Jungshik Shin | 87232d8 | 2017-05-13 21:10:13 -0700 | [diff] [blame] | 1 | // © 2016 and later: Unicode, Inc. and others. |
Jungshik Shin | 5feb9ad | 2016-10-21 12:52:48 -0700 | [diff] [blame] | 2 | // License & terms of use: http://www.unicode.org/copyright.html |
jshin@chromium.org | 6f31ac3 | 2014-03-26 22:15:14 +0000 | [diff] [blame] | 3 | /* |
| 4 | ******************************************************************************* |
| 5 | * |
Jungshik Shin | 70f8250 | 2016-01-29 00:32:36 -0800 | [diff] [blame] | 6 | * Copyright (C) 1999-2015, International Business Machines |
jshin@chromium.org | 6f31ac3 | 2014-03-26 22:15:14 +0000 | [diff] [blame] | 7 | * Corporation and others. All Rights Reserved. |
| 8 | * |
| 9 | ******************************************************************************* |
| 10 | * file name: package.cpp |
Jungshik Shin | 87232d8 | 2017-05-13 21:10:13 -0700 | [diff] [blame] | 11 | * encoding: UTF-8 |
jshin@chromium.org | 6f31ac3 | 2014-03-26 22:15:14 +0000 | [diff] [blame] | 12 | * tab size: 8 (not used) |
| 13 | * indentation:4 |
| 14 | * |
| 15 | * created on: 2005aug25 |
| 16 | * created by: Markus W. Scherer |
| 17 | * |
| 18 | * Read, modify, and write ICU .dat data package files. |
| 19 | * This is an integral part of the icupkg tool, moved to the toolutil library |
| 20 | * because parts of tool implementations tend to be later shared by |
| 21 | * other tools. |
| 22 | * Subsumes functionality and implementation code from |
| 23 | * gencmn, decmn, and icuswap tools. |
| 24 | */ |
| 25 | |
| 26 | #include "unicode/utypes.h" |
| 27 | #include "unicode/putil.h" |
| 28 | #include "unicode/udata.h" |
| 29 | #include "cstring.h" |
| 30 | #include "uarrsort.h" |
| 31 | #include "ucmndata.h" |
| 32 | #include "udataswp.h" |
| 33 | #include "swapimpl.h" |
| 34 | #include "toolutil.h" |
| 35 | #include "package.h" |
| 36 | #include "cmemory.h" |
| 37 | |
| 38 | #include <stdio.h> |
| 39 | #include <stdlib.h> |
| 40 | #include <string.h> |
| 41 | |
| 42 | |
| 43 | static const int32_t kItemsChunk = 256; /* How much to increase the filesarray by each time */ |
| 44 | |
| 45 | // general definitions ----------------------------------------------------- *** |
| 46 | |
jshin@chromium.org | 6f31ac3 | 2014-03-26 22:15:14 +0000 | [diff] [blame] | 47 | /* UDataInfo cf. udata.h */ |
| 48 | static const UDataInfo dataInfo={ |
| 49 | (uint16_t)sizeof(UDataInfo), |
| 50 | 0, |
| 51 | |
| 52 | U_IS_BIG_ENDIAN, |
| 53 | U_CHARSET_FAMILY, |
| 54 | (uint8_t)sizeof(UChar), |
| 55 | 0, |
| 56 | |
| 57 | {0x43, 0x6d, 0x6e, 0x44}, /* dataFormat="CmnD" */ |
| 58 | {1, 0, 0, 0}, /* formatVersion */ |
| 59 | {3, 0, 0, 0} /* dataVersion */ |
| 60 | }; |
| 61 | |
| 62 | U_CDECL_BEGIN |
| 63 | static void U_CALLCONV |
| 64 | printPackageError(void *context, const char *fmt, va_list args) { |
| 65 | vfprintf((FILE *)context, fmt, args); |
| 66 | } |
| 67 | U_CDECL_END |
| 68 | |
| 69 | static uint16_t |
| 70 | readSwapUInt16(uint16_t x) { |
| 71 | return (uint16_t)((x<<8)|(x>>8)); |
| 72 | } |
| 73 | |
| 74 | // platform types ---------------------------------------------------------- *** |
| 75 | |
| 76 | static const char *types="lb?e"; |
| 77 | |
| 78 | enum { TYPE_L, TYPE_B, TYPE_LE, TYPE_E, TYPE_COUNT }; |
| 79 | |
| 80 | static inline int32_t |
| 81 | makeTypeEnum(uint8_t charset, UBool isBigEndian) { |
| 82 | return 2*(int32_t)charset+isBigEndian; |
| 83 | } |
| 84 | |
| 85 | static inline int32_t |
| 86 | makeTypeEnum(char type) { |
| 87 | return |
| 88 | type == 'l' ? TYPE_L : |
| 89 | type == 'b' ? TYPE_B : |
| 90 | type == 'e' ? TYPE_E : |
| 91 | -1; |
| 92 | } |
| 93 | |
| 94 | static inline char |
| 95 | makeTypeLetter(uint8_t charset, UBool isBigEndian) { |
| 96 | return types[makeTypeEnum(charset, isBigEndian)]; |
| 97 | } |
| 98 | |
| 99 | static inline char |
| 100 | makeTypeLetter(int32_t typeEnum) { |
| 101 | return types[typeEnum]; |
| 102 | } |
| 103 | |
| 104 | static void |
| 105 | makeTypeProps(char type, uint8_t &charset, UBool &isBigEndian) { |
| 106 | int32_t typeEnum=makeTypeEnum(type); |
| 107 | charset=(uint8_t)(typeEnum>>1); |
| 108 | isBigEndian=(UBool)(typeEnum&1); |
| 109 | } |
| 110 | |
| 111 | U_CFUNC const UDataInfo * |
| 112 | getDataInfo(const uint8_t *data, int32_t length, |
| 113 | int32_t &infoLength, int32_t &headerLength, |
| 114 | UErrorCode *pErrorCode) { |
| 115 | const DataHeader *pHeader; |
| 116 | const UDataInfo *pInfo; |
| 117 | |
| 118 | if(pErrorCode==NULL || U_FAILURE(*pErrorCode)) { |
| 119 | return NULL; |
| 120 | } |
| 121 | if( data==NULL || |
| 122 | (length>=0 && length<(int32_t)sizeof(DataHeader)) |
| 123 | ) { |
| 124 | *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR; |
| 125 | return NULL; |
| 126 | } |
| 127 | |
| 128 | pHeader=(const DataHeader *)data; |
| 129 | pInfo=&pHeader->info; |
| 130 | if( (length>=0 && length<(int32_t)sizeof(DataHeader)) || |
| 131 | pHeader->dataHeader.magic1!=0xda || |
| 132 | pHeader->dataHeader.magic2!=0x27 || |
| 133 | pInfo->sizeofUChar!=2 |
| 134 | ) { |
| 135 | *pErrorCode=U_UNSUPPORTED_ERROR; |
| 136 | return NULL; |
| 137 | } |
| 138 | |
| 139 | if(pInfo->isBigEndian==U_IS_BIG_ENDIAN) { |
| 140 | headerLength=pHeader->dataHeader.headerSize; |
| 141 | infoLength=pInfo->size; |
| 142 | } else { |
| 143 | headerLength=readSwapUInt16(pHeader->dataHeader.headerSize); |
| 144 | infoLength=readSwapUInt16(pInfo->size); |
| 145 | } |
| 146 | |
| 147 | if( headerLength<(int32_t)sizeof(DataHeader) || |
| 148 | infoLength<(int32_t)sizeof(UDataInfo) || |
| 149 | headerLength<(int32_t)(sizeof(pHeader->dataHeader)+infoLength) || |
| 150 | (length>=0 && length<headerLength) |
| 151 | ) { |
| 152 | *pErrorCode=U_UNSUPPORTED_ERROR; |
| 153 | return NULL; |
| 154 | } |
| 155 | |
| 156 | return pInfo; |
| 157 | } |
| 158 | |
| 159 | static int32_t |
| 160 | getTypeEnumForInputData(const uint8_t *data, int32_t length, |
| 161 | UErrorCode *pErrorCode) { |
| 162 | const UDataInfo *pInfo; |
| 163 | int32_t infoLength, headerLength; |
| 164 | |
| 165 | /* getDataInfo() checks for illegal arguments */ |
| 166 | pInfo=getDataInfo(data, length, infoLength, headerLength, pErrorCode); |
| 167 | if(pInfo==NULL) { |
| 168 | return -1; |
| 169 | } |
| 170 | |
| 171 | return makeTypeEnum(pInfo->charsetFamily, (UBool)pInfo->isBigEndian); |
| 172 | } |
| 173 | |
| 174 | // file handling ----------------------------------------------------------- *** |
| 175 | |
| 176 | static void |
