| 1 | // © 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:   UTF-8 | 
|---|
| 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 | * NOTE: For proper logging, please use the res_getInt() constexpr | 
|---|
| 74 | */ | 
|---|
| 75 | #if U_SIGNED_RIGHT_SHIFT_IS_ARITHMETIC | 
|---|
| 76 | #   define RES_GET_INT_NO_TRACE(res) (((int32_t)((res)<<4L))>>4L) | 
|---|
| 77 | #else | 
|---|
| 78 | #   define RES_GET_INT_NO_TRACE(res) (int32_t)(((res)&0x08000000) ? (res)|0xf0000000 : (res)&0x07ffffff) | 
|---|
| 79 | #endif | 
|---|
| 80 |  | 
|---|
| 81 | #define RES_GET_UINT_NO_TRACE(res) ((res)&0x0fffffff) | 
|---|
| 82 |  | 
|---|
| 83 | #define URES_IS_ARRAY(type) ((int32_t)(type)==URES_ARRAY || (int32_t)(type)==URES_ARRAY16) | 
|---|
| 84 | #define URES_IS_TABLE(type) ((int32_t)(type)==URES_TABLE || (int32_t)(type)==URES_TABLE16 || (int32_t)(type)==URES_TABLE32) | 
|---|
| 85 | #define URES_IS_CONTAINER(type) (URES_IS_TABLE(type) || URES_IS_ARRAY(type)) | 
|---|
| 86 |  | 
|---|
| 87 | #define URES_MAKE_RESOURCE(type, offset) (((Resource)(type)<<28)|(Resource)(offset)) | 
|---|
| 88 | #define URES_MAKE_EMPTY_RESOURCE(type) ((Resource)(type)<<28) | 
|---|
| 89 |  | 
|---|
| 90 | /* indexes[] value names; indexes are generally 32-bit (Resource) indexes */ | 
|---|
| 91 | enum { | 
|---|
| 92 | /** | 
|---|
| 93 | * [0] contains the length of indexes[] | 
|---|
| 94 | * which is at most URES_INDEX_TOP of the latest format version | 
|---|
| 95 | * | 
|---|
| 96 | * formatVersion==1: all bits contain the length of indexes[] | 
|---|
| 97 | *   but the length is much less than 0xff; | 
|---|
| 98 | * formatVersion>1: | 
|---|
| 99 | *   only bits  7..0 contain the length of indexes[], | 
|---|
| 100 | *        bits 31..8 are reserved and set to 0 | 
|---|
| 101 | * formatVersion>=3: | 
|---|
| 102 | *        bits 31..8 poolStringIndexLimit bits 23..0 | 
|---|
| 103 | */ | 
|---|
| 104 | URES_INDEX_LENGTH, | 
|---|
| 105 | /** | 
|---|
| 106 | * [1] contains the top of the key strings, | 
|---|
| 107 | *     same as the bottom of resources or UTF-16 strings, rounded up | 
|---|
| 108 | */ | 
|---|
| 109 | URES_INDEX_KEYS_TOP, | 
|---|
| 110 | /** [2] contains the top of all resources */ | 
|---|
| 111 | URES_INDEX_RESOURCES_TOP, | 
|---|
| 112 | /** | 
|---|
| 113 | * [3] contains the top of the bundle, | 
|---|
| 114 | *     in case it were ever different from [2] | 
|---|
| 115 | */ | 
|---|
| 116 | URES_INDEX_BUNDLE_TOP, | 
|---|
| 117 | /** [4] max. length of any table */ | 
|---|
| 118 | URES_INDEX_MAX_TABLE_LENGTH, | 
|---|
| 119 | /** | 
|---|
| 120 | * [5] attributes bit set, see URES_ATT_* (new in formatVersion 1.2) | 
|---|
| 121 | * | 
|---|
| 122 | * formatVersion>=3: | 
|---|
| 123 | *   bits 31..16 poolStringIndex16Limit | 
|---|
| 124 | *   bits 15..12 poolStringIndexLimit bits 27..24 | 
|---|
| 125 | */ | 
|---|
| 126 | URES_INDEX_ATTRIBUTES, | 
|---|
| 127 | /** | 
|---|
| 128 | * [6] top of the 16-bit units (UTF-16 string v2 UChars, URES_TABLE16, URES_ARRAY16), | 
|---|
| 129 | *     rounded up (new in formatVersion 2.0, ICU 4.4) | 
|---|
| 130 | */ | 
|---|
| 131 | URES_INDEX_16BIT_TOP, | 
|---|
| 132 | /** [7] checksum of the pool bundle (new in formatVersion 2.0, ICU 4.4) */ | 
|---|
| 133 | URES_INDEX_POOL_CHECKSUM, | 
|---|
| 134 | URES_INDEX_TOP | 
|---|
| 135 | }; | 
|---|
| 136 |  | 
|---|
| 137 | /* | 
|---|
| 138 | * Nofallback attribute, attribute bit 0 in indexes[URES_INDEX_ATTRIBUTES]. | 
|---|
| 139 | * New in formatVersion 1.2 (ICU 3.6). | 
|---|
| 140 | * | 
|---|
| 141 | * If set, then this resource bundle is a standalone bundle. | 
|---|
