1 // Copyright (C) 2016 and later: Unicode, Inc. and others. 2 // License & terms of use: http://www.unicode.org/copyright.html 3 /* 4 ****************************************************************************** 5 * Copyright (C) 1999-2016, International Business Machines 6 * Corporation and others. All Rights Reserved. 7 ****************************************************************************** 8 * file name: uresdata.h 9 * encoding: US-ASCII 10 * tab size: 8 (not used) 11 * indentation:4 12 * 13 * created on: 1999dec08 14 * created by: Markus W. Scherer 15 * 06/24/02 weiv Added support for resource sharing 16 */ 17 18 #ifndef __RESDATA_H__ 19 #define __RESDATA_H__ 20 21 #include "unicode/utypes.h" 22 #include "unicode/udata.h" 23 #include "unicode/ures.h" 24 #include "putilimp.h" 25 #include "udataswp.h" 26 27 /** 28 * Numeric constants for internal-only types of resource items. 29 * These must use different numeric values than UResType constants 30 * because they are used together. 31 * Internal types are never returned by ures_getType(). 32 */ 33 typedef enum { 34 /** Include a negative value so that the compiler uses the same int type as for UResType. */ 35 URES_INTERNAL_NONE=-1, 36 37 /** Resource type constant for tables with 32-bit count, key offsets and values. */ 38 URES_TABLE32=4, 39 40 /** 41 * Resource type constant for tables with 16-bit count, key offsets and values. 42 * All values are URES_STRING_V2 strings. 43 */ 44 URES_TABLE16=5, 45 46 /** Resource type constant for 16-bit Unicode strings in formatVersion 2. */ 47 URES_STRING_V2=6, 48 49 /** 50 * Resource type constant for arrays with 16-bit count and values. 51 * All values are URES_STRING_V2 strings. 52 */ 53 URES_ARRAY16=9 54 55 /* Resource type 15 is not defined but effectively used by RES_BOGUS=0xffffffff. */ 56 } UResInternalType; 57 58 /* 59 * A Resource is a 32-bit value that has 2 bit fields: 60 * 31..28 4-bit type, see enum below 61 * 27..0 28-bit four-byte-offset or value according to the type 62 */ 63 typedef uint32_t Resource; 64 65 #define RES_BOGUS 0xffffffff 66 #define RES_MAX_OFFSET 0x0fffffff 67 68 #define RES_GET_TYPE(res) ((int32_t)((res)>>28UL)) 69 #define RES_GET_OFFSET(res) ((res)&0x0fffffff) 70 #define RES_GET_POINTER(pRoot, res) ((pRoot)+RES_GET_OFFSET(res)) 71 72 /* get signed and unsigned integer values directly from the Resource handle */ 73 #if U_SIGNED_RIGHT_SHIFT_IS_ARITHMETIC 74 # define RES_GET_INT(res) (((int32_t)((res)<<4L))>>4L) 75 #else 76 # define RES_GET_INT(res) (int32_t)(((res)&0x08000000) ? (res)|0xf0000000 : (res)&0x07ffffff) 77 #endif 78 79 #define RES_GET_UINT(res) ((res)&0x0fffffff) 80 81 #define URES_IS_ARRAY(type) ((int32_t)(type)==URES_ARRAY || (int32_t)(type)==URES_ARRAY16) 82 #define URES_IS_TABLE(type) ((int32_t)(type)==URES_TABLE || (int32_t)(type)==URES_TABLE16 || (int32_t)(type)==URES_TABLE32) 83 #define URES_IS_CONTAINER(type) (URES_IS_TABLE(type) || URES_IS_ARRAY(type)) 84 85 #define URES_MAKE_RESOURCE(type, offset) (((Resource)(type)<<28)|(Resource)(offset)) 86 #define URES_MAKE_EMPTY_RESOURCE(type) ((Resource)(type)<<28) 87 88 /* indexes[] value names; indexes are generally 32-bit (Resource) indexes */ 89 enum { 90 /** 91 * [0] contains the length of indexes[] 92 * which is at most URES_INDEX_TOP of the latest format version 93 * 94 * formatVersion==1: all bits contain the length of indexes[] 95 * but the length is much less than 0xff; 96 * formatVersion>1: 97 * only bits 7..0 contain the length of indexes[], 98 * bits 31..8 are reserved and set to 0 99 * formatVersion>=3: 100 * bits 31..8 poolStringIndexLimit bits 23..0 101 */ 102 URES_INDEX_LENGTH, 103 /** 104 * [1] contains the top of the key strings, 105 * same as the bottom