// © 2016 and later: Unicode, Inc. and others. // License & terms of use: http://www.unicode.org/copyright.html /* ****************************************************************************** * * Copyright (C) 1999-2014, International Business Machines * Corporation and others. All Rights Reserved. * ****************************************************************************** * file name: udata.h * encoding: UTF-8 * tab size: 8 (not used) * indentation:4 * * created on: 1999oct25 * created by: Markus W. Scherer */ #ifndef __UDATA_H__ #define __UDATA_H__ #include "unicode/utypes.h" #if U_SHOW_CPLUSPLUS_API #include "unicode/localpointer.h" #endif // U_SHOW_CPLUSPLUS_API U_CDECL_BEGIN /** * \file * \brief C API: Data loading interface * *

Information about data loading interface

* * This API is used to find and efficiently load data for ICU and applications * using ICU. It provides an abstract interface that specifies a data type and * name to find and load the data. Normally this API is used by other ICU APIs * to load required data out of the ICU data library, but it can be used to * load data out of other places. * * See the User Guide Data Management chapter. */ #ifndef U_HIDE_INTERNAL_API /** * Character used to separate package names from tree names * @internal ICU 3.0 */ #define U_TREE_SEPARATOR '-' /** * String used to separate package names from tree names * @internal ICU 3.0 */ #define U_TREE_SEPARATOR_STRING "-" /** * Character used to separate parts of entry names * @internal ICU 3.0 */ #define U_TREE_ENTRY_SEP_CHAR '/' /** * String used to separate parts of entry names * @internal ICU 3.0 */ #define U_TREE_ENTRY_SEP_STRING "/" /** * Alias for standard ICU data * @internal ICU 3.0 */ #define U_ICUDATA_ALIAS "ICUDATA" #endif /* U_HIDE_INTERNAL_API */ /** * UDataInfo contains the properties about the requested data. * This is meta data. * *

This structure may grow in the future, indicated by the * size field.

* *

ICU data must be at least 8-aligned, and should be 16-aligned. * The UDataInfo struct begins 4 bytes after the start of the data item, * so it is 4-aligned. * *

The platform data property fields help determine if a data * file can be efficiently used on a given machine. * The particular fields are of importance only if the data * is affected by the properties - if there is integer data * with word sizes > 1 byte, char* text, or UChar* text.

* *

The implementation for the udata_open[Choice]() * functions may reject data based on the value in isBigEndian. * No other field is used by the udata API implementation.

* *

The dataFormat may be used to identify * the kind of data, e.g. a converter table.

* *

The formatVersion field should be used to * make sure that the format can be interpreted. * It may be a good idea to check only for the one or two highest * of the version elements to allow the data memory to * get more or somewhat rearranged contents, for as long * as the using code can still interpret the older contents.

* *

The dataVersion field is intended to be a * common place to store the source version of the data; * for data from the Unicode character database, this could * reflect the Unicode version.

* * @stable ICU 2.0 */ typedef struct { /** sizeof(UDataInfo) * @stable ICU 2.0 */ uint16_t size; /** unused, set to 0 * @stable ICU 2.0*/ uint16_t reservedWord; /* platform data properties */ /** 0 for little-endian machine, 1 for big-endian * @stable ICU 2.0 */ uint8_t isBigEndian; /** see U_CHARSET_FAMILY values in utypes.h * @stable ICU 2.0*/ uint8_t charsetFamily; /** sizeof(UChar), one of { 1, 2, 4 } * @stable ICU 2.0*/ uint8_t sizeofUChar; /** unused, set to 0 * @stable ICU 2.0*/ uint8_t reservedByte; /** data format identifier * @stable ICU 2.0*/ uint8_t dataFormat[4]; /** versions: [0] major [1] minor [2] milli [3] micro * @stable ICU 2.0*/ uint8_t formatVersion[4]; /** versions: [0] major [1] minor [2] milli [3] micro * @stable ICU 2.0*/ uint8_t dataVersion[4]; } UDataInfo; /* API for reading data -----------------------------------------------------*/ /** * Forward declaration of the data memory type. * @stable ICU 2.0 */ typedef struct UDataMemory UDataMemory; /** * Callback function for udata_openChoice(). * @param context parameter passed into udata_openChoice(). * @param type The type of the data as passed into udata_openChoice(). * It may be NULL. * @param name The name of the data as passed into udata_openChoice(). * @param pInfo A pointer to the UDataInfo structure * of data that has been loaded and will be returned * by udata_openChoice() if this function * returns true. * @return true if the current data memory is acceptable * @stable ICU 2.0 */ typedef UBool U_CALLCONV UDataMemoryIsAcceptable(void *context, const char *type, const char *name, const UDataInfo *pInfo); /** * Convenience function. * This function works the same as udata_openChoice * except that any data that matches the type and name * is assumed to be acceptable. * @param path Specifies an absolute path and/or a basename for the * finding of the data in the file system. * NULL for ICU data. * @param type A string that specifies the type of data to be loaded. * For example, resource bundles are loaded with type "res", * conversion tables with type "cnv". * This may be NULL or empty. * @param name A string that specifies the name of the data. * @param pErrorCode An ICU UErrorCode parameter. It must not be NULL. * @return A pointer (handle) to a data memory object, or NULL * if an error occurs. Call udata_getMemory() * to get a pointer to the actual data. * * @see udata_openChoice * @stable ICU 2.0 */ U_CAPI UDataMemory * U_EXPORT2 udata_open(const char *path, const char *type, const char *name, UErrorCode *pErrorCode); /** * Data loading function. * This function is used to find and load efficiently data for * ICU and applications using ICU. * It provides an abstract interface that allows to specify a data * type and name to find and load the data. * *

