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