| 1 | /* Copyright (C) 2010-2020 The RetroArch team |
| 2 | * |
| 3 | * --------------------------------------------------------------------------------------- |
| 4 | * The following license statement only applies to this file (encoding_utf.c). |
| 5 | * --------------------------------------------------------------------------------------- |
| 6 | * |
| 7 | * Permission is hereby granted, free of charge, |
| 8 | * to any person obtaining a copy of this software and associated documentation files (the "Software"), |
| 9 | * to deal in the Software without restriction, including without limitation the rights to |
| 10 | * use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, |
| 11 | * and to permit persons to whom the Software is furnished to do so, subject to the following conditions: |
| 12 | * |
| 13 | * The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. |
| 14 | * |
| 15 | * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, |
| 16 | * INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, |
| 17 | * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. |
| 18 | * IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, |
| 19 | * WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, |
| 20 | * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. |
| 21 | */ |
| 22 | |
| 23 | #include <stdint.h> |
| 24 | #include <stdlib.h> |
| 25 | #include <stddef.h> |
| 26 | #include <string.h> |
| 27 | |
| 28 | #include <boolean.h> |
| 29 | #include <compat/strl.h> |
| 30 | #include <retro_inline.h> |
| 31 | |
| 32 | #include <encodings/utf.h> |
| 33 | |
| 34 | #if defined(_WIN32) && !defined(_XBOX) |
| 35 | #include <windows.h> |
| 36 | #elif defined(_XBOX) |
| 37 | #include <xtl.h> |
| 38 | #endif |
| 39 | |
| 40 | #define UTF8_WALKBYTE(string) (*((*(string))++)) |
| 41 | |
| 42 | static unsigned leading_ones(uint8_t c) |
| 43 | { |
| 44 | unsigned ones = 0; |
| 45 | while (c & 0x80) |
| 46 | { |
| 47 | ones++; |
| 48 | c <<= 1; |
| 49 | } |
| 50 | |
| 51 | return ones; |
| 52 | } |
| 53 | |
| 54 | /* Simple implementation. Assumes the sequence is |
| 55 | * properly synchronized and terminated. */ |
| 56 | |
| 57 | size_t utf8_conv_utf32(uint32_t *out, size_t out_chars, |
| 58 | const char *in, size_t in_size) |
| 59 | { |
| 60 | unsigned i; |
| 61 | size_t ret = 0; |
| 62 | while (in_size && out_chars) |
| 63 | { |
| 64 | unsigned extra, shift; |
| 65 | uint32_t c; |
| 66 | uint8_t first = *in++; |
| 67 | unsigned ones = leading_ones(first); |
| 68 | |
| 69 | if (ones > 6 || ones == 1) /* Invalid or desync. */ |
| 70 | break; |
| 71 | |
| 72 | extra = ones ? ones - 1 : ones; |
| 73 | if (1 + extra > in_size) /* Overflow. */ |
| 74 | break; |
| 75 | |
| 76 | shift = (extra - 1) * 6; |
| 77 | c = (first & ((1 << (7 - ones)) - 1)) << (6 * extra); |
| 78 | |
| 79 | for (i = 0; i < extra; i++, in++, shift -= 6) |
| 80 | c |= (*in & 0x3f) << shift; |
| 81 | |
| 82 | *out++ = c; |
