source3/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  * We have to use strcasecmp here as the character conversions
  26  * haven't been initialised yet. JRA.
  27  */
  28
  29 #undef strcasecmp
  30
  31 /**
  32  * @file
  33  *
  34  * @brief Samba wrapper/stub for iconv character set conversion.
  35  *
  36  * iconv is the XPG2 interface for converting between character
  37  * encodings.  This file provides a Samba wrapper around it, and also
  38  * a simple reimplementation that is used if the system does not
  39  * implement iconv.
  40  *
  41  * Samba only works with encodings that are supersets of ASCII: ascii
  42  * characters like whitespace can be tested for directly, multibyte
  43  * sequences start with a byte with the high bit set, and strings are
  44  * terminated by a nul byte.
  45  *
  46  * Note that the only function provided by iconv is conversion between
  47  * characters.  It doesn't directly support operations like
  48  * uppercasing or comparison.  We have to convert to UCS-2 and compare
  49  * there.
  50  *
  51  * @sa Samba Developers Guide
  52  **/
  53
  54 static size_t ascii_pull(void *,char **, size_t *, char **, size_t *);
  55 static size_t ascii_push(void *,char **, size_t *, char **, size_t *);
  56 static size_t latin1_push(void *,char **, size_t *, char **, size_t *);
  57 static size_t  utf8_pull(void *,char **, size_t *, char **, size_t *);
  58 static size_t  utf8_push(void *,char **, size_t *, char **, size_t *);
  59 static size_t ucs2hex_pull(void *,char **, size_t *, char **, size_t *);
  60 static size_t ucs2hex_push(void *,char **, size_t *, char **, size_t *);
  61 static size_t iconv_copy(void *,char **, size_t *, char **, size_t *);
  62
  63 static struct charset_functions builtin_functions[] = {
  64         {"UCS-2LE",  iconv_copy, iconv_copy},
  65         {"UTF8",   utf8_pull,  utf8_push},
  66         {"ASCII", ascii_pull, ascii_push},
  67         {"646", ascii_pull, ascii_push},
  68         {"ISO-8859-1", ascii_pull, latin1_push},
  69         {"UCS2-HEX", ucs2hex_pull, ucs2hex_push},
  70         {NULL, NULL, NULL}
  71 };
  72
  73 static struct charset_functions *charsets = NULL;
  74
  75 static struct charset_functions *find_charset_functions(const char *name)
  76 {
  77         struct charset_functions *c = charsets;
  78
  79         while(c) {
  80                 if (strcasecmp(name, c->name) == 0) {
  81                         return c;
  82                 }
  83                 c = c->next;
  84         }
  85
  86         return NULL;
  87 }
  88
  89 NTSTATUS smb_register_charset(struct charset_functions *funcs)
  90 {
  91         if (!funcs) {
  92                 return NT_STATUS_INVALID_PARAMETER;
  93         }
  94
  95         DEBUG(5, ("Attempting to register new charset %s\n", funcs->name));
  96         /* Check whether we already have this charset... */
  97         if (find_charset_functions(funcs->name)) {
  98                 DEBUG(0, ("Duplicate charset %s, not registering\n", funcs->name));
  99                 return NT_STATUS_OBJECT_NAME_COLLISION;
 100         }
 101
 102         funcs->next = funcs->prev = NULL;
 103         DEBUG(5, ("Registered charset %s\n", funcs->name));
 104         DLIST_ADD(charsets, funcs);
 105         return NT_STATUS_OK;
 106 }
 107
 108 void lazy_initialize_iconv(void)
 109 {
 110         static BOOL initialized;
 111         int i;
 112
 113         if (!initialized) {
 114                 initialized = True;
 115                 for(i = 0; builtin_functions[i].name; i++)
 116                         smb_register_charset(&builtin_functions[i]);
 117                 static_init_charset;
 118         }
 119 }
 120
 121 /* if there was an error then reset the internal state,
 122    this ensures that we don't have a shift state remaining for
 123    character sets like SJIS */
 124 static size_t sys_iconv(void *cd,
 125                         char **inbuf, size_t *inbytesleft,
 126                         char **outbuf, size_t *outbytesleft)
 127 {
 128 #ifdef HAVE_NATIVE_ICONV
 129         size_t ret = iconv((iconv_t)cd,
 130                            inbuf, inbytesleft,
 131                            outbuf, outbytesleft);
 132         if (ret == (size_t)-1) iconv(cd, NULL, NULL, NULL, NULL);
 133         return ret;
 134 #else
 135         errno = EINVAL;
 136         return -1;
 137 #endif
 138 }
 139
 140 /**
 141  * This is a simple portable iconv() implementaion.