| 177 | extractPackageName(const char *filename, |
| 178 | char pkg[], int32_t capacity) { |
| 179 | const char *basename; |
| 180 | int32_t len; |
| 181 | |
| 182 | basename=findBasename(filename); |
| 183 | len=(int32_t)strlen(basename)-4; /* -4: subtract the length of ".dat" */ |
| 184 | |
| 185 | if(len<=0 || 0!=strcmp(basename+len, ".dat")) { |
| 186 | fprintf(stderr, "icupkg: \"%s\" is not recognized as a package filename (must end with .dat)\n", |
| 187 | basename); |
| 188 | exit(U_ILLEGAL_ARGUMENT_ERROR); |
| 189 | } |
| 190 | |
| 191 | if(len>=capacity) { |
| 192 | fprintf(stderr, "icupkg: the package name \"%s\" is too long (>=%ld)\n", |
| 193 | basename, (long)capacity); |
| 194 | exit(U_ILLEGAL_ARGUMENT_ERROR); |
| 195 | } |
| 196 | |
| 197 | memcpy(pkg, basename, len); |
| 198 | pkg[len]=0; |
| 199 | } |
| 200 | |
| 201 | static int32_t |
| 202 | getFileLength(FILE *f) { |
| 203 | int32_t length; |
| 204 | |
| 205 | fseek(f, 0, SEEK_END); |
| 206 | length=(int32_t)ftell(f); |
| 207 | fseek(f, 0, SEEK_SET); |
| 208 | return length; |
| 209 | } |
| 210 | |
| 211 | /* |
| 212 | * Turn tree separators and alternate file separators into normal file separators. |
| 213 | */ |
| 214 | #if U_TREE_ENTRY_SEP_CHAR==U_FILE_SEP_CHAR && U_FILE_ALT_SEP_CHAR==U_FILE_SEP_CHAR |
| 215 | #define treeToPath(s) |
| 216 | #else |
| 217 | static void |
| 218 | treeToPath(char *s) { |
| 219 | char *t; |
| 220 | |
| 221 | for(t=s; *t!=0; ++t) { |
| 222 | if(*t==U_TREE_ENTRY_SEP_CHAR || *t==U_FILE_ALT_SEP_CHAR) { |
| 223 | *t=U_FILE_SEP_CHAR; |
| 224 | } |
| 225 | } |
| 226 | } |
| 227 | #endif |
| 228 | |
| 229 | /* |
| 230 | * Turn file separators into tree separators. |
| 231 | */ |
| 232 | #if U_TREE_ENTRY_SEP_CHAR==U_FILE_SEP_CHAR && U_FILE_ALT_SEP_CHAR==U_FILE_SEP_CHAR |
| 233 | #define pathToTree(s) |
| 234 | #else |
| 235 | static void |
| 236 | pathToTree(char *s) { |
| 237 | char *t; |
| 238 | |
| 239 | for(t=s; *t!=0; ++t) { |
| 240 | if(*t==U_FILE_SEP_CHAR || *t==U_FILE_ALT_SEP_CHAR) { |
| 241 | *t=U_TREE_ENTRY_SEP_CHAR; |
| 242 | } |
| 243 | } |
| 244 | } |
| 245 | #endif |
| 246 | |
| 247 | /* |
| 248 | * Prepend the path (if any) to the name and run the name through treeToName(). |
| 249 | */ |
| 250 | static void |
| 251 | makeFullFilename(const char *path, const char *name, |
| 252 | char *filename, int32_t capacity) { |
| 253 | char *s; |
| 254 | |
| 255 | // prepend the path unless NULL or empty |
| 256 | if(path!=NULL && path[0]!=0) { |
| 257 | if((int32_t)(strlen(path)+1)>=capacity) { |
| 258 | fprintf(stderr, "pathname too long: \"%s\"\n", path); |
| 259 | exit(U_BUFFER_OVERFLOW_ERROR); |
| 260 | } |
| 261 | strcpy(filename, path); |
| 262 | |
| 263 | // make sure the path ends with a file separator |
| 264 | s=strchr(filename, 0); |
| 265 | if(*(s-1)!=U_FILE_SEP_CHAR && *(s-1)!=U_FILE_ALT_SEP_CHAR) { |
| 266 | *s++=U_FILE_SEP_CHAR; |
| 267 | } |
| 268 | } else { |
| 269 | s=filename; |
| 270 | } |
| 271 | |
| 272 | // turn the name into a filename, turn tree separators into file separators |
| 273 | if((int32_t)((s-filename)+strlen(name))>=capacity) { |
| 274 | fprintf(stderr, "path/filename too long: \"%s%s\"\n", filename, name); |
| 275 | exit(U_BUFFER_OVERFLOW_ERROR); |
| 276 | } |
| 277 | strcpy(s, name); |
| 278 | treeToPath(s); |
| 279 | } |
| 280 | |
| 281 | static void |
| 282 | makeFullFilenameAndDirs(const char *path, const char *name, |
| 283 | char *filename, int32_t capacity) { |
| 284 | char *sep; |
| 285 | UErrorCode errorCode; |
| 286 | |
| 287 | makeFullFilename(path, name, filename, capacity); |
| 288 | |
| 289 | // make tree directories |
| 290 | errorCode=U_ZERO_ERROR; |
| 291 | sep=strchr(filename, 0)-strlen(name); |
| 292 | while((sep=strchr(sep, U_FILE_SEP_CHAR))!=NULL) { |
| 293 | if(sep!=filename) { |
| 294 | *sep=0; // truncate temporarily |
| 295 | uprv_mkdir(filename, &errorCode); |
| 296 | if(U_FAILURE(errorCode)) { |
| 297 | fprintf(stderr, "icupkg: unable to create tree directory \"%s\"\n", filename); |
| 298 | exit(U_FILE_ACCESS_ERROR); |
| 299 | } |
| 300 | } |
| 301 | *sep++=U_FILE_SEP_CHAR; // restore file separator character |
| 302 | } |
| 303 | } |
| 304 | |
| 305 | static uint8_t * |
| 306 | readFile(const char *path, const char *name, int32_t &length, char &type) { |
| 307 | char filename[1024]; |
| 308 | FILE *file; |
jshin@chromium.org | 6f31ac3 | 2014-03-26 22:15:14 +0000 | [diff] [blame] | 309 | UErrorCode errorCode; |
| 310 | int32_t fileLength, typeEnum; |
| 311 | |
| 312 | makeFullFilename(path, name, filename, (int32_t)sizeof(filename)); |
| 313 | |
| 314 | /* open the input file, get its length, allocate memory for it, read the file */ |
| 315 | file=fopen(filename, "rb"); |
| 316 | if(file==NULL) { |
| 317 | fprintf(stderr, "icupkg: unable to open input file \"%s\"\n", filename); |
| 318 | exit(U_FILE_ACCESS_ERROR); |
| 319 | } |
| 320 | |
| 321 | /* get the file length */ |
| 322 | fileLength=getFileLength(file); |
| 323 | if(ferror(file) || fileLength<=0) { |
| 324 | fprintf(stderr, "icupkg: empty input file \"%s\"\n", filename); |
| 325 | fclose(file); |
| 326 | exit(U_FILE_ACCESS_ERROR); |
| 327 | } |
| 328 | |
| 329 | /* allocate the buffer, pad to multiple of 16 */ |
| 330 | length=(fileLength+0xf)&~0xf; |
Jungshik Shin | 70f8250 | 2016-01-29 00:32:36 -0800 | [diff] [blame] | 331 | icu::LocalMemory<uint8_t> data((uint8_t *)uprv_malloc(length)); |
| 332 | if(data.isNull()) { |
jshin@chromium.org | 6f31ac3 | 2014-03-26 22:15:14 +0000 | [diff] [blame] | 333 | fclose(file); |
| 334 | fprintf(stderr, "icupkg: malloc error allocating %d bytes.\n", (int)length); |
| 335 | exit(U_MEMORY_ALLOCATION_ERROR); |
| 336 | } |
| 337 | |
| 338 | /* read the file */ |
Jungshik Shin | 70f8250 | 2016-01-29 00:32:36 -0800 | [diff] [blame] | 339 | if(fileLength!=(int32_t)fread(data.getAlias(), 1, fileLength, file)) { |
jshin@chromium.org | 6f31ac3 | 2014-03-26 22:15:14 +0000 | [diff] [blame] | 340 | fprintf(stderr, "icupkg: error reading \"%s\"\n", filename); |
| 341 | fclose(file); |
jshin@chromium.org | 6f31ac3 | 2014-03-26 22:15:14 +0000 | [diff] [blame] | 342 | exit(U_FILE_ACCESS_ERROR); |
| 343 | } |
| 344 | |
| 345 | /* pad the file to a multiple of 16 using the usual padding byte */ |
| 346 | if(fileLength<length) { |
Jungshik Shin | 70f8250 | 2016-01-29 00:32:36 -0800 | [diff] [blame] | 347 | memset(data.getAlias()+fileLength, 0xaa, length-fileLength); |
jshin@chromium.org | 6f31ac3 | 2014-03-26 22:15:14 +0000 | [diff] [blame] | 348 | } |
