]>
Commit | Line | Data |
---|---|---|
21d3294c | 1 | /* |
2 | ** $Id: lstrlib.c,v 1.132.1.4 2008/07/11 17:27:21 roberto Exp $ | |
3 | ** Standard library for string operations and pattern-matching | |
4 | ** See Copyright Notice in lua.h | |
5 | */ | |
6 | ||
7 | ||
8 | #include <ctype.h> | |
9 | #include <stddef.h> | |
10 | #include <stdio.h> | |
11 | #include <stdlib.h> | |
12 | #include <string.h> | |
13 | ||
14 | #define lstrlib_c | |
15 | #define LUA_LIB | |
16 | ||
17 | #include "lua.h" | |
18 | ||
19 | #include "lauxlib.h" | |
20 | #include "lualib.h" | |
21 | ||
22 | ||
23 | /* macro to `unsign' a character */ | |
24 | #define uchar(c) ((unsigned char)(c)) | |
25 | ||
26 | ||
27 | ||
28 | static int str_len (lua_State *L) { | |
29 | size_t l; | |
30 | luaL_checklstring(L, 1, &l); | |
31 | lua_pushinteger(L, l); | |
32 | return 1; | |
33 | } | |
34 | ||
35 | ||
36 | static ptrdiff_t posrelat (ptrdiff_t pos, size_t len) { | |
37 | /* relative string position: negative means back from end */ | |
38 | if (pos < 0) pos += (ptrdiff_t)len + 1; | |
39 | return (pos >= 0) ? pos : 0; | |
40 | } | |
41 | ||
42 | ||
43 | static int str_sub (lua_State *L) { | |
44 | size_t l; | |
45 | const char *s = luaL_checklstring(L, 1, &l); | |
46 | ptrdiff_t start = posrelat(luaL_checkinteger(L, 2), l); | |
47 | ptrdiff_t end = posrelat(luaL_optinteger(L, 3, -1), l); | |
48 | if (start < 1) start = 1; | |
49 | if (end > (ptrdiff_t)l) end = (ptrdiff_t)l; | |
50 | if (start <= end) | |
51 | lua_pushlstring(L, s+start-1, end-start+1); | |
52 | else lua_pushliteral(L, ""); | |
53 | return 1; | |
54 | } | |
55 | ||
56 | ||
57 | static int str_reverse (lua_State *L) { | |
58 | size_t l; | |
59 | luaL_Buffer b; | |
60 | const char *s = luaL_checklstring(L, 1, &l); | |
61 | luaL_buffinit(L, &b); | |
62 | while (l--) luaL_addchar(&b, s[l]); | |
63 | luaL_pushresult(&b); | |
64 | return 1; | |
65 | } | |
66 | ||
67 | ||
68 | static int str_lower (lua_State *L) { | |
69 | size_t l; | |
70 | size_t i; | |
71 | luaL_Buffer b; | |
72 | const char *s = luaL_checklstring(L, 1, &l); | |
73 | luaL_buffinit(L, &b); | |
74 | for (i=0; i<l; i++) | |
75 | luaL_addchar(&b, tolower(uchar(s[i]))); | |
76 | luaL_pushresult(&b); | |
77 | return 1; | |
78 | } | |
79 | ||
80 | ||
81 | static int str_upper (lua_State *L) { | |
82 | size_t l; | |
83 | size_t i; | |
84 | luaL_Buffer b; | |
85 | const char *s = luaL_checklstring(L, 1, &l); | |
86 | luaL_buffinit(L, &b); | |
87 | for (i=0; i<l; i++) | |
88 | luaL_addchar(&b, toupper(uchar(s[i]))); | |
89 | luaL_pushresult(&b); | |
90 | return 1; | |
91 | } | |
92 | ||
93 | static int str_rep (lua_State *L) { | |
94 | size_t l; | |
95 | luaL_Buffer b; | |
96 | const char *s = luaL_checklstring(L, 1, &l); | |
97 | int n = luaL_checkint(L, 2); | |
98 | luaL_buffinit(L, &b); | |
99 | while (n-- > 0) | |
100 | luaL_addlstring(&b, s, l); | |
101 | luaL_pushresult(&b); | |
102 | return 1; | |
103 | } | |
104 | ||
105 | ||
106 | static int str_byte (lua_State *L) { | |
107 | size_t l; | |
108 | const char *s = luaL_checklstring(L, 1, &l); | |
109 | ptrdiff_t posi = posrelat(luaL_optinteger(L, 2, 1), l); | |
110 | ptrdiff_t pose = posrelat(luaL_optinteger(L, 3, posi), l); | |
111 | int n, i; | |
112 | if (posi <= 0) posi = 1; | |
113 | if ((size_t)pose > l) pose = l; | |
114 | if (posi > pose) return 0; /* empty interval; return no values */ | |
115 | n = (int)(pose - posi + 1); | |
116 | if (posi + n <= pose) /* overflow? */ | |
117 | luaL_error(L, "string slice too long"); | |
118 | luaL_checkstack(L, n, "string slice too long"); | |
