1 // © 2016 and later: Unicode, Inc. and others. 2 // License & terms of use: http://www.unicode.org/copyright.html 3 /* 4 ******************************************************************************* 5 * 6 * Copyright (C) 2008-2011, International Business Machines 7 * Corporation, Google and others. All Rights Reserved. 8 * 9 ******************************************************************************* 10 */ 11 /* 12 * Author : eldawy@google.com (Mohamed Eldawy) 13 * ucnvsel.h 14 * 15 * Purpose: To generate a list of encodings capable of handling 16 * a given Unicode text 17 * 18 * Started 09-April-2008 19 */ 20 21 #ifndef __ICU_UCNV_SEL_H__ 22 #define __ICU_UCNV_SEL_H__ 23 24 #include "unicode/utypes.h" 25 26 #if !UCONFIG_NO_CONVERSION 27 28 #include "unicode/uset.h" 29 #include "unicode/utf16.h" 30 #include "unicode/uenum.h" 31 #include "unicode/ucnv.h" 32 33 #if U_SHOW_CPLUSPLUS_API 34 #include "unicode/localpointer.h" 35 #endif // U_SHOW_CPLUSPLUS_API 36 37 /** 38 * \file 39 * 40 * A converter selector is built with a set of encoding/charset names 41 * and given an input string returns the set of names of the 42 * corresponding converters which can convert the string. 43 * 44 * A converter selector can be serialized into a buffer and reopened 45 * from the serialized form. 46 */ 47 48 struct UConverterSelector; 49 /** 50 * @{ 51 * Typedef for selector data structure. 52 */ 53 typedef struct UConverterSelector UConverterSelector; 54 /** @} */ 55 56 /** 57 * Open a selector. 58 * If converterListSize is 0, build for all available converters. 59 * If excludedCodePoints is NULL, don't exclude any code points. 60 * 61 * @param converterList a pointer to encoding names needed to be involved. 62 * Can be NULL if converterListSize==0. 63 * The list and the names will be cloned, and the caller 64 * retains ownership of the original. 65 * @param converterListSize number of encodings in above list. 66 * If 0, builds a selector for all available converters. 67 * @param excludedCodePoints a set of code points to be excluded from consideration. 68 * That is, excluded code points in a string do not change 69 * the selection result. (They might be handled by a callback.) 70 * Use NULL to exclude nothing. 71 * @param whichSet what converter set to use? Use this to determine whether 72 * to consider only roundtrip mappings or also fallbacks. 73 * @param status an in/out ICU UErrorCode 74 * @return the new selector 75 * 76 * @stable ICU 4.2 77 */ 78 U_CAPI UConverterSelector* U_EXPORT2 79 ucnvsel_open(const char* const* converterList, int32_t converterListSize, 80 const USet* excludedCodePoints, 81 const UConverterUnicodeSet whichSet, UErrorCode* status); 82 83 /** 84 * Closes a selector. 85 * If any Enumerations were returned by ucnv_select*, they become invalid. 86 * They can be closed before or after calling ucnv_closeSelector, 87 * but should never be used after the selector is closed. 88 * 89 * @see ucnv_selectForString 90 * @see ucnv_selectForUTF8 91 * 92 * @param sel selector to close 93 * 94 * @stable ICU 4.2 95 */ 96 U_CAPI void U_EXPORT2 97 ucnvsel_close(UConverterSelector *sel); 98 99 #if U_SHOW_CPLUSPLUS_API 100 101 U_NAMESPACE_BEGIN 102 103 /** 104 * \class LocalUConverterSelectorPointer 105 * "Smart pointer" class, closes a UConverterSelector via ucnvsel_close(). 106 * For most methods see the LocalPointerBase base class. 107 * 108 * @see LocalPointerBase 109 * @see LocalPointer 110 * @stable ICU 4.4 111 */ 112 U_DEFINE_LOCAL_OPEN_POINTER(LocalUConverterSelectorPointer, UConverterSelector, ucnvsel_close); 113 114 U_NAMESPACE_END 115 116 #endif 117 118 /** 119 * Open a selector from its serialized form. 120 * The buffer must remain valid and unchanged for the lifetime of the selector. 121 * This is much faster than creating a selector from scratch. 122 * Using a serialized form from a different machine (endianness/charset) is supported. 123 * 124 * @param buffer pointer to the serialized form of a converter selector; 125 * must be 32-bit-aligned 126 * @param length the capacity of this buffer (can be equal to or larger than 127 * the actual data length) 128 * @param status an in/out ICU UErrorCode 129 * @return the new selector 130 * 131 * @stable ICU 4.2 132 */ 133 U_CAPI UConverterSelector* U_EXPORT2 134 ucnvsel_openFromSerialized(const void* buffer, int32_t length, UErrorCode* status); 135 136 /** 137 * Serialize a selector into a linear buffer. 138 * The serialized form is portable to different machines. 139 * 140 * @param sel selector to consider 141 * @param buffer pointer to 32-bit-aligned memory to be filled with the 142 * serialized form of this converter selector 143 * @param bufferCapacity the capacity of this buffer 144 * @param status an in/out ICU UErrorCode 145 * @return the required buffer capacity to hold serialize data (even if the call fails 146 * with a U_BUFFER_OVERFLOW_ERROR, it will return the required capacity) 147 * 148 * @stable ICU 4.2 149 */ 150 U_CAPI int32_t U_EXPORT2 151 ucnvsel_serialize(const UConverterSelector* sel, 152 void* buffer, int32_t bufferCapacity, UErrorCode* status); 153 154 /** 155 * Select converters that can map all characters in a UTF-16 string, 156 * ignoring the excluded code points. 157 * 158 * @param sel a selector 159 * @param s UTF-16 string 160 * @param length length of the string, or -1 if NUL-terminated 161 * @param status an in/out ICU UErrorCode 162 * @return an enumeration containing encoding names. 163 * The returned encoding names and their order will be the same as 164 * supplied when building the selector. 165 * 166 * @stable ICU 4.2 167 */ 168 U_CAPI UEnumeration * U_EXPORT2 169 ucnvsel_selectForString(const UConverterSelector* sel, 170 const UChar *s, int32_t length, UErrorCode *status); 171 172 /** 173 * Select converters that can map all characters in a UTF-8 string, 174 * ignoring the excluded code points. 175 * 176 * @param sel a selector 177 * @param s UTF-8 string 178 * @param length length of the string, or -1 if NUL-terminated 179 * @param status an in/out ICU UErrorCode 180 * @return an enumeration containing encoding names. 181 * The returned encoding names and their order will be the same as 182 * supplied when building the selector. 183 * 184 * @stable ICU 4.2 185 */ 186 U_CAPI UEnumeration * U_EXPORT2 187 ucnvsel_selectForUTF8(const UConverterSelector* sel, 188 const char *s, int32_t length, UErrorCode *status); 189 190 #endif /* !UCONFIG_NO_CONVERSION */ 191 192 #endif /* __ICU_UCNV_SEL_H__ */ 193