| 1 | // © 2016 and later: Unicode, Inc. and others. | 
|---|
| 2 | // License & terms of use: http://www.unicode.org/copyright.html | 
|---|
| 3 | /* | 
|---|
| 4 | ******************************************************************************* | 
|---|
| 5 | * | 
|---|
| 6 | *   Copyright (C) 2003-2015, International Business Machines | 
|---|
| 7 | *   Corporation and others.  All Rights Reserved. | 
|---|
| 8 | * | 
|---|
| 9 | ******************************************************************************* | 
|---|
| 10 | *   file name:  ucol_swp.cpp | 
|---|
| 11 | *   encoding:   UTF-8 | 
|---|
| 12 | *   tab size:   8 (not used) | 
|---|
| 13 | *   indentation:4 | 
|---|
| 14 | * | 
|---|
| 15 | *   created on: 2003sep10 | 
|---|
| 16 | *   created by: Markus W. Scherer | 
|---|
| 17 | * | 
|---|
| 18 | *   Swap collation binaries. | 
|---|
| 19 | */ | 
|---|
| 20 |  | 
|---|
| 21 | #include "unicode/udata.h" /* UDataInfo */ | 
|---|
| 22 | #include "utrie.h" | 
|---|
| 23 | #include "utrie2.h" | 
|---|
| 24 | #include "udataswp.h" | 
|---|
| 25 | #include "cmemory.h" | 
|---|
| 26 | #include "ucol_data.h" | 
|---|
| 27 | #include "ucol_swp.h" | 
|---|
| 28 |  | 
|---|
| 29 | /* swapping ----------------------------------------------------------------- */ | 
|---|
| 30 |  | 
|---|
| 31 | #if !UCONFIG_NO_COLLATION | 
|---|
| 32 |  | 
|---|
| 33 | U_CAPI UBool U_EXPORT2 | 
|---|
| 34 | ucol_looksLikeCollationBinary(const UDataSwapper *ds, | 
|---|
| 35 | const void *inData, int32_t length) { | 
|---|
| 36 | if(ds==nullptr || inData==nullptr || length<-1) { | 
|---|
| 37 | return false; | 
|---|
| 38 | } | 
|---|
| 39 |  | 
|---|
| 40 | // First check for format version 4+ which has a standard data header. | 
|---|
| 41 | UErrorCode errorCode=U_ZERO_ERROR; | 
|---|
| 42 | (void)udata_swapDataHeader(ds, inData, -1, nullptr, &errorCode); | 
|---|
| 43 | if(U_SUCCESS(errorCode)) { | 
|---|
| 44 | const UDataInfo &info=*(const UDataInfo *)((const char *)inData+4); | 
|---|
| 45 | if(info.dataFormat[0]==0x55 &&   // dataFormat="UCol" | 
|---|
| 46 | info.dataFormat[1]==0x43 && | 
|---|
| 47 | info.dataFormat[2]==0x6f && | 
|---|
| 48 | info.dataFormat[3]==0x6c) { | 
|---|
| 49 | return true; | 
|---|
| 50 | } | 
|---|
| 51 | } | 
|---|
| 52 |  | 
|---|
| 53 | // Else check for format version 3. | 
|---|
| 54 | const UCATableHeader *inHeader=(const UCATableHeader *)inData; | 
|---|
| 55 |  | 
|---|
| 56 | /* | 
|---|
| 57 | * The collation binary must contain at least the UCATableHeader, | 
|---|
| 58 | * starting with its size field. | 
|---|
| 59 | * sizeof(UCATableHeader)==42*4 in ICU 2.8 | 
|---|
| 60 | * check the length against the header size before reading the size field | 
|---|
| 61 | */ | 
|---|
| 62 | UCATableHeader header; | 
|---|
| 63 | uprv_memset(&header, 0, sizeof(header)); | 
|---|
| 64 | if(length<0) { | 
|---|
| 65 | header.size=udata_readInt32(ds, inHeader->size); | 
|---|
| 66 | } else if((length<(42*4) || length<(header.size=udata_readInt32(ds, inHeader->size)))) { | 
|---|
| 67 | return false; | 
|---|
| 68 | } | 
|---|
| 69 |  | 
|---|
| 70 | header.magic=ds->readUInt32(inHeader->magic); | 
|---|
| 71 | if(!( | 
|---|
| 72 | header.magic==UCOL_HEADER_MAGIC && | 
|---|
| 73 | inHeader->formatVersion[0]==3 /*&& | 
|---|
| 74 | inHeader->formatVersion[1]>=0*/ | 
|---|
| 75 | )) { | 
|---|
| 76 | return false; | 
|---|
| 77 | } | 
|---|
| 78 |  | 
|---|
| 79 | if(inHeader->isBigEndian!=ds->inIsBigEndian || inHeader->charSetFamily!=ds->inCharset) { | 
|---|
| 80 | return false; | 
|---|
| 81 | } | 
|---|
| 82 |  | 
|---|
| 83 | return true; | 
|---|
| 84 | } | 
|---|
| 85 |  | 
|---|
| 86 | namespace { | 
|---|
| 87 |  | 
|---|