119 | for (i=0; i<n; i++) | |
120 | lua_pushinteger(L, uchar(s[posi+i-1])); | |
121 | return n; | |
122 | } | |
123 | ||
124 | ||
125 | static int str_char (lua_State *L) { | |
126 | int n = lua_gettop(L); /* number of arguments */ | |
127 | int i; | |
128 | luaL_Buffer b; | |
129 | luaL_buffinit(L, &b); | |
130 | for (i=1; i<=n; i++) { | |
131 | int c = luaL_checkint(L, i); | |
132 | luaL_argcheck(L, uchar(c) == c, i, "invalid value"); | |
133 | luaL_addchar(&b, uchar(c)); | |
134 | } | |
135 | luaL_pushresult(&b); | |
136 | return 1; | |
137 | } | |
138 | ||
139 | ||
140 | static int writer (lua_State *L, const void* b, size_t size, void* B) { | |
141 | (void)L; | |
142 | luaL_addlstring((luaL_Buffer*) B, (const char *)b, size); | |
143 | return 0; | |
144 | } | |
145 | ||
146 | ||
147 | static int str_dump (lua_State *L) { | |
148 | luaL_Buffer b; | |
149 | luaL_checktype(L, 1, LUA_TFUNCTION); | |
150 | lua_settop(L, 1); | |
151 | luaL_buffinit(L,&b); | |
152 | if (lua_dump(L, writer, &b) != 0) | |
153 | luaL_error(L, "unable to dump given function"); | |
154 | luaL_pushresult(&b); | |
155 | return 1; | |
156 | } | |
157 | ||
158 | ||
159 | ||
160 | /* | |
161 | ** {====================================================== | |
162 | ** PATTERN MATCHING | |
163 | ** ======================================================= | |
164 | */ | |
165 | ||
166 | ||
167 | #define CAP_UNFINISHED (-1) | |
168 | #define CAP_POSITION (-2) | |
169 | ||
170 | typedef struct MatchState { | |
171 | const char *src_init; /* init of source string */ | |
172 | const char *src_end; /* end (`\0') of source string */ | |
173 | lua_State *L; | |
174 | int level; /* total number of captures (finished or unfinished) */ | |
175 | struct { | |
176 | const char *init; | |
177 | ptrdiff_t len; | |
178 | } capture[LUA_MAXCAPTURES]; | |
179 | } MatchState; | |
180 | ||
181 | ||
182 | #define L_ESC '%' | |
183 | #define SPECIALS "^$*+?.([%-" | |
184 | ||
185 | ||
186 | static int check_capture (MatchState *ms, int l) { | |
187 | l -= '1'; | |
188 | if (l < 0 || l >= ms->level || ms->capture[l].len == CAP_UNFINISHED) | |
189 | return luaL_error(ms->L, "invalid capture index"); | |
190 | return l; | |
191 | } | |
192 | ||
193 | ||
194 | static int capture_to_close (MatchState *ms) { | |
195 | int level = ms->level; | |
196 | for (level--; level>=0; level--) | |
197 | if (ms->capture[level].len == CAP_UNFINISHED) return level; | |
198 | return luaL_error(ms->L, "invalid pattern capture"); | |
199 | } | |
200 | ||
201 | ||
202 | static const char *classend (MatchState *ms, const char *p) { | |
203 | switch (*p++) { | |
204 | case L_ESC: { | |
205 | if (*p == '\0') | |
206 | luaL_error(ms->L, "malformed pattern (ends with " LUA_QL("%%") ")"); | |
207 | return p+1; | |
208 | } | |
209 | case '[': { | |
210 | if (*p == '^') p++; | |
211 | do { /* look for a `]' */ | |
212 | if (*p == '\0') | |
213 | luaL_error(ms->L, "malformed pattern (missing " LUA_QL("]") ")"); | |
214 | if (*(p++) == L_ESC && *p != '\0') | |
215 | p++; /* skip escapes (e.g. `%]') */ | |
216 | } while (*p != ']'); | |
217 | return p+1; | |
218 | } | |
219 | default: { | |
220 | return p; | |
221 | } | |
222 | } | |
223 | } | |
224 | ||
225 | ||
226 | static int match_class (int c, int cl) { | |
227 | int res; | |
228 | switch (tolower(cl)) { | |
229 | case 'a' : res = isalpha(c); break; | |
230 | case 'c' : res = iscntrl(c); break; | |
231 | case 'd' : res = isdigit(c); break; | |
232 | case 'l' : res = islower(c); break; | |
233 | case 'p' : res = ispunct(c); break; | |