 142  *
 143  * It only knows about a very small number of character sets - just
 144  * enough that Samba works on systems that don't have iconv.
 145  **/
 146 size_t smb_iconv(smb_iconv_t cd,
 147                  char **inbuf, size_t *inbytesleft,
 148                  char **outbuf, size_t *outbytesleft)
 149 {
 150         char cvtbuf[2048];
 151         char *bufp = cvtbuf;
 152         size_t bufsize;
 153
 154         /* in many cases we can go direct */
 155         if (cd->direct) {
 156                 return cd->direct(cd->cd_direct,
 157                                   (char **)inbuf, inbytesleft, outbuf, outbytesleft);
 158         }
 159
 160
 161         /* otherwise we have to do it chunks at a time */
 162         while (*inbytesleft > 0) {
 163                 bufp = cvtbuf;
 164                 bufsize = sizeof(cvtbuf);
 165
 166                 if (cd->pull(cd->cd_pull,
 167                              (char **)inbuf, inbytesleft, &bufp, &bufsize) == -1
 168                     && errno != E2BIG) return -1;
 169
 170                 bufp = cvtbuf;
 171                 bufsize = sizeof(cvtbuf) - bufsize;
 172
 173                 if (cd->push(cd->cd_push,
 174                              &bufp, &bufsize,
 175                              outbuf, outbytesleft) == -1) return -1;
 176         }
 177
 178         return 0;
 179 }
 180
 181 /*
 182   simple iconv_open() wrapper
 183  */
 184 smb_iconv_t smb_iconv_open(const char *tocode, const char *fromcode)
 185 {
 186         smb_iconv_t ret;
 187         struct charset_functions *from, *to;
 188
 189         lazy_initialize_iconv();
 190         from = charsets;
 191         to = charsets;
 192
 193         ret = (smb_iconv_t)malloc(sizeof(*ret));
 194         if (!ret) {
 195                 errno = ENOMEM;
 196                 return (smb_iconv_t)-1;
 197         }
 198         memset(ret, 0, sizeof(*ret));
 199
 200         ret->from_name = strdup(fromcode);
 201         ret->to_name = strdup(tocode);
 202
 203         /* check for the simplest null conversion */
 204         if (strcasecmp(fromcode, tocode) == 0) {
 205                 ret->direct = iconv_copy;
 206                 return ret;
 207         }
 208
 209         /* check if we have a builtin function for this conversion */
 210         from = find_charset_functions(fromcode);
 211         if(from)ret->pull = from->pull;
 212
 213         to = find_charset_functions(tocode);
 214         if(to)ret->push = to->push;
 215
 216         /* check if we can use iconv for this conversion */
 217 #ifdef HAVE_NATIVE_ICONV
 218         if (!ret->pull) {
 219                 ret->cd_pull = iconv_open("UCS-2LE", fromcode);
 220                 if (ret->cd_pull != (iconv_t)-1)
 221                         ret->pull = sys_iconv;
 222         }
 223
 224         if (!ret->push) {
 225                 ret->cd_push = iconv_open(tocode, "UCS-2LE");
 226                 if (ret->cd_push != (iconv_t)-1)
 227                         ret->push = sys_iconv;
 228         }
 229 #endif
 230
 231         /* check if there is a module available that can do this conversion */
 232         if (!ret->pull && NT_STATUS_IS_OK(smb_probe_module("charset", fromcode))) {
 233                 if(!(from = find_charset_functions(fromcode)))
 234                         DEBUG(0, ("Module %s doesn't provide charset %s!\n", fromcode, fromcode));
 235                 else
 236                         ret->pull = from->pull;
 237         }
 238
 239         if (!ret->push && NT_STATUS_IS_OK(smb_probe_module("charset", tocode))) {
 240                 if(!(to = find_charset_functions(tocode)))
 241                         DEBUG(0, ("Module %s doesn't provide charset %s!\n", tocode, tocode));
 242                 else
 243                         ret->push = to->push;
 244         }
 245
 246         if (!ret->push || !ret->pull) {