| 83 | in_size -= 1 + extra; |
| 84 | out_chars--; |
| 85 | ret++; |
| 86 | } |
| 87 | |
| 88 | return ret; |
| 89 | } |
| 90 | |
| 91 | bool utf16_conv_utf8(uint8_t *out, size_t *out_chars, |
| 92 | const uint16_t *in, size_t in_size) |
| 93 | { |
| 94 | size_t out_pos = 0; |
| 95 | size_t in_pos = 0; |
| 96 | static const |
| 97 | uint8_t utf8_limits[5] = { 0xC0, 0xE0, 0xF0, 0xF8, 0xFC }; |
| 98 | |
| 99 | for (;;) |
| 100 | { |
| 101 | unsigned num_adds; |
| 102 | uint32_t value; |
| 103 | |
| 104 | if (in_pos == in_size) |
| 105 | { |
| 106 | *out_chars = out_pos; |
| 107 | return true; |
| 108 | } |
| 109 | value = in[in_pos++]; |
| 110 | if (value < 0x80) |
| 111 | { |
| 112 | if (out) |
| 113 | out[out_pos] = (char)value; |
| 114 | out_pos++; |
| 115 | continue; |
| 116 | } |
| 117 | |
| 118 | if (value >= 0xD800 && value < 0xE000) |
| 119 | { |
| 120 | uint32_t c2; |
| 121 | |
| 122 | if (value >= 0xDC00 || in_pos == in_size) |
| 123 | break; |
| 124 | c2 = in[in_pos++]; |
| 125 | if (c2 < 0xDC00 || c2 >= 0xE000) |
| 126 | break; |
| 127 | value = (((value - 0xD800) << 10) | (c2 - 0xDC00)) + 0x10000; |
| 128 | } |
| 129 | |
| 130 | for (num_adds = 1; num_adds < 5; num_adds++) |
| 131 | if (value < (((uint32_t)1) << (num_adds * 5 + 6))) |
| 132 | break; |
| 133 | if (out) |
| 134 | out[out_pos] = (char)(utf8_limits[num_adds - 1] |
| 135 | + (value >> (6 * num_adds))); |
| 136 | out_pos++; |
| 137 | do |
| 138 | { |
| 139 | num_adds--; |
| 140 | if (out) |
| 141 | out[out_pos] = (char)(0x80 |
| 142 | + ((value >> (6 * num_adds)) & 0x3F)); |
| 143 | out_pos++; |
| 144 | }while (num_adds != 0); |
| 145 | } |
| 146 | |
| 147 | *out_chars = out_pos; |
| 148 | return false; |
| 149 | } |
| 150 | |
| 151 | /* Acts mostly like strlcpy. |
| 152 | * |
| 153 | * Copies the given number of UTF-8 characters, |
| 154 | * but at most d_len bytes. |
| 155 | * |
| 156 | * Always NULL terminates. |
| 157 | * Does not copy half a character. |
| 158 | * |
| 159 | * Returns number of bytes. 's' is assumed valid UTF-8. |
| 160 | * Use only if 'chars' is considerably less than 'd_len'. */ |
| 161 | size_t utf8cpy(char *d, size_t d_len, const char *s, size_t chars) |
| 162 | { |
| 163 | const uint8_t *sb = (const uint8_t*)s; |
| 164 | const uint8_t *sb_org = sb; |
| 165 | |
| 166 | if (!s) |
| 167 | return 0; |
| 168 | |
| 169 | while (*sb && chars-- > 0) |
| 170 | { |
| 171 | sb++; |
| 172 | while ((*sb & 0xC0) == 0x80) |
| 173 | sb++; |
| 174 | } |
| 175 | |
| 176 | if ((size_t)(sb - sb_org) > d_len-1 /* NUL */) |
| 177 | { |
| 178 | sb = sb_org + d_len-1; |
| 179 | while ((*sb & 0xC0) == 0x80) |
| 180 | sb--; |
| 181 | } |
| 182 | |
| 183 | memcpy(d, sb_org, sb-sb_org); |
| 184 | d[sb-sb_org] = '\0'; |
| 185 | |
| 186 | return sb-sb_org; |
| 187 | } |
| 188 | |
| 189 | const char *utf8skip(const char *str, size_t chars) |
| 190 | { |
| 191 | const uint8_t *strb = (const uint8_t*)str; |
