| 1 | /* |
| 2 | Open Asset Import Library (assimp) |
| 3 | ---------------------------------------------------------------------- |
| 4 | |
| 5 | Copyright (c) 2006-2025, assimp team |
| 6 | |
| 7 | All rights reserved. |
| 8 | |
| 9 | Redistribution and use of this software in source and binary forms, |
| 10 | with or without modification, are permitted provided that the |
| 11 | following conditions are met: |
| 12 | |
| 13 | * Redistributions of source code must retain the above |
| 14 | copyright notice, this list of conditions and the |
| 15 | following disclaimer. |
| 16 | |
| 17 | * Redistributions in binary form must reproduce the above |
| 18 | copyright notice, this list of conditions and the |
| 19 | following disclaimer in the documentation and/or other |
| 20 | materials provided with the distribution. |
| 21 | |
| 22 | * Neither the name of the assimp team, nor the names of its |
| 23 | contributors may be used to endorse or promote products |
| 24 | derived from this software without specific prior |
| 25 | written permission of the assimp team. |
| 26 | |
| 27 | THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS |
| 28 | "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT |
| 29 | LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR |
| 30 | A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT |
| 31 | OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, |
| 32 | SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT |
| 33 | LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, |
| 34 | DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY |
| 35 | THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT |
| 36 | (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE |
| 37 | OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. |
| 38 | |
| 39 | ---------------------------------------------------------------------- |
| 40 | */ |
| 41 | /** @file FBXBinaryTokenizer.cpp |
| 42 | * @brief Implementation of a fake lexer for binary fbx files - |
| 43 | * we emit tokens so the parser needs almost no special handling |
| 44 | * for binary files. |
| 45 | */ |
| 46 | |
| 47 | #ifndef ASSIMP_BUILD_NO_FBX_IMPORTER |
| 48 | |
| 49 | #include "FBXTokenizer.h" |
| 50 | #include "FBXUtil.h" |
| 51 | #include <assimp/defs.h> |
| 52 | #include <stdint.h> |
| 53 | #include <cstdint> |
| 54 | #include <assimp/Exceptional.h> |
| 55 | #include <assimp/ByteSwapper.h> |
| 56 | #include <assimp/DefaultLogger.hpp> |
| 57 | #include <assimp/StringUtils.h> |
| 58 | |
| 59 | namespace Assimp { |
| 60 | namespace FBX { |
| 61 | |
| 62 | // ------------------------------------------------------------------------------------------------ |
| 63 | Token::Token(const char* sbegin, const char* send, TokenType type, size_t offset) : |
| 64 | #ifdef DEBUG |
| 65 | contents(sbegin, static_cast<size_t>(send-sbegin)), |
| 66 | #endif |
| 67 | sbegin(sbegin), |
| 68 | send(send), |
| 69 | type(type), |
| 70 | line(offset), |
| 71 | column(BINARY_MARKER) { |
| 72 | ai_assert(sbegin); |
| 73 | ai_assert(send); |
| 74 | |
| 75 | // binary tokens may have zero length because they are sometimes dummies |
| 76 | // inserted by TokenizeBinary() |
| 77 | ai_assert(send >= sbegin); |
| 78 | } |
| 79 | |
| 80 | |
| 81 | namespace { |
| 82 | |
| 83 | // ------------------------------------------------------------------------------------------------ |
| 84 | // signal tokenization error, this is always unrecoverable. Throws DeadlyImportError. |
| 85 | AI_WONT_RETURN void TokenizeError(const std::string& message, size_t offset) AI_WONT_RETURN_SUFFIX; |
| 86 | AI_WONT_RETURN void TokenizeError(const std::string& message, size_t offset) |