 247                 SAFE_FREE(ret->from_name);
 248                 SAFE_FREE(ret->to_name);
 249                 SAFE_FREE(ret);
 250                 errno = EINVAL;
 251                 return (smb_iconv_t)-1;
 252         }
 253
 254         /* check for conversion to/from ucs2 */
 255         if (strcasecmp(fromcode, "UCS-2LE") == 0 && to) {
 256                 ret->direct = to->push;
 257                 ret->push = ret->pull = NULL;
 258                 return ret;
 259         }
 260
 261         if (strcasecmp(tocode, "UCS-2LE") == 0 && from) {
 262                 ret->direct = from->pull;
 263                 ret->push = ret->pull = NULL;
 264                 return ret;
 265         }
 266
 267         /* Check if we can do the conversion direct */
 268 #ifdef HAVE_NATIVE_ICONV
 269         if (strcasecmp(fromcode, "UCS-2LE") == 0) {
 270                 ret->direct = sys_iconv;
 271                 ret->cd_direct = ret->cd_push;
 272                 ret->cd_push = NULL;
 273                 return ret;
 274         }
 275         if (strcasecmp(tocode, "UCS-2LE") == 0) {
 276                 ret->direct = sys_iconv;
 277                 ret->cd_direct = ret->cd_pull;
 278                 ret->cd_pull = NULL;
 279                 return ret;
 280         }
 281 #endif
 282
 283         return ret;
 284 }
 285
 286 /*
 287   simple iconv_close() wrapper
 288 */
 289 int smb_iconv_close (smb_iconv_t cd)
 290 {
 291 #ifdef HAVE_NATIVE_ICONV
 292         if (cd->cd_direct) iconv_close((iconv_t)cd->cd_direct);
 293         if (cd->cd_pull) iconv_close((iconv_t)cd->cd_pull);
 294         if (cd->cd_push) iconv_close((iconv_t)cd->cd_push);
 295 #endif
 296
 297         SAFE_FREE(cd->from_name);
 298         SAFE_FREE(cd->to_name);
 299
 300         memset(cd, 0, sizeof(*cd));
 301         SAFE_FREE(cd);
 302         return 0;
 303 }
 304
 305
 306 /**********************************************************************
 307  the following functions implement the builtin character sets in Samba
 308  and also the "test" character sets that are designed to test
 309  multi-byte character set support for english users
 310 ***********************************************************************/
 311
 312 static size_t ascii_pull(void *cd, char **inbuf, size_t *inbytesleft,
 313                          char **outbuf, size_t *outbytesleft)
 314 {
 315         while (*inbytesleft >= 1 && *outbytesleft >= 2) {
 316                 (*outbuf)[0] = (*inbuf)[0];
 317                 (*outbuf)[1] = 0;
 318                 (*inbytesleft)  -= 1;
 319                 (*outbytesleft) -= 2;
 320                 (*inbuf)  += 1;
 321                 (*outbuf) += 2;
 322         }
 323
 324         if (*inbytesleft > 0) {
 325                 errno = E2BIG;
 326                 return -1;
 327         }
 328
 329         return 0;
 330 }
 331
 332 static size_t ascii_push(void *cd, char **inbuf, size_t *inbytesleft,
 333                          char **outbuf, size_t *outbytesleft)
 334 {
 335         int ir_count=0;
 336
 337         while (*inbytesleft >= 2 && *outbytesleft >= 1) {
 338                 (*outbuf)[0] = (*inbuf)[0] & 0x7F;
 339                 if ((*inbuf)[1]) ir_count++;
 340                 (*inbytesleft)  -= 2;
 341                 (*outbytesleft) -= 1;
 342                 (*inbuf)  += 2;
 343                 (*outbuf) += 1;
 344         }
 345
 346         if (*inbytesleft == 1) {
 347                 errno = EINVAL;
 348                 return -1;
 349         }
 350
 351         if (*inbytesleft > 1) {
 352                 errno = E2BIG;
 353                 return -1;
 354         }
 355
 356         return ir_count;
 357 }
 358
 359 static size_t latin1_push(void *cd, char **inbuf, size_t *inbytesleft,
 360                          char **outbuf, size_t *outbytesleft)
 361 {
 362         int ir_count=0;
 363
 364         while (*inbytesleft >= 2 && *outbytesleft >= 1) {