The implementation depends on platform properties and user preferences * and may involve loading shared libraries (DLLs), mapping * files into memory, or fopen()/fread() files. * It may also involve using static memory or database queries etc. * Several or all data items may be combined into one entity * (DLL, memory-mappable file).

* *

The data is always preceded by a header that includes * a UDataInfo structure. * The caller's isAcceptable() function is called to make * sure that the data is useful. It may be called several times if it * rejects the data and there is more than one location with data * matching the type and name.

* *

If path==NULL, then ICU data is loaded. * Otherwise, it is separated into a basename and a basename-less directory string. * The basename is used as the data package name, and the directory is * logically prepended to the ICU data directory string.

* *

For details about ICU data loading see the User Guide * Data Management chapter. (https://unicode-org.github.io/icu/userguide/icu_data/)

* * @param path Specifies an absolute path and/or a basename for the * finding of the data in the file system. * NULL for ICU data. * @param type A string that specifies the type of data to be loaded. * For example, resource bundles are loaded with type "res", * conversion tables with type "cnv". * This may be NULL or empty. * @param name A string that specifies the name of the data. * @param isAcceptable This function is called to verify that loaded data * is useful for the client code. If it returns false * for all data items, then udata_openChoice() * will return with an error. * @param context Arbitrary parameter to be passed into isAcceptable. * @param pErrorCode An ICU UErrorCode parameter. It must not be NULL. * @return A pointer (handle) to a data memory object, or NULL * if an error occurs. Call udata_getMemory() * to get a pointer to the actual data. * @stable ICU 2.0 */ U_CAPI UDataMemory * U_EXPORT2 udata_openChoice(const char *path, const char *type, const char *name, UDataMemoryIsAcceptable *isAcceptable, void *context, UErrorCode *pErrorCode); /** * Close the data memory. * This function must be called to allow the system to * release resources associated with this data memory. * @param pData The pointer to data memory object * @stable ICU 2.0 */ U_CAPI void U_EXPORT2 udata_close(UDataMemory *pData); /** * Get the pointer to the actual data inside the data memory. * The data is read-only. * * ICU data must be at least 8-aligned, and should be 16-aligned. * * @param pData The pointer to data memory object * @stable ICU 2.0 */ U_CAPI const void * U_EXPORT2 udata_getMemory(UDataMemory *pData); /** * Get the information from the data memory header. * This allows to get access to the header containing * platform data properties etc. which is not part of * the data itself and can therefore not be accessed * via the pointer that udata_getMemory() returns. * * @param pData pointer to the data memory object * @param pInfo pointer to a UDataInfo object; * its size field must be set correctly, * typically to sizeof(UDataInfo). * * *pInfo will be filled with the UDataInfo structure * in the data memory object. If this structure is smaller than * pInfo->size, then the size will be * adjusted and only part of the structure will be filled. * @stable ICU 2.0 */ U_CAPI void U_EXPORT2 udata_getInfo(UDataMemory *pData, UDataInfo *pInfo); /** * This function bypasses the normal ICU data loading process and * allows you to force ICU's system data to come out of a user-specified * area in memory. * * ICU data must be at least 8-aligned, and should be 16-aligned. * See https://unicode-org.github.io/icu/userguide/icudata * * The format of this data is that of the icu common data file, as is * generated by the pkgdata tool with mode=common or mode=dll. * You can read in a whole common mode file and pass the address to the start of the * data, or (with the appropriate link options) pass in the pointer to * the data that has been loaded from a dll by the operating system, * as shown in this code: * * extern const char U_IMPORT U_ICUDATA_ENTRY_POINT []; * // U_ICUDATA_ENTRY_POINT is same as entry point specified to pkgdata tool * UErrorCode status = U_ZERO_ERROR; * * udata_setCommonData(&U_ICUDATA_ENTRY_POINT, &status); * * It is important that the declaration be as above. The entry point * must not be declared as an extern void*. * * Starting with ICU 4.4, it is possible to set several data packages, * one per call to this function. * udata_open() will look for data in the multiple data packages in the order * in which they were set. * The position of the linked-in or default-name ICU .data package in the * search list depends on when the first data item is loaded that is not contained * in the already explicitly set packages. * If data was loaded implicitly before the first call to this function * (for example, via opening a converter, constructing a UnicodeString * from default-codepage data, using formatting or collation APIs, etc.), * then the default data will be first in the list. * * This function has no effect on application (non ICU) data. See udata_setAppData() * for similar functionality for application data. * * @param data pointer to ICU common data * @param err outgoing error status U_USING_DEFAULT_WARNING, U_UNSUPPORTED_ERROR * @stable ICU 2.0 */ U_CAPI void U_EXPORT2 udata_setCommonData(const void *data, UErrorCode *err); /** * This function bypasses the normal ICU data loading process for application-specific * data and allows you to force the it to come out of a user-specified * pointer. * * ICU data must be at least 8-aligned, and should be 16-aligned. * See https://unicode-org.github.io/icu/userguide/icudata * * The format of this data is that of the icu common data file, like 'icudt26l.dat' * or the corresponding shared library (DLL) file. * The application must read in or otherwise construct an image of the data and then * pass the address of it to this function. * * * Warning: setAppData will set a U_USING_DEFAULT_WARNING code if * data with the specified path that has already been opened, or * if setAppData with the same path has already been called. * Any such calls to setAppData will have no effect. * * * @param packageName the package name by which the application will refer * to (open) this data * @param data pointer to the data * @param err outgoing error status U_USING_DEFAULT_WARNING, U_UNSUPPORTED_ERROR * @see udata_setCommonData * @stable ICU 2.0 */ U_CAPI void U_EXPORT2 udata_setAppData(const char *packageName, const void *data, UErrorCode *err); /** * Possible settings for udata_setFileAccess() * @see udata_setFileAccess * @stable ICU 3.4 */ typedef enum UDataFileAccess { /** ICU looks for data in single files first, then in packages. (default) @stable ICU 3.4 */ UDATA_FILES_FIRST, /** An alias for the default access mode. @stable ICU 3.4 */ UDATA_DEFAULT_ACCESS = UDATA_FILES_FIRST, /** ICU only loads data from packages, not from single files. @stable ICU 3.4 */ UDATA_ONLY_PACKAGES, /** ICU loads data from packages first, and only from single files if the data cannot be found in a package. @stable ICU 3.4 */ UDATA_PACKAGES_FIRST, /** ICU does not access the file system for data loading. @stable ICU 3.4 */ UDATA_NO_FILES, #ifndef U_HIDE_DEPRECATED_API /** * Number of real UDataFileAccess values. * @deprecated ICU 58 The numeric value may change over time, see ICU ticket #12420. */ UDATA_FILE_ACCESS_COUNT #endif // U_HIDE_DEPRECATED_API } UDataFileAccess; /** * This function may be called to control how ICU loads data. It must be called * before any ICU data is loaded, including application data loaded with * ures/ResourceBundle or udata APIs. This function is not multithread safe. * The results of calling it while other threads are loading data are undefined. * @param access The type of file access to be used * @param status Error code. * @see UDataFileAccess * @stable ICU 3.4 */ U_CAPI void U_EXPORT2 udata_setFileAccess(UDataFileAccess access, UErrorCode *status); U_CDECL_END #if U_SHOW_CPLUSPLUS_API U_NAMESPACE_BEGIN /** * \class LocalUDataMemoryPointer * "Smart pointer" class, closes a UDataMemory via udata_close(). * For most methods see the LocalPointerBase base class. * * @see LocalPointerBase * @see LocalPointer * @stable ICU 4.4 */ U_DEFINE_LOCAL_OPEN_POINTER(LocalUDataMemoryPointer, UDataMemory, udata_close); U_NAMESPACE_END #endif // U_SHOW_CPLUSPLUS_API #endif