| 87 | { |
| 88 | throw DeadlyImportError("FBX-Tokenize" , Util::GetOffsetText(offset), message); |
| 89 | } |
| 90 | |
| 91 | |
| 92 | // ------------------------------------------------------------------------------------------------ |
| 93 | size_t Offset(const char* begin, const char* cursor) { |
| 94 | ai_assert(begin <= cursor); |
| 95 | |
| 96 | return cursor - begin; |
| 97 | } |
| 98 | |
| 99 | // ------------------------------------------------------------------------------------------------ |
| 100 | AI_WONT_RETURN void TokenizeError(const std::string& message, const char* begin, const char* cursor) AI_WONT_RETURN_SUFFIX; |
| 101 | void TokenizeError(const std::string& message, const char* begin, const char* cursor) { |
| 102 | TokenizeError(message, offset: Offset(begin, cursor)); |
| 103 | } |
| 104 | |
| 105 | // ------------------------------------------------------------------------------------------------ |
| 106 | uint32_t ReadWord(const char* input, const char*& cursor, const char* end) { |
| 107 | const size_t k_to_read = sizeof( uint32_t ); |
| 108 | if(Offset(begin: cursor, cursor: end) < k_to_read ) { |
| 109 | TokenizeError(message: "cannot ReadWord, out of bounds" ,begin: input, cursor); |
| 110 | } |
| 111 | |
| 112 | uint32_t word; |
| 113 | ::memcpy(dest: &word, src: cursor, n: 4); |
| 114 | AI_SWAP4(word); |
| 115 | |
| 116 | cursor += k_to_read; |
| 117 | |
| 118 | return word; |
| 119 | } |
| 120 | |
| 121 | // ------------------------------------------------------------------------------------------------ |
| 122 | uint64_t ReadDoubleWord(const char* input, const char*& cursor, const char* end) { |
| 123 | const size_t k_to_read = sizeof(uint64_t); |
| 124 | if(Offset(begin: cursor, cursor: end) < k_to_read) { |
| 125 | TokenizeError(message: "cannot ReadDoubleWord, out of bounds" ,begin: input, cursor); |
| 126 | } |
| 127 | |
| 128 | uint64_t dword /*= *reinterpret_cast<const uint64_t*>(cursor)*/; |
| 129 | ::memcpy( dest: &dword, src: cursor, n: sizeof( uint64_t ) ); |
| 130 | AI_SWAP8(dword); |
| 131 | |
| 132 | cursor += k_to_read; |
| 133 | |
| 134 | return dword; |
| 135 | } |
| 136 | |
| 137 | // ------------------------------------------------------------------------------------------------ |
| 138 | uint8_t ReadByte(const char* input, const char*& cursor, const char* end) { |
| 139 | if(Offset(begin: cursor, cursor: end) < sizeof( uint8_t ) ) { |
| 140 | TokenizeError(message: "cannot ReadByte, out of bounds" ,begin: input, cursor); |
| 141 | } |
| 142 | |
| 143 | uint8_t word;/* = *reinterpret_cast< const uint8_t* >( cursor )*/ |
| 144 | ::memcpy( dest: &word, src: cursor, n: sizeof( uint8_t ) ); |
| 145 | ++cursor; |
| 146 | |
| 147 | return word; |
| 148 | } |
| 149 | |
| 150 | // ------------------------------------------------------------------------------------------------ |
| 151 | unsigned int ReadString(const char*& sbegin_out, const char*& send_out, const char* input, |
| 152 | const char*& cursor, const char* end, bool long_length = false, bool allow_null = false) { |
| 153 | const uint32_t len_len = long_length ? 4 : 1; |
| 154 | if(Offset(begin: cursor, cursor: end) < len_len) { |
| 155 | TokenizeError(message: "cannot ReadString, out of bounds reading length" ,begin: input, cursor); |
| 156 | } |
| 157 | |
| 158 | const uint32_t length = long_length ? ReadWord(input, cursor, end) : ReadByte(input, cursor, end); |
| 159 | |
| 160 | if (Offset(begin: cursor, cursor: end) < length) { |
| 161 | TokenizeError(message: "cannot ReadString, length is out of bounds" ,begin: input, cursor); |
| 162 | } |
| 163 | |
| 164 | sbegin_out = cursor; |
| 165 | cursor += length; |