| 349 | |
| 350 | fclose(file); |
| 351 | |
| 352 | // minimum check for ICU-format data |
| 353 | errorCode=U_ZERO_ERROR; |
Jungshik Shin | 70f8250 | 2016-01-29 00:32:36 -0800 | [diff] [blame] | 354 | typeEnum=getTypeEnumForInputData(data.getAlias(), length, &errorCode); |
jshin@chromium.org | 6f31ac3 | 2014-03-26 22:15:14 +0000 | [diff] [blame] | 355 | if(typeEnum<0 || U_FAILURE(errorCode)) { |
| 356 | fprintf(stderr, "icupkg: not an ICU data file: \"%s\"\n", filename); |
jshin@chromium.org | 6f31ac3 | 2014-03-26 22:15:14 +0000 | [diff] [blame] | 357 | #if !UCONFIG_NO_LEGACY_CONVERSION |
| 358 | exit(U_INVALID_FORMAT_ERROR); |
| 359 | #else |
| 360 | fprintf(stderr, "U_INVALID_FORMAT_ERROR occurred but UCONFIG_NO_LEGACY_CONVERSION is on so this is expected.\n"); |
| 361 | exit(0); |
| 362 | #endif |
| 363 | } |
| 364 | type=makeTypeLetter(typeEnum); |
| 365 | |
Jungshik Shin | 70f8250 | 2016-01-29 00:32:36 -0800 | [diff] [blame] | 366 | return data.orphan(); |
jshin@chromium.org | 6f31ac3 | 2014-03-26 22:15:14 +0000 | [diff] [blame] | 367 | } |
| 368 | |
| 369 | // .dat package file representation ---------------------------------------- *** |
| 370 | |
| 371 | U_CDECL_BEGIN |
| 372 | |
| 373 | static int32_t U_CALLCONV |
| 374 | compareItems(const void * /*context*/, const void *left, const void *right) { |
| 375 | U_NAMESPACE_USE |
| 376 | |
| 377 | return (int32_t)strcmp(((Item *)left)->name, ((Item *)right)->name); |
| 378 | } |
| 379 | |
| 380 | U_CDECL_END |
| 381 | |
| 382 | U_NAMESPACE_BEGIN |
| 383 | |
Jungshik Shin | 87232d8 | 2017-05-13 21:10:13 -0700 | [diff] [blame] | 384 | Package::Package() |
Frank Tang | 1f164ee | 2022-11-08 12:31:27 -0800 | [diff] [blame^] | 385 | : doAutoPrefix(false), prefixEndsWithType(false) { |
jshin@chromium.org | 6f31ac3 | 2014-03-26 22:15:14 +0000 | [diff] [blame] | 386 | inPkgName[0]=0; |
| 387 | pkgPrefix[0]=0; |
| 388 | inData=NULL; |
| 389 | inLength=0; |
| 390 | inCharset=U_CHARSET_FAMILY; |
| 391 | inIsBigEndian=U_IS_BIG_ENDIAN; |
| 392 | |
| 393 | itemCount=0; |
| 394 | itemMax=0; |
| 395 | items=NULL; |
| 396 | |
| 397 | inStringTop=outStringTop=0; |
| 398 | |
| 399 | matchMode=0; |
| 400 | findPrefix=findSuffix=NULL; |
| 401 | findPrefixLength=findSuffixLength=0; |
| 402 | findNextIndex=-1; |
| 403 | |
| 404 | // create a header for an empty package |
| 405 | DataHeader *pHeader; |
| 406 | pHeader=(DataHeader *)header; |
| 407 | pHeader->dataHeader.magic1=0xda; |
| 408 | pHeader->dataHeader.magic2=0x27; |
| 409 | memcpy(&pHeader->info, &dataInfo, sizeof(dataInfo)); |
| 410 | headerLength=(int32_t)(4+sizeof(dataInfo)); |
| 411 | if(headerLength&0xf) { |
| 412 | /* NUL-pad the header to a multiple of 16 */ |
| 413 | int32_t length=(headerLength+0xf)&~0xf; |
| 414 | memset(header+headerLength, 0, length-headerLength); |
| 415 | headerLength=length; |
| 416 | } |
| 417 | pHeader->dataHeader.headerSize=(uint16_t)headerLength; |
| 418 | } |
| 419 | |
| 420 | Package::~Package() { |
| 421 | int32_t idx; |
| 422 | |
Jungshik Shin | 70f8250 | 2016-01-29 00:32:36 -0800 | [diff] [blame] | 423 | uprv_free(inData); |
jshin@chromium.org | 6f31ac3 | 2014-03-26 22:15:14 +0000 | [diff] [blame] | 424 | |
| 425 | for(idx=0; idx<itemCount; ++idx) { |
| 426 | if(items[idx].isDataOwned) { |
Jungshik Shin | 70f8250 | 2016-01-29 00:32:36 -0800 | [diff] [blame] | 427 | uprv_free(items[idx].data); |
jshin@chromium.org | 6f31ac3 | 2014-03-26 22:15:14 +0000 | [diff] [blame] | 428 | } |
| 429 | } |
| 430 | |
| 431 | uprv_free((void*)items); |
| 432 | } |
| 433 | |
| 434 | void |
| 435 | Package::setPrefix(const char *p) { |
| 436 | if(strlen(p)>=sizeof(pkgPrefix)) { |
| 437 | fprintf(stderr, "icupkg: --toc_prefix %s too long\n", p); |
| 438 | exit(U_ILLEGAL_ARGUMENT_ERROR); |
| 439 | } |
| 440 | strcpy(pkgPrefix, p); |
| 441 | } |
| 442 | |
| 443 | void |
| 444 | Package::readPackage(const char *filename) { |
| 445 | UDataSwapper *ds; |
| 446 | const UDataInfo *pInfo; |
| 447 | UErrorCode errorCode; |
| 448 | |
| 449 | const uint8_t *inBytes; |
| 450 | |
| 451 | int32_t length, offset, i; |
| 452 | int32_t itemLength, typeEnum; |
| 453 | char type; |
| 454 | |
| 455 | const UDataOffsetTOCEntry *inEntries; |
| 456 | |
| 457 | extractPackageName(filename, inPkgName, (int32_t)sizeof(inPkgName)); |
| 458 | |
| 459 | /* read the file */ |
| 460 | inData=readFile(NULL, filename, inLength, type); |
| 461 | length=inLength; |
| 462 | |
| 463 | /* |
| 464 | * swap the header - even if the swapping itself is a no-op |
| 465 | * because it tells us the header length |
| 466 | */ |
| 467 | errorCode=U_ZERO_ERROR; |
| 468 | makeTypeProps(type, inCharset, inIsBigEndian); |
| 469 | ds=udata_openSwapper(inIsBigEndian, inCharset, U_IS_BIG_ENDIAN, U_CHARSET_FAMILY, &errorCode); |
| 470 | if(U_FAILURE(errorCode)) { |
| 471 | fprintf(stderr, "icupkg: udata_openSwapper(\"%s\") failed - %s\n", |
| 472 | filename, u_errorName(errorCode)); |
| 473 | exit(errorCode); |
| 474 | } |
| 475 | |
| 476 | ds->printError=printPackageError; |
| 477 | ds->printErrorContext=stderr; |
| 478 | |
| 479 | headerLength=sizeof(header); |
| 480 | if(length<headerLength) { |
| 481 | headerLength=length; |
| 482 | } |
| 483 | headerLength=udata_swapDataHeader(ds, inData, headerLength, header, &errorCode); |
| 484 | if(U_FAILURE(errorCode)) { |
| 485 | exit(errorCode); |
| 486 | } |
| 487 | |
| 488 | /* check data format and format version */ |
| 489 | pInfo=(const UDataInfo *)((const char *)inData+4); |
| 490 | if(!( |
| 491 | pInfo->dataFormat[0]==0x43 && /* dataFormat="CmnD" */ |
| 492 | pInfo->dataFormat[1]==0x6d && |
| 493 | pInfo->dataFormat[2]==0x6e && |
| 494 | pInfo->dataFormat[3]==0x44 && |
| 495 | pInfo->formatVersion[0]==1 |
| 496 | )) { |
| 497 | fprintf(stderr, "icupkg: data format %02x.%02x.%02x.%02x (format version %02x) is not recognized as an ICU .dat package\n", |
| 498 | pInfo->dataFormat[0], pInfo->dataFormat[1], |
| 499 | pInfo->dataFormat[2], pInfo->dataFormat[3], |
| 500 | pInfo->formatVersion[0]); |
| 501 | exit(U_UNSUPPORTED_ERROR); |
| 502 | } |
| 503 | inIsBigEndian=(UBool)pInfo->isBigEndian; |
| 504 | inCharset=pInfo->charsetFamily; |
| 505 | |
| 506 | inBytes=(const uint8_t *)inData+headerLength; |
| 507 | inEntries=(const UDataOffsetTOCEntry *)(inBytes+4); |
| 508 | |
| 509 | /* check that the itemCount fits, then the ToC table, then at least the header of the last item */ |
| 510 | length-=headerLength; |
