source/texk/web2c/luatexdir/lua/lstrlibextjit.c

   1 /* lstrlibext.c for luajitex
   2
   3    Copyright 2013 Luigi Scarso
   4
   5    Code from lstrlibext.c for LuaTeX
   6    Original version copyright 2012 Taco Hoekwater <taco@luatex.org>
   7
   8    This file is part of LuajitTeX.
   9
  10    LuajitTeX is free software; you can redistribute it and/or modify it under
  11    the terms of the GNU General Public License as published by the Free
  12    Software Foundation; either version 2 of the License, or (at your
  13    option) any later version.
  14
  15    LuajitTeX is distributed in the hope that it will be useful, but WITHOUT
  16    ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
  17    FITNESS FOR A PARTICULAR PURPOSE.  See the GNU Lesser General Public
  18    License for more details.
  19
  20    You should have received a copy of the GNU General Public License along
  21    with LuaTeX; if not, see <http://www.gnu.org/licenses/>.
  22 */
  23
  24 #include "ptexlib.h"
  25 #include <stdio.h>
  26
  27 #define LUA_CORE
  28
  29 #include "lua.h"
  30 #include "lauxlib.h"
  31 #include "lualib.h"
  32
  33
  34
  35
  36 /*#include "lua51/lj_obj.h"*/
  37 /*#include "lua51/lj_gc.h"*/
  38 /*#include "lua51/lj_err.h"*/
  39 /*#include "lua51/lj_str.h"*/
  40 /*#include "lua51/lj_tab.h"*/
  41 /*#include "lua51/lj_meta.h"*/
  42 /*#include "lua51/lj_state.h"*/
  43 /*#include "lua51/lj_ff.h" */
  44 /*#include "lua51/lj_bcdump.h"*/
  45 /*#include "lua51/lj_char.h"*/
  46 /*#include "lua51/lj_lib.h"*/
  47
  48
  49
  50 static int bytepairs_aux (lua_State *L) {
  51   size_t ls;
  52   unsigned char i;
  53   const char *s = lua_tolstring(L, lua_upvalueindex(1), &ls);
  54   int ind       = lua_tointeger(L, lua_upvalueindex(2));
  55   if (ind<(int)ls) {
  56         if (ind+1<(int)ls) {
  57           lua_pushinteger(L, (ind+2));  /* iterator */
  58         } else {
  59           lua_pushinteger(L, (ind+1));  /* iterator */
  60         }
  61         lua_replace(L, lua_upvalueindex(2));
  62         i = (unsigned char)*(s+ind);
  63         lua_pushinteger(L, i);     /* byte one */
  64         if (ind+1<(int)ls) {
  65           i = (unsigned char)*(s+ind+1);
  66           lua_pushinteger(L, i);     /* byte two */
  67         } else {
  68           lua_pushnil(L);     /* odd string length */
  69         }
  70         return 2;
  71   }
  72   return 0;  /* string ended */
  73 }
  74
  75
  76 static int str_bytepairs (lua_State *L) {
  77   luaL_checkstring(L, 1);
  78   lua_settop(L, 1);
  79   lua_pushinteger(L, 0);
  80   lua_pushcclosure(L, bytepairs_aux, 2);
  81   return 1;
  82 }
  83
  84
  85 static int bytes_aux (lua_State *L) {
  86   size_t ls;
  87   unsigned char i;
  88   const char *s = lua_tolstring(L, lua_upvalueindex(1), &ls);
  89   int ind       = lua_tointeger(L, lua_upvalueindex(2));
  90   if (ind<(int)ls) {
  91     lua_pushinteger(L, (ind+1));  /* iterator */
  92         lua_replace(L, lua_upvalueindex(2));
  93         i = (unsigned char)*(s+ind);
