michael@0: /*
michael@0: ******************************************************************************
michael@0: *
michael@0: *   Copyright (C) 1999-2012, International Business Machines
michael@0: *   Corporation and others.  All Rights Reserved.
michael@0: *
michael@0: ******************************************************************************
michael@0: *   file name:  udata.h
michael@0: *   encoding:   US-ASCII
michael@0: *   tab size:   8 (not used)
michael@0: *   indentation:4
michael@0: *
michael@0: *   created on: 1999oct25
michael@0: *   created by: Markus W. Scherer
michael@0: */
michael@0: 
michael@0: #ifndef __UDATA_H__
michael@0: #define __UDATA_H__
michael@0: 
michael@0: #include "unicode/utypes.h"
michael@0: #include "unicode/localpointer.h"
michael@0: 
michael@0: U_CDECL_BEGIN
michael@0: 
michael@0: /**
michael@0:  * \file
michael@0:  * \brief C API: Data loading interface
michael@0:  *
michael@0:  * <h2>Information about data loading interface</h2>
michael@0:  *
michael@0:  * This API is used to find and efficiently load data for ICU and applications
michael@0:  * using ICU. It provides an abstract interface that specifies a data type and
michael@0:  * name to find and load the data. Normally this API is used by other ICU APIs
michael@0:  * to load required data out of the ICU data library, but it can be used to
michael@0:  * load data out of other places.
michael@0:  *
michael@0:  * See the User Guide Data Management chapter.
michael@0:  */
michael@0:  
michael@0: #ifndef U_HIDE_INTERNAL_API
michael@0: /**
michael@0:  * Character used to separate package names from tree names 
michael@0:  * @internal ICU 3.0
michael@0:  */
michael@0: #define U_TREE_SEPARATOR '-'
michael@0: 
michael@0: /**
michael@0:  * String used to separate package names from tree names 
michael@0:  * @internal ICU 3.0
michael@0:  */
michael@0: #define U_TREE_SEPARATOR_STRING "-"
michael@0: 
michael@0: /**
michael@0:  * Character used to separate parts of entry names
michael@0:  * @internal ICU 3.0
michael@0:  */
michael@0: #define U_TREE_ENTRY_SEP_CHAR '/'
michael@0: 
michael@0: /**
michael@0:  * String used to separate parts of entry names
michael@0:  * @internal ICU 3.0
michael@0:  */
michael@0: #define U_TREE_ENTRY_SEP_STRING "/"
michael@0: 
michael@0: /**
michael@0:  * Alias for standard ICU data 
michael@0:  * @internal ICU 3.0
michael@0:  */
michael@0: #define U_ICUDATA_ALIAS "ICUDATA"
michael@0: 
michael@0: #endif /* U_HIDE_INTERNAL_API */
michael@0: 
michael@0: /**
michael@0:  * UDataInfo contains the properties about the requested data.
michael@0:  * This is meta data.
michael@0:  *
michael@0:  * <p>This structure may grow in the future, indicated by the
michael@0:  * <code>size</code> field.</p>
michael@0:  *
michael@0:  * <p>The platform data property fields help determine if a data
michael@0:  * file can be efficiently used on a given machine.
michael@0:  * The particular fields are of importance only if the data
michael@0:  * is affected by the properties - if there is integer data
michael@0:  * with word sizes > 1 byte, char* text, or UChar* text.</p>
michael@0:  *
michael@0:  * <p>The implementation for the <code>udata_open[Choice]()</code>
michael@0:  * functions may reject data based on the value in <code>isBigEndian</code>.
michael@0:  * No other field is used by the <code>udata</code> API implementation.</p>
michael@0:  *
michael@0:  * <p>The <code>dataFormat</code> may be used to identify
michael@0:  * the kind of data, e.g. a converter table.</p>
michael@0:  *
michael@0:  * <p>The <code>formatVersion</code> field should be used to
michael@0:  * make sure that the format can be interpreted.