234 | case 's' : res = isspace(c); break; | |
235 | case 'u' : res = isupper(c); break; | |
236 | case 'w' : res = isalnum(c); break; | |
237 | case 'x' : res = isxdigit(c); break; | |
238 | case 'z' : res = (c == 0); break; | |
239 | default: return (cl == c); | |
240 | } | |
241 | return (islower(cl) ? res : !res); | |
242 | } | |
243 | ||
244 | ||
245 | static int matchbracketclass (int c, const char *p, const char *ec) { | |
246 | int sig = 1; | |
247 | if (*(p+1) == '^') { | |
248 | sig = 0; | |
249 | p++; /* skip the `^' */ | |
250 | } | |
251 | while (++p < ec) { | |
252 | if (*p == L_ESC) { | |
253 | p++; | |
254 | if (match_class(c, uchar(*p))) | |
255 | return sig; | |
256 | } | |
257 | else if ((*(p+1) == '-') && (p+2 < ec)) { | |
258 | p+=2; | |
259 | if (uchar(*(p-2)) <= c && c <= uchar(*p)) | |
260 | return sig; | |
261 | } | |
262 | else if (uchar(*p) == c) return sig; | |
263 | } | |
264 | return !sig; | |
265 | } | |
266 | ||
267 | ||
268 | static int singlematch (int c, const char *p, const char *ep) { | |
269 | switch (*p) { | |
270 | case '.': return 1; /* matches any char */ | |
271 | case L_ESC: return match_class(c, uchar(*(p+1))); | |
272 | case '[': return matchbracketclass(c, p, ep-1); | |
273 | default: return (uchar(*p) == c); | |
274 | } | |
275 | } | |
276 | ||
277 | ||
278 | static const char *match (MatchState *ms, const char *s, const char *p); | |
279 | ||
280 | ||
281 | static const char *matchbalance (MatchState *ms, const char *s, | |
282 | const char *p) { | |
283 | if (*p == 0 || *(p+1) == 0) | |
284 | luaL_error(ms->L, "unbalanced pattern"); | |
285 | if (*s != *p) return NULL; | |
286 | else { | |
287 | int b = *p; | |
288 | int e = *(p+1); | |
289 | int cont = 1; | |
290 | while (++s < ms->src_end) { | |
291 | if (*s == e) { | |
292 | if (--cont == 0) return s+1; | |
293 | } | |
294 | else if (*s == b) cont++; | |
295 | } | |
296 | } | |
297 | return NULL; /* string ends out of balance */ | |
298 | } | |
299 | ||
300 | ||
301 | static const char *max_expand (MatchState *ms, const char *s, | |
302 | const char *p, const char *ep) { | |
303 | ptrdiff_t i = 0; /* counts maximum expand for item */ | |
304 | while ((s+i)<ms->src_end && singlematch(uchar(*(s+i)), p, ep)) | |
305 | i++; | |
306 | /* keeps trying to match with the maximum repetitions */ | |
307 | while (i>=0) { | |
308 | const char *res = match(ms, (s+i), ep+1); | |
309 | if (res) return res; | |
310 | i--; /* else didn't match; reduce 1 repetition to try again */ | |
311 | } | |
312 | return NULL; | |
313 | } | |
314 | ||
315 | ||
316 | static const char *min_expand (MatchState *ms, const char *s, | |
317 | const char *p, const char *ep) { | |
318 | for (;;) { | |
319 | const char *res = match(ms, s, ep+1); | |
320 | if (res != NULL) | |
321 | return res; | |
322 | else if (s<ms->src_end && singlematch(uchar(*s), p, ep)) | |
323 | s++; /* try with one more repetition */ | |
324 | else return NULL; | |
325 | } | |
326 | } | |
327 | ||
328 | ||
329 | static const char *start_capture (MatchState *ms, const char *s, | |
330 | const char *p, int what) { | |
331 | const char *res; | |
332 | int level = ms->level; | |
333 | if (level >= LUA_MAXCAPTURES) luaL_error(ms->L, "too many captures"); | |
334 | ms->capture[level].init = s; | |
335 | ms->capture[level].len = what; | |
336 | ms->level = level+1; | |
337 | if ((res=match(ms, s, p)) == NULL) /* match failed? */ | |
338 | ms->level--; /* undo capture */ | |
339 | return res; | |
340 | } | |
341 | ||
342 | ||
343 | static const char *end_capture (MatchState *ms, const char *s, | |
