diff src/lua/lstrlib.c @ 11:27763b933818

raise lua sources up one level
author Robert McIntyre <rlm@mit.edu>
date Sat, 03 Mar 2012 11:07:39 -0600
parents src/lua/src/lstrlib.c@f9f4f1b99eed
children
line wrap: on
line diff
     1.1 --- /dev/null	Thu Jan 01 00:00:00 1970 +0000
     1.2 +++ b/src/lua/lstrlib.c	Sat Mar 03 11:07:39 2012 -0600
     1.3 @@ -0,0 +1,869 @@
     1.4 +/*
     1.5 +** $Id: lstrlib.c,v 1.132.1.4 2008/07/11 17:27:21 roberto Exp $
     1.6 +** Standard library for string operations and pattern-matching
     1.7 +** See Copyright Notice in lua.h
     1.8 +*/
     1.9 +
    1.10 +
    1.11 +#include <ctype.h>
    1.12 +#include <stddef.h>
    1.13 +#include <stdio.h>
    1.14 +#include <stdlib.h>
    1.15 +#include <string.h>
    1.16 +
    1.17 +#define lstrlib_c
    1.18 +#define LUA_LIB
    1.19 +
    1.20 +#include "lua.h"
    1.21 +
    1.22 +#include "lauxlib.h"
    1.23 +#include "lualib.h"
    1.24 +
    1.25 +
    1.26 +/* macro to `unsign' a character */
    1.27 +#define uchar(c)        ((unsigned char)(c))
    1.28 +
    1.29 +
    1.30 +
    1.31 +static int str_len (lua_State *L) {
    1.32 +  size_t l;
    1.33 +  luaL_checklstring(L, 1, &l);
    1.34 +  lua_pushinteger(L, l);
    1.35 +  return 1;
    1.36 +}
    1.37 +
    1.38 +
    1.39 +static ptrdiff_t posrelat (ptrdiff_t pos, size_t len) {
    1.40 +  /* relative string position: negative means back from end */
    1.41 +  if (pos < 0) pos += (ptrdiff_t)len + 1;
    1.42 +  return (pos >= 0) ? pos : 0;
    1.43 +}
    1.44 +
    1.45 +
    1.46 +static int str_sub (lua_State *L) {
    1.47 +  size_t l;
    1.48 +  const char *s = luaL_checklstring(L, 1, &l);
    1.49 +  ptrdiff_t start = posrelat(luaL_checkinteger(L, 2), l);
    1.50 +  ptrdiff_t end = posrelat(luaL_optinteger(L, 3, -1), l);
    1.51 +  if (start < 1) start = 1;
    1.52 +  if (end > (ptrdiff_t)l) end = (ptrdiff_t)l;
    1.53 +  if (start <= end)
    1.54 +    lua_pushlstring(L, s+start-1, end-start+1);
    1.55 +  else lua_pushliteral(L, "");
    1.56 +  return 1;
    1.57 +}
    1.58 +
    1.59 +
    1.60 +static int str_reverse (lua_State *L) {
    1.61 +  size_t l;
    1.62 +  luaL_Buffer b;
    1.63 +  const char *s = luaL_checklstring(L, 1, &l);
    1.64 +  luaL_buffinit(L, &b);
    1.65 +  while (l--) luaL_addchar(&b, s[l]);
    1.66 +  luaL_pushresult(&b);
    1.67 +  return 1;
    1.68 +}
    1.69 +
    1.70 +
    1.71 +static int str_lower (lua_State *L) {
    1.72 +  size_t l;
    1.73 +  size_t i;
    1.74 +  luaL_Buffer b;
    1.75 +  const char *s = luaL_checklstring(L, 1, &l);
    1.76 +  luaL_buffinit(L, &b);
    1.77 +  for (i=0; i<l; i++)
    1.78 +    luaL_addchar(&b, tolower(uchar(s[i])));
    1.79 +  luaL_pushresult(&b);
    1.80 +  return 1;
    1.81 +}
    1.82 +
    1.83 +
    1.84 +static int str_upper (lua_State *L) {
    1.85 +  size_t l;
    1.86 +  size_t i;
    1.87 +  luaL_Buffer b;
    1.88 +  const char *s = luaL_checklstring(L, 1, &l);
    1.89 +  luaL_buffinit(L, &b);
    1.90 +  for (i=0; i<l; i++)
    1.91 +    luaL_addchar(&b, toupper(uchar(s[i])));
    1.92 +  luaL_pushresult(&b);
    1.93 +  return 1;
    1.94 +}
    1.95 +
    1.96 +static int str_rep (lua_State *L) {
    1.97 +  size_t l;
    1.98 +  luaL_Buffer b;
    1.99 +  const char *s = luaL_checklstring(L, 1, &l);
   1.100 +  int n = luaL_checkint(L, 2);
   1.101 +  luaL_buffinit(L, &b);
   1.102 +  while (n-- > 0)
   1.103 +    luaL_addlstring(&b, s, l);
   1.104 +  luaL_pushresult(&b);
   1.105 +  return 1;
   1.106 +}
   1.107 +
   1.108 +
   1.109 +static int str_byte (lua_State *L) {
   1.110 +  size_t l;
   1.111 +  const char *s = luaL_checklstring(L, 1, &l);
   1.112 +  ptrdiff_t posi = posrelat(luaL_optinteger(L, 2, 1), l);
   1.113 +  ptrdiff_t pose = posrelat(luaL_optinteger(L, 3, posi), l);
   1.114 +  int n, i;
   1.115 +  if (posi <= 0) posi = 1;
   1.116 +  if ((size_t)pose > l) pose = l;
   1.117 +  if (posi > pose) return 0;  /* empty interval; return no values */
   1.118 +  n = (int)(pose -  posi + 1);
