/*
|
|
721 |
- slow - step through the string more deliberately
|
|
722 |
== static char *slow(register struct match *m, char *start, \
|
|
723 |
== char *stop, sopno startst, sopno stopst);
|
|
724 |
*/
|
|
725 |
static char * /* where it ended */ |
|
726 |
slow(charset, m, start, stop, startst, stopst) |
|
727 |
CHARSET_INFO *charset; |
|
728 |
register struct match *m; |
|
729 |
char *start; |
|
730 |
char *stop; |
|
731 |
sopno startst; |
|
732 |
sopno stopst; |
|
733 |
{
|
|
734 |
register states st = m->st; |
|
735 |
register states empty = m->empty; |
|
736 |
register states tmp = m->tmp; |
|
737 |
register char *p = start; |
|
738 |
register int c = (start == m->beginp) ? OUT : *(start-1); |
|
739 |
register int lastc; /* previous c */ |
|
740 |
register int flagch; |
|
741 |
register int i; |
|
742 |
register char *matchp; /* last p at which a match ended */ |
|
743 |
||
744 |
AT("slow", start, stop, startst, stopst); |
|
745 |
CLEAR(st); |
|
746 |
SET1(st, startst); |
|
747 |
SP("sstart", st, *p); |
|
748 |
st = step(m->g, startst, stopst, st, NOTHING, st); |
|
749 |
matchp = NULL; |
|
750 |
for (;;) { |
|
751 |
/* next character */
|
|
752 |
lastc = c; |
|
753 |
c = (p == m->endp) ? OUT : *p; |
|
754 |
||
755 |
/* is there an EOL and/or BOL between lastc and c? */
|
|
756 |
flagch = '\0'; |
|
757 |
i = 0; |
|
758 |
if ( (lastc == '\n' && m->g->cflags®_NEWLINE) || |
|
759 |
(lastc == OUT && !(m->eflags®_NOTBOL)) ) { |
|
760 |
flagch = BOL; |
|
761 |
i = m->g->nbol; |
|
762 |
}
|
|
763 |
if ( (c == '\n' && m->g->cflags®_NEWLINE) || |
|
764 |
(c == OUT && !(m->eflags®_NOTEOL)) ) { |
|
765 |
flagch = (flagch == BOL) ? BOLEOL : EOL; |
|
766 |
i += m->g->neol; |
|
767 |
}
|
|
768 |
if (i != 0) { |
|
769 |
for (; i > 0; i--) |
|
770 |
st = step(m->g, startst, stopst, st, flagch, st); |
|
771 |
SP("sboleol", st, c); |
|
772 |
}
|
|
773 |
||
774 |
/* how about a word boundary? */
|
|
775 |
if ( (flagch == BOL || (lastc != OUT && !ISWORD(charset,lastc))) && |
|
776 |
(c != OUT && ISWORD(charset,c)) ) { |
|
777 |
flagch = BOW; |
|
778 |
}
|
|
779 |
if ( (lastc != OUT && ISWORD(charset,lastc)) && |
|
780 |
(flagch == EOL || (c != OUT && !ISWORD(charset,c))) ) { |
|
781 |
flagch = EOW; |
|
782 |
}
|
|
783 |
if (flagch == BOW || flagch == EOW) { |
|
784 |
st = step(m->g, startst, stopst, st, flagch, st); |
|
785 |
SP("sboweow", st, c); |
|
786 |
}
|
|
787 |
||
788 |
/* are we done? */
|
|
789 |
if (ISSET(st, stopst)) |
|
790 |
matchp = p; |
|
791 |
if (EQ(st, empty) || p == stop) |
|
792 |
break; /* NOTE BREAK OUT */ |
|
793 |
||
794 |
/* no, we must deal with this character */
|
|
795 |
ASSIGN(tmp, st); |
|
796 |
ASSIGN(st, empty); |
|
797 |
assert(c != OUT); |
|
798 |
st = step(m->g, startst, stopst, tmp, c, st); |
|
799 |
SP("saft", st, c); |
|
800 |
assert(EQ(step(m->g, startst, stopst, st, NOTHING, st), st)); |
|
801 |
p++; |
|
802 |
}
|
|
803 |
||
804 |
return(matchp); |
|
805 |
}
|
|
806 |
||
807 |
||
808 |
/*
|
|
809 |
- step - map set of states reachable before char to set reachable after
|
|
810 |
== static states step(register struct re_guts *g, sopno start, sopno stop, \
|
|
811 |
== register states bef, int ch, register states aft);
|
|
812 |
== #define BOL (OUT+1)
|
|
813 |
== #define EOL (BOL+1)
|
|
814 |
== #define BOLEOL (BOL+2)
|
|
815 |
