1 /* BEGIN CSTYLED */ 2 /* 3 ** $Id: lstrlib.c,v 1.178.1.1 2013/04/12 18:48:47 roberto Exp $ 4 ** Standard library for string operations and pattern-matching 5 ** See Copyright Notice in lua.h 6 */ 7 8 9 #define lstrlib_c 10 #define LUA_LIB 11 12 #include <sys/lua/lua.h> 13 14 #include <sys/lua/lauxlib.h> 15 #include <sys/lua/lualib.h> 16 17 18 /* 19 ** maximum number of captures that a pattern can do during 20 ** pattern-matching. This limit is arbitrary. 21 */ 22 #if !defined(LUA_MAXCAPTURES) 23 #define LUA_MAXCAPTURES 16 24 #endif 25 26 27 /* macro to `unsign' a character */ 28 #define uchar(c) ((unsigned char)(c)) 29 30 /* 31 * The provided version of sprintf returns a char *, but str_format expects 32 * it to return the number of characters printed. This version has the expected 33 * behavior. 34 */ 35 static size_t str_sprintf(char *buf, const char *fmt, ...) { 36 va_list args; 37 size_t len; 38 39 va_start(args, fmt); 40 len = vsnprintf(buf, INT_MAX, fmt, args); 41 va_end(args); 42 43 return len; 44 } 45 46 47 static int str_len (lua_State *L) { 48 size_t l; 49 luaL_checklstring(L, 1, &l); 50 lua_pushinteger(L, (lua_Integer)l); 51 return 1; 52 } 53 54 55 /* translate a relative string position: negative means back from end */ 56 static size_t posrelat (ptrdiff_t pos, size_t len) { 57 if (pos >= 0) return (size_t)pos; 58 else if (0u - (size_t)pos > len) return 0; 59 else return len - ((size_t)-pos) + 1; 60 } 61 62 63 static int str_sub (lua_State *L) { 64 size_t l; 65 const char *s = luaL_checklstring(L, 1, &l); 66 size_t start = posrelat(luaL_checkinteger(L, 2), l); 67 size_t end = posrelat(luaL_optinteger(L, 3, -1), l); 68 if (start < 1) start = 1; 69 if (end > l) end = l; 70 if (start <= end) 71 lua_pushlstring(L, s + start - 1, end - start + 1); 72 else lua_pushliteral(L, ""); 73 return 1; 74 } 75 76 77 static int str_reverse (lua_State *L) { 78 size_t l, i; 79 luaL_Buffer b; 80 const char *s = luaL_checklstring(L, 1, &l); 81 char *p = luaL_buffinitsize(L, &b, l); 82 for (i = 0; i < l; i++) 83 p[i] = s[l - i - 1]; 84 luaL_pushresultsize(&b, l); 85 return 1; 86 } 87 88 89 static int str_lower (lua_State *L) { 90 size_t l; 91 size_t i; 92 luaL_Buffer b; 93 const char *s = luaL_checklstring(L, 1, &l); 94 char *p = luaL_buffinitsize(L, &b, l); 95 for (i=0; i<l; i++) 96 p[i] = tolower(uchar(s[i])); 97 luaL_pushresultsize(&b, l); 98 return 1; 99 } 100 101 102 static int str_upper (lua_State *L) { 103 size_t l; 104 size_t i; 105 luaL_Buffer b; 106 const char *s = luaL_checklstring(L, 1, &l); 107 char *p = luaL_buffinitsize(L, &b, l); 108 for (i=0; i<l; i++) 109 p[i] = toupper(uchar(s[i])); 110 luaL_pushresultsize(&b, l); 111 return 1; 112 } 113 114 115 /* reasonable limit to avoid arithmetic overflow */ 116 #define MAXSIZE ((~(size_t)0) >> 1) 117 118 static int str_rep (lua_State *L) { 119 size_t l, lsep; 120 const char *s = luaL_checklstring(L, 1, &l); 121 int n = luaL_checkint(L, 2); 122 const char *sep = luaL_optlstring(L, 3, "", &lsep); 123 if (n <= 0) lua_pushliteral(L, ""); 124 else if (l + lsep < l || l + lsep >= MAXSIZE / n) /* may overflow? */ 125 return luaL_error(L, "resulting string too large"); 126 else { 127 size_t totallen = n * l + (n - 1) * lsep; 128 luaL_Buffer b; 129 char *p = luaL_buffinitsize(L, &b, totallen); 130 while (n-- > 1) { /* first n-1 copies (followed by separator) */ 131 memcpy(p, s, l * sizeof(char)); p += l; 132 if (lsep > 0) { /* avoid empty 'memcpy' (may be expensive) */ 133 memcpy(p, sep, lsep * sizeof(char)); p += lsep; 134 } 135 } 136 memcpy(p, s, l * sizeof(char)); /* last copy (not followed by