| 265 | } |
| 266 | |
| 267 | size_t utf32_to_utf8(const char32_t* buf, size_t len, char* utf8_output) { |
| 268 | const uint32_t* data = reinterpret_cast<const uint32_t*>(buf); |
| 269 | size_t pos = 0; |
| 270 | const char* start{utf8_output}; |
| 271 | while (pos < len) { |
| 272 | // try to convert the next block of 2 ASCII characters |
| 273 | if (pos + 2 <= len) { // if it is safe to read 8 more |
| 274 | // bytes, check that they are ascii |
| 275 | uint64_t v; |
| 276 | std::memcpy(&v, data + pos, sizeof(uint64_t)); |
| 277 | if ((v & 0xFFFFFF80FFFFFF80) == 0) { |
| 278 | *utf8_output++ = char(buf[pos]); |
| 279 | *utf8_output++ = char(buf[pos + 1]); |
| 280 | pos += 2; |
| 281 | continue; |
| 282 | } |
| 283 | } |
| 284 | uint32_t word = data[pos]; |
| 285 | if ((word & 0xFFFFFF80) == 0) { |
| 286 | // will generate one UTF-8 bytes |
| 287 | *utf8_output++ = char(word); |
| 288 | pos++; |
| 289 | } else if ((word & 0xFFFFF800) == 0) { |
| 290 | // will generate two UTF-8 bytes |
| 291 | // we have 0b110XXXXX 0b10XXXXXX |
| 292 | *utf8_output++ = char((word >> 6) | 0b11000000); |
| 293 | *utf8_output++ = char((word & 0b111111) | 0b10000000); |
| 294 | pos++; |
| 295 | } else if ((word & 0xFFFF0000) == 0) { |
| 296 | // will generate three UTF-8 bytes |
| 297 | // we have 0b1110XXXX 0b10XXXXXX 0b10XXXXXX |
| 298 | if (word >= 0xD800 && word <= 0xDFFF) { |
| 299 | return 0; |
| 300 | } |
| 301 | *utf8_output++ = char((word >> 12) | 0b11100000); |
| 302 | *utf8_output++ = char(((word >> 6) & 0b111111) | 0b10000000); |
| 303 | *utf8_output++ = char((word & 0b111111) | 0b10000000); |
| 304 | pos++; |
| 305 | } else { |
| 306 | // will generate four UTF-8 bytes |
| 307 | // we have 0b11110XXX 0b10XXXXXX 0b10XXXXXX |
| 308 | // 0b10XXXXXX |
| 309 | if (word > 0x10FFFF) { |
| 310 | return 0; |
| 311 | } |
| 312 | *utf8_output++ = char((word >> 18) | 0b11110000); |
| 313 | *utf8_output++ = char(((word >> 12) & 0b111111) | 0b10000000); |
| 314 | *utf8_output++ = char(((word >> 6) & 0b111111) | 0b10000000); |
| 315 | *utf8_output++ = char((word & 0b111111) | 0b10000000); |
| 316 | pos++; |
| 317 | } |
| 318 | } |
| 319 | return utf8_output - start; |
| 320 | } |
| 321 | } // namespace ada::idna |
| 322 | /* end file src/unicode_transcoding.cpp */ |
| 323 | /* begin file src/mapping.cpp */ |