source/lib/iconv.c

   1 /*
   2    Unix SMB/CIFS implementation.
   3    minimal iconv implementation
   4    Copyright (C) Andrew Tridgell 2001
   5    Copyright (C) Jelmer Vernooij 2002,2003
   6
   7    This program is free software; you can redistribute it and/or modify
   8    it under the terms of the GNU General Public License as published by
   9    the Free Software Foundation; either version 2 of the License, or
  10    (at your option) any later version.
  11
  12    This program is distributed in the hope that it will be useful,
  13    but WITHOUT ANY WARRANTY; without even the implied warranty of
  14    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  15    GNU General Public License for more details.
  16
  17    You should have received a copy of the GNU General Public License
  18    along with this program; if not, write to the Free Software
  19    Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
  20 */
  21
  22 #include "includes.h"
  23
  24
  25 /**
  26  * @file
  27  *
  28  * @brief Samba wrapper/stub for iconv character set conversion.
  29  *
  30  * iconv is the XPG2 interface for converting between character
  31  * encodings.  This file provides a Samba wrapper around it, and also
  32  * a simple reimplementation that is used if the system does not
  33  * implement iconv.
  34  *
  35  * Samba only works with encodings that are supersets of ASCII: ascii
  36  * characters like whitespace can be tested for directly, multibyte
  37  * sequences start with a byte with the high bit set, and strings are
  38  * terminated by a nul byte.
  39  *
  40  * Note that the only function provided by iconv is conversion between
  41  * characters.  It doesn't directly support operations like
  42  * uppercasing or comparison.  We have to convert to UCS-2 and compare
  43  * there.
  44  *
  45  * @sa Samba Developers Guide
  46  **/
  47
  48 static size_t ascii_pull(void *,char **, size_t *, char **, size_t *);
  49 static size_t ascii_push(void *,char **, size_t *, char **, size_t *);
  50 static size_t  utf8_pull(void *,char **, size_t *, char **, size_t *);
  51 static size_t  utf8_push(void *,char **, size_t *, char **, size_t *);
  52 static size_t ucs2hex_pull(void *,char **, size_t *, char **, size_t *);
  53 static size_t ucs2hex_push(void *,char **, size_t *, char **, size_t *);
  54 static size_t iconv_copy(void *,char **, size_t *, char **, size_t *);
  55
  56 static struct charset_functions builtin_functions[] = {
  57         {"UCS-2LE",  iconv_copy, iconv_copy},
  58         {"UTF8",   utf8_pull,  utf8_push},
  59         {"ASCII", ascii_pull, ascii_push},
  60         {"646", ascii_pull, ascii_push},
  61         {"UCS2-HEX", ucs2hex_pull, ucs2hex_push},
  62         {NULL, NULL, NULL}
  63 };
  64
  65 static struct charset_functions *charsets = NULL;
  66
  67 static struct charset_functions *find_charset_functions(const char *name)
  68 {
  69         struct charset_functions *c = charsets;
  70
  71         while(c) {
  72                 if (strcasecmp(name, c->name) == 0) {
  73                         return c;
  74                 }
  75                 c = c->next;
  76         }
  77
  78         return NULL;
  79 }
  80
  81 NTSTATUS smb_register_charset(struct charset_functions *funcs)
  82 {
  83         if (!funcs) {
  84                 return NT_STATUS_INVALID_PARAMETER;
  85         }
  86
  87         DEBUG(5, ("Attempting to register new charset %s\n", funcs->name));
  88         /* Check whether we already have this charset... */
  89         if (find_charset_functions(funcs->name)) {
  90                 DEBUG(0, ("Duplicate charset %s, not registering\n", funcs->name));
  91                 return NT_STATUS_OBJECT_NAME_COLLISION;
  92         }
  93
  94         funcs->next = funcs->prev = NULL;
  95         DEBUG(5, ("Registered charset %s\n", funcs->name));
  96         DLIST_ADD(charsets, funcs);
  97         return NT_STATUS_OK;
  98 }
  99
 100 void lazy_initialize_iconv(void)
 101 {
 102         static BOOL initialized;
 103         int i;
 104
 105         if (!initialized) {
 106                 initialized = True;
 107                 for(i = 0; builtin_functions[i].name; i++)
 108                         smb_register_charset(&builtin_functions[i]);
 109                 static_init_charset;
 110         }
 111 }
 112
 113 /* if there was an error then reset the internal state,
 114    this ensures that we don't have a shift state remaining for
 115    character sets like SJIS */
 116 static size_t sys_iconv(void *cd,
 117                         char **inbuf, size_t *inbytesleft,
 118                         char **outbuf, size_t *outbytesleft)
 119 {
 120 #ifdef HAVE_NATIVE_ICONV
 121         size_t ret = iconv((iconv_t)cd,
 122                            inbuf, inbytesleft,
 123                            outbuf, outbytesleft);
 124         if (ret == (size_t)-1) iconv(cd, NULL, NULL, NULL, NULL);
 125         return ret;
 126 #else
 127         errno = EINVAL;
 128         return -1;
 129 #endif
 130 }
 131
 132 /**
 133  * This is a simple portable iconv() implementaion.