| 511 | if(length<4) { |
| 512 | /* itemCount does not fit */ |
| 513 | offset=0x7fffffff; |
| 514 | } else { |
| 515 | itemCount=udata_readInt32(ds, *(const int32_t *)inBytes); |
| 516 | setItemCapacity(itemCount); /* resize so there's space */ |
| 517 | if(itemCount==0) { |
| 518 | offset=4; |
| 519 | } else if(length<(4+8*itemCount)) { |
| 520 | /* ToC table does not fit */ |
| 521 | offset=0x7fffffff; |
| 522 | } else { |
| 523 | /* offset of the last item plus at least 20 bytes for its header */ |
| 524 | offset=20+(int32_t)ds->readUInt32(inEntries[itemCount-1].dataOffset); |
| 525 | } |
| 526 | } |
| 527 | if(length<offset) { |
| 528 | fprintf(stderr, "icupkg: too few bytes (%ld after header) for a .dat package\n", |
| 529 | (long)length); |
| 530 | exit(U_INDEX_OUTOFBOUNDS_ERROR); |
| 531 | } |
| 532 | /* do not modify the package length variable until the last item's length is set */ |
| 533 | |
| 534 | if(itemCount<=0) { |
| 535 | if(doAutoPrefix) { |
| 536 | fprintf(stderr, "icupkg: --auto_toc_prefix[_with_type] but the input package is empty\n"); |
| 537 | exit(U_INVALID_FORMAT_ERROR); |
| 538 | } |
| 539 | } else { |
| 540 | char prefix[MAX_PKG_NAME_LENGTH+4]; |
| 541 | char *s, *inItemStrings; |
| 542 | |
| 543 | if(itemCount>itemMax) { |
| 544 | fprintf(stderr, "icupkg: too many items, maximum is %d\n", itemMax); |
| 545 | exit(U_BUFFER_OVERFLOW_ERROR); |
| 546 | } |
| 547 | |
| 548 | /* swap the item name strings */ |
| 549 | int32_t stringsOffset=4+8*itemCount; |
| 550 | itemLength=(int32_t)(ds->readUInt32(inEntries[0].dataOffset))-stringsOffset; |
| 551 | |
| 552 | // don't include padding bytes at the end of the item names |
| 553 | while(itemLength>0 && inBytes[stringsOffset+itemLength-1]!=0) { |
| 554 | --itemLength; |
| 555 | } |
| 556 | |
| 557 | if((inStringTop+itemLength)>STRING_STORE_SIZE) { |
| 558 | fprintf(stderr, "icupkg: total length of item name strings too long\n"); |
| 559 | exit(U_BUFFER_OVERFLOW_ERROR); |
| 560 | } |
| 561 | |
| 562 | inItemStrings=inStrings+inStringTop; |
| 563 | ds->swapInvChars(ds, inBytes+stringsOffset, itemLength, inItemStrings, &errorCode); |
| 564 | if(U_FAILURE(errorCode)) { |
| 565 | fprintf(stderr, "icupkg failed to swap the input .dat package item name strings\n"); |
| 566 | exit(U_INVALID_FORMAT_ERROR); |
| 567 | } |
| 568 | inStringTop+=itemLength; |
| 569 | |
| 570 | // reset the Item entries |
| 571 | memset(items, 0, itemCount*sizeof(Item)); |
| 572 | |
| 573 | /* |
| 574 | * Get the common prefix of the items. |
| 575 | * New-style ICU .dat packages use tree separators ('/') between package names, |
| 576 | * tree names, and item names, |
| 577 | * while old-style ICU .dat packages (before multi-tree support) |
| 578 | * use an underscore ('_') between package and item names. |
| 579 | */ |
| 580 | offset=(int32_t)ds->readUInt32(inEntries[0].nameOffset)-stringsOffset; |
| 581 | s=inItemStrings+offset; // name of the first entry |
| 582 | int32_t prefixLength; |
| 583 | if(doAutoPrefix) { |
| 584 | // Use the first entry's prefix. Must be a new-style package. |
| 585 | const char *prefixLimit=strchr(s, U_TREE_ENTRY_SEP_CHAR); |
| 586 | if(prefixLimit==NULL) { |
| 587 | fprintf(stderr, |
| 588 | "icupkg: --auto_toc_prefix[_with_type] but " |
| 589 | "the first entry \"%s\" does not contain a '%c'\n", |
| 590 | s, U_TREE_ENTRY_SEP_CHAR); |
| 591 | exit(U_INVALID_FORMAT_ERROR); |
| 592 | } |
| 593 | prefixLength=(int32_t)(prefixLimit-s); |
Jungshik Shin (jungshik at google) | 0f8746a | 2015-01-08 15:46:45 -0800 | [diff] [blame] | 594 | if(prefixLength==0 || prefixLength>=UPRV_LENGTHOF(pkgPrefix)) { |
jshin@chromium.org | 6f31ac3 | 2014-03-26 22:15:14 +0000 | [diff] [blame] | 595 | fprintf(stderr, |
| 596 | "icupkg: --auto_toc_prefix[_with_type] but " |
| 597 | "the prefix of the first entry \"%s\" is empty or too long\n", |
| 598 | s); |
| 599 | exit(U_INVALID_FORMAT_ERROR); |
| 600 | } |
| 601 | if(prefixEndsWithType && s[prefixLength-1]!=type) { |
| 602 | fprintf(stderr, |
| 603 | "icupkg: --auto_toc_prefix_with_type but " |
| 604 | "the prefix of the first entry \"%s\" does not end with '%c'\n", |
| 605 | s, type); |
| 606 | exit(U_INVALID_FORMAT_ERROR); |
| 607 | } |
| 608 | memcpy(pkgPrefix, s, prefixLength); |
Jungshik Shin (jungshik at google) | 0f8746a | 2015-01-08 15:46:45 -0800 | [diff] [blame] | 609 | pkgPrefix[prefixLength]=0; |
jshin@chromium.org | 6f31ac3 | 2014-03-26 22:15:14 +0000 | [diff] [blame] | 610 | memcpy(prefix, s, ++prefixLength); // include the / |
| 611 | } else { |
| 612 | // Use the package basename as prefix. |
Jungshik Shin | 42d5027 | 2018-10-24 01:22:09 -0700 | [diff] [blame] | 613 | int32_t inPkgNameLength= static_cast<int32_t>(strlen(inPkgName)); |
jshin@chromium.org | 6f31ac3 | 2014-03-26 22:15:14 +0000 | [diff] [blame] | 614 | memcpy(prefix, inPkgName, inPkgNameLength); |
| 615 | prefixLength=inPkgNameLength; |
| 616 | |
| 617 | if( (int32_t)strlen(s)>=(inPkgNameLength+2) && |
| 618 | 0==memcmp(s, inPkgName, inPkgNameLength) && |
| 619 | s[inPkgNameLength]=='_' |
| 620 | ) { |
| 621 | // old-style .dat package |
| 622 | prefix[prefixLength++]='_'; |
| 623 | } else { |
| 624 | // new-style .dat package |
| 625 | prefix[prefixLength++]=U_TREE_ENTRY_SEP_CHAR; |
| 626 | // if it turns out to not contain U_TREE_ENTRY_SEP_CHAR |
| 627 | // then the test in the loop below will fail |
| 628 | } |
| 629 | } |
| 630 | prefix[prefixLength]=0; |
| 631 | |
| 632 | /* read the ToC table */ |
| 633 | for(i=0; i<itemCount; ++i) { |
| 634 | // skip the package part of the item name, error if it does not match the actual package name |
| 635 | // or if nothing follows the package name |
| 636 | offset=(int32_t)ds->readUInt32(inEntries[i].nameOffset)-stringsOffset; |
| 637 | s=inItemStrings+offset; |
| 638 | if(0!=strncmp(s, prefix, prefixLength) || s[prefixLength]==0) { |
| 639 | fprintf(stderr, "icupkg: input .dat item name \"%s\" does not start with \"%s\"\n", |
| 640 | s, prefix); |
| 641 | exit(U_INVALID_FORMAT_ERROR); |
| 642 | } |
| 643 | items[i].name=s+prefixLength; |
| 644 | |
| 645 | // set the item's data |
| 646 | items[i].data=(uint8_t *)inBytes+ds->readUInt32(inEntries[i].dataOffset); |
| 647 | if(i>0) { |
| 648 | items[i-1].length=(int32_t)(items[i].data-items[i-1].data); |
| 649 | |
| 650 | // set the previous item's platform type |
| 651 | typeEnum=getTypeEnumForInputData(items[i-1].data, items[i-1].length, &errorCode); |
| 652 | if(typeEnum<0 || U_FAILURE(errorCode)) { |