of resources or UTF-16 strings, rounded up 106 */ 107 URES_INDEX_KEYS_TOP, 108 /** [2] contains the top of all resources */ 109 URES_INDEX_RESOURCES_TOP, 110 /** 111 * [3] contains the top of the bundle, 112 * in case it were ever different from [2] 113 */ 114 URES_INDEX_BUNDLE_TOP, 115 /** [4] max. length of any table */ 116 URES_INDEX_MAX_TABLE_LENGTH, 117 /** 118 * [5] attributes bit set, see URES_ATT_* (new in formatVersion 1.2) 119 * 120 * formatVersion>=3: 121 * bits 31..16 poolStringIndex16Limit 122 * bits 15..12 poolStringIndexLimit bits 27..24 123 */ 124 URES_INDEX_ATTRIBUTES, 125 /** 126 * [6] top of the 16-bit units (UTF-16 string v2 UChars, URES_TABLE16, URES_ARRAY16), 127 * rounded up (new in formatVersion 2.0, ICU 4.4) 128 */ 129 URES_INDEX_16BIT_TOP, 130 /** [7] checksum of the pool bundle (new in formatVersion 2.0, ICU 4.4) */ 131 URES_INDEX_POOL_CHECKSUM, 132 URES_INDEX_TOP 133 }; 134 135 /* 136 * Nofallback attribute, attribute bit 0 in indexes[URES_INDEX_ATTRIBUTES]. 137 * New in formatVersion 1.2 (ICU 3.6). 138 * 139 * If set, then this resource bundle is a standalone bundle. 140 * If not set, then the bundle participates in locale fallback, eventually 141 * all the way to the root bundle. 142 * If indexes[] is missing or too short, then the attribute cannot be determined 143 * reliably. Dependency checking should ignore such bundles, and loading should 144 * use fallbacks. 145 */ 146 #define URES_ATT_NO_FALLBACK 1 147 148 /* 149 * Attributes for bundles that are, or use, a pool bundle. 150 * A pool bundle provides key strings that are shared among several other bundles 151 * to reduce their total size. 152 * New in formatVersion 2 (ICU 4.4). 153 */ 154 #define URES_ATT_IS_POOL_BUNDLE 2 155 #define URES_ATT_USES_POOL_BUNDLE 4 156 157 /* 158 * File format for .res resource bundle files 159 * 160 * ICU 56: New in formatVersion 3 compared with 2: ------------- 161 * 162 * Resource bundles can optionally use shared string-v2 values 163 * stored in the pool bundle. 164 * If so, then the indexes[] contain two new values 165 * in previously-unused bits of existing indexes[] slots: 166 * - poolStringIndexLimit: 167 * String-v2 offsets (in 32-bit Resource words) below this limit 168 * point to pool bundle string-v2 values. 169 * - poolStringIndex16Limit: 170 * Resource16 string-v2 offsets below this limit 171 * point to pool bundle string-v2 values. 172 * Guarantee: poolStringIndex16Limit <= poolStringIndexLimit 173 * 174 * The local bundle's poolStringIndexLimit is greater than 175 * any pool bundle string index used in the local bundle. 176 * The poolStringIndexLimit should not be greater than 177 * the maximum possible pool bundle string index. 178 * 179 * The maximum possible pool bundle string index is the index to the last non-NUL 180 * pool string character, due to suffix sharing. 181 * 182 * In the pool bundle, there is no structure that lists the strings. 183 * (The root resource is an empty Table.) 184 * If the strings need to be enumerated (as genrb --usePoolBundle does), 185 * then iterate through the pool bundle's 16-bit-units array from the beginning. 186 * Stop at the end of the array, or when an explicit or implicit string length 187 * would lead beyond the end of the array, 188 * or when an apparent string is not NUL-terminated. 189 * (Future genrb version might terminate the strings with 190 * what looks like a large explicit string length.) 191 * 192 * ICU 4.4: New in formatVersion 2 compared with 1.3: ------------- 193 * 194 * Three new resource types -- String-v2, Table16 and Array16 -- have their 195 * values stored in a new array of 16-bit units between the table key strings 196 * and the start of the other resources. 