| 1 | /* Character set conversion. |
| 2 | Copyright (C) 1999-2001, 2007 Free Software Foundation, Inc. |
| 3 | |
| 4 | This program is free software; you can redistribute it and/or modify |
| 5 | it under the terms of the GNU Lesser General Public License as published by |
| 6 | the Free Software Foundation; either version 2, or (at your option) |
| 7 | any later version. |
| 8 | |
| 9 | This program is distributed in the hope that it will be useful, |
| 10 | but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 11 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 12 | GNU Lesser General Public License for more details. |
| 13 | |
| 14 | You should have received a copy of the GNU Lesser General Public License along |
| 15 | with this program; if not, write to the Free Software Foundation, |
| 16 | Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. */ |
| 17 | |
| 18 | #include <config.h> |
| 19 | |
| 20 | /* Specification. */ |
| 21 | #include <iconv.h> |
| 22 | |
| 23 | #include <stddef.h> |
| 24 | |
| 25 | #if REPLACE_ICONV_UTF |
| 26 | # include <errno.h> |
| 27 | # include <stdint.h> |
| 28 | # include <stdlib.h> |
| 29 | # include "unistr.h" |
| 30 | # ifndef uintptr_t |
| 31 | # define uintptr_t unsigned long |
| 32 | # endif |
| 33 | #endif |
| 34 | |
| 35 | #if REPLACE_ICONV_UTF |
| 36 | |
| 37 | /* UTF-{16,32}{BE,LE} converters taken from GNU libiconv 1.11. */ |
| 38 | |
| 39 | /* Return code if invalid. (xxx_mbtowc) */ |
| 40 | # define RET_ILSEQ -1 |
| 41 | /* Return code if no bytes were read. (xxx_mbtowc) */ |
| 42 | # define RET_TOOFEW -2 |
| 43 | |
| 44 | /* Return code if invalid. (xxx_wctomb) */ |
| 45 | # define RET_ILUNI -1 |
| 46 | /* Return code if output buffer is too small. (xxx_wctomb, xxx_reset) */ |
| 47 | # define RET_TOOSMALL -2 |
| 48 | |
| 49 | /* |
| 50 | * UTF-16BE |
| 51 | */ |
| 52 | |
| 53 | /* Specification: RFC 2781 */ |
| 54 | |
| 55 | static int |
| 56 | utf16be_mbtowc (ucs4_t *pwc, const unsigned char *s, size_t n) |
| 57 | { |
| 58 | if (n >= 2) |
| 59 | { |
| 60 | ucs4_t wc = (s[0] << 8) + s[1]; |
| 61 | if (wc >= 0xd800 && wc < 0xdc00) |
| 62 | { |
| 63 | if (n >= 4) |
| 64 | { |
| 65 | ucs4_t wc2 = (s[2] << 8) + s[3]; |
| 66 | if (!(wc2 >= 0xdc00 && wc2 < 0xe000)) |
| 67 | return RET_ILSEQ; |
| 68 | *pwc = 0x10000 + ((wc - 0xd800) << 10) + (wc2 - 0xdc00); |
| 69 | return 4; |
| 70 | } |
| 71 | } |
| 72 | else if (wc >= 0xdc00 && wc < 0xe000) |
| 73 | { |
| 74 | return RET_ILSEQ; |
| 75 | } |
| 76 | else |
| 77 | { |
| 78 | *pwc = wc; |
| 79 | return 2; |
| 80 | } |
| 81 | } |
| 82 | return RET_TOOFEW; |
| 83 | } |
| 84 | |
| 85 | static int |
| 86 | utf16be_wctomb (unsigned char *r, ucs4_t wc, size_t n) |
| 87 | { |
| 88 | if (!(wc >= 0xd800 && wc < 0xe000)) |
| 89 | { |
| 90 | if (wc < 0x10000) |
| 91 | { |
| 92 | if (n >= 2) |
| 93 | { |
| 94 | r[0] = (unsigned char) (wc >> 8); |
| 95 | r[1] = (unsigned char) wc; |
| 96 | return 2; |
| 97 | } |
| 98 | else |