michael@0:  * I may be a good idea to check only for the one or two highest
michael@0:  * of the version elements to allow the data memory to
michael@0:  * get more or somewhat rearranged contents, for as long
michael@0:  * as the using code can still interpret the older contents.</p>
michael@0:  *
michael@0:  * <p>The <code>dataVersion</code> field is intended to be a
michael@0:  * common place to store the source version of the data;
michael@0:  * for data from the Unicode character database, this could
michael@0:  * reflect the Unicode version.</p>
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: typedef struct {
michael@0:     /** sizeof(UDataInfo)
michael@0:      *  @stable ICU 2.0 */
michael@0:     uint16_t size;
michael@0: 
michael@0:     /** unused, set to 0 
michael@0:      *  @stable ICU 2.0*/
michael@0:     uint16_t reservedWord;
michael@0: 
michael@0:     /* platform data properties */
michael@0:     /** 0 for little-endian machine, 1 for big-endian
michael@0:      *  @stable ICU 2.0 */
michael@0:     uint8_t isBigEndian;
michael@0: 
michael@0:     /** see U_CHARSET_FAMILY values in utypes.h 
michael@0:      *  @stable ICU 2.0*/
michael@0:     uint8_t charsetFamily;
michael@0: 
michael@0:     /** sizeof(UChar), one of { 1, 2, 4 } 
michael@0:      *  @stable ICU 2.0*/
michael@0:     uint8_t sizeofUChar;
michael@0: 
michael@0:     /** unused, set to 0 
michael@0:      *  @stable ICU 2.0*/
michael@0:     uint8_t reservedByte;
michael@0: 
michael@0:     /** data format identifier 
michael@0:      *  @stable ICU 2.0*/
michael@0:     uint8_t dataFormat[4];
michael@0: 
michael@0:     /** versions: [0] major [1] minor [2] milli [3] micro 
michael@0:      *  @stable ICU 2.0*/
michael@0:     uint8_t formatVersion[4];
michael@0: 
michael@0:     /** versions: [0] major [1] minor [2] milli [3] micro 
michael@0:      *  @stable ICU 2.0*/
michael@0:     uint8_t dataVersion[4];
michael@0: } UDataInfo;
michael@0: 
michael@0: /* API for reading data -----------------------------------------------------*/
michael@0: 
michael@0: /**
michael@0:  * Forward declaration of the data memory type.
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: typedef struct UDataMemory UDataMemory;
michael@0: 
michael@0: /**
michael@0:  * Callback function for udata_openChoice().
michael@0:  * @param context parameter passed into <code>udata_openChoice()</code>.
michael@0:  * @param type The type of the data as passed into <code>udata_openChoice()</code>.
michael@0:  *             It may be <code>NULL</code>.
michael@0:  * @param name The name of the data as passed into <code>udata_openChoice()</code>.
michael@0:  * @param pInfo A pointer to the <code>UDataInfo</code> structure
michael@0:  *              of data that has been loaded and will be returned
michael@0:  *              by <code>udata_openChoice()</code> if this function
michael@0:  *              returns <code>TRUE</code>.
michael@0:  * @return TRUE if the current data memory is acceptable
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: typedef UBool U_CALLCONV
michael@0: UDataMemoryIsAcceptable(void *context,
michael@0:                         const char *type, const char *name,
michael@0:                         const UDataInfo *pInfo);
michael@0: 
michael@0: 
michael@0: /**
michael@0:  * Convenience function.
michael@0:  * This function works the same as <code>udata_openChoice</code>
michael@0:  * except that any data that matches the type and name
michael@0:  * is assumed to be acceptable.
michael@0:  * @param path Specifies an absolute path and/or a basename for the
michael@0:  *             finding of the data in the file system.
michael@0:  *             <code>NULL</code> for ICU data.
michael@0:  * @param type A string that specifies the type of data to be loaded.
michael@0:  *             For example, resource bundles are loaded with type "res",
michael@0:  *             conversion tables with type "cnv".