| 142 | * If not set, then the bundle participates in locale fallback, eventually | 
|---|
| 143 | * all the way to the root bundle. | 
|---|
| 144 | * If indexes[] is missing or too short, then the attribute cannot be determined | 
|---|
| 145 | * reliably. Dependency checking should ignore such bundles, and loading should | 
|---|
| 146 | * use fallbacks. | 
|---|
| 147 | */ | 
|---|
| 148 | #define URES_ATT_NO_FALLBACK 1 | 
|---|
| 149 |  | 
|---|
| 150 | /* | 
|---|
| 151 | * Attributes for bundles that are, or use, a pool bundle. | 
|---|
| 152 | * A pool bundle provides key strings that are shared among several other bundles | 
|---|
| 153 | * to reduce their total size. | 
|---|
| 154 | * New in formatVersion 2 (ICU 4.4). | 
|---|
| 155 | */ | 
|---|
| 156 | #define URES_ATT_IS_POOL_BUNDLE 2 | 
|---|
| 157 | #define URES_ATT_USES_POOL_BUNDLE 4 | 
|---|
| 158 |  | 
|---|
| 159 | /* | 
|---|
| 160 | * File format for .res resource bundle files | 
|---|
| 161 | * | 
|---|
| 162 | * ICU 56: New in formatVersion 3 compared with 2: ------------- | 
|---|
| 163 | * | 
|---|
| 164 | * Resource bundles can optionally use shared string-v2 values | 
|---|
| 165 | * stored in the pool bundle. | 
|---|
| 166 | * If so, then the indexes[] contain two new values | 
|---|
| 167 | * in previously-unused bits of existing indexes[] slots: | 
|---|
| 168 | * - poolStringIndexLimit: | 
|---|
| 169 | *     String-v2 offsets (in 32-bit Resource words) below this limit | 
|---|
| 170 | *     point to pool bundle string-v2 values. | 
|---|
| 171 | * - poolStringIndex16Limit: | 
|---|
| 172 | *     Resource16 string-v2 offsets below this limit | 
|---|
| 173 | *     point to pool bundle string-v2 values. | 
|---|
| 174 | * Guarantee: poolStringIndex16Limit <= poolStringIndexLimit | 
|---|
| 175 | * | 
|---|
| 176 | * The local bundle's poolStringIndexLimit is greater than | 
|---|
| 177 | * any pool bundle string index used in the local bundle. | 
|---|
| 178 | * The poolStringIndexLimit should not be greater than | 
|---|
| 179 | * the maximum possible pool bundle string index. | 
|---|
| 180 | * | 
|---|
| 181 | * The maximum possible pool bundle string index is the index to the last non-NUL | 
|---|
| 182 | * pool string character, due to suffix sharing. | 
|---|
| 183 | * | 
|---|
| 184 | * In the pool bundle, there is no structure that lists the strings. | 
|---|
| 185 | * (The root resource is an empty Table.) | 
|---|
| 186 | * If the strings need to be enumerated (as genrb --usePoolBundle does), | 
|---|
| 187 | * then iterate through the pool bundle's 16-bit-units array from the beginning. | 
|---|
| 188 | * Stop at the end of the array, or when an explicit or implicit string length | 
|---|
| 189 | * would lead beyond the end of the array, | 
|---|
| 190 | * or when an apparent string is not NUL-terminated. | 
|---|
| 191 | * (Future genrb version might terminate the strings with | 
|---|
| 192 | * what looks like a large explicit string length.) | 
|---|
| 193 | * | 
|---|
| 194 | * ICU 4.4: New in formatVersion 2 compared with 1.3: ------------- | 
|---|
| 195 | * | 
|---|
| 196 | * Three new resource types -- String-v2, Table16 and Array16 -- have their | 
|---|
| 197 | * values stored in a new array of 16-bit units between the table key strings | 
|---|
| 198 | * and the start of the other resources. | 
|---|
| 199 | * | 
|---|
| 200 | * genrb eliminates duplicates among Unicode string-v2 values. | 
|---|
| 201 | * Multiple Unicode strings may use the same offset and string data, | 
|---|
| 202 | * or a short string may point to the suffix of a longer string. ("Suffix sharing") | 
|---|