   1.119 +  if (posi + n <= pose)  /* overflow? */
   1.120 +    luaL_error(L, "string slice too long");
   1.121 +  luaL_checkstack(L, n, "string slice too long");
   1.122 +  for (i=0; i<n; i++)
   1.123 +    lua_pushinteger(L, uchar(s[posi+i-1]));
   1.124 +  return n;
   1.125 +}
   1.126 +
   1.127 +
   1.128 +static int str_char (lua_State *L) {
   1.129 +  int n = lua_gettop(L);  /* number of arguments */
   1.130 +  int i;
   1.131 +  luaL_Buffer b;
   1.132 +  luaL_buffinit(L, &b);
   1.133 +  for (i=1; i<=n; i++) {
   1.134 +    int c = luaL_checkint(L, i);
   1.135 +    luaL_argcheck(L, uchar(c) == c, i, "invalid value");
   1.136 +    luaL_addchar(&b, uchar(c));
   1.137 +  }
   1.138 +  luaL_pushresult(&b);
   1.139 +  return 1;
   1.140 +}
   1.141 +
   1.142 +
   1.143 +static int writer (lua_State *L, const void* b, size_t size, void* B) {
   1.144 +  (void)L;
   1.145 +  luaL_addlstring((luaL_Buffer*) B, (const char *)b, size);
   1.146 +  return 0;
   1.147 +}
   1.148 +
   1.149 +
   1.150 +static int str_dump (lua_State *L) {
   1.151 +  luaL_Buffer b;
   1.152 +  luaL_checktype(L, 1, LUA_TFUNCTION);
   1.153 +  lua_settop(L, 1);
   1.154 +  luaL_buffinit(L,&b);
   1.155 +  if (lua_dump(L, writer, &b) != 0)
   1.156 +    luaL_error(L, "unable to dump given function");
   1.157 +  luaL_pushresult(&b);
   1.158 +  return 1;
   1.159 +}
   1.160 +
   1.161 +
   1.162 +
   1.163 +/*
   1.164 +** {======================================================
   1.165 +** PATTERN MATCHING
   1.166 +** =======================================================
   1.167 +*/
   1.168 +
   1.169 +
   1.170 +#define CAP_UNFINISHED	(-1)
   1.171 +#define CAP_POSITION	(-2)
   1.172 +
   1.173 +typedef struct MatchState {
   1.174 +  const char *src_init;  /* init of source string */
   1.175 +  const char *src_end;  /* end (`\0') of source string */
   1.176 +  lua_State *L;
   1.177 +  int level;  /* total number of captures (finished or unfinished) */
   1.178 +  struct {
   1.179 +    const char *init;
   1.180 +    ptrdiff_t len;
   1.181 +  } capture[LUA_MAXCAPTURES];
   1.182 +} MatchState;
   1.183 +
   1.184 +
   1.185 +#define L_ESC		'%'
   1.186 +#define SPECIALS	"^$*+?.([%-"
   1.187 +
   1.188 +
   1.189 +static int check_capture (MatchState *ms, int l) {
   1.190 +  l -= '1';
   1.191 +  if (l < 0 || l >= ms->level || ms->capture[l].len == CAP_UNFINISHED)
   1.192 +    return luaL_error(ms->L, "invalid capture index");
   1.193 +  return l;
   1.194 +}
   1.195 +
   1.196 +
   1.197 +static int capture_to_close (MatchState *ms) {
   1.198 +  int level = ms->level;
   1.199 +  for (level--; level>=0; level--)
   1.200 +    if (ms->capture[level].len == CAP_UNFINISHED) return level;
   1.201 +  return luaL_error(ms->L, "invalid pattern capture");
   1.202 +}
   1.203 +
   1.204 +
   1.205 +static const char *classend (MatchState *ms, const char *p) {
   1.206 +  switch (*p++) {
   1.207 +    case L_ESC: {
   1.208 +      if (*p == '\0')
   1.209 +        luaL_error(ms->L, "malformed pattern (ends with " LUA_QL("%%") ")");
   1.210 +      return p+1;
   1.211 +    }
   1.212 +    case '[': {
   1.213 +      if (*p == '^') p++;
   1.214 +      do {  /* look for a `]' */
   1.215 +        if (*p == '\0')
   1.216 +          luaL_error(ms->L, "malformed pattern (missing " LUA_QL("]") ")");
   1.217 +        if (*(p++) == L_ESC && *p != '\0')
   1.218 +          p++;  /* skip escapes (e.g. `%]') */
   1.219 +      } while (*p != ']');
   1.220 +      return p+1;
   1.221 +    }
   1.222 +    default: {
   1.223 +      return p;
   1.224 +    }
   1.225 +  }
   1.226 +}
   1.227 +
   1.228 +
   1.229 +static int match_class (int c, int cl) {
   1.230 +  int res;
   1.231 +  switch (tolower(cl)) {
   1.232 +    case 'a' : res = isalpha(c); break;
   1.233 +    case 'c' : res = iscntrl(c); break;
   1.234 +    case 'd' : res = isdigit(c); break;
   1.235 +    case 'l' : res = islower(c); break;
   1.236 +    case 'p' : res = ispunct(c); break;
   1.237 +    case 's' : res = isspace(c); break;