separator) */ 137 luaL_pushresultsize(&b, totallen); 138 } 139 return 1; 140 } 141 142 143 static int str_byte (lua_State *L) { 144 size_t l; 145 const char *s = luaL_checklstring(L, 1, &l); 146 size_t posi = posrelat(luaL_optinteger(L, 2, 1), l); 147 size_t pose = posrelat(luaL_optinteger(L, 3, posi), l); 148 int n, i; 149 if (posi < 1) posi = 1; 150 if (pose > l) pose = l; 151 if (posi > pose) return 0; /* empty interval; return no values */ 152 n = (int)(pose - posi + 1); 153 if (posi + n <= pose) /* (size_t -> int) overflow? */ 154 return luaL_error(L, "string slice too long"); 155 luaL_checkstack(L, n, "string slice too long"); 156 for (i=0; i<n; i++) 157 lua_pushinteger(L, uchar(s[posi+i-1])); 158 return n; 159 } 160 161 162 static int str_char (lua_State *L) { 163 int n = lua_gettop(L); /* number of arguments */ 164 int i; 165 luaL_Buffer b; 166 char *p = luaL_buffinitsize(L, &b, n); 167 for (i=1; i<=n; i++) { 168 int c = luaL_checkint(L, i); 169 luaL_argcheck(L, uchar(c) == c, i, "value out of range"); 170 p[i - 1] = uchar(c); 171 } 172 luaL_pushresultsize(&b, n); 173 return 1; 174 } 175 176 177 #if defined(LUA_USE_DUMP) 178 static int writer (lua_State *L, const void* b, size_t size, void* B) { 179 (void)L; 180 luaL_addlstring((luaL_Buffer*) B, (const char *)b, size); 181 return 0; 182 } 183 184 185 static int str_dump (lua_State *L) { 186 luaL_Buffer b; 187 luaL_checktype(L, 1, LUA_TFUNCTION); 188 lua_settop(L, 1); 189 luaL_buffinit(L,&b); 190 if (lua_dump(L, writer, &b) != 0) 191 return luaL_error(L, "unable to dump given function"); 192 luaL_pushresult(&b); 193 return 1; 194 } 195 #endif 196 197 198 /* 199 ** {====================================================== 200 ** PATTERN MATCHING 201 ** ======================================================= 202 */ 203 204 205 #define CAP_UNFINISHED (-1) 206 #define CAP_POSITION (-2) 207 208 209 typedef struct MatchState { 210 int matchdepth; /* control for recursive depth (to avoid C stack overflow) */ 211 const char *src_init; /* init of source string */ 212 const char *src_end; /* end ('\0') of source string */ 213 const char *p_end; /* end ('\0') of pattern */ 214 lua_State *L; 215 int level; /* total number of captures (finished or unfinished) */ 216 struct { 217 const char *init; 218 ptrdiff_t len; 219 } capture[LUA_MAXCAPTURES]; 220 } MatchState; 221 222 223 /* recursive function */ 224 static const char *match (MatchState *ms, const char *s, const char *p); 225 226 227 /* maximum recursion depth for 'match' */ 228 #if !defined(MAXCCALLS) 229 #define MAXCCALLS 200 230 #endif 231 232 233 #define L_ESC '%' 234 #define SPECIALS "^$*+?.([%-" 235 236 237 static int check_capture (MatchState *ms, int l) { 238 l -= '1'; 239 if (l < 0 || l >= ms->level || ms->capture[l].len == CAP_UNFINISHED) 240 return luaL_error(ms->L, "invalid capture index %%%d", l + 1); 241 return l; 242 } 243 244 245 static int capture_to_close (MatchState *ms) { 246 int level = ms->level; 247 for (level--; level>=0; level--) 248 if (ms->capture[level].len == CAP_UNFINISHED) return level; 249 return luaL_error(ms->L, "invalid pattern capture"); 250 } 251 252 253 static const char *classend (MatchState *ms, const char *p) { 254 switch (*p++) { 255 case L_ESC: { 256 if (p == ms->p_end) 257 luaL_error(ms->L, "malformed pattern (ends with " LUA_QL("%%") ")"); 258 return p+1; 259 } 260 case '[': { 261 if (*p == '^') p++; 262 do { /* look for a `]' */ 263 if (p == ms->p_end) 264 luaL_error(ms->L, "malformed pattern (missing " LUA_QL("]") ")"); 265 if (*(p++) == L_ESC && p < ms->p_end) 266 p++; /* skip escapes (e.g. `%]') */ 