344 | const char *p) { | |
345 | int l = capture_to_close(ms); | |
346 | const char *res; | |
347 | ms->capture[l].len = s - ms->capture[l].init; /* close capture */ | |
348 | if ((res = match(ms, s, p)) == NULL) /* match failed? */ | |
349 | ms->capture[l].len = CAP_UNFINISHED; /* undo capture */ | |
350 | return res; | |
351 | } | |
352 | ||
353 | ||
354 | static const char *match_capture (MatchState *ms, const char *s, int l) { | |
355 | size_t len; | |
356 | l = check_capture(ms, l); | |
357 | len = ms->capture[l].len; | |
358 | if ((size_t)(ms->src_end-s) >= len && | |
359 | memcmp(ms->capture[l].init, s, len) == 0) | |
360 | return s+len; | |
361 | else return NULL; | |
362 | } | |
363 | ||
364 | ||
365 | static const char *match (MatchState *ms, const char *s, const char *p) { | |
366 | init: /* using goto's to optimize tail recursion */ | |
367 | switch (*p) { | |
368 | case '(': { /* start capture */ | |
369 | if (*(p+1) == ')') /* position capture? */ | |
370 | return start_capture(ms, s, p+2, CAP_POSITION); | |
371 | else | |
372 | return start_capture(ms, s, p+1, CAP_UNFINISHED); | |
373 | } | |
374 | case ')': { /* end capture */ | |
375 | return end_capture(ms, s, p+1); | |
376 | } | |
377 | case L_ESC: { | |
378 | switch (*(p+1)) { | |
379 | case 'b': { /* balanced string? */ | |
380 | s = matchbalance(ms, s, p+2); | |
381 | if (s == NULL) return NULL; | |
382 | p+=4; goto init; /* else return match(ms, s, p+4); */ | |
383 | } | |
384 | case 'f': { /* frontier? */ | |
385 | const char *ep; char previous; | |
386 | p += 2; | |
387 | if (*p != '[') | |
388 | luaL_error(ms->L, "missing " LUA_QL("[") " after " | |
389 | LUA_QL("%%f") " in pattern"); | |
390 | ep = classend(ms, p); /* points to what is next */ | |
391 | previous = (s == ms->src_init) ? '\0' : *(s-1); | |
392 | if (matchbracketclass(uchar(previous), p, ep-1) || | |
393 | !matchbracketclass(uchar(*s), p, ep-1)) return NULL; | |
394 | p=ep; goto init; /* else return match(ms, s, ep); */ | |
395 | } | |
396 | default: { | |
397 | if (isdigit(uchar(*(p+1)))) { /* capture results (%0-%9)? */ | |
398 | s = match_capture(ms, s, uchar(*(p+1))); | |
399 | if (s == NULL) return NULL; | |
400 | p+=2; goto init; /* else return match(ms, s, p+2) */ | |
401 | } | |
402 | goto dflt; /* case default */ | |
403 | } | |
404 | } | |
405 | } | |
406 | case '\0': { /* end of pattern */ | |
407 | return s; /* match succeeded */ | |
408 | } | |
409 | case '$': { | |
410 | if (*(p+1) == '\0') /* is the `$' the last char in pattern? */ | |
411 | return (s == ms->src_end) ? s : NULL; /* check end of string */ | |
412 | else goto dflt; | |
413 | } | |
414 | default: dflt: { /* it is a pattern item */ | |
415 | const char *ep = classend(ms, p); /* points to what is next */ | |
416 | int m = s<ms->src_end && singlematch(uchar(*s), p, ep); | |
417 | switch (*ep) { | |
418 | case '?': { /* optional */ | |
419 | const char *res; | |
420 | if (m && ((res=match(ms, s+1, ep+1)) != NULL)) | |
421 | return res; | |
422 | p=ep+1; goto init; /* else return match(ms, s, ep+1); */ | |
423 | } | |
424 | case '*': { /* 0 or more repetitions */ | |
425 | return max_expand(ms, s, p, ep); | |
426 | } | |
427 | case '+': { /* 1 or more repetitions */ | |
428 | return (m ? max_expand(ms, s+1, p, ep) : NULL); | |
429 | } | |
430 | case '-': { /* 0 or more repetitions (minimum) */ | |
431 | return min_expand(ms, s, p, ep); | |
432 | } | |
433 | default: { | |
434 | if (!m) return NULL; | |
435 | s++; p=ep; goto init; /* else return match(ms, s+1, ep); */ | |