| 653 | fprintf(stderr, "icupkg: not an ICU data file: item \"%s\" in \"%s\"\n", items[i-1].name, filename); |
| 654 | exit(U_INVALID_FORMAT_ERROR); |
| 655 | } |
| 656 | items[i-1].type=makeTypeLetter(typeEnum); |
| 657 | } |
Frank Tang | 1f164ee | 2022-11-08 12:31:27 -0800 | [diff] [blame^] | 658 | items[i].isDataOwned=false; |
jshin@chromium.org | 6f31ac3 | 2014-03-26 22:15:14 +0000 | [diff] [blame] | 659 | } |
| 660 | // set the last item's length |
| 661 | items[itemCount-1].length=length-ds->readUInt32(inEntries[itemCount-1].dataOffset); |
| 662 | |
| 663 | // set the last item's platform type |
| 664 | typeEnum=getTypeEnumForInputData(items[itemCount-1].data, items[itemCount-1].length, &errorCode); |
| 665 | if(typeEnum<0 || U_FAILURE(errorCode)) { |
Jungshik Shin | b318966 | 2017-11-07 11:18:34 -0800 | [diff] [blame] | 666 | fprintf(stderr, "icupkg: not an ICU data file: item \"%s\" in \"%s\"\n", items[itemCount-1].name, filename); |
jshin@chromium.org | 6f31ac3 | 2014-03-26 22:15:14 +0000 | [diff] [blame] | 667 | exit(U_INVALID_FORMAT_ERROR); |
| 668 | } |
| 669 | items[itemCount-1].type=makeTypeLetter(typeEnum); |
| 670 | |
| 671 | if(type!=U_ICUDATA_TYPE_LETTER[0]) { |
| 672 | // sort the item names for the local charset |
| 673 | sortItems(); |
| 674 | } |
| 675 | } |
| 676 | |
| 677 | udata_closeSwapper(ds); |
| 678 | } |
| 679 | |
| 680 | char |
| 681 | Package::getInType() { |
| 682 | return makeTypeLetter(inCharset, inIsBigEndian); |
| 683 | } |
| 684 | |
| 685 | void |
| 686 | Package::writePackage(const char *filename, char outType, const char *comment) { |
| 687 | char prefix[MAX_PKG_NAME_LENGTH+4]; |
| 688 | UDataOffsetTOCEntry entry; |
| 689 | UDataSwapper *dsLocalToOut, *ds[TYPE_COUNT]; |
| 690 | FILE *file; |
| 691 | Item *pItem; |
| 692 | char *name; |
| 693 | UErrorCode errorCode; |
| 694 | int32_t i, length, prefixLength, maxItemLength, basenameOffset, offset, outInt32; |
| 695 | uint8_t outCharset; |
| 696 | UBool outIsBigEndian; |
| 697 | |
| 698 | extractPackageName(filename, prefix, MAX_PKG_NAME_LENGTH); |
| 699 | |
| 700 | // if there is an explicit comment, then use it, else use what's in the current header |
| 701 | if(comment!=NULL) { |
| 702 | /* get the header size minus the current comment */ |
| 703 | DataHeader *pHeader; |
| 704 | int32_t length; |
| 705 | |
| 706 | pHeader=(DataHeader *)header; |
| 707 | headerLength=4+pHeader->info.size; |
| 708 | length=(int32_t)strlen(comment); |
| 709 | if((int32_t)(headerLength+length)>=(int32_t)sizeof(header)) { |
| 710 | fprintf(stderr, "icupkg: comment too long\n"); |
| 711 | exit(U_BUFFER_OVERFLOW_ERROR); |
| 712 | } |
| 713 | memcpy(header+headerLength, comment, length+1); |
| 714 | headerLength+=length; |
| 715 | if(headerLength&0xf) { |
| 716 | /* NUL-pad the header to a multiple of 16 */ |
| 717 | length=(headerLength+0xf)&~0xf; |
| 718 | memset(header+headerLength, 0, length-headerLength); |
| 719 | headerLength=length; |
| 720 | } |
| 721 | pHeader->dataHeader.headerSize=(uint16_t)headerLength; |
| 722 | } |
| 723 | |
| 724 | makeTypeProps(outType, outCharset, outIsBigEndian); |
| 725 | |
| 726 | // open (TYPE_COUNT-2) swappers |
| 727 | // one is a no-op for local type==outType |
| 728 | // one type (TYPE_LE) is bogus |
| 729 | errorCode=U_ZERO_ERROR; |
| 730 | i=makeTypeEnum(outType); |
Frank Tang | 1f164ee | 2022-11-08 12:31:27 -0800 | [diff] [blame^] | 731 | ds[TYPE_B]= i==TYPE_B ? NULL : udata_openSwapper(true, U_ASCII_FAMILY, outIsBigEndian, outCharset, &errorCode); |
| 732 | ds[TYPE_L]= i==TYPE_L ? NULL : udata_openSwapper(false, U_ASCII_FAMILY, outIsBigEndian, outCharset, &errorCode); |
jshin@chromium.org | 6f31ac3 | 2014-03-26 22:15:14 +0000 | [diff] [blame] | 733 | ds[TYPE_LE]=NULL; |
Frank Tang | 1f164ee | 2022-11-08 12:31:27 -0800 | [diff] [blame^] | 734 | ds[TYPE_E]= i==TYPE_E ? NULL : udata_openSwapper(true, U_EBCDIC_FAMILY, outIsBigEndian, outCharset, &errorCode); |
jshin@chromium.org | 6f31ac3 | 2014-03-26 22:15:14 +0000 | [diff] [blame] | 735 | if(U_FAILURE(errorCode)) { |
| 736 | fprintf(stderr, "icupkg: udata_openSwapper() failed - %s\n", u_errorName(errorCode)); |
| 737 | exit(errorCode); |
| 738 | } |
| 739 | for(i=0; i<TYPE_COUNT; ++i) { |
| 740 | if(ds[i]!=NULL) { |
| 741 | ds[i]->printError=printPackageError; |
| 742 | ds[i]->printErrorContext=stderr; |
| 743 | } |
| 744 | } |
| 745 | |
| 746 | dsLocalToOut=ds[makeTypeEnum(U_CHARSET_FAMILY, U_IS_BIG_ENDIAN)]; |
| 747 | |
| 748 | // create the file and write its contents |
| 749 | file=fopen(filename, "wb"); |
| 750 | if(file==NULL) { |
| 751 | fprintf(stderr, "icupkg: unable to create file \"%s\"\n", filename); |
| 752 | exit(U_FILE_ACCESS_ERROR); |
| 753 | } |
| 754 | |
| 755 | // swap and write the header |
| 756 | if(dsLocalToOut!=NULL) { |
| 757 | udata_swapDataHeader(dsLocalToOut, header, headerLength, header, &errorCode); |
| 758 | if(U_FAILURE(errorCode)) { |
| 759 | fprintf(stderr, "icupkg: udata_swapDataHeader(local to out) failed - %s\n", u_errorName(errorCode)); |
| 760 | exit(errorCode); |
| 761 | } |
| 762 | } |
| 763 | length=(int32_t)fwrite(header, 1, headerLength, file); |
| 764 | if(length!=headerLength) { |
| 765 | fprintf(stderr, "icupkg: unable to write complete header to file \"%s\"\n", filename); |
| 766 | exit(U_FILE_ACCESS_ERROR); |
| 767 | } |
| 768 | |
| 769 | // prepare and swap the package name with a tree separator |
| 770 | // for prepending to item names |
| 771 | if(pkgPrefix[0]==0) { |
| 772 | prefixLength=(int32_t)strlen(prefix); |
| 773 | } else { |
| 774 | prefixLength=(int32_t)strlen(pkgPrefix); |
| 775 | memcpy(prefix, pkgPrefix, prefixLength); |
| 776 | if(prefixEndsWithType) { |
| 777 | prefix[prefixLength-1]=outType; |
| 778 | } |
| 779 | } |
| 780 | prefix[prefixLength++]=U_TREE_ENTRY_SEP_CHAR; |
| 781 | prefix[prefixLength]=0; |
| 782 | if(dsLocalToOut!=NULL) { |
| 783 | dsLocalToOut->swapInvChars(dsLocalToOut, prefix, prefixLength, prefix, &errorCode); |
| 784 | if(U_FAILURE(errorCode)) { |
| 785 | fprintf(stderr, "icupkg: swapInvChars(output package name) failed - %s\n", u_errorName(errorCode)); |
| 786 | exit(errorCode); |
| 787 | } |
| 788 | |
| 789 | // swap and sort the item names (sorting needs to be done in the output charset) |
| 790 | dsLocalToOut->swapInvChars(dsLocalToOut, inStrings, inStringTop, inStrings, &errorCode); |
| 791 | if(U_FAILURE(errorCode)) { |
| 792 | fprintf(stderr, "icupkg: swapInvChars(item names) failed - %s\n", u_errorName(errorCode)); |