| 166 | |
| 167 | send_out = cursor; |
| 168 | |
| 169 | if(!allow_null) { |
| 170 | for (unsigned int i = 0; i < length; ++i) { |
| 171 | if(sbegin_out[i] == '\0') { |
| 172 | TokenizeError(message: "failed ReadString, unexpected NUL character in string" ,begin: input, cursor); |
| 173 | } |
| 174 | } |
| 175 | } |
| 176 | |
| 177 | return length; |
| 178 | } |
| 179 | |
| 180 | // ------------------------------------------------------------------------------------------------ |
| 181 | void ReadData(const char*& sbegin_out, const char*& send_out, const char* input, const char*& cursor, const char* end) { |
| 182 | if(Offset(begin: cursor, cursor: end) < 1) { |
| 183 | TokenizeError(message: "cannot ReadData, out of bounds reading length" ,begin: input, cursor); |
| 184 | } |
| 185 | |
| 186 | const char type = *cursor; |
| 187 | sbegin_out = cursor++; |
| 188 | |
| 189 | switch(type) |
| 190 | { |
| 191 | // 16 bit int |
| 192 | case 'Y': |
| 193 | cursor += 2; |
| 194 | break; |
| 195 | |
| 196 | // 1 bit bool flag (yes/no) |
| 197 | case 'C': |
| 198 | cursor += 1; |
| 199 | break; |
| 200 | |
| 201 | // 32 bit int |
| 202 | case 'I': |
| 203 | // <- fall through |
| 204 | |
| 205 | // float |
| 206 | case 'F': |
| 207 | cursor += 4; |
| 208 | break; |
| 209 | |
| 210 | // double |
| 211 | case 'D': |
| 212 | cursor += 8; |
| 213 | break; |
| 214 | |
| 215 | // 64 bit int |
| 216 | case 'L': |
| 217 | cursor += 8; |
| 218 | break; |
| 219 | |
| 220 | // note: do not write cursor += ReadWord(...cursor) as this would be UB |
| 221 | |
| 222 | // raw binary data |
| 223 | case 'R': |
| 224 | { |
| 225 | const uint32_t length = ReadWord(input, cursor, end); |
| 226 | cursor += length; |
| 227 | break; |
| 228 | } |
| 229 | |
| 230 | case 'b': |
| 231 | // TODO: what is the 'b' type code? Right now we just skip over it / |
| 232 | // take the full range we could get |
| 233 | cursor = end; |
| 234 | break; |
| 235 | |
| 236 | // array of * |
| 237 | case 'f': |
| 238 | case 'd': |
| 239 | case 'l': |
| 240 | case 'i': |
| 241 | case 'c': { |
| 242 | const uint32_t length = ReadWord(input, cursor, end); |
| 243 | const uint32_t encoding = ReadWord(input, cursor, end); |
| 244 | |
| 245 | const uint32_t comp_len = ReadWord(input, cursor, end); |
| 246 | |
| 247 | // compute length based on type and check against the stored value |
| 248 | if(encoding == 0) { |
| 249 | uint32_t stride = 0; |
| 250 | switch(type) |
| 251 | { |
| 252 | case 'f': |
| 253 | case 'i': |
| 254 | stride = 4; |
| 255 | break; |
| 256 | |
| 257 | case 'd': |
| 258 | case 'l': |
| 259 | stride = 8; |
| 260 | break; |
| 261 | |
| 262 | case 'c': |
| 263 | stride = 1; |
| 264 | break; |
| 265 | |
| 266 | default: |
| 267 | ai_assert(false); |
| 268 | }; |
| 269 | ai_assert(stride > 0); |
| 270 | if(length * stride != comp_len) { |
| 271 | TokenizeError(message: "cannot ReadData, calculated data stride differs from what the file claims" ,begin: input, cursor); |
| 272 | } |
| 273 | } |
| 274 | // zip/deflate algorithm (encoding==1)? take given length. anything else? die |
| 275 | else if (encoding != 1) { |
| 276 | TokenizeError(message: "cannot ReadData, unknown encoding" ,begin: input, cursor); |
| 277 | } |
| 278 | cursor += comp_len; |
| 279 | break; |
| 280 | } |
| 281 | |
| 282 | // string |
| 283 | case 'S': { |
| 284 | const char* sb, *se; |
| 285 | // 0 characters can legally happen in such strings |
| 286 | ReadString(sbegin_out&: sb, send_out&: se, input, cursor, end, long_length: true, allow_null: true); |
| 287 | break; |
| 288 | } |
| 289 | default: |
| 290 | TokenizeError(message: "cannot ReadData, unexpected type code: " + std::string(&type, 1),begin: input, cursor); |