436 | } | |
437 | } | |
438 | } | |
439 | } | |
440 | } | |
441 | ||
442 | ||
443 | ||
444 | static const char *lmemfind (const char *s1, size_t l1, | |
445 | const char *s2, size_t l2) { | |
446 | if (l2 == 0) return s1; /* empty strings are everywhere */ | |
447 | else if (l2 > l1) return NULL; /* avoids a negative `l1' */ | |
448 | else { | |
449 | const char *init; /* to search for a `*s2' inside `s1' */ | |
450 | l2--; /* 1st char will be checked by `memchr' */ | |
451 | l1 = l1-l2; /* `s2' cannot be found after that */ | |
452 | while (l1 > 0 && (init = (const char *)memchr(s1, *s2, l1)) != NULL) { | |
453 | init++; /* 1st char is already checked */ | |
454 | if (memcmp(init, s2+1, l2) == 0) | |
455 | return init-1; | |
456 | else { /* correct `l1' and `s1' to try again */ | |
457 | l1 -= init-s1; | |
458 | s1 = init; | |
459 | } | |
460 | } | |
461 | return NULL; /* not found */ | |
462 | } | |
463 | } | |
464 | ||
465 | ||
466 | static void push_onecapture (MatchState *ms, int i, const char *s, | |
467 | const char *e) { | |
468 | if (i >= ms->level) { | |
469 | if (i == 0) /* ms->level == 0, too */ | |
470 | lua_pushlstring(ms->L, s, e - s); /* add whole match */ | |
471 | else | |
472 | luaL_error(ms->L, "invalid capture index"); | |
473 | } | |
474 | else { | |
475 | ptrdiff_t l = ms->capture[i].len; | |
476 | if (l == CAP_UNFINISHED) luaL_error(ms->L, "unfinished capture"); | |
477 | if (l == CAP_POSITION) | |
478 | lua_pushinteger(ms->L, ms->capture[i].init - ms->src_init + 1); | |
479 | else | |
480 | lua_pushlstring(ms->L, ms->capture[i].init, l); | |
481 | } | |
482 | } | |
483 | ||
484 | ||
485 | static int push_captures (MatchState *ms, const char *s, const char *e) { | |
486 | int i; | |
487 | int nlevels = (ms->level == 0 && s) ? 1 : ms->level; | |
488 | luaL_checkstack(ms->L, nlevels, "too many captures"); | |
489 | for (i = 0; i < nlevels; i++) | |
490 | push_onecapture(ms, i, s, e); | |
491 | return nlevels; /* number of strings pushed */ | |
492 | } | |
493 | ||
494 | ||
495 | static int str_find_aux (lua_State *L, int find) { | |
496 | size_t l1, l2; | |
497 | const char *s = luaL_checklstring(L, 1, &l1); | |
498 | const char *p = luaL_checklstring(L, 2, &l2); | |
499 | ptrdiff_t init = posrelat(luaL_optinteger(L, 3, 1), l1) - 1; | |
500 | if (init < 0) init = 0; | |
501 | else if ((size_t)(init) > l1) init = (ptrdiff_t)l1; | |
502 | if (find && (lua_toboolean(L, 4) || /* explicit request? */ | |
503 | strpbrk(p, SPECIALS) == NULL)) { /* or no special characters? */ | |
504 | /* do a plain search */ | |
505 | const char *s2 = lmemfind(s+init, l1-init, p, l2); | |
506 | if (s2) { | |
507 | lua_pushinteger(L, s2-s+1); | |
508 | lua_pushinteger(L, s2-s+l2); | |
509 | return 2; | |
510 | } | |
511 | } | |
512 | else { | |
513 | MatchState ms; | |
514 | int anchor = (*p == '^') ? (p++, 1) : 0; | |
515 | const char *s1=s+init; | |
516 | ms.L = L; | |
517 | ms.src_init = s; | |
518 | ms.src_end = s+l1; | |
519 | do { | |
520 | const char *res; | |
521 | ms.level = 0; | |
522 | if ((res=match(&ms, s1, p)) != NULL) { | |
523 | if (find) { | |
524 | lua_pushinteger(L, s1-s+1); /* start */ | |
525 | lua_pushinteger(L, res-s); /* end */ | |
526 | return push_captures(&ms, NULL, 0) + 2; | |
527 | } | |
528 | else | |
529 | return push_captures(&ms, s1, res); | |
530 | } | |
531 | } while (s1++ < ms.src_end && !anchor); | |
532 | } | |
533 | lua_pushnil(L); /* not found */ | |
534 | return 1; | |
535 | } | |
536 | ||
537 | ||
538 | static int str_find (lua_State *L) { | |