| 793 | exit(errorCode); |
| 794 | } |
| 795 | sortItems(); |
| 796 | } |
| 797 | |
| 798 | // create the output item names in sorted order, with the package name prepended to each |
| 799 | for(i=0; i<itemCount; ++i) { |
| 800 | length=(int32_t)strlen(items[i].name); |
Frank Tang | 1f164ee | 2022-11-08 12:31:27 -0800 | [diff] [blame^] | 801 | name=allocString(false, length+prefixLength); |
jshin@chromium.org | 6f31ac3 | 2014-03-26 22:15:14 +0000 | [diff] [blame] | 802 | memcpy(name, prefix, prefixLength); |
| 803 | memcpy(name+prefixLength, items[i].name, length+1); |
| 804 | items[i].name=name; |
| 805 | } |
| 806 | |
| 807 | // calculate offsets for item names and items, pad to 16-align items |
| 808 | // align only the first item; each item's length is a multiple of 16 |
| 809 | basenameOffset=4+8*itemCount; |
| 810 | offset=basenameOffset+outStringTop; |
| 811 | if((length=(offset&15))!=0) { |
| 812 | length=16-length; |
Frank Tang | 1f164ee | 2022-11-08 12:31:27 -0800 | [diff] [blame^] | 813 | memset(allocString(false, length-1), 0xaa, length); |
jshin@chromium.org | 6f31ac3 | 2014-03-26 22:15:14 +0000 | [diff] [blame] | 814 | offset+=length; |
| 815 | } |
| 816 | |
| 817 | // write the table of contents |
| 818 | // first the itemCount |
| 819 | outInt32=itemCount; |
| 820 | if(dsLocalToOut!=NULL) { |
| 821 | dsLocalToOut->swapArray32(dsLocalToOut, &outInt32, 4, &outInt32, &errorCode); |
| 822 | if(U_FAILURE(errorCode)) { |
| 823 | fprintf(stderr, "icupkg: swapArray32(item count) failed - %s\n", u_errorName(errorCode)); |
| 824 | exit(errorCode); |
| 825 | } |
| 826 | } |
| 827 | length=(int32_t)fwrite(&outInt32, 1, 4, file); |
| 828 | if(length!=4) { |
| 829 | fprintf(stderr, "icupkg: unable to write complete item count to file \"%s\"\n", filename); |
| 830 | exit(U_FILE_ACCESS_ERROR); |
| 831 | } |
| 832 | |
| 833 | // then write the item entries (and collect the maxItemLength) |
| 834 | maxItemLength=0; |
| 835 | for(i=0; i<itemCount; ++i) { |
| 836 | entry.nameOffset=(uint32_t)(basenameOffset+(items[i].name-outStrings)); |
| 837 | entry.dataOffset=(uint32_t)offset; |
| 838 | if(dsLocalToOut!=NULL) { |
| 839 | dsLocalToOut->swapArray32(dsLocalToOut, &entry, 8, &entry, &errorCode); |
| 840 | if(U_FAILURE(errorCode)) { |
| 841 | fprintf(stderr, "icupkg: swapArray32(item entry %ld) failed - %s\n", (long)i, u_errorName(errorCode)); |
| 842 | exit(errorCode); |
| 843 | } |
| 844 | } |
| 845 | length=(int32_t)fwrite(&entry, 1, 8, file); |
| 846 | if(length!=8) { |
| 847 | fprintf(stderr, "icupkg: unable to write complete item entry %ld to file \"%s\"\n", (long)i, filename); |
| 848 | exit(U_FILE_ACCESS_ERROR); |
| 849 | } |
| 850 | |
| 851 | length=items[i].length; |
| 852 | if(length>maxItemLength) { |
| 853 | maxItemLength=length; |
| 854 | } |
| 855 | offset+=length; |
| 856 | } |
| 857 | |
| 858 | // write the item names |
| 859 | length=(int32_t)fwrite(outStrings, 1, outStringTop, file); |
| 860 | if(length!=outStringTop) { |
| 861 | fprintf(stderr, "icupkg: unable to write complete item names to file \"%s\"\n", filename); |
| 862 | exit(U_FILE_ACCESS_ERROR); |
| 863 | } |
| 864 | |
| 865 | // write the items |
| 866 | for(pItem=items, i=0; i<itemCount; ++pItem, ++i) { |
| 867 | int32_t type=makeTypeEnum(pItem->type); |
| 868 | if(ds[type]!=NULL) { |
| 869 | // swap each item from its platform properties to the desired ones |
| 870 | udata_swap( |
| 871 | ds[type], |
| 872 | pItem->data, pItem->length, pItem->data, |
| 873 | &errorCode); |
| 874 | if(U_FAILURE(errorCode)) { |
| 875 | fprintf(stderr, "icupkg: udata_swap(item %ld) failed - %s\n", (long)i, u_errorName(errorCode)); |
| 876 | exit(errorCode); |
| 877 | } |
| 878 | } |
| 879 | length=(int32_t)fwrite(pItem->data, 1, pItem->length, file); |
| 880 | if(length!=pItem->length) { |
| 881 | fprintf(stderr, "icupkg: unable to write complete item %ld to file \"%s\"\n", (long)i, filename); |
| 882 | exit(U_FILE_ACCESS_ERROR); |
| 883 | } |
| 884 | } |
| 885 | |
| 886 | if(ferror(file)) { |
| 887 | fprintf(stderr, "icupkg: unable to write complete file \"%s\"\n", filename); |
| 888 | exit(U_FILE_ACCESS_ERROR); |
| 889 | } |
| 890 | |
| 891 | fclose(file); |
| 892 | for(i=0; i<TYPE_COUNT; ++i) { |
| 893 | udata_closeSwapper(ds[i]); |
| 894 | } |
| 895 | } |
| 896 | |
| 897 | int32_t |
| 898 | Package::findItem(const char *name, int32_t length) const { |
| 899 | int32_t i, start, limit; |
| 900 | int result; |
| 901 | |
| 902 | /* do a binary search for the string */ |
| 903 | start=0; |
| 904 | limit=itemCount; |
| 905 | while(start<limit) { |
| 906 | i=(start+limit)/2; |
| 907 | if(length>=0) { |
| 908 | result=strncmp(name, items[i].name, length); |
| 909 | } else { |
| 910 | result=strcmp(name, items[i].name); |
| 911 | } |
| 912 | |
| 913 | if(result==0) { |
| 914 | /* found */ |
| 915 | if(length>=0) { |
| 916 | /* |
| 917 | * if we compared just prefixes, then we may need to back up |
| 918 | * to the first item with this prefix |
| 919 | */ |
| 920 | while(i>0 && 0==strncmp(name, items[i-1].name, length)) { |
| 921 | --i; |
| 922 | } |
| 923 | } |
| 924 | return i; |
| 925 | } else if(result<0) { |
| 926 | limit=i; |
| 927 | } else /* result>0 */ { |
| 928 | start=i+1; |
| 929 | } |
| 930 | } |
| 931 | |
| 932 | return ~start; /* not found, return binary-not of the insertion point */ |
| 933 | } |
| 934 | |
| 935 | void |
| 936 | Package::findItems(const char *pattern) { |
| 937 | const char *wild; |
| 938 | |
| 939 | if(pattern==NULL || *pattern==0) { |
| 940 | findNextIndex=-1; |
| 941 | return; |
| 942 | } |
| 943 | |
| 944 | findPrefix=pattern; |
| 945 | findSuffix=NULL; |
| 946 | findSuffixLength=0; |
| 947 | |
| 948 | wild=strchr(pattern, '*'); |
| 949 | if(wild==NULL) { |
| 950 | // no wildcard |
| 951 | findPrefixLength=(int32_t)strlen(pattern); |
| 952 | } else { |
| 953 | // one wildcard |
| 954 | findPrefixLength=(int32_t)(wild-pattern); |
| 955 | findSuffix=wild+1; |
| 956 | findSuffixLength=(int32_t)strlen(findSuffix); |
| 957 | if(NULL!=strchr(findSuffix, '*')) { |
| 958 | // two or more wildcards |
| 959 | fprintf(stderr, "icupkg: syntax error (more than one '*') in item pattern \"%s\"\n", pattern); |
| 960 | exit(U_PARSE_ERROR); |
| 961 | } |
| 962 | } |
| 963 | |
| 964 | if(findPrefixLength==0) { |
| 965 | findNextIndex=0; |
| 966 | } else { |
| 967 | findNextIndex=findItem(findPrefix, findPrefixLength); |
| 968 | } |
| 969 | } |