| 291 | } |
| 292 | |
| 293 | if(cursor > end) { |
| 294 | TokenizeError(message: "cannot ReadData, the remaining size is too small for the data type: " + std::string(&type, 1),begin: input, cursor); |
| 295 | } |
| 296 | |
| 297 | // the type code is contained in the returned range |
| 298 | send_out = cursor; |
| 299 | } |
| 300 | |
| 301 | |
| 302 | // ------------------------------------------------------------------------------------------------ |
| 303 | bool ReadScope(TokenList &output_tokens, StackAllocator &token_allocator, const char *input, const char *&cursor, const char *end, bool const is64bits) { |
| 304 | // the first word contains the offset at which this block ends |
| 305 | const uint64_t end_offset = is64bits ? ReadDoubleWord(input, cursor, end) : ReadWord(input, cursor, end); |
| 306 | |
| 307 | // we may get 0 if reading reached the end of the file - |
| 308 | // fbx files have a mysterious extra footer which I don't know |
| 309 | // how to extract any information from, but at least it always |
| 310 | // starts with a 0. |
| 311 | if(!end_offset) { |
| 312 | return false; |
| 313 | } |
| 314 | |
| 315 | if(end_offset > Offset(begin: input, cursor: end)) { |
| 316 | TokenizeError(message: "block offset is out of range" ,begin: input, cursor); |
| 317 | } |
| 318 | else if(end_offset < Offset(begin: input, cursor)) { |
| 319 | TokenizeError(message: "block offset is negative out of range" ,begin: input, cursor); |
| 320 | } |
| 321 | |
| 322 | // the second data word contains the number of properties in the scope |
| 323 | const uint64_t prop_count = is64bits ? ReadDoubleWord(input, cursor, end) : ReadWord(input, cursor, end); |
| 324 | |
| 325 | // the third data word contains the length of the property list |
| 326 | const uint64_t prop_length = is64bits ? ReadDoubleWord(input, cursor, end) : ReadWord(input, cursor, end); |
| 327 | |
| 328 | // now comes the name of the scope/key |
| 329 | const char* sbeg, *send; |
| 330 | ReadString(sbegin_out&: sbeg, send_out&: send, input, cursor, end); |
| 331 | |
| 332 | output_tokens.push_back(new_Token(sbeg, send, TokenType_KEY, Offset(begin: input, cursor) )); |
| 333 | |
| 334 | // now come the individual properties |
| 335 | const char* begin_cursor = cursor; |
| 336 | |
| 337 | if ((begin_cursor + prop_length) > end) { |
| 338 | TokenizeError(message: "property length out of bounds reading length " , begin: input, cursor); |
| 339 | } |
| 340 | |
| 341 | for (unsigned int i = 0; i < prop_count; ++i) { |
| 342 | ReadData(sbegin_out&: sbeg, send_out&: send, input, cursor, end: begin_cursor + prop_length); |
| 343 | |
| 344 | output_tokens.push_back(new_Token(sbeg, send, TokenType_DATA, Offset(begin: input, cursor) )); |
| 345 | |
| 346 | if(i != prop_count-1) { |
| 347 | output_tokens.push_back(new_Token(cursor, cursor + 1, TokenType_COMMA, Offset(begin: input, cursor) )); |
| 348 | } |
| 349 | } |
| 350 | |
| 351 | if (Offset(begin: begin_cursor, cursor) != prop_length) { |
| 352 | TokenizeError(message: "property length not reached, something is wrong" ,begin: input, cursor); |
| 353 | } |
| 354 | |
| 355 | // at the end of each nested block, there is a NUL record to indicate |
| 356 | // that the sub-scope exists (i.e. to distinguish between P: and P : {}) |
| 357 | // this NUL record is 13 bytes long on 32 bit version and 25 bytes long on 64 bit. |
| 358 | const size_t sentinel_block_length = is64bits ? (sizeof(uint64_t)* 3 + 1) : (sizeof(uint32_t)* 3 + 1); |
| 359 | |
| 360 | if (Offset(begin: input, cursor) < end_offset) { |
| 361 | if (end_offset - Offset(begin: input, cursor) < sentinel_block_length) { |