539 | return str_find_aux(L, 1); | |
540 | } | |
541 | ||
542 | ||
543 | static int str_match (lua_State *L) { | |
544 | return str_find_aux(L, 0); | |
545 | } | |
546 | ||
547 | ||
548 | static int gmatch_aux (lua_State *L) { | |
549 | MatchState ms; | |
550 | size_t ls; | |
551 | const char *s = lua_tolstring(L, lua_upvalueindex(1), &ls); | |
552 | const char *p = lua_tostring(L, lua_upvalueindex(2)); | |
553 | const char *src; | |
554 | ms.L = L; | |
555 | ms.src_init = s; | |
556 | ms.src_end = s+ls; | |
557 | for (src = s + (size_t)lua_tointeger(L, lua_upvalueindex(3)); | |
558 | src <= ms.src_end; | |
559 | src++) { | |
560 | const char *e; | |
561 | ms.level = 0; | |
562 | if ((e = match(&ms, src, p)) != NULL) { | |
563 | lua_Integer newstart = e-s; | |
564 | if (e == src) newstart++; /* empty match? go at least one position */ | |
565 | lua_pushinteger(L, newstart); | |
566 | lua_replace(L, lua_upvalueindex(3)); | |
567 | return push_captures(&ms, src, e); | |
568 | } | |
569 | } | |
570 | return 0; /* not found */ | |
571 | } | |
572 | ||
573 | ||
574 | static int gmatch (lua_State *L) { | |
575 | luaL_checkstring(L, 1); | |
576 | luaL_checkstring(L, 2); | |
577 | lua_settop(L, 2); | |
578 | lua_pushinteger(L, 0); | |
579 | lua_pushcclosure(L, gmatch_aux, 3); | |
580 | return 1; | |
581 | } | |
582 | ||
583 | ||
584 | static int gfind_nodef (lua_State *L) { | |
585 | return luaL_error(L, LUA_QL("string.gfind") " was renamed to " | |
586 | LUA_QL("string.gmatch")); | |
587 | } | |
588 | ||
589 | ||
590 | static void add_s (MatchState *ms, luaL_Buffer *b, const char *s, | |
591 | const char *e) { | |
592 | size_t l, i; | |
593 | const char *news = lua_tolstring(ms->L, 3, &l); | |
594 | for (i = 0; i < l; i++) { | |
595 | if (news[i] != L_ESC) | |
596 | luaL_addchar(b, news[i]); | |
597 | else { | |
598 | i++; /* skip ESC */ | |
599 | if (!isdigit(uchar(news[i]))) | |
600 | luaL_addchar(b, news[i]); | |
601 | else if (news[i] == '0') | |
602 | luaL_addlstring(b, s, e - s); | |
603 | else { | |
604 | push_onecapture(ms, news[i] - '1', s, e); | |
605 | luaL_addvalue(b); /* add capture to accumulated result */ | |
606 | } | |
607 | } | |
608 | } | |
609 | } | |
610 | ||
611 | ||
612 | static void add_value (MatchState *ms, luaL_Buffer *b, const char *s, | |
613 | const char *e) { | |
614 | lua_State *L = ms->L; | |
615 | switch (lua_type(L, 3)) { | |
616 | case LUA_TNUMBER: | |
617 | case LUA_TSTRING: { | |
618 | add_s(ms, b, s, e); | |
619 | return; | |
620 | } | |
621 | case LUA_TFUNCTION: { | |
622 | int n; | |
623 | lua_pushvalue(L, 3); | |
624 | n = push_captures(ms, s, e); | |
625 | lua_call(L, n, 1); | |
626 | break; | |
627 | } | |
628 | case LUA_TTABLE: { | |
629 | push_onecapture(ms, 0, s, e); | |
630 | lua_gettable(L, 3); | |
631 | break; | |
632 | } | |
633 | } | |
634 | if (!lua_toboolean(L, -1)) { /* nil or false? */ | |
635 | lua_pop(L, 1); | |
636 | lua_pushlstring(L, s, e - s); /* keep original text */ | |
637 | } | |
638 | else if (!lua_isstring(L, -1)) | |
639 | luaL_error(L, "invalid replacement value (a %s)", luaL_typename(L, -1)); | |
640 | luaL_addvalue(b); /* add result to accumulator */ | |
641 | } | |
642 | ||
643 | ||
644 | static int str_gsub (lua_State *L) { | |
645 | size_t srcl; | |
646 | const char *src = luaL_checklstring(L, 1, &srcl); | |
647 | const char *p = luaL_checkstring(L, 2); | |
648 | int tr = lua_type(L, 3); | |
649 | int max_s = luaL_optint(L, 4, srcl+1); | |
650 | int anchor = (*p == '^') ? (p++, 1) : 0; | |
651 | int n = 0; | |
652 | MatchState ms; | |
653 | luaL_Buffer b; | |