267 } while (*p != ']'); 268 return p+1; 269 } 270 default: { 271 return p; 272 } 273 } 274 } 275 276 277 static int match_class (int c, int cl) { 278 int res; 279 switch (tolower(cl)) { 280 case 'a' : res = isalpha(c); break; 281 case 'c' : res = iscntrl(c); break; 282 case 'd' : res = isdigit(c); break; 283 case 'g' : res = isgraph(c); break; 284 case 'l' : res = islower(c); break; 285 case 'p' : res = ispunct(c); break; 286 case 's' : res = isspace(c); break; 287 case 'u' : res = isupper(c); break; 288 case 'w' : res = isalnum(c); break; 289 case 'x' : res = isxdigit(c); break; 290 case 'z' : res = (c == 0); break; /* deprecated option */ 291 default: return (cl == c); 292 } 293 return (islower(cl) ? res : !res); 294 } 295 296 297 static int matchbracketclass (int c, const char *p, const char *ec) { 298 int sig = 1; 299 if (*(p+1) == '^') { 300 sig = 0; 301 p++; /* skip the `^' */ 302 } 303 while (++p < ec) { 304 if (*p == L_ESC) { 305 p++; 306 if (match_class(c, uchar(*p))) 307 return sig; 308 } 309 else if ((*(p+1) == '-') && (p+2 < ec)) { 310 p+=2; 311 if (uchar(*(p-2)) <= c && c <= uchar(*p)) 312 return sig; 313 } 314 else if (uchar(*p) == c) return sig; 315 } 316 return !sig; 317 } 318 319 320 static int singlematch (MatchState *ms, const char *s, const char *p, 321 const char *ep) { 322 if (s >= ms->src_end) 323 return 0; 324 else { 325 int c = uchar(*s); 326 switch (*p) { 327 case '.': return 1; /* matches any char */ 328 case L_ESC: return match_class(c, uchar(*(p+1))); 329 case '[': return matchbracketclass(c, p, ep-1); 330 default: return (uchar(*p) == c); 331 } 332 } 333 } 334 335 336 static const char *matchbalance (MatchState *ms, const char *s, 337 const char *p) { 338 if (p >= ms->p_end - 1) 339 luaL_error(ms->L, "malformed pattern " 340 "(missing arguments to " LUA_QL("%%b") ")"); 341 if (*s != *p) return NULL; 342 else { 343 int b = *p; 344 int e = *(p+1); 345 int cont = 1; 346 while (++s < ms->src_end) { 347 if (*s == e) { 348 if (--cont == 0) return s+1; 349 } 350 else if (*s == b) cont++; 351 } 352 } 353 return NULL; /* string ends out of balance */ 354 } 355 356 357 static const char *max_expand (MatchState *ms, const char *s, 358 const char *p, const char *ep) { 359 ptrdiff_t i = 0; /* counts maximum expand for item */ 360 while (singlematch(ms, s + i, p, ep)) 361 i++; 362 /* keeps trying to match with the maximum repetitions */ 363 while (i>=0) { 364 const char *res = match(ms, (s+i), ep+1); 365 if (res) return res; 366 i--; /* else didn't match; reduce 1 repetition to try again */ 367 } 368 return NULL; 369 } 370 371 372 static const char *min_expand (MatchState *ms, const char *s, 373 const char *p, const char *ep) { 374 for (;;) { 375 const char *res = match(ms, s, ep+1); 376 if (res != NULL) 377 return res; 378 else if (singlematch(ms, s, p, ep)) 379 s++; /* try with one more repetition */ 380 else return NULL; 381 } 382 } 383 384 385 static const char *start_capture (MatchState *ms, const char *s, 386 const char *p, int what) { 387 const char *res; 388 int level = ms->level; 389 if (level >= LUA_MAXCAPTURES) luaL_error(ms->L, "too many captures"); 390 ms->capture[level].init = s; 391 ms->capture[level].len = what; 392 ms->level = level+1; 393 if ((res=match(ms, s, p)) == NULL) /* match failed? */ 394 ms->level--; /* undo capture */ 395 return res; 396 } 397 398 399 static const char *end_capture (MatchState *ms, const char *s, 400 const char *p) { 401 int l = capture_to_close(ms); 402 const char *res; 403 ms->capture[l].len = s - ms->capture[l].init; /* close capture */ 404 if ((res = match(ms, s, p)) == NULL) /* match