   1.238 +    case 'u' : res = isupper(c); break;
   1.239 +    case 'w' : res = isalnum(c); break;
   1.240 +    case 'x' : res = isxdigit(c); break;
   1.241 +    case 'z' : res = (c == 0); break;
   1.242 +    default: return (cl == c);
   1.243 +  }
   1.244 +  return (islower(cl) ? res : !res);
   1.245 +}
   1.246 +
   1.247 +
   1.248 +static int matchbracketclass (int c, const char *p, const char *ec) {
   1.249 +  int sig = 1;
   1.250 +  if (*(p+1) == '^') {
   1.251 +    sig = 0;
   1.252 +    p++;  /* skip the `^' */
   1.253 +  }
   1.254 +  while (++p < ec) {
   1.255 +    if (*p == L_ESC) {
   1.256 +      p++;
   1.257 +      if (match_class(c, uchar(*p)))
   1.258 +        return sig;
   1.259 +    }
   1.260 +    else if ((*(p+1) == '-') && (p+2 < ec)) {
   1.261 +      p+=2;
   1.262 +      if (uchar(*(p-2)) <= c && c <= uchar(*p))
   1.263 +        return sig;
   1.264 +    }
   1.265 +    else if (uchar(*p) == c) return sig;
   1.266 +  }
   1.267 +  return !sig;
   1.268 +}
   1.269 +
   1.270 +
   1.271 +static int singlematch (int c, const char *p, const char *ep) {
   1.272 +  switch (*p) {
   1.273 +    case '.': return 1;  /* matches any char */
   1.274 +    case L_ESC: return match_class(c, uchar(*(p+1)));
   1.275 +    case '[': return matchbracketclass(c, p, ep-1);
   1.276 +    default:  return (uchar(*p) == c);
   1.277 +  }
   1.278 +}
   1.279 +
   1.280 +
   1.281 +static const char *match (MatchState *ms, const char *s, const char *p);
   1.282 +
   1.283 +
   1.284 +static const char *matchbalance (MatchState *ms, const char *s,
   1.285 +                                   const char *p) {
   1.286 +  if (*p == 0 || *(p+1) == 0)
   1.287 +    luaL_error(ms->L, "unbalanced pattern");
   1.288 +  if (*s != *p) return NULL;
   1.289 +  else {
   1.290 +    int b = *p;
   1.291 +    int e = *(p+1);
   1.292 +    int cont = 1;
   1.293 +    while (++s < ms->src_end) {
   1.294 +      if (*s == e) {
   1.295 +        if (--cont == 0) return s+1;
   1.296 +      }
   1.297 +      else if (*s == b) cont++;
   1.298 +    }
   1.299 +  }
   1.300 +  return NULL;  /* string ends out of balance */
   1.301 +}
   1.302 +
   1.303 +
   1.304 +static const char *max_expand (MatchState *ms, const char *s,
   1.305 +                                 const char *p, const char *ep) {
   1.306 +  ptrdiff_t i = 0;  /* counts maximum expand for item */
   1.307 +  while ((s+i)<ms->src_end && singlematch(uchar(*(s+i)), p, ep))
   1.308 +    i++;
   1.309 +  /* keeps trying to match with the maximum repetitions */
   1.310 +  while (i>=0) {
   1.311 +    const char *res = match(ms, (s+i), ep+1);
   1.312 +    if (res) return res;
   1.313 +    i--;  /* else didn't match; reduce 1 repetition to try again */
   1.314 +  }
   1.315 +  return NULL;
   1.316 +}
   1.317 +
   1.318 +
   1.319 +static const char *min_expand (MatchState *ms, const char *s,
   1.320 +                                 const char *p, const char *ep) {
   1.321 +  for (;;) {
   1.322 +    const char *res = match(ms, s, ep+1);
   1.323 +    if (res != NULL)
   1.324 +      return res;
   1.325 +    else if (s<ms->src_end && singlematch(uchar(*s), p, ep))
   1.326 +      s++;  /* try with one more repetition */
   1.327 +    else return NULL;
   1.328 +  }
   1.329 +}
   1.330 +
   1.331 +
   1.332 +static const char *start_capture (MatchState *ms, const char *s,
   1.333 +                                    const char *p, int what) {
   1.334 +  const char *res;
   1.335 +  int level = ms->level;
   1.336 +  if (level >= LUA_MAXCAPTURES) luaL_error(ms->L, "too many captures");
   1.337 +  ms->capture[level].init = s;
   1.338 +  ms->capture[level].len = what;
   1.339 +  ms->level = level+1;
   1.340 +  if ((res=match(ms, s, p)) == NULL)  /* match failed? */
   1.341 +    ms->level--;  /* undo capture */
   1.342 +  return res;
   1.343 +}
   1.344 +
   1.345 +
   1.346 +static const char *end_capture (MatchState *ms, const char *s,
   1.347 +                                  const char *p) {
   1.348 +  int l = capture_to_close(ms);