| 203 | * For example, one string "abc" may be reused for another string "bc" by pointing | 
|---|
| 204 | * to the second character. (Short strings-v2 are NUL-terminated | 
|---|
| 205 | * and not preceded by an explicit length value.) | 
|---|
| 206 | * | 
|---|
| 207 | * It is allowed for all resource types to share values. | 
|---|
| 208 | * The swapper code (ures_swap()) has been modified so that it swaps each item | 
|---|
| 209 | * exactly once. | 
|---|
| 210 | * | 
|---|
| 211 | * A resource bundle may use a special pool bundle. Some or all of the table key strings | 
|---|
| 212 | * of the using-bundle are omitted, and the key string offsets for such key strings refer | 
|---|
| 213 | * to offsets in the pool bundle. | 
|---|
| 214 | * The using-bundle's and the pool-bundle's indexes[URES_INDEX_POOL_CHECKSUM] values | 
|---|
| 215 | * must match. | 
|---|
| 216 | * Two bits in indexes[URES_INDEX_ATTRIBUTES] indicate whether a resource bundle | 
|---|
| 217 | * is or uses a pool bundle. | 
|---|
| 218 | * | 
|---|
| 219 | * Table key strings must be compared in ASCII order, even if they are not | 
|---|
| 220 | * stored in ASCII. | 
|---|
| 221 | * | 
|---|
| 222 | * New in formatVersion 1.3 compared with 1.2: ------------- | 
|---|
| 223 | * | 
|---|
| 224 | * genrb eliminates duplicates among key strings. | 
|---|
| 225 | * Multiple table items may share one key string, or one item may point | 
|---|
| 226 | * to the suffix of another's key string. ("Suffix sharing") | 
|---|
| 227 | * For example, one key "abc" may be reused for another key "bc" by pointing | 
|---|
| 228 | * to the second character. (Key strings are NUL-terminated.) | 
|---|
| 229 | * | 
|---|
| 230 | * ------------- | 
|---|
| 231 | * | 
|---|
| 232 | * An ICU4C resource bundle file (.res) is a binary, memory-mappable file | 
|---|
| 233 | * with nested, hierarchical data structures. | 
|---|
| 234 | * It physically contains the following: | 
|---|
| 235 | * | 
|---|
| 236 | *   Resource root; -- 32-bit Resource item, root item for this bundle's tree; | 
|---|
| 237 | *                     currently, the root item must be a table or table32 resource item | 
|---|
| 238 | *   int32_t indexes[indexes[0]]; -- array of indexes for friendly | 
|---|
| 239 | *                                   reading and swapping; see URES_INDEX_* above | 
|---|
| 240 | *                                   new in formatVersion 1.1 (ICU 2.8) | 
|---|
| 241 | *   char keys[]; -- characters for key strings | 
|---|
| 242 | *                   (formatVersion 1.0: up to 65k of characters; 1.1: <2G) | 
|---|
| 243 | *                   (minus the space for root and indexes[]), | 
|---|
| 244 | *                   which consist of invariant characters (ASCII/EBCDIC) and are NUL-terminated; | 
|---|
| 245 | *                   padded to multiple of 4 bytes for 4-alignment of the following data | 
|---|
| 246 | *   uint16_t 16BitUnits[]; -- resources that are stored entirely as sequences of 16-bit units | 
|---|
| 247 | *                             (new in formatVersion 2/ICU 4.4) | 
|---|
| 248 | *                             data is indexed by the offset values in 16-bit resource types, | 
|---|
| 249 | *                             with offset 0 pointing to the beginning of this array; | 
|---|
| 250 | *                             there is a 0 at offset 0, for empty resources; | 
|---|
| 251 | *                             padded to multiple of 4 bytes for 4-alignment of the following data | 
|---|
| 252 | *   data; -- data directly and indirectly indexed by the root item; | 
|---|
| 253 | *            the structure is determined by walking the tree | 
