michael@0: /*
michael@0: *******************************************************************************
michael@0: *
michael@0: *   Copyright (C) 2008-2011, International Business Machines
michael@0: *   Corporation, Google and others.  All Rights Reserved.
michael@0: *
michael@0: *******************************************************************************
michael@0: */
michael@0: /*
michael@0:  * Author : eldawy@google.com (Mohamed Eldawy)
michael@0:  * ucnvsel.h
michael@0:  *
michael@0:  * Purpose: To generate a list of encodings capable of handling
michael@0:  * a given Unicode text
michael@0:  *
michael@0:  * Started 09-April-2008
michael@0:  */
michael@0: 
michael@0: #ifndef __ICU_UCNV_SEL_H__
michael@0: #define __ICU_UCNV_SEL_H__
michael@0: 
michael@0: #include "unicode/utypes.h"
michael@0: 
michael@0: #if !UCONFIG_NO_CONVERSION
michael@0: 
michael@0: #include "unicode/uset.h"
michael@0: #include "unicode/utf16.h"
michael@0: #include "unicode/uenum.h"
michael@0: #include "unicode/ucnv.h"
michael@0: #include "unicode/localpointer.h"
michael@0: 
michael@0: /**
michael@0:  * \file
michael@0:  *
michael@0:  * A converter selector is built with a set of encoding/charset names
michael@0:  * and given an input string returns the set of names of the
michael@0:  * corresponding converters which can convert the string.
michael@0:  *
michael@0:  * A converter selector can be serialized into a buffer and reopened
michael@0:  * from the serialized form.
michael@0:  */
michael@0: 
michael@0: /**
michael@0:  * @{
michael@0:  * The selector data structure
michael@0:  */
michael@0: struct UConverterSelector;
michael@0: typedef struct UConverterSelector UConverterSelector;
michael@0: /** @} */
michael@0: 
michael@0: /**
michael@0:  * Open a selector.
michael@0:  * If converterListSize is 0, build for all available converters.
michael@0:  * If excludedCodePoints is NULL, don't exclude any code points.
michael@0:  *
michael@0:  * @param converterList a pointer to encoding names needed to be involved. 
michael@0:  *                      Can be NULL if converterListSize==0.
michael@0:  *                      The list and the names will be cloned, and the caller
michael@0:  *                      retains ownership of the original.
michael@0:  * @param converterListSize number of encodings in above list.
michael@0:  *                          If 0, builds a selector for all available converters.
michael@0:  * @param excludedCodePoints a set of code points to be excluded from consideration.
michael@0:  *                           That is, excluded code points in a string do not change
michael@0:  *                           the selection result. (They might be handled by a callback.)
michael@0:  *                           Use NULL to exclude nothing.
michael@0:  * @param whichSet what converter set to use? Use this to determine whether
michael@0:  *                 to consider only roundtrip mappings or also fallbacks.
michael@0:  * @param status an in/out ICU UErrorCode
michael@0:  * @return the new selector
michael@0:  *
michael@0:  * @stable ICU 4.2
michael@0:  */
michael@0: U_STABLE UConverterSelector* U_EXPORT2
michael@0: ucnvsel_open(const char* const*  converterList, int32_t converterListSize,
michael@0:              const USet* excludedCodePoints,
michael@0:              const UConverterUnicodeSet whichSet, UErrorCode* status);
michael@0: 
michael@0: /**
michael@0:  * Closes a selector.
michael@0:  * If any Enumerations were returned by ucnv_select*, they become invalid.
michael@0:  * They can be closed before or after calling ucnv_closeSelector,
michael@0:  * but should never be used after the selector is closed.
michael@0:  *
michael@0:  * @see ucnv_selectForString
michael@0:  * @see ucnv_selectForUTF8
michael@0:  *
michael@0:  * @param sel selector to close
michael@0:  *
michael@0:  * @stable ICU 4.2
michael@0:  */
michael@0: U_STABLE void U_EXPORT2
michael@0: ucnvsel_close(UConverterSelector *sel);
michael@0: 
michael@0: #if U_SHOW_CPLUSPLUS_API
michael@0: 
michael@0: U_NAMESPACE_BEGIN
michael@0: 
michael@0: /**
michael@0:  * \class LocalUConverterSelectorPointer
michael@0:  * "Smart pointer" class, closes a UConverterSelector via ucnvsel_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(LocalUConverterSelectorPointer, UConverterSelector, ucnvsel_close);
michael@0: 
michael@0: U_NAMESPACE_END
michael@0: 
michael@0: #endif
michael@0: 
michael@0: /**
michael@0:  * Open a selector from its serialized form.