197 * 198 * genrb eliminates duplicates among Unicode string-v2 values. 199 * Multiple Unicode strings may use the same offset and string data, 200 * or a short string may point to the suffix of a longer string. ("Suffix sharing") 201 * For example, one string "abc" may be reused for another string "bc" by pointing 202 * to the second character. (Short strings-v2 are NUL-terminated 203 * and not preceded by an explicit length value.) 204 * 205 * It is allowed for all resource types to share values. 206 * The swapper code (ures_swap()) has been modified so that it swaps each item 207 * exactly once. 208 * 209 * A resource bundle may use a special pool bundle. Some or all of the table key strings 210 * of the using-bundle are omitted, and the key string offsets for such key strings refer 211 * to offsets in the pool bundle. 212 * The using-bundle's and the pool-bundle's indexes[URES_INDEX_POOL_CHECKSUM] values 213 * must match. 214 * Two bits in indexes[URES_INDEX_ATTRIBUTES] indicate whether a resource bundle 215 * is or uses a pool bundle. 216 * 217 * Table key strings must be compared in ASCII order, even if they are not 218 * stored in ASCII. 219 * 220 * New in formatVersion 1.3 compared with 1.2: ------------- 221 * 222 * genrb eliminates duplicates among key strings. 223 * Multiple table items may share one key string, or one item may point 224 * to the suffix of another's key string. ("Suffix sharing") 225 * For example, one key "abc" may be reused for another key "bc" by pointing 226 * to the second character. (Key strings are NUL-terminated.) 227 * 228 * ------------- 229 * 230 * An ICU4C resource bundle file (.res) is a binary, memory-mappable file 231 * with nested, hierarchical data structures. 232 * It physically contains the following: 233 * 234 * Resource root; -- 32-bit Resource item, root item for this bundle's tree; 235 * currently, the root item must be a table or table32 resource item 236 * int32_t indexes[indexes[0]]; -- array of indexes for friendly 237 * reading and swapping; see URES_INDEX_* above 238 * new in formatVersion 1.1 (ICU 2.8) 239 * char keys[]; -- characters for key strings 240 * (formatVersion 1.0: up to 65k of characters; 1.1: <2G) 241 * (minus the space for root and indexes[]), 242 * which consist of invariant characters (ASCII/EBCDIC) and are NUL-terminated; 243 * padded to multiple of 4 bytes for 4-alignment of the following data 244 * uint16_t 16BitUnits[]; -- resources that are stored entirely as sequences of 16-bit units 245 * (new in formatVersion 2/ICU 4.4) 246 * data is indexed by the offset values in 16-bit resource types, 247 * with offset 0 pointing to the beginning of this array; 248 * there is a 0 at offset 0, for empty resources; 249 * padded to multiple of 4 bytes for 4-alignment of the following data 250 * data; -- data directly and indirectly indexed by the root item; 251 * the structure is determined by walking the tree 252 * 253 * Each resource bundle item has a 32-bit Resource handle (see typedef above) 254 * which contains the item type number in its upper 4 bits (31..28) and either 255 * an offset or a direct value in its lower 28 bits (27..0). 256 * The order of items is undefined and only determined by walking the tree. 257 * Leaves of the tree may be stored first or last or anywhere in between, 258 * and it is in theory possible to have unreferenced holes in the file. 259 * 260 * 16-bit-unit values: 261 * Starting with formatVersion 2/ICU 4.4, some resources are stored in a special 262 * array of 16-bit units. Each resource value is a sequence of 16-bit units, 263 * with no per-resource padding to a 4-byte boundary. 264 * 16-bit container types (Table16 and Array16) contain Resource16 values 265 * which are offsets to String-v2 resources in the same 16-bit-units array. 