| 192 | |
| 193 | if (!chars) |
| 194 | return str; |
| 195 | |
| 196 | do |
| 197 | { |
| 198 | strb++; |
| 199 | while ((*strb & 0xC0)==0x80) |
| 200 | strb++; |
| 201 | chars--; |
| 202 | }while (chars); |
| 203 | |
| 204 | return (const char*)strb; |
| 205 | } |
| 206 | |
| 207 | size_t utf8len(const char *string) |
| 208 | { |
| 209 | size_t ret = 0; |
| 210 | |
| 211 | if (!string) |
| 212 | return 0; |
| 213 | |
| 214 | while (*string) |
| 215 | { |
| 216 | if ((*string & 0xC0) != 0x80) |
| 217 | ret++; |
| 218 | string++; |
| 219 | } |
| 220 | return ret; |
| 221 | } |
| 222 | |
| 223 | /* Does not validate the input, returns garbage if it's not UTF-8. */ |
| 224 | uint32_t utf8_walk(const char **string) |
| 225 | { |
| 226 | uint8_t first = UTF8_WALKBYTE(string); |
| 227 | uint32_t ret = 0; |
| 228 | |
| 229 | if (first < 128) |
| 230 | return first; |
| 231 | |
| 232 | ret = (ret << 6) | (UTF8_WALKBYTE(string) & 0x3F); |
| 233 | if (first >= 0xE0) |
| 234 | { |
| 235 | ret = (ret << 6) | (UTF8_WALKBYTE(string) & 0x3F); |
| 236 | if (first >= 0xF0) |
| 237 | { |
| 238 | ret = (ret << 6) | (UTF8_WALKBYTE(string) & 0x3F); |
| 239 | return ret | (first & 7) << 18; |
| 240 | } |
| 241 | return ret | (first & 15) << 12; |
| 242 | } |
| 243 | |
| 244 | return ret | (first & 31) << 6; |
| 245 | } |
| 246 | |
| 247 | static bool utf16_to_char(uint8_t **utf_data, |
| 248 | size_t *dest_len, const uint16_t *in) |
| 249 | { |
| 250 | unsigned len = 0; |
| 251 | |
| 252 | while (in[len] != '\0') |
| 253 | len++; |
| 254 | |
| 255 | utf16_conv_utf8(NULL, dest_len, in, len); |
| 256 | *dest_len += 1; |
| 257 | *utf_data = (uint8_t*)malloc(*dest_len); |
| 258 | if (*utf_data == 0) |
| 259 | return false; |
| 260 | |
| 261 | return utf16_conv_utf8(*utf_data, dest_len, in, len); |
| 262 | } |
| 263 | |
| 264 | bool utf16_to_char_string(const uint16_t *in, char *s, size_t len) |
| 265 | { |
| 266 | size_t dest_len = 0; |
| 267 | uint8_t *utf16_data = NULL; |
| 268 | bool ret = utf16_to_char(&utf16_data, &dest_len, in); |
| 269 | |
| 270 | if (ret) |
| 271 | { |
| 272 | utf16_data[dest_len] = 0; |
| 273 | strlcpy(s, (const char*)utf16_data, len); |
| 274 | } |
| 275 | |
| 276 | free(utf16_data); |
| 277 | utf16_data = NULL; |
| 278 | |
| 279 | return ret; |
| 280 | } |
| 281 | |
| 282 | #if defined(_WIN32) && !defined(_XBOX) && !defined(UNICODE) |
| 283 | /* Returned pointer MUST be freed by the caller if non-NULL. */ |
| 284 | static char *mb_to_mb_string_alloc(const char *str, |
| 285 | enum CodePage cp_in, enum CodePage cp_out) |
| 286 | { |
| 287 | wchar_t *path_buf_wide = NULL; |
| 288 | int path_buf_wide_len = MultiByteToWideChar(cp_in, 0, str, -1, NULL, 0); |
| 289 | |
| 290 | /* Windows 95 will return 0 from these functions with |
| 291 | * a UTF8 codepage set without MSLU. |
| 292 | * |
| 293 | * From an unknown MSDN version (others omit this info): |
| 294 | * - CP_UTF8 Windows 98/Me, Windows NT 4.0 and later: |