 134  *
 135  * It only knows about a very small number of character sets - just
 136  * enough that Samba works on systems that don't have iconv.
 137  **/
 138 size_t smb_iconv(smb_iconv_t cd,
 139                  char **inbuf, size_t *inbytesleft,
 140                  char **outbuf, size_t *outbytesleft)
 141 {
 142         char cvtbuf[2048];
 143         char *bufp = cvtbuf;
 144         size_t bufsize;
 145
 146         /* in many cases we can go direct */
 147         if (cd->direct) {
 148                 return cd->direct(cd->cd_direct,
 149                                   (char **)inbuf, inbytesleft, outbuf, outbytesleft);
 150         }
 151
 152
 153         /* otherwise we have to do it chunks at a time */
 154         while (*inbytesleft > 0) {
 155                 bufp = cvtbuf;
 156                 bufsize = sizeof(cvtbuf);
 157
 158                 if (cd->pull(cd->cd_pull,
 159                              (char **)inbuf, inbytesleft, &bufp, &bufsize) == -1
 160                     && errno != E2BIG) return -1;
 161
 162                 bufp = cvtbuf;
 163                 bufsize = sizeof(cvtbuf) - bufsize;
 164
 165                 if (cd->push(cd->cd_push,
 166                              &bufp, &bufsize,
 167                              outbuf, outbytesleft) == -1) return -1;
 168         }
 169
 170         return 0;
 171 }
 172
 173 /*
 174   simple iconv_open() wrapper
 175  */
 176 smb_iconv_t smb_iconv_open(const char *tocode, const char *fromcode)
 177 {
 178         smb_iconv_t ret;
 179         struct charset_functions *from, *to;
 180
 181         lazy_initialize_iconv();
 182         from = charsets;
 183         to = charsets;
 184
 185         ret = (smb_iconv_t)malloc(sizeof(*ret));
 186         if (!ret) {
 187                 errno = ENOMEM;
 188                 return (smb_iconv_t)-1;
 189         }
 190         memset(ret, 0, sizeof(*ret));
 191
 192         ret->from_name = strdup(fromcode);
 193         ret->to_name = strdup(tocode);
 194
 195         /* check for the simplest null conversion */
 196         if (strcasecmp(fromcode, tocode) == 0) {
 197                 ret->direct = iconv_copy;
 198                 return ret;
 199         }
 200
 201         /* check if we have a builtin function for this conversion */
 202         from = find_charset_functions(fromcode);
 203         if(from)ret->pull = from->pull;
 204
 205         to = find_charset_functions(tocode);
 206         if(to)ret->push = to->push;
 207
 208         /* check if we can use iconv for this conversion */
 209 #ifdef HAVE_NATIVE_ICONV
 210         if (!ret->pull) {
 211                 ret->cd_pull = iconv_open("UCS-2LE", fromcode);
 212                 if (ret->cd_pull != (iconv_t)-1)
 213                         ret->pull = sys_iconv;
 214         }
 215
 216         if (!ret->push) {
 217                 ret->cd_push = iconv_open(tocode, "UCS-2LE");
 218                 if (ret->cd_push != (iconv_t)-1)
 219                         ret->push = sys_iconv;
 220         }
 221 #endif
 222
 223         /* check if there is a module available that can do this conversion */
 224         if (!ret->pull && NT_STATUS_IS_OK(smb_probe_module("charset", fromcode))) {
 225                 if(!(from = find_charset_functions(fromcode)))
 226                         DEBUG(0, ("Module %s doesn't provide charset %s!\n", fromcode, fromcode));
 227                 else
 228                         ret->pull = from->pull;
 229         }
 230
 231         if (!ret->push && NT_STATUS_IS_OK(smb_probe_module("charset", tocode))) {
 232                 if(!(to = find_charset_functions(tocode)))
 233                         DEBUG(0, ("Module %s doesn't provide charset %s!\n", tocode, tocode));