266 * 267 * Direct values: 268 * - Empty Unicode strings have an offset value of 0 in the Resource handle itself. 269 * - Starting with formatVersion 2/ICU 4.4, an offset value of 0 for 270 * _any_ resource type indicates an empty value. 271 * - Integer values are 28-bit values stored in the Resource handle itself; 272 * the interpretation of unsigned vs. signed integers is up to the application. 273 * 274 * All other types and values use 28-bit offsets to point to the item's data. 275 * The offset is an index to the first 32-bit word of the value, relative to the 276 * start of the resource data (i.e., the root item handle is at offset 0). 277 * To get byte offsets, the offset is multiplied by 4 (or shifted left by 2 bits). 278 * All resource item values are 4-aligned. 279 * 280 * New in formatVersion 2/ICU 4.4: Some types use offsets into the 16-bit-units array, 281 * indexing 16-bit units in that array. 282 * 283 * The structures (memory layouts) for the values for each item type are listed 284 * in the table below. 285 * 286 * Nested, hierarchical structures: ------------- 287 * 288 * Table items contain key-value pairs where the keys are offsets to char * key strings. 289 * The values of these pairs are either Resource handles or 290 * offsets into the 16-bit-units array, depending on the table type. 291 * 292 * Array items are simple vectors of Resource handles, 293 * or of offsets into the 16-bit-units array, depending on the array type. 294 * 295 * Table key string offsets: ------- 296 * 297 * Key string offsets are relative to the start of the resource data (of the root handle), 298 * i.e., the first string has an offset of 4+sizeof(indexes). 299 * (After the 4-byte root handle and after the indexes array.) 300 * 301 * If the resource bundle uses a pool bundle, then some key strings are stored 302 * in the pool bundle rather than in the local bundle itself. 303 * - In a Table or Table16, the 16-bit key string offset is local if it is 304 * less than indexes[URES_INDEX_KEYS_TOP]<<2. 305 * Otherwise, subtract indexes[URES_INDEX_KEYS_TOP]<<2 to get the offset into 306 * the pool bundle key strings. 307 * - In a Table32, the 32-bit key string offset is local if it is non-negative. 308 * Otherwise, reset bit 31 to get the pool key string offset. 309 * 310 * Unlike the local offset, the pool key offset is relative to 311 * the start of the key strings, not to the start of the bundle. 312 * 313 * An alias item is special (and new in ICU 2.4): -------------- 314 * 315 * Its memory layout is just like for a UnicodeString, but at runtime it resolves to 316 * another resource bundle's item according to the path in the string. 317 * This is used to share items across bundles that are in different lookup/fallback 318 * chains (e.g., large collation data among zh_TW and zh_HK). 319 * This saves space (for large items) and maintenance effort (less duplication of data). 320 * 321 * -------------------------------------------------------------------------- 322 * 323 * Resource types: 324 * 325 * Most resources have their values stored at four-byte offsets from the start 326 * of the resource data. These values are at least 4-aligned. 327 * Some resource values are stored directly in the offset field of the Resource itself. 328 * See UResType in unicode/ures.h for enumeration constants for Resource types. 329 * 330 * Some resources have their values stored as sequences of 16-bit units, 331 * at 2-byte offsets from the start of a contiguous 16-bit-unit array between 332 * the table key strings and the other resources. (new in formatVersion 2/ICU 4.4) 333 * At offset 0 of that array is a 16-bit zero value for empty 16-bit resources. 