failed? */ 405 ms->capture[l].len = CAP_UNFINISHED; /* undo capture */ 406 return res; 407 } 408 409 410 static const char *match_capture (MatchState *ms, const char *s, int l) { 411 size_t len; 412 l = check_capture(ms, l); 413 len = ms->capture[l].len; 414 if ((size_t)(ms->src_end-s) >= len && 415 memcmp(ms->capture[l].init, s, len) == 0) 416 return s+len; 417 else return NULL; 418 } 419 420 421 static const char *match (MatchState *ms, const char *s, const char *p) { 422 if (ms->matchdepth-- == 0) 423 luaL_error(ms->L, "pattern too complex"); 424 init: /* using goto's to optimize tail recursion */ 425 if (p != ms->p_end) { /* end of pattern? */ 426 switch (*p) { 427 case '(': { /* start capture */ 428 if (*(p + 1) == ')') /* position capture? */ 429 s = start_capture(ms, s, p + 2, CAP_POSITION); 430 else 431 s = start_capture(ms, s, p + 1, CAP_UNFINISHED); 432 break; 433 } 434 case ')': { /* end capture */ 435 s = end_capture(ms, s, p + 1); 436 break; 437 } 438 case '$': { 439 if ((p + 1) != ms->p_end) /* is the `$' the last char in pattern? */ 440 goto dflt; /* no; go to default */ 441 s = (s == ms->src_end) ? s : NULL; /* check end of string */ 442 break; 443 } 444 case L_ESC: { /* escaped sequences not in the format class[*+?-]? */ 445 switch (*(p + 1)) { 446 case 'b': { /* balanced string? */ 447 s = matchbalance(ms, s, p + 2); 448 if (s != NULL) { 449 p += 4; goto init; /* return match(ms, s, p + 4); */ 450 } /* else fail (s == NULL) */ 451 break; 452 } 453 case 'f': { /* frontier? */ 454 const char *ep; char previous; 455 p += 2; 456 if (*p != '[') 457 luaL_error(ms->L, "missing " LUA_QL("[") " after " 458 LUA_QL("%%f") " in pattern"); 459 ep = classend(ms, p); /* points to what is next */ 460 previous = (s == ms->src_init) ? '\0' : *(s - 1); 461 if (!matchbracketclass(uchar(previous), p, ep - 1) && 462 matchbracketclass(uchar(*s), p, ep - 1)) { 463 p = ep; goto init; /* return match(ms, s, ep); */ 464 } 465 s = NULL; /* match failed */ 466 break; 467 } 468 case '0': case '1': case '2': case '3': 469 case '4': case '5': case '6': case '7': 470 case '8': case '9': { /* capture results (%0-%9)? */ 471 s = match_capture(ms, s, uchar(*(p + 1))); 472 if (s != NULL) { 473 p += 2; goto init; /* return match(ms, s, p + 2) */ 474 } 475 break; 476 } 477 default: goto dflt; 478 } 479 break; 480 } 481 default: dflt: { /* pattern class plus optional suffix */ 482 const char *ep = classend(ms, p); /* points to optional suffix */ 483 /* does not match at least once? */ 484 if (!singlematch(ms, s, p, ep)) { 485 if (*ep == '*' || *ep == '?' || *ep == '-') { /* accept empty? */ 486 p = ep + 1; goto init; /* return match(ms, s, ep + 1); */ 487 } 488 else /* '+' or no suffix */ 489 s = NULL; /* fail */ 490 } 491 else { /* matched once */ 492 switch (*ep) { /* handle optional suffix */ 493 case '?': { /* optional */ 494 const char *res; 495 if ((res = match(ms, s + 1, ep + 1)) != NULL) 496 s = res; 497 else { 498 p = ep + 1; goto init; /* else return match(ms, s, ep + 1); */ 499 } 500 break; 501 } 502 case '+': /* 1 or more repetitions */ 503 s++; /* 1 match already done */ 504 /* FALLTHROUGH */ 505 case '*': /* 0 or more repetitions */ 506 s = max_expand(ms, s, p, ep); 507 break; 508 case '-': /* 0 or more repetitions (minimum) */ 509 s = min_expand(ms, s, p, ep); 510 break; 511 default: /* no suffix */ 512 s++; p = ep; goto init; /* return match(ms, s + 1, ep); */ 513 } 514 } 515 break; 516 } 517 } 518 } 519 ms->matchdepth++; 520 return s; 521 } 522 523 524 525 static const char *lmemfind (const char *s1, size_t l1, 526 const char *s2, size_t