|---|
| 254 | * | 
|---|
| 255 | * Each resource bundle item has a 32-bit Resource handle (see typedef above) | 
|---|
| 256 | * which contains the item type number in its upper 4 bits (31..28) and either | 
|---|
| 257 | * an offset or a direct value in its lower 28 bits (27..0). | 
|---|
| 258 | * The order of items is undefined and only determined by walking the tree. | 
|---|
| 259 | * Leaves of the tree may be stored first or last or anywhere in between, | 
|---|
| 260 | * and it is in theory possible to have unreferenced holes in the file. | 
|---|
| 261 | * | 
|---|
| 262 | * 16-bit-unit values: | 
|---|
| 263 | * Starting with formatVersion 2/ICU 4.4, some resources are stored in a special | 
|---|
| 264 | * array of 16-bit units. Each resource value is a sequence of 16-bit units, | 
|---|
| 265 | * with no per-resource padding to a 4-byte boundary. | 
|---|
| 266 | * 16-bit container types (Table16 and Array16) contain Resource16 values | 
|---|
| 267 | * which are offsets to String-v2 resources in the same 16-bit-units array. | 
|---|
| 268 | * | 
|---|
| 269 | * Direct values: | 
|---|
| 270 | * - Empty Unicode strings have an offset value of 0 in the Resource handle itself. | 
|---|
| 271 | * - Starting with formatVersion 2/ICU 4.4, an offset value of 0 for | 
|---|
| 272 | *   _any_ resource type indicates an empty value. | 
|---|
| 273 | * - Integer values are 28-bit values stored in the Resource handle itself; | 
|---|
| 274 | *   the interpretation of unsigned vs. signed integers is up to the application. | 
|---|
| 275 | * | 
|---|
| 276 | * All other types and values use 28-bit offsets to point to the item's data. | 
|---|
| 277 | * The offset is an index to the first 32-bit word of the value, relative to the | 
|---|
| 278 | * start of the resource data (i.e., the root item handle is at offset 0). | 
|---|
| 279 | * To get byte offsets, the offset is multiplied by 4 (or shifted left by 2 bits). | 
|---|
| 280 | * All resource item values are 4-aligned. | 
|---|
| 281 | * | 
|---|
| 282 | * New in formatVersion 2/ICU 4.4: Some types use offsets into the 16-bit-units array, | 
|---|
| 283 | * indexing 16-bit units in that array. | 
|---|
| 284 | * | 
|---|
| 285 | * The structures (memory layouts) for the values for each item type are listed | 
|---|
| 286 | * in the table below. | 
|---|
| 287 | * | 
|---|
| 288 | * Nested, hierarchical structures: ------------- | 
|---|
| 289 | * | 
|---|
| 290 | * Table items contain key-value pairs where the keys are offsets to char * key strings. | 
|---|
| 291 | * The values of these pairs are either Resource handles or | 
|---|
| 292 | * offsets into the 16-bit-units array, depending on the table type. | 
|---|
| 293 | * | 
|---|
| 294 | * Array items are simple vectors of Resource handles, | 
|---|
| 295 | * or of offsets into the 16-bit-units array, depending on the array type. | 
|---|
| 296 | * | 
|---|
| 297 | * Table key string offsets: ------- | 
|---|
| 298 | * | 
|---|
| 299 | * Key string offsets are relative to the start of the resource data (of the root handle), | 
|---|
| 300 | * i.e., the first string has an offset of 4+sizeof(indexes). | 
|---|
| 301 | * (After the 4-byte root handle and after the indexes array.) | 
|---|
| 302 | * | 
|---|
| 303 | * If the resource bundle uses a pool bundle, then some key strings are stored | 
|---|
| 304 | * in the pool bundle rather than in the local bundle itself. | 
|---|
| 305 | * - In a Table or Table16, the 16-bit key string offset is local if it is | 
|---|
| 306 | *   less than indexes[URES_INDEX_KEYS_TOP]<<2. | 
|---|