  94         lua_pushinteger(L, i);     /* byte */
  95         return 1;
  96   }
  97   return 0;  /* string ended */
  98 }
  99
 100
 101 static int str_bytes (lua_State *L) {
 102   luaL_checkstring(L, 1);
 103   lua_settop(L, 1);
 104   lua_pushinteger(L, 0);
 105   lua_pushcclosure(L, bytes_aux, 2);
 106   return 1;
 107 }
 108
 109
 110 static int utf_failed(lua_State *L, int new_ind) {
 111   static char fffd [3] = {0xEF,0xBF,0xBD};
 112   lua_pushinteger(L, new_ind);  /* iterator */
 113   lua_replace(L, lua_upvalueindex(2));
 114   lua_pushlstring(L, fffd, 3);
 115   return 1;
 116 }
 117
 118
 119 static int utfcharacters_aux (lua_State *L) {
 120   static const unsigned char mask[4] = {0x80,0xE0,0xF0,0xF8};
 121   static const unsigned char mequ[4] = {0x00,0xC0,0xE0,0xF0};
 122   size_t ls;
 123   unsigned char c;
 124   int j;
 125   const char *s = lua_tolstring(L, lua_upvalueindex(1), &ls);
 126   int       ind = lua_tointeger(L, lua_upvalueindex(2));
 127   if (ind>=(int)ls) return 0; /* end of string */
 128   c = (unsigned) s[ind];
 129   for (j=0;j<4;j++) {
 130     if ((c&mask[j])==mequ[j]) {
 131       int k;
 132       if (ind+1+j>(int)ls) return utf_failed(L,ls); /* will not fit */
 133       for (k=1; k<=j; k++) {
 134         c = (unsigned) s[ind+k];
 135         if ((c&0xC0)!=0x80) return utf_failed(L,ind+k); /* bad follow */
 136       }
 137       lua_pushinteger(L, ind+1+j);  /* iterator */
 138       lua_replace(L, lua_upvalueindex(2));
 139       lua_pushlstring(L, s+ind, 1+j);
 140       return 1;
 141     }
 142   }
 143   return utf_failed(L,ind+1); /* we found a follow byte! */
 144 }
 145
 146
 147 static int str_utfcharacters (lua_State *L) {
 148   luaL_checkstring(L, 1);
 149   lua_settop(L, 1);
 150   lua_pushinteger(L, 0);
 151   lua_pushcclosure(L, utfcharacters_aux, 2);
 152   return 1;
 153 }
 154
 155
 156
 157 static int utfvalues_aux (lua_State *L) {
 158   size_t ls;
 159   unsigned char i = 0;
 160   unsigned char j = 0;
 161   unsigned char k = 0;
 162   unsigned char l = 0;
 163   unsigned int  v = 0xFFFD;
 164   int numbytes = 1;
 165   const char *s = lua_tolstring(L, lua_upvalueindex(1), &ls);
 166   int ind       = lua_tointeger(L, lua_upvalueindex(2));
 167
 168   if (ind<(int)ls) {
 169         i = *(s+ind);
 170         if (i<0x80) {
 171           v = i;
 172         } else if (i>=0xF0) {
 173           if ((ind+3)<(int)ls && ((unsigned)*(s+ind+1))>=0x80
 174                   && ((unsigned)*(s+ind+2))>=0x80 && ((unsigned)*(s+ind+3))>=0x80) {
 175                 numbytes  = 4;
 176                 j = ((unsigned)*(s+ind+1))-128;
 177                 k = ((unsigned)*(s+ind+2))-128;
 178                 l = ((unsigned)*(s+ind+3))-128;
 179                 v = (((((i-0xF0)*64) + j)*64) + k)*64 + l;
 180           }
 181         } else if (i>=0xE0) {
 182           if ((ind+2)<(int)ls && ((unsigned)*(s+ind+1))>=0x80 && ((unsigned)*(s+ind+2))>=0x80) {
 183                 numbytes  = 3;
 184                 j = ((unsigned)*(s+ind+1))-128;