| 88 | /* swap a header-less collation formatVersion=3 binary, inside a resource bundle or ucadata.icu */ | 
|---|
| 89 | int32_t | 
|---|
| 90 | swapFormatVersion3(const UDataSwapper *ds, | 
|---|
| 91 | const void *inData, int32_t length, void *outData, | 
|---|
| 92 | UErrorCode *pErrorCode) { | 
|---|
| 93 | const uint8_t *inBytes; | 
|---|
| 94 | uint8_t *outBytes; | 
|---|
| 95 |  | 
|---|
| 96 | const UCATableHeader *inHeader; | 
|---|
| 97 | UCATableHeader *outHeader; | 
|---|
| 98 | UCATableHeader header; | 
|---|
| 99 |  | 
|---|
| 100 | uint32_t count; | 
|---|
| 101 |  | 
|---|
| 102 | /* argument checking in case we were not called from ucol_swap() */ | 
|---|
| 103 | if(U_FAILURE(*pErrorCode)) { | 
|---|
| 104 | return 0; | 
|---|
| 105 | } | 
|---|
| 106 | if(ds==nullptr || inData==nullptr || length<-1 || (length>0 && outData==nullptr)) { | 
|---|
| 107 | *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR; | 
|---|
| 108 | return 0; | 
|---|
| 109 | } | 
|---|
| 110 |  | 
|---|
| 111 | inBytes=(const uint8_t *)inData; | 
|---|
| 112 | outBytes=(uint8_t *)outData; | 
|---|
| 113 |  | 
|---|
| 114 | inHeader=(const UCATableHeader *)inData; | 
|---|
| 115 | outHeader=(UCATableHeader *)outData; | 
|---|
| 116 |  | 
|---|
| 117 | /* | 
|---|
| 118 | * The collation binary must contain at least the UCATableHeader, | 
|---|
| 119 | * starting with its size field. | 
|---|
| 120 | * sizeof(UCATableHeader)==42*4 in ICU 2.8 | 
|---|
| 121 | * check the length against the header size before reading the size field | 
|---|
| 122 | */ | 
|---|
| 123 | uprv_memset(&header, 0, sizeof(header)); | 
|---|
| 124 | if(length<0) { | 
|---|
| 125 | header.size=udata_readInt32(ds, inHeader->size); | 
|---|
| 126 | } else if((length<(42*4) || length<(header.size=udata_readInt32(ds, inHeader->size)))) { | 
|---|
| 127 | udata_printError(ds, "ucol_swap(formatVersion=3): too few bytes (%d after header) for collation data\n", | 
|---|
| 128 | length); | 
|---|
| 129 | *pErrorCode=U_INDEX_OUTOFBOUNDS_ERROR; | 
|---|
| 130 | return 0; | 
|---|
| 131 | } | 
|---|
| 132 |  | 
|---|
| 133 | header.magic=ds->readUInt32(inHeader->magic); | 
|---|
| 134 | if(!( | 
|---|
| 135 | header.magic==UCOL_HEADER_MAGIC && | 
|---|
| 136 | inHeader->formatVersion[0]==3 /*&& | 
|---|
| 137 | inHeader->formatVersion[1]>=0*/ | 
|---|
| 138 | )) { | 
|---|
| 139 | udata_printError(ds, "ucol_swap(formatVersion=3): magic 0x%08x or format version %02x.%02x is not a collation binary\n", | 
|---|
| 140 | header.magic, | 
|---|
| 141 | inHeader->formatVersion[0], inHeader->formatVersion[1]); | 
|---|
| 142 | *pErrorCode=U_UNSUPPORTED_ERROR; | 
|---|
| 143 | return 0; | 
|---|
| 144 | } | 
|---|
| 145 |  | 
|---|
| 146 | if(inHeader->isBigEndian!=ds->inIsBigEndian || inHeader->charSetFamily!=ds->inCharset) { | 
|---|
| 147 | udata_printError(ds, "ucol_swap(formatVersion=3): endianness %d or charset %d does not match the swapper\n", | 
|---|
| 148 | inHeader->isBigEndian, inHeader->charSetFamily); | 
|---|
| 149 | *pErrorCode=U_INVALID_FORMAT_ERROR; | 
|---|
| 150 | return 0; | 
|---|
| 151 | } | 
|---|
| 152 |  | 
|---|
| 153 | if(length>=0) { | 
|---|
| 154 | /* copy everything, takes care of data that needs no swapping */ | 
|---|
| 155 | if(inBytes!=outBytes) { | 
|---|
| 156 | uprv_memcpy(outBytes, inBytes, header.size); | 
|---|
| 157 | } | 
|---|
| 158 |  | 
|---|
| 159 | /* swap the necessary pieces in the order of their occurrence in the data */ | 
|---|
| 160 |  | 
|---|
| 161 | /* read more of the UCATableHeader (the size field was read above) */ | 
|---|
| 162 | header.options=                 ds->readUInt32(inHeader->options); | 
|---|
| 163 | header.UCAConsts=               ds->readUInt32(inHeader->UCAConsts); | 