 234                 else
 235                         ret->push = to->push;
 236         }
 237
 238         if (!ret->push || !ret->pull) {
 239                 SAFE_FREE(ret->from_name);
 240                 SAFE_FREE(ret->to_name);
 241                 SAFE_FREE(ret);
 242                 errno = EINVAL;
 243                 return (smb_iconv_t)-1;
 244         }
 245
 246         /* check for conversion to/from ucs2 */
 247         if (strcasecmp(fromcode, "UCS-2LE") == 0 && to) {
 248                 ret->direct = to->push;
 249                 ret->push = ret->pull = NULL;
 250                 return ret;
 251         }
 252
 253         if (strcasecmp(tocode, "UCS-2LE") == 0 && from) {
 254                 ret->direct = from->pull;
 255                 ret->push = ret->pull = NULL;
 256                 return ret;
 257         }
 258
 259         /* Check if we can do the conversion direct */
 260 #ifdef HAVE_NATIVE_ICONV
 261         if (strcasecmp(fromcode, "UCS-2LE") == 0) {
 262                 ret->direct = sys_iconv;
 263                 ret->cd_direct = ret->cd_push;
 264                 ret->cd_push = NULL;
 265                 return ret;
 266         }
 267         if (strcasecmp(tocode, "UCS-2LE") == 0) {
 268                 ret->direct = sys_iconv;
 269                 ret->cd_direct = ret->cd_pull;
 270                 ret->cd_pull = NULL;
 271                 return ret;
 272         }
 273 #endif
 274
 275         return ret;
 276 }
 277
 278 /*
 279   simple iconv_close() wrapper
 280 */
 281 int smb_iconv_close (smb_iconv_t cd)
 282 {
 283 #ifdef HAVE_NATIVE_ICONV
 284         if (cd->cd_direct) iconv_close((iconv_t)cd->cd_direct);
 285         if (cd->cd_pull) iconv_close((iconv_t)cd->cd_pull);
 286         if (cd->cd_push) iconv_close((iconv_t)cd->cd_push);
 287 #endif
 288
 289         SAFE_FREE(cd->from_name);
 290         SAFE_FREE(cd->to_name);
 291
 292         memset(cd, 0, sizeof(*cd));
 293         SAFE_FREE(cd);
 294         return 0;
 295 }
 296
 297
 298 /**********************************************************************
 299  the following functions implement the builtin character sets in Samba
 300  and also the "test" character sets that are designed to test
 301  multi-byte character set support for english users
 302 ***********************************************************************/
 303
 304 static size_t ascii_pull(void *cd, char **inbuf, size_t *inbytesleft,
 305                          char **outbuf, size_t *outbytesleft)
 306 {
 307         while (*inbytesleft >= 1 && *outbytesleft >= 2) {
 308                 (*outbuf)[0] = (*inbuf)[0];
 309                 (*outbuf)[1] = 0;
 310                 (*inbytesleft)  -= 1;
 311                 (*outbytesleft) -= 2;
 312                 (*inbuf)  += 1;
 313                 (*outbuf) += 2;
 314         }
 315
 316         if (*inbytesleft > 0) {
 317                 errno = E2BIG;
 318                 return -1;
 319         }
 320
 321         return 0;
 322 }
 323
 324 static size_t ascii_push(void *cd, char **inbuf, size_t *inbytesleft,
 325                          char **outbuf, size_t *outbytesleft)
 326 {
 327         int ir_count=0;
 328
 329         while (*inbytesleft >= 2 && *outbytesleft >= 1) {
 330                 (*outbuf)[0] = (*inbuf)[0] & 0x7F;
 331                 if ((*inbuf)[1]) ir_count++;
 332                 (*inbytesleft)  -= 2;
 333                 (*outbytesleft) -= 1;
 334                 (*inbuf)  += 2;
 335                 (*outbuf) += 1;
 336         }
 337
 338         if (*inbytesleft == 1) {
 339                 errno = EINVAL;
 340                 return -1;
 341         }
 342