654 | luaL_argcheck(L, tr == LUA_TNUMBER || tr == LUA_TSTRING || | |
655 | tr == LUA_TFUNCTION || tr == LUA_TTABLE, 3, | |
656 | "string/function/table expected"); | |
657 | luaL_buffinit(L, &b); | |
658 | ms.L = L; | |
659 | ms.src_init = src; | |
660 | ms.src_end = src+srcl; | |
661 | while (n < max_s) { | |
662 | const char *e; | |
663 | ms.level = 0; | |
664 | e = match(&ms, src, p); | |
665 | if (e) { | |
666 | n++; | |
667 | add_value(&ms, &b, src, e); | |
668 | } | |
669 | if (e && e>src) /* non empty match? */ | |
670 | src = e; /* skip it */ | |
671 | else if (src < ms.src_end) | |
672 | luaL_addchar(&b, *src++); | |
673 | else break; | |
674 | if (anchor) break; | |
675 | } | |
676 | luaL_addlstring(&b, src, ms.src_end-src); | |
677 | luaL_pushresult(&b); | |
678 | lua_pushinteger(L, n); /* number of substitutions */ | |
679 | return 2; | |
680 | } | |
681 | ||
682 | /* }====================================================== */ | |
683 | ||
684 | ||
685 | /* maximum size of each formatted item (> len(format('%99.99f', -1e308))) */ | |
686 | #define MAX_ITEM 512 | |
687 | /* valid flags in a format specification */ | |
688 | #define FLAGS "-+ #0" | |
689 | /* | |
690 | ** maximum size of each format specification (such as '%-099.99d') | |
691 | ** (+10 accounts for %99.99x plus margin of error) | |
692 | */ | |
693 | #define MAX_FORMAT (sizeof(FLAGS) + sizeof(LUA_INTFRMLEN) + 10) | |
694 | ||
695 | ||
696 | static void addquoted (lua_State *L, luaL_Buffer *b, int arg) { | |
697 | size_t l; | |
698 | const char *s = luaL_checklstring(L, arg, &l); | |
699 | luaL_addchar(b, '"'); | |
700 | while (l--) { | |
701 | switch (*s) { | |
702 | case '"': case '\\': case '\n': { | |
703 | luaL_addchar(b, '\\'); | |
704 | luaL_addchar(b, *s); | |
705 | break; | |
706 | } | |
707 | case '\r': { | |
708 | luaL_addlstring(b, "\\r", 2); | |
709 | break; | |
710 | } | |
711 | case '\0': { | |
712 | luaL_addlstring(b, "\\000", 4); | |
713 | break; | |
714 | } | |
715 | default: { | |
716 | luaL_addchar(b, *s); | |
717 | break; | |
718 | } | |
719 | } | |
720 | s++; | |
721 | } | |
722 | luaL_addchar(b, '"'); | |
723 | } | |
724 | ||
725 | static const char *scanformat (lua_State *L, const char *strfrmt, char *form) { | |
726 | const char *p = strfrmt; | |
727 | while (*p != '\0' && strchr(FLAGS, *p) != NULL) p++; /* skip flags */ | |
728 | if ((size_t)(p - strfrmt) >= sizeof(FLAGS)) | |
729 | luaL_error(L, "invalid format (repeated flags)"); | |
730 | if (isdigit(uchar(*p))) p++; /* skip width */ | |
731 | if (isdigit(uchar(*p))) p++; /* (2 digits at most) */ | |
732 | if (*p == '.') { | |
733 | p++; | |
734 | if (isdigit(uchar(*p))) p++; /* skip precision */ | |
735 | if (isdigit(uchar(*p))) p++; /* (2 digits at most) */ | |
736 | } | |
737 | if (isdigit(uchar(*p))) | |
738 | luaL_error(L, "invalid format (width or precision too long)"); | |
739 | *(form++) = '%'; | |
740 | strncpy(form, strfrmt, p - strfrmt + 1); | |
741 | form += p - strfrmt + 1; | |
742 | *form = '\0'; | |
743 | return p; | |
744 | } | |
745 | ||
746 | ||
747 | static void addintlen (char *form) { | |
748 | size_t l = strlen(form); | |
749 | char spec = form[l - 1]; | |
750 | strcpy(form + l - 1, LUA_INTFRMLEN); | |
751 | form[l + sizeof(LUA_INTFRMLEN) - 2] = spec; | |
752 | form[l + sizeof(LUA_INTFRMLEN) - 1] = '\0'; | |
753 | } | |
754 | ||
755 | ||
756 | static int str_format (lua_State *L) { | |
757 | int arg = 1; | |
758 | size_t sfl; | |
759 | const char *strfrmt = luaL_checklstring(L, arg, &sfl); | |
760 | const char *strfrmt_end = strfrmt+sfl; | |