| 970 | |
| 971 | int32_t |
| 972 | Package::findNextItem() { |
| 973 | const char *name, *middle, *treeSep; |
| 974 | int32_t idx, nameLength, middleLength; |
| 975 | |
| 976 | if(findNextIndex<0) { |
| 977 | return -1; |
| 978 | } |
| 979 | |
| 980 | while(findNextIndex<itemCount) { |
| 981 | idx=findNextIndex++; |
| 982 | name=items[idx].name; |
| 983 | nameLength=(int32_t)strlen(name); |
| 984 | if(nameLength<(findPrefixLength+findSuffixLength)) { |
| 985 | // item name too short for prefix & suffix |
| 986 | continue; |
| 987 | } |
| 988 | if(findPrefixLength>0 && 0!=memcmp(findPrefix, name, findPrefixLength)) { |
| 989 | // left the range of names with this prefix |
| 990 | break; |
| 991 | } |
| 992 | middle=name+findPrefixLength; |
| 993 | middleLength=nameLength-findPrefixLength-findSuffixLength; |
| 994 | if(findSuffixLength>0 && 0!=memcmp(findSuffix, name+(nameLength-findSuffixLength), findSuffixLength)) { |
| 995 | // suffix does not match |
| 996 | continue; |
| 997 | } |
| 998 | // prefix & suffix match |
| 999 | |
| 1000 | if(matchMode&MATCH_NOSLASH) { |
| 1001 | treeSep=strchr(middle, U_TREE_ENTRY_SEP_CHAR); |
| 1002 | if(treeSep!=NULL && (treeSep-middle)<middleLength) { |
| 1003 | // the middle (matching the * wildcard) contains a tree separator / |
| 1004 | continue; |
| 1005 | } |
| 1006 | } |
| 1007 | |
| 1008 | // found a matching item |
| 1009 | return idx; |
| 1010 | } |
| 1011 | |
| 1012 | // no more items |
| 1013 | findNextIndex=-1; |
| 1014 | return -1; |
| 1015 | } |
| 1016 | |
| 1017 | void |
| 1018 | Package::setMatchMode(uint32_t mode) { |
| 1019 | matchMode=mode; |
| 1020 | } |
| 1021 | |
| 1022 | void |
| 1023 | Package::addItem(const char *name) { |
Frank Tang | 1f164ee | 2022-11-08 12:31:27 -0800 | [diff] [blame^] | 1024 | addItem(name, NULL, 0, false, U_ICUDATA_TYPE_LETTER[0]); |
jshin@chromium.org | 6f31ac3 | 2014-03-26 22:15:14 +0000 | [diff] [blame] | 1025 | } |
| 1026 | |
| 1027 | void |
| 1028 | Package::addItem(const char *name, uint8_t *data, int32_t length, UBool isDataOwned, char type) { |
| 1029 | int32_t idx; |
| 1030 | |
| 1031 | idx=findItem(name); |
| 1032 | if(idx<0) { |
| 1033 | // new item, make space at the insertion point |
| 1034 | ensureItemCapacity(); |
| 1035 | // move the following items down |
| 1036 | idx=~idx; |
| 1037 | if(idx<itemCount) { |
| 1038 | memmove(items+idx+1, items+idx, (itemCount-idx)*sizeof(Item)); |
| 1039 | } |
| 1040 | ++itemCount; |
| 1041 | |
| 1042 | // reset this Item entry |
| 1043 | memset(items+idx, 0, sizeof(Item)); |
| 1044 | |
| 1045 | // copy the item's name |
Frank Tang | 1f164ee | 2022-11-08 12:31:27 -0800 | [diff] [blame^] | 1046 | items[idx].name=allocString(true, static_cast<int32_t>(strlen(name))); |
jshin@chromium.org | 6f31ac3 | 2014-03-26 22:15:14 +0000 | [diff] [blame] | 1047 | strcpy(items[idx].name, name); |
| 1048 | pathToTree(items[idx].name); |
| 1049 | } else { |
| 1050 | // same-name item found, replace it |
| 1051 | if(items[idx].isDataOwned) { |
Jungshik Shin | 70f8250 | 2016-01-29 00:32:36 -0800 | [diff] [blame] | 1052 | uprv_free(items[idx].data); |
jshin@chromium.org | 6f31ac3 | 2014-03-26 22:15:14 +0000 | [diff] [blame] | 1053 | } |
| 1054 | |
| 1055 | // keep the item's name since it is the same |
| 1056 | } |
| 1057 | |
| 1058 | // set the item's data |
| 1059 | items[idx].data=data; |
| 1060 | items[idx].length=length; |
| 1061 | items[idx].isDataOwned=isDataOwned; |
| 1062 | items[idx].type=type; |
| 1063 | } |
| 1064 | |
| 1065 | void |
| 1066 | Package::addFile(const char *filesPath, const char *name) { |
| 1067 | uint8_t *data; |
| 1068 | int32_t length; |
| 1069 | char type; |
| 1070 | |
| 1071 | data=readFile(filesPath, name, length, type); |
| 1072 | // readFile() exits the tool if it fails |
Frank Tang | 1f164ee | 2022-11-08 12:31:27 -0800 | [diff] [blame^] | 1073 | addItem(name, data, length, true, type); |
jshin@chromium.org | 6f31ac3 | 2014-03-26 22:15:14 +0000 | [diff] [blame] | 1074 | } |
| 1075 | |
| 1076 | void |
| 1077 | Package::addItems(const Package &listPkg) { |
| 1078 | const Item *pItem; |
| 1079 | int32_t i; |
| 1080 | |
| 1081 | for(pItem=listPkg.items, i=0; i<listPkg.itemCount; ++pItem, ++i) { |
Frank Tang | 1f164ee | 2022-11-08 12:31:27 -0800 | [diff] [blame^] | 1082 | addItem(pItem->name, pItem->data, pItem->length, false, pItem->type); |
jshin@chromium.org | 6f31ac3 | 2014-03-26 22:15:14 +0000 | [diff] [blame] | 1083 | } |
| 1084 | } |
| 1085 | |
| 1086 | void |
| 1087 | Package::removeItem(int32_t idx) { |
| 1088 | if(idx>=0) { |
| 1089 | // remove the item |
| 1090 | if(items[idx].isDataOwned) { |
Jungshik Shin | 70f8250 | 2016-01-29 00:32:36 -0800 | [diff] [blame] | 1091 | uprv_free(items[idx].data); |
jshin@chromium.org | 6f31ac3 | 2014-03-26 22:15:14 +0000 | [diff] [blame] | 1092 | } |
| 1093 | |
| 1094 | // move the following items up |
| 1095 | if((idx+1)<itemCount) { |
| 1096 | memmove(items+idx, items+idx+1, (itemCount-(idx+1))*sizeof(Item)); |
| 1097 | } |
| 1098 | --itemCount; |
| 1099 | |
| 1100 | if(idx<=findNextIndex) { |
| 1101 | --findNextIndex; |
| 1102 | } |
| 1103 | } |
| 1104 | } |
| 1105 | |
| 1106 | void |
| 1107 | Package::removeItems(const char *pattern) { |
| 1108 | int32_t idx; |
| 1109 | |
| 1110 | findItems(pattern); |
| 1111 | while((idx=findNextItem())>=0) { |
| 1112 | removeItem(idx); |
| 1113 | } |
| 1114 | } |
| 1115 | |
| 1116 | void |
| 1117 | Package::removeItems(const Package &listPkg) { |
| 1118 | const Item *pItem; |
| 1119 | int32_t i; |
| 1120 | |
| 1121 | for(pItem=listPkg.items, i=0; i<listPkg.itemCount; ++pItem, ++i) { |
| 1122 | removeItems(pItem->name); |
| 1123 | } |
| 1124 | } |
| 1125 | |
| 1126 | void |
| 1127 | Package::extractItem(const char *filesPath, const char *outName, int32_t idx, char outType) { |
| 1128 | char filename[1024]; |
| 1129 | UDataSwapper *ds; |
| 1130 | FILE *file; |
| 1131 | Item *pItem; |
| 1132 | int32_t fileLength; |
| 1133 | uint8_t itemCharset, outCharset; |
| 1134 | UBool itemIsBigEndian, outIsBigEndian; |
| 1135 | |
| 1136 | if(idx<0 || itemCount<=idx) { |
| 1137 | return; |
| 1138 | } |
| 1139 | pItem=items+idx; |
| 1140 | |
| 1141 | // swap the data to the outType |
| 1142 | // outType==0: don't swap |
| 1143 | if(outType!=0 && pItem->type!=outType) { |
| 1144 | // open the swapper |
| 1145 | UErrorCode errorCode=U_ZERO_ERROR; |
| 1146 | makeTypeProps(pItem->type, itemCharset, itemIsBigEndian); |