   1.349 +  const char *res;
   1.350 +  ms->capture[l].len = s - ms->capture[l].init;  /* close capture */
   1.351 +  if ((res = match(ms, s, p)) == NULL)  /* match failed? */
   1.352 +    ms->capture[l].len = CAP_UNFINISHED;  /* undo capture */
   1.353 +  return res;
   1.354 +}
   1.355 +
   1.356 +
   1.357 +static const char *match_capture (MatchState *ms, const char *s, int l) {
   1.358 +  size_t len;
   1.359 +  l = check_capture(ms, l);
   1.360 +  len = ms->capture[l].len;
   1.361 +  if ((size_t)(ms->src_end-s) >= len &&
   1.362 +      memcmp(ms->capture[l].init, s, len) == 0)
   1.363 +    return s+len;
   1.364 +  else return NULL;
   1.365 +}
   1.366 +
   1.367 +
   1.368 +static const char *match (MatchState *ms, const char *s, const char *p) {
   1.369 +  init: /* using goto's to optimize tail recursion */
   1.370 +  switch (*p) {
   1.371 +    case '(': {  /* start capture */
   1.372 +      if (*(p+1) == ')')  /* position capture? */
   1.373 +        return start_capture(ms, s, p+2, CAP_POSITION);
   1.374 +      else
   1.375 +        return start_capture(ms, s, p+1, CAP_UNFINISHED);
   1.376 +    }
   1.377 +    case ')': {  /* end capture */
   1.378 +      return end_capture(ms, s, p+1);
   1.379 +    }
   1.380 +    case L_ESC: {
   1.381 +      switch (*(p+1)) {
   1.382 +        case 'b': {  /* balanced string? */
   1.383 +          s = matchbalance(ms, s, p+2);
   1.384 +          if (s == NULL) return NULL;
   1.385 +          p+=4; goto init;  /* else return match(ms, s, p+4); */
   1.386 +        }
   1.387 +        case 'f': {  /* frontier? */
   1.388 +          const char *ep; char previous;
   1.389 +          p += 2;
   1.390 +          if (*p != '[')
   1.391 +            luaL_error(ms->L, "missing " LUA_QL("[") " after "
   1.392 +                               LUA_QL("%%f") " in pattern");
   1.393 +          ep = classend(ms, p);  /* points to what is next */
   1.394 +          previous = (s == ms->src_init) ? '\0' : *(s-1);
   1.395 +          if (matchbracketclass(uchar(previous), p, ep-1) ||
   1.396 +             !matchbracketclass(uchar(*s), p, ep-1)) return NULL;
   1.397 +          p=ep; goto init;  /* else return match(ms, s, ep); */
   1.398 +        }
   1.399 +        default: {
   1.400 +          if (isdigit(uchar(*(p+1)))) {  /* capture results (%0-%9)? */
   1.401 +            s = match_capture(ms, s, uchar(*(p+1)));
   1.402 +            if (s == NULL) return NULL;
   1.403 +            p+=2; goto init;  /* else return match(ms, s, p+2) */
   1.404 +          }
   1.405 +          goto dflt;  /* case default */
   1.406 +        }
   1.407 +      }
   1.408 +    }
   1.409 +    case '\0': {  /* end of pattern */
   1.410 +      return s;  /* match succeeded */
   1.411 +    }
   1.412 +    case '$': {
   1.413 +      if (*(p+1) == '\0')  /* is the `$' the last char in pattern? */
   1.414 +        return (s == ms->src_end) ? s : NULL;  /* check end of string */
   1.415 +      else goto dflt;
   1.416 +    }
   1.417 +    default: dflt: {  /* it is a pattern item */
   1.418 +      const char *ep = classend(ms, p);  /* points to what is next */
   1.419 +      int m = s<ms->src_end && singlematch(uchar(*s), p, ep);
   1.420 +      switch (*ep) {
   1.421 +        case '?': {  /* optional */
   1.422 +          const char *res;
   1.423 +          if (m && ((res=match(ms, s+1, ep+1)) != NULL))
   1.424 +            return res;
   1.425 +          p=ep+1; goto init;  /* else return match(ms, s, ep+1); */
   1.426 +        }
   1.427 +        case '*': {  /* 0 or more repetitions */
   1.428 +          return max_expand(ms, s, p, ep);
   1.429 +        }
   1.430 +        case '+': {  /* 1 or more repetitions */
   1.431 +          return (m ? max_expand(ms, s+1, p, ep) : NULL);
   1.432 +        }
   1.433 +        case '-': {  /* 0 or more repetitions (minimum) */
   1.434 +          return min_expand(ms, s, p, ep);
   1.435 +        }
   1.436 +        default: {
   1.437 +          if (!m) return NULL;