|---|
| 164 | header.contractionUCACombos=    ds->readUInt32(inHeader->contractionUCACombos); | 
|---|
| 165 | header.mappingPosition=         ds->readUInt32(inHeader->mappingPosition); | 
|---|
| 166 | header.expansion=               ds->readUInt32(inHeader->expansion); | 
|---|
| 167 | header.contractionIndex=        ds->readUInt32(inHeader->contractionIndex); | 
|---|
| 168 | header.contractionCEs=          ds->readUInt32(inHeader->contractionCEs); | 
|---|
| 169 | header.contractionSize=         ds->readUInt32(inHeader->contractionSize); | 
|---|
| 170 | header.endExpansionCE=          ds->readUInt32(inHeader->endExpansionCE); | 
|---|
| 171 | header.expansionCESize=         ds->readUInt32(inHeader->expansionCESize); | 
|---|
| 172 | header.endExpansionCECount=     udata_readInt32(ds, inHeader->endExpansionCECount); | 
|---|
| 173 | header.contractionUCACombosSize=udata_readInt32(ds, inHeader->contractionUCACombosSize); | 
|---|
| 174 | header.scriptToLeadByte=        ds->readUInt32(inHeader->scriptToLeadByte); | 
|---|
| 175 | header.leadByteToScript=        ds->readUInt32(inHeader->leadByteToScript); | 
|---|
| 176 |  | 
|---|
| 177 | /* swap the 32-bit integers in the header */ | 
|---|
| 178 | ds->swapArray32(ds, inHeader, (int32_t)((const char *)&inHeader->jamoSpecial-(const char *)inHeader), | 
|---|
| 179 | outHeader, pErrorCode); | 
|---|
| 180 | ds->swapArray32(ds, &(inHeader->scriptToLeadByte), sizeof(header.scriptToLeadByte) + sizeof(header.leadByteToScript), | 
|---|
| 181 | &(outHeader->scriptToLeadByte), pErrorCode); | 
|---|
| 182 | /* set the output platform properties */ | 
|---|
| 183 | outHeader->isBigEndian=ds->outIsBigEndian; | 
|---|
| 184 | outHeader->charSetFamily=ds->outCharset; | 
|---|
| 185 |  | 
|---|
| 186 | /* swap the options */ | 
|---|
| 187 | if(header.options!=0) { | 
|---|
| 188 | ds->swapArray32(ds, inBytes+header.options, header.expansion-header.options, | 
|---|
| 189 | outBytes+header.options, pErrorCode); | 
|---|
| 190 | } | 
|---|
| 191 |  | 
|---|
| 192 | /* swap the expansions */ | 
|---|
| 193 | if(header.mappingPosition!=0 && header.expansion!=0) { | 
|---|
| 194 | if(header.contractionIndex!=0) { | 
|---|
| 195 | /* expansions bounded by contractions */ | 
|---|
| 196 | count=header.contractionIndex-header.expansion; | 
|---|
| 197 | } else { | 
|---|
| 198 | /* no contractions: expansions bounded by the main trie */ | 
|---|
| 199 | count=header.mappingPosition-header.expansion; | 
|---|
| 200 | } | 
|---|
| 201 | ds->swapArray32(ds, inBytes+header.expansion, (int32_t)count, | 
|---|
| 202 | outBytes+header.expansion, pErrorCode); | 
|---|
| 203 | } | 
|---|
| 204 |  | 
|---|
| 205 | /* swap the contractions */ | 
|---|
| 206 | if(header.contractionSize!=0) { | 
|---|
| 207 | /* contractionIndex: char16_t[] */ | 
|---|
| 208 | ds->swapArray16(ds, inBytes+header.contractionIndex, header.contractionSize*2, | 
|---|
| 209 | outBytes+header.contractionIndex, pErrorCode); | 
|---|
| 210 |  | 
|---|
| 211 | /* contractionCEs: CEs[] */ | 
|---|
| 212 | ds->swapArray32(ds, inBytes+header.contractionCEs, header.contractionSize*4, | 
|---|
| 213 | outBytes+header.contractionCEs, pErrorCode); | 
|---|
| 214 | } | 
|---|
| 215 |  | 
|---|
| 216 | /* swap the main trie */ | 
|---|
| 217 | if(header.mappingPosition!=0) { | 
|---|
| 218 | count=header.endExpansionCE-header.mappingPosition; | 
|---|
| 219 | utrie_swap(ds, inBytes+header.mappingPosition, (int32_t)count, | 
|---|
| 220 | outBytes+header.mappingPosition, pErrorCode); | 
|---|
| 221 | } | 
|---|
| 222 |  | 
|---|
| 223 | /* swap the max expansion table */ | 
|---|
| 224 | if(header.endExpansionCECount!=0) { | 
|---|