 185                 k = ((unsigned)*(s+ind+2))-128;
 186                 v = (((i-0xE0)*64) + j)*64 + k;
 187           }
 188
 189         } else if (i>=0xC0) {
 190           if ((ind+1)<(int)ls && ((unsigned)*(s+ind+1))>=0x80) {
 191                 numbytes  = 2;
 192                 j = ((unsigned)*(s+ind+1))-128;
 193                 v = ((i-0xC0)*64) + j;
 194           }
 195         }
 196         lua_pushinteger(L, (ind+numbytes));  /* iterator */
 197         lua_replace(L, lua_upvalueindex(2));
 198         lua_pushinteger(L, v);
 199         return 1;
 200   }
 201   return 0;  /* string ended */
 202 }
 203
 204
 205 static int str_utfvalues (lua_State *L) {
 206   luaL_checkstring(L, 1);
 207   lua_settop(L, 1);
 208   lua_pushinteger(L, 0);
 209   lua_pushcclosure(L, utfvalues_aux, 2);
 210   return 1;
 211 }
 212
 213
 214 static int characterpairs_aux (lua_State *L) {
 215   size_t ls;
 216   char b[2];
 217   const char *s = lua_tolstring(L, lua_upvalueindex(1), &ls);
 218   int ind       = lua_tointeger(L, lua_upvalueindex(2));
 219   if (ind<(int)ls) {
 220         if (ind+1<(int)ls) {
 221           lua_pushinteger(L, (ind+2));  /* iterator */
 222         } else {
 223           lua_pushinteger(L, (ind+1));  /* iterator */
 224         }
 225         lua_replace(L, lua_upvalueindex(2));
 226         b[0] = *(s+ind); b[1] = 0;
 227         lua_pushlstring(L, b, 1);
 228         if (ind+1<(int)ls) {
 229           b[0] = *(s+ind+1);
 230           lua_pushlstring(L, b, 1);
 231         } else {
 232           lua_pushlstring(L, b+1, 0);
 233         }
 234         return 2;
 235   }
 236   return 0;  /* string ended */
 237 }
 238
 239
 240 static int str_characterpairs (lua_State *L) {
 241   luaL_checkstring(L, 1);
 242   lua_settop(L, 1);
 243   lua_pushinteger(L, 0);
 244   lua_pushcclosure(L, characterpairs_aux, 2);
 245   return 1;
 246 }
 247
 248
 249 static int characters_aux (lua_State *L) {
 250   size_t ls;
 251   char b[2];
 252   const char *s = lua_tolstring(L, lua_upvalueindex(1), &ls);
 253   int ind  = lua_tointeger(L, lua_upvalueindex(2));
 254   if (ind<(int)ls) {
 255     lua_pushinteger(L, (ind+1));  /* iterator */
 256         lua_replace(L, lua_upvalueindex(2));
 257         b[0] = *(s+ind); b[1] = 0;
 258         lua_pushlstring(L, b, 1);
 259         return 1;
 260   }
 261   return 0;  /* string ended */
 262 }
 263
 264
 265 static int str_characters (lua_State *L) {
 266   luaL_checkstring(L, 1);
 267   lua_settop(L, 1);
 268   lua_pushinteger(L, 0);
 269   lua_pushcclosure(L, characters_aux, 2);
 270   return 1;
 271 }
 272
 273 static int str_split (lua_State *L) {
 274   size_t l;
 275   size_t i;
 276   int n;
 277   char *q, *p, *orig;
 278   int mult = 0;
 279   const char *s = luaL_checklstring(L, 1, &l);
 280   const char *joiner = luaL_optstring(L, 2, " +");
 281   lua_newtable(L);
 282   if (l == 0) {
 283         lua_pushvalue(L,1);
 284         lua_rawseti(L,-2,1);
 285         return 1;
 286   }
 287   orig = p = malloc(l+1);
