icuSources/common/unicode/udata.h

   1 /*
   2 ******************************************************************************
   3 *
   4 *   Copyright (C) 1999-2010, International Business Machines
   5 *   Corporation and others.  All Rights Reserved.
   6 *
   7 ******************************************************************************
   8 *   file name:  udata.h
   9 *   encoding:   US-ASCII
  10 *   tab size:   8 (not used)
  11 *   indentation:4
  12 *
  13 *   created on: 1999oct25
  14 *   created by: Markus W. Scherer
  15 */
  16
  17 #ifndef __UDATA_H__
  18 #define __UDATA_H__
  19
  20 #include "unicode/utypes.h"
  21 #include "unicode/localpointer.h"
  22
  23 U_CDECL_BEGIN
  24
  25 /**
  26  * \file
  27  * \brief C API: Data loading interface
  28  *
  29  * <h2>Information about data loading interface</h2>
  30  *
  31  * This API is used to find and efficiently load data for ICU and applications
  32  * using ICU. It provides an abstract interface that specifies a data type and
  33  * name to find and load the data. Normally this API is used by other ICU APIs
  34  * to load required data out of the ICU data library, but it can be used to
  35  * load data out of other places.
  36  *
  37  * See the User Guide Data Management chapter.
  38  */
  39
  40 #ifndef U_HIDE_INTERNAL_API
  41 /**
  42  * Character used to separate package names from tree names
  43  * @internal ICU 3.0
  44  */
  45 #define U_TREE_SEPARATOR '-'
  46
  47 /**
  48  * String used to separate package names from tree names
  49  * @internal ICU 3.0
  50  */
  51 #define U_TREE_SEPARATOR_STRING "-"
  52
  53 /**
  54  * Character used to separate parts of entry names
  55  * @internal ICU 3.0
  56  */
  57 #define U_TREE_ENTRY_SEP_CHAR '/'
  58
  59 /**
  60  * String used to separate parts of entry names
  61  * @internal ICU 3.0
  62  */
  63 #define U_TREE_ENTRY_SEP_STRING "/"
  64
  65 /**
  66  * Alias for standard ICU data
  67  * @internal ICU 3.0
  68  */
  69 #define U_ICUDATA_ALIAS "ICUDATA"
  70
  71 #endif /* U_HIDE_INTERNAL_API */
  72
  73 /**
  74  * UDataInfo contains the properties about the requested data.
  75  * This is meta data.
  76  *
  77  * <p>This structure may grow in the future, indicated by the
  78  * <code>size</code> field.</p>
  79  *
  80  * <p>The platform data property fields help determine if a data
  81  * file can be efficiently used on a given machine.
  82  * The particular fields are of importance only if the data
  83  * is affected by the properties - if there is integer data
  84  * with word sizes > 1 byte, char* text, or UChar* text.</p>
  85  *
  86  * <p>The implementation for the <code>udata_open[Choice]()</code>
  87  * functions may reject data based on the value in <code>isBigEndian</code>.
  88  * No other field is used by the <code>udata</code> API implementation.</p>
  89  *
  90  * <p>The <code>dataFormat</code> may be used to identify
  91  * the kind of data, e.g. a converter table.</p>
  92  *
  93  * <p>The <code>formatVersion</code> field should be used to
  94  * make sure that the format can be interpreted.
  95  * I may be a good idea to check only for the one or two highest
  96  * of the version elements to allow the data memory to
  97  * get more or somewhat rearranged contents, for as long
  98  * as the using code can still interpret the older contents.</p>
  99  *
 100  * <p>The <code>dataVersion</code> field is intended to be a
 101  * common place to store the source version of the data;
 102  * for data from the Unicode character database, this could
 103  * reflect the Unicode version.</p>
 104  * @stable ICU 2.0
 105  */
 106 typedef struct {
 107     /** sizeof(UDataInfo)
 108      *  @stable ICU 2.0 */
 109     uint16_t size;
 110
 111     /** unused, set to 0
 112      *  @stable ICU 2.0*/
 113     uint16_t reservedWord;
 114
 115     /* platform data properties */
 116     /** 0 for little-endian machine, 1 for big-endian
 117      *  @stable ICU 2.0 */
 118     uint8_t isBigEndian;
 119
 120     /** see U_CHARSET_FAMILY values in utypes.h
 121      *  @stable ICU 2.0*/
 122     uint8_t charsetFamily;
 123
 124     /** sizeof(UChar), one of { 1, 2, 4 }
 125      *  @stable ICU 2.0*/
 126     uint8_t sizeofUChar;
 127
 128     /** unused, set to 0
 129      *  @stable ICU 2.0*/
 130     uint8_t reservedByte;
 131
 132     /** data format identifier
 133      *  @stable ICU 2.0*/
 134     uint8_t dataFormat[4];
 135
 136     /** versions: [0] major [1] minor [2] milli [3] micro
 137      *  @stable ICU 2.0*/
 138     uint8_t formatVersion[4];
 139
 140     /** versions: [0] major [1] minor [2] milli [3] micro
 141      *  @stable ICU 2.0*/
 142     uint8_t dataVersion[4];
 143 } UDataInfo;
 144
 145 /* API for reading data -----------------------------------------------------*/
 146
 147 /**
 148  * Forward declaration of the data memory type.