| 1147 | makeTypeProps(outType, outCharset, outIsBigEndian); |
| 1148 | ds=udata_openSwapper(itemIsBigEndian, itemCharset, outIsBigEndian, outCharset, &errorCode); |
| 1149 | if(U_FAILURE(errorCode)) { |
| 1150 | fprintf(stderr, "icupkg: udata_openSwapper(item %ld) failed - %s\n", |
| 1151 | (long)idx, u_errorName(errorCode)); |
| 1152 | exit(errorCode); |
| 1153 | } |
| 1154 | |
| 1155 | ds->printError=printPackageError; |
| 1156 | ds->printErrorContext=stderr; |
| 1157 | |
| 1158 | // swap the item from its platform properties to the desired ones |
| 1159 | udata_swap(ds, pItem->data, pItem->length, pItem->data, &errorCode); |
| 1160 | if(U_FAILURE(errorCode)) { |
| 1161 | fprintf(stderr, "icupkg: udata_swap(item %ld) failed - %s\n", (long)idx, u_errorName(errorCode)); |
| 1162 | exit(errorCode); |
| 1163 | } |
| 1164 | udata_closeSwapper(ds); |
| 1165 | pItem->type=outType; |
| 1166 | } |
| 1167 | |
| 1168 | // create the file and write its contents |
| 1169 | makeFullFilenameAndDirs(filesPath, outName, filename, (int32_t)sizeof(filename)); |
| 1170 | file=fopen(filename, "wb"); |
| 1171 | if(file==NULL) { |
| 1172 | fprintf(stderr, "icupkg: unable to create file \"%s\"\n", filename); |
| 1173 | exit(U_FILE_ACCESS_ERROR); |
| 1174 | } |
| 1175 | fileLength=(int32_t)fwrite(pItem->data, 1, pItem->length, file); |
| 1176 | |
| 1177 | if(ferror(file) || fileLength!=pItem->length) { |
| 1178 | fprintf(stderr, "icupkg: unable to write complete file \"%s\"\n", filename); |
| 1179 | exit(U_FILE_ACCESS_ERROR); |
| 1180 | } |
| 1181 | fclose(file); |
| 1182 | } |
| 1183 | |
| 1184 | void |
| 1185 | Package::extractItem(const char *filesPath, int32_t idx, char outType) { |
| 1186 | extractItem(filesPath, items[idx].name, idx, outType); |
| 1187 | } |
| 1188 | |
| 1189 | void |
| 1190 | Package::extractItems(const char *filesPath, const char *pattern, char outType) { |
| 1191 | int32_t idx; |
| 1192 | |
| 1193 | findItems(pattern); |
| 1194 | while((idx=findNextItem())>=0) { |
| 1195 | extractItem(filesPath, idx, outType); |
| 1196 | } |
| 1197 | } |
| 1198 | |
| 1199 | void |
| 1200 | Package::extractItems(const char *filesPath, const Package &listPkg, char outType) { |
| 1201 | const Item *pItem; |
| 1202 | int32_t i; |
| 1203 | |
| 1204 | for(pItem=listPkg.items, i=0; i<listPkg.itemCount; ++pItem, ++i) { |
| 1205 | extractItems(filesPath, pItem->name, outType); |
| 1206 | } |
| 1207 | } |
| 1208 | |
| 1209 | int32_t |
| 1210 | Package::getItemCount() const { |
| 1211 | return itemCount; |
| 1212 | } |
| 1213 | |
| 1214 | const Item * |
| 1215 | Package::getItem(int32_t idx) const { |
| 1216 | if (0 <= idx && idx < itemCount) { |
| 1217 | return &items[idx]; |
| 1218 | } |
| 1219 | return NULL; |
| 1220 | } |
| 1221 | |
| 1222 | void |
| 1223 | Package::checkDependency(void *context, const char *itemName, const char *targetName) { |
| 1224 | // check dependency: make sure the target item is in the package |
| 1225 | Package *me=(Package *)context; |
| 1226 | if(me->findItem(targetName)<0) { |
Frank Tang | 1f164ee | 2022-11-08 12:31:27 -0800 | [diff] [blame^] | 1227 | me->isMissingItems=true; |
jshin@chromium.org | 6f31ac3 | 2014-03-26 22:15:14 +0000 | [diff] [blame] | 1228 | fprintf(stderr, "Item %s depends on missing item %s\n", itemName, targetName); |
| 1229 | } |
| 1230 | } |
| 1231 | |
| 1232 | UBool |
| 1233 | Package::checkDependencies() { |
Frank Tang | 1f164ee | 2022-11-08 12:31:27 -0800 | [diff] [blame^] | 1234 | isMissingItems=false; |
jshin@chromium.org | 6f31ac3 | 2014-03-26 22:15:14 +0000 | [diff] [blame] | 1235 | enumDependencies(this, checkDependency); |
| 1236 | return (UBool)!isMissingItems; |
| 1237 | } |
| 1238 | |
| 1239 | void |
| 1240 | Package::enumDependencies(void *context, CheckDependency check) { |
| 1241 | int32_t i; |
| 1242 | |
| 1243 | for(i=0; i<itemCount; ++i) { |
| 1244 | enumDependencies(items+i, context, check); |
| 1245 | } |
| 1246 | } |
| 1247 | |
| 1248 | char * |
| 1249 | Package::allocString(UBool in, int32_t length) { |
| 1250 | char *p; |
| 1251 | int32_t top; |
| 1252 | |
| 1253 | if(in) { |
| 1254 | top=inStringTop; |
| 1255 | p=inStrings+top; |
| 1256 | } else { |
| 1257 | top=outStringTop; |
| 1258 | p=outStrings+top; |
| 1259 | } |
| 1260 | top+=length+1; |
| 1261 | |
| 1262 | if(top>STRING_STORE_SIZE) { |
| 1263 | fprintf(stderr, "icupkg: string storage overflow\n"); |
| 1264 | exit(U_BUFFER_OVERFLOW_ERROR); |
| 1265 | } |
| 1266 | if(in) { |
| 1267 | inStringTop=top; |
| 1268 | } else { |
| 1269 | outStringTop=top; |
| 1270 | } |
| 1271 | return p; |
| 1272 | } |
| 1273 | |
| 1274 | void |
| 1275 | Package::sortItems() { |
| 1276 | UErrorCode errorCode=U_ZERO_ERROR; |
Frank Tang | 1f164ee | 2022-11-08 12:31:27 -0800 | [diff] [blame^] | 1277 | uprv_sortArray(items, itemCount, (int32_t)sizeof(Item), compareItems, NULL, false, &errorCode); |
jshin@chromium.org | 6f31ac3 | 2014-03-26 22:15:14 +0000 | [diff] [blame] | 1278 | if(U_FAILURE(errorCode)) { |
| 1279 | fprintf(stderr, "icupkg: sorting item names failed - %s\n", u_errorName(errorCode)); |
| 1280 | exit(errorCode); |
| 1281 | } |
| 1282 | } |
| 1283 | |
Jungshik Shin | 87232d8 | 2017-05-13 21:10:13 -0700 | [diff] [blame] | 1284 | void Package::setItemCapacity(int32_t max) |
jshin@chromium.org | 6f31ac3 | 2014-03-26 22:15:14 +0000 | [diff] [blame] | 1285 | { |
| 1286 | if(max<=itemMax) { |
| 1287 | return; |
| 1288 | } |
| 1289 | Item *newItems = (Item*)uprv_malloc(max * sizeof(items[0])); |
| 1290 | Item *oldItems = items; |
| 1291 | if(newItems == NULL) { |
Jungshik Shin | 87232d8 | 2017-05-13 21:10:13 -0700 | [diff] [blame] | 1292 | fprintf(stderr, "icupkg: Out of memory trying to allocate %lu bytes for %d items\n", |
| 1293 | (unsigned long)(max*sizeof(items[0])), max); |
jshin@chromium.org | 6f31ac3 | 2014-03-26 22:15:14 +0000 | [diff] [blame] | 1294 | exit(U_MEMORY_ALLOCATION_ERROR); |
| 1295 | } |
| 1296 | if(items && itemCount>0) { |
Jungshik Shin | 5feb9ad | 2016-10-21 12:52:48 -0700 | [diff] [blame] | 1297 | uprv_memcpy(newItems, items, (size_t)itemCount*sizeof(items[0])); |
jshin@chromium.org | 6f31ac3 | 2014-03-26 22:15:14 +0000 | [diff] [blame] | 1298 | } |
| 1299 | itemMax = max; |
| 1300 | items = newItems; |
| 1301 | uprv_free(oldItems); |
| 1302 | } |
| 1303 | |
| 1304 | void Package::ensureItemCapacity() |
| 1305 | { |
| 1306 | if((itemCount+1)>itemMax) { |
| 1307 | setItemCapacity(itemCount+kItemsChunk); |
| 1308 | } |
| 1309 | } |
| 1310 | |
| 1311 | U_NAMESPACE_END |