   1.438 +          s++; p=ep; goto init;  /* else return match(ms, s+1, ep); */
   1.439 +        }
   1.440 +      }
   1.441 +    }
   1.442 +  }
   1.443 +}
   1.444 +
   1.445 +
   1.446 +
   1.447 +static const char *lmemfind (const char *s1, size_t l1,
   1.448 +                               const char *s2, size_t l2) {
   1.449 +  if (l2 == 0) return s1;  /* empty strings are everywhere */
   1.450 +  else if (l2 > l1) return NULL;  /* avoids a negative `l1' */
   1.451 +  else {
   1.452 +    const char *init;  /* to search for a `*s2' inside `s1' */
   1.453 +    l2--;  /* 1st char will be checked by `memchr' */
   1.454 +    l1 = l1-l2;  /* `s2' cannot be found after that */
   1.455 +    while (l1 > 0 && (init = (const char *)memchr(s1, *s2, l1)) != NULL) {
   1.456 +      init++;   /* 1st char is already checked */
   1.457 +      if (memcmp(init, s2+1, l2) == 0)
   1.458 +        return init-1;
   1.459 +      else {  /* correct `l1' and `s1' to try again */
   1.460 +        l1 -= init-s1;
   1.461 +        s1 = init;
   1.462 +      }
   1.463 +    }
   1.464 +    return NULL;  /* not found */
   1.465 +  }
   1.466 +}
   1.467 +
   1.468 +
   1.469 +static void push_onecapture (MatchState *ms, int i, const char *s,
   1.470 +                                                    const char *e) {
   1.471 +  if (i >= ms->level) {
   1.472 +    if (i == 0)  /* ms->level == 0, too */
   1.473 +      lua_pushlstring(ms->L, s, e - s);  /* add whole match */
   1.474 +    else
   1.475 +      luaL_error(ms->L, "invalid capture index");
   1.476 +  }
   1.477 +  else {
   1.478 +    ptrdiff_t l = ms->capture[i].len;
   1.479 +    if (l == CAP_UNFINISHED) luaL_error(ms->L, "unfinished capture");
   1.480 +    if (l == CAP_POSITION)
   1.481 +      lua_pushinteger(ms->L, ms->capture[i].init - ms->src_init + 1);
   1.482 +    else
   1.483 +      lua_pushlstring(ms->L, ms->capture[i].init, l);
   1.484 +  }
   1.485 +}
   1.486 +
   1.487 +
   1.488 +static int push_captures (MatchState *ms, const char *s, const char *e) {
   1.489 +  int i;
   1.490 +  int nlevels = (ms->level == 0 && s) ? 1 : ms->level;
   1.491 +  luaL_checkstack(ms->L, nlevels, "too many captures");
   1.492 +  for (i = 0; i < nlevels; i++)
   1.493 +    push_onecapture(ms, i, s, e);
   1.494 +  return nlevels;  /* number of strings pushed */
   1.495 +}
   1.496 +
   1.497 +
   1.498 +static int str_find_aux (lua_State *L, int find) {
   1.499 +  size_t l1, l2;
   1.500 +  const char *s = luaL_checklstring(L, 1, &l1);
   1.501 +  const char *p = luaL_checklstring(L, 2, &l2);
   1.502 +  ptrdiff_t init = posrelat(luaL_optinteger(L, 3, 1), l1) - 1;
   1.503 +  if (init < 0) init = 0;
   1.504 +  else if ((size_t)(init) > l1) init = (ptrdiff_t)l1;
   1.505 +  if (find && (lua_toboolean(L, 4) ||  /* explicit request? */
   1.506 +      strpbrk(p, SPECIALS) == NULL)) {  /* or no special characters? */
   1.507 +    /* do a plain search */
   1.508 +    const char *s2 = lmemfind(s+init, l1-init, p, l2);
   1.509 +    if (s2) {
   1.510 +      lua_pushinteger(L, s2-s+1);
   1.511 +      lua_pushinteger(L, s2-s+l2);
   1.512 +      return 2;
   1.513 +    }
   1.514 +  }
   1.515 +  else {
   1.516 +    MatchState ms;
   1.517 +    int anchor = (*p == '^') ? (p++, 1) : 0;
   1.518 +    const char *s1=s+init;
   1.519 +    ms.L = L;
   1.520 +    ms.src_init = s;
   1.521 +    ms.src_end = s+l1;
   1.522 +    do {
   1.523 +      const char *res;
   1.524 +      ms.level = 0;
   1.525 +      if ((res=match(&ms, s1, p)) != NULL) {
   1.526 +        if (find) {
   1.527 +          lua_pushinteger(L, s1-s+1);  /* start */
   1.528 +          lua_pushinteger(L, res-s);   /* end */
   1.529 +          return push_captures(&ms, NULL, 0) + 2;
   1.530 +        }
   1.531 +        else
   1.532 +          return push_captures(&ms, s1, res);
   1.533 +      }
   1.534 +    } while (s1++ < ms.src_end && !anchor);
   1.535 +  }
   1.536 +  lua_pushnil(L);  /* not found */
   1.537 +  return 1;
   1.538 +}
   1.539 +
   1.540 +
   1.541 +static int str_find (lua_State *L) {