== #define NOTHING (BOL+3)
|
|
816 |
== #define BOW (BOL+4)
|
|
817 |
== #define EOW (BOL+5)
|
|
818 |
== #define CODEMAX (BOL+5) // highest code used
|
|
819 |
== #define NONCHAR(c) ((c) > CHAR_MAX)
|
|
820 |
== #define NNONCHAR (CODEMAX-CHAR_MAX)
|
|
821 |
*/
|
|
822 |
static states |
|
823 |
step(g, start, stop, bef, ch, aft) |
|
824 |
register struct re_guts *g; |
|
825 |
sopno start; /* start state within strip */ |
|
826 |
sopno stop; /* state after stop state within strip */ |
|
827 |
register states bef; /* states reachable before */ |
|
828 |
int ch; /* character or NONCHAR code */ |
|
829 |
register states aft; /* states already known reachable after */ |
|
830 |
{
|
|
831 |
register cset *cs; |
|
832 |
register sop s; |
|
833 |
register sopno pc; |
|
834 |
register onestate here; /* note, macros know this name */ |
|
835 |
register sopno look; |
|
836 |
register onestate i; /* Changed from int by Monty */ |
|
837 |
||
838 |
for (pc = start, INIT(here, pc); pc != stop; pc++, INC(here)) { |
|
839 |
s = g->strip[pc]; |
|
840 |
switch (OP(s)) { |
|
841 |
case OEND: |
|
842 |
assert(pc == stop-1); |
|
843 |
break; |
|
844 |
case OCHAR: |
|
845 |
/* only characters can match */
|
|
846 |
assert(!NONCHAR(ch) || ch != (char)OPND(s)); |
|
847 |
if (ch == (char)OPND(s)) |
|
848 |
FWD(aft, bef, 1); |
|
849 |
break; |
|
850 |
case OBOL: |
|
851 |
if (ch == BOL || ch == BOLEOL) |
|
852 |
FWD(aft, bef, 1); |
|
853 |
break; |
|
854 |
case OEOL: |
|
855 |
if (ch == EOL || ch == BOLEOL) |
|
856 |
FWD(aft, bef, 1); |
|
857 |
break; |
|
858 |
case OBOW: |
|
859 |
if (ch == BOW) |
|
860 |
FWD(aft, bef, 1); |
|
861 |
break; |
|
862 |
case OEOW: |
|
863 |
if (ch == EOW) |
|
864 |
FWD(aft, bef, 1); |
|
865 |
break; |
|
866 |
case OANY: |
|
867 |
if (!NONCHAR(ch)) |
|
868 |
FWD(aft, bef, 1); |
|
869 |
break; |
|
870 |
case OANYOF: |
|
871 |
cs = &g->sets[OPND(s)]; |
|
872 |
if (!NONCHAR(ch) && CHIN(cs, ch)) |
|
873 |
FWD(aft, bef, 1); |
|
874 |
break; |
|
875 |
case OBACK_: /* ignored here */ |
|
876 |
case O_BACK: |
|
877 |
FWD(aft, aft, 1); |
|
878 |
break; |
|
879 |
case OPLUS_: /* forward, this is just an empty */ |
|
880 |
FWD(aft, aft, 1); |
|
881 |
break; |
|
882 |
case O_PLUS: /* both forward and back */ |
|
883 |
FWD(aft, aft, 1); |
|
884 |
i = ISSETBACK(aft, OPND(s)); |
|
885 |
BACK(aft, aft, OPND(s)); |
|
886 |
if (!i && ISSETBACK(aft, OPND(s))) { |
|
887 |
/* oho, must reconsider loop body */
|
|
888 |
pc -= OPND(s) + 1; |
|
889 |
INIT(here, pc); |
|
890 |
}
|
|
891 |
break; |
|
892 |
case OQUEST_: /* two branches, both forward */ |
|
893 |
FWD(aft, aft, 1); |
|
894 |
FWD(aft, aft, OPND(s)); |
|
895 |
break; |
|
896 |
case O_QUEST: /* just an empty */ |
|
897 |
FWD(aft, aft, 1); |
|
898 |
break; |
|
899 |
case OLPAREN: /* not significant here */ |
|
900 |
case ORPAREN: |
|
901 |
FWD(aft, aft, 1); |
|
902 |
break; |
|
903 |
case OCH_: /* mark the first two branches */ |
|
904 |
FWD(aft, aft, 1); |
|
905 |
assert(OP(g->strip[pc+OPND(s)]) == OOR2); |
|
906 |
FWD(aft, aft, OPND(s)); |
|
907 |
break; |
|
908 |
case OOR1: /* done a branch, find the O_CH */ |
|
909 |
if (ISSTATEIN(aft, here)) { |
|
910 |
for (look = 1; |
|
911 |
OP(s = g->strip[pc+look]) != O_CH; |
|
912 |
look += OPND(s)) |
|
913 |
assert(OP(s) == OOR2); |
|
914 |
FWD(aft, aft, look); |
|
915 |
}
|
|
916 |
break; |