 288   if (p==NULL) {
 289         fprintf(stderr, "fatal: memory exhausted (malloc of %u bytes).\n",(int)(l+1));
 290         exit(EXIT_FAILURE);
 291   }
 292   strcpy(p,s);
 293   n = 1;
 294   q = p;
 295
 296   if (*joiner == 0) {
 297         for (i=0;i<l;i++) {
 298           lua_pushlstring(L,q,1); q++;
 299           lua_rawseti(L,-2,n); n++;
 300         }
 301         free(orig);
 302         return 1;
 303   }
 304   if (*(joiner+1) == '+') {
 305         mult = 1;
 306         while(*p==*joiner) {
 307           p++;
 308           l--;
 309         }
 310         q = p;
 311   }
 312   for (i=0;i<l;i++) {
 313         if (*(p+i)==*joiner) {
 314           *(p+i) = 0;
 315           lua_pushlstring(L,q,((p+i)-q));
 316           lua_rawseti(L,-2,n); n++;
 317           if (mult) {
 318                 while(*(p+i+1)==*joiner) {
 319                   i++;
 320                 }
 321           }
 322           q = p+i+1;
 323         }
 324   }
 325   if (mult && q==(p+l)) {
 326         free(orig);
 327         return 1;
 328   }
 329   if(q<=(p+l)) {
 330         lua_pushlstring(L,q,strlen(q));
 331         lua_rawseti(L,-2,n);
 332   }
 333   free(orig);
 334   return 1;
 335 }
 336
 337
 338 static const luaL_Reg strlibext[] = {
 339   {"utfvalues", str_utfvalues},
 340   {"utfcharacters", str_utfcharacters},
 341   {"characters", str_characters},
 342   {"characterpairs", str_characterpairs},
 343   {"bytes", str_bytes},
 344   {"bytepairs", str_bytepairs},
 345   {"explode", str_split},
 346   /* {"dump", str_dump} already in luajit */
 347   {NULL, NULL}
 348 };
 349
 350
 351
 352
 353 /* ------------------------------------------------------------------------ */
 354
 355 /* lj_libdef.h is generated by buildvm, it's not available on source */
 356 /* #include "lua51/lj_libdef.h" */
 357
 358 /* LUALIB_API int luaopen_string(lua_State *L) */
 359 /* { */
 360 /*   GCtab *mt; */
 361 /*   global_State *g; */
 362 /*   LJ_LIB_REG(L, LUA_STRLIBNAME, string); */
 363 /*   luaL_register(L, LUA_STRLIBNAME, strlib); */
 364 /*   //LJ_LIB_REG(L, LUA_STRLIBNAME, strlib); */
 365 /* #if defined(LUA_COMPAT_GFIND) && !LJ_52 */
 366 /*   lua_getfield(L, -1, "gmatch"); */
 367 /*   lua_setfield(L, -2, "gfind"); */
 368 /* #endif */
 369 /*   mt = lj_tab_new(L, 0, 1); */
 370 /*   /\* NOBARRIER: basemt is a GC root. *\/ */
 371 /*   g = G(L); */
 372 /*   setgcref(basemt_it(g, LJ_TSTR), obj2gco(mt)); */
 373 /*   settabV(L, lj_tab_setstr(L, mt, mmname_str(g, MM_index)), tabV(L->top-1)); */
 374 /*   mt->nomm = (uint8_t)(~(1u<<MM_index)); */
 375 /*   return 1; */
 376 /* } */
 377
 378 /* void open_strlibext(lua_State *L) */
 379 /* { */
 380 /*   int v; */
 381 /*   v = luaopen_string(L); */
 382 /* } */
 383
 384
 385 void open_strlibext(lua_State * L)
 386 {
 387     const luaL_Reg *lib;
 388     lua_getglobal(L, "string");
 389     for (lib=strlibext;lib->name;lib++) {
 390         lua_pushcfunction(L, lib->func);
 391         lua_setfield(L, -2, lib->name);
 392     }
 393     lua_pop(L,1);
 394 }
 395