334 * 335 * Resource16 values in Table16 and Array16 are 16-bit offsets to String-v2 336 * resources, with the offsets relative to the start of the 16-bit-units array. 337 * Starting with formatVersion 3/ICU 56, if offset<poolStringIndex16Limit 338 * then use the pool bundle's 16-bit-units array, 339 * otherwise subtract that limit and use the local 16-bit-units array. 340 * 341 * Type Name Memory layout of values 342 * (in parentheses: scalar, non-offset values) 343 * 344 * 0 Unicode String: int32_t length, UChar[length], (UChar)0, (padding) 345 * or (empty string ("") if offset==0) 346 * 1 Binary: int32_t length, uint8_t[length], (padding) 347 * - the start of the bytes is 16-aligned - 348 * 2 Table: uint16_t count, uint16_t keyStringOffsets[count], (uint16_t padding), Resource[count] 349 * 3 Alias: (physically same value layout as string, new in ICU 2.4) 350 * 4 Table32: int32_t count, int32_t keyStringOffsets[count], Resource[count] 351 * (new in formatVersion 1.1/ICU 2.8) 352 * 5 Table16: uint16_t count, uint16_t keyStringOffsets[count], Resource16[count] 353 * (stored in the 16-bit-units array; new in formatVersion 2/ICU 4.4) 354 * 6 Unicode String-v2:UChar[length], (UChar)0; length determined by the first UChar: 355 * - if first is not a trail surrogate, then the length is implicit 356 * and u_strlen() needs to be called 357 * - if first<0xdfef then length=first&0x3ff (and skip first) 358 * - if first<0xdfff then length=((first-0xdfef)<<16) | second UChar 359 * - if first==0xdfff then length=((second UChar)<<16) | third UChar 360 * (stored in the 16-bit-units array; new in formatVersion 2/ICU 4.4) 361 * 362 * Starting with formatVersion 3/ICU 56, if offset<poolStringIndexLimit 363 * then use the pool bundle's 16-bit-units array, 364 * otherwise subtract that limit and use the local 16-bit-units array. 365 * (Note different limits for Resource16 vs. Resource.) 366 * 367 * 7 Integer: (28-bit offset is integer value) 368 * 8 Array: int32_t count, Resource[count] 369 * 9 Array16: uint16_t count, Resource16[count] 370 * (stored in the 16-bit-units array; new in formatVersion 2/ICU 4.4) 371 * 14 Integer Vector: int32_t length, int32_t[length] 372 * 15 Reserved: This value denotes special purpose resources and is for internal use. 373 * 374 * Note that there are 3 types with data vector values: 375 * - Vectors of 8-bit bytes stored as type Binary. 376 * - Vectors of 16-bit words stored as type Unicode String or Unicode String-v2 377 * (no value restrictions, all values 0..ffff allowed!). 378 * - Vectors of 32-bit words stored as type Integer Vector. 379 */ 380 381 /* 382 * Structure for a single, memory-mapped ResourceBundle. 383 */ 384 typedef struct ResourceData { 385 UDataMemory *data; 386 const int32_t *pRoot; 387 const uint16_t *p16BitUnits; 388 const char *poolBundleKeys; 389 Resource rootRes; 390 int32_t localKeyLimit; 391 const uint16_t *poolBundleStrings; 392 int32_t poolStringIndexLimit; 393 int32_t poolStringIndex16Limit; 394 UBool noFallback; /* see URES_ATT_NO_FALLBACK */ 395 UBool isPoolBundle; 396 UBool usesPoolBundle; 397 UBool useNativeStrcmp; 398 } ResourceData; 399 400 /* 401 * Read a resource bundle from memory. 402 */ 403 U_INTERNAL void U_EXPORT2 404 res_read(ResourceData *pResData, 405 const UDataInfo *pInfo, const void *inBytes, int32_t length, 406 UErrorCode *errorCode); 407 408 /* 409 * Load a resource bundle file. 410 * The ResourceData structure must be allocated externally. 411 */ 412 U_CFUNC void 413 res_load(ResourceData *pResData, 414 const char *path, const char *name, UErrorCode *errorCode); 415 416 /* 417 * Release a resource bundle file. 418 * This does not release the ResourceData structure itself. 