l2) { 527 if (l2 == 0) return s1; /* empty strings are everywhere */ 528 else if (l2 > l1) return NULL; /* avoids a negative `l1' */ 529 else { 530 const char *init; /* to search for a `*s2' inside `s1' */ 531 l2--; /* 1st char will be checked by `memchr' */ 532 l1 = l1-l2; /* `s2' cannot be found after that */ 533 while (l1 > 0 && (init = (const char *)memchr(s1, *s2, l1)) != NULL) { 534 init++; /* 1st char is already checked */ 535 if (memcmp(init, s2+1, l2) == 0) 536 return init-1; 537 else { /* correct `l1' and `s1' to try again */ 538 l1 -= init-s1; 539 s1 = init; 540 } 541 } 542 return NULL; /* not found */ 543 } 544 } 545 546 547 static void push_onecapture (MatchState *ms, int i, const char *s, 548 const char *e) { 549 if (i >= ms->level) { 550 if (i == 0) /* ms->level == 0, too */ 551 lua_pushlstring(ms->L, s, e - s); /* add whole match */ 552 else 553 luaL_error(ms->L, "invalid capture index"); 554 } 555 else { 556 ptrdiff_t l = ms->capture[i].len; 557 if (l == CAP_UNFINISHED) luaL_error(ms->L, "unfinished capture"); 558 if (l == CAP_POSITION) 559 lua_pushinteger(ms->L, ms->capture[i].init - ms->src_init + 1); 560 else 561 lua_pushlstring(ms->L, ms->capture[i].init, l); 562 } 563 } 564 565 566 static int push_captures (MatchState *ms, const char *s, const char *e) { 567 int i; 568 int nlevels = (ms->level == 0 && s) ? 1 : ms->level; 569 luaL_checkstack(ms->L, nlevels, "too many captures"); 570 for (i = 0; i < nlevels; i++) 571 push_onecapture(ms, i, s, e); 572 return nlevels; /* number of strings pushed */ 573 } 574 575 576 /* check whether pattern has no special characters */ 577 static int nospecials (const char *p, size_t l) { 578 size_t upto = 0; 579 do { 580 if (strpbrk(p + upto, SPECIALS)) 581 return 0; /* pattern has a special character */ 582 upto += strlen(p + upto) + 1; /* may have more after \0 */ 583 } while (upto <= l); 584 return 1; /* no special chars found */ 585 } 586 587 588 static int str_find_aux (lua_State *L, int find) { 589 size_t ls, lp; 590 const char *s = luaL_checklstring(L, 1, &ls); 591 const char *p = luaL_checklstring(L, 2, &lp); 592 size_t init = posrelat(luaL_optinteger(L, 3, 1), ls); 593 if (init < 1) init = 1; 594 else if (init > ls + 1) { /* start after string's end? */ 595 lua_pushnil(L); /* cannot find anything */ 596 return 1; 597 } 598 /* explicit request or no special characters? */ 599 if (find && (lua_toboolean(L, 4) || nospecials(p, lp))) { 600 /* do a plain search */ 601 const char *s2 = lmemfind(s + init - 1, ls - init + 1, p, lp); 602 if (s2) { 603 lua_pushinteger(L, s2 - s + 1); 604 lua_pushinteger(L, s2 - s + lp); 605 return 2; 606 } 607 } 608 else { 609 MatchState ms; 610 const char *s1 = s + init - 1; 611 int anchor = (*p == '^'); 612 if (anchor) { 613 p++; lp--; /* skip anchor character */ 614 } 615 ms.L = L; 616 ms.matchdepth = MAXCCALLS; 617 ms.src_init = s; 618 ms.src_end = s + ls; 619 ms.p_end = p + lp; 620 do { 621 const char *res; 622 ms.level = 0; 623 lua_assert(ms.matchdepth == MAXCCALLS); 624 if ((res=match(&ms, s1, p)) != NULL) { 625 if (find) { 626 lua_pushinteger(L, s1 - s + 1); /* start */ 627 lua_pushinteger(L, res - s); /* end */ 628 return push_captures(&ms, NULL, 0) + 2; 629 } 630 else 631 return push_captures(&ms, s1, res); 632 } 633 } while (s1++ < ms.src_end && !anchor); 634 } 635 lua_pushnil(L); /* not found */ 636 return 1; 637 } 638 639 640 static int str_find (lua_State *L) { 641 return str_find_aux(L, 1); 642 } 643 644 645 static int str_match (lua_State *L) { 646 return str_find_aux(L, 0); 647 } 648 649 650 static int gmatch_aux (lua_State *L) { 651 MatchState