1 /* 2 * Summary: interface for the encoding conversion functions 3 * Description: interface for the encoding conversion functions needed for 4 * XML basic encoding and iconv() support. 5 * 6 * Related specs are 7 * rfc2044 (UTF-8 and UTF-16) F. Yergeau Alis Technologies 8 * [ISO-10646] UTF-8 and UTF-16 in Annexes 9 * [ISO-8859-1] ISO Latin-1 characters codes. 10 * [UNICODE] The Unicode Consortium, "The Unicode Standard -- 11 * Worldwide Character Encoding -- Version 1.0", Addison- 12 * Wesley, Volume 1, 1991, Volume 2, 1992. UTF-8 is 13 * described in Unicode Technical Report #4. 14 * [US-ASCII] Coded Character Set--7-bit American Standard Code for 15 * Information Interchange, ANSI X3.4-1986. 16 * 17 * Copy: See Copyright for the status of this software. 18 * 19 * Author: Daniel Veillard 20 */ 21 22 #ifndef __XML_CHAR_ENCODING_H__ 23 #define __XML_CHAR_ENCODING_H__ 24 25 #include <libxml/xmlversion.h> 26 27 #ifdef LIBXML_ICONV_ENABLED 28 #include <iconv.h> 29 #endif 30 31 #ifdef __cplusplus 32 extern "C" { 33 #endif 34 35 /* 36 * xmlCharEncoding: 37 * 38 * Predefined values for some standard encodings. 39 * Libxml does not do beforehand translation on UTF8 and ISOLatinX. 40 * It also supports ASCII, ISO-8859-1, and UTF16 (LE and BE) by default. 41 * 42 * Anything else would have to be translated to UTF8 before being 43 * given to the parser itself. The BOM for UTF16 and the encoding 44 * declaration are looked at and a converter is looked for at that 45 * point. If not found the parser stops here as asked by the XML REC. A 46 * converter can be registered by the user using xmlRegisterCharEncodingHandler 47 * but the current form doesn't allow stateful transcoding (a serious 48 * problem agreed !). If iconv has been found it will be used 49 * automatically and allow stateful transcoding, the simplest is then 50 * to be sure to enable iconv and to provide iconv libs for the encoding 51 * support needed. 52 * 53 * Note that the generic "UTF-16" is not a predefined value. Instead, only 54 * the specific UTF-16LE and UTF-16BE are present. 55 */ 56 typedef enum { 57 XML_CHAR_ENCODING_ERROR= -1, /* No char encoding detected */ 58 XML_CHAR_ENCODING_NONE= 0, /* No char encoding detected */ 59 XML_CHAR_ENCODING_UTF8= 1, /* UTF-8 */ 60 XML_CHAR_ENCODING_UTF16LE= 2, /* UTF-16 little endian */ 61 XML_CHAR_ENCODING_UTF16BE= 3, /* UTF-16 big endian */ 62 XML_CHAR_ENCODING_UCS4LE= 4, /* UCS-4 little endian */ 63 XML_CHAR_ENCODING_UCS4BE= 5, /* UCS-4 big endian */ 64 XML_CHAR_ENCODING_EBCDIC= 6, /* EBCDIC uh! */ 65 XML_CHAR_ENCODING_UCS4_2143=7, /* UCS-4 unusual ordering */ 66 XML_CHAR_ENCODING_UCS4_3412=8, /* UCS-4 unusual ordering */ 67 XML_CHAR_ENCODING_UCS2= 9, /* UCS-2 */ 68 XML_CHAR_ENCODING_8859_1= 10,/* ISO-8859-1 ISO Latin 1 */ 69 XML_CHAR_ENCODING_8859_2= 11,/* ISO-8859-2 ISO Latin 2 */ 70 XML_CHAR_ENCODING_8859_3= 12,/* ISO-8859-3 */ 71 XML_CHAR_ENCODING_8859_4= 13,/* ISO-8859-4 */ 72 XML_CHAR_ENCODING_8859_5= 14,/* ISO-8859-5 */ 73 XML_CHAR_ENCODING_8859_6= 15,/* ISO-8859-6 */ 74 XML_CHAR_ENCODING_8859_7= 16,/* ISO-8859-7 */ 75 XML_CHAR_ENCODING_8859_8= 17,/* ISO-8859-8 */ 76 XML_CHAR_ENCODING_8859_9= 18,/* ISO-8859-9 */ 77 XML_CHAR_ENCODING_2022_JP= 19,/* ISO-2022-JP */ 78 XML_CHAR_ENCODING_SHIFT_JIS=20,/* Shift_JIS */ 79 XML_CHAR_ENCODING_EUC_JP= 21,/* EUC-JP */ 80 XML_CHAR_ENCODING_ASCII= 22 /* pure ASCII */ 81 } xmlCharEncoding; 82 83 /** 84 * xmlCharEncodingInputFunc: 85 * @out: a pointer to an array of bytes to store the UTF-8 result 86 * @outlen: the length of @out 87 * @in: a pointer to an array of chars in the original encoding 88 * @inlen: the length of @in 89 * 90 * Take a block of chars in the original encoding and try to convert 91 * it to an UTF-8 block of chars out. 92 * 93 * Returns the number of bytes written, -1 if lack of space, or -2 94 * if the transcoding failed. 95 * The value of @inlen after return is the number of octets consumed 96 * if the return value is positive, else unpredictiable. 97 * The value of @outlen after return is the number of octets consumed. 