michael@0:  * The buffer must remain valid and unchanged for the lifetime of the selector.
michael@0:  * This is much faster than creating a selector from scratch.
michael@0:  * Using a serialized form from a different machine (endianness/charset) is supported.
michael@0:  *
michael@0:  * @param buffer pointer to the serialized form of a converter selector;
michael@0:  *               must be 32-bit-aligned
michael@0:  * @param length the capacity of this buffer (can be equal to or larger than
michael@0:  *               the actual data length)
michael@0:  * @param status an in/out ICU UErrorCode
michael@0:  * @return the new selector
michael@0:  *
michael@0:  * @stable ICU 4.2
michael@0:  */
michael@0: U_STABLE UConverterSelector* U_EXPORT2
michael@0: ucnvsel_openFromSerialized(const void* buffer, int32_t length, UErrorCode* status);
michael@0: 
michael@0: /**
michael@0:  * Serialize a selector into a linear buffer.
michael@0:  * The serialized form is portable to different machines.
michael@0:  *
michael@0:  * @param sel selector to consider
michael@0:  * @param buffer pointer to 32-bit-aligned memory to be filled with the
michael@0:  *               serialized form of this converter selector
michael@0:  * @param bufferCapacity the capacity of this buffer
michael@0:  * @param status an in/out ICU UErrorCode
michael@0:  * @return the required buffer capacity to hold serialize data (even if the call fails
michael@0:  *         with a U_BUFFER_OVERFLOW_ERROR, it will return the required capacity)
michael@0:  *
michael@0:  * @stable ICU 4.2
michael@0:  */
michael@0: U_STABLE int32_t U_EXPORT2
michael@0: ucnvsel_serialize(const UConverterSelector* sel,
michael@0:                   void* buffer, int32_t bufferCapacity, UErrorCode* status);
michael@0: 
michael@0: /**
michael@0:  * Select converters that can map all characters in a UTF-16 string,
michael@0:  * ignoring the excluded code points.
michael@0:  *
michael@0:  * @param sel a selector
michael@0:  * @param s UTF-16 string
michael@0:  * @param length length of the string, or -1 if NUL-terminated
michael@0:  * @param status an in/out ICU UErrorCode
michael@0:  * @return an enumeration containing encoding names.
michael@0:  *         The returned encoding names and their order will be the same as
michael@0:  *         supplied when building the selector.
michael@0:  *
michael@0:  * @stable ICU 4.2
michael@0:  */
michael@0: U_STABLE UEnumeration * U_EXPORT2
michael@0: ucnvsel_selectForString(const UConverterSelector* sel,
michael@0:                         const UChar *s, int32_t length, UErrorCode *status);
michael@0: 
michael@0: /**
michael@0:  * Select converters that can map all characters in a UTF-8 string,
michael@0:  * ignoring the excluded code points.
michael@0:  *
michael@0:  * @param sel a selector
michael@0:  * @param s UTF-8 string
michael@0:  * @param length length of the string, or -1 if NUL-terminated
michael@0:  * @param status an in/out ICU UErrorCode
michael@0:  * @return an enumeration containing encoding names.
michael@0:  *         The returned encoding names and their order will be the same as
michael@0:  *         supplied when building the selector.
michael@0:  *
michael@0:  * @stable ICU 4.2
michael@0:  */
michael@0: U_STABLE UEnumeration * U_EXPORT2
michael@0: ucnvsel_selectForUTF8(const UConverterSelector* sel,
michael@0:                       const char *s, int32_t length, UErrorCode *status);
michael@0: 
michael@0: #endif  /* !UCONFIG_NO_CONVERSION */
michael@0: 
michael@0: #endif  /* __ICU_UCNV_SEL_H__ */