ms; 652 size_t ls, lp; 653 const char *s = lua_tolstring(L, lua_upvalueindex(1), &ls); 654 const char *p = lua_tolstring(L, lua_upvalueindex(2), &lp); 655 const char *src; 656 ms.L = L; 657 ms.matchdepth = MAXCCALLS; 658 ms.src_init = s; 659 ms.src_end = s+ls; 660 ms.p_end = p + lp; 661 for (src = s + (size_t)lua_tointeger(L, lua_upvalueindex(3)); 662 src <= ms.src_end; 663 src++) { 664 const char *e; 665 ms.level = 0; 666 lua_assert(ms.matchdepth == MAXCCALLS); 667 if ((e = match(&ms, src, p)) != NULL) { 668 lua_Integer newstart = e-s; 669 if (e == src) newstart++; /* empty match? go at least one position */ 670 lua_pushinteger(L, newstart); 671 lua_replace(L, lua_upvalueindex(3)); 672 return push_captures(&ms, src, e); 673 } 674 } 675 return 0; /* not found */ 676 } 677 678 679 static int str_gmatch (lua_State *L) { 680 luaL_checkstring(L, 1); 681 luaL_checkstring(L, 2); 682 lua_settop(L, 2); 683 lua_pushinteger(L, 0); 684 lua_pushcclosure(L, gmatch_aux, 3); 685 return 1; 686 } 687 688 689 static void add_s (MatchState *ms, luaL_Buffer *b, const char *s, 690 const char *e) { 691 size_t l, i; 692 const char *news = lua_tolstring(ms->L, 3, &l); 693 for (i = 0; i < l; i++) { 694 if (news[i] != L_ESC) 695 luaL_addchar(b, news[i]); 696 else { 697 i++; /* skip ESC */ 698 if (!isdigit(uchar(news[i]))) { 699 if (news[i] != L_ESC) 700 luaL_error(ms->L, "invalid use of " LUA_QL("%c") 701 " in replacement string", L_ESC); 702 luaL_addchar(b, news[i]); 703 } 704 else if (news[i] == '0') 705 luaL_addlstring(b, s, e - s); 706 else { 707 push_onecapture(ms, news[i] - '1', s, e); 708 luaL_addvalue(b); /* add capture to accumulated result */ 709 } 710 } 711 } 712 } 713 714 715 static void add_value (MatchState *ms, luaL_Buffer *b, const char *s, 716 const char *e, int tr) { 717 lua_State *L = ms->L; 718 switch (tr) { 719 case LUA_TFUNCTION: { 720 int n; 721 lua_pushvalue(L, 3); 722 n = push_captures(ms, s, e); 723 lua_call(L, n, 1); 724 break; 725 } 726 case LUA_TTABLE: { 727 push_onecapture(ms, 0, s, e); 728 lua_gettable(L, 3); 729 break; 730 } 731 default: { /* LUA_TNUMBER or LUA_TSTRING */ 732 add_s(ms, b, s, e); 733 return; 734 } 735 } 736 if (!lua_toboolean(L, -1)) { /* nil or false? */ 737 lua_pop(L, 1); 738 lua_pushlstring(L, s, e - s); /* keep original text */ 739 } 740 else if (!lua_isstring(L, -1)) 741 luaL_error(L, "invalid replacement value (a %s)", luaL_typename(L, -1)); 742 luaL_addvalue(b); /* add result to accumulator */ 743 } 744 745 746 static int str_gsub (lua_State *L) { 747 size_t srcl, lp; 748 const char *src = luaL_checklstring(L, 1, &srcl); 749 const char *p = luaL_checklstring(L, 2, &lp); 750 int tr = lua_type(L, 3); 751 size_t max_s = luaL_optinteger(L, 4, srcl+1); 752 int anchor = (*p == '^'); 753 size_t n = 0; 754 MatchState ms; 755 luaL_Buffer b; 756 luaL_argcheck(L, tr == LUA_TNUMBER || tr == LUA_TSTRING || 757 tr == LUA_TFUNCTION || tr == LUA_TTABLE, 3, 758 "string/function/table expected"); 759 luaL_buffinit(L, &b); 760 if (anchor) { 761 p++; lp--; /* skip anchor character */ 762 } 763 ms.L = L; 764 ms.matchdepth = MAXCCALLS; 765 ms.src_init = src; 766 ms.src_end = src+srcl; 767 ms.p_end = p + lp; 768 while (n < max_s) { 769 const char *e; 770 ms.level = 0; 771 lua_assert(ms.matchdepth == MAXCCALLS); 772 e = match(&ms, src, p); 773 if (e) { 774 n++; 775 add_value(&ms, &b, src, e, tr); 776 } 777 if (e && e>src) /* non empty match? */ 778 src = e; /* skip it */ 779 else if (src < ms.src_end) 780 luaL_addchar(&b, *src++); 781 else break; 782 if (anchor) break; 783 } 784 luaL_addlstring(&b, src, ms.src_end-src); 785 