| 307 | *   Otherwise, subtract indexes[URES_INDEX_KEYS_TOP]<<2 to get the offset into | 
|---|
| 308 | *   the pool bundle key strings. | 
|---|
| 309 | * - In a Table32, the 32-bit key string offset is local if it is non-negative. | 
|---|
| 310 | *   Otherwise, reset bit 31 to get the pool key string offset. | 
|---|
| 311 | * | 
|---|
| 312 | * Unlike the local offset, the pool key offset is relative to | 
|---|
| 313 | * the start of the key strings, not to the start of the bundle. | 
|---|
| 314 | * | 
|---|
| 315 | * An alias item is special (and new in ICU 2.4): -------------- | 
|---|
| 316 | * | 
|---|
| 317 | * Its memory layout is just like for a UnicodeString, but at runtime it resolves to | 
|---|
| 318 | * another resource bundle's item according to the path in the string. | 
|---|
| 319 | * This is used to share items across bundles that are in different lookup/fallback | 
|---|
| 320 | * chains (e.g., large collation data among zh_TW and zh_HK). | 
|---|
| 321 | * This saves space (for large items) and maintenance effort (less duplication of data). | 
|---|
| 322 | * | 
|---|
| 323 | * -------------------------------------------------------------------------- | 
|---|
| 324 | * | 
|---|
| 325 | * Resource types: | 
|---|
| 326 | * | 
|---|
| 327 | * Most resources have their values stored at four-byte offsets from the start | 
|---|
| 328 | * of the resource data. These values are at least 4-aligned. | 
|---|
| 329 | * Some resource values are stored directly in the offset field of the Resource itself. | 
|---|
| 330 | * See UResType in unicode/ures.h for enumeration constants for Resource types. | 
|---|
| 331 | * | 
|---|
| 332 | * Some resources have their values stored as sequences of 16-bit units, | 
|---|
| 333 | * at 2-byte offsets from the start of a contiguous 16-bit-unit array between | 
|---|
| 334 | * the table key strings and the other resources. (new in formatVersion 2/ICU 4.4) | 
|---|
| 335 | * At offset 0 of that array is a 16-bit zero value for empty 16-bit resources. | 
|---|
| 336 | * | 
|---|
| 337 | * Resource16 values in Table16 and Array16 are 16-bit offsets to String-v2 | 
|---|
| 338 | * resources, with the offsets relative to the start of the 16-bit-units array. | 
|---|
| 339 | * Starting with formatVersion 3/ICU 56, if offset<poolStringIndex16Limit | 
|---|
| 340 | * then use the pool bundle's 16-bit-units array, | 
|---|
| 341 | * otherwise subtract that limit and use the local 16-bit-units array. | 
|---|
| 342 | * | 
|---|
| 343 | * Type Name            Memory layout of values | 
|---|
| 344 | *                      (in parentheses: scalar, non-offset values) | 
|---|
| 345 | * | 
|---|
| 346 | * 0  Unicode String:   int32_t length, UChar[length], (UChar)0, (padding) | 
|---|
| 347 | *                  or  (empty string ("") if offset==0) | 
|---|
| 348 | * 1  Binary:           int32_t length, uint8_t[length], (padding) | 
|---|
| 349 | *                      - the start of the bytes is 16-aligned - | 
|---|
| 350 | * 2  Table:            uint16_t count, uint16_t keyStringOffsets[count], (uint16_t padding), Resource[count] | 
|---|
| 351 | * 3  Alias:            (physically same value layout as string, new in ICU 2.4) | 
|---|
| 352 | * 4  Table32:          int32_t count, int32_t keyStringOffsets[count], Resource[count] | 
|---|
| 353 | *                      (new in formatVersion 1.1/ICU 2.8) | 
|---|
| 354 | * 5  Table16:          uint16_t count, uint16_t keyStringOffsets[count], Resource16[count] | 
|---|
| 355 | *                      (stored in the 16-bit-units array; new in formatVersion 2/ICU 4.4) | 
|---|