 149  * @stable ICU 2.0
 150  */
 151 typedef struct UDataMemory UDataMemory;
 152
 153 /**
 154  * Callback function for udata_openChoice().
 155  * @param context parameter passed into <code>udata_openChoice()</code>.
 156  * @param type The type of the data as passed into <code>udata_openChoice()</code>.
 157  *             It may be <code>NULL</code>.
 158  * @param name The name of the data as passed into <code>udata_openChoice()</code>.
 159  * @param pInfo A pointer to the <code>UDataInfo</code> structure
 160  *              of data that has been loaded and will be returned
 161  *              by <code>udata_openChoice()</code> if this function
 162  *              returns <code>TRUE</code>.
 163  * @return TRUE if the current data memory is acceptable
 164  * @stable ICU 2.0
 165  */
 166 typedef UBool U_CALLCONV
 167 UDataMemoryIsAcceptable(void *context,
 168                         const char *type, const char *name,
 169                         const UDataInfo *pInfo);
 170
 171
 172 /**
 173  * Convenience function.
 174  * This function works the same as <code>udata_openChoice</code>
 175  * except that any data that matches the type and name
 176  * is assumed to be acceptable.
 177  * @param path Specifies an absolute path and/or a basename for the
 178  *             finding of the data in the file system.
 179  *             <code>NULL</code> for ICU data.
 180  * @param type A string that specifies the type of data to be loaded.
 181  *             For example, resource bundles are loaded with type "res",
 182  *             conversion tables with type "cnv".
 183  *             This may be <code>NULL</code> or empty.
 184  * @param name A string that specifies the name of the data.
 185  * @param pErrorCode An ICU UErrorCode parameter. It must not be <code>NULL</code>.
 186  * @return A pointer (handle) to a data memory object, or <code>NULL</code>
 187  *         if an error occurs. Call <code>udata_getMemory()</code>
 188  *         to get a pointer to the actual data.
 189  *
 190  * @see udata_openChoice
 191  * @stable ICU 2.0
 192  */
 193 U_STABLE UDataMemory * U_EXPORT2
 194 udata_open(const char *path, const char *type, const char *name,
 195            UErrorCode *pErrorCode);
 196
 197 /**
 198  * Data loading function.
 199  * This function is used to find and load efficiently data for
 200  * ICU and applications using ICU.
 201  * It provides an abstract interface that allows to specify a data
 202  * type and name to find and load the data.
 203  *
 204  * <p>The implementation depends on platform properties and user preferences
 205  * and may involve loading shared libraries (DLLs), mapping
 206  * files into memory, or fopen()/fread() files.
 207  * It may also involve using static memory or database queries etc.
 208  * Several or all data items may be combined into one entity
 209  * (DLL, memory-mappable file).</p>
 210  *
 211  * <p>The data is always preceded by a header that includes
 212  * a <code>UDataInfo</code> structure.
 213  * The caller's <code>isAcceptable()</code> function is called to make
 214  * sure that the data is useful. It may be called several times if it
 215  * rejects the data and there is more than one location with data
 216  * matching the type and name.</p>
 217  *
 218  * <p>If <code>path==NULL</code>, then ICU data is loaded.
 219  * Otherwise, it is separated into a basename and a basename-less directory string.