luaL_pushresult(&b); 786 lua_pushinteger(L, n); /* number of substitutions */ 787 return 2; 788 } 789 790 /* }====================================================== */ 791 792 793 794 /* 795 ** {====================================================== 796 ** STRING FORMAT 797 ** ======================================================= 798 */ 799 800 /* 801 ** LUA_INTFRMLEN is the length modifier for integer conversions in 802 ** 'string.format'; LUA_INTFRM_T is the integer type corresponding to 803 ** the previous length 804 */ 805 #if !defined(LUA_INTFRMLEN) /* { */ 806 #if defined(LUA_USE_LONGLONG) 807 808 #define LUA_INTFRMLEN "ll" 809 #define LUA_INTFRM_T long long 810 811 #else 812 813 #define LUA_INTFRMLEN "l" 814 #define LUA_INTFRM_T long 815 816 #endif 817 #endif /* } */ 818 819 820 /* 821 ** LUA_FLTFRMLEN is the length modifier for float conversions in 822 ** 'string.format'; LUA_FLTFRM_T is the float type corresponding to 823 ** the previous length 824 */ 825 #if !defined(LUA_FLTFRMLEN) 826 827 #define LUA_FLTFRMLEN "" 828 #define LUA_FLTFRM_T double 829 830 #endif 831 832 833 /* maximum size of each formatted item (> len(format('%99.99f', -1e308))) */ 834 #define MAX_ITEM 512 835 /* valid flags in a format specification */ 836 #define FLAGS "-+ #0" 837 /* 838 ** maximum size of each format specification (such as '%-099.99d') 839 ** (+10 accounts for %99.99x plus margin of error) 840 */ 841 #define MAX_FORMAT (sizeof(FLAGS) + sizeof(LUA_INTFRMLEN) + 10) 842 843 844 static void addquoted (lua_State *L, luaL_Buffer *b, int arg) { 845 size_t l; 846 const char *s = luaL_checklstring(L, arg, &l); 847 luaL_addchar(b, '"'); 848 while (l--) { 849 if (*s == '"' || *s == '\\' || *s == '\n') { 850 luaL_addchar(b, '\\'); 851 luaL_addchar(b, *s); 852 } 853 else if (*s == '\0' || iscntrl(uchar(*s))) { 854 char buff[10]; 855 if (!isdigit(uchar(*(s+1)))) 856 snprintf(buff, sizeof(buff), "\\%d", (int)uchar(*s)); 857 else 858 snprintf(buff, sizeof(buff), "\\%03d", (int)uchar(*s)); 859 luaL_addstring(b, buff); 860 } 861 else 862 luaL_addchar(b, *s); 863 s++; 864 } 865 luaL_addchar(b, '"'); 866 } 867 868 static const char *scanformat (lua_State *L, const char *strfrmt, char *form) { 869 const char *p = strfrmt; 870 while (*p != '\0' && strchr(FLAGS, *p) != NULL) p++; /* skip flags */ 871 if ((size_t)(p - strfrmt) >= sizeof(FLAGS)/sizeof(char)) 872 luaL_error(L, "invalid format (repeated flags)"); 873 if (isdigit(uchar(*p))) p++; /* skip width */ 874 if (isdigit(uchar(*p))) p++; /* (2 digits at most) */ 875 if (*p == '.') { 876 p++; 877 if (isdigit(uchar(*p))) p++; /* skip precision */ 878 if (isdigit(uchar(*p))) p++; /* (2 digits at most) */ 879 } 880 if (isdigit(uchar(*p))) 881 luaL_error(L, "invalid format (width or precision too long)"); 882 *(form++) = '%'; 883 memcpy(form, strfrmt, (p - strfrmt + 1) * sizeof(char)); 884 form += p - strfrmt + 1; 885 *form = '\0'; 886 return p; 887 } 888 889 890 /* 891 ** add length modifier into formats 892 */ 893 static void addlenmod (char *form, const char *lenmod, size_t size) { 894 size_t l = strlen(form); 895 size_t lm = strlen(lenmod); 896 char spec = form[l - 1]; 897 strlcpy(form + l - 1, lenmod, size - (l - 1)); 898 form[l + lm - 1] = spec; 899 form[l + lm] = '\0'; 900 } 901 902 903 static int str_format (lua_State *L) { 904 int top = lua_gettop(L); 905 int arg = 1; 906 size_t sfl; 907 const char *strfrmt = luaL_checklstring(L, arg, &sfl); 908 const char *strfrmt_end = strfrmt+sfl; 909 luaL_Buffer b; 910 luaL_buffinit(L, &b); 911 while (strfrmt < strfrmt_end) { 912 if (*strfrmt != L_ESC) 913 luaL_addchar(&b, *strfrmt++); 