   1.542 +  return str_find_aux(L, 1);
   1.543 +}
   1.544 +
   1.545 +
   1.546 +static int str_match (lua_State *L) {
   1.547 +  return str_find_aux(L, 0);
   1.548 +}
   1.549 +
   1.550 +
   1.551 +static int gmatch_aux (lua_State *L) {
   1.552 +  MatchState ms;
   1.553 +  size_t ls;
   1.554 +  const char *s = lua_tolstring(L, lua_upvalueindex(1), &ls);
   1.555 +  const char *p = lua_tostring(L, lua_upvalueindex(2));
   1.556 +  const char *src;
   1.557 +  ms.L = L;
   1.558 +  ms.src_init = s;
   1.559 +  ms.src_end = s+ls;
   1.560 +  for (src = s + (size_t)lua_tointeger(L, lua_upvalueindex(3));
   1.561 +       src <= ms.src_end;
   1.562 +       src++) {
   1.563 +    const char *e;
   1.564 +    ms.level = 0;
   1.565 +    if ((e = match(&ms, src, p)) != NULL) {
   1.566 +      lua_Integer newstart = e-s;
   1.567 +      if (e == src) newstart++;  /* empty match? go at least one position */
   1.568 +      lua_pushinteger(L, newstart);
   1.569 +      lua_replace(L, lua_upvalueindex(3));
   1.570 +      return push_captures(&ms, src, e);
   1.571 +    }
   1.572 +  }
   1.573 +  return 0;  /* not found */
   1.574 +}
   1.575 +
   1.576 +
   1.577 +static int gmatch (lua_State *L) {
   1.578 +  luaL_checkstring(L, 1);
   1.579 +  luaL_checkstring(L, 2);
   1.580 +  lua_settop(L, 2);
   1.581 +  lua_pushinteger(L, 0);
   1.582 +  lua_pushcclosure(L, gmatch_aux, 3);
   1.583 +  return 1;
   1.584 +}
   1.585 +
   1.586 +
   1.587 +static int gfind_nodef (lua_State *L) {
   1.588 +  return luaL_error(L, LUA_QL("string.gfind") " was renamed to "
   1.589 +                       LUA_QL("string.gmatch"));
   1.590 +}
   1.591 +
   1.592 +
   1.593 +static void add_s (MatchState *ms, luaL_Buffer *b, const char *s,
   1.594 +                                                   const char *e) {
   1.595 +  size_t l, i;
   1.596 +  const char *news = lua_tolstring(ms->L, 3, &l);
   1.597 +  for (i = 0; i < l; i++) {
   1.598 +    if (news[i] != L_ESC)
   1.599 +      luaL_addchar(b, news[i]);
   1.600 +    else {
   1.601 +      i++;  /* skip ESC */
   1.602 +      if (!isdigit(uchar(news[i])))
   1.603 +        luaL_addchar(b, news[i]);
   1.604 +      else if (news[i] == '0')
   1.605 +          luaL_addlstring(b, s, e - s);
   1.606 +      else {
   1.607 +        push_onecapture(ms, news[i] - '1', s, e);
   1.608 +        luaL_addvalue(b);  /* add capture to accumulated result */
   1.609 +      }
   1.610 +    }
   1.611 +  }
   1.612 +}
   1.613 +
   1.614 +
   1.615 +static void add_value (MatchState *ms, luaL_Buffer *b, const char *s,
   1.616 +                                                       const char *e) {
   1.617 +  lua_State *L = ms->L;
   1.618 +  switch (lua_type(L, 3)) {
   1.619 +    case LUA_TNUMBER:
   1.620 +    case LUA_TSTRING: {
   1.621 +      add_s(ms, b, s, e);
   1.622 +      return;
   1.623 +    }
   1.624 +    case LUA_TFUNCTION: {
   1.625 +      int n;
   1.626 +      lua_pushvalue(L, 3);
   1.627 +      n = push_captures(ms, s, e);
   1.628 +      lua_call(L, n, 1);
   1.629 +      break;
   1.630 +    }
   1.631 +    case LUA_TTABLE: {
   1.632 +      push_onecapture(ms, 0, s, e);
   1.633 +      lua_gettable(L, 3);
   1.634 +      break;
   1.635 +    }
   1.636 +  }
   1.637 +  if (!lua_toboolean(L, -1)) {  /* nil or false? */
   1.638 +    lua_pop(L, 1);
   1.639 +    lua_pushlstring(L, s, e - s);  /* keep original text */
   1.640 +  }
   1.641 +  else if (!lua_isstring(L, -1))
   1.642 +    luaL_error(L, "invalid replacement value (a %s)", luaL_typename(L, -1)); 
   1.643 +  luaL_addvalue(b);  /* add result to accumulator */
   1.644 +}
   1.645 +
   1.646 +
   1.647 +static int str_gsub (lua_State *L) {
   1.648 +  size_t srcl;
   1.649 +  const char *src = luaL_checklstring(L, 1, &srcl);
   1.650 +  const char *p = luaL_checkstring(L, 2);
   1.651 +  int  tr = lua_type(L, 3);
   1.652 +  int max_s = luaL_optint(L, 4, srcl+1);
   1.653 +  int anchor = (*p == '^') ? (p++, 1) : 0;
   1.654 +  int n = 0;
   1.655 +  MatchState ms;