 220  * The basename is used as the data package name, and the directory is
 221  * logically prepended to the ICU data directory string.</p>
 222  *
 223  * <p>For details about ICU data loading see the User Guide
 224  * Data Management chapter. (http://icu-project.org/userguide/icudata.html)</p>
 225  *
 226  * @param path Specifies an absolute path and/or a basename for the
 227  *             finding of the data in the file system.
 228  *             <code>NULL</code> for ICU data.
 229  * @param type A string that specifies the type of data to be loaded.
 230  *             For example, resource bundles are loaded with type "res",
 231  *             conversion tables with type "cnv".
 232  *             This may be <code>NULL</code> or empty.
 233  * @param name A string that specifies the name of the data.
 234  * @param isAcceptable This function is called to verify that loaded data
 235  *                     is useful for the client code. If it returns FALSE
 236  *                     for all data items, then <code>udata_openChoice()</code>
 237  *                     will return with an error.
 238  * @param context Arbitrary parameter to be passed into isAcceptable.
 239  * @param pErrorCode An ICU UErrorCode parameter. It must not be <code>NULL</code>.
 240  * @return A pointer (handle) to a data memory object, or <code>NULL</code>
 241  *         if an error occurs. Call <code>udata_getMemory()</code>
 242  *         to get a pointer to the actual data.
 243  * @stable ICU 2.0
 244  */
 245 U_STABLE UDataMemory * U_EXPORT2
 246 udata_openChoice(const char *path, const char *type, const char *name,
 247                  UDataMemoryIsAcceptable *isAcceptable, void *context,
 248                  UErrorCode *pErrorCode);
 249
 250 /**
 251  * Close the data memory.
 252  * This function must be called to allow the system to
 253  * release resources associated with this data memory.
 254  * @param pData The pointer to data memory object
 255  * @stable ICU 2.0
 256  */
 257 U_STABLE void U_EXPORT2
 258 udata_close(UDataMemory *pData);
 259
 260 #if U_SHOW_CPLUSPLUS_API
 261
 262 U_NAMESPACE_BEGIN
 263
 264 /**
 265  * \class LocalUDataMemoryPointer
 266  * "Smart pointer" class, closes a UDataMemory via udata_close().
 267  * For most methods see the LocalPointerBase base class.
 268  *
 269  * @see LocalPointerBase
 270  * @see LocalPointer
 271  * @stable ICU 4.4
 272  */
 273 U_DEFINE_LOCAL_OPEN_POINTER(LocalUDataMemoryPointer, UDataMemory, udata_close);
 274
 275 U_NAMESPACE_END
 276
 277 #endif
 278
 279 /**
 280  * Get the pointer to the actual data inside the data memory.
 281  * The data is read-only.
 282  * @param pData The pointer to data memory object
 283  * @stable ICU 2.0
 284  */
 285 U_STABLE const void * U_EXPORT2
 286 udata_getMemory(UDataMemory *pData);
 287
 288 /**
 289  * Get the information from the data memory header.
 290  * This allows to get access to the header containing
 291  * platform data properties etc. which is not part of
 292  * the data itself and can therefore not be accessed
 293  * via the pointer that <code>udata_getMemory()</code> returns.
 294  *
 295  * @param pData pointer to the data memory object
 296  * @param pInfo pointer to a UDataInfo object;
 297  *              its <code>size</code> field must be set correctly,
 298  *              typically to <code>sizeof(UDataInfo)</code>.
 299  *
 300  * <code>*pInfo</code> will be filled with the UDataInfo structure
 301  * in the data memory object. If this structure is smaller than
 302  * <code>pInfo->size</code>, then the <code>size</code> will be
 303  * adjusted and only part of the structure will be filled.
 304  * @stable ICU 2.0
 305  */
 306 U_STABLE void U_EXPORT2
 307 udata_getInfo(UDataMemory *pData, UDataInfo *pInfo);
 308
 309 /**
 310  * This function bypasses the normal ICU data loading process and
 311  * allows you to force ICU's system data to come out of a user-specified
 312  * area in memory.
 313  *
 314  * The format of this data is that of the icu common data file, as is
 315  * generated by the pkgdata tool with mode=common or mode=dll.