914 else if (*++strfrmt == L_ESC) 915 luaL_addchar(&b, *strfrmt++); /* %% */ 916 else { /* format item */ 917 char form[MAX_FORMAT]; /* to store the format (`%...') */ 918 char *buff = luaL_prepbuffsize(&b, MAX_ITEM); /* to put formatted item */ 919 int nb = 0; /* number of bytes in added item */ 920 if (++arg > top) 921 luaL_argerror(L, arg, "no value"); 922 strfrmt = scanformat(L, strfrmt, form); 923 switch (*strfrmt++) { 924 case 'c': { 925 nb = str_sprintf(buff, form, luaL_checkint(L, arg)); 926 break; 927 } 928 case 'd': case 'i': { 929 lua_Number n = luaL_checknumber(L, arg); 930 LUA_INTFRM_T ni = (LUA_INTFRM_T)n; 931 lua_Number diff = n - (lua_Number)ni; 932 luaL_argcheck(L, -1 < diff && diff < 1, arg, 933 "not a number in proper range"); 934 addlenmod(form, LUA_INTFRMLEN, MAX_FORMAT); 935 nb = str_sprintf(buff, form, ni); 936 break; 937 } 938 case 'o': case 'u': case 'x': case 'X': { 939 lua_Number n = luaL_checknumber(L, arg); 940 unsigned LUA_INTFRM_T ni = (unsigned LUA_INTFRM_T)n; 941 lua_Number diff = n - (lua_Number)ni; 942 luaL_argcheck(L, -1 < diff && diff < 1, arg, 943 "not a non-negative number in proper range"); 944 addlenmod(form, LUA_INTFRMLEN, MAX_FORMAT); 945 nb = str_sprintf(buff, form, ni); 946 break; 947 } 948 #if defined(LUA_USE_FLOAT_FORMATS) 949 case 'e': case 'E': case 'f': 950 #if defined(LUA_USE_AFORMAT) 951 case 'a': case 'A': 952 #endif 953 case 'g': case 'G': { 954 addlenmod(form, LUA_FLTFRMLEN, MAX_FORMAT); 955 nb = str_sprintf(buff, form, (LUA_FLTFRM_T)luaL_checknumber(L, arg)); 956 break; 957 } 958 #endif 959 case 'q': { 960 addquoted(L, &b, arg); 961 break; 962 } 963 case 's': { 964 size_t l; 965 const char *s = luaL_tolstring(L, arg, &l); 966 if (!strchr(form, '.') && l >= 100) { 967 /* no precision and string is too long to be formatted; 968 keep original string */ 969 luaL_addvalue(&b); 970 break; 971 } 972 else { 973 nb = str_sprintf(buff, form, s); 974 lua_pop(L, 1); /* remove result from 'luaL_tolstring' */ 975 break; 976 } 977 } 978 default: { /* also treat cases `pnLlh' */ 979 return luaL_error(L, "invalid option " LUA_QL("%%%c") " to " 980 LUA_QL("format"), *(strfrmt - 1)); 981 } 982 } 983 luaL_addsize(&b, nb); 984 } 985 } 986 luaL_pushresult(&b); 987 return 1; 988 } 989 990 /* }====================================================== */ 991 992 993 static const luaL_Reg strlib[] = { 994 {"byte", str_byte}, 995 {"char", str_char}, 996 #if defined(LUA_USE_DUMP) 997 {"dump", str_dump}, 998 #endif 999 {"find", str_find}, 1000 {"format", str_format}, 1001 {"gmatch", str_gmatch}, 1002 {"gsub", str_gsub}, 1003 {"len", str_len}, 1004 {"lower", str_lower}, 1005 {"match", str_match}, 1006 {"rep", str_rep}, 1007 {"reverse", str_reverse}, 1008 {"sub", str_sub}, 1009 {"upper", str_upper}, 1010 {NULL, NULL} 1011 }; 1012 1013 1014 static void createmetatable (lua_State *L) { 1015 lua_createtable(L, 0, 1); /* table to be metatable for strings */ 1016 lua_pushliteral(L, ""); /* dummy string */ 1017 lua_pushvalue(L, -2); /* copy table */ 1018 lua_setmetatable(L, -2); /* set table as metatable for strings */ 1019 lua_pop(L, 1); /* pop dummy string */ 1020 lua_pushvalue(L, -2); /* get string library */ 1021 lua_setfield(L, -2, "__index"); /* metatable.__index = string */ 1022 lua_pop(L, 1); /* pop metatable */ 1023 } 1024 1025 1026 /* 1027 ** Open string library 1028 */ 1029 LUAMOD_API int luaopen_string (lua_State *L) { 1030 luaL_newlib(L, strlib); 1031 createmetatable(L); 1032 return 1; 1033 } 1034 1035 #if defined(_KERNEL) 1036 1037 EXPORT_SYMBOL(luaopen_string); 1038 1039 #endif 1040 /* END CSTYLED */ 1041