michael@0:  *             This may be <code>NULL</code> or empty.
michael@0:  * @param name A string that specifies the name of the data.
michael@0:  * @param pErrorCode An ICU UErrorCode parameter. It must not be <code>NULL</code>.
michael@0:  * @return A pointer (handle) to a data memory object, or <code>NULL</code>
michael@0:  *         if an error occurs. Call <code>udata_getMemory()</code>
michael@0:  *         to get a pointer to the actual data.
michael@0:  *
michael@0:  * @see udata_openChoice
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE UDataMemory * U_EXPORT2
michael@0: udata_open(const char *path, const char *type, const char *name,
michael@0:            UErrorCode *pErrorCode);
michael@0: 
michael@0: /**
michael@0:  * Data loading function.
michael@0:  * This function is used to find and load efficiently data for
michael@0:  * ICU and applications using ICU.
michael@0:  * It provides an abstract interface that allows to specify a data
michael@0:  * type and name to find and load the data.
michael@0:  *
michael@0:  * <p>The implementation depends on platform properties and user preferences
michael@0:  * and may involve loading shared libraries (DLLs), mapping
michael@0:  * files into memory, or fopen()/fread() files.
michael@0:  * It may also involve using static memory or database queries etc.
michael@0:  * Several or all data items may be combined into one entity
michael@0:  * (DLL, memory-mappable file).</p>
michael@0:  *
michael@0:  * <p>The data is always preceded by a header that includes
michael@0:  * a <code>UDataInfo</code> structure.
michael@0:  * The caller's <code>isAcceptable()</code> function is called to make
michael@0:  * sure that the data is useful. It may be called several times if it
michael@0:  * rejects the data and there is more than one location with data
michael@0:  * matching the type and name.</p>
michael@0:  *
michael@0:  * <p>If <code>path==NULL</code>, then ICU data is loaded.
michael@0:  * Otherwise, it is separated into a basename and a basename-less directory string.
michael@0:  * The basename is used as the data package name, and the directory is
michael@0:  * logically prepended to the ICU data directory string.</p>
michael@0:  *
michael@0:  * <p>For details about ICU data loading see the User Guide
michael@0:  * Data Management chapter. (http://icu-project.org/userguide/icudata.html)</p>
michael@0:  *
michael@0:  * @param path Specifies an absolute path and/or a basename for the
michael@0:  *             finding of the data in the file system.
michael@0:  *             <code>NULL</code> for ICU data.
michael@0:  * @param type A string that specifies the type of data to be loaded.
michael@0:  *             For example, resource bundles are loaded with type "res",
michael@0:  *             conversion tables with type "cnv".
michael@0:  *             This may be <code>NULL</code> or empty.
michael@0:  * @param name A string that specifies the name of the data.
michael@0:  * @param isAcceptable This function is called to verify that loaded data
michael@0:  *                     is useful for the client code. If it returns FALSE
michael@0:  *                     for all data items, then <code>udata_openChoice()</code>
michael@0:  *                     will return with an error.
michael@0:  * @param context Arbitrary parameter to be passed into isAcceptable.
michael@0:  * @param pErrorCode An ICU UErrorCode parameter. It must not be <code>NULL</code>.
michael@0:  * @return A pointer (handle) to a data memory object, or <code>NULL</code>
michael@0:  *         if an error occurs. Call <code>udata_getMemory()</code>
michael@0:  *         to get a pointer to the actual data.