| 356 | * 6  Unicode String-v2:UChar[length], (UChar)0; length determined by the first UChar: | 
|---|
| 357 | *                      - if first is not a trail surrogate, then the length is implicit | 
|---|
| 358 | *                        and u_strlen() needs to be called | 
|---|
| 359 | *                      - if first<0xdfef then length=first&0x3ff (and skip first) | 
|---|
| 360 | *                      - if first<0xdfff then length=((first-0xdfef)<<16) | second UChar | 
|---|
| 361 | *                      - if first==0xdfff then length=((second UChar)<<16) | third UChar | 
|---|
| 362 | *                      (stored in the 16-bit-units array; new in formatVersion 2/ICU 4.4) | 
|---|
| 363 | * | 
|---|
| 364 | *                      Starting with formatVersion 3/ICU 56, if offset<poolStringIndexLimit | 
|---|
| 365 | *                      then use the pool bundle's 16-bit-units array, | 
|---|
| 366 | *                      otherwise subtract that limit and use the local 16-bit-units array. | 
|---|
| 367 | *                      (Note different limits for Resource16 vs. Resource.) | 
|---|
| 368 | * | 
|---|
| 369 | * 7  Integer:          (28-bit offset is integer value) | 
|---|
| 370 | * 8  Array:            int32_t count, Resource[count] | 
|---|
| 371 | * 9  Array16:          uint16_t count, Resource16[count] | 
|---|
| 372 | *                      (stored in the 16-bit-units array; new in formatVersion 2/ICU 4.4) | 
|---|
| 373 | * 14 Integer Vector:   int32_t length, int32_t[length] | 
|---|
| 374 | * 15 Reserved:         This value denotes special purpose resources and is for internal use. | 
|---|
| 375 | * | 
|---|
| 376 | * Note that there are 3 types with data vector values: | 
|---|
| 377 | * - Vectors of 8-bit bytes stored as type Binary. | 
|---|
| 378 | * - Vectors of 16-bit words stored as type Unicode String or Unicode String-v2 | 
|---|
| 379 | *                     (no value restrictions, all values 0..ffff allowed!). | 
|---|
| 380 | * - Vectors of 32-bit words stored as type Integer Vector. | 
|---|
| 381 | */ | 
|---|
| 382 |  | 
|---|
| 383 | /* | 
|---|
| 384 | * Structure for a single, memory-mapped ResourceBundle. | 
|---|
| 385 | */ | 
|---|
| 386 | typedef struct ResourceData { | 
|---|
| 387 | UDataMemory *data; | 
|---|
| 388 | const int32_t *pRoot; | 
|---|
| 389 | const uint16_t *p16BitUnits; | 
|---|
| 390 | const char *poolBundleKeys; | 
|---|
| 391 | Resource rootRes; | 
|---|
| 392 | int32_t localKeyLimit; | 
|---|
| 393 | const uint16_t *poolBundleStrings; | 
|---|
| 394 | int32_t poolStringIndexLimit; | 
|---|
| 395 | int32_t poolStringIndex16Limit; | 
|---|
| 396 | UBool noFallback; /* see URES_ATT_NO_FALLBACK */ | 
|---|
| 397 | UBool isPoolBundle; | 
|---|
| 398 | UBool usesPoolBundle; | 
|---|
| 399 | UBool useNativeStrcmp; | 
|---|
| 400 | } ResourceData; | 
|---|
| 401 |  | 
|---|
| 402 | /* | 
|---|
| 403 | * Read a resource bundle from memory. | 
|---|
| 404 | */ | 
|---|
| 405 | U_INTERNAL void U_EXPORT2 | 
|---|
| 406 | res_read(ResourceData *pResData, | 
|---|
| 407 | const UDataInfo *pInfo, const void *inBytes, int32_t length, | 
|---|
| 408 | UErrorCode *errorCode); | 
|---|
| 409 |  | 
|---|
| 410 | /* | 
|---|
| 411 | * Load a resource bundle file. | 
|---|
| 412 | * The ResourceData structure must be allocated externally. | 
|---|
| 413 | */ | 
|---|
| 414 | U_CFUNC void | 
|---|
| 415 | res_load(ResourceData *pResData, | 
|---|
| 416 | const char *path, const char *name, UErrorCode *errorCode); | 
|---|
| 417 |  | 
|---|
| 418 | /* | 
|---|
| 419 | * Release a resource bundle file. | 
|---|
| 420 | * This does not release the ResourceData structure itself. | 