| 295 | * Translate using UTF-8. When this is set, dwFlags must be zero. |
| 296 | * - Windows 95: Under the Microsoft Layer for Unicode, |
| 297 | * MultiByteToWideChar also supports CP_UTF7 and CP_UTF8. |
| 298 | */ |
| 299 | |
| 300 | if (!path_buf_wide_len) |
| 301 | return strdup(str); |
| 302 | |
| 303 | path_buf_wide = (wchar_t*) |
| 304 | calloc(path_buf_wide_len + sizeof(wchar_t), sizeof(wchar_t)); |
| 305 | |
| 306 | if (path_buf_wide) |
| 307 | { |
| 308 | MultiByteToWideChar(cp_in, 0, |
| 309 | str, -1, path_buf_wide, path_buf_wide_len); |
| 310 | |
| 311 | if (*path_buf_wide) |
| 312 | { |
| 313 | int path_buf_len = WideCharToMultiByte(cp_out, 0, |
| 314 | path_buf_wide, -1, NULL, 0, NULL, NULL); |
| 315 | |
| 316 | if (path_buf_len) |
| 317 | { |
| 318 | char *path_buf = (char*) |
| 319 | calloc(path_buf_len + sizeof(char), sizeof(char)); |
| 320 | |
| 321 | if (path_buf) |
| 322 | { |
| 323 | WideCharToMultiByte(cp_out, 0, |
| 324 | path_buf_wide, -1, path_buf, |
| 325 | path_buf_len, NULL, NULL); |
| 326 | |
| 327 | free(path_buf_wide); |
| 328 | |
| 329 | if (*path_buf) |
| 330 | return path_buf; |
| 331 | |
| 332 | free(path_buf); |
| 333 | return NULL; |
| 334 | } |
| 335 | } |
| 336 | else |
| 337 | { |
| 338 | free(path_buf_wide); |
| 339 | return strdup(str); |
| 340 | } |
| 341 | } |
| 342 | |
| 343 | free(path_buf_wide); |
| 344 | } |
| 345 | |
| 346 | return NULL; |
| 347 | } |
| 348 | #endif |
| 349 | |
| 350 | /* Returned pointer MUST be freed by the caller if non-NULL. */ |
| 351 | char* utf8_to_local_string_alloc(const char *str) |
| 352 | { |
| 353 | if (str && *str) |
| 354 | { |
| 355 | #if defined(_WIN32) && !defined(_XBOX) && !defined(UNICODE) |
| 356 | return mb_to_mb_string_alloc(str, CODEPAGE_UTF8, CODEPAGE_LOCAL); |
| 357 | #else |
| 358 | /* assume string needs no modification if not on Windows */ |
| 359 | return strdup(str); |
| 360 | #endif |
| 361 | } |
| 362 | return NULL; |
| 363 | } |
| 364 | |
| 365 | /* Returned pointer MUST be freed by the caller if non-NULL. */ |
| 366 | char* local_to_utf8_string_alloc(const char *str) |
| 367 | { |
| 368 | if (str && *str) |
| 369 | { |
| 370 | #if defined(_WIN32) && !defined(_XBOX) && !defined(UNICODE) |
| 371 | return mb_to_mb_string_alloc(str, CODEPAGE_LOCAL, CODEPAGE_UTF8); |
| 372 | #else |
| 373 | /* assume string needs no modification if not on Windows */ |
| 374 | return strdup(str); |
| 375 | #endif |
| 376 | } |
| 377 | return NULL; |
| 378 | } |
| 379 | |
| 380 | /* Returned pointer MUST be freed by the caller if non-NULL. */ |
| 381 | wchar_t* utf8_to_utf16_string_alloc(const char *str) |
| 382 | { |
| 383 | #ifdef _WIN32 |
| 384 | int len = 0; |
| 385 | int out_len = 0; |
| 386 | #else |
| 387 | size_t len = 0; |
| 388 | size_t out_len = 0; |
| 389 | #endif |
| 390 | wchar_t *buf = NULL; |
| 391 | |
| 392 | if (!str || !*str) |
| 393 | return NULL; |
| 394 | |