 365                 (*outbuf)[0] = (*inbuf)[0];
 366                 if ((*inbuf)[1]) ir_count++;
 367                 (*inbytesleft)  -= 2;
 368                 (*outbytesleft) -= 1;
 369                 (*inbuf)  += 2;
 370                 (*outbuf) += 1;
 371         }
 372
 373         if (*inbytesleft == 1) {
 374                 errno = EINVAL;
 375                 return -1;
 376         }
 377
 378         if (*inbytesleft > 1) {
 379                 errno = E2BIG;
 380                 return -1;
 381         }
 382
 383         return ir_count;
 384 }
 385
 386 static size_t ucs2hex_pull(void *cd, char **inbuf, size_t *inbytesleft,
 387                          char **outbuf, size_t *outbytesleft)
 388 {
 389         while (*inbytesleft >= 1 && *outbytesleft >= 2) {
 390                 unsigned v;
 391
 392                 if ((*inbuf)[0] != '@') {
 393                         /* seven bit ascii case */
 394                         (*outbuf)[0] = (*inbuf)[0];
 395                         (*outbuf)[1] = 0;
 396                         (*inbytesleft)  -= 1;
 397                         (*outbytesleft) -= 2;
 398                         (*inbuf)  += 1;
 399                         (*outbuf) += 2;
 400                         continue;
 401                 }
 402                 /* it's a hex character */
 403                 if (*inbytesleft < 5) {
 404                         errno = EINVAL;
 405                         return -1;
 406                 }
 407
 408                 if (sscanf(&(*inbuf)[1], "%04x", &v) != 1) {
 409                         errno = EILSEQ;
 410                         return -1;
 411                 }
 412
 413                 (*outbuf)[0] = v&0xff;
 414                 (*outbuf)[1] = v>>8;
 415                 (*inbytesleft)  -= 5;
 416                 (*outbytesleft) -= 2;
 417                 (*inbuf)  += 5;
 418                 (*outbuf) += 2;
 419         }
 420
 421         if (*inbytesleft > 0) {
 422                 errno = E2BIG;
 423                 return -1;
 424         }
 425
 426         return 0;
 427 }
 428
 429 static size_t ucs2hex_push(void *cd, char **inbuf, size_t *inbytesleft,
 430                            char **outbuf, size_t *outbytesleft)
 431 {
 432         while (*inbytesleft >= 2 && *outbytesleft >= 1) {
 433                 char buf[6];
 434
 435                 if ((*inbuf)[1] == 0 &&
 436                     ((*inbuf)[0] & 0x80) == 0 &&
 437                     (*inbuf)[0] != '@') {
 438                         (*outbuf)[0] = (*inbuf)[0];
 439                         (*inbytesleft)  -= 2;
 440                         (*outbytesleft) -= 1;
 441                         (*inbuf)  += 2;
 442                         (*outbuf) += 1;
 443                         continue;
 444                 }
 445                 if (*outbytesleft < 5) {
 446                         errno = E2BIG;
 447                         return -1;
 448                 }
 449                 snprintf(buf, 6, "@%04x", SVAL(*inbuf, 0));
 450                 memcpy(*outbuf, buf, 5);
 451                 (*inbytesleft)  -= 2;
 452                 (*outbytesleft) -= 5;
 453                 (*inbuf)  += 2;
 454                 (*outbuf) += 5;
 455         }
 456
 457         if (*inbytesleft == 1) {
 458                 errno = EINVAL;
 459                 return -1;
 460         }
 461
 462         if (*inbytesleft > 1) {
 463                 errno = E2BIG;
 464                 return -1;
 465         }
 466
 467         return 0;
 468 }
 469
 470
 471 static size_t iconv_copy(void *cd, char **inbuf, size_t *inbytesleft,
 472                          char **outbuf, size_t *outbytesleft)
 473 {
 474         int n;
 475
 476         n = MIN(*inbytesleft, *outbytesleft);
 477
 478         memmove(*outbuf, *inbuf, n);
 479
 480         (*inbytesleft) -= n;
 481         (*outbytesleft) -= n;
 482         (*inbuf) += n;