|---|
| 421 | */ | 
|---|
| 422 | U_CFUNC void | 
|---|
| 423 | res_unload(ResourceData *pResData); | 
|---|
| 424 |  | 
|---|
| 425 | U_INTERNAL UResType U_EXPORT2 | 
|---|
| 426 | res_getPublicType(Resource res); | 
|---|
| 427 |  | 
|---|
| 428 | /////////////////////////////////////////////////////////////////////////// | 
|---|
| 429 | // To enable tracing, use the inline versions of the res_get* functions. // | 
|---|
| 430 | /////////////////////////////////////////////////////////////////////////// | 
|---|
| 431 |  | 
|---|
| 432 | /* | 
|---|
| 433 | * Return a pointer to a zero-terminated, const UChar* string | 
|---|
| 434 | * and set its length in *pLength. | 
|---|
| 435 | * Returns NULL if not found. | 
|---|
| 436 | */ | 
|---|
| 437 | U_INTERNAL const UChar * U_EXPORT2 | 
|---|
| 438 | res_getStringNoTrace(const ResourceData *pResData, Resource res, int32_t *pLength); | 
|---|
| 439 |  | 
|---|
| 440 | U_INTERNAL const uint8_t * U_EXPORT2 | 
|---|
| 441 | res_getBinaryNoTrace(const ResourceData *pResData, Resource res, int32_t *pLength); | 
|---|
| 442 |  | 
|---|
| 443 | U_INTERNAL const int32_t * U_EXPORT2 | 
|---|
| 444 | res_getIntVectorNoTrace(const ResourceData *pResData, Resource res, int32_t *pLength); | 
|---|
| 445 |  | 
|---|
| 446 | U_INTERNAL const UChar * U_EXPORT2 | 
|---|
| 447 | res_getAlias(const ResourceData *pResData, Resource res, int32_t *pLength); | 
|---|
| 448 |  | 
|---|
| 449 | U_INTERNAL Resource U_EXPORT2 | 
|---|
| 450 | res_getResource(const ResourceData *pResData, const char *key); | 
|---|
| 451 |  | 
|---|
| 452 | U_INTERNAL int32_t U_EXPORT2 | 
|---|
| 453 | res_countArrayItems(const ResourceData *pResData, Resource res); | 
|---|
| 454 |  | 
|---|
| 455 | U_INTERNAL Resource U_EXPORT2 | 
|---|
| 456 | res_getArrayItem(const ResourceData *pResData, Resource array, int32_t indexS); | 
|---|
| 457 |  | 
|---|
| 458 | U_INTERNAL Resource U_EXPORT2 | 
|---|
| 459 | res_getTableItemByIndex(const ResourceData *pResData, Resource table, int32_t indexS, const char ** key); | 
|---|
| 460 |  | 
|---|
| 461 | U_INTERNAL Resource U_EXPORT2 | 
|---|
| 462 | res_getTableItemByKey(const ResourceData *pResData, Resource table, int32_t *indexS, const char* * key); | 
|---|
| 463 |  | 
|---|
| 464 | /** | 
|---|
| 465 | * Iterates over the path and stops when a scalar resource is found. | 
|---|
| 466 | * Follows aliases. | 
|---|
| 467 | * Modifies the contents of *path (replacing separators with NULs), | 
|---|
| 468 | * and also moves *path forward while it finds items. | 
|---|
| 469 | * | 
|---|
| 470 | * @param path input: "CollationElements/Sequence" or "zoneStrings/3/2" etc.; | 
|---|
| 471 | *             output: points to the part that has not yet been processed | 
|---|
| 472 | */ | 
|---|
| 473 | U_CFUNC Resource res_findResource(const ResourceData *pResData, Resource r, | 
|---|
| 474 | char** path, const char** key); | 
|---|
| 475 |  | 
|---|
| 476 | #ifdef __cplusplus | 
|---|
| 477 |  | 
|---|
| 478 | #include "resource.h" | 
|---|
| 479 | #include "restrace.h" | 
|---|
| 480 |  | 
|---|
| 481 | U_NAMESPACE_BEGIN | 
|---|
| 482 |  | 
|---|
| 483 | inline const UChar* res_getString(const ResourceTracer& traceInfo, | 
|---|
| 484 | const ResourceData *pResData, Resource res, int32_t *pLength) { | 
|---|
| 485 | traceInfo.trace( "string"); | 
|---|
| 486 | return res_getStringNoTrace(pResData, res, pLength); | 
|---|
| 487 | } | 
|---|
| 488 |  | 
|---|
| 489 | inline const uint8_t* res_getBinary(const ResourceTracer& traceInfo, | 
|---|
| 490 | const ResourceData *pResData, Resource res, int32_t *pLength) { | 
|---|