 343         if (*inbytesleft > 1) {
 344                 errno = E2BIG;
 345                 return -1;
 346         }
 347
 348         return ir_count;
 349 }
 350
 351
 352 static size_t ucs2hex_pull(void *cd, char **inbuf, size_t *inbytesleft,
 353                          char **outbuf, size_t *outbytesleft)
 354 {
 355         while (*inbytesleft >= 1 && *outbytesleft >= 2) {
 356                 unsigned v;
 357
 358                 if ((*inbuf)[0] != '@') {
 359                         /* seven bit ascii case */
 360                         (*outbuf)[0] = (*inbuf)[0];
 361                         (*outbuf)[1] = 0;
 362                         (*inbytesleft)  -= 1;
 363                         (*outbytesleft) -= 2;
 364                         (*inbuf)  += 1;
 365                         (*outbuf) += 2;
 366                         continue;
 367                 }
 368                 /* it's a hex character */
 369                 if (*inbytesleft < 5) {
 370                         errno = EINVAL;
 371                         return -1;
 372                 }
 373
 374                 if (sscanf(&(*inbuf)[1], "%04x", &v) != 1) {
 375                         errno = EILSEQ;
 376                         return -1;
 377                 }
 378
 379                 (*outbuf)[0] = v&0xff;
 380                 (*outbuf)[1] = v>>8;
 381                 (*inbytesleft)  -= 5;
 382                 (*outbytesleft) -= 2;
 383                 (*inbuf)  += 5;
 384                 (*outbuf) += 2;
 385         }
 386
 387         if (*inbytesleft > 0) {
 388                 errno = E2BIG;
 389                 return -1;
 390         }
 391
 392         return 0;
 393 }
 394
 395 static size_t ucs2hex_push(void *cd, char **inbuf, size_t *inbytesleft,
 396                            char **outbuf, size_t *outbytesleft)
 397 {
 398         while (*inbytesleft >= 2 && *outbytesleft >= 1) {
 399                 char buf[6];
 400
 401                 if ((*inbuf)[1] == 0 &&
 402                     ((*inbuf)[0] & 0x80) == 0 &&
 403                     (*inbuf)[0] != '@') {
 404                         (*outbuf)[0] = (*inbuf)[0];
 405                         (*inbytesleft)  -= 2;
 406                         (*outbytesleft) -= 1;
 407                         (*inbuf)  += 2;
 408                         (*outbuf) += 1;
 409                         continue;
 410                 }
 411                 if (*outbytesleft < 5) {
 412                         errno = E2BIG;
 413                         return -1;
 414                 }
 415                 snprintf(buf, 6, "@%04x", SVAL(*inbuf, 0));
 416                 memcpy(*outbuf, buf, 5);
 417                 (*inbytesleft)  -= 2;
 418                 (*outbytesleft) -= 5;
 419                 (*inbuf)  += 2;
 420                 (*outbuf) += 5;
 421         }
 422
 423         if (*inbytesleft == 1) {
 424                 errno = EINVAL;
 425                 return -1;
 426         }
 427
 428         if (*inbytesleft > 1) {
 429                 errno = E2BIG;
 430                 return -1;
 431         }
 432
 433         return 0;
 434 }
 435
 436
 437 static size_t iconv_copy(void *cd, char **inbuf, size_t *inbytesleft,
 438                          char **outbuf, size_t *outbytesleft)
 439 {
 440         int n;
 441
 442         n = MIN(*inbytesleft, *outbytesleft);
 443
 444         memmove(*outbuf, *inbuf, n);
 445
 446         (*inbytesleft) -= n;
 447         (*outbytesleft) -= n;
 448         (*inbuf) += n;
 449         (*outbuf) += n;
 450
 451         if (*inbytesleft > 0) {
 452                 errno = E2BIG;
 453                 return -1;
 454         }
 455
 456         return 0;
 457 }
 458