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE UDataMemory * U_EXPORT2
michael@0: udata_openChoice(const char *path, const char *type, const char *name,
michael@0:                  UDataMemoryIsAcceptable *isAcceptable, void *context,
michael@0:                  UErrorCode *pErrorCode);
michael@0: 
michael@0: /**
michael@0:  * Close the data memory.
michael@0:  * This function must be called to allow the system to
michael@0:  * release resources associated with this data memory.
michael@0:  * @param pData The pointer to data memory object
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE void U_EXPORT2
michael@0: udata_close(UDataMemory *pData);
michael@0: 
michael@0: #if U_SHOW_CPLUSPLUS_API
michael@0: 
michael@0: U_NAMESPACE_BEGIN
michael@0: 
michael@0: /**
michael@0:  * \class LocalUDataMemoryPointer
michael@0:  * "Smart pointer" class, closes a UDataMemory via udata_close().
michael@0:  * For most methods see the LocalPointerBase base class.
michael@0:  *
michael@0:  * @see LocalPointerBase
michael@0:  * @see LocalPointer
michael@0:  * @stable ICU 4.4
michael@0:  */
michael@0: U_DEFINE_LOCAL_OPEN_POINTER(LocalUDataMemoryPointer, UDataMemory, udata_close);
michael@0: 
michael@0: U_NAMESPACE_END
michael@0: 
michael@0: #endif
michael@0: 
michael@0: /**
michael@0:  * Get the pointer to the actual data inside the data memory.
michael@0:  * The data is read-only.
michael@0:  * @param pData The pointer to data memory object
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE const void * U_EXPORT2
michael@0: udata_getMemory(UDataMemory *pData);
michael@0: 
michael@0: /**
michael@0:  * Get the information from the data memory header.
michael@0:  * This allows to get access to the header containing
michael@0:  * platform data properties etc. which is not part of
michael@0:  * the data itself and can therefore not be accessed
michael@0:  * via the pointer that <code>udata_getMemory()</code> returns.
michael@0:  *
michael@0:  * @param pData pointer to the data memory object
michael@0:  * @param pInfo pointer to a UDataInfo object;
michael@0:  *              its <code>size</code> field must be set correctly,
michael@0:  *              typically to <code>sizeof(UDataInfo)</code>.
michael@0:  *
michael@0:  * <code>*pInfo</code> will be filled with the UDataInfo structure
michael@0:  * in the data memory object. If this structure is smaller than
michael@0:  * <code>pInfo->size</code>, then the <code>size</code> will be
michael@0:  * adjusted and only part of the structure will be filled.
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE void U_EXPORT2
michael@0: udata_getInfo(UDataMemory *pData, UDataInfo *pInfo);
michael@0: 
michael@0: /**
michael@0:  * This function bypasses the normal ICU data loading process and
michael@0:  * allows you to force ICU's system data to come out of a user-specified
michael@0:  * area in memory.
michael@0:  *
michael@0:  * The format of this data is that of the icu common data file, as is
michael@0:  * generated by the pkgdata tool with mode=common or mode=dll.
michael@0:  * You can read in a whole common mode file and pass the address to the start of the
michael@0:  * data, or (with the appropriate link options) pass in the pointer to
michael@0:  * the data that has been loaded from a dll by the operating system,
michael@0:  * as shown in this code:
michael@0:  *
michael@0:  *       extern const char U_IMPORT U_ICUDATA_ENTRY_POINT [];
michael@0:  *        // U_ICUDATA_ENTRY_POINT is same as entry point specified to pkgdata tool
michael@0:  *       UErrorCode  status = U_ZERO_ERROR;
michael@0:  *
michael@0:  *       udata_setCommonData(&U_ICUDATA_ENTRY_POINT, &status);
michael@0:  *
michael@0:  * It is important that the declaration be as above. The entry point
michael@0:  * must not be declared as an extern void*.
michael@0:  *
michael@0:  * Starting with ICU 4.4, it is possible to set several data packages,
michael@0:  * one per call to this function.
michael@0:  * udata_open() will look for data in the multiple data packages in the order
michael@0:  * in which they were set.
michael@0:  * The position of the linked-in or default-name ICU .data package in the
michael@0:  * search list depends on when the first data item is loaded that is not contained
michael@0:  * in the already explicitly set packages.
michael@0:  * If data was loaded implicitly before the first call to this function
michael@0:  * (for example, via opening a converter, constructing a UnicodeString
michael@0:  * from default-codepage data, using formatting or collation APIs, etc.),
michael@0:  * then the default data will be first in the list.