| 395 | #ifdef _WIN32 |
| 396 | len = MultiByteToWideChar(CP_UTF8, 0, str, -1, NULL, 0); |
| 397 | |
| 398 | if (len) |
| 399 | { |
| 400 | buf = (wchar_t*)calloc(len, sizeof(wchar_t)); |
| 401 | |
| 402 | if (!buf) |
| 403 | return NULL; |
| 404 | |
| 405 | out_len = MultiByteToWideChar(CP_UTF8, 0, str, -1, buf, len); |
| 406 | } |
| 407 | else |
| 408 | { |
| 409 | /* fallback to ANSI codepage instead */ |
| 410 | len = MultiByteToWideChar(CP_ACP, 0, str, -1, NULL, 0); |
| 411 | |
| 412 | if (len) |
| 413 | { |
| 414 | buf = (wchar_t*)calloc(len, sizeof(wchar_t)); |
| 415 | |
| 416 | if (!buf) |
| 417 | return NULL; |
| 418 | |
| 419 | out_len = MultiByteToWideChar(CP_ACP, 0, str, -1, buf, len); |
| 420 | } |
| 421 | } |
| 422 | |
| 423 | if (out_len < 0) |
| 424 | { |
| 425 | free(buf); |
| 426 | return NULL; |
| 427 | } |
| 428 | #else |
| 429 | /* NOTE: For now, assume non-Windows platforms' locale is already UTF-8. */ |
| 430 | len = mbstowcs(NULL, str, 0) + 1; |
| 431 | |
| 432 | if (len) |
| 433 | { |
| 434 | buf = (wchar_t*)calloc(len, sizeof(wchar_t)); |
| 435 | |
| 436 | if (!buf) |
| 437 | return NULL; |
| 438 | |
| 439 | out_len = mbstowcs(buf, str, len); |
| 440 | } |
| 441 | |
| 442 | if (out_len == (size_t)-1) |
| 443 | { |
| 444 | free(buf); |
| 445 | return NULL; |
| 446 | } |
| 447 | #endif |
| 448 | |
| 449 | return buf; |
| 450 | } |
| 451 | |
| 452 | /* Returned pointer MUST be freed by the caller if non-NULL. */ |
| 453 | char* utf16_to_utf8_string_alloc(const wchar_t *str) |
| 454 | { |
| 455 | #ifdef _WIN32 |
| 456 | int len = 0; |
| 457 | #else |
| 458 | size_t len = 0; |
| 459 | #endif |
| 460 | char *buf = NULL; |
| 461 | |
| 462 | if (!str || !*str) |
| 463 | return NULL; |
| 464 | |
| 465 | #ifdef _WIN32 |
| 466 | { |
| 467 | UINT code_page = CP_UTF8; |
| 468 | len = WideCharToMultiByte(code_page, |
| 469 | 0, str, -1, NULL, 0, NULL, NULL); |
| 470 | |
| 471 | /* fallback to ANSI codepage instead */ |
| 472 | if (!len) |
| 473 | { |
| 474 | code_page = CP_ACP; |
| 475 | len = WideCharToMultiByte(code_page, |
| 476 | 0, str, -1, NULL, 0, NULL, NULL); |
| 477 | } |
| 478 | |
| 479 | buf = (char*)calloc(len, sizeof(char)); |
| 480 | |
| 481 | if (!buf) |
| 482 | return NULL; |
| 483 | |
| 484 | if (WideCharToMultiByte(code_page, |
| 485 | 0, str, -1, buf, len, NULL, NULL) < 0) |
| 486 | { |
| 487 | free(buf); |
| 488 | return NULL; |
| 489 | } |
| 490 | } |
| 491 | #else |
| 492 | /* NOTE: For now, assume non-Windows platforms' |
| 493 | * locale is already UTF-8. */ |
| 494 | len = wcstombs(NULL, str, 0) + 1; |
| 495 | |
| 496 | if (len) |
| 497 | { |
| 498 | buf = (char*)calloc(len, sizeof(char)); |
| 499 | |
| 500 | if (!buf) |
| 501 | return NULL; |
| 502 | |
| 503 | if (wcstombs(buf, str, len) == (size_t)-1) |
| 504 | { |
| 505 | free(buf); |
| 506 | return NULL; |
| 507 | } |
| 508 | } |
| 509 | #endif |
| 510 | |
| 511 | return buf; |
| 512 | } |