 316  * You can read in a whole common mode file and pass the address to the start of the
 317  * data, or (with the appropriate link options) pass in the pointer to
 318  * the data that has been loaded from a dll by the operating system,
 319  * as shown in this code:
 320  *
 321  *       extern const char U_IMPORT U_ICUDATA_ENTRY_POINT [];
 322  *        // U_ICUDATA_ENTRY_POINT is same as entry point specified to pkgdata tool
 323  *       UErrorCode  status = U_ZERO_ERROR;
 324  *
 325  *       udata_setCommonData(&U_ICUDATA_ENTRY_POINT, &status);
 326  *
 327  * It is important that the declaration be as above. The entry point
 328  * must not be declared as an extern void*.
 329  *
 330  * Starting with ICU 4.4, it is possible to set several data packages,
 331  * one per call to this function.
 332  * udata_open() will look for data in the multiple data packages in the order
 333  * in which they were set.
 334  * The position of the linked-in or default-name ICU .data package in the
 335  * search list depends on when the first data item is loaded that is not contained
 336  * in the already explicitly set packages.
 337  * If data was loaded implicitly before the first call to this function
 338  * (for example, via opening a converter, constructing a UnicodeString
 339  * from default-codepage data, using formatting or collation APIs, etc.),
 340  * then the default data will be first in the list.
 341  *
 342  * This function has no effect on application (non ICU) data.  See udata_setAppData()
 343  * for similar functionality for application data.
 344  *
 345  * @param data pointer to ICU common data
 346  * @param err outgoing error status <code>U_USING_DEFAULT_WARNING, U_UNSUPPORTED_ERROR</code>
 347  * @stable ICU 2.0
 348  */
 349 U_STABLE void U_EXPORT2
 350 udata_setCommonData(const void *data, UErrorCode *err);
 351
 352
 353 /**
 354  * This function bypasses the normal ICU data loading process for application-specific
 355  * data and allows you to force the it to come out of a user-specified
 356  * pointer.
 357  *
 358  * The format of this data is that of the icu common data file, like 'icudt26l.dat'
 359  * or the corresponding shared library (DLL) file.
 360  * The application must read in or otherwise construct an image of the data and then
 361  * pass the address of it to this function.
 362  *
 363  *
 364  * Warning:  setAppData will set a U_USING_DEFAULT_WARNING code if
 365  *           data with the specifed path that has already been opened, or
 366  *           if setAppData with the same path has already been called.
 367  *           Any such calls to setAppData will have no effect.
 368  *
 369  *
 370  * @param packageName the package name by which the application will refer
 371  *             to (open) this data
 372  * @param data pointer to the data
 373  * @param err outgoing error status <code>U_USING_DEFAULT_WARNING, U_UNSUPPORTED_ERROR</code>
 374  * @see udata_setCommonData
 375  * @stable ICU 2.0
 376  */
 377 U_STABLE void U_EXPORT2
 378 udata_setAppData(const char *packageName, const void *data, UErrorCode *err);
 379
 380 /**
 381  * Possible settings for udata_setFileAccess()
 382  * @see udata_setFileAccess
 383  * @stable ICU 3.4
 384  */
 385 typedef enum UDataFileAccess {
 386     /** ICU looks for data in single files first, then in packages. (default) */
 387     UDATA_FILES_FIRST,
 388     /** ICU only loads data from packages, not from single files. */
 389     UDATA_ONLY_PACKAGES,
 390     /** ICU loads data from packages first, and only from single files
 391         if the data cannot be found in a package. */
 392     UDATA_PACKAGES_FIRST,
 393     /** ICU does not access the file system for data loading. */
 394     UDATA_NO_FILES,
 395     /** An alias for the default access mode. */
 396     UDATA_DEFAULT_ACCESS = UDATA_FILES_FIRST,
 397     UDATA_FILE_ACCESS_COUNT
 398 } UDataFileAccess;
 399
 400 /**
 401  * This function may be called to control how ICU loads data. It must be called
 402  * before any ICU data is loaded, including application data loaded with
 403  * ures/ResourceBundle or udata APIs. This function is not multithread safe.
 404  * The results of calling it while other threads are loading data are undefined.
 405  * @param access The type of file access to be used
 406  * @param status Error code.
 407  * @see UDataFileAccess
 408  * @stable ICU 3.4
 409  */
 410 U_STABLE void U_EXPORT2
 411 udata_setFileAccess(UDataFileAccess access, UErrorCode *status);
 412
 413 U_CDECL_END
 414
 415 #endif