761 | luaL_Buffer b; | |
762 | luaL_buffinit(L, &b); | |
763 | while (strfrmt < strfrmt_end) { | |
764 | if (*strfrmt != L_ESC) | |
765 | luaL_addchar(&b, *strfrmt++); | |
766 | else if (*++strfrmt == L_ESC) | |
767 | luaL_addchar(&b, *strfrmt++); /* %% */ | |
768 | else { /* format item */ | |
769 | char form[MAX_FORMAT]; /* to store the format (`%...') */ | |
770 | char buff[MAX_ITEM]; /* to store the formatted item */ | |
771 | arg++; | |
772 | strfrmt = scanformat(L, strfrmt, form); | |
773 | switch (*strfrmt++) { | |
774 | case 'c': { | |
775 | sprintf(buff, form, (int)luaL_checknumber(L, arg)); | |
776 | break; | |
777 | } | |
778 | case 'd': case 'i': { | |
779 | addintlen(form); | |
780 | sprintf(buff, form, (LUA_INTFRM_T)luaL_checknumber(L, arg)); | |
781 | break; | |
782 | } | |
783 | case 'o': case 'u': case 'x': case 'X': { | |
784 | addintlen(form); | |
785 | sprintf(buff, form, (unsigned LUA_INTFRM_T)luaL_checknumber(L, arg)); | |
786 | break; | |
787 | } | |
788 | case 'e': case 'E': case 'f': | |
789 | case 'g': case 'G': { | |
790 | sprintf(buff, form, (double)luaL_checknumber(L, arg)); | |
791 | break; | |
792 | } | |
793 | case 'q': { | |
794 | addquoted(L, &b, arg); | |
795 | continue; /* skip the 'addsize' at the end */ | |
796 | } | |
797 | case 's': { | |
798 | size_t l; | |
799 | const char *s = luaL_checklstring(L, arg, &l); | |
800 | if (!strchr(form, '.') && l >= 100) { | |
801 | /* no precision and string is too long to be formatted; | |
802 | keep original string */ | |
803 | lua_pushvalue(L, arg); | |
804 | luaL_addvalue(&b); | |
805 | continue; /* skip the `addsize' at the end */ | |
806 | } | |
807 | else { | |
808 | sprintf(buff, form, s); | |
809 | break; | |
810 | } | |
811 | } | |
812 | default: { /* also treat cases `pnLlh' */ | |
813 | return luaL_error(L, "invalid option " LUA_QL("%%%c") " to " | |
814 | LUA_QL("format"), *(strfrmt - 1)); | |
815 | } | |
816 | } | |
817 | luaL_addlstring(&b, buff, strlen(buff)); | |
818 | } | |
819 | } | |
820 | luaL_pushresult(&b); | |
821 | return 1; | |
822 | } | |
823 | ||
824 | ||
825 | static const luaL_Reg strlib[] = { | |
826 | {"byte", str_byte}, | |
827 | {"char", str_char}, | |
828 | {"dump", str_dump}, | |
829 | {"find", str_find}, | |
830 | {"format", str_format}, | |
831 | {"gfind", gfind_nodef}, | |
832 | {"gmatch", gmatch}, | |
833 | {"gsub", str_gsub}, | |
834 | {"len", str_len}, | |
835 | {"lower", str_lower}, | |
836 | {"match", str_match}, | |
837 | {"rep", str_rep}, | |
838 | {"reverse", str_reverse}, | |
839 | {"sub", str_sub}, | |
840 | {"upper", str_upper}, | |
841 | {NULL, NULL} | |
842 | }; | |
843 | ||
844 | ||
845 | static void createmetatable (lua_State *L) { | |
846 | lua_createtable(L, 0, 1); /* create metatable for strings */ | |
847 | lua_pushliteral(L, ""); /* dummy string */ | |
848 | lua_pushvalue(L, -2); | |
849 | lua_setmetatable(L, -2); /* set string metatable */ | |
850 | lua_pop(L, 1); /* pop dummy string */ | |
851 | lua_pushvalue(L, -2); /* string library... */ | |
852 | lua_setfield(L, -2, "__index"); /* ...is the __index metamethod */ | |
853 | lua_pop(L, 1); /* pop metatable */ | |
854 | } | |
855 | ||
856 | ||
857 | /* | |
858 | ** Open string library | |
859 | */ | |
860 | LUALIB_API int luaopen_string (lua_State *L) { | |
861 | luaL_register(L, LUA_STRLIBNAME, strlib); | |
862 | #if defined(LUA_COMPAT_GFIND) | |
863 | lua_getfield(L, -1, "gmatch"); | |
864 | lua_setfield(L, -2, "gfind"); | |
865 | #endif | |
866 | createmetatable(L); | |
867 | return 1; | |
868 | } | |
869 |