| 362 | TokenizeError(message: "insufficient padding bytes at block end" ,begin: input, cursor); |
| 363 | } |
| 364 | |
| 365 | output_tokens.push_back(new_Token(cursor, cursor + 1, TokenType_OPEN_BRACKET, Offset(begin: input, cursor) )); |
| 366 | |
| 367 | // XXX this is vulnerable to stack overflowing .. |
| 368 | while(Offset(begin: input, cursor) < end_offset - sentinel_block_length) { |
| 369 | ReadScope(output_tokens, token_allocator, input, cursor, end: input + end_offset - sentinel_block_length, is64bits); |
| 370 | } |
| 371 | output_tokens.push_back(new_Token(cursor, cursor + 1, TokenType_CLOSE_BRACKET, Offset(begin: input, cursor) )); |
| 372 | |
| 373 | for (unsigned int i = 0; i < sentinel_block_length; ++i) { |
| 374 | if(cursor[i] != '\0') { |
| 375 | TokenizeError(message: "failed to read nested block sentinel, expected all bytes to be 0" ,begin: input, cursor); |
| 376 | } |
| 377 | } |
| 378 | cursor += sentinel_block_length; |
| 379 | } |
| 380 | |
| 381 | if (Offset(begin: input, cursor) != end_offset) { |
| 382 | TokenizeError(message: "scope length not reached, something is wrong" ,begin: input, cursor); |
| 383 | } |
| 384 | |
| 385 | return true; |
| 386 | } |
| 387 | |
| 388 | } // anonymous namespace |
| 389 | |
| 390 | // ------------------------------------------------------------------------------------------------ |
| 391 | // TODO: Test FBX Binary files newer than the 7500 version to check if the 64 bits address behaviour is consistent |
| 392 | void TokenizeBinary(TokenList &output_tokens, const char *input, size_t length, StackAllocator &token_allocator) { |
| 393 | ai_assert(input); |
| 394 | ASSIMP_LOG_DEBUG("Tokenizing binary FBX file" ); |
| 395 | |
| 396 | if(length < 0x1b) { |
| 397 | TokenizeError(message: "file is too short" ,offset: 0); |
| 398 | } |
| 399 | |
| 400 | //uint32_t offset = 0x15; |
| 401 | /* const char* cursor = input + 0x15; |
| 402 | |
| 403 | const uint32_t flags = ReadWord(input, cursor, input + length); |
| 404 | |
| 405 | const uint8_t padding_0 = ReadByte(input, cursor, input + length); // unused |
| 406 | const uint8_t padding_1 = ReadByte(input, cursor, input + length); // unused*/ |
| 407 | |
| 408 | if (strncmp(s1: input,s2: "Kaydara FBX Binary" ,n: 18)) { |
| 409 | TokenizeError(message: "magic bytes not found" ,offset: 0); |
| 410 | } |
| 411 | |
| 412 | const char* cursor = input + 18; |
| 413 | /*Result ignored*/ ReadByte(input, cursor, end: input + length); |
| 414 | /*Result ignored*/ ReadByte(input, cursor, end: input + length); |
| 415 | /*Result ignored*/ ReadByte(input, cursor, end: input + length); |
| 416 | /*Result ignored*/ ReadByte(input, cursor, end: input + length); |
| 417 | /*Result ignored*/ ReadByte(input, cursor, end: input + length); |
| 418 | const uint32_t version = ReadWord(input, cursor, end: input + length); |
| 419 | ASSIMP_LOG_DEBUG("FBX version: " , version); |
| 420 | const bool is64bits = version >= 7500; |
| 421 | const char *end = input + length; |
| 422 | try |
| 423 | { |
| 424 | while (cursor < end ) { |
| 425 | if (!ReadScope(output_tokens, token_allocator, input, cursor, end: input + length, is64bits)) { |
| 426 | break; |
| 427 | } |
| 428 | } |
| 429 | } |
| 430 | catch (const DeadlyImportError& e) |
| 431 | { |
| 432 | if (!is64bits && (length > std::numeric_limits<uint32_t>::max())) { |
| 433 | throw DeadlyImportError("The FBX file is invalid. This may be because the content is too big for this older version (" , ai_to_string(value: version), ") of the FBX format. (" , e.what(), ")" ); |
| 434 | } |
| 435 | throw; |
| 436 | } |
| 437 | } |
| 438 | |
| 439 | } // !FBX |
| 440 | } // !Assimp |
| 441 | |
| 442 | #endif |
| 443 | |