| 99 | return RET_TOOSMALL; |
| 100 | } |
| 101 | else if (wc < 0x110000) |
| 102 | { |
| 103 | if (n >= 4) |
| 104 | { |
| 105 | ucs4_t wc1 = 0xd800 + ((wc - 0x10000) >> 10); |
| 106 | ucs4_t wc2 = 0xdc00 + ((wc - 0x10000) & 0x3ff); |
| 107 | r[0] = (unsigned char) (wc1 >> 8); |
| 108 | r[1] = (unsigned char) wc1; |
| 109 | r[2] = (unsigned char) (wc2 >> 8); |
| 110 | r[3] = (unsigned char) wc2; |
| 111 | return 4; |
| 112 | } |
| 113 | else |
| 114 | return RET_TOOSMALL; |
| 115 | } |
| 116 | } |
| 117 | return RET_ILUNI; |
| 118 | } |
| 119 | |
| 120 | /* |
| 121 | * UTF-16LE |
| 122 | */ |
| 123 | |
| 124 | /* Specification: RFC 2781 */ |
| 125 | |
| 126 | static int |
| 127 | utf16le_mbtowc (ucs4_t *pwc, const unsigned char *s, size_t n) |
| 128 | { |
| 129 | if (n >= 2) |
| 130 | { |
| 131 | ucs4_t wc = s[0] + (s[1] << 8); |
| 132 | if (wc >= 0xd800 && wc < 0xdc00) |
| 133 | { |
| 134 | if (n >= 4) |
| 135 | { |
| 136 | ucs4_t wc2 = s[2] + (s[3] << 8); |
| 137 | if (!(wc2 >= 0xdc00 && wc2 < 0xe000)) |
| 138 | return RET_ILSEQ; |
| 139 | *pwc = 0x10000 + ((wc - 0xd800) << 10) + (wc2 - 0xdc00); |
| 140 | return 4; |
| 141 | } |
| 142 | } |
| 143 | else if (wc >= 0xdc00 && wc < 0xe000) |
| 144 | { |
| 145 | return RET_ILSEQ; |
| 146 | } |
| 147 | else |
| 148 | { |
| 149 | *pwc = wc; |
| 150 | return 2; |
| 151 | } |
| 152 | } |
| 153 | return RET_TOOFEW; |
| 154 | } |
| 155 | |
| 156 | static int |
| 157 | utf16le_wctomb (unsigned char *r, ucs4_t wc, size_t n) |
| 158 | { |
| 159 | if (!(wc >= 0xd800 && wc < 0xe000)) |
| 160 | { |
| 161 | if (wc < 0x10000) |
| 162 | { |
| 163 | if (n >= 2) |
| 164 | { |
| 165 | r[0] = (unsigned char) wc; |
| 166 | r[1] = (unsigned char) (wc >> 8); |
| 167 | return 2; |
| 168 | } |
| 169 | else |
| 170 | return RET_TOOSMALL; |
| 171 | } |
| 172 | else if (wc < 0x110000) |
| 173 | { |
| 174 | if (n >= 4) |
| 175 | { |
| 176 | ucs4_t wc1 = 0xd800 + ((wc - 0x10000) >> 10); |
| 177 | ucs4_t wc2 = 0xdc00 + ((wc - 0x10000) & 0x3ff); |
| 178 | r[0] = (unsigned char) wc1; |
| 179 | r[1] = (unsigned char) (wc1 >> 8); |
| 180 | r[2] = (unsigned char) wc2; |
| 181 | r[3] = (unsigned char) (wc2 >> 8); |
| 182 | return 4; |
| 183 | } |
| 184 | else |
| 185 | return RET_TOOSMALL; |
| 186 | } |
| 187 | } |
| 188 | return RET_ILUNI; |
| 189 | } |
| 190 | |
| 191 | /* |
| 192 | * UTF-32BE |
| 193 | */ |
| 194 | |
| 195 | /* Specification: Unicode 3.1 Standard Annex #19 */ |
| 196 | |
| 197 | static int |
| 198 | utf32be_mbtowc (ucs4_t *pwc, const unsigned char *s, size_t n) |
| 199 | { |
| 200 | if (n >= 4) |
| 201 | { |
| 202 | ucs4_t wc = (s[0] << 24) + (s[1] << 16) + (s[2] << 8) + s[3]; |
| 203 | if (wc < 0x110000 && !(wc >= 0xd800 && wc < 0xe000)) |
| 204 | { |
| 205 | *pwc = wc; |
| 206 | return 4; |
| 207 | } |
| 208 | else |
| 209 | return RET_ILSEQ; |
| 210 | } |
| 211 | return RET_TOOFEW; |
| 212 | } |
| 213 | |
| 214 | static int |
| 215 | utf32be_wctomb (unsigned char *r, ucs4_t wc, size_t n) |