| 225 | ds->swapArray32(ds, inBytes+header.endExpansionCE, header.endExpansionCECount*4, | 
|---|
| 226 | outBytes+header.endExpansionCE, pErrorCode); | 
|---|
| 227 | } | 
|---|
| 228 |  | 
|---|
| 229 | /* expansionCESize, unsafeCP, contrEndCP: uint8_t[], no need to swap */ | 
|---|
| 230 |  | 
|---|
| 231 | /* swap UCA constants */ | 
|---|
| 232 | if(header.UCAConsts!=0) { | 
|---|
| 233 | /* | 
|---|
| 234 | * if UCAConsts!=0 then contractionUCACombos because we are swapping | 
|---|
| 235 | * the UCA data file, and we know that the UCA contains contractions | 
|---|
| 236 | */ | 
|---|
| 237 | ds->swapArray32(ds, inBytes+header.UCAConsts, header.contractionUCACombos-header.UCAConsts, | 
|---|
| 238 | outBytes+header.UCAConsts, pErrorCode); | 
|---|
| 239 | } | 
|---|
| 240 |  | 
|---|
| 241 | /* swap UCA contractions */ | 
|---|
| 242 | if(header.contractionUCACombosSize!=0) { | 
|---|
| 243 | count=header.contractionUCACombosSize*inHeader->contractionUCACombosWidth*U_SIZEOF_UCHAR; | 
|---|
| 244 | ds->swapArray16(ds, inBytes+header.contractionUCACombos, (int32_t)count, | 
|---|
| 245 | outBytes+header.contractionUCACombos, pErrorCode); | 
|---|
| 246 | } | 
|---|
| 247 |  | 
|---|
| 248 | /* swap the script to lead bytes */ | 
|---|
| 249 | if(header.scriptToLeadByte!=0) { | 
|---|
| 250 | int indexCount = ds->readUInt16(*((uint16_t*)(inBytes+header.scriptToLeadByte))); // each entry = 2 * uint16 | 
|---|
| 251 | int dataCount = ds->readUInt16(*((uint16_t*)(inBytes+header.scriptToLeadByte + 2))); // each entry = uint16 | 
|---|
| 252 | ds->swapArray16(ds, inBytes+header.scriptToLeadByte, | 
|---|
| 253 | 4 + (4 * indexCount) + (2 * dataCount), | 
|---|
| 254 | outBytes+header.scriptToLeadByte, pErrorCode); | 
|---|
| 255 | } | 
|---|
| 256 |  | 
|---|
| 257 | /* swap the lead byte to scripts */ | 
|---|
| 258 | if(header.leadByteToScript!=0) { | 
|---|
| 259 | int indexCount = ds->readUInt16(*((uint16_t*)(inBytes+header.leadByteToScript))); // each entry = uint16 | 
|---|
| 260 | int dataCount = ds->readUInt16(*((uint16_t*)(inBytes+header.leadByteToScript + 2))); // each entry = uint16 | 
|---|
| 261 | ds->swapArray16(ds, inBytes+header.leadByteToScript, | 
|---|
| 262 | 4 + (2 * indexCount) + (2 * dataCount), | 
|---|
| 263 | outBytes+header.leadByteToScript, pErrorCode); | 
|---|
| 264 | } | 
|---|
| 265 | } | 
|---|
| 266 |  | 
|---|
| 267 | return header.size; | 
|---|
| 268 | } | 
|---|
| 269 |  | 
|---|
| 270 | // swap formatVersion 4 or 5 ----------------------------------------------- *** | 
|---|
| 271 |  | 
|---|
| 272 | // The following are copied from CollationDataReader, trading an awkward copy of constants | 
|---|
| 273 | // for an awkward relocation of the i18n collationdatareader.h file into the common library. | 
|---|
| 274 | // Keep them in sync! | 
|---|
| 275 |  | 
|---|
| 276 | enum { | 
|---|
| 277 | IX_INDEXES_LENGTH,  // 0 | 
|---|
| 278 | IX_OPTIONS, | 
|---|
| 279 | IX_RESERVED2, | 
|---|
| 280 | IX_RESERVED3, | 
|---|
| 281 |  | 
|---|
| 282 | IX_JAMO_CE32S_START,  // 4 | 
|---|
| 283 | IX_REORDER_CODES_OFFSET, | 
|---|
| 284 | IX_REORDER_TABLE_OFFSET, | 
|---|
| 285 | IX_TRIE_OFFSET, | 
|---|
| 286 |  | 
|---|
| 287 | IX_RESERVED8_OFFSET,  // 8 | 
|---|
| 288 | IX_CES_OFFSET, | 
|---|
| 289 | IX_RESERVED10_OFFSET, | 
|---|
| 290 | IX_CE32S_OFFSET, | 
|---|
| 291 |  | 
|---|
| 292 | IX_ROOT_ELEMENTS_OFFSET,  // 12 | 
|---|
| 293 | IX_CONTEXTS_OFFSET, | 
|---|
| 294 | IX_UNSAFE_BWD_OFFSET, | 
|---|
| 295 | IX_FAST_LATIN_TABLE_OFFSET, | 
|---|
| 296 |  | 
|---|
| 297 | IX_SCRIPTS_OFFSET,  // 16 | 
|---|
| 298 | IX_COMPRESSIBLE_BYTES_OFFSET, | 
|---|
| 299 | IX_RESERVED18_OFFSET, | 
|---|
| 300 | IX_TOTAL_SIZE | 