98 */ 99 typedef int (* xmlCharEncodingInputFunc)(unsigned char *out, int *outlen, 100 const unsigned char *in, int *inlen); 101 102 103 /** 104 * xmlCharEncodingOutputFunc: 105 * @out: a pointer to an array of bytes to store the result 106 * @outlen: the length of @out 107 * @in: a pointer to an array of UTF-8 chars 108 * @inlen: the length of @in 109 * 110 * Take a block of UTF-8 chars in and try to convert it to another 111 * encoding. 112 * Note: a first call designed to produce heading info is called with 113 * in = NULL. If stateful this should also initialize the encoder state. 114 * 115 * Returns the number of bytes written, -1 if lack of space, or -2 116 * if the transcoding failed. 117 * The value of @inlen after return is the number of octets consumed 118 * if the return value is positive, else unpredictiable. 119 * The value of @outlen after return is the number of octets produced. 120 */ 121 typedef int (* xmlCharEncodingOutputFunc)(unsigned char *out, int *outlen, 122 const unsigned char *in, int *inlen); 123 124 125 /* 126 * Block defining the handlers for non UTF-8 encodings. 127 * If iconv is supported, there are two extra fields. 128 */ 129 typedef struct _xmlCharEncodingHandler xmlCharEncodingHandler; 130 typedef xmlCharEncodingHandler *xmlCharEncodingHandlerPtr; 131 struct _xmlCharEncodingHandler { 132 char *name; 133 xmlCharEncodingInputFunc input; 134 xmlCharEncodingOutputFunc output; 135 #ifdef LIBXML_ICONV_ENABLED 136 iconv_t iconv_in; 137 iconv_t iconv_out; 138 #endif /* LIBXML_ICONV_ENABLED */ 139 #ifdef LIBXML_ICU_ENABLED 140 struct _uconv_t *uconv_in; 141 struct _uconv_t *uconv_out; 142 #endif /* LIBXML_ICU_ENABLED */ 143 }; 144 145 #ifdef __cplusplus 146 } 147 #endif 148 #include <libxml/tree.h> 149 #ifdef __cplusplus 150 extern "C" { 151 #endif 152 153 /* 154 * Interfaces for encoding handlers. 155 */ 156 XML_DEPRECATED 157 XMLPUBFUN void XMLCALL 158 xmlInitCharEncodingHandlers (void); 159 XML_DEPRECATED 160 XMLPUBFUN void XMLCALL 161 xmlCleanupCharEncodingHandlers (void); 162 XMLPUBFUN void XMLCALL 163 xmlRegisterCharEncodingHandler (xmlCharEncodingHandlerPtr handler); 164 XMLPUBFUN xmlCharEncodingHandlerPtr XMLCALL 165 xmlGetCharEncodingHandler (xmlCharEncoding enc); 166 XMLPUBFUN xmlCharEncodingHandlerPtr XMLCALL 167 xmlFindCharEncodingHandler (const char *name); 168 XMLPUBFUN xmlCharEncodingHandlerPtr XMLCALL 169 xmlNewCharEncodingHandler (const char *name, 170 xmlCharEncodingInputFunc input, 171 xmlCharEncodingOutputFunc output); 172 173 /* 174 * Interfaces for encoding names and aliases. 175 */ 176 XMLPUBFUN int XMLCALL 177 xmlAddEncodingAlias (const char *name, 178 const char *alias); 179 XMLPUBFUN int XMLCALL 180 xmlDelEncodingAlias (const char *alias); 181 XMLPUBFUN const char * XMLCALL 182 xmlGetEncodingAlias (const char *alias); 183 XMLPUBFUN void XMLCALL 184 xmlCleanupEncodingAliases (void); 185 XMLPUBFUN xmlCharEncoding XMLCALL 186 xmlParseCharEncoding (const char *name); 187 XMLPUBFUN const char * XMLCALL 188 xmlGetCharEncodingName (xmlCharEncoding enc); 189 190 /* 191 * Interfaces directly used by the parsers. 192 */ 193 XMLPUBFUN xmlCharEncoding XMLCALL 194 xmlDetectCharEncoding (const unsigned char *in, 195 int len); 196 197 XMLPUBFUN int XMLCALL 198 xmlCharEncOutFunc (xmlCharEncodingHandler *handler, 199 xmlBufferPtr out, 200 xmlBufferPtr in); 201 202 XMLPUBFUN int XMLCALL 203 xmlCharEncInFunc (xmlCharEncodingHandler *handler, 204 xmlBufferPtr out, 205 xmlBufferPtr in); 206 XMLPUBFUN int XMLCALL 207 xmlCharEncFirstLine (xmlCharEncodingHandler *handler, 208 xmlBufferPtr out, 209 xmlBufferPtr in); 210 XMLPUBFUN int XMLCALL 211 xmlCharEncCloseFunc (xmlCharEncodingHandler *handler); 212 213 /* 214 * Export a few useful functions 215 */ 216 #ifdef LIBXML_OUTPUT_ENABLED 217 XMLPUBFUN int XMLCALL 218 UTF8Toisolat1 (unsigned char *out, 219 int *outlen, 220 const unsigned char *in, 221 int *inlen); 222 #endif /* LIBXML_OUTPUT_ENABLED */ 223 XMLPUBFUN int XMLCALL 224 isolat1ToUTF8 (unsigned char *out, 225 int *outlen, 226 const unsigned char *in, 227 int *inlen); 228 #ifdef __cplusplus 229 } 230 #endif 231 232 #endif /* __XML_CHAR_ENCODING_H__ */ 233