419 */ 420 U_CFUNC void 421 res_unload(ResourceData *pResData); 422 423 U_INTERNAL UResType U_EXPORT2 424 res_getPublicType(Resource res); 425 426 /* 427 * Return a pointer to a zero-terminated, const UChar* string 428 * and set its length in *pLength. 429 * Returns NULL if not found. 430 */ 431 U_INTERNAL const UChar * U_EXPORT2 432 res_getString(const ResourceData *pResData, Resource res, int32_t *pLength); 433 434 U_INTERNAL const UChar * U_EXPORT2 435 res_getAlias(const ResourceData *pResData, Resource res, int32_t *pLength); 436 437 U_INTERNAL const uint8_t * U_EXPORT2 438 res_getBinary(const ResourceData *pResData, Resource res, int32_t *pLength); 439 440 U_INTERNAL const int32_t * U_EXPORT2 441 res_getIntVector(const ResourceData *pResData, Resource res, int32_t *pLength); 442 443 U_INTERNAL Resource U_EXPORT2 444 res_getResource(const ResourceData *pResData, const char *key); 445 446 U_INTERNAL int32_t U_EXPORT2 447 res_countArrayItems(const ResourceData *pResData, Resource res); 448 449 U_INTERNAL Resource U_EXPORT2 450 res_getArrayItem(const ResourceData *pResData, Resource array, int32_t indexS); 451 452 U_INTERNAL Resource U_EXPORT2 453 res_getTableItemByIndex(const ResourceData *pResData, Resource table, int32_t indexS, const char ** key); 454 455 U_INTERNAL Resource U_EXPORT2 456 res_getTableItemByKey(const ResourceData *pResData, Resource table, int32_t *indexS, const char* * key); 457 458 /** 459 * Iterates over the path and stops when a scalar resource is found. 460 * Follows aliases. 461 * Modifies the contents of *path (replacing separators with NULs), 462 * and also moves *path forward while it finds items. 463 * 464 * @param path input: "CollationElements/Sequence" or "zoneStrings/3/2" etc.; 465 * output: points to the part that has not yet been processed 466 */ 467 U_CFUNC Resource res_findResource(const ResourceData *pResData, Resource r, 468 char** path, const char** key); 469 470 #ifdef __cplusplus 471 472 #include "resource.h" 473 474 U_NAMESPACE_BEGIN 475 476 class ResourceDataValue : public ResourceValue { 477 public: ResourceDataValue()478 ResourceDataValue() : pResData(NULL), res(URES_NONE) {} 479 virtual ~ResourceDataValue(); 480 setData(const ResourceData * data)481 void setData(const ResourceData *data) { pResData = data; } setResource(Resource r)482 void setResource(Resource r) { res = r; } 483 484 virtual UResType getType() const; 485 virtual const UChar *getString(int32_t &length, UErrorCode &errorCode) const; 486 virtual const UChar *getAliasString(int32_t &length, UErrorCode &errorCode) const; 487 virtual int32_t getInt(UErrorCode &errorCode) const; 488 virtual uint32_t getUInt(UErrorCode &errorCode) const; 489 virtual const int32_t *getIntVector(int32_t &length, UErrorCode &errorCode) const; 490 virtual const uint8_t *getBinary(int32_t &length, UErrorCode &errorCode) const; 491 virtual ResourceArray getArray(UErrorCode &errorCode) const; 492 virtual ResourceTable getTable(UErrorCode &errorCode) const; 493 virtual UBool isNoInheritanceMarker() const; 494 virtual int32_t getStringArray(UnicodeString *dest, int32_t capacity, 495 UErrorCode &errorCode) const; 496 virtual int32_t getStringArrayOrStringAsArray(UnicodeString *dest, int32_t capacity, 497 UErrorCode &errorCode) const; 498 virtual UnicodeString getStringOrFirstOfArray(UErrorCode &errorCode) const; 499 500 const ResourceData *pResData; 501 502 private: 503 Resource res; 504 }; 505 506 U_NAMESPACE_END 507 508 #endif /* __cplusplus */ 509 510 /** 511 * Swap an ICU resource bundle. See udataswp.h. 512 * @internal 513 */ 514 U_CAPI int32_t U_EXPORT2 515 ures_swap(const UDataSwapper *ds, 516 const void *inData, int32_t length, void *outData, 517 UErrorCode *pErrorCode); 518 519 #endif 520