| 216 | { |
| 217 | if (wc < 0x110000 && !(wc >= 0xd800 && wc < 0xe000)) |
| 218 | { |
| 219 | if (n >= 4) |
| 220 | { |
| 221 | r[0] = 0; |
| 222 | r[1] = (unsigned char) (wc >> 16); |
| 223 | r[2] = (unsigned char) (wc >> 8); |
| 224 | r[3] = (unsigned char) wc; |
| 225 | return 4; |
| 226 | } |
| 227 | else |
| 228 | return RET_TOOSMALL; |
| 229 | } |
| 230 | return RET_ILUNI; |
| 231 | } |
| 232 | |
| 233 | /* |
| 234 | * UTF-32LE |
| 235 | */ |
| 236 | |
| 237 | /* Specification: Unicode 3.1 Standard Annex #19 */ |
| 238 | |
| 239 | static int |
| 240 | utf32le_mbtowc (ucs4_t *pwc, const unsigned char *s, size_t n) |
| 241 | { |
| 242 | if (n >= 4) |
| 243 | { |
| 244 | ucs4_t wc = s[0] + (s[1] << 8) + (s[2] << 16) + (s[3] << 24); |
| 245 | if (wc < 0x110000 && !(wc >= 0xd800 && wc < 0xe000)) |
| 246 | { |
| 247 | *pwc = wc; |
| 248 | return 4; |
| 249 | } |
| 250 | else |
| 251 | return RET_ILSEQ; |
| 252 | } |
| 253 | return RET_TOOFEW; |
| 254 | } |
| 255 | |
| 256 | static int |
| 257 | utf32le_wctomb (unsigned char *r, ucs4_t wc, size_t n) |
| 258 | { |
| 259 | if (wc < 0x110000 && !(wc >= 0xd800 && wc < 0xe000)) |
| 260 | { |
| 261 | if (n >= 4) |
| 262 | { |
| 263 | r[0] = (unsigned char) wc; |
| 264 | r[1] = (unsigned char) (wc >> 8); |
| 265 | r[2] = (unsigned char) (wc >> 16); |
| 266 | r[3] = 0; |
| 267 | return 4; |
| 268 | } |
| 269 | else |
| 270 | return RET_TOOSMALL; |
| 271 | } |
| 272 | return RET_ILUNI; |
| 273 | } |
| 274 | |
| 275 | #endif |
| 276 | |
| 277 | size_t |
| 278 | rpl_iconv (iconv_t cd, |
| 279 | ICONV_CONST char **inbuf, size_t *inbytesleft, |
| 280 | char **outbuf, size_t *outbytesleft) |
| 281 | #undef iconv |
| 282 | { |
| 283 | #if REPLACE_ICONV_UTF |
| 284 | switch ((uintptr_t) cd) |
| 285 | { |
| 286 | { |
| 287 | int (*xxx_wctomb) (unsigned char *, ucs4_t, size_t); |
| 288 | |
| 289 | case (uintptr_t) _ICONV_UTF8_UTF16BE: |
| 290 | xxx_wctomb = utf16be_wctomb; |
| 291 | goto loop_from_utf8; |
| 292 | case (uintptr_t) _ICONV_UTF8_UTF16LE: |
| 293 | xxx_wctomb = utf16le_wctomb; |
| 294 | goto loop_from_utf8; |
| 295 | case (uintptr_t) _ICONV_UTF8_UTF32BE: |
| 296 | xxx_wctomb = utf32be_wctomb; |
| 297 | goto loop_from_utf8; |
| 298 | case (uintptr_t) _ICONV_UTF8_UTF32LE: |
| 299 | xxx_wctomb = utf32le_wctomb; |
| 300 | goto loop_from_utf8; |
| 301 | |
| 302 | loop_from_utf8: |
| 303 | if (inbuf == NULL || *inbuf == NULL) |
| 304 | return 0; |
| 305 | { |
| 306 | ICONV_CONST char *inptr = *inbuf; |
| 307 | size_t inleft = *inbytesleft; |
| 308 | char *outptr = *outbuf; |
| 309 | size_t outleft = *outbytesleft; |
| 310 | size_t res = 0; |
| 311 | while (inleft > 0) |
| 312 | { |
| 313 | ucs4_t uc; |
| 314 | int m = u8_mbtoucr (&uc, (const uint8_t *) inptr, inleft); |
| 315 | if (m <= 0) |
| 316 | { |
| 317 | if (m == -1) |
| 318 | { |
| 319 | errno = EILSEQ; |
| 320 | res = (size_t)(-1); |
| 321 | break; |
| 322 | } |
| 323 | if (m == -2) |
| 324 | { |
| 325 | errno = EINVAL; |