|---|
| 301 | }; | 
|---|
| 302 |  | 
|---|
| 303 | int32_t | 
|---|
| 304 | swapFormatVersion4(const UDataSwapper *ds, | 
|---|
| 305 | const void *inData, int32_t length, void *outData, | 
|---|
| 306 | UErrorCode &errorCode) { | 
|---|
| 307 | if(U_FAILURE(errorCode)) { return 0; } | 
|---|
| 308 |  | 
|---|
| 309 | const uint8_t *inBytes=(const uint8_t *)inData; | 
|---|
| 310 | uint8_t *outBytes=(uint8_t *)outData; | 
|---|
| 311 |  | 
|---|
| 312 | const int32_t *inIndexes=(const int32_t *)inBytes; | 
|---|
| 313 | int32_t indexes[IX_TOTAL_SIZE+1]; | 
|---|
| 314 |  | 
|---|
| 315 | // Need at least IX_INDEXES_LENGTH and IX_OPTIONS. | 
|---|
| 316 | if(0<=length && length<8) { | 
|---|
| 317 | udata_printError(ds, "ucol_swap(formatVersion=4): too few bytes " | 
|---|
| 318 | "(%d after header) for collation data\n", | 
|---|
| 319 | length); | 
|---|
| 320 | errorCode=U_INDEX_OUTOFBOUNDS_ERROR; | 
|---|
| 321 | return 0; | 
|---|
| 322 | } | 
|---|
| 323 |  | 
|---|
| 324 | int32_t indexesLength=indexes[0]=udata_readInt32(ds, inIndexes[0]); | 
|---|
| 325 | if(0<=length && length<(indexesLength*4)) { | 
|---|
| 326 | udata_printError(ds, "ucol_swap(formatVersion=4): too few bytes " | 
|---|
| 327 | "(%d after header) for collation data\n", | 
|---|
| 328 | length); | 
|---|
| 329 | errorCode=U_INDEX_OUTOFBOUNDS_ERROR; | 
|---|
| 330 | return 0; | 
|---|
| 331 | } | 
|---|
| 332 |  | 
|---|
| 333 | for(int32_t i=1; i<=IX_TOTAL_SIZE && i<indexesLength; ++i) { | 
|---|
| 334 | indexes[i]=udata_readInt32(ds, inIndexes[i]); | 
|---|
| 335 | } | 
|---|
| 336 | for(int32_t i=indexesLength; i<=IX_TOTAL_SIZE; ++i) { | 
|---|
| 337 | indexes[i]=-1; | 
|---|
| 338 | } | 
|---|
| 339 | inIndexes=nullptr;  // Make sure we do not accidentally use these instead of indexes[]. | 
|---|
| 340 |  | 
|---|
| 341 | // Get the total length of the data. | 
|---|
| 342 | int32_t size; | 
|---|
| 343 | if(indexesLength>IX_TOTAL_SIZE) { | 
|---|
| 344 | size=indexes[IX_TOTAL_SIZE]; | 
|---|
| 345 | } else if(indexesLength>IX_REORDER_CODES_OFFSET) { | 
|---|
| 346 | size=indexes[indexesLength-1]; | 
|---|
| 347 | } else { | 
|---|
| 348 | size=indexesLength*4; | 
|---|
| 349 | } | 
|---|
| 350 | if(length<0) { return size; } | 
|---|
| 351 |  | 
|---|
| 352 | if(length<size) { | 
|---|
| 353 | udata_printError(ds, "ucol_swap(formatVersion=4): too few bytes " | 
|---|
| 354 | "(%d after header) for collation data\n", | 
|---|
| 355 | length); | 
|---|
| 356 | errorCode=U_INDEX_OUTOFBOUNDS_ERROR; | 
|---|
| 357 | return 0; | 
|---|
| 358 | } | 
|---|
| 359 |  | 
|---|
| 360 | // Copy the data for inaccessible bytes and arrays of bytes. | 
|---|
| 361 | if(inBytes!=outBytes) { | 
|---|
| 362 | uprv_memcpy(outBytes, inBytes, size); | 
|---|
| 363 | } | 
|---|
| 364 |  | 
|---|
| 365 | // Swap the int32_t indexes[]. | 
|---|
| 366 | ds->swapArray32(ds, inBytes, indexesLength * 4, outBytes, &errorCode); | 
|---|
| 367 |  | 
|---|
| 368 | // The following is a modified version of CollationDataReader::read(). | 
|---|
| 369 | // Here we use indexes[] not inIndexes[] because | 
|---|
| 370 | // the inIndexes[] may not be in this machine's endianness. | 
|---|
| 371 | int32_t index;  // one of the indexes[] slots | 
|---|
| 372 | int32_t offset;  // byte offset for the index part | 
|---|
| 373 | // int32_t length;  // number of bytes in the index part | 
|---|
| 374 |  | 
|---|
| 375 | index = IX_REORDER_CODES_OFFSET; | 
|---|
| 376 | offset = indexes[index]; | 
|---|
| 377 | length = indexes[index + 1] - offset; | 
|---|
| 378 | if(length > 0) { | 
|---|
| 379 | ds->swapArray32(ds, inBytes + offset, length, outBytes + offset, &errorCode); | 
|---|
| 380 | } | 
|---|
| 381 |  | 