|
917 |
case OOR2: /* propagate OCH_'s marking */ |
|
918 |
FWD(aft, aft, 1); |
|
919 |
if (OP(g->strip[pc+OPND(s)]) != O_CH) { |
|
920 |
assert(OP(g->strip[pc+OPND(s)]) == OOR2); |
|
921 |
FWD(aft, aft, OPND(s)); |
|
922 |
}
|
|
923 |
break; |
|
924 |
case O_CH: /* just empty */ |
|
925 |
FWD(aft, aft, 1); |
|
926 |
break; |
|
927 |
default: /* ooooops... */ |
|
928 |
assert(nope); |
|
929 |
break; |
|
930 |
}
|
|
931 |
}
|
|
932 |
||
933 |
return(aft); |
|
934 |
}
|
|
935 |
||
936 |
#ifdef REDEBUG
|
|
937 |
/*
|
|
938 |
- print - print a set of states
|
|
939 |
== #ifdef REDEBUG
|
|
940 |
== static void print(struct match *m, char *caption, states st, \
|
|
941 |
== int ch, FILE *d);
|
|
942 |
== #endif
|
|
943 |
*/
|
|
944 |
static void |
|
945 |
print(m, caption, st, ch, d) |
|
946 |
struct match *m; |
|
947 |
char *caption; |
|
948 |
states st; |
|
949 |
int ch; |
|
950 |
FILE *d; |
|
951 |
{
|
|
952 |
register struct re_guts *g = m->g; |
|
953 |
register int i; |
|
954 |
register int first = 1; |
|
955 |
char buf[10]; |
|
956 |
||
957 |
if (!(m->eflags®_TRACE)) |
|
958 |
return; |
|
959 |
||
960 |
fprintf(d, "%s", caption); |
|
961 |
if (ch != '\0') |
|
962 |
fprintf(d, " %s", printchar(ch,buf)); |
|
963 |
for (i = 0; i < g->nstates; i++) |
|
964 |
if (ISSET(st, i)) { |
|
965 |
fprintf(d, "%s%d", (first) ? "\t" : ", ", i); |
|
966 |
first = 0; |
|
967 |
}
|
|
968 |
fprintf(d, "\n"); |
|
969 |
}
|
|
970 |
||
971 |
/*
|
|
972 |
- at - print current situation
|
|
973 |
== #ifdef REDEBUG
|
|
974 |
== static void at(struct match *m, char *title, char *start, char *stop, \
|
|
975 |
== sopno startst, sopno stopst);
|
|
976 |
== #endif
|
|
977 |
*/
|
|
978 |
static void |
|
979 |
at(m, title, start, stop, startst, stopst) |
|
980 |
struct match *m; |
|
981 |
char *title; |
|
982 |
char *start; |
|
983 |
char *stop; |
|
984 |
sopno startst; |
|
985 |
sopno stopst; |
|
986 |
{
|
|
987 |
char buf[10]; |
|
988 |
if (!(m->eflags®_TRACE)) |
|
989 |
return; |
|
990 |
||
991 |
printf("%s %s-", title, printchar(*start,buf)); |
|
992 |
printf("%s ", printchar(*stop,buf)); |
|
993 |
printf("%ld-%ld\n", (long)startst, (long)stopst,buf); |
|
994 |
}
|
|
995 |
||
996 |
#ifndef PCHARDONE
|
|
997 |
#define PCHARDONE /* never again */ |
|
998 |
/*
|
|
999 |
- printchar - make a character printable
|
|
1000 |
== #ifdef REDEBUG
|
|
1001 |
== static char *printchar(int ch);
|
|
1002 |
== #endif
|
|
1003 |
*
|
|
1004 |
* Is this identical to regchar() over in debug.c? Well, yes. But a
|
|
1005 |
* duplicate here avoids having a debugging-capable regexec.o tied to
|
|
1006 |
* a matching debug.o, and this is convenient. It all disappears in
|
|
1007 |
* the non-debug compilation anyway, so it doesn't matter much.
|
|
1008 |
*/
|
|
1009 |
static char * /* -> representation */ |
|
1010 |
printchar(ch,pbuf) |
|
1011 |
int ch; |
|
1012 |
char *pbuf; |
|
1013 |
{
|
|
1014 |
if (isprint(ch) || ch == ' ') |
|
1015 |
sprintf(pbuf, "%c", ch); |
|
1016 |
else
|
|
1017 |
sprintf(pbuf, "\\%o", ch); |
|
1018 |
return(pbuf); |
|
1019 |
}
|
|
1020 |
#endif
|
|
1021 |
#endif
|
|
1022 |
||
1023 |
#undef matcher
|
|
1024 |
#undef fast
|
|
1025 |
#undef slow
|
|
1026 |
#undef dissect
|
|
1027 |
#undef backref
|
|
1028 |
#undef step
|
|
1029 |
#undef print
|
|
1030 |
#undef at
|
|
1031 |
#undef match
|