   1.656 +  luaL_Buffer b;
   1.657 +  luaL_argcheck(L, tr == LUA_TNUMBER || tr == LUA_TSTRING ||
   1.658 +                   tr == LUA_TFUNCTION || tr == LUA_TTABLE, 3,
   1.659 +                      "string/function/table expected");
   1.660 +  luaL_buffinit(L, &b);
   1.661 +  ms.L = L;
   1.662 +  ms.src_init = src;
   1.663 +  ms.src_end = src+srcl;
   1.664 +  while (n < max_s) {
   1.665 +    const char *e;
   1.666 +    ms.level = 0;
   1.667 +    e = match(&ms, src, p);
   1.668 +    if (e) {
   1.669 +      n++;
   1.670 +      add_value(&ms, &b, src, e);
   1.671 +    }
   1.672 +    if (e && e>src) /* non empty match? */
   1.673 +      src = e;  /* skip it */
   1.674 +    else if (src < ms.src_end)
   1.675 +      luaL_addchar(&b, *src++);
   1.676 +    else break;
   1.677 +    if (anchor) break;
   1.678 +  }
   1.679 +  luaL_addlstring(&b, src, ms.src_end-src);
   1.680 +  luaL_pushresult(&b);
   1.681 +  lua_pushinteger(L, n);  /* number of substitutions */
   1.682 +  return 2;
   1.683 +}
   1.684 +
   1.685 +/* }====================================================== */
   1.686 +
   1.687 +
   1.688 +/* maximum size of each formatted item (> len(format('%99.99f', -1e308))) */
   1.689 +#define MAX_ITEM	512
   1.690 +/* valid flags in a format specification */
   1.691 +#define FLAGS	"-+ #0"
   1.692 +/*
   1.693 +** maximum size of each format specification (such as '%-099.99d')
   1.694 +** (+10 accounts for %99.99x plus margin of error)
   1.695 +*/
   1.696 +#define MAX_FORMAT	(sizeof(FLAGS) + sizeof(LUA_INTFRMLEN) + 10)
   1.697 +
   1.698 +
   1.699 +static void addquoted (lua_State *L, luaL_Buffer *b, int arg) {
   1.700 +  size_t l;
   1.701 +  const char *s = luaL_checklstring(L, arg, &l);
   1.702 +  luaL_addchar(b, '"');
   1.703 +  while (l--) {
   1.704 +    switch (*s) {
   1.705 +      case '"': case '\\': case '\n': {
   1.706 +        luaL_addchar(b, '\\');
   1.707 +        luaL_addchar(b, *s);
   1.708 +        break;
   1.709 +      }
   1.710 +      case '\r': {
   1.711 +        luaL_addlstring(b, "\\r", 2);
   1.712 +        break;
   1.713 +      }
   1.714 +      case '\0': {
   1.715 +        luaL_addlstring(b, "\\000", 4);
   1.716 +        break;
   1.717 +      }
   1.718 +      default: {
   1.719 +        luaL_addchar(b, *s);
   1.720 +        break;
   1.721 +      }
   1.722 +    }
   1.723 +    s++;
   1.724 +  }
   1.725 +  luaL_addchar(b, '"');
   1.726 +}
   1.727 +
   1.728 +static const char *scanformat (lua_State *L, const char *strfrmt, char *form) {
   1.729 +  const char *p = strfrmt;
   1.730 +  while (*p != '\0' && strchr(FLAGS, *p) != NULL) p++;  /* skip flags */
   1.731 +  if ((size_t)(p - strfrmt) >= sizeof(FLAGS))
   1.732 +    luaL_error(L, "invalid format (repeated flags)");
   1.733 +  if (isdigit(uchar(*p))) p++;  /* skip width */
   1.734 +  if (isdigit(uchar(*p))) p++;  /* (2 digits at most) */
   1.735 +  if (*p == '.') {
   1.736 +    p++;
   1.737 +    if (isdigit(uchar(*p))) p++;  /* skip precision */
   1.738 +    if (isdigit(uchar(*p))) p++;  /* (2 digits at most) */
   1.739 +  }
   1.740 +  if (isdigit(uchar(*p)))
   1.741 +    luaL_error(L, "invalid format (width or precision too long)");
   1.742 +  *(form++) = '%';
   1.743 +  strncpy(form, strfrmt, p - strfrmt + 1);
   1.744 +  form += p - strfrmt + 1;
   1.745 +  *form = '\0';
   1.746 +  return p;
   1.747 +}
   1.748 +
   1.749 +
   1.750 +static void addintlen (char *form) {
   1.751 +  size_t l = strlen(form);
   1.752 +  char spec = form[l - 1];
   1.753 +  strcpy(form + l - 1, LUA_INTFRMLEN);
   1.754 +  form[l + sizeof(LUA_INTFRMLEN) - 2] = spec;
   1.755 +  form[l + sizeof(LUA_INTFRMLEN) - 1] = '\0';
   1.756 +}
   1.757 +
   1.758 +
   1.759 +static int str_format (lua_State *L) {
   1.760 +  int arg = 1;
   1.761 +  size_t sfl;
   1.762 +  const char *strfrmt = luaL_checklstring(L, arg, &sfl);
   1.763 +  const char *strfrmt_end = strfrmt+sfl;
   1.764 +  luaL_Buffer b;
   1.765 +  luaL_buffinit(L, &b);
   1.766 +  while (strfrmt < strfrmt_end) {