|---|
| 382 | // Skip the IX_REORDER_TABLE_OFFSET byte array. | 
|---|
| 383 |  | 
|---|
| 384 | index = IX_TRIE_OFFSET; | 
|---|
| 385 | offset = indexes[index]; | 
|---|
| 386 | length = indexes[index + 1] - offset; | 
|---|
| 387 | if(length > 0) { | 
|---|
| 388 | utrie2_swap(ds, inBytes + offset, length, outBytes + offset, &errorCode); | 
|---|
| 389 | } | 
|---|
| 390 |  | 
|---|
| 391 | index = IX_RESERVED8_OFFSET; | 
|---|
| 392 | offset = indexes[index]; | 
|---|
| 393 | length = indexes[index + 1] - offset; | 
|---|
| 394 | if(length > 0) { | 
|---|
| 395 | udata_printError(ds, "ucol_swap(formatVersion=4): unknown data at IX_RESERVED8_OFFSET\n", length); | 
|---|
| 396 | errorCode = U_UNSUPPORTED_ERROR; | 
|---|
| 397 | return 0; | 
|---|
| 398 | } | 
|---|
| 399 |  | 
|---|
| 400 | index = IX_CES_OFFSET; | 
|---|
| 401 | offset = indexes[index]; | 
|---|
| 402 | length = indexes[index + 1] - offset; | 
|---|
| 403 | if(length > 0) { | 
|---|
| 404 | ds->swapArray64(ds, inBytes + offset, length, outBytes + offset, &errorCode); | 
|---|
| 405 | } | 
|---|
| 406 |  | 
|---|
| 407 | index = IX_RESERVED10_OFFSET; | 
|---|
| 408 | offset = indexes[index]; | 
|---|
| 409 | length = indexes[index + 1] - offset; | 
|---|
| 410 | if(length > 0) { | 
|---|
| 411 | udata_printError(ds, "ucol_swap(formatVersion=4): unknown data at IX_RESERVED10_OFFSET\n", length); | 
|---|
| 412 | errorCode = U_UNSUPPORTED_ERROR; | 
|---|
| 413 | return 0; | 
|---|
| 414 | } | 
|---|
| 415 |  | 
|---|
| 416 | index = IX_CE32S_OFFSET; | 
|---|
| 417 | offset = indexes[index]; | 
|---|
| 418 | length = indexes[index + 1] - offset; | 
|---|
| 419 | if(length > 0) { | 
|---|
| 420 | ds->swapArray32(ds, inBytes + offset, length, outBytes + offset, &errorCode); | 
|---|
| 421 | } | 
|---|
| 422 |  | 
|---|
| 423 | index = IX_ROOT_ELEMENTS_OFFSET; | 
|---|
| 424 | offset = indexes[index]; | 
|---|
| 425 | length = indexes[index + 1] - offset; | 
|---|
| 426 | if(length > 0) { | 
|---|
| 427 | ds->swapArray32(ds, inBytes + offset, length, outBytes + offset, &errorCode); | 
|---|
| 428 | } | 
|---|
| 429 |  | 
|---|
| 430 | index = IX_CONTEXTS_OFFSET; | 
|---|
| 431 | offset = indexes[index]; | 
|---|
| 432 | length = indexes[index + 1] - offset; | 
|---|
| 433 | if(length > 0) { | 
|---|
| 434 | ds->swapArray16(ds, inBytes + offset, length, outBytes + offset, &errorCode); | 
|---|
| 435 | } | 
|---|
| 436 |  | 
|---|
| 437 | index = IX_UNSAFE_BWD_OFFSET; | 
|---|
| 438 | offset = indexes[index]; | 
|---|
| 439 | length = indexes[index + 1] - offset; | 
|---|
| 440 | if(length > 0) { | 
|---|
| 441 | ds->swapArray16(ds, inBytes + offset, length, outBytes + offset, &errorCode); | 
|---|
| 442 | } | 
|---|
| 443 |  | 
|---|
| 444 | index = IX_FAST_LATIN_TABLE_OFFSET; | 
|---|
| 445 | offset = indexes[index]; | 
|---|
| 446 | length = indexes[index + 1] - offset; | 
|---|
| 447 | if(length > 0) { | 
|---|
| 448 | ds->swapArray16(ds, inBytes + offset, length, outBytes + offset, &errorCode); | 
|---|
| 449 | } | 
|---|
| 450 |  | 
|---|
| 451 | index = IX_SCRIPTS_OFFSET; | 
|---|
| 452 | offset = indexes[index]; | 
|---|
| 453 | length = indexes[index + 1] - offset; | 
|---|
| 454 | if(length > 0) { | 
|---|
| 455 | ds->swapArray16(ds, inBytes + offset, length, outBytes + offset, &errorCode); | 
|---|
| 456 | } | 
|---|
| 457 |  | 
|---|
| 458 | // Skip the  IX_COMPRESSIBLE_BYTES_OFFSET byte array. | 
|---|
| 459 |  | 
|---|
| 460 | index = IX_RESERVED18_OFFSET; | 
|---|
| 461 | offset = indexes[index]; | 
|---|
| 462 | length = indexes[index + 1] - offset; | 
|---|
| 463 | if(length > 0) { | 
|---|
| 464 | udata_printError(ds, "ucol_swap(formatVersion=4): unknown data at IX_RESERVED18_OFFSET\n", length); | 