 459 static size_t utf8_pull(void *cd, char **inbuf, size_t *inbytesleft,
 460                          char **outbuf, size_t *outbytesleft)
 461 {
 462         while (*inbytesleft >= 1 && *outbytesleft >= 2) {
 463                 unsigned char *c = (unsigned char *)*inbuf;
 464                 unsigned char *uc = (unsigned char *)*outbuf;
 465                 int len = 1;
 466
 467                 if ((c[0] & 0x80) == 0) {
 468                         uc[0] = c[0];
 469                         uc[1] = 0;
 470                 } else if ((c[0] & 0xf0) == 0xe0) {
 471                         if (*inbytesleft < 3) {
 472                                 DEBUG(0,("short utf8 char\n"));
 473                                 goto badseq;
 474                         }
 475                         uc[1] = ((c[0]&0xF)<<4) | ((c[1]>>2)&0xF);
 476                         uc[0] = (c[1]<<6) | (c[2]&0x3f);
 477                         len = 3;
 478                 } else if ((c[0] & 0xe0) == 0xc0) {
 479                         if (*inbytesleft < 2) {
 480                                 DEBUG(0,("short utf8 char\n"));
 481                                 goto badseq;
 482                         }
 483                         uc[1] = (c[0]>>2) & 0x7;
 484                         uc[0] = (c[0]<<6) | (c[1]&0x3f);
 485                         len = 2;
 486                 }
 487
 488                 (*inbuf)  += len;
 489                 (*inbytesleft)  -= len;
 490                 (*outbytesleft) -= 2;
 491                 (*outbuf) += 2;
 492         }
 493
 494         if (*inbytesleft > 0) {
 495                 errno = E2BIG;
 496                 return -1;
 497         }
 498
 499         return 0;
 500
 501 badseq:
 502         errno = EINVAL;
 503         return -1;
 504 }
 505
 506 static size_t utf8_push(void *cd, char **inbuf, size_t *inbytesleft,
 507                          char **outbuf, size_t *outbytesleft)
 508 {
 509         while (*inbytesleft >= 2 && *outbytesleft >= 1) {
 510                 unsigned char *c = (unsigned char *)*outbuf;
 511                 unsigned char *uc = (unsigned char *)*inbuf;
 512                 int len=1;
 513
 514                 if (uc[1] & 0xf8) {
 515                         if (*outbytesleft < 3) {
 516                                 DEBUG(0,("short utf8 write\n"));
 517                                 goto toobig;
 518                         }
 519                         c[0] = 0xe0 | (uc[1]>>4);
 520                         c[1] = 0x80 | ((uc[1]&0xF)<<2) | (uc[0]>>6);
 521                         c[2] = 0x80 | (uc[0]&0x3f);
 522                         len = 3;
 523                 } else if (uc[1] | (uc[0] & 0x80)) {
 524                         if (*outbytesleft < 2) {
 525                                 DEBUG(0,("short utf8 write\n"));
 526                                 goto toobig;
 527                         }
 528                         c[0] = 0xc0 | (uc[1]<<2) | (uc[0]>>6);
 529                         c[1] = 0x80 | (uc[0]&0x3f);
 530                         len = 2;
 531                 } else {
 532                         c[0] = uc[0];
 533                 }
 534
 535
 536                 (*inbytesleft)  -= 2;
 537                 (*outbytesleft) -= len;
 538                 (*inbuf)  += 2;
 539                 (*outbuf) += len;
 540         }
 541
 542         if (*inbytesleft == 1) {
 543                 errno = EINVAL;
 544                 return -1;
 545         }
 546
 547         if (*inbytesleft > 1) {
 548                 errno = E2BIG;
 549                 return -1;
 550         }
 551
 552         return 0;
 553
 554 toobig:
 555         errno = E2BIG;
 556         return -1;
 557 }
 558