   1.767 +    if (*strfrmt != L_ESC)
   1.768 +      luaL_addchar(&b, *strfrmt++);
   1.769 +    else if (*++strfrmt == L_ESC)
   1.770 +      luaL_addchar(&b, *strfrmt++);  /* %% */
   1.771 +    else { /* format item */
   1.772 +      char form[MAX_FORMAT];  /* to store the format (`%...') */
   1.773 +      char buff[MAX_ITEM];  /* to store the formatted item */
   1.774 +      arg++;
   1.775 +      strfrmt = scanformat(L, strfrmt, form);
   1.776 +      switch (*strfrmt++) {
   1.777 +        case 'c': {
   1.778 +          sprintf(buff, form, (int)luaL_checknumber(L, arg));
   1.779 +          break;
   1.780 +        }
   1.781 +        case 'd':  case 'i': {
   1.782 +          addintlen(form);
   1.783 +          sprintf(buff, form, (LUA_INTFRM_T)luaL_checknumber(L, arg));
   1.784 +          break;
   1.785 +        }
   1.786 +        case 'o':  case 'u':  case 'x':  case 'X': {
   1.787 +          addintlen(form);
   1.788 +          sprintf(buff, form, (unsigned LUA_INTFRM_T)luaL_checknumber(L, arg));
   1.789 +          break;
   1.790 +        }
   1.791 +        case 'e':  case 'E': case 'f':
   1.792 +        case 'g': case 'G': {
   1.793 +          sprintf(buff, form, (double)luaL_checknumber(L, arg));
   1.794 +          break;
   1.795 +        }
   1.796 +        case 'q': {
   1.797 +          addquoted(L, &b, arg);
   1.798 +          continue;  /* skip the 'addsize' at the end */
   1.799 +        }
   1.800 +        case 's': {
   1.801 +          size_t l;
   1.802 +          const char *s = luaL_checklstring(L, arg, &l);
   1.803 +          if (!strchr(form, '.') && l >= 100) {
   1.804 +            /* no precision and string is too long to be formatted;
   1.805 +               keep original string */
   1.806 +            lua_pushvalue(L, arg);
   1.807 +            luaL_addvalue(&b);
   1.808 +            continue;  /* skip the `addsize' at the end */
   1.809 +          }
   1.810 +          else {
   1.811 +            sprintf(buff, form, s);
   1.812 +            break;
   1.813 +          }
   1.814 +        }
   1.815 +        default: {  /* also treat cases `pnLlh' */
   1.816 +          return luaL_error(L, "invalid option " LUA_QL("%%%c") " to "
   1.817 +                               LUA_QL("format"), *(strfrmt - 1));
   1.818 +        }
   1.819 +      }
   1.820 +      luaL_addlstring(&b, buff, strlen(buff));
   1.821 +    }
   1.822 +  }
   1.823 +  luaL_pushresult(&b);
   1.824 +  return 1;
   1.825 +}
   1.826 +
   1.827 +
   1.828 +static const luaL_Reg strlib[] = {
   1.829 +  {"byte", str_byte},
   1.830 +  {"char", str_char},
   1.831 +  {"dump", str_dump},
   1.832 +  {"find", str_find},
   1.833 +  {"format", str_format},
   1.834 +  {"gfind", gfind_nodef},
   1.835 +  {"gmatch", gmatch},
   1.836 +  {"gsub", str_gsub},
   1.837 +  {"len", str_len},
   1.838 +  {"lower", str_lower},
   1.839 +  {"match", str_match},
   1.840 +  {"rep", str_rep},
   1.841 +  {"reverse", str_reverse},
   1.842 +  {"sub", str_sub},
   1.843 +  {"upper", str_upper},
   1.844 +  {NULL, NULL}
   1.845 +};
   1.846 +
   1.847 +
   1.848 +static void createmetatable (lua_State *L) {
   1.849 +  lua_createtable(L, 0, 1);  /* create metatable for strings */
   1.850 +  lua_pushliteral(L, "");  /* dummy string */
   1.851 +  lua_pushvalue(L, -2);
   1.852 +  lua_setmetatable(L, -2);  /* set string metatable */
   1.853 +  lua_pop(L, 1);  /* pop dummy string */
   1.854 +  lua_pushvalue(L, -2);  /* string library... */
   1.855 +  lua_setfield(L, -2, "__index");  /* ...is the __index metamethod */
   1.856 +  lua_pop(L, 1);  /* pop metatable */
   1.857 +}
   1.858 +
   1.859 +
   1.860 +/*
   1.861 +** Open string library
   1.862 +*/
   1.863 +LUALIB_API int luaopen_string (lua_State *L) {
   1.864 +  luaL_register(L, LUA_STRLIBNAME, strlib);
   1.865 +#if defined(LUA_COMPAT_GFIND)
   1.866 +  lua_getfield(L, -1, "gmatch");
   1.867 +  lua_setfield(L, -2, "gfind");
   1.868 +#endif
   1.869 +  createmetatable(L);
   1.870 +  return 1;
   1.871 +}
   1.872 +