|---|
| 465 | errorCode = U_UNSUPPORTED_ERROR; | 
|---|
| 466 | return 0; | 
|---|
| 467 | } | 
|---|
| 468 |  | 
|---|
| 469 | return size; | 
|---|
| 470 | } | 
|---|
| 471 |  | 
|---|
| 472 | }  // namespace | 
|---|
| 473 |  | 
|---|
| 474 | /* swap ICU collation data like ucadata.icu */ | 
|---|
| 475 | U_CAPI int32_t U_EXPORT2 | 
|---|
| 476 | ucol_swap(const UDataSwapper *ds, | 
|---|
| 477 | const void *inData, int32_t length, void *outData, | 
|---|
| 478 | UErrorCode *pErrorCode) { | 
|---|
| 479 | if(U_FAILURE(*pErrorCode)) { return 0; } | 
|---|
| 480 |  | 
|---|
| 481 | /* udata_swapDataHeader checks the arguments */ | 
|---|
| 482 | int32_t headerSize=udata_swapDataHeader(ds, inData, length, outData, pErrorCode); | 
|---|
| 483 | if(U_FAILURE(*pErrorCode)) { | 
|---|
| 484 | // Try to swap the old format version which did not have a standard data header. | 
|---|
| 485 | *pErrorCode=U_ZERO_ERROR; | 
|---|
| 486 | return swapFormatVersion3(ds, inData, length, outData, pErrorCode); | 
|---|
| 487 | } | 
|---|
| 488 |  | 
|---|
| 489 | /* check data format and format version */ | 
|---|
| 490 | const UDataInfo &info=*(const UDataInfo *)((const char *)inData+4); | 
|---|
| 491 | if(!( | 
|---|
| 492 | info.dataFormat[0]==0x55 &&   // dataFormat="UCol" | 
|---|
| 493 | info.dataFormat[1]==0x43 && | 
|---|
| 494 | info.dataFormat[2]==0x6f && | 
|---|
| 495 | info.dataFormat[3]==0x6c && | 
|---|
| 496 | (3<=info.formatVersion[0] && info.formatVersion[0]<=5) | 
|---|
| 497 | )) { | 
|---|
| 498 | udata_printError(ds, "ucol_swap(): data format %02x.%02x.%02x.%02x " | 
|---|
| 499 | "(format version %02x.%02x) is not recognized as collation data\n", | 
|---|
| 500 | info.dataFormat[0], info.dataFormat[1], | 
|---|
| 501 | info.dataFormat[2], info.dataFormat[3], | 
|---|
| 502 | info.formatVersion[0], info.formatVersion[1]); | 
|---|
| 503 | *pErrorCode=U_UNSUPPORTED_ERROR; | 
|---|
| 504 | return 0; | 
|---|
| 505 | } | 
|---|
| 506 |  | 
|---|
| 507 | inData=(const char *)inData+headerSize; | 
|---|
| 508 | if(length>=0) { length-=headerSize; } | 
|---|
| 509 | outData=(outData == nullptr) ? nullptr : (char *)outData+headerSize; | 
|---|
| 510 | int32_t collationSize; | 
|---|
| 511 | if(info.formatVersion[0]>=4) { | 
|---|
| 512 | collationSize=swapFormatVersion4(ds, inData, length, outData, *pErrorCode); | 
|---|
| 513 | } else { | 
|---|
| 514 | collationSize=swapFormatVersion3(ds, inData, length, outData, pErrorCode); | 
|---|
| 515 | } | 
|---|
| 516 | if(U_SUCCESS(*pErrorCode)) { | 
|---|
| 517 | return headerSize+collationSize; | 
|---|
| 518 | } else { | 
|---|
| 519 | return 0; | 
|---|
| 520 | } | 
|---|
| 521 | } | 
|---|
| 522 |  | 
|---|
| 523 | /* swap inverse UCA collation data (invuca.icu) */ | 
|---|
| 524 | U_CAPI int32_t U_EXPORT2 | 
|---|
| 525 | ucol_swapInverseUCA(const UDataSwapper *ds, | 
|---|
| 526 | const void *inData, int32_t length, void *outData, | 
|---|
| 527 | UErrorCode *pErrorCode) { | 
|---|
| 528 | const UDataInfo *pInfo; | 
|---|
| 529 | int32_t headerSize; | 
|---|
| 530 |  | 
|---|
| 531 | const uint8_t *inBytes; | 
|---|
| 532 | uint8_t *outBytes; | 
|---|
| 533 |  | 
|---|
| 534 | const InverseUCATableHeader *inHeader; | 
|---|
| 535 | InverseUCATableHeader *outHeader; | 
|---|
| 536 | InverseUCATableHeader header={ 0,0,0,0,0,{0,0,0,0},{0,0,0,0,0,0,0,0} }; | 
|---|
| 537 |  | 
|---|
| 538 | /* udata_swapDataHeader checks the arguments */ | 
|---|
| 539 | headerSize=udata_swapDataHeader(ds, inData, length, outData, pErrorCode); | 
|---|
| 540 | if(pErrorCode==nullptr || U_FAILURE(*pErrorCode)) { | 
|---|
| 541 | return 0; | 
|---|
| 542 | } | 
|---|
| 543 |  | 
|---|