| 326 | res = (size_t)(-1); |
| 327 | break; |
| 328 | } |
| 329 | abort (); |
| 330 | } |
| 331 | else |
| 332 | { |
| 333 | int n = xxx_wctomb ((uint8_t *) outptr, uc, outleft); |
| 334 | if (n < 0) |
| 335 | { |
| 336 | if (n == RET_ILUNI) |
| 337 | { |
| 338 | errno = EILSEQ; |
| 339 | res = (size_t)(-1); |
| 340 | break; |
| 341 | } |
| 342 | if (n == RET_TOOSMALL) |
| 343 | { |
| 344 | errno = E2BIG; |
| 345 | res = (size_t)(-1); |
| 346 | break; |
| 347 | } |
| 348 | abort (); |
| 349 | } |
| 350 | else |
| 351 | { |
| 352 | inptr += m; |
| 353 | inleft -= m; |
| 354 | outptr += n; |
| 355 | outleft -= n; |
| 356 | } |
| 357 | } |
| 358 | } |
| 359 | *inbuf = inptr; |
| 360 | *inbytesleft = inleft; |
| 361 | *outbuf = outptr; |
| 362 | *outbytesleft = outleft; |
| 363 | return res; |
| 364 | } |
| 365 | } |
| 366 | |
| 367 | { |
| 368 | int (*xxx_mbtowc) (ucs4_t *, const unsigned char *, size_t); |
| 369 | |
| 370 | case (uintptr_t) _ICONV_UTF16BE_UTF8: |
| 371 | xxx_mbtowc = utf16be_mbtowc; |
| 372 | goto loop_to_utf8; |
| 373 | case (uintptr_t) _ICONV_UTF16LE_UTF8: |
| 374 | xxx_mbtowc = utf16le_mbtowc; |
| 375 | goto loop_to_utf8; |
| 376 | case (uintptr_t) _ICONV_UTF32BE_UTF8: |
| 377 | xxx_mbtowc = utf32be_mbtowc; |
| 378 | goto loop_to_utf8; |
| 379 | case (uintptr_t) _ICONV_UTF32LE_UTF8: |
| 380 | xxx_mbtowc = utf32le_mbtowc; |
| 381 | goto loop_to_utf8; |
| 382 | |
| 383 | loop_to_utf8: |
| 384 | if (inbuf == NULL || *inbuf == NULL) |
| 385 | return 0; |
| 386 | { |
| 387 | ICONV_CONST char *inptr = *inbuf; |
| 388 | size_t inleft = *inbytesleft; |
| 389 | char *outptr = *outbuf; |
| 390 | size_t outleft = *outbytesleft; |
| 391 | size_t res = 0; |
| 392 | while (inleft > 0) |
| 393 | { |
| 394 | ucs4_t uc; |
| 395 | int m = xxx_mbtowc (&uc, (const uint8_t *) inptr, inleft); |
| 396 | if (m <= 0) |
| 397 | { |
| 398 | if (m == RET_ILSEQ) |
| 399 | { |
| 400 | errno = EILSEQ; |
| 401 | res = (size_t)(-1); |
| 402 | break; |
| 403 | } |
| 404 | if (m == RET_TOOFEW) |
| 405 | { |
| 406 | errno = EINVAL; |
| 407 | res = (size_t)(-1); |
| 408 | break; |
| 409 | } |
| 410 | abort (); |
| 411 | } |
| 412 | else |
| 413 | { |
| 414 | int n = u8_uctomb ((uint8_t *) outptr, uc, outleft); |
| 415 | if (n < 0) |
| 416 | { |
| 417 | if (n == -1) |
| 418 | { |
| 419 | errno = EILSEQ; |
| 420 | res = (size_t)(-1); |
| 421 | break; |
| 422 | } |
| 423 | if (n == -2) |
| 424 | { |
| 425 | errno = E2BIG; |
| 426 | res = (size_t)(-1); |
| 427 | break; |
| 428 | } |
| 429 | abort (); |
| 430 | } |
| 431 | else |
| 432 | { |
| 433 | inptr += m; |
| 434 | inleft -= m; |
| 435 | outptr += n; |
| 436 | outleft -= n; |
| 437 | } |
| 438 | } |
| 439 | } |
| 440 | *inbuf = inptr; |
| 441 | *inbytesleft = inleft; |
| 442 | *outbuf = outptr; |
| 443 | *outbytesleft = outleft; |
| 444 | return res; |
| 445 | } |
| 446 | } |
| 447 | } |
| 448 | #endif |
| 449 | return iconv (cd, inbuf, inbytesleft, outbuf, outbytesleft); |
| 450 | } |