| 491 | traceInfo.trace( "binary"); | 
|---|
| 492 | return res_getBinaryNoTrace(pResData, res, pLength); | 
|---|
| 493 | } | 
|---|
| 494 |  | 
|---|
| 495 | inline const int32_t* res_getIntVector(const ResourceTracer& traceInfo, | 
|---|
| 496 | const ResourceData *pResData, Resource res, int32_t *pLength) { | 
|---|
| 497 | traceInfo.trace( "intvector"); | 
|---|
| 498 | return res_getIntVectorNoTrace(pResData, res, pLength); | 
|---|
| 499 | } | 
|---|
| 500 |  | 
|---|
| 501 | inline int32_t res_getInt(const ResourceTracer& traceInfo, Resource res) { | 
|---|
| 502 | traceInfo.trace( "int"); | 
|---|
| 503 | return RES_GET_INT_NO_TRACE(res); | 
|---|
| 504 | } | 
|---|
| 505 |  | 
|---|
| 506 | inline uint32_t res_getUInt(const ResourceTracer& traceInfo, Resource res) { | 
|---|
| 507 | traceInfo.trace( "uint"); | 
|---|
| 508 | return RES_GET_UINT_NO_TRACE(res); | 
|---|
| 509 | } | 
|---|
| 510 |  | 
|---|
| 511 | class ResourceDataValue : public ResourceValue { | 
|---|
| 512 | public: | 
|---|
| 513 | ResourceDataValue() : | 
|---|
| 514 | res(static_cast<Resource>(URES_NONE)), | 
|---|
| 515 | fTraceInfo() {} | 
|---|
| 516 | virtual ~ResourceDataValue(); | 
|---|
| 517 |  | 
|---|
| 518 | void setData(const ResourceData *data) { | 
|---|
| 519 | resData = *data; | 
|---|
| 520 | } | 
|---|
| 521 |  | 
|---|
| 522 | void setResource(Resource r, ResourceTracer&& traceInfo) { | 
|---|
| 523 | res = r; | 
|---|
| 524 | fTraceInfo = traceInfo; | 
|---|
| 525 | } | 
|---|
| 526 |  | 
|---|
| 527 | const ResourceData &getData() const { return resData; } | 
|---|
| 528 | virtual UResType getType() const; | 
|---|
| 529 | virtual const UChar *getString(int32_t &length, UErrorCode &errorCode) const; | 
|---|
| 530 | virtual const UChar *getAliasString(int32_t &length, UErrorCode &errorCode) const; | 
|---|
| 531 | virtual int32_t getInt(UErrorCode &errorCode) const; | 
|---|
| 532 | virtual uint32_t getUInt(UErrorCode &errorCode) const; | 
|---|
| 533 | virtual const int32_t *getIntVector(int32_t &length, UErrorCode &errorCode) const; | 
|---|
| 534 | virtual const uint8_t *getBinary(int32_t &length, UErrorCode &errorCode) const; | 
|---|
| 535 | virtual ResourceArray getArray(UErrorCode &errorCode) const; | 
|---|
| 536 | virtual ResourceTable getTable(UErrorCode &errorCode) const; | 
|---|
| 537 | virtual UBool isNoInheritanceMarker() const; | 
|---|
| 538 | virtual int32_t getStringArray(UnicodeString *dest, int32_t capacity, | 
|---|
| 539 | UErrorCode &errorCode) const; | 
|---|
| 540 | virtual int32_t getStringArrayOrStringAsArray(UnicodeString *dest, int32_t capacity, | 
|---|
| 541 | UErrorCode &errorCode) const; | 
|---|
| 542 | virtual UnicodeString getStringOrFirstOfArray(UErrorCode &errorCode) const; | 
|---|
| 543 |  | 
|---|
| 544 | private: | 
|---|
| 545 | // TODO(ICU-20769): If UResourceBundle.fResData becomes a pointer, | 
|---|
| 546 | // then remove this value field again and just store a pResData pointer. | 
|---|
| 547 | ResourceData resData; | 
|---|
| 548 | Resource res; | 
|---|
| 549 | ResourceTracer fTraceInfo; | 
|---|
| 550 | }; | 
|---|
| 551 |  | 
|---|
| 552 | U_NAMESPACE_END | 
|---|
| 553 |  | 
|---|
| 554 | #endif  /* __cplusplus */ | 
|---|
| 555 |  | 
|---|
| 556 | /** | 
|---|
| 557 | * Swap an ICU resource bundle. See udataswp.h. | 
|---|
| 558 | * @internal | 
|---|
| 559 | */ | 
|---|
| 560 | U_CAPI int32_t U_EXPORT2 | 
|---|
| 561 | ures_swap(const UDataSwapper *ds, | 
|---|
| 562 | const void *inData, int32_t length, void *outData, | 
|---|
| 563 | UErrorCode *pErrorCode); | 
|---|
| 564 |  | 
|---|
| 565 | #endif | 
|---|
| 566 |  | 
|---|