| 544 | /* check data format and format version */ | 
|---|
| 545 | pInfo=(const UDataInfo *)((const char *)inData+4); | 
|---|
| 546 | if(!( | 
|---|
| 547 | pInfo->dataFormat[0]==0x49 &&   /* dataFormat="InvC" */ | 
|---|
| 548 | pInfo->dataFormat[1]==0x6e && | 
|---|
| 549 | pInfo->dataFormat[2]==0x76 && | 
|---|
| 550 | pInfo->dataFormat[3]==0x43 && | 
|---|
| 551 | pInfo->formatVersion[0]==2 && | 
|---|
| 552 | pInfo->formatVersion[1]>=1 | 
|---|
| 553 | )) { | 
|---|
| 554 | udata_printError(ds, "ucol_swapInverseUCA(): data format %02x.%02x.%02x.%02x (format version %02x.%02x) is not an inverse UCA collation file\n", | 
|---|
| 555 | pInfo->dataFormat[0], pInfo->dataFormat[1], | 
|---|
| 556 | pInfo->dataFormat[2], pInfo->dataFormat[3], | 
|---|
| 557 | pInfo->formatVersion[0], pInfo->formatVersion[1]); | 
|---|
| 558 | *pErrorCode=U_UNSUPPORTED_ERROR; | 
|---|
| 559 | return 0; | 
|---|
| 560 | } | 
|---|
| 561 |  | 
|---|
| 562 | inBytes=(const uint8_t *)inData+headerSize; | 
|---|
| 563 | outBytes=(uint8_t *)outData+headerSize; | 
|---|
| 564 |  | 
|---|
| 565 | inHeader=(const InverseUCATableHeader *)inBytes; | 
|---|
| 566 | outHeader=(InverseUCATableHeader *)outBytes; | 
|---|
| 567 |  | 
|---|
| 568 | /* | 
|---|
| 569 | * The inverse UCA collation binary must contain at least the InverseUCATableHeader, | 
|---|
| 570 | * starting with its size field. | 
|---|
| 571 | * sizeof(UCATableHeader)==8*4 in ICU 2.8 | 
|---|
| 572 | * check the length against the header size before reading the size field | 
|---|
| 573 | */ | 
|---|
| 574 | if(length<0) { | 
|---|
| 575 | header.byteSize=udata_readInt32(ds, inHeader->byteSize); | 
|---|
| 576 | } else if( | 
|---|
| 577 | ((length-headerSize)<(8*4) || | 
|---|
| 578 | (uint32_t)(length-headerSize)<(header.byteSize=udata_readInt32(ds, inHeader->byteSize))) | 
|---|
| 579 | ) { | 
|---|
| 580 | udata_printError(ds, "ucol_swapInverseUCA(): too few bytes (%d after header) for inverse UCA collation data\n", | 
|---|
| 581 | length); | 
|---|
| 582 | *pErrorCode=U_INDEX_OUTOFBOUNDS_ERROR; | 
|---|
| 583 | return 0; | 
|---|
| 584 | } | 
|---|
| 585 |  | 
|---|
| 586 | if(length>=0) { | 
|---|
| 587 | /* copy everything, takes care of data that needs no swapping */ | 
|---|
| 588 | if(inBytes!=outBytes) { | 
|---|
| 589 | uprv_memcpy(outBytes, inBytes, header.byteSize); | 
|---|
| 590 | } | 
|---|
| 591 |  | 
|---|
| 592 | /* swap the necessary pieces in the order of their occurrence in the data */ | 
|---|
| 593 |  | 
|---|
| 594 | /* read more of the InverseUCATableHeader (the byteSize field was read above) */ | 
|---|
| 595 | header.tableSize=   ds->readUInt32(inHeader->tableSize); | 
|---|
| 596 | header.contsSize=   ds->readUInt32(inHeader->contsSize); | 
|---|
| 597 | header.table=       ds->readUInt32(inHeader->table); | 
|---|
| 598 | header.conts=       ds->readUInt32(inHeader->conts); | 
|---|
| 599 |  | 
|---|
| 600 | /* swap the 32-bit integers in the header */ | 
|---|
| 601 | ds->swapArray32(ds, inHeader, 5*4, outHeader, pErrorCode); | 
|---|
| 602 |  | 
|---|
| 603 | /* swap the inverse table; tableSize counts uint32_t[3] rows */ | 
|---|
| 604 | ds->swapArray32(ds, inBytes+header.table, header.tableSize*3*4, | 
|---|
| 605 | outBytes+header.table, pErrorCode); | 
|---|
| 606 |  | 
|---|
| 607 | /* swap the continuation table; contsSize counts UChars */ | 
|---|
| 608 | ds->swapArray16(ds, inBytes+header.conts, header.contsSize*U_SIZEOF_UCHAR, | 
|---|
| 609 | outBytes+header.conts, pErrorCode); | 
|---|
| 610 | } | 
|---|
| 611 |  | 
|---|
| 612 | return headerSize+header.byteSize; | 
|---|
| 613 | } | 
|---|
| 614 |  | 
|---|
| 615 | #endif /* #if !UCONFIG_NO_COLLATION */ | 
|---|
| 616 |  | 
|---|