 483         (*outbuf) += n;
 484
 485         if (*inbytesleft > 0) {
 486                 errno = E2BIG;
 487                 return -1;
 488         }
 489
 490         return 0;
 491 }
 492
 493 static size_t utf8_pull(void *cd, char **inbuf, size_t *inbytesleft,
 494                          char **outbuf, size_t *outbytesleft)
 495 {
 496         while (*inbytesleft >= 1 && *outbytesleft >= 2) {
 497                 unsigned char *c = (unsigned char *)*inbuf;
 498                 unsigned char *uc = (unsigned char *)*outbuf;
 499                 int len = 1;
 500
 501                 if ((c[0] & 0x80) == 0) {
 502                         uc[0] = c[0];
 503                         uc[1] = 0;
 504                 } else if ((c[0] & 0xf0) == 0xe0) {
 505                         if (*inbytesleft < 3) {
 506                                 DEBUG(0,("short utf8 char\n"));
 507                                 goto badseq;
 508                         }
 509                         uc[1] = ((c[0]&0xF)<<4) | ((c[1]>>2)&0xF);
 510                         uc[0] = (c[1]<<6) | (c[2]&0x3f);
 511                         len = 3;
 512                 } else if ((c[0] & 0xe0) == 0xc0) {
 513                         if (*inbytesleft < 2) {
 514                                 DEBUG(0,("short utf8 char\n"));
 515                                 goto badseq;
 516                         }
 517                         uc[1] = (c[0]>>2) & 0x7;
 518                         uc[0] = (c[0]<<6) | (c[1]&0x3f);
 519                         len = 2;
 520                 }
 521
 522                 (*inbuf)  += len;
 523                 (*inbytesleft)  -= len;
 524                 (*outbytesleft) -= 2;
 525                 (*outbuf) += 2;
 526         }
 527
 528         if (*inbytesleft > 0) {
 529                 errno = E2BIG;
 530                 return -1;
 531         }
 532
 533         return 0;
 534
 535 badseq:
 536         errno = EINVAL;
 537         return -1;
 538 }
 539
 540 static size_t utf8_push(void *cd, char **inbuf, size_t *inbytesleft,
 541                          char **outbuf, size_t *outbytesleft)
 542 {
 543         while (*inbytesleft >= 2 && *outbytesleft >= 1) {
 544                 unsigned char *c = (unsigned char *)*outbuf;
 545                 unsigned char *uc = (unsigned char *)*inbuf;
 546                 int len=1;
 547
 548                 if (uc[1] & 0xf8) {
 549                         if (*outbytesleft < 3) {
 550                                 DEBUG(0,("short utf8 write\n"));
 551                                 goto toobig;
 552                         }
 553                         c[0] = 0xe0 | (uc[1]>>4);
 554                         c[1] = 0x80 | ((uc[1]&0xF)<<2) | (uc[0]>>6);
 555                         c[2] = 0x80 | (uc[0]&0x3f);
 556                         len = 3;
 557                 } else if (uc[1] | (uc[0] & 0x80)) {
 558                         if (*outbytesleft < 2) {
 559                                 DEBUG(0,("short utf8 write\n"));
 560                                 goto toobig;
 561                         }
 562                         c[0] = 0xc0 | (uc[1]<<2) | (uc[0]>>6);
 563                         c[1] = 0x80 | (uc[0]&0x3f);
 564                         len = 2;
 565                 } else {
 566                         c[0] = uc[0];
 567                 }
 568
 569
 570                 (*inbytesleft)  -= 2;
 571                 (*outbytesleft) -= len;
 572                 (*inbuf)  += 2;
 573                 (*outbuf) += len;
 574         }
 575
 576         if (*inbytesleft == 1) {
 577                 errno = EINVAL;
 578                 return -1;
 579         }
 580
 581         if (*inbytesleft > 1) {
 582                 errno = E2BIG;
 583                 return -1;
 584         }
 585
 586         return 0;
 587
 588 toobig:
 589         errno = E2BIG;
 590         return -1;
 591 }
 592