michael@0:  *
michael@0:  * This function has no effect on application (non ICU) data.  See udata_setAppData()
michael@0:  * for similar functionality for application data.
michael@0:  *
michael@0:  * @param data pointer to ICU common data
michael@0:  * @param err outgoing error status <code>U_USING_DEFAULT_WARNING, U_UNSUPPORTED_ERROR</code>
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE void U_EXPORT2
michael@0: udata_setCommonData(const void *data, UErrorCode *err);
michael@0: 
michael@0: 
michael@0: /**
michael@0:  * This function bypasses the normal ICU data loading process for application-specific
michael@0:  * data and allows you to force the it to come out of a user-specified
michael@0:  * pointer.
michael@0:  *
michael@0:  * The format of this data is that of the icu common data file, like 'icudt26l.dat'
michael@0:  * or the corresponding shared library (DLL) file.
michael@0:  * The application must read in or otherwise construct an image of the data and then
michael@0:  * pass the address of it to this function.
michael@0:  *
michael@0:  *
michael@0:  * Warning:  setAppData will set a U_USING_DEFAULT_WARNING code if
michael@0:  *           data with the specifed path that has already been opened, or
michael@0:  *           if setAppData with the same path has already been called.
michael@0:  *           Any such calls to setAppData will have no effect.
michael@0:  *
michael@0:  *
michael@0:  * @param packageName the package name by which the application will refer
michael@0:  *             to (open) this data
michael@0:  * @param data pointer to the data
michael@0:  * @param err outgoing error status <code>U_USING_DEFAULT_WARNING, U_UNSUPPORTED_ERROR</code>
michael@0:  * @see udata_setCommonData
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE void U_EXPORT2
michael@0: udata_setAppData(const char *packageName, const void *data, UErrorCode *err);
michael@0: 
michael@0: /**
michael@0:  * Possible settings for udata_setFileAccess()
michael@0:  * @see udata_setFileAccess
michael@0:  * @stable ICU 3.4
michael@0:  */
michael@0: typedef enum UDataFileAccess {
michael@0:     /** ICU looks for data in single files first, then in packages. (default) @stable ICU 3.4 */
michael@0:     UDATA_FILES_FIRST,
michael@0:     /** An alias for the default access mode. @stable ICU 3.4 */
michael@0:     UDATA_DEFAULT_ACCESS = UDATA_FILES_FIRST,
michael@0:     /** ICU only loads data from packages, not from single files. @stable ICU 3.4 */
michael@0:     UDATA_ONLY_PACKAGES,
michael@0:     /** ICU loads data from packages first, and only from single files
michael@0:         if the data cannot be found in a package. @stable ICU 3.4 */
michael@0:     UDATA_PACKAGES_FIRST,
michael@0:     /** ICU does not access the file system for data loading. @stable ICU 3.4 */
michael@0:     UDATA_NO_FILES,
michael@0:     /** Number of real UDataFileAccess values. @stable ICU 3.4 */
michael@0:     UDATA_FILE_ACCESS_COUNT
michael@0: } UDataFileAccess;
michael@0: 
michael@0: /**
michael@0:  * This function may be called to control how ICU loads data. It must be called
michael@0:  * before any ICU data is loaded, including application data loaded with 
michael@0:  * ures/ResourceBundle or udata APIs. This function is not multithread safe.  
michael@0:  * The results of calling it while other threads are loading data are undefined.
michael@0:  * @param access The type of file access to be used
michael@0:  * @param status Error code.
michael@0:  * @see UDataFileAccess
michael@0:  * @stable ICU 3.4 
michael@0:  */
michael@0: U_STABLE void U_EXPORT2
michael@0: udata_setFileAccess(UDataFileAccess access, UErrorCode *status);
michael@0: 
michael@0: U_CDECL_END
michael@0: 
michael@0: #endif